1: \documentclass[12pt,preprint2]{aastex}
2:
3: \usepackage{epsfig}
4:
5: \shorttitle{Carbonates in space}
6: \shortauthors{Posch et al.}
7:
8: \begin{document}
9:
10: \title{Carbonates in space -- \\
11: The challenge of low temperature data\footnote{Based on observations
12: with ISO, an ESA project with instruments funded
13: by ESA Member States (especially the PI countries: France,
14: Germany, the Netherlands and the United Kingdom) and with the
15: participation of ISAS and NASA.}}
16: \author{Th. Posch, A. Baier}
17: \affil{Institut f\"ur Astronomie, T\"urkenschanzstra{\ss}e 17, A-1180 Wien, Austria}
18: \email{posch@astro.univie.ac.at}
19:
20: \and
21:
22: \author{H. Mutschke}
23: \affil{Astrophysikalisches Institut,
24: Schillerg\"asschen 2-3, D-07745 Jena, Germany}
25:
26: \and
27:
28: \author{Th. Henning}
29: \affil{Max-Planck-Institut f\"ur Astronomie (MPIA), K\"onigstuhl 17,
30: D-69117 Heidelberg, Germany}
31:
32:
33: %------------------------------------------------------------------------------
34:
35: \begin{abstract}
36:
37: Carbonates have repeatedly been discussed as possible carriers of stardust emission bands.
38: However, the band assignments proposed so far were mainly based on room
39: temperature powder transmission spectra of the respective minerals.
40: Since very cold calcite grains have been claimed to be present in
41: protostars and in Planetary Nebulae such as NGC 6302, the changes
42: of their dielectric functions at low temperatures are relevant from
43: an astronomical point of view.
44:
45: We have derived the IR optical constants of calcite and dolomite from
46: reflectance spectra -- measured at 300, 200, 100 and 10\,K -- and calculated
47: small particle spectra for different grain shapes, with the following results:
48: i) The absorption efficiency factors both of calcite and dolomite are extremely
49: dependent on the particle shapes. This is due to the high peak values of the
50: optical constants of CaCO$_3$ and CaMg[CO$_3$]$_2$.
51: ii) The far infrared properties of calcite and dolomite
52: depend also very significantly on the temperature. Below 200\,K,
53: a pronounced sharpening and increase in the band strengths of the FIR resonances occurs.
54: iii) In view of the intrinsic strength and sharpening of the $\sim$44\,$\mu$m
55: band of calcite at 200--100\,K, the absence of this band -- inferred from
56: Infrared Space Observatory data -- in PNe requires dust temperatures below
57: 45\,K.
58: iv) Calcite grains at such low temperatures can account for the
59: `92'\,$\mu$m band, while our data rule out dolomite as the carrier
60: of the 60--65\,$\mu$m band.
61:
62: The optical constants here presented are publicly available
63: in the electronic database
64: {\texttt{http://www.astro.uni-jena.de/Laboratory/OCDB}.}
65:
66:
67: \end{abstract}
68:
69: \keywords{circumstellar matter --- infrared: stars --- methods: laboratory ---
70: stars: AGB and post-AGB --- stars: atmospheres}
71:
72:
73: %------------------------------------------------------------------------------
74:
75:
76: \section{Introduction}
77:
78: In the 1970s, carbonates have been proposed as carriers of mid-
79: infrared features detected in astronomical spectra, such as the 11.3\,$\mu$m
80: emission band seen in NGC 7027 and in the 'Red Rectangle'
81: (Gillet, Forrest \& Merrill 1973; Bregman \& Rank 1975;
82: \cite{Breg77}).
83:
84: In parallel to that, reflectance spectra of calcite (CaCO$_3$) and magnesite
85: (MgCO$_3$) have been measured in the laboratory. \cite{H70} did the measurements
86: in the whole mid- and far-IR range for calcite, magnesite and dolomite
87: (CaMg[CO$_3$]$_2$)), while \cite{P76} examined calcite and magnesite
88: in the 5--50\,$\mu$m region. All these measurements have been carried
89: out at room temperature.
90:
91: Carbonate grains have been found in interplanetary dust particles (IDPs).
92: \cite{San86} and \cite{TB86} reported the identification of carbonate minerals
93: in IDPs both by a chemical dissolution experiment and by transmission electron
94: microscopy. As shown by \cite{TB86}, the mineralogy of the IDP carbonates is
95: essentially that of breunnerite -- an Mg-rich solid solution of magnesite
96: and siderite (FeCO$_3$). Solid solutions of Ca- and Mg-carbonates are rare
97: in the IDPs examined so far (see also \cite{JB01}). It is noteworthy that
98: the IDPs show a 6.8\,$\mu$m absorption feature which is probably due
99: to the stretching vibration of carbonate anions (Sandford 1986).
100: It has to be noted however, that there are doubts on this identification
101: because the carbonate band at 11.4\,$\mu$m was not detected in a
102: study on five IDPs (Quirico et al.\ 2000). A band similar to the 6.8\,$\mu$m
103: feature is supposedly seen in the spectra of the ``Deep Impact'' ejecta
104: from comet 9P/Tempel 1 measured with the {\em Spitzer}\/ IRS spectrometer
105: (Lisse et al.\ 2006). In these spectra, also bands at 11-12 and
106: 13-14\,$\mu$m have been assigned to carbonates.
107:
108: While the identification of the 11.3\,$\mu$m band with carbonates
109: (\cite{Breg77}) has been confuted by McCarthy, Forrest \& Houck (1978)
110: and subsequently been replaced by an identification
111: with hydrocarbons (\cite{Coh86}), there has been continuous and recently
112: even increased interest in carbonates as stardust and planetary dust
113: components. Carbonate bands at wavelengths larger than 25\,$\mu$m came
114: to the fore:
115: \cite{Er00} and \cite{Lel00} suggested calcite and siderite (FeCO$_3$) as
116: carriers of aerosol absorption bands of the Martian atmosphere {\em inter
117: alia}\/ at 26.5, 30--31 and 43.5\,$\mu$m.
118:
119: \cite{K02aunda} and \cite{K02nat}, by a careful analysis of the 2.4--200\,$\mu$m
120: data of the Planetary Nebula NGC 6302 taken with the Infrared Space Observatory
121: (ISO), were able to show that the previously unidentified `92'\,$\mu$m emission
122: band and the 62\,$\mu$m component of the 60\,$\mu$m emission complex in this
123: nebula can possibly be attributed to calcite and dolomite, respectively. According to their model fit, the carbonates belong to the cold (30--60\,K)
124: dust component of this nebula.
125:
126: \cite{Chia05} studied the 60--180\,$\mu$m ISO spectra of more than 30 protostars
127: and detected a solid state feature peaking between 90 and 110\,$\mu$m in 17
128: of these spectra. Again, calcite has been suggested as the feature carrier,
129: even though the accompanying characteristic resonance features of calcite
130: could not be detected in the respective sources for different reasons
131: (partly related to the sensitivity of ISO's detectors). For two star forming
132: regions -- the Carina Nebula and Sharpless 171 -- this band had already been
133: discovered by \cite{OO03} but had been ascribed to onion structured carbon grains.
134: The variation in the band position has been explained by Chiavassa et al.\
135: by varying amounts of other than calcium ions (e.g.\ Mg or Fe ions)
136: in the carbonatic carrier.
137:
138: The astronomical observations summarized so far indicate a need for more detailed investigations in two respects: (a) measurement of IR properties below 0\,K and (b) investigation of variations in the band profiles due to particle shape and size effects. Both is best achieved on the basis of determining optical constants for low temperatures. Furthermore, this approach makes it possible to get rid of the possible influence of the matrix effect which may have influenced part of the previous comparisons between laboratory and observational data on carbonates.
139:
140: Given that it is unclear how carbonates can form in a circumstellar
141: environment, \cite{FG05} examined the formation of calcium carbonate in an
142: oxygen-rich stellar wind (C/O $<$ 1) by thermodynamic calculations, assuming the
143: formation of calcite from Ca, H$_2$O and CO. They came to the conclusion
144: that calcite -- if at all -- would rather form in Asymptotic Giant Branch
145: (AGB) stars with {\em low}\/ mass loss rates (approximately 1 $\times$ 10$^{-6}$ M$_{\odot}$/y) than in AGB stars in their superwind phase (of which planetary nebulae like NGC 6302 are the remnants however). But also in low-mass-loss-rate AGB stars, less than 1\% of the total amount of dust would be carbonate dust. According to \cite{FG05}, it is mainly the relatively low condensation temperature of carbonates which makes their condensation in circumstellar environments so inefficient: for this temperature range to be reached in an expanding circumstellar shell, the dust must have a substantially larger
146: distance from the central star (and, accordingly, a substantially
147: reduced material density) than for the silicates.
148:
149: { Toppani et al.\ (2005), however, suggest a possibly (more) efficient
150: formation mechanism of carbonates in the Post-AGB phase of stellar
151: evolution. Based on laboratory experiments, they predict carbonate formation
152: by reactions between a comparatively hot silicate gas and a 300-500\,K
153: H$_2$O-CO$_2$-rich gas under conditions far from local thermal equilibrium
154: (LTE). They assume this non-LTE carbonate formation to take place when high-velocity winds of forming PNe interact with dense remnants of the
155: slow-wind-AGB phase.}
156:
157: A major goal of our study is to contribute to an assessment of the validity
158: of the `carbonate hypotheses' from the spectroscopic point of view.
159:
160: The structure of the present paper is the following. In \S 2, we describe the derivation of optical constants of CaCO$_3$ and CaMg[CO$_3$]$_2$ from reflectance measurements, both for room temperature and for cryogenic temperatures. In \S 3, absorption efficiency spectra of small calcite and dolomite grains are presented. Finally, in \S 4, we compare these small particle
161: spectra with infrared bands of astronomical objects which have been attributed
162: to carbonates in recent studies. For the case of calcite, we present new
163: constraints on the particle shapes and temperatures which are necessary to
164: reconcile the `92'\,$\mu$m band assignments by \cite{K02aunda} and \cite{Chia05}
165: with currently available laboratory data.
166:
167:
168: %------------------------------------------------------------------------------
169:
170:
171: \section{Reflectance spectroscopy and de\-rivation of the optical
172: constants \label{s:OC}}
173:
174: \subsection{Crystallographic properties and sample preparation}
175:
176: % 1.3. 2005
177: Calcite and dolomite both belong to the trigonal crystal class; hence, they
178: are both uniaxial, and polarized reflectance measurements are required to
179: determine their optical constants.\footnote{Interestingly, the phenomenon of
180: {\em birefringence}\/ has been first observed in calcite crystals.}
181: The space group of calcite is R$\bar{3}$c, while dolomite belongs to the
182: space group R$\bar{3}$ and has a reduced symmetry compared to its chemically
183: simpler relative. While the calcium ions occupy
184: crystallographically equivalent sites in CaCO$_3$, half of the places occupied
185: by Ca ions in calcite are occupied by Mg ions in dolomite. The reduced symmetry
186: of dolomite compared to calcite leads to the occurrence of an additional
187: IR band (which is located close to 33\,$\mu$m) in its spectrum.
188: For more details on the lattice structure of Mg and Ca carbonates, see
189: \cite{H70} and references therein (e.g. \cite{W64}).
190:
191: In order to perform polarized reflectance spectroscopy, planes containing the
192: respective crystal's c-axes need to be found or to be produced. This
193: is a difficult task in the case of ordinary CaCO$_3$ or CaMg[CO$_3$]$_2$
194: rhombohedra, since none of their surfaces fulfill this condition. The c-axis
195: runs diagonally through the obtuse corners of cleaved rhombohedra.
196:
197: In the case of our dolomite rhombohedron -- a 2$\times$2$\times$2\,cm$^3$
198: crystal from Eugui, Navarra, Spain --, it was therefore necessary to cut it in order
199: to produce a c-parallel surface, which has been polished subsequently.
200: Polarized reflectance measurements on the polished surfaces showed no `mixing'
201: of the ordinary and extraordinary ray, which proved the cut to be parallel
202: to the c-axis indeed.
203: In the case of our calcite crystal -- a 4$\times$2$\times$1\,cm$^3$
204: fragment of a prism found at Sweetwater, Missouri, USA -- no cut was necessary
205: because this individual crystal had a relatively rare stature, its long edges
206: being grown parallel to the symmetry axis. Hence, polishing one of the lateral
207: surfaces of this crystal was sufficient to obtain a large measuring spot
208: containing the c-axis. Indeed, polarized spectroscopy on that spot did not
209: show significant traces of mixing of the two principal orientations of the electric
210: field vector relative to c (E$\|$c, E$\bot$c; see also below, Fig.\
211: \ref{f:Ref_Calc}).
212:
213: %..............................................................................
214:
215:
216: \subsection{Reflectance spectroscopy}
217:
218: The reflectance spectra of our two samples have been measured
219: at near-normal incidence using an FTIR spectrometer Bruker 113v in the
220: wavelength range 2--120\,$\mu$m at a resolution of 2\,cm$^{-1}$. Gold
221: mirrors have been used as references. A turnable grid polarizer on
222: polyethylene substrate positioned at the entrance of the sample compartment
223: was used to select polarization directions parallel and normal to the
224: plane of incidence.
225:
226: \begin{figure*}[htbp]
227: \epsscale{1.45}
228: \plotone{f1.eps}
229: \caption{Polarized reflectance spectra of calcite for T\,=\,300 K
230: (solid line), 200 K (dashed), 100 K (dash-dotted) and
231: 10 K (dotted line).
232: The spectra for E$\bot$c were shifted vertically by an
233: amount of 1.0.
234: \label{f:Ref_Calc}}
235: \end{figure*}
236:
237: Fig.\ \ref{f:Ref_Calc} shows reflectance spectra of calcite in the
238: 5--120\,$\mu$m range which comprises all IR bands of this mineral. In the
239: case of the reflectance for the ordinary ray (E$\bot$c), an offset of 1.0
240: has been added to all reflectance values. The different linestyles correspond
241: to the different temperatures (10--300\,K) as indicated.
242:
243: For the ordinary ray, five reflection bands can be discerned in the
244: infrared spectrum of calcite, which are centered at 6.8, 14.0, 30.7,
245: 44.0 and 91.4\,$\mu$m for room temperature.
246: For the extraordinary ray, the respective reflectance band positions
247: amount to 11.4, 29.6 and 91.4\,$\mu$m (again for room temperature).
248:
249: \begin{figure*}[htbp]
250: \epsscale{1.45}
251: \plotone{f2.eps}
252: \caption{Polarized reflectance spectra of dolomite for the same
253: series of temperatures as in the case of calcite.
254: \label{f:Ref_Dol}}
255: \end{figure*}
256:
257: In Fig.\ \ref{f:Ref_Dol}, the reflectance spectra of dolomite are plotted
258: on the same scale as for calcite. Dolomite has a slightly more complex
259: band pattern than calcite, which is a consequence of its comparatively
260: reduced lattice symmetry mentioned above. The reflectance maxima are located
261: at 6.6, 13.8, 25.5, 38.4 and 61.7\,$\mu$m for the ordinary ray (at T = 300\,K),
262: while for the extraordinary ray, the respective positions are 11.3, 25.4, 31.6,
263: 56.7 and 60.0\,$\mu$m.
264: %Both for dolomite and calcite, all bands that have
265: %been studied by separate cryogenic measurements shift to shorter wavelenghs
266: %and become narrower as the samples are cooled down.
267:
268: % 7.2. / 30.3. 2006
269: For the cryogenic measurements, the experimental setup was the following:
270: the samples have been placed into a continuous-flow liquid-helium
271: cryostat with contact-gas cooling (CryoVac KONTI Spectro B)
272: equipped with polyethylene windows. It was inserted into the FTIR
273: spectrometer's sample compartment at the same position where the room
274: temperature reflection measurements had been performed. The cryostat
275: windows did not allow measurements at wavelengths shorter than 15\,$\mu$m.
276: The cryostat possesses two sample mounts which can alternatingly be
277: moved into the infrared beam, thus allowing subsequent measurements on
278: the sample and the reference mirror at each temperature. A third measurement
279: with the sample being tilted to suppress the reflection was necessary
280: (again at each temperature) in order to determine the reflectivity of
281: the cryostat windows (``zero signal''), which had to be subtracted from
282: both the sample and the reference spectra. The temperature was controlled
283: by means of a CryoVac temperature controller using an ohmic heater and
284: {two silicon diode temperature sensors connected to the liquid-helium cooled
285: walls of the contact gas chamber and to the sample mount, respectively.}
286: The temperatures have been stable to about $\pm$0.5~K during the measurements.
287:
288: Spectra have been taken at 200\,K, 100\,K, and 10\,K during cooling down the samples.
289: Two cooling cycles have been carried out for each sample with the two polarizer
290: orientations. The wavelength- and temperature-dependent reflectances have been
291: calculated by dividing the ``zero-corrected'' sample and reference spectra at each
292: temperature.
293:
294: It is found that both for dolomite and calcite all bands in the above
295: indicated wavelength range undergo a shift to shorter wavelengths,
296: a bandwidth decrease and band strength increase with decreasing temperatures.
297: These effects are especially significant for calcite's 44\,$\mu$m reflectance
298: band and for dolomite's 38\,$\mu$m reflectance band. A more detailed discussion
299: is given in \S 3.
300:
301: %.............................................................................
302:
303:
304: \subsection{Lorentz oscillator fits and derived optical constants}
305:
306: The polarized reflectance spectra of calcite and dolomite can be quite
307: accurately reproduced by Lorentz oscillator fits, as has been shown, e.g.,
308: by \cite{H70}. Using the ansatz
309:
310: \begin{eqnarray} \label{Lor}
311: \epsilon(\nu) = \epsilon_{\infty} + \sum_{j} \frac{\Omega_j^2}{TO_j^2 - \nu^2 - i \gamma_j \nu},
312: \end{eqnarray}
313:
314: where $\epsilon_{\infty}$ is the dielectric constant for frequencies large
315: compared to all infrared wavelengths, $\Omega_j$ is the strength, TO$_j$ the
316: transverse optical {phonon} frequency, $\gamma_j$ the damping constant of the $j$-th
317: oscillator, respectively.
318:
319: \begin{figure*}[htbp]
320: %\epsscale{1.6}
321: \plotone{f3a.eps}
322: \plotone{f3b.eps}
323: \caption{Optical constants of calcite for room temperature, 200\,K, 100\,K
324: and 10\,K for the ordinary ray (E$\bot$c) (left) and for the extraordinary ray
325: (E$\|$c) (right). \label{f:nks_Calc}}
326: \end{figure*}
327:
328: %\begin{figure}[htbp]
329: %\plotone{f4.eps}
330: %\caption{Optical constants of calcite for room temperature, 200\,K, 100\,K and
331: % 10\,K for the extraordinary ray (E$\|$c).
332: % \label{f:nkp_Calc}}
333: %\end{figure}
334:
335: Figure \ref{f:nks_Calc} shows the optical constants of
336: calcite which have been derived from the polarized reflectance spectra by
337: means of Lorentz oscillator fits. The oscillator parameters which were used
338: for these fits -- five oscillators for the ordinary ray and three for the
339: extraordinary ray -- are listed in Tab.\ \ref{t:cal-RT-Lor}.
340:
341: For dolomite, we only list its oscillator parameters -- six oscillators
342: for the ordinary ray and five for the extraordinary ray (see Tab.\
343: \ref{t:dol-RT-Lor}) --, while the optical constants are not shown.
344: However, both for calcite and dolomite, the optical constants can be
345: retrieved from the electronic database
346: { \texttt{http://www.astro.uni-jena.de/Labo\-ratory/OCDB} },
347: which also includes graphs.
348:
349: %As mentioned in the introduction, dolo\-mite has a slightly more complex infrared
350: %spectrum than calcite. This can easily be verified by counting the respective
351: %numbers of oscillators. To enable a satisfactory reproduction of the mid-IR
352: %reflectance spectra of dolomite, six oscillators are needed for the ordinary
353: %ray, and five are required for the exrtraordinary ray (instead of five and
354: %three, respectively, in the case of calcite).
355:
356: Apart from the different number of oscillators, another important difference
357: between calcite's and dolomite's IR spectra is the absence of any phonon
358: band for the latter at wavelengths larger than 67\,$\mu$m (or wavenumbers
359: smaller than 150\,cm$^{-1}$). In other words, it is noteworthy that calcite's
360: IR bands extend to wavelengths roughly 1.5 times larger than dolomite's
361: `reddest' band.
362:
363: For 300\,K, our oscillator parameters are well comparable with those
364: derived by \cite{H70} except for some minor bands which are missing
365: in their measurements (e.g.\ our `4$_{ER}$' for dolomite),
366: probably due to a somewhat too small spectral resolution.
367:
368: \begin{table*}[htbp]
369: \begin{center}
370: \caption{Lorentz oscillator parameters derived for calcite.
371: `OR' refers to the ordinary ray, `ER' to the extraordinary ray.
372: Where no temperature is indicated as subscript, oscillator
373: parameters have been derived from room temperature measurements.
374: % Where a temperature is additionally indicated as a subscript,
375: % oscillator parameters have been derived both from the
376: % room temperature and from the cryogenic measurements, respectively.
377: % The resonances 1$_{OR}$, 2$_{OR}$ and 1$_{ER}$ are less temperature
378: % sensitive, hence only the oscillator parameters were only derived from
379: % the room temperature spectra.
380: For the dielectric background, we chose the
381: value $\epsilon_{\inf, \rm OR}$ = 2.6 in the case of OR and
382: $\epsilon_{\inf, \rm ER}$ = 2.3 in the case of ER. All values
383: are given in units of cm$^{-1}$.
384: \label{t:cal-RT-Lor}}
385: \begin{tabular}{lccc}
386: \hline\hline
387: %Calcit, TP, Osz.par., März 2005 / Februar 2006
388: $j$ & TO$_j$ & $\Omega_j$ & $\gamma_j$ \\
389: \tableline
390: 1$_{\rm OR}$ & 1407.4 & 1039.6 & 14.8 \\
391: 2$_{\rm OR}$ & 712.0 & 71.4 & 4.0 \\
392: 3$_{\rm OR, 300K}$ & 297.0 & 377.7 & 15.7 \\
393: 3$_{\rm OR, 200K}$ & 302.5 & 385.4 & 12.0 \\
394: 3$_{\rm OR, 100K}$ & 306.0 & 387.0 & 8.87 \\
395: 3$_{\rm OR, 10K}$ & 307.1 & 387.6 & 7.34 \\
396: 4$_{\rm OR, 300K}$ & 222.8 & 206.1 & 11.0 \\
397: 4$_{\rm OR, 200K}$ & 226.5 & 211.2 & 8.35 \\
398: 4$_{\rm OR, 100K}$ & 228.8 & 219.1 & 6.14 \\
399: 4$_{\rm OR, 10K}$ & 229.7 & 219.7 & 4.54 \\
400: 5$_{\rm OR, 300K}$ & 104.7 & 148.1 & 6.00 \\
401: 5$_{\rm OR, 200K}$ & 103.2 & 165.7 & 5.35 \\
402: 5$_{\rm OR, 100K}$ & 103.8 & 158.0 & 3.57 \\
403: 5$_{\rm OR, 10K}$ & 104.2 & 167.6 & 2.45 \\
404: \tableline
405: %Calcit, Osz.par., TP, März 2005 / Februar 2006
406: 1$_{ER}$ & 872.0 & 250.8 & 2.09 \\
407: 2$_{ER,300K}$ & 306.7 & 368.0 & 16.9 \\
408: 2$_{ER,200K}$ & 312.6 & 372.2 & 11.6 \\
409: 2$_{ER,100K}$ & 317.6 & 377.6 & 9.19 \\
410: 2$_{ER,10K}$ & 318.9 & 378.7 & 7.30 \\
411: 3$_{ER,300K}$ & 94.3 & 200.6 & 7.10 \\
412: 3$_{ER,200K}$ & 93.3 & 203.2 & 5.91 \\
413: 3$_{ER,100K}$ & 94.6 & 203.6 & 4.37 \\
414: 3$_{ER,10K}$ & 95.0 & 202.4 & 2.86 \\
415: \tableline
416: \end{tabular}
417: \end{center}
418: \end{table*}
419:
420: \begin{table*}[htbp]
421: \begin{center}
422: \caption{The set of Lorentz oscillator parameters derived
423: for dolomite by fitting the room temperature reflectance
424: spectra. All designations have the same meaning as in the
425: previous table, but here, we chose
426: $\epsilon_{\inf, \rm OR}$ = 2.7 and
427: $\epsilon_{\inf, \rm ER}$ = 2.4.
428: \label{t:dol-RT-Lor}}
429: \begin{tabular}{lccc}
430: \hline\hline
431: %Dolomit, TP, Osz.par., März 2005 / Februar 2006
432: $j$ & TO$_j$ & $\Omega_j$ & $\gamma_j$ \\
433: \tableline
434: 1$_{\rm OR}$ & 1453.8 & 399.2 & 23.8 \\
435: 2$_{\rm OR}$ & 1425.5 & 980.0 & 15.0 \\
436: 3$_{\rm OR}$ & 727.3 & 98.8 & 6.14 \\
437: 4$_{\rm OR, 300K}$ & 348.7 & 459.0 & 22.1 \\
438: 4$_{\rm OR, 200K}$ & 351.5 & 462.5 & 16.0 \\
439: 4$_{\rm OR, 100K}$ & 356.0 & 464.4 & 12.0 \\
440: 4$_{\rm OR, 10K}$ & 355.9 & 464.5 & 11.9 \\
441: 5$_{\rm OR, 300K}$ & 259.1 & 190.6 & 9.86 \\
442: 5$_{\rm OR, 200K}$ & 261.1 & 201.9 & 7.97 \\
443: 5$_{\rm OR, 100K}$ & 263.4 & 201.5 & 5.58 \\
444: 5$_{\rm OR, 10K}$ & 263.4 & 201.4 & 5.54 \\
445: 6$_{\rm OR, 300K}$ & 152.9 & 209.1 & 5.18 \\
446: 6$_{\rm OR, 200K}$ & 152.9 & 214.2 & 3.42 \\
447: 6$_{\rm OR, 100K}$ & 153.5 & 214.9 & 2.26 \\
448: 6$_{\rm OR, 10K}$ & 153.5 & 214.7 & 2.20 \\
449: \tableline
450: %Dolomit, Osz.par., TP, März 2005 / Februar 2006
451: 1$_{ER}$ & 874.9 & 301.0 & 2.98 \\
452: 2$_{ER,300K}$ & 364.8 & 335.8 & 20.4 \\
453: 2$_{ER,200K}$ & 368.5 & 341.8 & 16.3 \\
454: 2$_{ER,100K}$ & 370.7 & 349.8 & 12.7 \\
455: 2$_{ER,10K}$ & 371.5 & 356.4 & 11.9 \\
456: 3$_{ER,300K}$ & 316.8 & 179.5 & 9.13 \\
457: 3$_{ER,200K}$ & 319.4 & 175.6 & 8.24 \\
458: 3$_{ER,100K}$ & 321.3 & 168.1 & 6.05 \\
459: 3$_{ER,10K}$ & 321.6 & 178.0 & 5.05 \\
460: 4$_{ER,300K}$ & 171.9 & 30.5 & 8.76 \\
461: 4$_{ER,200K}$ & 174.0 & 29.2 & 6.42 \\
462: 4$_{ER,100K}$ & 174.4 & 31.1 & 5.56 \\
463: 4$_{ER,10K}$ & 174.8 & 34.4 & 4.97 \\
464: 5$_{ER,300K}$ & 150.3 & 243.2 & 3.35 \\
465: 5$_{ER,200K}$ & 151.2 & 247.5 & 3.29 \\
466: 5$_{ER,100K}$ & 151.4 & 249.8 & 2.31 \\
467: 5$_{ER,10K}$ & 152.2 & 260.4 & 2.17 \\
468: \tableline
469: \end{tabular}
470: \end{center}
471: \end{table*}
472:
473:
474: %------------------------------------------------------------------------------
475:
476:
477: \section{Small particle spectra for different shape distributions \label{s:mie}}
478:
479: From the optical constants shown and discussed in the previous section, the
480: absorption cross sections per particle volume -- C$_{abs}$/V -- have been
481: calculated for {grains of different shapes. Since the focus of this paper is
482: on bands at large wavelengths, we choose the limit of particles being small
483: compared to the wavelength (Rayleigh limit, size parameter
484: x=2$\pi$a/$\lambda\ll$ 1 and $\vert\epsilon^{1/2}\vert$ x $\ll$ 1,
485: $a$ being the grain radius or largest semidiameter).
486:
487: The following grain shapes have been taken into account: spheres (according to Mie
488: theory) and continuous distributions of ellipsoids (CDEs). As for the latter, two
489: different modifications of it have been used: a CDE with equal probability for all
490: ellipsoid axis ratios (`mean CDE', after \cite{BH83}) and a CDE with maximum
491: probability for spherical grains (`weighted CDE', after \cite{Osk92}).
492:
493:
494: \subsection{Temperature effects for spherical grains}
495:
496: {Figures \ref{f:Cabs_Cal} and \ref{f:Cabs_Dol} show the volume normalized absorption cross
497: sections C$_{abs}$/V calculated for spherical calcite and dolomite grains, respectively,
498: depending on temperature.
499: The figures demonstrate that the effect of the cooling on the band positions
500: and band widths is present also in the small particle spectra. Similarly to the
501: reflectance spectra, this effect is especially pronounced for the case of the
502: $\sim$40\,$\mu$m band of calcite, which shifts from 42.6\,$\mu$m for 300\,K to
503: 41.4\,$\mu$m for 10\,K, while at the same time, the C$_{abs}$/V peak value increases
504: by a factor of 2.4. For the calcite bands around 30 and 80\,$\mu$m, the
505: temperature dependence is not equally strong, but still significant.
506:
507: In the case of CaMg[CO$_3$]$_2$, the tempera\-ture-dependent evolution of the spectra
508: is similar apart from the effect that there is only an insignificant change between
509: 100\,K and 10\,K (cf. Fig.\ \ref{f:Cabs_Dol}). For the additional band in the dolomite
510: spectrum (compared to calcite) around 31\,$\mu$m, the temperature dependence is rather
511: moderate compared to the other bands. In the wavelength range
512: $\lambda$\,$<$\,15\,$\mu$m, the dolomite and the calcite spectra resemble
513: each other very closely.
514:
515: \begin{figure*}[htbp]
516: \epsscale{1.6}
517: \plotone{f4.eps}
518: \caption{Volume normalized absorption cross sections of spherical calcite
519: grains for room temperature, 200\,K, 100\,K and 10\,K in the wavelength
520: range 5--110\,$\mu$m.
521: \label{f:Cabs_Cal}}
522: \end{figure*}
523:
524: \begin{figure*}[htbp]
525: \epsscale{1.6}
526: \plotone{f5.eps}
527: \caption{Volume normalized absorption cross sections of spherical dolomite
528: grains for room temperature, 200\,K, 100\,K and 10\,K.
529: \label{f:Cabs_Dol}}
530: \end{figure*}
531:
532:
533: \subsection{Grain shape effects at low temperatures}
534:
535: Due to the high oscillator strengths characterizing most of the infrared bands of both
536: calcite and dolomite, the grain shape dependence of the band profiles is very strong.
537: For the FIR bands of calcite, this is evident from the comparison of
538: Fig.\ \ref{f:Cabs_Cal} with Fig.\ \ref{f:Cabs_Cal_CDE} as well as from the
539: peak wavelengths given in Tab.\ \ref{t:shapes}. Calcite's 30\,$\mu$m band
540: is shifted to the `red' by about 2\,$\mu$m for a mean CDE compared to the case of
541: spherical grains, and at the same time its width is increased by factor of four.
542: The 80-90\,$\mu$m double-band of calcite degenerates into a broad single bump
543: centered at 90.1\,$\mu$m with two wings for the mean CDE.
544: For the weighted CDE case, the maximum of the absorption
545: efficiency is located at 86.3\,$\mu$m for a grain temperature of 300\,K.
546: For the 40\,$\mu$m band, the broadening due to the shape distribution
547: is not equally strong as for the 80-90\,$\mu$m band, and the shift of
548: the band position to longer wavelengths is limited to 1\,$\mu$m for
549: a given temperature (compared to more than 5\,$\mu$m for the 80-90\,$\mu$m
550: band).
551:
552: \begin{table*}[htbp]
553: \begin{center}
554: \caption{Positions of the maxima of C$_{abs}$/V for
555: calcite grains at different temperatures and for different
556: shape distributions
557: \label{t:shapes}}
558: \begin{tabular}{lccc}
559: \hline\hline
560: & 20\,$\mu$m complex & 40\,$\mu$m & 80--90\,$\mu$m \\
561: Peak position & [$\mu$m] & [$\mu$m] & [$\mu$m] \\
562: \tableline
563: \textbf{sphere} & & & \\
564: 10\,K & 27.05, 27.84 & 41.40 & 79.62, 83.77 \\
565: 300\,K & 28.58 & 42.64 & 80.56, 84.71 \\
566: \tableline
567: \textbf{weighted CDE} & & & \\
568: 10\,K & 28.18 & 41.75 & 84.91 \\
569: 300\,K & 29.13 & 43.09 & 86.25 \\
570: \tableline
571: \textbf{mean CDE} & & & \\
572: 10\,K & 30.02 & 42.54 & 90.31 \\
573: 300\,K & 30.66 & 43.68 & 90.14 \\
574: \tableline
575: \end{tabular}
576: \end{center}
577: \end{table*}
578:
579: Comparing the band properties for different temperatures within a given grain model,
580: we find that the shift of the peak position is very prominent in the case of
581: the 40\,$\mu$m feature, even for the grain shape distributions. It clearly dominates
582: over shape effects for this band, whereas this is not true
583: for the 30 and 80--90\,$\mu$m features. In the case of the latter both the shape effects
584: are stronger due to the higher oscillator strengths and the temperature effects are
585: intrinsically smaller as has been discussed in the previous section.
586:
587: \begin{figure}[htbp]
588: \epsscale{1.0}
589: \plotone{f6.eps}
590: \caption{C$_{abs}$/V profiles for a distribution of ellipsoidally
591: shaped calcite grains (see text for details). Only the wavelength range
592: 20--120\,$\mu$m and only the two temperature `extremes'
593: T=10\,K and T=300\,K have been depicted here.
594: \label{f:Cabs_Cal_CDE}}
595: \end{figure}
596:
597:
598: \subsection{Comparison to PE powder spectra}
599:
600: Since powder spectra of calcite and dolomite -- measured at room
601: temperature -- have been utilized by both Kemper et al.\ (2002a) and
602: Chiavassa et al.\ (2005) in order to argue for the presence of carbonate dust in
603: astronomical objects, it is important to discuss the relation between such
604: powder spectra and the small particle spectra based on our optical constants.
605: By `powder spectra' or `PE spectra', we understand the transmission spectra
606: of sub-micron sized ground powder embedded in polyethylene (PE),
607: transformed into absorption cross sections.
608:
609: Powder spectra generally depend on various circumstances such as particle size
610: and shape, orientation, agglomeration degree, and the refractive index of the
611: surrounding medium (e.g.\ \cite{BH83}).
612:
613: Fig.\ \ref{f:PE_CDE} shows both the powder spectrum published by Kemper
614: et al.\ (2002a) -- measured on a sample of unsedimented calcite grains
615: (dash-dotted-line) -- and of a sample of calcite grains which had previously
616: been subjected to sedimentation in a liquid (acetone), in order to remove large grains (solid line).
617: The peak of the latter spectrum is shifted by about 2\,$\mu$m to shorter wavelengths with respect to the data published by Kemper et al.\ (2002a), although both have been measured at room temperature.
618: { This is due to the presence of large grains (several $\mu$m in diameter)
619: in the unsedimented powder.}
620: The spectrum of the sedimented powder is in good agreement with the result of the calculation for a {\em mean}\/ CDE with PE environment (see Fig.\ \ref{f:PE_CDE}), indicating the presence of a considerable fraction of grains with
621: shapes far away from spherical symmetry.
622:
623: Conversely, we find that the mean CDE is a good representation of the shape
624: distribution of a real calcite powder with grain sizes within the Rayleigh
625: limit. This justifies the use of this shape distribution for the comparison with
626: observed data at large wavelengths. In the following (see Sect.\,4.1) we will
627: apply this model to simulate dust spectra at low temperatures.
628:
629: \begin{figure}[htbp]
630: \plotone{f7.eps}
631: \caption{Comparison of calcite's FIR bands measured by PE powder transmission
632: spectroscopy (solid and dashed-dotted lines; see text for details) with the absorption efficiency spectrum calculated from the optical constants for a mean continuous distribution of ellipsoids (dashed) and for a weighted CDE (dotted),
633: both in PE.
634: \label{f:PE_CDE}}
635: \end{figure}
636:
637:
638: %---------------------------------------------------------------------------
639:
640:
641: \section{Comparison with astronomical observations \label{s:comp}}
642:
643: %30.1.2007
644: The laboratory spectra presented so far are especially relevant for a comparison
645: with FIR spectra of cold dust in objects such as the planetary nebula NGC 6302.
646: Hence, we shall first give a brief review of some key properties of the dust
647: spectrum of this object.
648:
649: Barlow (1997) observed features 65 and 69\,$\mu$m and another possible broad
650: band feature in the range of 88-98\,$\mu$m. While the 65\,$\mu$m feature was
651: attributed to crystalline ice (see also Waters et al.\ 1996), the 69\,$\mu$m
652: feature was suspected to be due to crystalline forsterite and the question
653: for the carrier of the 88-98\,$\mu$m band remained open.
654:
655: Molster et al.\ (2001) identified the $\sim$65\,$\mu$m band as a blend of diopside
656: and crystalline water ice with enstatite. They also confirmed the reality of the
657: broad feature around 90\,$\mu$m and suspected that there should be a very cold dust
658: component present in the nebula.
659:
660: %12.1.2007
661: Kemper et al.\ (2002a,b) first assigned { the $\sim$65\,$\mu$m feature
662: to cold dolomite (with additional contributions by diopside and water ice)}
663: and the broad emission band around 90\,$\mu$m to cold calcite dust. As already
664: mentioned above, the temperature of the carbonate dust was assumed to be in the
665: 30-60\,K range. A mass fraction of less than 0.3\% (for calcite as well as for dolomite) has been derived by these authors.
666:
667: We retrieved an ISO-SWS and an ISO-LWS spectrum of NGC 6302 from the ISO archive and reduced it by means of the OLP version 10.0.
668:
669: From the composite spectrum (ranging from 2.4--200\,$\mu$m), we subtracted a combination of Planck functions for temperatures of 30, 55 and 95\,K. The remaining `residual dust emission' is shown in Fig.\ \ref{f:resid1} for the
670: 35--105\,$\mu$m range; the individual broad peaks of the gray line result from various ice and dust species such as H$_2$O, forsterite, enstatite, diopside, and -- possibly -- carbonates (see Kemper et al.\ 2002a, Tab.\ 2).
671:
672: \begin{figure}[htbp]
673: \plotone{f8.eps}
674: \caption{Comparison between the residual dust emission of NGC 6302
675: and synthetic dust emission spectra (for 30\,K and 60\,K) of
676: calcite and dolomite in the 35-105\,$\mu$m range.
677: \label{f:resid1}}
678: \end{figure}
679:
680: Our aim has been to examine whether the assignments of the 60-65\,$\mu$m
681: and `92'\,$\mu$m bands to dolomite and calcite is compatible with the
682: optical constants and powder spectra presented in Sect.\ 3, and, if so,
683: for which temperatures and particle shapes this is the case. For
684: answering this questions, the detectability of bands at wavelengths
685: between 35 and 45\,$\mu$m depending on the dust temperature also
686: plays a key role.
687:
688:
689: \subsection{The `92'\,$\mu$m band in NGC 6302}
690:
691: A detailed view of the `92'\,$\mu$m band is given in Fig.\ \ref{f:resid2}.
692: It can be seen from this figure (when compared to Fig.\ \ref{f:Cabs_Cal_CDE})
693: that only for a mean CDE particle shape distribution of calcite grains,
694: the bandwidth is as large as in the case of the NGC 6302 spectrum.
695: For a weighted CDE, the bandwidth is smaller by 45\% (at the relevant
696: cryogenic temperatures). The best fit is achieved by the C$_{abs}$/V function
697: for a mean CDE multiplied by a Planck function for a temperature
698: of 30\,K. A dust temperature of 60\,K is hardly compatible with the observations,
699: both with respect to the `92'\,$\mu$m band profile and with respect to the
700: emergence of a $\sim$42\,$\mu$m feature (see dotted lines in Figs.\
701: \ref{f:resid1} and \ref{f:resid2}). The calcite data derived from
702: transmission spectroscopy of {\em sedimented}\/ powder (dash-dotted line
703: in Fig.\ \ref{f:resid2}) roughly reproduce the observed band profile as
704: well. It is noteworthy that the `92'\,$\mu$m band profile of NGC 6302
705: obtained by Kemper et al.\ (2002a) is better fitted by the PE spectrum
706: of an {\em unsedimented}\/ calcite powder sample which is broader and peaks
707: at a larger wavelength. This may be due to the use of different ISO LWS spectra
708: (a mean of seven vs.\ a single) or due to some uncertainty in the continuum
709: subtraction.
710:
711: \begin{figure}[htbp]
712: \plotone{f9.eps}
713: \caption{Comparison between the resi\-dual dust emission of NGC 6302
714: and synthetic dust emission spectra of ensembles of ellipsoidally
715: shaped calcite grains for 30\,K and 60\,K in the 75--105\,$\mu$m range.
716: \label{f:resid2}}
717: \end{figure}
718:
719:
720: \subsection{The 60-65\,$\mu$m emission complex in NGC 6302}
721:
722: In Fig.\ \ref{f:resid3} a close-up view of the 60\,$\mu$m emission complex is shown.
723: Although the bandwidth of a mean CDE particle shape distribution can reproduce the
724: observed bandwidth of the 60\,$\mu$m feature, neither for a mean CDE nor for a
725: weighted CDE -- both at a temperature of 45\,K -- the {\em positions}\/ coincide with the
726: observed one. The difference between the band positions resulting from
727: laboratory measurements and the observed band position ($\sim$\,64\,$\mu$m) is
728: too large as to be compensated by temperature effects -- see also Fig.\
729: \ref{f:resid1}, where a comparison with dolomite dust at T=30\,K is made.
730: Dust temperatures lower than 40\,K would allow to suppress the dolomite
731: emission band at $\sim$37\,$\mu$m (see again Fig.\ \ref{f:resid1}), but the
732: position of the 60\,$\mu$m feature would still not match.
733: Therefore this feature cannot be fit only by dolomite,
734: while it is not excluded that dolomite contributes to the `blue wing' of
735: the 60\,$\mu$m emission complex. Our study confirms that the 60-65\,$\mu$m
736: band must have other major carriers, e.g.\ diopside as suggested by
737: Kemper et al.\ (2002a). A recent publication by Chihara et al.\
738: (2007) shows that a mixture of diopside and {\aa}kermanite (Ca$_2$MgSi$_2$O$_7$)
739: at a temperature of 30\,K provides a good match of the 60-65\,$\mu$m feature.
740:
741:
742: \begin{figure}[htbp]
743: \plotone{f10.eps}
744: \caption{Comparison between the residual dust emission of NGC 6302
745: and synthetic dust emission spectra of ensembles of ellipsoidally
746: shaped dolomite grains for 30\,K and 50\,K in the 49--81\,$\mu$m range.
747: \label{f:resid3}}
748: \end{figure}
749:
750: \subsection{The 90--110\,$\mu$m band in protostars}
751:
752: Chiavassa et al.\ (2005) studied a sample of 32 low and intermediate
753: mass protostars (mostly Class 0 and Herbig Ae/Be objects). Out of this
754: sample, 17 sources show a broad band feature between 90 and 110\,$\mu$m.
755: The dust temperature in these objects is found by blackbody fits to the
756: continuum to be below 50\,K (except for emission from the innermost regions).
757: Calcite is suggested to be a possible carrier of this feature, which seems
758: to be supported by the laboratory measurements of Kemper et al.\ (2002a)
759: that show a peak at 93\,$\mu$m and a width (FWHM) of 16\,$\mu$m.
760:
761: Our data show that for PE powder measurements on small grains and also for
762: calculated spectra assuming a mean CDE, the maximum of the `92'$\mu$m
763: calcite band is located at 90$\mu$m.
764: For other particle shape distributions, it is narrower and located
765: at much smaller wavelengths, especially for the temperature range relevant
766: for PNe and protostars. The transverse optical resonance wavelength
767: of mode 5$_{OR,10K}$ of about 96\,$\mu$m (Tab.\ 1) can be considered
768: as the maximum peak position for this band (cf.\ \cite{BH83}). The
769: feature produced by the 3$_{ER,10K}$ mode with a TO resonance wavelength
770: of about 105\,$\mu$m is much broader and therefore comparably weak
771: and cannot dominate the calcite grain emission even if alignment
772: effects were present.
773:
774: Hence we conclude that the bands observed for the Herbig Ae/Be objects
775: may in principle be compatible with emission from calcite grains, if sufficiently
776: extreme shape distributions (much more extreme than the mean CDE) and/or
777: large grains are assumed (cf.\ Fig.\ 7 in Chiavassa et al.\ 2005).
778: For the Class 0 objects, in the contrary, we confirm that the feature
779: characteristics are incompatible with laboratory data for calcite.
780: Whether the incorporation of other ions into a carbonate can shift the
781: band to still larger wavelengths has to be left open at this point.
782:
783: %-----------------------------------------------------------------------------
784:
785: \section{Conclusions \label{s:conc}}
786:
787: While previous studies on carbonate dust in astronomical environments were
788: largely based on powder measurements performed at room temperature,
789: we have on the one hand derived optical constants of calcite and dolomite
790: and on the other hand carried out measurements at cryogenic temperatures
791: (200, 100 and 10\,K). From these optical constants, small particle spectra
792: (absorption efficiencies as functions of the wavelength and temperature)
793: were calculated for different grain shapes.
794:
795: Since the optical constants of calcite and dolomite reach peak values of
796: 5--10 (corresponding to maxima of the imaginary part of the dielectric function
797: of up to 150), their absorption efficiencies are extremely sensitive to particle
798: shape effects. For the `92'\,$\mu$m band of calcite, e.g., the negativity
799: range of the real part of the dielectric function is as large as
800: 73--96\,$\mu$m for the ordinary ray (and even 75--105\,$\mu$m for
801: the extraordinary ray). The former values approximately indicate
802: how large the range is in which calcite's `92'\,$\mu$m band can shift due
803: to particle shape effects, since the ordinary ray is weighted more
804: strongly than the extraordinary ray.
805:
806: At cryogenic temperatures, the FIR bands of calcite and dolomite increase
807: their intensity, get narrower and shift to shorter wavelengths. The
808: temperature-related shift in position is largest for calcite's 40\,$\mu$m
809: band: in this case, it reaches 0.5\,$\mu$m per 100\,K.
810:
811: As for the `92'\,$\mu$m feature detected by Kemper et al.\ (2002a) in ISO-LWS
812: spectra of NGC 6302, this feature can only be assigned to CaCO$_3$ grains
813: if the assumption of a mean CDE -- with equal probability of all grain shapes
814: -- is made.
815:
816: Furthermore, dust temperatures have to be well below 40\,K in order
817: to suppress the 30\,$\mu$m and 42\,$\mu$m bands of calcite which accompany
818: the `92'\,$\mu$m band but are not seen in the spectra of NGC 6302.
819:
820: %-----------------------------------------------------------------------------
821:
822:
823: \begin{small}
824: {\acknowledgments
825:
826: We thank Gabriele Born and Walter Teu\-schel, Jena, for
827: sample preparation and help with the cryogenic measurements.
828: Ingrid Hodou\v s kindly provided the ISO-spectrum of
829: NGC 6302, reduced with the ISO spectral analysis
830: package ISAP. HM acknowledges support by DFG grant
831: Mu \mbox{1164/5.} TP acknowledges support by the DFG
832: `Forschergruppe' {\em laboratory astrophysics}\/,
833: by a grant from the Max-Planck-Society. AB has been supported
834: by the Austrian `Fonds zur F\"orderung der wissenschaftlichen
835: Forschung' (FWF; project number P18939-N16).
836: Comments by an anonymous referee helped to improve
837: the original version of this paper.}
838:
839: \end{small}
840:
841: %======================================================
842:
843: \begin{thebibliography}
844:
845: {\bibitem [Barlow(1997)] {B97}
846: Barlow, M.J., 1997, Ap\&{}SS 255, 315}
847:
848: {\bibitem [Bohren \& Huffman(1983)] {BH83}
849: Bohren, C.F., Huffman, D.R., 1983, Absorption and Scattering of Light
850: by Small Particles, John Wiley, New York}
851:
852: {\bibitem [Bregman \& Rank(1975)] {BR75}
853: Bregman, J.D. \& Rank, D.M., 1975, \apj 195, L125}
854: %Identification of the 890cm-1 carbonate signature in NGC 7072
855:
856: {\bibitem [Bregman(1977)] {Breg77}
857: Bregman, J.D., 1977, PASP 89, 335}
858:
859: {\bibitem [Ceccarelli et al.(2002)] {Cecc02}
860: Ceccarelli, C., Caux, E., Tielens, A.G.G.M., Kemper, F., Waters, L.B.F.M.,
861: Phillips, T., 2002, \aap 395, L29}
862:
863: {\bibitem [Chiavassa et al.(2005)] {Chia05}
864: Chiavassa A., Ceccarelli, C., Tielens, A.G.G.M., Caux, E., Maret, S.,
865: 2005,\aap 432, 547}
866:
867: {\bibitem [Chihara et al.(2007)] {Chih07}
868: Chihara H., Koike, C., Tsuchiyama, A., 2007,
869: \aap 464, 229}
870:
871: {\bibitem[Cohen et al.(1986)]{Coh86}
872: Cohen, M., Allamandola, L., Tielens, A.G.G.M., et al., 1986, \apj 302, 737}
873:
874: {\bibitem[Erard et al.(2000)]{Er00}
875: Erard, S., Lellouch, E., Encrenaz, T., Morris, P., de Graauw, T.,
876: \& Burgdorf, M., 2000, LPI Conf.\ Abstr.\, 31, 1325}
877:
878: {\bibitem [Ferrarotti \& Gail(2005)] {FG05}
879: Ferrarotti, A.S., \& Gail H.-P., 2005, \aap 430, 959}
880:
881: {\bibitem [Gillet, Forrest \& Merrill 1973] {GFM73}
882: Gillett, F.C., Forrest, W.J., Merrill, K.M., 1973, \apj 183, 87}
883: %8-13-micron spectra of NGC 7027, BD +30°3639, and NGC 6572
884:
885: {\bibitem [Hellwege et al.(1970)] {H70}
886: Hellwege, K.H., Lesch, W., Plihal, M., Schaack, G., 1970, Z.\ Physik 232, 61}
887:
888: {\bibitem [Hofmeister, Wopenka \& Lockock(2004)] {Hof04}
889: Hofmeister, A.M., Wopenka, B., Locock, A.J., 2004, \gca 68, 4485}
890:
891: {\bibitem [Joswiak \& Brownlee(2001)] {JB01}
892: Joswiak, D.J., Brownlee, D.E., 2001, Lunar and Planetary Institute
893: Conference Abstracts, 32, 1998}
894:
895: {\bibitem [Kemper et al.(2002a)] {K02aunda}
896: Kemper, F., Molster, F.J., J\"ager, C., Waters, L.B.F.M.,
897: 2002a, \aap 394, 679}
898:
899: {\bibitem [Kemper et al.(2002b)] {K02nat}
900: Kemper, F., J\"ager, C., Waters, L.B.F.M., et al., 2002b, Nature 415, 295}
901:
902: {\bibitem [Lellouch, Encrenaz \& de Graauw(2000)] {Lel00}
903: Lellouch, E., Encrenaz, T., de Grauuw, T., et al., 2000,
904: Planet.\ Space Sci., 48, 1393}
905:
906: {\bibitem [Lisse et al.(2006)] {Liss06}
907: Lisse C.M., VanCleve J., Adams A.C., Science 313, 635}
908:
909: {\bibitem [McCarthy, Forrest \& Houck(1978)] {MFH78}
910: McCarthy, J.F., Forrest, W.J., Houck, J.R., 1978, \apj 224, 109}
911: %16-38 micron spectroscopy of NGC 7027
912:
913: {\bibitem [Molster et al.(2001)] {M01}
914: Molster, F.J., Lim, T.L., Silvester, R.J., et al., 2001,
915: \aap 372, 165}
916:
917: {\bibitem [Onaka \& Okada(2003)] {OO03}
918: Onaka, T., \& Okada, Y., 2003, \apj 585, 872}
919:
920: {\bibitem [Orofino et al.(1998)] {Or98}
921: Orofino, V., Blanco, A., Fonti, S., Proce, R., Rotundi, A., 1998,
922: Planet.\ Space Sci., 46, 1659}
923:
924: {\bibitem [Ossenkopf, Henning \& Mathis(1992)] {Osk92}
925: Ossenkopf, V., Henning, Th., Mathis, J.S., 1992, \aap 261, 567}
926:
927: {\bibitem [Penman(1976)] {P76}
928: Penman, J.M., 1976, \mnras 176, 539}
929:
930: {\bibitem [Quirico et al.(2000)] {Q00}
931: Quirico, E., Raynal, P.-I., Borg, J., et al., 2000,
932: Lunar and Planetary Institute Conference Abstracts, 31, 1260}
933:
934: {\bibitem [Sandford(1986)] {San86}
935: Sandford, S.A., 1986, Science, 231, 1540}
936:
937: {\bibitem [Tomeoka \& Buseck(1986)] {TB86}
938: Tomeoka, K., \& Buseck, P.R., 1986, Science, 231, 1544}
939:
940: { {\bibitem [Toppani et al. (2005)] {T05}
941: Toppani A., Robert F., Libourel G., et al., 2005, Nature, 437, 1121}}
942:
943: {\bibitem [Waters(1996)] {W96}
944: Waters L.B.F.M, Molster F., de Jong T., et al., 1996,
945: \aap 315, L361}
946:
947: {\bibitem [Wyckhoff(1964)] {W64}
948: Wyckhoff, R.W.G., 1964, Crystal structures, 2nd ed.,
949: vol.\ 2, New York and London: Sydney Interscience Publ.}
950:
951: \end{thebibliography}
952:
953: \end{document}