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

2: %%

3: %% Modified 2005 December 5

4: %%

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

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

7: 

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

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

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

11: 

12: %% The command below calls the preprint style

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

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

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

16: %%

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

18: 

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

20: 

21: %%\documentclass[manuscript]{aastex}

22: %\usepackage{rotating}

23: %\usepackage{natbib}

24: %%\usepackage{url}

25: %\bibliographystyle{apj}

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

27: 

28: %% \documentclass[preprint2]{aastex}

29: 

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

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

32: %% use the longabstract style option.

33: 

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

35: 

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

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

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

39: %%

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

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

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

43: %% for information.

44: 

45: \newcommand{\vdag}{(v)^\dagger}

46: \newcommand{\myemail}{skywalker@galaxy.far.far.away}

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{Mid-IR interferometry of massive stars}

60: \shortauthors{Rajagopal 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{Mid-Infrared interferometry of dust around massive evolved stars \\

72:     }

73: 

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

75: %% author and affiliation information.

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

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

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

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

80: 

81: \author{Jayadev Rajagopal\altaffilmark{1}, Jean-Luc Menut\altaffilmark{2}, 

82: D. Wallace\altaffilmark{3}, W.C. Danchi\altaffilmark{3},

83: O. Chesneau\altaffilmark{4}, B. Lopez\altaffilmark{2},

84: J.D. Monnier\altaffilmark{5}, M. Ireland\altaffilmark{6}, P.G. Tuthill\altaffilmark{7}}

85: \email{jrajagopal@ctio.noao.edu}

86: 

87: 

88: 

89: %% Notice that each of these authors has alternate affiliations, which

90: %% are identified by the \altaffilmark after each name.  Specify alternate

91: %% affiliation information with \altaffiltext, with one command per each

92: %% affiliation.

93: 

94: \altaffiltext{1}{Michelson Fellow, University of Maryland, College Park MD 20742, \\

95: Present address: CTIO, Casilla 603, La Serena, Chile}

96: \altaffiltext{3}{NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA}

97: \altaffiltext{2}{Observatoire de la C\^ote d'Azur, D\'epartment G\'emini, UMR 6203, Nice, France}

98: \altaffiltext{4}{Observatoire de la C\^ote d'Azur, D\'epartment G\'emini, Avenue Copernic, Grasse, 

99: France}

100: \altaffiltext{5}{Department of Astronomy, University of Michigan, 501

101: East University Avenue, Ann Arbor, MI 4809, USA}

102: \altaffiltext{7}{School of Physics, Sydney University, NSW, 2006, Australia}

103: \altaffiltext{6}{Planetary Science MC 150-21, California Institute of Technology, Pasadena CA 91125}

104:     

105: 

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

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

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

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

110: %% editorial office after submission.

111: 

112: \begin{abstract}

113: We report long-baseline interferometric measurements of circumstellar 

114: dust around massive 

115: evolved stars with the MIDI instrument on the Very Large Telescope 

116: Interferometer and provide spectrally dispersed visibilities in the 8-13 \micron\ 

117: wavelength band. We also present diffraction-limited observations at

118: 10.7 \micron\ on the Keck Telescope with baselines up to 8.7 m which

119: explore larger scale structure. We have resolved the dust shells around 

120: the late type WC stars 

121: WR 106 and WR 95, and the enigmatic NaSt1 (formerly WR 122), suspected to   

122: have recently evolved from a Luminous 

123: Blue Variable (LBV) stage. For AG Car, 

124: the protoypical 

125: LBV in our sample, we marginally resolve structure close to the star, 

126: distinct from the well-studied detached nebula.  

127: The dust shells  around the two WC stars show fairly 

128: constant

129: size in the 8-13 \micron\ MIDI band, with gaussian half-widths of $\sim$ 25 to 

130: 40 mas, and the  Keck observations 

131: reveal an  additional extended structure around WR 106. 

132: The visibility profiles for NaSt1 obtained from two MIDI baselines 

133: indicate a compact source embedded in an extended structure. The compact

134: dust we detect around NaSt1 and AG Car favors recent or ongoing dust formation.

135: 

136: Using the measured

137: visibilities, we 

138: build spherically symmetric radiative transfer models of the WC

139: dust shells which enable detailed comparison with existing SED-based models. 

140: Our results indicate that the inner radii of the shells are

141: within a few tens of AU from the stars. 

142: In addition, our models favor grain size distributions with large

143: ($\sim$ 1 \micron) dust grains.

144: This proximity of the inner dust to the hot central star emphasises

145: the difficulty faced by current theories in forming dust in the hostile

146: environment around WR stars. 

147: Although we detect no direct evidence for binarity for these objects, dust 

148: production in a 

149: colliding-wind interface in a binary system is a feasible

150: mechanism in WR systems under these conditions. 

151: 

152: \end{abstract}

153: 

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

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

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

157: %% what keyword punctuation is appropriate.

158: 

159: \keywords{Interferometry: general --- Wolf-Rayet stars: individual(WR 95,

160: WR 106, NaSt1) --- Luminous Blue Variables: individual(AG Car)}

161: 

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

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

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

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

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

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

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

169: %% each reference.

170: 

171: 

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

173: %% linked in the electronic edition to a data center may do so by tagging

174: %% their objects with \objectname{} or \object{}.  Each macro takes the

175: %% object name as its required argument. The optional, square-bracket 

176: %% argument should be used in cases where the data center identification

177: %% differs from what is to be printed in the paper.  The text appearing 

178: %% in curly braces is what will appear in print in the published paper. 

179: %% If the object name is recognized by the data centers, it will be linked

180: %% in the electronic edition to the object data available at the data centers  

181: %%

182: %% Note that for sources with brackets in their names, e.g. [WEG2004] 14h-090,

183: %% the brackets must be escaped with backslashes when used in the first

184: %% square-bracket argument, for instance, \object[\[WEG2004\] 14h-090]{90}).

185: %%  Otherwise, LaTeX will issue an error. 

186: 

187: \section{Introduction}

188: The most massive stars ($\gtrsim 40 $M$_\sun$) will likely pass through a 

189: Luminous Blue Variable (LBV) and Wolf-Rayet (WR) phase before ending their lives

190: in a supernova explosion. LBVs and some WRs are known to show infrared 

191: excess associated with circumstellar dust. The LBV phase is short-lived and

192: unstable with irregular mass ejections of $10^{-5}$ to $10^{-4} $M$_\sun$ for

193: each outburst \citep{hum94}, resulting in circumstellar nebulae 

194: and

195: associated dust. The dust in this phase is consistent 

196: with grains of differing compositions ranging over metallic, silicate and 

197: carbon-rich \citep{voo99,cla03}.  The WR stage, characterised by massive 

198: winds (of the order $\dot{\rm M} = 10^{-5}\rm M_\sun {\rm yr}^{-1}$) is divided into two 

199: main phases \citep{smi68}, 

200: the WN phase with

201: strong spectral lines of N and CNO-cycle products, and the later WC phase 

202: with strong lines of C and other He-burning products.

203: Dust formation 

204: is associated with the WC phase, consistent with the over-abundance of C, and is

205: notably absent in the WN phase \citep{vdh01b}, although recent {\it Spitzer} observations

206: \citep{bar06} may be the first indications to the contrary. Long term and detailed 

207: infrared photometry of WC stars \citep[hereafter WvdHT]{will87},  has

208: led to considerable progress in understanding the nature of the dust and its 

209: formation. The favored mechanism for dust-formation in this stage

210: is through wind-wind collision in a WR-OB binary system \citep{uso91}. For

211: the episodic dust producers WR 140 and WR 137, photometric and spectrosocopic 

212: data \citep{will90,will01} had

213: already indicated binarity which has been confirmed by long-baseline 

214: interferometric observations for WR 140 \citep{mon04a}. 

215: High resolution images in the near-IR using aperture-

216: masking techniques with the 

217: Keck telescope provide direct

218: evidence for binarity in the case of the constantly dusty, late-type WC stars 

219: WR 98a, WR 104 \citep{mon99,tut99}, and more recently, the ``cocoon stars'' in the 

220: Quintuplet cluster \citep{tut06}. 

221: \Citet{mon07} present a strong case for binarity for all the dusty WRs in the Keck 

222: aperture-masking sample.  However, whether all 

223: dusty late type WCs are binaries is an open question and dust formation in 

224: single WR stars has not been

225: ruled out \citep{che00}.       

226: 

227: In this paper, we present and interpret interferometric mid infrared 

228: observations of the      

229: late type WC stars \object{WR 95} and \object{WR 106}, the suspected 

230: LBV-WR transition

231: star \object{NaSt1}, and the LBV \object{AG Car}. The main goals of this 

232: campaign were

233: to resolve the sizes of the dust shells and characterize the physical 

234: distribution of dust. All the objects were observed with the MIDI instrument

235: at the Very large Telescope Interferometer (VLTI) 

236: in the 8-13 \micron\ wavelength range, typically with a baseline of $\sim$ 47 m. 

237: NaSt1 and WR 106 were 

238: also observed at 10.7 \micron\ with a multi-aperture interferometric technique 

239: using individual

240: segments of the Keck-I telescope mirror to get complementary 

241: data from baselines up to the 10 m diameter of the Keck mirror. 

242: In Sections 2 and 3, we describe the observations, data reduction and 

243: present the results. 

244: Section 4 details the dust models we have constructed and the

245: comparison with existing models based on the spectral energy distribution.

246: Section 5 summarizes these results and presents the conclusions.

247: 

248: \section{Observations and Data Reduction}

249: \subsection{MIDI}

250: Most targets were observed over four nights in 2004 July, except for AG Car 

251: which was allocated one night in 2004 April (Table 1). NaSt1 was also 

252: observed on two nights in 2004 September. 

253: All observations were with the $\sim$ 47 m long UT2-UT3 baseline, except for the

254: 2004 September observations of NaSt1 which used the $\sim$ 89 m UT2-UT4 baseline.

255: The MIDI instrument 

256: \citep{lei03} on the VLTI combines the light from a given 

257: pair of 8.2 m telescopes (UTs)

258: and produces spectrally dispersed fringes in the 8-13 \micron\ band. A detailed

259: description of the observing technique and data reduction can be found in

260: \citet{lei04}. Here we limit ourselves to a brief summary.

261: 

262: The observing sequence starts with acquiring the target on each

263: individual telescope with the 8.7 \micron\ filter (1.4 \micron\ wide) and attaining 

264: maximal overlap of the two images, while chopping the secondary at

265: $\sim$ 2 Hz over 10\arcsec. 

266: The images are 

267: combined to

268: produce the two complementary interferograms and 

269: dispersed through a prism with a resolution of $\lambda/\delta\lambda \sim$ 30. 

270: The delay lines are scanned to find and track the fringes 

271: and fringe visibilities are recorded while stepping the delay through $\sim$

272: 10 wavelengths ($\lambda = $10 \micron) in steps of 2 \micron\ with a piezo-driven mirror.

273: The scan rate is tuned to

274: the timescale of instrumental and atmospheric fluctuations of delay and

275: a typical measurement records

276: $\sim$ 200 scans in about 5 minutes. After the fringe measurement, a photometric

277: measurement is carried out by blocking the light from one and then the other

278: telescope, while chopping the secondary mirror. No chopping is done during

279: the fringe-tracking scans. The pointing is maintained by the Coude guiding with

280: only tip-tilt correction; higher order AO was yet to be integrated into the system.

281: A sensitivity better than $\sim$ 1 Jy is achieved for this

282: mode of fringe measurement. Given that these observations were during the

283: first regular semester after MIDI commisioning, the accuracy that we achieve

284: for the visibility measurements is of the order of 10\% and is

285: mostly set by photometry, guiding (overlap of the two images) and 

286: fringe tracking errors which impose limits on the accuracy of calibration. 

287: We calculate the errors on the visibility as the standard deviation

288: over the multiple calibrations for each measurement and conservatively quote 

289: the higher of this number or 10\% of the visibility amplitude. 

290: The above is true for the {\it absolute} visibility amplitude errors; the

291: relative errors are less and the slope of the visibility in the 8 to

292: 13 \micron\ band is accurate to $\sim$3\%.

293:  

294: 

295: Standard calibrator stars of known diameter from the MIDI catalogue\footnote{  

296: \url[HREF]{http://www.eso.org/projects/vlti/instru/midi/midi\_calib.txt}} 

297: were observed

298: immediately before and after the target (Table 1). Given the relatively large systematic

299: errors inherent in the visibility determination, we have also used calibrators from 

300: other programs within a 

301: span of 2-5 hours (depending on atmospheric conditions).    

302: 

303: The visibility reduction method generally follows the now well-established algorithms

304: for optical/IR interferometry (e.g., \citet{for97} for fiber stellar interferometers: the

305: basic principles apply for non-fiber systems like MIDI) 

306: and was carried out using custom software

307: provided by the VLTI consortium written in the {\it IDL} language. The spectra

308: are extracted with a typical binning of 4 pixels, with each pixel being 

309: $\sim$ 0.05 \micron\ wide. 

310: The two complementary

311: outputs of the combiner are differenced to remove the incoherent flux including

312: overall variations in 

313: the contribution from the two telescopes. Next, the fringe power is estimated

314: from the Fourier transform and the visibility amplitude\footnote{The usual estimator

315: used in optical/infrared interferometry is the square of the visibility 

316: amplitude (V$^2$) since

317: this quantity can be corrected for detector noise bias. However the MIDI 

318: measurement error in the mode we used (``High-Sens'', with no simultaneous photometry) is dominated by calibration error, 

319: not detector noise.}, normalized by the photometric fluxes, is calculated.

320: The visibility is then calibrated using the instrument visibility measured 

321: on the calibrator stars. 

322: 

323: \subsection{Keck}

324: We carried out observations at 10.7 \micron\ with the Keck-I telescope 

325: on two stars in our sample, WR 106 and NaSt1, in 2005 May (Table 2). These observations

326: were carried out in a ``segment-tilting'' mode \citep{mon04b,wei06} to achieve close to diffraction

327: limited resolution. 

328: Imaging at 10 \micron\ with the full pupil has, in

329: practice, failed to reliably yield diffraction limited images. This

330: is in part attributable to the ``seeing spike" problem, where the

331: visibilities at shorter baselines are difficult to calibrate in the

332: presence of changes in the seeing between the target and calibrator.

333: The results shown here demonstrate the utility of this new method in

334: alleviating some of these problems. In this technique, 

335: custom software is used to  

336: reconfigure the segmented primary mirror so as to allow selected groups of

337: segments to focus on the detector (the Long-Wavelength Spectrograph camera). 

338: Each group (four in our case) of

339: 6 segments is chosen so as to form a non-redundant sparse-aperture 

340: array. The 1.8 m hexagonal segments are individually

341: phased up so that each group focuses on a separate spot on the detector. 

342: Thus in a given frame, 24 out of 36

343: segments are used in these four ``Fizeau arrays'' (the rest of the segments are

344: pointed away from the detector). This arrangement measures 15 baselines

345: ranging from about 2 m to 9 m and 20 closure phases. Each frame is of

346: the order of 10 ms and we co-add 9 frames, giving an

347: effective exposure of 90 ms. Chopping is done at a 5 Hz rate. Calibrators 

348: are observed before and

349: after the targets and each target is visited at least twice during

350: a night. The analysis of

351: the data closely follows the aperture-masking analysis described in 

352: detail in numerous publications \citep[e.g.,][]{tut00}.

353: These measurements complement

354: the long-baseline VLTI data by adding in visibilities at shorter baselines

355: up to 9 m. 

356: \section{Results}

357: \subsection{The WC stars}

358: The two WC type stars in our sample are WR 95 and WR 106. They are both

359: classified as late type (WC 9) and are known to be persistent dust producers

360: from IR excess detected over long-term photometry \citep{wil95}. 

361: The photometric data do not have any periodicity and neither do the spectra 

362: indicate any clear signs of binarity. 

363: We obtained two MIDI visibility measurements

364: on WR 95 on consecutive nights in July 2004. Though the projected baseline

365: and position angle differ slightly between these two measurements,

366: the visibility values were remarkably similar (Figure 1, broken lines). Though this might indicate a

367: small degree of asymmetry, the accuracy of these measurements are not high

368: enough to justify this interpretation. Instead we have chosen the average of 

369: the two sets to represent the spectrally dispersed visibility (Figure 1, solid line). 

370: As can be seen in Figure 1, we clearly resolve the WR 95 dust shell with 

371: visibility values ranging from $\sim$ 0.2 to $\sim$ 0.6 over the 

372: 8-13 \micron\ band. In the absence of detailed knowledge of the source geometry

373: we have assumed a gaussian profile and the resulting FWHM in milliarcseconds (versus wavelength)

374: is shown by the dotted line. WR 95 was not observed with the Keck segment-tilting method 

375: since it was below the sensitivity limit.  

376: 

377: Figure 2 shows the MIDI results for WR 106. Though

378: we obtained two measurements in July 2004 (see Table 1), poor seeing conditions

379: resulted in variations in fringe tracking and calibration between the two

380: observations. In the absence of a clear criterion to select one data set

381: over the other, we average the two and the resulting larger error bars are

382: shown in the figure. However, this target

383: is well-resolved and meaningful conclusions can be drawn even from the

384: less-than-optimum data. The visibility variation with wavelength is similar

385: to that for WR 95. 

386: 

387: WR 106 was resolved 

388: by the Keck single aperture (Table 2). This was surprising in that, given the visibility 

389: values measured by the VLTI, one would have expected an unresolved source 

390: at the comparatively short baselines offered by the Keck mirror.  

391: This indicates that there is an extended 

392: component in the dust distribution which was over-resolved and hence failed

393: to show up in the VLTI

394: observations. We present these results in  Section 4 where

395: we discuss some simple models.

396: 

397: In Table 3, we list the angular sizes (gaussian FWHM) measured with MIDI at 10.5 \micron\ 

398: and the corresponding

399: linear sizes of the objects, assuming 

400: distances from the literature (see table footnote). 

401: The parameters directly constrained by this

402: data set are the size of the dust shells and its variation with

403: wavelength. The radius at which the dust grains start to condense is a much-debated

404: and crucial topic in the theory of dust-formation around WR stars 

405: \citep{zub98,cro03}. The major difficulty has been that

406: the equilibrium temperatures at the radii predicted by the SED-based models

407: are often above the sublimation temperature for carbon dust. 

408: Dust formation in the shock interface of a colliding-wind system

409: has found some favor in this context. A recent summary of results from

410: a long-term

411: campaign of aperture-masking interferometry at the Keck \citep{mon07}

412: provides the near-IR sizes for a number of WR stars, including those in

413: our sample. The authors point to a tight correlation between the

414: the surface-brightness in the 1.5 - 3.0 \micron\ band and the sizes (gaussian

415: FWHM) of the objects in their sample to argue for a common underlying 

416: dust-production mechanism.

417: Our direct measurement of {\em the mid-IR} size-scale for these

418: objects is an

419: important addition to this debate. 

420: From our gaussian FWHM curves, it is clear that the

421: size of the emitting region is more or less invariant in the 8 - 13 \micron\ band. 

422: This could indicate that the emission is dominated by an inner

423: edge or ring of material. 

424: In fact the sizes for WR 95 at the shorter wavelengths (1.6 and 2.2 \micron) reported 

425: in \citet{mon07} are quite similar to our values. For WR 106, the mid-IR size

426: is larger, as expected from the presence of the extension seen in the Keck measurement.

427: In section 4, we present radiative transfer

428: models for the dust shells based on our measurements and derive 

429: inner edge radii 

430: which we compare with values from SED-based models. 

431: 

432: 

433: \subsection{NaSt1}

434: NaSt1 has a checkered history of classification. It was initially typed as

435: a late-type WN star, WR 122 \citep{vdh81}, but subsequently this classification

436: has been questioned. The

437: most recent study \citep{cro99} has resulted in this star not

438: being considered a Wolf-Rayet any longer. Crowther and Smith find the chemical

439: composition of the optical nebulosity sorrounding the star to be similar in

440: many respects to an LBV, like $\eta$ Carina, and conjecture that the obscured

441: central star could be an early type WN, presumably just leaving the LBV phase.

442: The uncertainty in classification is in part due to

443: the rarity of these very massive stars, especially one caught in transition.

444: As such our measurement of the extent of the close-in dust and its

445: variation with wavelength is of significance.

446: The source is unresolved by MIDI in single dish imaging mode and the flux measured within the $\sim$300 mas beam is similar to the the value quoted in Smith and Houck (2001, Figure  4 in their paper) from a bigger beam, indicating that all the  mid-IR emission 

447: originates from a compact object.  Another notable feature is the absence (as in the WCs) of any silicate

448: feature in the visibility profile (Figure 3) or the MIDI spectrum (Figure 4).

449: Although this does not rule

450: out the presence of silicate dust, it attests to the fact that there are no structures with silicate emission at the MIDI resolution scale. This is in agreement with the finding of \citet{cro99} that the nebula is highly enriched in CNO processed material.

451: 

452: NaSt1 was not resolved with the Keck single aperture. In our experience,

453: the segment-tilting method is capable of resolving sizes down to  

454: 40-50 mas at the 3-sigma level. At 10.7 \micron, this is approximately a factor of 6 better than 

455: the formal diffraction

456: limit (1.22 $\lambda$/D) for the 10 m aperture. A more rigorous analysis

457: for the resolution limit for aperture-masking observations at Keck 

458: (the near-infrared analogue of segment-tilting) estimates it to be 4 times better than the

459: diffraction limit \citep{mon07}. We estimate an upper limit of

460: 50 mas for NaSt1 from the Keck experiment.

461: 

462: NaSt1 was observed with MIDI at both the 

463: UT2-UT3 (projected baseline: 44.0 m, PA: 42\degr)

464: and the longer UT2-UT4 (projected baseline: 63.0 m, PA: 78\degr).

465: Figure 3 shows the visibility values and corresponding gaussian sizes for NaSt1.

466: The visibility varies with wavelength,

467: decreasing towards the red end

468: of the spectrum. This typically indicates cooler material further away

469: from the star and is in contrast to the behavior we see in the

470: WC stars. At least part of the dust around LBVs could be swept from the

471: ISM and/or formed in a cooler environment in the

472: episodic ejecta from the LBV phase and could lead to a less-defined inner edge and 

473: more spread in the dust temperature. 

474: The visibility on the PA 42\degr\ shorter baseline is very similar to that on the longer one, 

475: and if we intrepret this

476: directly as a size, may indicate an elongation in this direction. The optical image of NaSt1 \citep{cro99}

477: in the NII $\lambda$6583 line

478: shows an elliptical nebula with the major axis at a PA of 30\degr and an aspect ratio of 0.6, though the size

479: scale ($\sim$ 8 arcsec) is several orders of magnitude larger than the

480: milliarcsecond scales we measure.

481: 

482: However, it must be stressed that the visibility curves from the two different 

483: baselines are almost identical and the 

484: probability of obtaining the same profile for a spatially complex object (even if as 

485: simple as an ellipse) by pure chance is 

486: very low.

487: A possible alternative interpretation is that, following the 

488: suggestion of \citet{cro99}, the mid-IR 

489: source could be split into two components: a compact, unresolved part, with a correlated flux of 

490: about 5~Jy from 8 to 13 \micron\

491: that corresponds to the central source and perhaps some hot dust in its immediate surrounding, 

492: and a detached 

493: $\eta$ Car-like nebula, fully resolved by the interferometer. The extension of this nebula may be of the same order as 

494: that of the region contributing the [NeII] 12.8 \micron\ line which is seen in the single dish spectrum, but not in the 

495: correlated (i.e., arising in the unresolved core) flux (Figure 4). However, the upper limit 

496: on the size

497: of this component would be limited to $\sim$ 50-60 mas at most; anything bigger would 

498: have been resolved by the

499: Keck observations. Given the 10\% error on the MIDI visibilities, this could still be

500: over-resolved, though marginally, even at the shorter 44 m baseline. 

501: The size of the compact component is possibly less than 5 mas or so,

502: given that higher values would start to affect the relative slope of the visibilities

503: between the two baselines. 

504: Qualitatively, this hypothesis offers a natural explanation of the measurements: 

505: the central source is unresolved and the extension 

506: overresolved whatever the baseline length (44 m or 63 m). However, the

507: tight constraints on the extended component weaken the case. Clearly,

508: more observations are needed to clarify the picture.  

509: A very hot source (T$\sim$100,000 K) is required for the fitting the object's SED 

510: by \citet{cro99}, and the 5 Jy (1.5 $\times$ 10$^{-13}$ Wm$^{-2}\mu$m$^{-1}$) correlated flux we measure at 10 $\mu$m 

511: is not consistent with this. A constant 

512: correlated flux from 8 to 13$\mu$m (Figure  4) can hardly be attributed to a source at a single 

513: temperature (say, $\sim$ 2000 K) either. In order to 

514: decompose the SED into multiple sources at different temperatures it is 

515: important to study this object in the near-IR using the AMBER instrument 

516: at the VLTI to resolve the K band 

517: continuum and the very strong HeI 2.058 \micron\ line that may 

518: originate in the close vicinity of the 

519: star \citep[their Figure 6]{cro99}.

520: 

521: 

522: At a distance of 2 kpc, the maximum linear size from

523: the gaussian model is $\sim$ 40 AU and with a two-component model it is $\sim$ 100 AU (for comparison, 

524: the average extension of the optical nebula is $\sim$ 13600 AU at 2 kpc from \citeauthor{cro99}). If we take the same 

525: characteristic expansion velocity  as

526: \citeauthor{cro99} (15kms$^{-1}$) then the dynamical age of the dusty 

527: nebula at 100 AU from the source is about 30 yr. 

528: This shortens to a couple of years if instead we assume dust grains forming in the wind at a velocity of $\sim$ 100 -300 

529: kms$^{-1}$\citep{zub98,cro99}.  

530: Hence the dust that we resolve points to very recent or ongoing dust formation close to the star which has implications for its 

531: current poorly-known evolutionary state. 

532: 

533: \subsection{AG Car}

534: AG Car is one of the best studied LBVs and a prototype of its class. 

535: AG Car has been extensively observed from UV to radio wavelengths with photometric, spectroscopic, and polarimetric techniques 

536: over several decades \citep{hum89}, which locates it in the H-R diagram 

537: very close to the Humphreys-Davidson limit, an empirical luminosity boundary 

538: for evolved massive stars \citep{hum79}.

539: 

540: AG Car is surrounded by a bright extended bipolar nebula, 30" $\times$ 40" 

541: in size \citep{not92}, perhaps from a massive 

542: erruptive event 

543: 10$^4$ years ago \citep{lam01}. The kinematics of the bipolar nebula is consistent with a two-lobe polar ejection viewed nearly from 

544: the equator \citep{not92}. Spectroscopic analysis by \citet{gro06} 

545: supports the bi-polar geometry and shows the 50-100 R$_\sun$ supergiant 

546: central star to be rotating

547: close to its break-up velocity. 

548: The high rotational velocity after the huge LBV outburst may be understood if the original explosion expelled the mass and momentum 

549: preferentially toward the poles. The wind at smaller latitude may not have formed or may have launched material with velocities lower than 

550: the escape velocity which fell back onto the star. 

551: The nebula contains a large amount of dust 

552: \citep[$\sim$0.2 M$_\sun$]{voo00}, which implies a 

553: huge mass loss rate during the eruption (at least 8-15 M$_\sun$).

554: 

555: Our measurements are set in context by two studies in particular; \citet{voo00} and \citet{not02}.

556: The images in \citet{voo00} obtained with the 10 \micron\ TIMMI camera on the 3.6 m NTT at La Silla 

557: reveal a central peak and a detached nebula with a radius 

558: of $\sim$ 10\arcsec. Their SED analysis

559: reveals a complex situation: 

560: the dust is 

561: dominantly oxygen-rich but Polycyclic Aromatic Hydrocarbon (PAH) features are 

562: observed at 3.3, 7.7, 8.6 and 11.3 \micron\ revealing some amount of 

563: carbon dust. 

564: Moreover, a population of very large grains ($>$ 10\micron) is 

565: necessary to explain the flux levels at long wavelengths whereas a 

566: population of small warm grains is invoked to explain the flux between 5 and 20\micron.  

567: 

568: There is evidence for CO J = 1$\to$0 and 2$\to$1 emission associated with AG Car \citep{not02} and the authors conclude that the

569: unresolved CO 

570: emission must arise close to the star in a high density warm region and not in the detached nebulae resolved in optical 

571: emission-line images. They postulate that a warm and dense equatorial disk exists close to the star.

572: 

573: MIDI observations of AG Car were carried out with the UT2-UT3 baseline on two nights, with projected baselines of 42.0 m and 35.6 m and

574: PA of 41\degr\ and 69\degr\ respectively. The visibilites and corresponding gaussian sizes are

575: shown in Figure 5 and the spectrum in Figure 6. AG Car is marginally resolved. The measured size ($\sim$ 11 mas) is similar

576: for both observations and 

577: fairly uniform with wavelength. 

578: The \citet{not02} optical images show strongly enhanced H$\alpha$ 

579: brightness at PA$\sim$35-225\degr, and the morphology is similar 

580: at 12$\mu$m \citep{voo00}. This direction is also the waist of the nebula, the largest extension being seen perpendicular to it.

581: The MIDI baselines are therefore oriented close to an optimal direction to detect equatorial material and we are 

582: very likely resolving the equatorial disk, though we cannot presently distinguish between a disk and shell structure. 

583: This is the first 

584: direct evidence for dust in a compact structure. 

585: The size we detect (11 mas) is $\sim$ 3 orders of magnitude smaller than the upper limit (3'') to the size of the (unresolved) CO outflow in 

586: \citet{not02}. 

587: The corresponding linear size is of

588: the order of only 100 AU at a distance of 6 kpc; very close to the central star. Nevertheless, this detection adds  

589: weight to the \citet{not02} claim that the CO emission arises from molecular gas close to the star and is shielded by a disk structure. 

590: The presence of dust has implications also for the origin of this compact structure. \Citet{not02} favor a scenario where the

591: the star goes through a Red Super Giant (RSG) or RSG-like phase \citep{smi98,lam01} during which the molecular outflow is formed. 

592: Dust formation is also commonly observed during the RSG stage \citep{mas05,dan94} and fits in with this hypothesis, 

593: though the $\sim$ 100 AU scale size indicates very recent or current dust formation as well.

594: 

595: There are few published spectra of AG Car in the mid-IR and the MIDI spectrum is a valuable addition. 

596: For comparison we also

597: show in Figure 6 the spectrum from the Short Wavelength Spectrometer (SWS) on the Infrared Space Observatory (ISO) 

598: in the 8-13 \micron\ range \citep{slo03}, scaled to roughly match the flux level of the MIDI spectrum.

599: The ISO spectrum, with a much bigger aperture, exhibits a slightly different slope from 

600: the MIDI measurement, but the striking difference is that the MIDI spectrum does not show the 11.3\micron\ PAH feature, implying that 

601: the PAH originates from further out in the disk/shell as conjectured by \citeauthor{voo00}.   

602: 

603: \section{The WC Models}

604: The dust around late type WR stars has been closely studied over the last

605: two decades or so and presents a challenging problem pertaining to

606: its formation, geometry and evolution. There have been quite a few

607: attempts to model the dust in some detail. WvdHT 

608: model these as spherical shells based on extensive IR

609: spectroscopy. \citet{zub98} provides dust

610: models with emphasis on the physics of grain formation and growth 

611: taking into account

612: the grain dynamics in the stellar wind. \citet{har04} take advantage of

613: the high resolution near-IR images of WR 104 \citep{tut99}, which show a ``pinwheel'' 

614: of dust formed in a binary wind interface, to construct a 3D radiative transfer model for this

615: object. As mentioned earlier, a large percentage of the known dust-forming

616: late type WC stars are known to be binaries, WR 104 being a prime example. 

617: The formation of dust in

618: these stars is usually interpreted using the wind-wind collision model

619: of \citet{uso91}. None of the stars in our sample show definite evidence  

620: of being  binaries \citep{wil00}, though this remains a strong possibility. 

621: 

622: It has been difficult to explain dust 

623: formation in the hostile environment of the wind from single WR stars.

624: SED-based models typically have dust forming at distances from 

625: 300 up to a 1000 times the stellar radius \citep{zub98}. The 

626: consensus on the material of the dust, from the SED signature, is amorphous

627: carbon (WvdHT). However, at these

628: distances the electron temperatures in the winds of WRs are likely as high as

629: 10$^4$ K \citep{zub98}, implying that carbon is almost fully 

630: ionised. 

631: 

632: We present here 

633: radiative transfer models for the WC stars in our sample with an aim

634: to use the interferometric measurements to answer the crucial question as

635: formation radius of the inner edges of these dust shells.  

636: To this end, we use a 2D Monte Carlo radiative transfer code, MCTRANSF \citep{nic03},

637: to construct spherically

638: symmetric dust distributions consistent with the measured visibilities.  

639: 

640: The parameters of the models are the star diameter and effective temperature, the

641: internal and external radius of the envelope, its density profile and 

642: optical depth, the minimal and maximal radius and size distribution of the 

643: grains and the optical indices of

644: the dust. 

645: We start with the basic parameters for the

646: dust shells around WR 95 and WR 106 mostly taken from the comprehensive 

647: SED-based models of WvdHT (see Table 4). 

648: By propagating energy packets in the sampled envelope (circular grid), 

649: taking into account the absorption, emission and scattering characteristics 

650: of the dust, MCTRANSF computes the temperature of the envelope, the SED of 

651: the flux received by the observer directly from the star, as well as the flux 

652: emitted and 

653: scattered by the envelope. 

654: The code has been described in detail in \citet{nic03} (see

655: \citeauthor{pas04}, \citeyear{pas04}, for benchmarking tests). 

656: An example application in

657: modeling a circumstellar dust shell is found in \citet{woi05}.

658: 

659: 

660: We then explore the parameter

661: space around this model by primarily varying the location of the inner edge of

662: the disk to see what values are most consistent with both our measured 

663: visibilities and spectra. We also tried limited variations of the optical depth, 

664: grain size and

665: distribution to quantify their effect on the visibility profile and spectra.

666: The radius of the central star and the composition of the dust were not

667: changed in these trials. 

668: MCTRANSF does not explicitly incorporate the sublimation temperature of the 

669: grain as an input.  The dust temperature in the nucleation zone 

670: from existing observational and modeling constraints ranges from 1000 - 1500 K 

671: or so \citep{zub98, har04} and in the process of iteration we discard those

672: models which produce untenably high temperatures at the inner edge of the shell.

673: We stress that the modeling is not a ``fitting'' exercise. That does not

674: seem justified, given 

675: the limited visibility data and the fact that the actual geometry of the dust is

676: most likely more complex than simple spherical shell models. Our goal here

677: is to use the new measurements

678: to constrain the vitally important distance to the dust-forming

679: zone from the star. 

680: %The radiative transfer code we use has been 

681: %described in detail in Niccolini et al. (2003) (see 

682: %Pascucci et al 2004 for benchmarking tests). An example application in

683: %modeling a circumstellar dust shell are found in Woitke and Niccolini (2005).  

684:  

685: \subsection{WR 106}

686: Extensive studies of the SED of this well known dust-producing WC9 star exist

687: in the literature and long-term IR photometry is summarised in \citet{vdh01b}.

688: Infrared (J-Q bands) measurements are presented in

689: WvdHT.  More recent mid-IR ISO spectra are found in \citet{smi01}.     

690: Given the difficulty in standardising mid-IR photometry from

691: various sources and the fact that dusty WC stars can have differing

692: degrees of variability, we rely mostly on the 8-13 \micron\ 

693: spectra obtained with the MIDI instrument simultaneously with the

694: visibilities. \citet{lei04} have confirmed the reliability

695: of spectra from MIDI. We used standard reduction techniques as

696: they describe to extract the spectra; most of our interferometric 

697: calibrators were also chosen to be IR spectroscopic standards

698: \footnote{\url[HREF]{http://ssc.spitzer.caltech.edu/irs/calib/templ/cohen\_models/}} and 

699: as such

700: provide ideal templates to calibrate the spectra. 

701: De-reddening WR spectra has been non-trivial, primarily because these stars are

702: grouped towards the Galactic center with large columns of intervening

703: interstellar dust in addition to the circumstellar dust. The wide interstellar silicate feature centered 

704: around 9.7 \micron\ complicates the issue \citep{smi01}. To 

705: provide a uniform comparison, we have de-reddened the 8-13 \micron\

706: band with values of A$_\lambda$/A$_{\rm v}$ from WvdHT and the van der Hulst

707: extinction curve 15 for the shorter wavelengths. 

708: 

709: For WR 106, the model visibilities from our ``best'' model are shown in Figure 7, 

710: along with the Keck measurements as well as the MIDI visibility at 10.5 \micron.

711: 

712: The MIDI visibilities across the entire band (8-13 \micron) are compared

713: with the model in Figure 8. 

714: The closure phases measured by the Keck segment-tilting experiment did not

715: show any significant deviation from zero, signifying a lack of asymmetry at 

716: this resolution and sensitivity. We therefore plot the Keck visibilities 

717: treating the baselines as scalars (i.e., ignoring the orientation). 

718: Figure 9 shows the model SED along with our measurements and those from 

719: the literature\footnote{From catalogs available at the {\it VizieR} online database \citep{och00}. Optical fluxes: Homogeneous Means in 

720: the UBV System (Mermilliod, 1991); Hipparcos Input Catalogue, Version 2 (Turon et al., 1993); The USNO-B1.0 Catalog 

721: (Monet et al., 2003); IR fluxes: Catalog of Infrared Observations, Edition 5 (Gezari et al., 1999); Third release of DENIS (DENIS consortium, 2005).} 

722: spanning wavelengths from the optical to the far-infrared for comparison. 

723: The temperature profile of the dust in our model is shown in Figure 10. 

724: The parameters for the model are as below:

725: 

726: 

727: \subsubsection*{Central Star}

728: For the central star, we have adopted a blackbody spectrum at the

729: effective temperature listed below. The temperature of the

730: WR star has been assigned a range of values in the literature.

731: Most dust-formation models adopt a value between 18000 and 25000 K

732: (e.g., WvdHT). These values

733: have been criticized as being under-estimates from recent 

734: studies of the WR spectra \citep{cro03}. However,

735: we find in our models that higher temperatures (we tested up to 40000 K)

736: do not fit the 8-13 \micron\ spectra and also lead to  dust temperatures 

737: above the 1500 K or so sublimation temperature usually adopted for amorphous 

738: carbon grains.

739: This indeed has been

740: the reason that previous modeling attempts favour the lower values as 

741: well.

742: Although more sophisticated 

743: SEDs have been used for the stellar photosphere \citep[WdvHT]{har04}, 

744: a simple blackbody

745: is likely adequate for our purpose. The radius of the star is

746: is assumed to be 14.6R$_\sun$ (WdvHT) at a distance of 

747: 2.09 Kpc \citep{vdh01a}.

748: \begin{itemize}

749: \item[*]Temperature of central star (T$_{eff}$) = 19000 K

750: \item[*]Angular radius of central star (R$_*$) = 3.23$\times$10$^{-5}$ arcsecond

751: \end{itemize}

752: \subsubsection*{Dust Shell physical properties}

753: \begin{itemize}

754: \item[*]Composition: Amorphous carbon (optical constants from 

755: \citeauthor{rol91}, 1991)

756: \item[*]Grain size: Ranges from 0.5 to 0.6 \micron

757: \item[*]Optical depth ($\tau$) = 0.015 at 10 \micron

758: \end{itemize}

759: \subsubsection*{Dust Shell geometry}

760: \begin{itemize}

761: \item[*]Inner radius of dust shell (R$_{int}$) = 280 R$_*$(multiples of stellar radius)

762: \item[*]Outer radius of dust shell = 8400 R$_*$

763: \item[*]Density profile: Proportional to  r$^{-2}$, 

764: except in the region between

765: 1.0 R$_{int}$ and 14.0 R$_{int}$ where we introduce an over-dense

766: region with density proportional to 7$\times$r$^{-2}$. 

767: \end{itemize}

768: \subsection{Discussion}

769: Most of the starting parameters of the dust shell model were based on 

770: WvdHT.

771: They establish the composition to be amorphous C

772: at least in the case of the episodic dust-producer WR 140. The lack of 

773: silicate features in the mid-IR region also points to amorphous C.

774: The grain size and distribution for the shells are not well known. 

775: The \citet{zub98} grain-formation model predicts a fairly small maximum grain size($\sim$ 100 - 200{\AA}). 

776: \Citet{har04} model

777: the {\it inner part} of the WR 104 pinwheel nebula with small ($\sim$ 100{\AA}) grains, mainly to match the

778: 2-10 \micron\ flux.

779: However, there is evidence in the SED of WR nebulae pointing towards large grains of $\sim$ 0.1 to 1

780: \micron\ \citep{chi01,vee98} and \citet{mar02} find a characteristic size of 0.5 \micron\ for the

781: extended emission in WR 112. 

782: The model presented here also favors large grains with a quasi-uniform 

783: grain size ranging from

784: 0.5 to 0.6 \micron, with a standard interstellar distribution scaling 

785: as a$^{-3.5}$\citep{mat77}. 

786: A model with sizes ranging from 0.03 to 2.0 \micron 

787: also shows reasonable agreement with the data for similar values

788: of the other parameters. Despite

789: extensive trials we were unable to find a distribution of smaller grain sizes

790: which agreed with the data, in particular the flux at the shorter infrared wavelengths. 

791: %We point out that the visibility profiles 

792: %generated by our models 

793: %are fairly insensitive to the grain size distribution which is largely

794: %constrained by the near and mid-IR spectra. 

795: 

796: The optical depth for these dust shells in the mid IR is fairly low 

797: \citep{zub98}.

798: We use here a value of 0.015. 

799: Our primary goal of matching the measured visibilities is not strongly

800: affected by the optical depth and it is the spectra that constrain this

801: parameter. 

802: 

803: The important constraint imposed by the visibility profile 

804: is on the range of {\em inner edge radii} of the dust shell. The ``best''

805: model in Figure 7 has a value of 280 R$_*$ for the inner edge radius, which 

806: is the same as in the WvdHT

807: model. We show in Figure 8 the model visibility profile

808: for  values ranging from 230 R$_*$ to 370 R$_*$.

809: Even for this narrow range, the 

810: visibility profiles are clearly different, 

811: indicating the fairly strong constraints imposed by the measurements.

812:     

813: The over-density (denser than the r$^{-2}$ profile) that we introduce in 

814: the shell is to accomodate the 

815: visibilities measured with the smaller baselines at the Keck,

816: which are lower than expected for a r$^{-2}$ profile consistent with the

817: longer baseline MIDI visibilities. The over-dense

818: region in the dust shell extends from the inner edge (R$_{int}$) up to 14 R$_{int}$ in 

819: the model,

820: i.e., size scales from 9 to 125 mas. This enhancement at the spatial

821: scales probed by the segment-tilting experiment (resolution $\sim$ 50 mas, see Section 3.2) 

822: decreases the visibility at the low spatial frequencies.

823: This of course is not a unique solution, but offers an

824: intuitive physical model.

825: The over-density could indicate an 

826: episode of more vigorous dust production. It could also be an artifact of

827: applying an overly simplistic spherically symmetric model to a possibly 

828: complex shell geometry as in, for example, a ``pin-wheel'' structure. 

829: In the absence

830: of more complete visibility sampling, we limit ourselves to simply

831: pointing out the existence of an extended structure in the dust around WR 106.      

832: 

833: \subsection {WR 95}

834: WR 95 is also a late-type dust-forming WC, similar to 

835: WR 106. As in the case of WR 106, there is no definite evidence so far,

836: either spectroscopic or photometric, of binarity. WR 95 falls below

837: the sensitivity limit of the Keck segment-tilting experiment and so

838: we base our models solely on the MIDI visibilities and spectra. Our best

839: model visibilities along with the measured values in the 8-13 \micron\ MIDI

840: band are shown in Figure 11.

841: Figure 12 shows the measured 8-13 \micron\  SED, the near-IR and far-IR photometry from the literature, and the model SED profile.

842: The dust temperature profile 

843: is as in Figure 13. 

844: The parameters of the model are as below:

845: 

846: \subsubsection*{Central Star}

847: \begin {itemize}  

848: \item[*] Effective temperature (T$_{eff}$): 25000 K

849: \item[*] Angular radius of star (R$_*$): 3.2$\times$10$^{-5}$ arcsecond

850: \end{itemize}

851: \subsubsection*{Dust shell physical properties}

852: \begin{itemize}

853: \item[*]Composition: Amorphous carbon

854: \item[*]Grain size: Ranges from 0.5 to 0.6 \micron

855: \item[*]Optical depth ($\tau$) = 0.005 at 10 \micron

856: \end{itemize}

857: \subsubsection*{Dust Shell geometry}

858: \begin{itemize}

859: \item[*]Inner radius of dust shell (R$_{int}$) = 480 R$_*$(multiples of stellar radius)

860: \item[*]Outer radius of dust shell = 1230 R$_*$

861: \item[*]Density profile: Proportional to  r$^{-2}$ 

862: \end{itemize}

863: As in the case of WR 106, we base our initial model parameters mostly on 

864: WvdHT. 

865: To match the near IR photometry and 

866: our visibility measurements we had to adopt higher

867: temperature of 25000 K for the WR star. The optical depth

868: for our best model, 0.005 at 10 \micron, is lower than that for WR 106. 

869: Since we do not have shorter baseline Keck

870: observations in this case to detect extended structure, a uniform r$^{-2}$

871: density profile suffices. The nominal inner radius of the shell is 480 R$_*$ (Figure 11 shows how the model visibilities

872: vary for slightly different values of this parameter), compared

873: to 410 R$_*$ derived by WvdHT. As in the case of WR 106, a dust grain distribution with sizes ranging from 0.3 to 2.0 \micron\ also yields a fair fit 

874: to the measurements.

875: \section{Conclusions}

876: High-resolution measurements of dust around massive stars are sparse in the

877: literature. Coupled with the rarity of the objects themselves, it is not

878: surprising that so much remains unclear about the mechanisms of dust-formation

879: and the nature of the dust itself. Only a handful of WR stars known

880: to produce dust have been imaged with any degree of detail. Three of these,

881: WR 104, WR 98a and WR 140 are binaries, producing dust at the wind-wind 

882: interface. The fourth, WR 112

883: reveals a multiple-shell dust structure, though at much 

884: larger (arcsecond) scales than discussed here for our targets, 

885: suggestive also of being a binary \citep{mar02}. Hence the measurements presented here, albeit

886: with limited baselines, of

887: similar objects with no previous evidence of binarity, 

888: are an important step in probing the nature of these

889: dust shells. 

890: 

891: We observed two WC9 stars (WR 106 and WR 95), an LBV (AG Car) and an object which may be 

892: in a stage of transition from an LBV to a WR phase, NaSt1. 

893: NaSt1 was well resolved and the

894: visibility measurements suggest increasing size with wavelength, a behavior 

895: unlike the WC 9 stars in our sample. We favor a

896: two-component model for the dust with a compact, unresolved

897: core and an extended component. Previous studies of the SED of this object have

898: also indicated this structure. For AG Car as well, our measurements resolve a compact

899: mid-IR source which is likely associated with a known molecular

900: gas component and is the first direct evidence for dust near the

901: central star. For both these objects, this study has revealed the presence

902: of close-in dust that has important implications for the state of evolution

903: of the central star.    

904: 

905: For the WC stars we construct spherically symmetric radiative transfer models

906: for the dust shells.

907: These models indicate the inner edge of these shells are at a few hundred

908: stellar radii (tens of AU) from the star.

909: Gaussian sizes (FWHM) directly estimated from the visibility 

910: measurement provide  model-independent upper limits which are consistent

911: with these values. 

912: At these distances, the dust temperature is fairly close to sublimation

913: values for carbonaceous dust. In addition, our results support previous

914: indications of fairly large grains ($\sim$ 1 \micron) despite these conditions.

915: These new results re-open an existing controversy: How does dust form in the 

916: hostile environment of a WR stellar wind at such high temperatures?  

917: The proximity of the inner edge to the star also forces us to adopt values

918: close to 20,000 K for the central star to keep the dust below the 

919: sublimation point. However, the spectra of late-type WC stars point to

920: higher (40,000 to 70,000 K) temperatures, rendering the situation even worse. 

921: Our

922: results certainly indicate that dust formation is likely occuring

923: under fairly extreme conditions. This may suggest a colliding-wind binary

924: model but direct evidence is still lacking for the objects in our sample. 

925: 

926: The dust around WR 106 seems to deviate from an r$^{-2}$ profile at

927: the shorter baseline Keck measurements. Our simple models suggest an over-dense

928: region fairly close to the inner edge. This could be also be a first indication of

929: of binarity with multiple shells from  episodic dust production or a more

930: complex geometry as in a pinwheel. But given

931: our limited baseline coverage, a caveat on modeling artifacts is warranted.

932: 

933: We have not attempted to estimate the dust mass from the models. Our 

934: measurements

935: are insensitive to the colder dust further from the star and it is unlikely

936: that we can improve on the existing dust mass estimates for these stars

937: from spectral profiles.

938: 

939: In the long-term, further progress would involve multiple baselines to image the dust 

940: structure. Observations of the outer dust at longer wavelengths are also required

941: to accurately estimate the dust mass and geometry. In the short-term, we aim to

942: look for asymmetries in the dust distribution by using the UT3-UT4

943: baseline at the VLTI, orthogonal to the one used here. 

944: 

945: 

946: % In a manner similar to \objectname authors can provide links to dataset

947: %% hosted at participating data centers via the \dataset{} command.  The

948: %% second curly bracket argument is printed in the text while the first

949: %% parentheses argument serves as the valid data set identifier.  Large

950: %% lists of data set are best provided in a table (see Table 3 for an example).

951: %% Valid data set identifiers should be obtained from the data center that

952: %% is currently hosting the data.

953: %%

954: %% Note that AASTeX interprets everything between the curly braces in the 

955: %% macro as regular text, so any special characters, e.g. "#" or "_," must be 

956: %% preceded by a backslash. Otherwise, you will get a LaTeX error when you 

957: %% compile your manuscript.  Special characters do not 

958: %% need to be escaped in the optional, square-bracket argument.

959: 

960: 

961: %% In this section, we use  the \subsection command to set off

962: %% a subsection.  \footnote is used to insert a footnote to the text.

963: 

964: %% Observe the use of the LaTeX \label

965: %% command after the \subsection to give a symbolic KEY to the

966: %% subsection for cross-referencing in a \ref command.

967: %% You can use LaTeX's \ref and \label commands to keep track of

968: %% cross-references to sections, equations, tables, and figures.

969: %% That way, if you change the order of any elements, LaTeX will

970: %% automatically renumber them.

971: 

972: %% This section also includes several of the displayed math environments

973: %% mentioned in the Author Guide.

974: 

975: 

976: \acknowledgments

977: 

978: %%We are grateful to V. Barger, T. Han, and R. J. N. Phillips for

979: %%doing the math in section~\ref{bozomath}.

980: %%More information on the AASTeX macros package is available \\ at

981: %%\url{http://www.aas.org/publications/aastex}.

982: %%For technical support, please write to

983: %%\email{aastex-help@aas.org}.

984: This work was performed in part under contract with the Jet Propulsion Laboratory (JPL)

985: funded by NASA through the Michelson Fellowship Program. JPL is managed for the

986: National Aeronautics and Space Administration (NASA) 

987: by the California Institute of Technology.

988: Some of the data presented herein were obtained at the W.M. Keck Observatory, 

989: which is operated as a scientific partnership among the California Institute of Technology, 

990: the University of California and NASA. 

991: The Observatory was made possible by the generous financial support of the W.M. Keck Foundation.

992: The authors wish to recognize and acknowledge the very significant cultural role and reverence 

993: that the summit of Mauna Kea has always had within the indigenous Hawaiian community. 

994: We are most fortunate to have the opportunity to conduct observations from this mountain.

995: We acknowledge the VLTI consortium for both technical and software support.  

996: 

997: We thank the referee for

998: comments which have significantly improved this article.

999: 

1000: %% To help institutions obtain information on the effectiveness of their

1001: %% telescopes, the AAS Journals has created a group of keywords for telescope

1002: %% facilities. A common set of keywords will make these types of searches

1003: %% significantly easier and more accurate. In addition, they will also be

1004: %% useful in linking papers together which utilize the same telescopes

1005: %% within the framework of the National Virtual Observatory.

1006: %% See the AASTeX Web site at http://www.journals.uchicago.edu/AAS/AASTeX

1007: %% for information on obtaining the facility keywords.

1008: 

1009: %% After the acknowledgments section, use the following syntax and the

1010: %% \facility{} macro to list the keywords of facilities used in the research

1011: %% for the paper.  Each keyword will be checked against the master list during

1012: %% copy editing.  Individual instruments or configurations can be provided 

1013: %% in parentheses, after the keyword, but they will not be verified.

1014: 

1015: {\it Facilities:} \facility{Keck}, \facility{VLTI (MIDI)} 

1016: 

1017: %% Appendix material should be preceded with a single \appendix command.

1018: %% There should be a \section command for each appendix. Mark appendix

1019: %% subsections with the same markup you use in the main body of the paper.

1020: 

1021: %% Each Appendix (indicated with \section) will be lettered A, B, C, etc.

1022: %% The equation counter will reset when it encounters the \appendix

1023: %% command and will number appendix equations (A1), (A2), etc.

1024: 

1025: 

1026: %% The reference list follows the main body and any appendices.

1027: %% Use LaTeX's thebibliography environment to mark up your reference list.

1028: %% Note \begin{thebibliography} is followed by an empty set of

1029: %% curly braces.  If you forget this, LaTeX will generate the error

1030: %% "Perhaps a missing \item?".

1031: %%

1032: %% thebibliography produces citations in the text using \bibitem-\cite

1033: %% cross-referencing. Each reference is preceded by a

1034: %% \bibitem command that defines in curly braces the KEY that corresponds

1035: %% to the KEY in the \cite commands (see the first section above).

1036: %% Make sure that you provide a unique KEY for every \bibitem or else the

1037: %% paper will not LaTeX. The square brackets should contain

1038: %% the citation text that LaTeX will insert in

1039: %% place of the \cite commands.

1040: 

1041: %% We have used macros to produce journal name abbreviations.

1042: %% AASTeX provides a number of these for the more frequently-cited journals.

1043: %% See the Author Guide for a list of them.

1044: 

1045: %% Note that the style of the \bibitem labels (in []) is slightly

1046: %% different from previous examples.  The natbib system solves a host

1047: %% of citation expression problems, but it is necessary to clearly

1048: %% delimit the year from the author name used in the citation.

1049: %% See the natbib documentation for more details and options.

1050: %%\bibliography{rajagopal_ref}

1051: %\input{rajagopal_ref.tex}

1052: 

1053: \begin{thebibliography}{}

1054: \bibitem[Barniske et al.(2006)]{bar06} Barniske, A., 

1055: Oskinova, L., Hamann, W.-R., \& Gr{\"a}fener, G.\ 2006, ASP Conf.~Ser.~353: 

1056: Stellar Evolution at Low Metallicity: Mass Loss, Explosions, Cosmology, 

1057: 353, 241

1058: \bibitem[Cherchneff et al.(2000)]{che00} Cherchneff, I., Le 

1059: Teuff, Y.~H., Williams, P.~M., \& Tielens, A.~G.~G.~M.\ 2000, \aap, 357, 

1060: 572 

1061: \bibitem[Chiar \& Tielens(2001)]{chi01} Chiar, J.~E., \& 

1062: Tielens, A.~G.~G.~M.\ 2001, \apjl, 550, L207

1063: \bibitem[Clark et al.(2003)]{cla03} Clark, J.~S., Egan, 

1064: M.~P., Crowther, P.~A., Mizuno, D.~R., Larionov, V.~M., \& Arkharov, A.\ 

1065: 2003, \aap, 412, 185 

1066: \bibitem[Coude Du Foresto et al.(1997)]{for97} Coude Du 

1067: Foresto, V., Ridgway, S., \& Mariotti, J.-M.\ 1997, \aaps, 121, 379 

1068: \bibitem[Crowther \& Smith(1999)]{cro99} Crowther, P.~A., \& 

1069: Smith, L.~J.\ 1999, \mnras, 308, 82 

1070: \bibitem[Crowther(2003)]{cro03} Crowther, P.~A.\ 2003, \apss, 

1071: 285, 677

1072: \bibitem[Danchi et al.(1994)]{dan94} Danchi, W.~C., Bester, 

1073: M., Degiacomi, C.~G., Greenhill, L.~J., \& Townes, C.~H.\ 1994, \aj, 107, 

1074: 1469 

1075: \bibitem[Groh et al.(2006)]{gro06} Groh, J.~H., Hillier, 

1076: D.~J., \& Damineli, A.\ 2006, \apjl, 638, L33

1077: \bibitem[Harries et al.(2004)]{har04} Harries, T.~J., 

1078: Monnier, J.~D., Symington, N.~H., \& Kurosawa, R.\ 2004, \mnras, 350, 565 

1079: \bibitem[Humphreys \& Davidson(1979)]{hum79} Humphreys, 

1080: R.~M., \& Davidson, K.\ 1979, \apj, 232, 409 

1081: \bibitem[Humphreys et al.(1989)]{hum89} Humphreys, R.~M., 

1082: Lamers, H.~J.~G.~L.~M., Hoekzema, N., \& Cassatella, A.\ 1989, \aap, 218, 

1083: L17 

1084: \bibitem[Humphreys \& Davidson(1994)]{hum94} Humphreys, 

1085: R.~M., \& Davidson, K.\ 1994, \pasp, 106, 1025 

1086: \bibitem[Lamers et al.(2001)]{lam01} Lamers, H.~J.~G.~L.~M., 

1087: Nota, A., Panagia, N., Smith, L.~J., \& Langer, N.\ 2001, \apj, 551, 764 

1088: \bibitem[Leinert et al.(2003)]{lei03} Leinert, C., et al.\ 

1089: 2003, \procspie, 4838, 893

1090: \bibitem[Leinert et al.(2004)]{lei04} Leinert, C., et al.\ 

1091: 2004, \aap, 423, 537

1092: \bibitem[Marchenko et al.(2002)]{mar02} Marchenko, S.~V., 

1093: Moffat, A.~F.~J., Vacca, W.~D., C{\^o}t{\'e}, S., \& Doyon, R.\ 2002, 

1094: \apjl, 565, L59

1095: \bibitem[Massey et al.(2005)]{mas05} Massey, P., Plez, B., 

1096: Levesque, E.~M., Olsen, K.~A.~G., Clayton, G.~C., \& Josselin, E.\ 2005, 

1097: \apj, 634, 1286M

1098: \bibitem[Mathis et al.(1977)]{mat77} Mathis, J.~S., Rumpl, 

1099: W., \& Nordsieck, K.~H.\ 1977, \apj, 217, 425 

1100: \bibitem[Monnier et al.(1999)]{mon99} Monnier, J.~D., 

1101: Tuthill, P.~G., \& Danchi, W.~C.\ 1999, \apjl, 525, L97

1102: \bibitem[Monnier et al.(2004a)]{mon04a} Monnier, J.~D., et al.\ 

1103: 2004a, \apjl, 602, L57

1104: \bibitem[Monnier et al.(2004b)]{mon04b} Monnier, J.~D., 

1105: Tuthill, P.~G., Ireland, M.~J., Cohen, R., \& Tannirkulam, A.\ 2004b, 

1106: Bulletin of the American Astronomical Society, 36, 1367 

1107: \bibitem[Monnier et al.(2007)]{mon07} Monnier, J.~D., 

1108: Tuthill, P.~G., Danchi, W.~C., Murphy, N., \& Harries, T.~J.\ 2007, \apj, 

1109: 655, 1033

1110: \bibitem[Niccolini et al.(2003)]{nic03} Niccolini, G., 

1111: Woitke, P., \& Lopez, B.\ 2003, \aap, 399, 703

1112: \bibitem[Nota et al.(1992)]{not92} Nota, A., Leitherer, C., 

1113: Clampin, M., Greenfield, P., \& Golimowski, D.~A.\ 1992, \apj, 398, 621

1114: \bibitem[Nota et al.(2002)]{not02} Nota, A., Pasquali, A., 

1115: Marston, A.~P., Lamers, H.~J.~G.~L.~M., Clampin, M., \& Schulte-Ladbeck, 

1116: R.~E.\ 2002, \aj, 124, 2920

1117: \bibitem[Ochsenbein et al.(2000)]{och00} Ochsenbein, F., 

1118: Bauer, P., \& Marcout, J.\ 2000, \aaps, 143, 23 

1119: \bibitem[Pascucci et al.(2004)]{pas04} Pascucci, I., Wolf, 

1120: S., Steinacker, J., Dullemond, C.~P., Henning, T., Niccolini, G., Woitke, 

1121: P., \& Lopez, B.\ 2004, \aap, 417, 793

1122: \bibitem[Rouleau \& Martin(1991)]{rol91} Rouleau, F., \& 

1123: Martin, P.~G.\ 1991, \apj, 377, 526 

1124: \bibitem[Sloan et al.(2003)]{slo03} Sloan, G.~C., Kraemer, 

1125: K.~E., Price, S.~D., \& Shipman, R.~F.\ 2003, \apjs, 147, 379

1126: \bibitem[Smith(1968)]{smi68} Smith, L.~F.\ 1968, \mnras, 138, 

1127: 109

1128: \bibitem[Smith et al.(1998)]{smi98} Smith, L.~J., Nota, A., 

1129: Pasquali, A., Leitherer, C., Clampin, M., \& Crowther, P.~A.\ 1998, \apj, 

1130: 503, 278

1131: \bibitem[Smith \& Houck(2001)]{smi01} Smith, J.~D.~T., \& 

1132: Houck, J.~R.\ 2001, \aj, 121, 2115 

1133: %\bibitem[Tannirkulam et al.(2005)]{tan05} Tannirkulam, A., 

1134: %Monnier, J., Ireland, M., Tuthill, P., Harries, T., \& Cohen, R.\ 2005, 

1135: %Protostars and Planets V, 8383

1136: \bibitem[Tuthill et al.(1999)]{tut99} Tuthill, P.~G., 

1137: Monnier, J.~D., \& Danchi, W.~C.\ 1999, \nat, 398, 487

1138: \bibitem[Tuthill et al.(2000)]{tut00} Tuthill, P.~G., 

1139: Monnier, J.~D., Danchi, W.~C., Wishnow, E.~H., \& Haniff, C.~A.\ 2000, 

1140: \pasp, 112, 555 

1141: \bibitem[Tuthill et al.(2006)]{tut06} Tuthill, P., Monnier, 

1142: J., Tanner, A., Figer, D., Ghez, A., \& Danchi, W.\ 2006, Science, 313, 935

1143: \bibitem[Usov(1991)]{uso91} Usov, V.~V.\ 1991, \mnras, 252, 

1144: 49

1145: \bibitem[van der Hucht et al.(1981)]{vdh81} van der Hucht, 

1146: K.~A., Conti, P.~S., Lundstrom, I., \& Stenholm, B.\ 1981, Space Science 

1147: Reviews, 28, 227 

1148: \bibitem[van der Hucht(2001)]{vdh01a} van der Hucht, K.~A.\ 

1149: 2001, New Astronomy Review, 45, 135

1150: \bibitem[van der Hucht et al.(2001)]{vdh01b} van der Hucht, 

1151: K.~A., Williams, P.~M., \& Morris, P.~W.\ 2001, ESA SP-460: The Promise of 

1152: the Herschel Space Observatory, 273

1153: \bibitem[Veen et al.(1998)]{vee98} Veen, P.~M., van Genderen, 

1154: A.~M., van der Hucht, K.~A., Li, A., Sterken, C., \& Dominik, C.\ 1998, 

1155: \aap, 329, 199 

1156: \bibitem[Voors et al.(1999)]{voo99} Voors, R.~H.~M., Waters, 

1157: L.~B.~F.~M., Morris, P.~W., Trams, N.~R., de Koter, A., \& Bouwman, J.\ 

1158: 1999, \aap, 341, L67 

1159: \bibitem[Voors et al.(2000)]{voo00} Voors, R.~H.~M., et al.\ 

1160: 2000, \aap, 356, 501 

1161: \bibitem[Weiner et al.(2006)]{wei06} Weiner, J., et al.\ 

1162: 2006, \apj, 636, 1067

1163: \bibitem[Williams et al.(1987)]{will87} Williams, P.~M., van 

1164: der Hucht, K.~A., \& The, P.~S.\ 1987, \aap, 182, 91

1165: \bibitem[Williams et al.(1990)]{will90} Williams, P.~M., van 

1166: der Hucht, K.~A., Pollock, A.~M.~T., Florkowski, D.~R., van der Woerd, H., 

1167: \& Wamsteker, W.~M.\ 1990, \mnras, 243, 662

1168: \bibitem[Williams(1995)]{wil95} Williams, P.~M.\ 1995, IAU 

1169: Symposium, 163, 335

1170: \bibitem[Williams \& van der Hucht(2000)]{wil00} Williams, 

1171: P.~M., \& van der Hucht, K.~A.\ 2000, \mnras, 314, 23

1172: \bibitem[Williams et al.(2001)]{will01} Williams, P.~M., et 

1173: al.\ 2001, \mnras, 324, 156

1174: \bibitem[Woitke \& Niccolini(2005)]{woi05} Woitke, P., \& 

1175: Niccolini, G.\ 2005, \aap, 433, 1101 

1176: \bibitem[Zubko(1998)]{zub98} Zubko, V.~G.\ 1998, \mnras, 295, 

1177: 109

1178: \end{thebibliography}

1179: 

1180: 

1181: %%\begin{thebibliography}{}

1182: 

1183: \clearpage

1184: \thispagestyle{empty}

1185: {\rotate

1186: \begin{table}

1187: \tabletypesize{\scriptsize}

1188: \begin{center}

1189: \caption{MIDI Observations \label{tbl-1}}

1190: \begin{tabular}{cccccc}

1191: \tableline

1192: Star & V  & IRAS 12\micron & Type & Date Observed & Calibrators\\  

1193:      &  mag  &    Jy            &      &            &             \\ 

1194: \tableline

1195: WR 95 & 13.4 & 4.56 & WC 9 & 07-30-2004 & HD 177716, HD 129456, HD 169916 \\

1196:       &      &      &      & 07-31-2004 & HD 169916, HD 165135, HD 192947 \\

1197: WR 106 & 11.93 & 11.67 & WC 9 & 07-09-2004 & HD 165135 \\

1198: NaSt1 & 15.4 & 14.5 & WN(?) & 07-30-2004 & Same as WR 95 above \\

1199:       &      &      &     & 09-29-2004\tablenotemark{a} & HD 168723 \\

1200:       &      &      &     & 09-30-2004\tablenotemark{a} & HD 168723 \\

1201: AG Car & 7.09 & 12.42 & LBV & 04-10-2004 & HD 81797, HD 107446, HD 129456, HD 139997\\

1202: \tableline

1203: \multicolumn{6}{l}{$^a$Baseline UT2-UT4}

1204: \end{tabular}

1205: %\tablenotetext{a}{Baseline UT2-UT4}

1206: %% Any table notes must follow the \end{tabular} command.

1207: %%generated with the \LaTeX\ table environment}

1208: %%\tablenotetext{b}{Yet another sample footnote for table~\ref{tbl-2}}

1209: %%\tablenotetext{c}{Another sample footnote for table~\ref{tbl-2}}

1210: %%\tablecomments{We can also attach a long-ish paragraph of explanatory

1211: %%material to a table.}

1212: \end{center}

1213: \end{table}}

1214: %%\end{rotate}

1215: %% If the table is more than one page long, the width of the table can vary

1216: %% from page to page when the default \tablewidth is used, as below.  The

1217: %% individual table widths for each page will be written to the log file; a

1218: %% maximum tablewidth for the table can be computed from these values.

1219: %% The \tablewidth argument can then be reset and the file reprocessed, so

1220: %% that the table is of uniform width throughout. Try getting the widths

1221: %% from the log file and changing the \tablewidth parameter to see how

1222: %% adjusting this value affects table formatting.

1223: 

1224: %% The \dataset{} macro has also been applied to a few of the objects to

1225: %% show how many observations can be tagged in a table.

1226: 

1227: 

1228: %% Tables may also be prepared as separate files. See the accompanying

1229: %% sample file table.tex for an example of an external table file.

1230: %% To include an external file in your main document, use the \input

1231: %% command. Uncomment the line below to include table.tex in this

1232: %% sample file. (Note that you will need to comment out the \documentclass,

1233: %% \begin{document}, and \end{document} commands from table.tex if you want

1234: %% to include it in this document.)

1235: 

1236: %%\input{table_sizes.tex}

1237: \clearpage

1238: 

1239: \begin{table}

1240: \begin{center}

1241: \caption{Keck Observations}

1242: \begin{tabular}{ccc}

1243: \tableline\tableline

1244: Star & Date of Observation & gaussian FWHM\\ 

1245:      &       &    mas (@ 10.7 \micron)  \\ 

1246: \tableline \tableline

1247: WR 106 &06-26-2005 & 56.0$\pm$6.9\tablenotemark{a} \\

1248: NaSt1 & 06-26-2005 & 50.0 \tablenotemark{b} \\

1249: \tableline

1250: \end{tabular}

1251: \tablenotetext{a}{Dominant error arises from calibration}

1252: \tablenotetext{b}{Object was unresolved. Value is an estimated upper limit}

1253: \end{center}

1254: \end{table}

1255: 

1256: \clearpage

1257: 

1258: \begin{table}

1259: \begin{center}

1260: \caption{Measured Sizes (MIDI)}

1261: \begin{tabular}{ccccc}

1262: \tableline\tableline

1263: Star & Distance & gaussian FWHM & Linear Size & References\\ 

1264:      &   Kpc    &    mas (@ 10.5 \micron)    &      AU    &  \\ 

1265: \tableline \tableline

1266: WR 95 & 2.09& 28.4$\pm$1.5 & 59.4 & 1  \\

1267: WR 106 & 2.3& 45.0$\pm$5.0 & 103.5  & 1 \\

1268: NaSt1 & 2.0& 20.1, 14.8 ($\pm$2.0)\tablenotemark{a} & 40.2, 29.6  & 2 \\

1269: AG Car & 6.0& 12.5$\pm$3.0 & 75 & 3 \\

1270: \tableline

1271: \end{tabular}

1272: \tablenotetext{a}{The two values are from two different baselines. Interpreting sizes from visibility may be inaccurate for this object; see text}

1273: \tablenotetext{}{References.-(1)\citet{vdh01a}; (2)\citet{cro99}; (3)\citet{hum89}}

1274: \end{center}

1275: \end{table}

1276: 

1277: \clearpage

1278: \begin{table}

1279: \begin{center}

1280: \caption{Starting Parameters for the WC Models}

1281: \begin{tabular}{lc}

1282: \tableline\tableline

1283: Parameter & Value\tablenotemark{a}\\

1284: \tableline\tableline

1285: T$_{{\rm eff}}$ & 1900 K \\

1286: R$_*$ & 14.6 R$_{\sun}$ \\

1287: Dust shell R$_{{\rm int}}$ & 410 R$_*$ (WR 95)\\

1288:                            &280 R$_*$ (WR 106)\\

1289: Dust shell thickness & 3 R$_{{\rm int}}$ (WR 95)\\

1290:                      & 30 R$_{{\rm int}}$ (WR 106)\\

1291: Dust composition & Amorphous C \\

1292: Density Profile & r$^{-2}$ \\

1293: Optical depth @ 10 \micron \tablenotemark{b}& 0.01 \\

1294: Grain sizes\tablenotemark{b} & 0.1 to 1 \micron\\

1295: \tableline

1296: \end{tabular}

1297: \tablenotetext{}{Unless stated otherwise, all values are from WvdHT}

1298: \tablenotetext{a}{Single values are common for both stars}

1299: \tablenotetext{b}{See text, Section 4.2, for references}

1300: \end{center}

1301: \end{table}

1302: 

1303: %% Use the figure environment and \plotone or \plottwo to include

1304: %% figures and captions in your electronic submission.

1305: %% To embed the sample graphics in

1306: %% the file, uncomment the \plotone, \plottwo, and

1307: %% \includegraphics commands

1308: %%

1309: %% If you need a layout that cannot be achieved with \plotone or

1310: %% \plottwo, you can invoke the graphicx package directly with the

1311: %% \includegraphics command or use \plotfiddle. For more information,

1312: %% please see the tutorial on "Using Electronic Art with AASTeX" in the

1313: %% documentation section at the AASTeX Web site,

1314: %% http://www.journals.uchicago.edu/AAS/AASTeX.

1315: %%

1316: %% The examples below also include sample markup for submission of

1317: %% supplemental electronic materials. As always, be sure to check

1318: %% the instructions to authors for the journal you are submitting to

1319: %% for specific submissions guidelines as they vary from

1320: %% journal to journal.

1321: 

1322: %% This example uses \plotone to include an EPS file scaled to

1323: %% 80% of its natural size with \epsscale. Its caption

1324: %% has been written to indicate that additional figure parts will be

1325: %% available in the electronic journal.

1326: 

1327: %%\begin{figure}

1328: %%\epsscale{.80}

1329: %%\plotone{f1.eps}

1330: %%\caption{Derived spectra for 3C138 \citep[see][]{heiles03}. Plots for all sources are available

1331: %%in the electronic edition of {\it The Astrophysical Journal}.\label{fig1}}

1332: %%\end{figure}

1333: 

1334: \clearpage

1335: 

1336: %% Here we use \plottwo to present two versions of the same figure,

1337: %% one in black and white for print the other in RGB color

1338: %% for online presentation. Note that the caption indicates

1339: %% that a color version of the figure will be available online.

1340: %%

1341: 

1342: \begin{figure}

1343: \plottwo{f1a.eps}{f1b.eps}

1344: \caption{Left: Visibility profile with wavelength for WR 95. The solid

1345: line is the mean of the two measurements. 

1346: Right: The corresponding (mean visibility) gaussian FWHM. \label{fig1}}

1347: \end{figure}

1348: 

1349: \clearpage

1350: 

1351: \begin{figure}

1352: \plottwo{f2a.eps}{f2b.eps}

1353: \caption{Left: Visibility profile with wavelength for WR 106 

1354: Right: The corresponding gaussian FWHM. \label{fig2}}

1355: \end{figure}

1356: 

1357: \clearpage

1358: 

1359: \begin{figure}

1360: \plottwo{f3a.eps}{f3b.eps}

1361: \caption{Left: Visibility profile with wavelength for NaSt1 (only one set of error bars shown for clarity). 

1362: Right: The corresponding gaussian FWHM. Transforming the visibility directly to a 

1363: gaussian size is, however, probably inaccurate for this object (see text). \label{fig3}}

1364: \end{figure}

1365: 

1366: \clearpage

1367: 

1368: \begin{figure}

1369: \plotone{f4.eps}

1370: \caption{MIDI spectrum and correlated flux for NaSt1. The Ne[II] 12.8 \micron\ feature is not seen in the correlated flux

1371: and is likely fully resolved by the interferometer as is a large fraction ($\sim 1/2$) of the total continuum flux. 

1372: \label{fig4}}

1373: \end{figure}

1374: 

1375: \clearpage

1376: 

1377: \begin{figure}

1378: \plottwo{f5a.eps}{f5b.eps}

1379: \caption{Left: Visibility profile with wavelength for AG Car. 

1380: Right: Corresponding gaussian FWHM. The error bars are similar for both measurements. For clarity only one set is shown.  \label{fig5}}

1381: \end{figure}

1382: 

1383: \clearpage

1384: 

1385: \begin{figure}

1386: \plotone{f6.eps}

1387: \caption{MIDI spectrum for AG Car, with the ISO spectrum (scaled overall to match the MIDI flux level) for comparison. 

1388: The PAH feature at 11.3 \micron\ is not seen in the MIDI spectrum.

1389:  \label{fig6}}

1390: \end{figure}

1391: 

1392: \clearpage

1393: 

1394: \begin{figure}

1395: \plotone{f7.eps}

1396: \caption{Visibility versus baseline for WR 106. The solid line is the model visibility, the filled circles are the Keck visibilities (upper left) 

1397: and the diamond, MIDI at 10.5\micron.\label{fig7}}

1398: \end{figure}

1399: 

1400: \clearpage

1401: 

1402: \begin{figure}

1403: \plotone{f8.eps}

1404: \caption{Visibility versus wavelength for WR 106. The line (with error bars) is the MIDI measurement.

1405: Models with varying inner radii, R$_{int}$ (in units of R$_*$), for the shell are shown for comparison.\label{fig8}}

1406: \end{figure}

1407: 

1408: \clearpage

1409: 

1410: \begin{figure}

1411: \plotone{f9.eps}

1412: \caption{SED of WR 106. The solid line is our best model, the grey circles are 

1413: the measurements from MIDI, the squares are measurements from the literature (see text

1414: for sources). \label{fig9}}

1415: \end{figure}

1416: 

1417: \clearpage

1418: 

1419: \begin{figure}

1420: \plotone{f10.eps}

1421: \caption{Dust temperature profile for the WR 106 best model.\label{fig10}}

1422: \end{figure}

1423: 

1424: %\begin{figure}

1425: %\plotone{visibase2.eps}

1426: %\caption{Visibilty vs Baseline: The dotted line is the model visibility for a model with 

1427: %R$_{int}$ = 350R$_*$, the dashed line is for a model with R$_{int}$ = 250R$_*$. 

1428: %Filled circles are Keck and the diamond, MIDI measurements.\label{fig11}}

1429: %\end{figure}

1430: 

1431: \clearpage

1432: 

1433: \begin{figure}

1434: \plotone{f11.eps}

1435: \caption{Visibility versus wavelength for WR 95. The line (with error bars) is the MIDI measurement, the filled circles are the visibility points 

1436: from the best model. For comparison we show visibilities for models with different inner radii (R$_{int}$) for the shell.\label{fig11}}

1437: \end{figure}

1438: 

1439: \clearpage

1440: 

1441: \begin{figure}

1442: \plotone{f12.eps}

1443: \caption{SED for WR 95: The solid line is our best model, the grey circles the measurements from MIDI (de-reddened as in text), 

1444: the diamonds are measurements from the literature (see text for sources)\label{fig12}}

1445: \end{figure}

1446: 

1447: \clearpage

1448: 

1449: \begin{figure}

1450: \plotone{f13.eps}

1451: \caption{Dust temperature profile of our WR 95 model.\label{fig13}}

1452: \end{figure}

1453: %% This figure uses \includegraphics to scale and rotate the still frame

1454: %% for an mpeg animation.

1455: 

1456: %%\begin{figure}

1457: %%\includegraphics[angle=90,scale=.50]{f3.eps}

1458: %%\caption{Animation still frame taken from \citet{kim03}.

1459: %%This figure is also available as an mpeg

1460: %%animation in the electronic edition of the

1461: %%{\it Astrophysical Journal}.}

1462: %%\end{figure}

1463: 

1464: %% If you are not including electonic art with your submission, you may

1465: %% mark up your captions using the \figcaption command. See the

1466: %% User Guide for details.

1467: %%

1468: %% No more than seven \figcaption commands are allowed per page,

1469: %% so if you have more than seven captions, insert a \clearpage

1470: %% after every seventh one.

1471: 

1472: %% Tables should be submitted one per page, so put a \clearpage before

1473: %% each one.

1474: 

1475: %% Two options are available to the author for producing tables:  the

1476: %% deluxetable environment provided by the AASTeX package or the LaTeX

1477: %% table environment.  Use of deluxetable is preferred.

1478: %%

1479: 

1480: %% Three table samples follow, two marked up in the deluxetable environment,

1481: %% one marked up as a LaTeX table.

1482: 

1483: %% In this first example, note that the \tabletypesize{}

1484: %% command has been used to reduce the font size of the table.

1485: %% We also use the \rotate command to rotate the table to

1486: %% landscape orientation since it is very wide even at the

1487: %% reduced font size.

1488: %%

1489: %% Note also that the \label command needs to be placed

1490: %% inside the \tablecaption.

1491: 

1492: %% This table also includes a table comment indicating that the full

1493: %% version will be available in machine-readable format in the electronic

1494: %% edition.

1495: 

1496: %% Do not put multiple tabular environments within a single table.

1497: %% The optional \label should appear inside the \caption command.

1498: 

1499: %%\clearpage

1500: %%\begin{rotate}

1501: 

1502: %% The following command ends your manuscript. LaTeX will ignore any text

1503: %% that appears after it.

1504: 

1505: \end{document}

1506: 

1507: %%

1508: %% End of file `sample.tex'.

1509: