1: \documentclass{emulateapj}

2: 

3: \newcommand{\be}{\begin{equation} }

4: \newcommand{\ee}{\end{equation} }

5: \newcommand{\bea}{\begin{eqnarray} }

6: \newcommand{\eea}{\end{eqnarray} }

7: \newcommand{\etal}{\mbox{\rm et al.~} }

8: \newcommand{\kms}{\mbox{km s$^{-1}~$} }

9: \newcommand{\msun}{M$_{\odot}~$ }

10: \newcommand{\rsun}{R$_{\odot}~$ }

11: \newcommand{\lsun}{L$_{\odot}~$ }

12: \newcommand{\chisq}{$\chi_{\nu}^2$ }

13: \newcommand{\teff}{${\rm T_{eff}}$ }

14: 

15: \slugcomment{Accepted to ApJL: 25 Aug 2008}

16: 

17: \shortauthors{Maness {\it et~al.\/}}

18: 

19: \shorttitle{Millimeter Imaging of HD 32297}

20: 

21: \begin{document}

22: 

23: \title{CARMA Millimeter-Wave Aperture Synthesis Imaging of the HD

24: 32297 Debris Disk}

25: 

26: \author{ 

27: H. L. Maness\altaffilmark{1},

28: M. P. Fitzgerald\altaffilmark{2}, 

29: R. Paladini\altaffilmark{3},

30: P. Kalas\altaffilmark{1},

31: G. Duchene \altaffilmark{1,4}, 

32: James R. Graham\altaffilmark{1}

33: } 

34: 

35: \altaffiltext{1}{Astronomy Dept, UC Berkeley, Berkeley, CA 94720}

36: \altaffiltext{2}{Michelson Fellow, IGPP, L-413, Lawrence Livermore

37: National Laboratory, 7000 East Ave, Livermore, CA 94550}

38: \altaffiltext{3}{Spitzer Science Center, California Institute of

39: Technology, Pasadena, CA 91125}

40: \altaffiltext{4}{Laboratoire d'Astrophysique de Grenoble, CNRS/UJF UMR

41: 5571, F-38041 Grenoble cedex 9, France}

42: 

43: \begin{abstract}

44: We present the first detection and mapping of the HD 32297 debris disk

45: at 1.3 mm with the Combined Array for Research in Millimeter-wave Astronomy (CARMA).

46: With a sub-arcsecond beam, this detection represents the highest

47: angular resolution (sub)mm debris disk observation made to date. Our

48: model fits to the spectral energy distribution from the CARMA flux and

49: new Spitzer MIPS photometry support the earlier suggestion that at

50: least two, possibly three, distinct grain populations are traced by

51: the current data.  The observed millimeter map shows an asymmetry

52: between the northeast and southwest disk lobes, suggesting large

53: grains may be trapped in resonance with an unseen exoplanet.

54: Alternatively, the observed morphology could result from the recent

55: breakup of a massive planetesimal.  A similar-scale asymmetry is also

56: observed in scattered light but not in the mid-infrared.  This

57: contrast between asymmetry at short and long wavelengths and symmetry

58: at intermediate wavelengths is in qualitative agreement with

59: predictions of resonant debris disk models.  With resolved

60: observations in several bands spanning over three decades in

61: wavelength, HD 32297 provides a unique testbed for theories of grain

62: and planetary dynamics, and could potentially provide strong

63: multi-wavelength evidence for an exoplanetary system.

64: \end{abstract}

65: 

66: \keywords{circumstellar matter $-$ planetary systems: formation $-$

67: planetary systems: protoplanetary disks $-$ stars: individual (HD

68: 32297)}

69: 

70: \section{Introduction}

71: 

72: Debris disks provide the principal means of studying the formation and

73: evolution of planetary systems on timescales of 10$-$100 Myr.

74: Evidence for exoplanets in these systems can be found by matching

75: density variations in debris disks to theoretical models of the

76: gravitational perturbations caused by planets (e.g., Reche et

77: al. 2008).  A modest sample of debris disks have now been imaged in

78: the visible, and many show substructure such as clumps, warps, and

79: offsets, consistent with dynamical perturbations by massive planets.

80: However, a wide variety of other mechanisms can produce similar

81: structures (Moro-Martin et al. 2007, and references therein).  As

82: different wavebands are sensitive to different grain sizes, which are

83: in turn subject to different dynamical influences, multi-wavelength

84: observations offer the most promising path towards definitively

85: classifying the physical mechanisms at work in these systems.  At

86: present, only a few debris disks have resolved observations spanning

87: more than a decade in wavelength.  A particularly critical, though

88: technologically challenging, deficit of observations lies at

89: (sub)millimeter wavelengths, which trace large grains primarily

90: affected by gravitational forces.  To date, bolometer arrays have

91: resolved 8 debris disks in the (sub)mm (e.g., Holland et al. 1998,

92: Greaves et al. 1998). However, such low-resolution

93: ($\theta_{\mathrm{beam}} \gtrsim 10''$) single-dish measurements are

94: limited to the largest, nearest disks.  Higher resolution

95: interferometric observations are needed to access the larger debris

96: disk population already imaged at shorter wavelengths. Some pioneering

97: work has been done in this area; OVRO and PdBI have detected and

98: resolved two debris disks (Vega: Koerner et al. 2001, Wilner et

99: al. 2002; HD 107146: Carpenter et al. 2005).  Recently,

100: \citet{Corder08} mapped HD 107146 with the Combined Array for Research

101: in Millimeter-wave Astronomy (CARMA) at 1.3 mm, providing the highest

102: fidelity interferometric debris disk map to date.  Here, we report the

103: near-simultaneous CARMA detection of HD 32297, the third debris disk

104: mapped with a (sub)mm interferometer.  With a sub-arcsecond beam, this

105: detection is the highest angular resolution (sub)mm debris disk

106: observation made to date.

107: 

108: HD 32297 is a $\sim$30 Myr A-star at $112_{-12}^{+15}$ pc

109: \citep{Perryman97}, first discovered to host a resolved debris disk

110: with HST/NICMOS near-infrared (NIR) imaging \citep{Schneider05}.  The

111: discovery image showed an edge-on debris disk extending to 400 AU

112: (3.3$''$), with an inner-disk brightness asymmetry inward of 60 AU

113: ($0.5''$).  \citet{Kalas05} subsequently imaged HD 32297 in the

114: optical, revealing an asymmetric, extended outer disk ($\sim$1700 AU,

115: 15$''$) likely interacting with the interstellar medium.  Later,

116: \citet{Redfield07} detected circumstellar gas in this system,

117: reporting the strongest Na I absorption measured toward any known

118: debris disk. Most recently, \citet{Fitzgerald07}, hereafter F07, and

119: \citet{Moerchen07} resolved HD 32297 in mid-infrared (MIR), thermal

120: emission.  Detailed analysis of the spectral energy distribution (SED)

121: by F07 showed that multiple grain populations may be present in the

122: disk.  The lack of long wavelength data needed to characterize the

123: large grain properties of HD 32297 motivated the CARMA observations

124: presented here.

125: 

126: \section{Observations and Data Reduction}

127: 

128: We observed HD 32297 with CARMA, located 7200 ft above sea level

129: outside Big Pine, California and consisting of six 10.4-m and nine

130: 6.1-m antennas, previously comprising the OVRO and BIMA arrays. We

131: used CARMA's D configuration (baselines: 11$-$148 m) on 2008 March 06

132: and the more extended C configuration (baselines: 26$-$370 m) on 2007

133: October 26 and 2007 November 08.  We tuned the receivers to a central

134: frequency of 227 GHz.  The total continuum bandwidth is 1.5 GHz,

135: contained in three 500 MHz bands.  Our total observation time was 19

136: hours in good weather with rms path errors of $\lesssim 175 \mu$m and

137: zenith opacities of $\tau_{230} \lesssim 0.35$.  Throughout our

138: observations, we used optical offset guiding to maximize pointing

139: accuracy (Corder, Carpenter \& Wright 2008, in prep).

140: 

141: We calibrated the data using the MIRIAD software package.  We

142: performed passband calibration with 15 minutes integrations on a bright

143: quasar (3C 84, 3C 111) observed at the start of each track.  We

144: derived time-dependent phase solutions from 3 minutes integrations on

145: J0530+135, observed following each 15 minutes integration on source.  In

146: addition, we used 1 minutes integrations on a secondary calibrator,

147: 3C 120 (observed every cycle), to test the astrometric

148: accuracy of the interferometer and the integrity of our phase

149: solutions.  Finally, we flux-calibrated the data using a 5 minutes

150: integration on a planet (Uranus, Mars) observed once per track.  The

151: systematic uncertainty in CARMA's absolute flux scale is $\sim$20\%

152: (W. Kwon, private communication).

153: 

154: The 1.3 mm map of HD 32297 combining all observations is shown in

155: Figure \ref{map}.  For this map, we adopted natural weighting to

156: provide optimal sensitivity and processed the visibility data as a

157: mosaic to accomodate the heterogeneous array.  We deconvolved the

158: dirty map using the Steer CLEAN algorithm for mosaics \citep{Steer84},

159: set to bring the rms in the cleaned region of the residual map to that

160: measured in an off-source region.  To emphasize possible resolved

161: structure, we restored the map with a circular beam with radius equal

162: to that of the semimajor axis of the naturally weighted beam

163: ($\theta_{\mathrm{beam}}=0.9''$).

164: 

165: In addition to the new CARMA data presented here, we also extracted

166: Spitzer MIPS photometry from director's discretionary time program 225

167: (PI: G. Schneider).  Using the MOPEX package, we manually calibrated

168: the 24 $\mu$m data to remove the strong background gradient and

169: ``jailbars'' evident in the pipeline processed image.  The 70 $\mu$m

170: data did not suffer severe artifacts; in this case, we simply

171: mosaicked the data. The 160 $\mu$m data are contaminated by a spectral

172: leak, which occurs for bright sources and is typically corrected using

173: observations of a calibration star.  However, no such calibration

174: observation was made in this program; we, therefore, include the 160

175: $\mu$m data in this analysis only as an upper-limit.  The MIPS fluxes

176: as derived from aperture photometry are $F_{\nu}$(23.68 $\mu$m) = 0.21

177: $\pm$ 0.01 Jy, $F_{\nu}$(71.42 $\mu$m) = 0.85 $\pm$ 0.06 Jy,

178: $F_{\nu}$(155.9 $\mu$m) $<$ 0.46 $\pm$ 0.06 Jy.  In the SED modeling

179: (\S 4), we adopt calibration uncertainties listed in the MIPS data

180: handbook and apply a color correction appropriate to the modeled grain

181: temepratures.

182: 

183: \section{Results}

184: 

185: \begin{figure}

186: \centering

187: \includegraphics[width=.45\textwidth,angle=0]{f1.eps} 

188: \caption{CARMA 1.3-mm continuum map of HD 32297.  The black cross

189: marks the stellar position, with the full-width representing 

190: four-times the total rms millimeter positional uncertainty.  Contours

191: begin at 2$\sigma$ and increase by 1$\sigma$ thereafter ($\sigma=$

192: 0.44 mJy beam$^{-1}, \theta_{\mathrm{beam}}=0.9''$). The source

193: morphology suggests that HD 32297 may be marginally resolved.  The

194: measured position angle is consistent with that observed in the NIR

195: and MIR ($45-57^{\circ}$).  The centroid of the disk emission is

196: offset from the stellar position at the 4$\sigma$ level.}

197: \label{map}

198: \end{figure}

199: 

200: Figure \ref{map} shows a 1.3 mm continuum source detected at the

201: 7$\sigma$ level.  The observed peak in the map is $2.8 \pm 0.4$ mJy

202: beam$^{-1}$.  The probability of detecting an unrelated background

203: source within 1$''$ of the HD 32297 stellar position is $\ll$1\%,

204: based on recent source counts at 850 $\mu$m (e.g., Scott et al. 2002).

205: Thus, we conclude that the detected emission is associated with the HD

206: 32297 debris disk.  The map morphology suggests the source may be

207: resolved (see below).  However, in the D-array data set alone the

208: visibility amplitudes are constant with baseline length (Figure

209: \ref{vis}), suggesting the source is unresolved.  We therefore, adopt

210: an integrated flux measured using the D-array data set only: $5.1 \pm

211: 1.1$ mJy.

212: 

213: \begin{figure}

214: \centering

215: \includegraphics[width=0.18\textwidth,angle=90]{f2.eps}

216: \caption{Vector-averaged visibility amplitudes as a function of

217: projected baseline length from phase center.  The combined data

218: amplitudes decrease with distance from phase center, suggesting the

219: source is resolved at the 2$\sigma$ level.  The D-array data, on the

220: other hand, are consistent with a point-source model (grey dashed

221: line), allowing a robust flux measurement.}

222: \label{vis}

223: \end{figure}

224: 

225: The source centroid in Figure \ref{map} appears offset from the

226: Hipparcos stellar position (located at phase center).  To quantify

227: this result, we fit a point source model to the visibility data; the

228: measured offset is $\Delta r=0.43 \pm 0.08''$ ($\Delta \alpha = -0.30

229: \pm 0.08''$, $\Delta \delta = -0.31 \pm 0.07''$).  This offset is

230: significantly larger than that observed for test calibrator, 3C 120,

231: observed each source cycle ($\Delta r=0.041 \pm 0.001''$, $\Delta

232: \alpha = -0.010 \pm 0.002''$, $\Delta \delta = 0.040 \pm 0.001''$).

233: In addition to systematic astrometric uncertainties, we estimate a

234: statistical positional error for HD 32297 of $\Delta \theta =

235: (4/\pi)^{1/4} / \sqrt{8 \ln 2} \cdot \theta_\mathrm{beam} / SNR \sim

236: 0.06''$, in agreement with the formal errors derived from fitting the

237: visibilities.  Combining the systematic and statistical errors, the

238: total rms positional uncertainty for HD 32297 is $0.09''$.  Thus, the

239: peak continuum emission is offset from the stellar position at the

240: 4$\sigma$ level.

241: 

242: As previously noted, the Figure \ref{map} map suggests the source is

243: marginally resolved.  To quantify this effect, in Figure \ref{vis}, we

244: show the visibility amplitudes for the combined data

245: set as a function of projected baseline length from phase center.  The

246: data are binned by baseline length in three circularly symmetric

247: annular bins, with widths chosen to provide approximately the same

248: number of visibilities per bin.  The errors in amplitude represent the

249: standard deviation in the mean of the visibilities in each bin.

250: Figure \ref{vis} shows that the source amplitudes decline with

251: baseline length, with a deviation from a point-source model of

252: approximately 2$\sigma$.  Thus, the source appears to be indeed

253: resolved, though additional data are needed to confirm this result.

254: In particular, we note that uncorrected atmospheric phase errors at mm

255: wavelengths can artificially enlarge targets. Still, the observed

256: morphology agrees with that observed at other wavelengths.  Fitting an

257: elliptical Gaussian to the visibilities yields a FWHM of $1.8 \pm

258: 0.6'' \times 0.2 \pm 0.6''$ with a position angle of $55 \pm

259: 10^\circ$, consistent to within 1$\sigma$ of the orientation observed

260: in the NIR and MIR (P.A.$=45-57^{\circ}$, Schneider et al. 2005, F07,

261: Moerchen et al. 2007).

262: 

263: \begin{figure}

264: \centering

265: \includegraphics[width=0.35\textwidth,angle=0]{f3.eps}

266: \caption{Two-population SED fit (see text and F07). The small-grain

267:   population is plotted in red, the large-grain population in blue,

268:   the stellar photosphere in grey, and the composite model in green.

269:   The relative shadings correspond to the 2-d histograms of the

270:   component SEDs for each model state in the Markov chains.

271:   Photometry references are as follows: MICHELLE (F07), T-ReCS

272:   (extracted from data in \citet{Moerchen07}), IRAS \citep{Moor06},

273:   MIPS (this paper), CARMA (this paper).  The model provides a good

274:   fit to the mid- and far-infrared data but underestimates the

275:   observed mm flux, potentially providing evidence for a third grain

276:   population.}

277: \label{sed}

278: \end{figure}

279: 

280: We observe no CO $J=2-1$ line emission in our data.  The 3$\sigma$

281: limit is 22 mJy beam$^{-1}$ (1.5 K, $\theta_{\mathrm{beam}}=0.70'' \times

282: 0.52''$) in a 42 km s$^{-1}$ channel.  Adopting the same assumptions

283: as \citet{Dent05}, the corresponding gas mass limit is $M_{gas}

284: \lesssim 0.3 M_{\mathrm{Jup}}$.

285: 

286: \section{Discussion}

287: 

288: F07 attempted to model the observed SED and $N'$-band image of HD

289: 32297 with a single ring of grains of characteristic size, but found

290: that a second population of grains was needed to adequately fit the

291: observed SED for $\lambda \gtrsim 25 \mu$m.  Indeed, the contrast

292: between the observed mm and $N'$-band morphologies (see below and

293: Figure \ref{multiwav}) suggests that the grains responsible for the

294: emission at each wavelength constitute separate populations.  To test

295: whether the population of mm-emitting grains is consistent with the

296: second, larger grain population proposed by F07 to fit the SED at 25

297: $\mu$m $\lesssim \lambda \lesssim$ 60 $\mu$m, we revisit the F07

298: model, adopting their data and fitting method, and incorporating the

299: Qa-band flux of \citet{Moerchen07} and the new MIPS and CARMA

300: fluxes. The free parameters in the single population model of F07 are

301: the disk inclination ($i$), position angle (PA), radii of the inner

302: and outer edges ($\varpi_0$, $\varpi_1$), surface density power-law

303: index ($\gamma$), vertical optical depth to absorption at the inner

304: edge ($\tau_0\equiv\tau_\perp^{\mathrm{abs}}(\varpi_0)$), stellar flux

305: factor ($\xi$), and effective grain size ($\lambda_\mathrm{sm}$). To

306: fit the long-wavelength SED ($\lambda \gtrsim 25 \mu$m), we augment

307: this model with a population of larger grains of effective size

308: $\lambda_\mathrm{lg}$ and total emitting area $A_\mathrm{lg}$, located

309: in a narrow ring at the small-grain inner-disk edge, $\varpi_0$.

310: These grains contribute flux, $F_{\nu,\mathrm{lg}}$, according to:

311: \begin{eqnarray}

312: \epsilon_{\nu,\mathrm{lg}} &=&\left\{

313: \begin{array}{ll}

314: \lambda_\mathrm{lg}/\lambda & \mbox{if $\lambda>\lambda_\mathrm{lg}$,} \\

315: 1 & \mbox{otherwise},

316: \end{array}\right. \\

317: T_\mathrm{lg}(r) &=& 468 \left(\frac{L_*/L_\sun}{\lambda_\mathrm{lg}/1\,\micron}\right)^{1/5} \left(\frac{r}{1\,\mathrm{AU}}\right)^{-2/5} \mathrm{K}, \\ 

318: F_{\nu,\mathrm{lg}} &=& \left(\frac{A_\mathrm{lg}}{d^2}\right) \epsilon_{\nu,\mathrm{lg}} B_\nu[T_\mathrm{lg}(\varpi_0)].

319: \end{eqnarray}

320: 

321: Following F07, we ran three Monte Carlo Markov chains with $3\times

322: 10^4$ samples, and simulataneously fit both the large and small grain

323: populations.  The results of this procedure are listed in Table

324: \ref{sed_table}, and the corresponding range of allowed dust emission

325: is plotted in Figure \ref{sed} (red: small grains, blue: large grains,

326: grey: photosphere, green: composite).  While the two-population model

327: provides a satisfactory fit to the mid- and far-infrared data, the mm

328: flux is underestimated at the 4$\sigma$ level, suggesting that the

329: second grain population proposed by F07 is not responsible for the

330: majority of the mm flux.  Nevertheless, the new MIPS data lend further

331: evidence to the F07 suggestion that two populations are needed to fit

332: the observed SED for $\lambda \lesssim 160 \mu$m.  Therefore, the fit

333: suggests at least three distinct populations are traced by the current

334: observations.  However, from the two-population model, only the 1.3 mm

335: flux appears to trace the putative third population.  Since this

336: population is described by both a mass/emitting area and a

337: size/temperature, we lack sufficient data to fully characterize it via

338: modeling. We note, though, that this population likely traces $\gtrsim

339: 95$\% of the total dust mass.  Assuming a characteristic stellocentric

340: distance of $\sim$50 AU (\S 3), $L_{*}=5.4L_{\sun}$ (F07), and an

341: effective grain size of 1.3-mm, the implied mm grain temperature is

342: $\sim$30 K, suggesting a dust mass of $M_{\mathrm{mm}}\sim M_{\earth}$

343: (adopting an opacity of 1.7 cm$^2$ g$^{-1}$).  This estimated mass is

344: among the highest observed for debris disks detected in the (sub)mm

345: and is two orders of magnitude larger than that implied for the

346: ``large-grain'' population in the SED fit:

347: $M_{\mathrm{lg}}\sim0.02M_{\earth}$ (using the fitted parameters in

348: Table \ref{sed_table} and assuming spherical grains with a density of

349: 1 g cm$^{-3}$).  Future far-infrared / sub-mm observations are needed

350: to confirm the three populations proposed here and better constrain

351: their properties.

352: 

353: \begin{deluxetable}{lcc}

354: \tablecaption{\label{sed_table} Best-Fit Model Parameters}

355: \tablewidth{0pt}

356: \tablehead{

357: \colhead{Parameter}  &

358: \colhead{Best-fit} &

359: \colhead{Description} }

360: \startdata

361: $i$ (deg) & $90  \pm 5$ & disk inclination \\

362: PA (deg) & $46 \pm 3$ & disk position angle \\

363: $\varpi_0$ (AU) & $70_{-10}^{+20}$ & inner edge \\

364: $\varpi_1$ (AU) & $>$ 1200 & outer edge \\

365: $\log_{10}\tau_0$ & $-2.4_{-0.2}^{+0.3}$ & vertical optical depth at $\varpi_0$ \\

366: $\gamma$ & $<$ -1.63 & surf density power-law index \\

367: $\xi / 7.22 \times 10^{-20}$ & $1.03 \pm 0.03$ & $(R_{*}/d)^2$, stellar flux factor \\

368: $\log_{10}(\lambda_\mathrm{sm}/1\,\micron)$ & $-1.4_{-1.3}^{+0.6}$ & small grain effective size \\

369: $\log_{10}(A_\mathrm{lg}/1\,\mathrm{cm}^2)$ & $28.9 \pm 0.2$ & large grain emitting area \\

370: $\log_{10}(\lambda_\mathrm{lg}/1\,\micron)$ & $1.3_{-0.1}^{+0.2}$ & large grain effective size

371: \enddata

372: \tablecomments{Confidence intervals are 95\% for marginal

373: posterior distributions.  Adopted priors are as in F07, except for

374: $\gamma$, which is constrained to be in the domain [-4,0] due to the

375: inconsistency of rising surface density with the scattered-light image

376: and CARMA map; the large grain effective size was constrained by a

377: log-uniform prior from 1\,nm to 100\,mm and a requirement that

378: $\lambda_{\mathrm{lg}} > \lambda_{\mathrm{sm}}$.}

379: \end{deluxetable} 

380: \begin{figure}

381: \centering

382: \includegraphics[width=0.23\textwidth,angle=90]{f4.eps}

383: \caption{{\it Left:} CARMA contours of HD 32297, overlaid on the NIR

384: scattered-light image from Schneider et al. (2005). {\it Right:}

385: Photosphere-subtracted MIR contours from F07 overlaid on the same

386: image. The asymmetry in the CARMA data between the northeast and

387: southwest lobes suggests the large, mm-sized grains may be trapped in

388: resonance with an unseen exoplanet.  A similar asymmetry is also

389: observed in scattered light but not in the MIR.  The contrast between

390: asymmetry at short and long wavelengths and symmetry at intermediate

391: wavelengths is a direct prediction of the resonant debris disk models

392: of Wyatt (2006).}

393: \label{multiwav}

394: \end{figure}

395: 

396: To address the observed mm morphology, in Figure \ref{multiwav}, we

397: qualitatively compare the CARMA mm map (contours, left panel) to the

398: MIR image of F07 (contours, right panel) and the NIR scattered-light

399: image of \citet{Schneider05} (color, both panels).  The images at

400: all three wavelengths are consistent with an edge-on disk.  However,

401: while both the NIR and CARMA data exhibit a brightness asymmetry

402: between the northeast and southwest lobes inward of $\sim$0.5$''$, the

403: F07 MIR image is consistent with azimuthal symmetry.  We note that

404: \citet{Moerchen07} found evidence for asymmetry in their Qa-band data,

405: though their data were not PSF subtracted, and the asymmetry was in

406: the opposite sense as observed in the NIR / CARMA data (NE lobe

407: brighter than SW).

408: 

409: Recently, \citet{Grigorieva07} used predictions from their numerical

410: model of collisional avalanches to suggest that the observed

411: scattered-light asymmetry in HD 32297 results from the breakup of a

412: large planetesimal.  However, while a massive collision can explain

413: the observed NIR and mm morphology, it is not clear why the MIR image

414: would not show a similar asymmetry.  An alternative hypothesis is that

415: the structure results from a planetary-induced resonance.  In this

416: case, the dust is either trapped in resonance as it drifts inward due

417: to Poynting-Robertson drag, or it remains locked in resonance after

418: being generated by parent planetesimals in resonance themselves (e.g.,

419: Krivov et al. 2007, and references therein).  Interestingly, a recent

420: study of the latter mechanism by \citet{Wyatt06} directly predicts a

421: contrast between asymmetry at short and long wavelengths and symmetry

422: at intermediate wavelengths, as observed in HD 32297.  In this model,

423: (sub)mm emission is dominated by large grains, which have the same

424: clumpy resonant distribution as the parent planetesimals.  Small

425: grains, traced at short wavelengths, exhibit a similar asymmetry, as

426: they are preferentially born in the high density, resonant structures

427: before being rapidly expelled from the system.  Lastly,

428: moderately-sized grains sampled at intermediate wavelengths remain

429: bound to the star, but have fallen out of resonance due to radiation

430: pressure and are subsequently scattered into an axisymmetric

431: morphology.  

432: 

433: This proposed scenario of \citet{Wyatt06} conveniently explains the

434: qualitative picture for HD 32297 depicted in Figure \ref{multiwav},

435: yet we caution that this suggestion is highly speculative, and future

436: rigorous modeling of this system is needed to draw firm conclusions

437: from the available data.  One potential problem with this hypothesis

438: is that the SED model predicts that the small ($N'$-band-emitting)

439: grains have sizes $\lesssim$ 1 $\mu$m (Table \ref{sed_table}), similar

440: to that expected to produce the NIR scattered-light image (see

441: discussion in F07).  However, the interpretation of Figure

442: \ref{multiwav} in terms of the \citet{Wyatt06} models requires that

443: the NIR-scattering and MIR-emitting grains have different sizes.  This

444: ambiguity could potentially be resolved through simultaneous modeling

445: of scattering and emission, incorporating the optical image presented

446: in \citet{Kalas05}. Most importantly, and independent of speculation,

447: HD 32297 is currently one of only a few debris disks with resolved

448: observations in four wavelength regimes (optical, NIR, MIR, mm).

449: Taken together, these observations can provide a unique testbed for

450: theories of grain and planetary dynamics.

451: 

452: \acknowledgments H.M. is funded by the GRFP at NSF and the GOPF at UC

453: Berkeley.  M.\,P.\,F. acknowledges support from the Michelson

454: Fellowship Program. Support for CARMA construction was derived from

455: the states of CA, IL, and MD, the Gordon and Betty Moore Foundation,

456: the Eileen and Kenneth Norris Foundation, the Caltech Associates, and

457: the National Science Foundation.

458: 

459: \begin{thebibliography}{}

460: \bibitem[Carpenter et al.(2005)]{Carpenter05} Carpenter, J.~M., et al.\ 2005, \aj, 129, 1049

461: \bibitem[Corder et al.(2008)]{Corder08} Corder, S.~A., et al.\ 2008, \apjl, submitted

462: \bibitem[Dent et al.(2005)]{Dent05} Dent, W.~R.~F., et al.\ 2005, \mnras, 359, 663 

463: \bibitem[Fitzgerald et al.(2007)]{Fitzgerald07} Fitzgerald, M.~P., et al.\ 2007, \apj, 670, 557 

464: \bibitem[Greaves et al.(1998)]{Greaves98} Greaves, J.~S., et al.\ 1998, \apjl, 5 06, L133

465: \bibitem[Grigorieva et al.(2007)]{Grigorieva07} Grigorieva, A., et al.\ 2007, \aap, 461, 537 

466: \bibitem[Holland et al.(1998)]{Holland98} Holland, W.~S., et al.\ 1998, \nat, 392, 788 

467: \bibitem[Kalas(2005)]{Kalas05} Kalas, P.\ 2005, \apjl, 635, L169 

468: \bibitem[Krivov et al.(2007)]{Krivov07} Krivov, A.~V. et al.\ 2007, \aap, 462, 199 

469: \bibitem[Koerner et al.(2001)]{Koerner01} Koerner, D.~W., et al.\ 2001, \apjl, 560, L181 

470: \bibitem[Moerchen et al.(2007)]{Moerchen07} Moerchen, M.~M., et al.\ 2007, \apjl, 666, L109 

471: \bibitem[Mo{\'o}r et al.(2006)]{Moor06} Mo{\'o}r, A., et al.\ 2006, \apj, 644, 525 

472: \bibitem[Moro-Martin et al.(2007)]{MoroMartin07} Moro-Martin, A., et al.\ 2007, ArXiv Astrophysics e-prints, arXiv:astro-ph/0703383 

473: \bibitem[Perryman et al.(1997)]{Perryman97} Perryman, M.~A.~C., et al.\ 1997, \aap, 323, L49 

474: \bibitem[Reche et al.(2008)]{Reche08} Reche, R., et al.\ 2008, \aap, 480, 551 

475: \bibitem[Redfield(2007)]{Redfield07} Redfield, S.\ 2007, \apjl, 656, L97 

476: \bibitem[Schneider et al.(2005)]{Schneider05} Schneider, G., et al.\ 2005, \apjl, 629, L117 

477: \bibitem[Scott et al.(2002)]{Scott02} Scott, S.~E., et al.\ 2002, \mnras, 331, 817

478: \bibitem[Steer et al.(1984)]{Steer84} Steer, D.~G., et al.\ 1984, \aap, 137, 159

479: \bibitem[Wilner et al.(2002)]{Wilner02} Wilner, D.~J., et al.\ 2002, \apjl, 569, L115 

480: \bibitem[Wyatt(2006)]{Wyatt06} Wyatt, M.~C.\ 2006, \apj, 639, 1153 

481: \end{thebibliography}{}

482: 

483: \end{document}

484: