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

2: %\documentclass{emulateapj}

3: 

4: %\usepackage{graphicx}

5: %\usepackage{lscape}

6: % general LaTeX mark--ups

7: %

8: 

9: \newcommand{\balq}{BAL~quasar}

10: \newcommand{\balQ}{BAL~Quasar}

11: \newcommand{\balqs}{BAL~quasars}

12: \newcommand{\balQs}{BAL~Quasars}

13: \newcommand{\uv}{ultraviolet}

14: \newcommand{\xray}{\hbox{X-ray}}

15: \newcommand{\asca}{{\emph{ASCA}}}

16: \newcommand{\chandra}{{\emph{Chandra}}}

17: \newcommand{\conx}{{\emph{Constellation-X}}}

18: \newcommand{\xmm}{\emph{XMM-Newton}}

19: \newcommand{\sax}{{\emph{Beppo-SAX}}}

20: \newcommand{\rosat}{{\emph{ROSAT}}}

21: \newcommand{\einstein}{{\emph{Einstein}}}

22: \newcommand{\heao}{\mbox{\emph{HEAO-1}}}

23: \newcommand{\hst}{{\emph{HST}}}

24: \newcommand{\iue}{{\emph{IUE}}}

25: \newcommand{\iras}{{\emph{IRAS}}}

26: \newcommand{\iso}{{\emph{ISO}}}

27: \newcommand{\spitzer}{{\emph{Spitzer}}}

28: 

29: \newcommand{\feka}{\mbox{Fe\,K$\alpha$}}

30: \newcommand{\kms}{\mbox{\,km\,s$^{-1}$}}

31: \newcommand{\cmsq}{\mbox{\,cm$^{-2}$}}

32: \newcommand{\cc}{\mbox{\,cm$^{-3}$}}

33: \newcommand{\flux}{\mbox{\,erg\,cm$^{-2}$\,s$^{-1}$}}

34: \newcommand{\flam}{\mbox{\,erg\,cm$^{-2}$\,s$^{-1}$\,\AA$^{-1}$}}

35: \newcommand{\fE}{\mbox{\,photon\,cm$^{-2}$\,s$^{-1}$\,keV$^{-1}$}}

36: \newcommand{\fnu}{\mbox{\,erg\,cm$^{-2}$\,s$^{-1}$\,Hz$^{-1}$}}

37: \newcommand{\lumin}{\mbox{\,erg~s$^{-1}$}}

38: \newcommand{\persec}{\mbox{\,s$^{-1}$}}

39: \newcommand{\nh}{\mbox{${N}_{\rm H}$}} 

40: \newcommand{\bj}{$B_{\rm J}$}

41: \newcommand{\ew}{EW$_{\rm a}$}

42: 

43: \newcommand{\aox}{$\alpha_{\rm ox}$}

44: \newcommand{\aoxp}{$\alpha^{\prime}_{\rm ox}$}

45: \newcommand{\aoxc}{$\alpha_{\rm ox}{\rm (corr)}$}

46: \newcommand{\aoxl}{$\alpha_{\rm ox}(l_{2500})$}

47: \newcommand{\fdeep}{$f_{\rm deep}$}

48: \newcommand{\auv}{$\alpha_{\rm UV}$}

49: \newcommand{\vmax}{$v_{\rm max}$}

50: \newcommand{\CIV}{\ion{C}{4}}

51: \newcommand{\SiIV}{\ion{Si}{4}}

52: \newcommand{\OVI}{\ion{O}{6}}

53: \newcommand{\MgII}{\ion{Mg}{2}}

54: \newcommand{\AlIII}{\ion{Al}{3}}

55: \newcommand{\OIII}{[\ion{O}{3}]}\newcommand{\HI}{\ion{H}{1}}

56: \newcommand{\Halpha}{H$\alpha$}

57: \newcommand{\delg}{$\Delta(g-i)$}

58: \newcommand{\daox}{$\Delta \alpha_{\rm ox}$}

59: \newcommand{\HR}{{HR}}

60: \newcommand{\GHR}{$\Gamma_{\rm HR}$}

61: \newcommand{\mrk}{{Mrk~231}}

62: \newcommand{\pgbal}{{PG~2112$+$059}}

63: \newcommand{\nim}{NIM}

64: \newcommand{\altcite}{\citealt}

65: \newcommand{\mathbold}{\boldmath}

66: \newcommand{\korista}{KVMW}

67: \newcommand{\weymann}{WMFH}

68: %%This paper

69: \newcommand{\total}{38}

70: \newcommand{\lbqstotal}{44}

71: \newcommand{\xdetected}{27}

72: \newcommand{\microns}{\micron}

73: \newcommand{\leight}{$l_{\rm 8\mu{m}}$}

74: \newcommand{\lfive}{$l_{\rm 5000\AA}$}

75: \newcommand{\ltwo}{$l_{\rm 2500\AA}$}

76: \newcommand{\alphao}{$\alpha_{\rm o}$}

77: \newcommand{\lsun}{$L_{\odot}$}

78: \newcommand{\msun}{$M_{\odot}$}

79: \newcommand{\hnot}{H$_{0}$}

80: \newcommand{\hunits}{\mbox{\kms\,Mpc$^{-1}$}}

81: \newcommand{\mdot}{\mbox{$M_{\odot}$ yr$^{-1}$}}

82: \newcommand{\rgsp}{Rev.~Geophys.~Space Phys.} % Reviews on Geophysics and Space Physics

83: 

84: 

85: 

86: 

87: %        **** DO NOT MODIFY BELOW ****   **** DO NOT MODIFY BELOW **** 

88: %

89: %

90: %-----------------------------------------------------------------------------

91: % LaTeX mark-ups for this paper

92: 

93: \newcommand{\todo}{\Huge{$\bullet$}\normalsize}

94: 

95: \setlength{\tabcolsep}{0.04in}

96: 

97: \begin{document}

98:  

99: %\received{date month year}

100: %\accepted{date month year}

101: %\journalid{number}{date month year}

102: %\articleid{number}{number}

103: %\slugcomment{submitted to: {\it The Astrophysical Journal Letters}}

104:  

105: \shortauthors{Markwick-Kemper et al.~}

106: \shorttitle{Dust in the wind of \pgbal}

107: 

108: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

109: \title{Dust in the wind: Crystalline silicates, corundum and periclase in \pgbal}

110: 

111: \author{F. Markwick-Kemper,\altaffilmark{1,2}

112: S.\ C. Gallagher,\altaffilmark{3}

113: D.\ C. Hines\altaffilmark{4}

114: \& J. Bouwman\altaffilmark{5}}

115: \altaffiltext{1}{School of Physics and Astronomy, University of Manchester, Sackville Street, PO Box 88, Manchester, M60 1QD, UK; {\em F.Markwick-Kemper@manchester.ac.uk}}

116: \altaffiltext{2}{University of Virginia, Department of Astronomy, PO Box

117:   400325, Charlottesville, VA 22904-4325, USA}

118: \altaffiltext{3}{Department of Physics \& Astronomy, University of

119:   California -- Los Angeles, 430 Portola Plaza, Box 951547, Los

120:   Angeles CA, 90095--1547, USA; {\em sgall@astro.ucla.edu}}

121: \altaffiltext{4}{Space Sciences Institute, 4750 Walnut Street, Suite 205

122: Boulder, CO 80301, USA; {\em dhines@as.arizona.edu}}

123: \altaffiltext{5}{Max-Planck-Institute for Astronomy, K\"onigstuhl 17, D-69117 Heidelberg, Germany; {\em bouwman@mpia.de}}

124: 

125: 

126: \begin{abstract}

127: We have determined the mineralogical composition of dust in the Broad

128: Absorption Line (BAL) quasar \pgbal\ using mid-infrared spectroscopy

129: obtained with the {\em Spitzer Space Telescope}. From spectral fitting

130: of the solid state features, we find evidence for Mg-rich amorphous

131: silicates with olivine stoichiometry, as well as the first detection

132: of corundum (Al$_2$O$_3$) and periclase (MgO) in quasars. This mixed

133: composition provides the first direct evidence for a clumpy density

134: structure of the grain forming region.  The silicates in total

135: encompass ($56.5 \pm 1.4$)\% of the identified dust mass, while

136: corundum takes up ($38 \pm 3$) wt.\%. Depending on the choice of

137: continuum, a range of mass fractions is observed for periclase ranging

138: from ($2.7 \pm 1.7$)\% in the most conservative case to ($9 \pm 2$)\% in

139: a less constrained continuum.  In addition, we identify a feature at

140: 11.2 $\mu$m as the crystalline silicate forsterite, with only a minor

141: contribution from polycyclic aromatic hydrocarbons. The ($5 \pm 3$)\%

142: crystalline silicate fraction requires high temperatures such as those

143: found in the immediate quasar environment in order to counteract rapid

144: destruction from cosmic rays.

145: 

146: \end{abstract}

147: \keywords{galaxies: active --- quasars: emission lines --- quasars:

148: individual (\pgbal) --- ISM: dust, extinction}

149: 

150: 

151: \section{Introduction}

152: \label{sec:intro}

153: 

154: In the local Universe, Asymptotic Giant Branch (AGB) stars are

155: believed to dominate dust production \citep[e.g.,][]{D_03_dust}.  AGB

156: stars are a late phase of stellar evolution for stars with initial

157: masses of $M$=1--8\msun, which develop outflows where the low

158: temperatures and high densities become optimal for dust condensation.

159: However, a significant amount of dust originating from AGB stars does

160: not build up within $\sim1$~Gyr of the birth of the first generation

161: of low-mass stars \citep{ME_03_earlydust}.  Nevertheless, large

162: quantities of dust are clearly present at these early times: quasar

163: host galaxies at $z\sim6$ show evidence for 10$^{8-9}$~\msun\ of dust

164: heated by star formation in their host galaxies as seen in submm and

165: far-infrared emission \citep{PriddeyEtal2003,BCB_06_dust}.

166: Furthermore, extinction curves for a $z=6.2$ quasar and GRB~050904

167: ($z=6.29$) are not consistent with those locally

168: \citep{maio04,stratta07}.  Another mechanism for dust production is

169: therefore required, and \citet{elvis-dust} explored the possibility

170: that quasar winds might provide environments suitable for efficient

171: dust formation.  They determined that the temperatures and pressures

172: in these regions could reach the values found in cool, dust-producing

173: stars.  The quasar winds are predicted to produce dust masses up to

174: $10^7$ \msun\ \citep{elvis-dust}, implying that the quasar wind may in

175: some cases only account for part of the dust production, and

176: supernovae have been suggested as an alternative source of dust in

177: high-z galaxies \citep{SEB_06_SNdust}.

178:  

179: Quasar outflows are most obviously manifested in Broad Absorption Line

180: (BAL) quasars.  This population, approximately 20\% of optically

181: selected type~1 quasar samples \citep[e.g.,][]{HewFol2003,reich+03a},

182: is notable for broad, blueshifted absorption evident in common

183: ultraviolet resonance transitions such as \CIV, Ly$\alpha$, and \OVI.

184: These P~Cygni-type features arise because the observer is looking

185: through an outflowing wind. \balqs\ are thus a natural population to

186: consider when investigating the grain properties of dust in the quasar

187: environment.  In this letter, we present a detailed analysis of the

188: mid-infrared spectrum of the luminous \balq, \pgbal, in order to

189: determine its dust composition.  \pgbal, an {\em IRAS} source, is

190: mid-infrared bright, and one of the most luminous low-redshift

191: ($z=0.466$) Bright Quasar Survey objects with $M_{\rm V}=-26.9$

192: (\hnot$=70$\hunits, $\Omega_{\rm M} = 0.3$, and $\Omega_{\Lambda} =

193: 0.7$ are assumed throughout).  \hst\ spectra revealed broad, shallow

194: \CIV\ absorption, and it has been well-studied in the UV and X-ray

195: \citep[e.g.,][]{gall+04}.

196: 

197: \section{Observations and Data Reduction}

198: 

199: \pgbal\ was observed on May 25, 2005, using the low resolution modules

200: of the InfraRed Spectrograph \citep[IRS;][]{HRV_04_IRS} on board

201: \spitzer\ \citep{WRL_04_Spitzer}, under AOR key 10949376.  We

202: performed 1 cycle of 240 sec.~integrations with the short low modules

203: (5.2--14.5 $\mu$m) and 2 cycles of 120 sec.\ with the long

204: low modules (14.0--38.0 $\mu$m).

205: 

206: We applied the data reduction, background subtraction and flux

207: calibration methods described in

208: \citet{BLD_06_etacha,BHH_submitted,SBA_submitted} to pipeline version

209: S14 of the data.  The resulting absolute flux is accurate to

210: $\sim$5\%, while the relative fluxes are accurate to $\sim$1.7\%.  The

211: resulting combined spectrum is shown in Fig.~\ref{fig:spectrum}.

212: 

213: \section{Analysis and Results}

214: 

215: The spectrum of \pgbal\ exhibits the broad emission bands at $\sim10$

216: and $\sim18$ $\mu$m generally ascribed to amorphous silicates

217: \citep[e.g.][]{D_03_dust}.  In

218: addition, the [Ne {\sc ii}] transition at 12.8 $\mu$m is visible, as

219: well as substructure in the solid state emission at $\sim11.2$ $\mu$m

220: (Fig.~\ref{fig:spectrum}).

221: 

222: \subsection{Continuum Determination}

223: \label{sec:cont}

224: 

225: Complementary to the IRS spectrum, we added mid-IR photometry from the 

226: literature as

227: compiled by NED\footnote{NASA/IPAC Extragalactic Database; {\tt

228: http://nedwww.ipac.caltech.edu}} (see Figure~\ref{fig:spectrum}).  In

229: wavelength regions in common, we found good agreement between (all

230: observed-frame wavelengths) 25\micron\ {\em IRAS} data

231: \citep{sanders+89}, 12 and 7.5\micron\ {\em ISO} data \citep{haas+00},

232: and 10\micron\ Palomar data \citep{neug+1987}.  To determine the

233: continuum underlying the dust emission features, we fit a power-law

234: model ($F_{\nu}=F_{\nu,0}\nu^{\alpha}$) using the photometric and

235: spectroscopic data (and uncertainties) from 5--8\microns\ and

236: 24.5--40\microns (rest-frame). As shown in Figure~\ref{fig:spectrum},

237: this provided a good fit to the continuum with

238: $\alpha=-0.617\pm0.004$.  We note that there is no evidence from

239: quasar composite spectral energy distributions for an inflection under

240: the dust emission features \citep[see][]{ric+06}, and a power-law

241: model fits the luminous quasar composite in this spectral regime quite

242: well.  Because the continuum is not well constrained at wavelengths

243: longer than the broad 18\micron\ emission feature, we performed a

244: second fit including the 13--14\micron\ IRS data at the dip between

245: the 10 and 18\micron\ emission features.  This pulled the continuum up

246: to the value of the 60\micron\ {\em IRAS} photometric datapoint plus

247: uncertainty for $\alpha=-0.674\pm0.003$.

248: 

249: \subsection{Spectral Feature Analysis}

250: \label{sec:feat}

251: 

252: Assuming the emission region is optically thin in the infrared, the

253: continuum-divided spectrum (Fig.~\ref{fig:fit}) yields the resulting

254: opacity of the dust. As at each wavelength, features and continuum

255: contributions are in equal amounts arising from the same temperature

256: components \citep[e.g.][]{SHK_05_quasar}, the temperature dependence

257: has been eliminated by dividing the the spectra by the continua. The

258: contribution of synchrotron emission to the continuum at these

259: wavelengths can be ignored \citep{HCM_98_quasars}.

260:  

261: We fitted the resulting opacity curves with laboratory spectroscopy of

262: minerals commonly observed in AGB stars, assuming that the resulting

263: opacity is a linear combination of the individual mineral opacities.

264: We also subtracted a spline-fit continuum between our fit boundaries

265: -- 8 and 25 $\mu$m -- of each mineral spectrum, to provide the same

266: baseline.  For each continuum divided spectrum, we ran 512 models with

267: varying dust compositions \citep[amorphous and crystalline olivine,

268: Al$_2$O$_3$, MgO, polycyclic aromatic hydrocarbons (PAHs); based on

269: prior observations and model calculations, see e.g.][]{D_03_dust} and

270: grain shape \citep[spherical or non-spherical;][]{BH_83_scattering},

271: to find the relative mass fractions in each component. The fit was

272: performed using $\chi^2$ minimization of each continuum-divided

273: spectrum, excluding the 12.8 $\mu$m [Ne{\sc ii}] line.

274: 

275: Fig.~\ref{fig:fit} shows the best fits for both continuum-divided

276: spectra, with $\Delta \chi^2$ values of 2.8 for continuum 1 and 2.6

277: for continuum 2.  If we allow a tolerance of a factor 2 for $\Delta

278: \chi^2$, we find classes of dust compositions containing 44 and 124

279: good fits; 30 and 47 of which require the presence of PAHs over the

280: 8--25 $\mu$m range, for the first and second continuum respectively.

281: 

282: The essential dust components in this class of fits are amorphous

283: olivine -- MgFeSiO$_4$ \citep{DBH_95_glasses} and/or Mg$_2$SiO$_4$

284: \citep{JDM_03_solgel} crystalline forsterite \citep{KTS_99_xsils},

285: corundum \citep[Al$_2$O$_3$;][]{BDH_97_12-17um} and periclase

286: \citep[MgO;][]{HKS_03_MgO}.  For the 20 $\mu$m excess in the spectrum

287: not explained by the silicates, we find that MgO in non-spherical

288: grains (represented by a common distribution of ellipsoids) provides a

289: good match. Both spherical and non-spherical grains give good results

290: for the amorphous olivines and corundum. Including grains with radii

291: $a \gtrsim 1$ $\mu$m greatly increased the $\Delta \chi^2$ of the

292: fits, and are therefore not likely to contribute to the observed

293: emission features. The fit results are summarized in

294: Table~\ref{tab:fitresults}.

295: 

296: The dust mixture is very similar for both continuum-divided spectra,

297: with the notable exception of the MgO mass fraction. In case of the

298: second continuum, where the 13-14 micron range was included in the

299: continuum fit, the MgO contribution is suppressed, but when the 13-14

300: micron range is --more realistically-- treated as dust feature

301: emission, and is not used to constrain the continuum, we find that

302: $\sim9$ wt.\% of the dust may be in the form of MgO.

303: 

304: To measure the contribution of PAHs to the emission, we used an

305: average interstellar profile for the PAHs from

306: \citet{HVP_01_OOPS}. Although they contribute to the 11.2 $\mu$m

307: feature, the PAHs do not reproduce its shape well, and additional

308: opacity from forsterite is needed.  We find that PAHs are responsible

309: for $\sim$30\% and $\sim$15\% for the first and second

310: continuum respectively.  This is consistent with the appearance of the

311: \pgbal\ spectrum in the 5--9 $\mu$m range (Fig.~\ref{fig:fit}).  The

312: 11.2 $\mu$m resonance in PAHs is likely due to neutral molecules,

313: while the features in the 5--9 $\mu$m range are predominantly carried

314: by ionized PAHs \citep[e.g.~][]{VPV_04_PAH}. A range of a factor

315: $\sim$6 in $I_{11.2}/I_{6.2}$ around the interstellar value is

316: observed in Galactic environments covering more than 4 orders of

317: magnitude in $G_0$ \citep{HVP_01_OOPS}. In case $G_0$ is

318: significantly stronger than the prevalent field in the Galaxy, 

319: $I_{11.2}/I_{6.2}$ decreases, further limiting the amount of PAHs

320: contributing to the 11.2 $\mu$m feature. 

321: 

322: \section{Dust composition}

323: 

324: The presence of silicates in local AGN has been known for many years

325: through resonant absorption at 9.7 and 18 $\mu$m

326: \citep[e.g.,][]{IU_00_cygA}, while more recently, silicate features

327: have also been detected in emission towards several quasars and AGN

328: \citep{SHK_05_quasar,SSL_05_emission,HSS_05_emission,SRH_06_AGN}.

329: 

330: While the spectral appearance of \pgbal\ in the infrared clearly

331: requires silicates, it is interesting that oxides appear to be present

332: as well. As a source of opacity, corundum and periclase cause the 18

333: $\mu$m feature to become stronger relative to the 9.7 $\mu$m

334: resonance, while at the same time, they do fill in the opacity trough

335: between the two silicate resonances \citep[see

336: also][]{JDM_03_solgel}. Corundum also causes a red wing on the 9.7

337: $\mu$m resonance, which fits well with the feature as observed in

338: \pgbal.  This is the first reported detection of periclase and

339: corundum as dust components in a quasar.

340: 

341: \subsection{Corundum}

342: 

343: The presence of Al$_2$O$_3$ is not surprising; it is commonly found in

344: the stellar winds of AGB stars \citep[e.g.~][]{SP_98_classification}.

345: In thermodynamic equilibrium, corundum (Al$_2$O$_3$) is expected to

346: condense at $\sim1500$ K, well above the stability limit for silicates

347: \citep{T_90_silmineralogy}. In a cooling gas newly condensed

348: Al$_2$O$_3$ grains will gradually be covered by silicate layers as the

349: temperature drops, provided the density of the gas does not decrease

350: too rapidly. In cases where the gas density has become too low for

351: significant dust growth once the gas temperature has reached the

352: stability limits for silicates, a \emph{freeze-out} of the dust

353: condensation sequence occurs, and Al$_2$O$_3$ remains the dominant

354: dust component. Indeed, AGB stars with low mass loss rates (10$^{-9}$

355: -- 10$^{-7}$ \mdot) contain significant fractions of Al$_2$O$_3$

356: \citep{BGO_06_bulge,HH_05_agb}, and a decrease in Al$_2$O$_3$ content

357: is observed with increasing mass loss rate for AGB stars in the LMC

358: \citep{DSR_05_lmc}.

359: 

360: Both the

361: luminosity and presence of silicates in emission in \pgbal\ (rather than

362: absorption; \citealt[e.g.,][]{smith07}) argue for geometrically

363: associating these grains with the quasar rather than the host galaxy

364: \citep{SRH_06_AGN}.  The presence of both corundum and olivines in the

365: spectrum indicate that there is a highly inhomogeneous -- clumpy --

366: density structure in the dust forming regions 

367: as expected from the broad-band spectral

368: modeling of \citet{nenkova02} and the theoretical work of

369: \citet{es06}.  As the gas travels outwards, certain pockets apparently

370: maintain high enough density throughout the cooling process that

371: silicates can be formed on top of the corundum, whereas in lower

372: density regions of the wind the dust condensation sequence gets

373: truncated after the formation of Al$_2$O$_3$. 

374: 

375: 

376: \subsection{Periclase}

377: 

378: Periclase (MgO) has its stability limits $\sim$50 K below the

379: stability limits of Mg-rich olivines ($\sim$1100 K), but just above

380: the evaporation temperature of Fe --

381: the temperature below which

382: Fe$^{2+}$ can be incorporated in the amorphous silicates

383: \citep{GS_99_condensation,FG_03_magnesiowuestite}.  In the

384: thermodynamic equilibrium of a slowly cooling gas, Mg$_2$SiO$_4$ would

385: form first, consuming all available gas-phase Mg in the process, thus

386: prohibiting the formation of MgO \citep{FG_03_magnesiowuestite}.  The

387: presence of solid MgO in \pgbal\ indicates that the dust forming gas

388: cooled rapidly to a temperature below these stability limits, 

389: allowing for the formation of MgO and Fe-containing silicates at the

390: same time as the formation of Mg$_2$SiO$_4$.

391: 

392: \subsection{Amorphous silicates}

393: 

394: The amorphous silicates in \pgbal\ have the stoichiometry of a 

395: Mg-rich olivine.  By using both

396: Mg$_2$SiO$_4$ and MgFeSiO$_4$ to fit the spectrum, we have determined

397: that the Fe-content of the amorphous silicates is low:

398: Mg$_{1.95}$Fe$_{0.05}$SiO$_4$, although using only Mg$_2$SiO$_4$ or

399: MgFeSiO$_4$ gives good results, too. The silicate grains are found to

400: be sub-micron-sized (at 10 $\mu$m the opacity of silicates is the same

401: for all grain sizes $\lesssim$ 0.5 $\mu$m), but the data did not allow

402: us to distinguish between spherical or non-spherical grains.  These

403: results compare well to the silicates in the Galactic interstellar

404: medium,which are composed of spherical olivine grains

405: \citep{KVT_04_GC}, or even more general, of non-spherical Mg-rich

406: olivines \citep{MWD_07_interstellar}.

407: 

408: Attempts to determine the silicate properties in active galaxies

409: include a fit to the 9.7 $\mu$m absorption feature seen in NGC~1068,

410: which shows evidence for the presence of Al-bearing species, either in

411: the form of aluminum-silicates or corundum, giving rise to a red wing

412: to the silicate resonance \citep{JMR_04_NGC1068}, while

413: \citet{SSL_05_emission} notice that the 9.7 and 18 $\mu$m emission

414: features seen in NGC~3998 have a different spectral appearance than

415: the silicates in the Galactic ISM. In particular, the 18 $\mu$m

416: feature is very strong with respect to the 9.7 $\mu$m feature, and the

417: peak positions of the resonances deviate as well. We suggest that in

418: NGC~3998, as well as in some quasars \citep{SRH_06_AGN}, the broad

419: `silicate' features might be more accurately modeled with a mixture of

420: silicates, corundum and MgO, not unlike \pgbal.

421: 

422: \subsection{Crystalline silicates}

423: 

424: A fraction of (5$\pm$3)\% by weight of the silicates is found to be

425: crystalline.  Evidence of crystalline silicates is not reported in any

426: other silicate emission spectra seen in quasars or AGN, though a

427: feature seen at 11.2 $\mu$m in some objects is identified with PAHs,

428: supported by the detection of additional PAH resonances at 6.2 and 7.7

429: $\mu$m \citep{schweitz06}. We only ascribe a small fraction of the

430: emission around 11.2 $\mu$m to PAHs (see Sect.~\ref{sec:feat}).

431: Crystalline silicates are not detected in the interstellar medium of

432: our own Galaxy \citep[$< 2.2$ wt.\%;][]{KVT_05_erratum}, due to

433: radiation damage caused by cosmic rays on a time scale of 40 Myr. In

434: quasars, the times scales may be even shorter if the cosmic ray flux

435: is higher, and crystalline silicates need to be reformed continuously

436: to explain the observed abundance. The formation of crystalline

437: silicates requires higher densities than amorphous silicates.  A

438: quasar wind origin may explain the presence of crystalline silicates

439: in a relatively small fraction (12 out 77) of ultraluminous infrared

440: galaxies \citep{STA_06_ulirgs}.

441: 

442: \section{Conclusion}

443: 

444: For the first time, the composition of the dust in a quasar has been

445: determined, albeit limited to the components with resonances in the

446: infrared. While the origin of the dust could lie in stellar ejecta,

447: the dust properties are consistent with their formation in the quasar

448: wind itself.  The dust in \pgbal\ clearly bears similarity to dust in

449: other astrophysical environments, i.e.~in the properties of the

450: amorphous silicates, but the presence of large amounts of MgO and

451: Al$_2$O$_3$ sets it apart from the composition of interstellar dust in

452: the local universe.

453: 

454: Assuming the dust is formed in the quasar

455: wind, the co-existence of the highly refractory corundum, the

456: crystalline silicates and and the less refractory MgO and amorphous

457: silicates indicates that the wind of \pgbal\ has an inhomogeneous

458: temperature and density structure.

459: 

460: \acknowledgements This work is based on observations made with the Spitzer

461: Space Telescope, which is operated by the Jet Propulsion Laboratory,

462: California Institute of Technology under a contract with NASA. Support for

463: this work was provided by NASA through an award issued by JPL/Caltech.

464: 

465: \facility{Spitzer(IRS)}

466: 

467: \begin{thebibliography}%{50}\setlength{\itemsep}{0mm}

468: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

469: \expandafter\ifx\csname url\endcsname\relax

470:   \def\url#1{{\tt #1}}\fi

471: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi

472: \providecommand{\eprint}[2][]{\url{#2}}

473: 

474: \bibitem[\protect\astroncite{{Beelen} et~al.}{2006}]{BCB_06_dust}

475: {Beelen}, A., {Cox}, P., {Benford}, D.~J., {Dowell}, C.~D., {Kov{\'a}cs}, A.,

476:   {Bertoldi}, F., {Omont}, A., \& {Carilli}, C.~L. 2006, \apj, 642, 694

477: 

478: \bibitem[\protect\astroncite{Begemann et~al.}{1997}]{BDH_97_12-17um}

479: Begemann, B., Dorschner, J., Henning, T., Mutschke, H., G\"urtler, J., K\"ompe,

480:   C., \& Nass, R. 1997, \apj, 476, 199

481: 

482: \bibitem[\protect\astroncite{Blommaert et~al.}{2006}]{BGO_06_bulge}

483: Blommaert, J. A. D.~L., et~al. 2006, \aap, 460, 555

484: 

485: \bibitem[\protect\astroncite{Bohren \& Huffman}{1983}]{BH_83_scattering}

486: Bohren, C.~F. \& Huffman, D.~R. 1983, Absorption and scattering of light by

487:   small particles (New York: Wiley)

488: 

489: \bibitem[\protect\astroncite{{Bouwman} et~al.}{2006}]{BLD_06_etacha}

490: {Bouwman}, J., {Lawson}, W.~A., {Dominik}, C., {Feigelson}, E.~D., {Henning},

491:   T., {Tielens}, A.~G.~G.~M., \& {Waters}, L.~B.~F.~M. 2006, \apjl, 653, L57

492: 

493: \bibitem[\protect\astroncite{Bouwman et~al.}{2007}]{BHH_submitted}

494: Bouwman, J., et~al. 2007, \apj, submitted

495: 

496: \bibitem[\protect\astroncite{Dijkstra et~al.}{2005}]{DSR_05_lmc}

497: Dijkstra, C., Speck, A.~K., Reid, R.~B., \& Abraham, P. 2005, \apjl, 633, L133

498: 

499: \bibitem[\protect\astroncite{Dorschner et~al.}{1995}]{DBH_95_glasses}

500: Dorschner, J., Begemann, B., Henning, T., J\"ager, C., \& Mutschke, H. 1995,

501:   \aap, 300, 503

502: 

503: \bibitem[\protect\astroncite{Draine}{2003}]{D_03_dust}

504: Draine, B.~T. 2003, \araa, 41, 241

505: 

506: \bibitem[\protect\astroncite{{Elitzur} \& {Shlosman}}{2006}]{es06}

507: {Elitzur}, M. \& {Shlosman}, I. 2006, \apjl, 648, L101

508: 

509: \bibitem[\protect\astroncite{{Elvis} et~al.}{2002}]{elvis-dust}

510: {Elvis}, M., {Marengo}, M., \& {Karovska}, M. 2002, \apjl, 567, L107

511: 

512: \bibitem[\protect\astroncite{Ferrarotti \& Gail}{2003}]{FG_03_magnesiowuestite}

513: Ferrarotti, A.~S. \& Gail, H.~P. 2003, \aap, 398, 1029

514: 

515: \bibitem[\protect\astroncite{Gail \& Sedlmayr}{1999}]{GS_99_condensation}

516: Gail, H.-P. \& Sedlmayr, E. 1999, \aap, 347, 594

517: 

518: \bibitem[\protect\astroncite{{Gallagher} et~al.}{2004}]{gall+04}

519: {Gallagher}, S.~C., {Brandt}, W.~N., {Wills}, B.~J., {Charlton}, J.~C.,

520:   {Chartas}, G., \& {Laor}, A. 2004, \apj, 603, 425

521: 

522: \bibitem[\protect\astroncite{{Haas} et~al.}{1998}]{HCM_98_quasars}

523: {Haas}, M., {Chini}, R., {Meisenheimer}, K., {Stickel}, M., {Lemke}, D.,

524:   {Klaas}, U., \& {Kreysa}, E. 1998, \apjl, 503, L109

525: 

526: \bibitem[\protect\astroncite{{Haas} et~al.}{2000}]{haas+00}

527: {Haas}, M., {M{\"u}ller}, S.~A.~H., {Chini}, R., {Meisenheimer}, K., {Klaas},

528:   U., {Lemke}, D., {Kreysa}, E., \& {Camenzind}, M. 2000, \aap, 354, 453

529: 

530: \bibitem[\protect\astroncite{Hao et~al.}{2005}]{HSS_05_emission}

531: Hao, L., et~al. 2005, \apjl, 625, L75

532: 

533: \bibitem[\protect\astroncite{Heras \& Hony}{2005}]{HH_05_agb}

534: Heras, A.~M. \& Hony, S. 2005, \aap, 439, 171

535: 

536: \bibitem[\protect\astroncite{{Hewett} \& {Foltz}}{2003}]{HewFol2003}

537: {Hewett}, P.~C. \& {Foltz}, C.~B. 2003, \aj, 125, 1784

538: 

539: \bibitem[\protect\astroncite{Hofmeister et~al.}{2003}]{HKS_03_MgO}

540: Hofmeister, A.~M., Keppel, E., \& Speck, A.~K. 2003, \mnras, 345, 16

541: 

542: \bibitem[\protect\astroncite{Hony et~al.}{2001}]{HVP_01_OOPS}

543: Hony, S., {Van Kerckhoven}, C., Peeters, E., Tielens, A. G. G.~M., Hudgins,

544:   D.~M., \& Allamandola, L.~J. 2001, \aap, 370, 1030

545: 

546: \bibitem[\protect\astroncite{Houck et~al.}{2004}]{HRV_04_IRS}

547: Houck, J.~R., et~al. 2004, \apjs, 154, 18

548: 

549: \bibitem[\protect\astroncite{Imanishi \& Ueno}{2000}]{IU_00_cygA}

550: Imanishi, M. \& Ueno, S. 2000, \apj, 535, 626

551: 

552: \bibitem[\protect\astroncite{Jaffe et~al.}{2004}]{JMR_04_NGC1068}

553: Jaffe, W., et~al. 2004, \nat, 429, 47

554: 

555: \bibitem[\protect\astroncite{J\"ager et~al.}{2003}]{JDM_03_solgel}

556: J\"ager, C., Dorschner, J., Mutschke, H., Posch, T., \& Henning, T. 2003, \aap,

557:   408, 193

558: 

559: \bibitem[\protect\astroncite{Kemper et~al.}{2004}]{KVT_04_GC}

560: Kemper, F., Vriend, W.~J., \& Tielens, A. G. G.~M. 2004, \apj, 609, 826

561: 

562: \bibitem[\protect\astroncite{Kemper et~al.}{2005}]{KVT_05_erratum}

563: --- 2005, \apj, 633, 534

564: 

565: \bibitem[\protect\astroncite{Koike et~al.}{1999}]{KTS_99_xsils}

566: Koike, C., Tsuchiyama, A., \& Suto, H. 1999, in Proc.~32nd ISAS Lunar

567:   Planet.~Symp., 175--178

568: 

569: \bibitem[\protect\astroncite{{Maiolino} et~al.}{2004}]{maio04}

570: {Maiolino}, R., {Schneider}, R., {Oliva}, E., {Bianchi}, S., {Ferrara}, A.,

571:   {Mannucci}, F., {Pedani}, M., \& {Roca Sogorb}, M. 2004, \nat, 431, 533

572: 

573: \bibitem[\protect\astroncite{Min et~al.}{2007}]{MWD_07_interstellar}

574: Min, M., Waters, L. B. F.~M., {de Koter}, A., Hovenier, J.~W., Keller, L.~P.,

575:   \& {Markwick-Kemper}, F. 2007, \aap, 462, 667

576: 

577: \bibitem[\protect\astroncite{{Morgan} \& {Edmunds}}{2003}]{ME_03_earlydust}

578: {Morgan}, H.~L. \& {Edmunds}, M.~G. 2003, \mnras, 343, 427

579: 

580: \bibitem[\protect\astroncite{{Nenkova} et~al.}{2002}]{nenkova02}

581: {Nenkova}, M., {Ivezi{\'c}}, {\v Z}., \& {Elitzur}, M. 2002, \apjl, 570, L9

582: 

583: \bibitem[\protect\astroncite{{Neugebauer} et~al.}{1987}]{neug+1987}

584: {Neugebauer}, G., {Green}, R.~F., {Matthews}, K., {Schmidt}, M., {Soifer},

585:   B.~T., \& {Bennett}, J. 1987, \apjs, 63, 615

586: 

587: \bibitem[\protect\astroncite{{Priddey} et~al.}{2003}]{PriddeyEtal2003}

588: {Priddey}, R.~S., {Isaak}, K.~G., {McMahon}, R.~G., {Robson}, E.~I., \&

589:   {Pearson}, C.~P. 2003, \mnras, 344, L74

590: 

591: \bibitem[\protect\astroncite{{Reichard} et~al.}{2003}]{reich+03a}

592: {Reichard}, T.~A., et~al. 2003, \aj, 125, 1711

593: 

594: \bibitem[\protect\astroncite{{Richards} et~al.}{2006}]{ric+06}

595: {Richards}, G.~T., et~al. 2006, \apjs, 166, 470

596: 

597: \bibitem[\protect\astroncite{{Sanders} et~al.}{1989}]{sanders+89}

598: {Sanders}, D.~B., {Phinney}, E.~S., {Neugebauer}, G., {Soifer}, B.~T., \&

599:   {Matthews}, K. 1989, \apj, 347, 29

600: 

601: \bibitem[\protect\astroncite{{Schweitzer} et~al.}{2006}]{schweitz06}

602: {Schweitzer}, M., et~al. 2006, \apj, 649, 79

603: 

604: \bibitem[\protect\astroncite{Shi et~al.}{2006}]{SRH_06_AGN}

605: Shi, Y., et~al. 2006, \apj, 653, 127

606: 

607: \bibitem[\protect\astroncite{Siebenmorgen et~al.}{2005}]{SHK_05_quasar}

608: Siebenmorgen, R., Haas, M., Kr\"ugel, E., \& Schulz, B. 2005, \aap, 436, L5

609: 

610: \bibitem[\protect\astroncite{Sloan \& Price}{1998}]{SP_98_classification}

611: Sloan, G.~C. \& Price, S.~D. 1998, \apjs, 119, 141

612: 

613: \bibitem[\protect\astroncite{{Smith} et~al.}{2007}]{smith07}

614: {Smith}, J.~D.~T., et~al. 2007, \apj, 656, 770

615: 

616: \bibitem[\protect\astroncite{Spoon et~al.}{2006}]{STA_06_ulirgs}

617: Spoon, H. W.~W., et~al. 2006, \apj, 638, 759

618: 

619: \bibitem[\protect\astroncite{{Stratta} et~al.}{2007}]{stratta07}

620: {Stratta}, G., {Maiolino}, R., {Fiore}, F., \& {D'Elia}, V. 2007, \apjl, 661,

621:   L9

622: 

623: \bibitem[\protect\astroncite{Sturm et~al.}{2005}]{SSL_05_emission}

624: Sturm, E., et~al. 2005, \apjl, 629, L21

625: 

626: \bibitem[\protect\astroncite{{Sugerman} et~al.}{2006}]{SEB_06_SNdust}

627: {Sugerman}, B.~E.~K., et~al. 2006, Science, 313, 196

628: 

629: \bibitem[\protect\astroncite{{Swain} et~al.}{2007}]{SBA_submitted}

630: {Swain}, M.~R., {Bouwman}, J., {Akeson}, R., {Lawler}, S., \& {Beichman}, C.

631:   2007, \eprint{astro-ph/0702593}

632: 

633: \bibitem[\protect\astroncite{Tielens}{1990}]{T_90_silmineralogy}

634: Tielens, A. G. G.~M. 1990, in From Miras to Planetary Nebulae: which path for

635:   stellar evolution?, eds. M.~O. Mennessier \& A.~Omont, 186--200

636: 

637: \bibitem[\protect\astroncite{{van Diedenhoven} et~al.}{2004}]{VPV_04_PAH}

638: {van Diedenhoven}, B., Peeters, E., {Van Kerckhoven}, C., Hony, S., Hudgins,

639:   D.~M., Allamandola, L.~J., \& Tielens, A. G. G.~M. 2004, \apj, 611, 928

640: 

641: \bibitem[\protect\astroncite{Werner et~al.}{2004}]{WRL_04_Spitzer}

642: Werner, M.~W., et~al. 2004, \apjs, 154, 1

643: 

644: \end{thebibliography}

645: 

646: \clearpage

647: 

648: \begin{deluxetable}{lccc}

649: \tablecolumns{4}

650: \tablewidth{0pc}

651: \tablecaption{Best $\Delta \chi^2$ fits to the two continuum-divided spectra

652: of \pgbal\ (Fig.~\ref{fig:fit}). The first column shows the dust fraction

653: considered, and the following columns show the mass fractions for

654: both continuum fits and the average. The standard deviation on the results

655: is determined from the spread of the mass fractions observed in 

656: those fits within the $\Delta \chi^2$ tolerance of a factor 2. 

657: The bottom line shows the average composition of the amorphous silicate

658: dust. \label{tab:fitresults}}

659: \tablehead{

660: \colhead{ } & \colhead{continuum 1} & \colhead{continuum 2} & \colhead{average}\\

661: \colhead{ } & \colhead{wt.\%} & \colhead{wt.\%} & \colhead{wt.\%}}

662: \startdata

663: silicates/total & ($54 \pm 1$)   & ($59 \pm 1$)   & ($56.5 \pm 1.4$)   \\

664: corundum/total & ($37 \pm 2$)   & ($39 \pm 2$)   & ($38 \pm 3$)  \\

665: MgO/total & ($9 \pm 2$)   &  ($2.7 \pm 1.7$)   & ($5.9 \pm 2.6$)  \\

666: crystalline/silicates  & ($4 \pm 2$)   & ($6 \pm 2$)   & ($5 \pm 3$)  \\

667: Mg$_2$SiO$_4$/am.~silicates & ($70 \pm 44$)   &  ($49 \pm 40$)   & \nodata \\

668: MgFeSiO$_4$/am.~silicates &  ($30 \pm 44$)   & ($51 \pm 40$)   & \nodata \\

669: \hline

670: am.~sil.~composition & Mg$_{1.9}$Fe$_{0.1}$SiO$_4$ & Mg$_{1.8}$Fe$_{0.2}$SiO$_4$ & Mg$_{1.85}$Fe$_{0.15}$SiO$_4$\\

671: \enddata

672: \end{deluxetable}

673: 

674: \clearpage

675: 

676: \begin{figure}

677: \plotone{f1.eps}

678: \caption{({\em Top panel}) Mid-infrared spectrum of PG2112+059; the

679:   abscissas are rest-frame values.  The diamonds are photometric data

680:   from the following sources (from left to right, given in

681:   observed-frame units): {\em IRAS} 60 and 25 \micron\ \citep{sanders+89}, 

682:   {\em ISO} 12 \micron\ \citep{haas+00}, Palomar 10

683:   \micron\ \citep{neug+1987}, {\em ISO} 7.5 \micron\ \citep{haas+00}, 

684:   Palomar 3.5\micron\ \citep[two observations;][]{neug+1987}.  

685:   Shown is the IRS spectrum (with error bars), and

686:   the dashed and thin solid curves are two power-law continuum models

687:   ({\em Bottom panel}) The IRS spectrum minus the dashed continuum fit

688:   (continuum 1; thick solid histogram) and thin solid continuum fit (

689:   continuum 2; thin

690:   histogram) shown in the top panel.  The crosses indicate the

691:   frequencies of the data used in the dashed continuum fit; for the

692:   second continuum fit, the IRS data from 13--14 \micron\

693:   ([2.30--2.14]$\times10^{13}$ Hz) were also included.}

694: \label{fig:spectrum}

695: \end{figure}

696: 

697: \clearpage

698: 

699: \begin{figure}

700: \includegraphics{f2.eps}

701: \caption{Best $8-25$ $\mu$m fits, from our suite of good fits, for the

702: continuum-divided spectrum of \pgbal\ (offset by -1) using the first and second

703: choice of continuum (top and bottom panel, respectively). The thick

704: solid line is the total fit to the data (histogram, with error bars)

705: composed of amorphous olivine (dotted line), forsterite (dashed line),

706: corundum (Al$_2$O$_3$, dash-dotted line), MgO (dash-triple-dotted

707: line), and the mean interstellar PAH spectrum (solid line; only in

708: panel 1).  The inset in panel 1 shows the strength of PAH features

709: expected in the 5.5--8.5 $\mu$m range for a radiation field comparable

710: to that of the Galaxy.}

711: \label{fig:fit}

712: \end{figure}

713: 

714: 

715: 

716: 

717: \end{document}

718: