1: \documentclass[preprint]{aastex}
2: \usepackage[]{epsfig}
3: %%%%%%%%%%%%%%%%%%%%% macro definitions %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4: \newcommand \beq {\begin{equation}}
5: \newcommand \eeq {\end{equation}}
6: \newcommand \Angstrom {\,{\rm \AA}}
7: \newcommand \C {{\rm C}}
8: \newcommand \cm {\,{\rm cm}}
9: \newcommand \erg {\,{\rm erg}}
10: \newcommand \eV {\,{\rm eV}\,}
11: \newcommand \g {\,{\rm g}}
12: \newcommand \rmH {{\rm H}}
13: \newcommand \K {\,{\rm K}}
14: \newcommand \kms {\,{\rm km \, s}^{-1}}
15: %\newcommand \micron {\,\mu{\rm m}}
16: \newcommand \nH {n_{\rm H}}
17: \newcommand \pc {\,{\rm pc}}
18: \newcommand \s {\,{\rm s}}
19: \newcommand \uHab {u_{\rm Hab}^{\rm uv}}
20: \newcommand \urad {u_{\rm rad}}
21: \newcommand \yr {\,{\rm yr}}
22: \newcommand \Qabs {Q_{\rm abs}}
23: \newcommand \simlt {\lesssim}
24: \newcommand \simgt {\gtrsim}
25: \newcommand \gtsim {\gtrsim}
26: \newcommand \ltsim {\lesssim}
27: \newcommand \fsi {{\rm Y_{\rm si}}}
28:
29: \newcommand{\figwidth}{11cm}
30: \newcommand{\figwidthland}{12cm}
31: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
32: %------- delete following for submission to ApJ --------
33: \pagestyle{myheadings}
34: % today's date
35: %use number register 200 for "decade"
36: \countdef\decade=200
37: \decade=0
38: \advance\decade by \year
39: %\advance\decade by -2000 %to suppress two leading digits of yearb
40: \countdef\hours=201
41: \hours=0
42: \advance\hours by \time
43: \divide\hours by 60
44: \countdef\mins=202
45: \mins=0
46: \advance\mins by \hours
47: \multiply\mins by 60
48: \multiply\hours by 100
49: \countdef\miltime=203
50: \miltime=0
51: \advance\miltime by \hours
52: \advance\miltime by \time
53: \advance\miltime by -\mins
54: \def\today{\number\decade.\number\month.\number\day.\number\miltime}
55:
56: %\markright{\today: DRAFT}
57:
58: \shorttitle{Ultrasmall Silicate Grains}
59:
60: \begin{document}
61:
62: \title{
63: %------------- enable for labelling preprint ---------------------------
64: \vspace*{-2.0em}
65: {\normalsize\rm submitted to {\it The Astrophysical Journal Letters}}\\
66: \vspace*{1.0em}
67: %-----------------------------------------------------------------------
68: On Ultrasmall Silicate Grains in the Diffuse Interstellar Medium\\
69: % {\small DRAFT: \today}%
70: }
71:
72:
73: \author{Aigen Li and B.T. Draine}
74: \affil{Princeton University Observatory, Peyton Hall,
75: Princeton, NJ 08544, USA;\\
76: {\sf agli@astro.princeton.edu,
77: draine@astro.princeton.edu}}
78:
79: \begin{abstract}
80: The abundance of both amorphous and crystalline silicates
81: in very small grains is limited by the fact
82: that the 10$\mu$m silicate emission feature is not detected in
83: the diffuse ISM.
84: On the basis of the observed IR emission spectrum
85: for the diffuse ISM,
86: the observed ultraviolet extinction curve,
87: and the 10$\mu$m silicate absorption profile, we obtain
88: upper limits on the abundances of ultrasmall
89: ($a\ltsim15\Angstrom$) amorphous and crystalline
90: silicate grains.
91: Contrary to previous work, as much as $\sim$20\% of
92: interstellar Si could be in $a\ltsim 15\Angstrom$ silicate grains
93: without violating observational constraints.
94: Not more than $\sim$5\% of the Si can be in crystalline silicates
95: (of any size).
96:
97: \end{abstract}
98:
99: \keywords{dust, extinction --- infrared: ISM: lines and bands
100: --- ISM: abundances --- ultraviolet: ISM}
101:
102: \section{Introduction}
103: The idea of interstellar silicate grains can be traced back to
104: the 1960's when Kamijo (1963) first proposed that SiO$_2$, condensed
105: in the atmospheres of cool stars and blown out into the interstellar
106: space, could provide condensation cores for the formation of ``dirty
107: ices''. It was later shown by Gilman (1969) that grains around
108: oxygen-rich cool giants are mainly silicates such as Al$_2$SiO$_3$
109: and Mg$_2$SiO$_4$. Silicates were first detected in emission in
110: M stars (Woolf \& Ney 1969; Knacke et al.\ 1969), in the Trapezium
111: region of the Orion Nebula (Stein \& Gillett 1969), and in comet
112: Bennett 1969i (Maas, Ney, \& Woolf 1970); in absorption toward the
113: Galactic Center (Hackwell, Gehrz, \& Woolf 1970), and toward the
114: Becklin-Neugebauer object and Kleinmann-Low Nebula (Gillett
115: \& Forrest 1973).
116: Silicates are now known to be ubiquitous, seen in
117: interstellar clouds, circumstellar disks around young stellar objects
118: (YSOs), main-sequence stars and evolved stars, in HII regions, and in
119: interplanetary and cometary dust.
120:
121: The observed broad, featureless profile of the 10$\mu$m Si-O stretching mode
122: indicated that interstellar silicates are predominantly amorphous.
123: With the advent of {\it Infrared Space Observatory} (ISO)
124: and the improved ground-based spectrometers, the presence of
125: crystalline silicates has been revealed in six comets (see Hanner 1999
126: for a summary), in dust disks around main-sequence stars (see Artymowicz
127: 2000 for a summary), YSOs, and evolved stars (see Waelkens, Malfait,
128: \& Waters 2000 for a summary), in interplanetary dust particles (IDPs)
129: (Bradley et al.\ 1999), and probably also in the Orion Nebula (Cesarsky
130: et al.\ 2000).
131:
132: If silicate grains with radii $a\ltsim 15\Angstrom$ -- which we will
133: refer to as ``ultrasmall'' -- are present,
134: single-photon heating by starlight (Greenberg 1968)
135: will cause these grains
136: to radiate in the 10$\micron$ feature
137: (see, e.g., Draine \& Anderson 1985).
138: The absence of a 10$\micron$ emission feature in IRAS spectra led
139: D\'esert et al.\ (1986) to conclude that not more than 1\% of Si could be
140: in $a < 15\Angstrom$ silicate grains, and on this basis recent grain models
141: (Duley, Jones, \& Williams 1989; D\'{e}sert, Boulanger, \& Puget 1990;
142: Siebenmorgen \& Kr\"{u}gel 1992; Mathis 1996; Li \& Greenberg 1997;
143: Dwek et al.\ 1997; Weingartner \& Draine 2001; Li \& Draine 2001)
144: have excluded ultrasmall silicate grains as a significant grain component.
145: Nondetection of
146: the 10$\mu$m silicate emission feature
147: in the diffuse ISM by ISO (Mattila et al.\ 1996)
148: and the {\it Infrared Telescope in Space} (IRTS) (Onaka et al.\ 1996)
149: appeared to confirm this conclusion.
150:
151: However, the presence of ultrasmall silicate grains can not be ruled out
152: since the 10$\mu$m silicate emission feature may be hidden by
153: the dominant PAH features.
154: In this {\it Letter} we place quantitative upper limits on
155: the abundances of both amorphous and crystalline ultrasmall silicate
156: grains by calculating spectra for such tiny grains heated
157: by starlight, and comparing to
158: measurements of the infrared (IR) {\it emission}
159: of the diffuse ISM by
160: IRTS (Onaka et al.\ 1996) and by
161: the {\it Diffuse Infrared Background
162: Experiment} (DIRBE) instrument on the {\it Cosmic Background Explorer}
163: (COBE) satellite.
164: The interstellar extinction curve and the 10$\mu$m silicate
165: {\it absorption} profile are also invoked to provide further
166: constraints.
167:
168: In \S\ref{sec:opct_si} we discuss the optical and thermal properties of
169: crystalline silicate material. In \S\ref{sec:upplimt} we calculate the
170: IR emission spectra of ultrasmall silicate grains and deduce upper limits
171: on their abundances based on the constraints provided by the observed IR
172: emission spectrum (\S\ref{sec:upplimt_emsn}), by the interstellar extinction
173: curve (\S\ref{sec:upplimt_ext}), and by the 10$\mu$m silicate absorption
174: profile (\S\ref{sec:upplimt_silabs}). In \S\ref{sec:discussion} we discuss
175: the effects of dust size distributions and the sources of uncertainty.
176: \S\ref{sec:summary} summarizes our main conclusions.
177:
178: \section{Optical and Thermal Properties of Amorphous and
179: Crystalline Silicates\label{sec:opct_si}}
180:
181: We approximate the grains as spherical
182: and use Mie theory to calculate their absorption
183: and scattering properties.
184: The complex refractive index
185: ${\rm m}(\lambda) = {\rm m}^\prime(\lambda) +
186: i~ {\rm m}^{\prime\prime}(\lambda)$ of amorphous
187: silicates is taken from Draine \& Lee (1984).
188:
189: For crystalline silicates, we proceed as follows:
190: for $\lambda \le 0.3\mu$m, we take ${\rm m}^{\prime\prime}$ from
191: Huffman \& Stapp (1973) for crystalline olivine (Mg,Fe)$_2$SiO$_4$;
192: for $0.3< \lambda \le 6\micron$, we take ${\rm m}^{\prime\prime}$ from
193: ``astronomical silicate'' (Draine \& Lee 1984);
194: for $7 < \lambda \le 200\micron$,
195: we take ${\rm m}^{\prime\prime}$ from Mukai \& Koike (1990) for crystalline
196: olivine Mg$_{1.8}$Fe$_{0.2}$SiO$_{4}$; for $\lambda > 200\mu$m,
197: we assume
198: ${\rm m}^{\prime\prime}(\lambda)={\rm m}^{\prime\prime}(200\mu {\rm m})
199: (200\micron/\lambda)^2$.
200: The real part ${\rm m}^\prime(\lambda)$
201: is obtained from ${\rm m}^{\prime\prime}(\lambda)$
202: using the Kramers-Kronig relation.
203: %(Bohren \& Huffman 1983).
204:
205: The specific heat of amorphous silicates is taken from
206: Draine \& Li (2001). For crystalline silicates, following
207: Draine \& Li (2001), we adopt a three-dimensional Debye model
208: with $\Theta \approx 720$K which closely reproduces
209: the specific heat of crystalline Mg$_2$SiO$_4$ measured by
210: Robie, Hemingway, \& Takei (1982).
211:
212: \section{Upper Limits on Ultrasmall Silicate Grains\label{sec:upplimt}}
213:
214: \subsection{IR Emission Spectra\label{sec:upplimt_emsn}}
215:
216: We will now derive upper limits on the abundances of ultrasmall silicate
217: grains in the diffuse ISM based on comparison of the {\it observed}
218: IR emission spectrum with the {\it calculated} emission spectrum of
219: ultrasmall silicate grains. For this purpose, we require a
220: line-of-sight with:
221: 1) a good quality mid-IR emission spectrum; and estimates of
222: 2) the starlight intensity
223: $\chi_{\rm MMP}$ (relative
224: to the Mathis, Mezger, \& Panagia [1983] radiation field),
225: and 3) the hydrogen column density $N_{\rm H}$.
226:
227: \begin{figure}[h]
228: \begin{center}
229: \epsfig{
230: % file=emsn_mirs.cps,
231: % file=emsn_mirs.ps,
232: file=f1.cps,
233: % file=f1.eps,
234: width=\figwidthland,angle=270}
235: \end{center}\vspace*{-1em}
236: \caption{
237: \label{fig:MIRmodel}
238: \footnotesize
239: Spectrum toward $l=44^{\rm o}20^\prime$,
240: $b=-0^{\rm o}20^\prime$ (see text).
241: Diamonds: DIRBE photometry
242: Heavy solid curve: MIRS/IRTS spectrum.
243: Light solid curve: emission calculated for the
244: carbonaceous/silicate grain model of
245: Li \& Draine (2001) heated by
246: diffuse starlight with an intensity $\chi_{\rm MMP}=2$ relative
247: to the Mathis, Mezger, \& Panagia (1983) radiation field.
248: Reddened starlight is labelled ``stellar''; curves labelled
249: B$_{\rm carb}$,
250: B$_{\rm sil}$ are contributions from ``big'' ($a > 250\Angstrom$)
251: carbonaceous and silicate grains; S$_{\rm carb}$, S$_{\rm sil}$
252: are contributions from ``small'' ($a < 250\Angstrom$) carbonaceous
253: and silicate grains.
254: }
255: \end{figure}
256:
257: The diffuse ISM toward
258: $l = 44^{\rm o}20^\prime$,
259: $b=-0^{\rm o}20^\prime$ has been observed
260: by
261: %the Diffuse Infrared Background Experiment (DIRBE) on
262: %the Cosmic Background Explorer (COBE),
263: DIRBE
264: (Hauser et al.\ 1998)
265: and the Mid-Infrared Spectrograph (MIRS) on
266: %the Infrared Telescope in Space
267: IRTS
268: (Onaka et al.\ 1996) has obtained the 4.7--11.7$\mu$m spectrum.
269: Li \& Draine (2001) have modelled the IR emission from the dust in this
270: direction.
271: In Figure \ref{fig:MIRmodel} we show the observed photometry and mid-IR
272: spectrum
273: %toward $l = 44^{\rm o}20^\prime$, $b=-0^{\rm o}20^\prime$,
274: together with the
275: Li \& Draine (2001) model, with a total gas column
276: $N_{\rm H}=4.3\times10^{22}\cm^{-2}$,
277: and a mixture of carbonaceous grains and silicate grains heated by
278: starlight with $\chi_{\rm MMP}\approx 2$.
279: This model is seen to be in good agreement with the
280: PAH emission features seen by IRTS, as well as the DIRBE photometry.
281: The total power radiated constrains the product
282: $\chi_{\rm MMP}\times N_{\rm H}$,
283: while the $\sim$100$\micron$
284: wavelength of peak emission constrains $\chi_{\rm MMP}$.
285:
286: Lacking {\it a priori} expectations concerning the form of the
287: size distribution for ultrasmall silicate grains,
288: we first consider grains of a single size $a$
289: (radius of an equal-volume sphere).
290: Suppose a fraction $\fsi(a)$ of the total silicon abundance
291: (Si/H $\approx 36\times 10^{-6}$; Grevesse \& Sauval 1998)
292: is contained in amorphous silicate grains of radius $a$.
293: Applying the {\it thermal-discrete} method developed in
294: Draine \& Li (2001), we calculate the energy probability
295: distribution function and the resulting IR emission
296: spectrum.
297: The resulting emission spectra are shown in
298: Figure \ref{fig:emsn_acsi}, where for each size $a$ the
299: value of $\fsi$ has been chosen such that this abundance of
300: ultrasmall amorphous silicate grains would not -- by itself --
301: exceed the observed emission at any wavelength.
302: Grains with $a \simlt10\Angstrom$
303: are heated by energetic photons to ``temperatures'' high enough
304: to emit in the 10$\micron$ silicate feature, and their abundances
305: are thus limited by the IRTS spectrum.
306: For grains with $a>10\Angstrom$ the emission peak shifts
307: to longer wavelength, and the DIRBE 25$\micron$ data provides stronger
308: constraints.
309:
310: From Figure \ref{fig:emsn_acsi}a
311: it can be seen that,
312: for $a < 10\Angstrom$, even if a fraction $\fsi=15\%$
313: were in the form of amorphous silicates, the resulting
314: 10$\mu$m Si-O emission feature would not have been apparent in the IRTS
315: spectrum.
316: For $a=10\Angstrom$, $\fsi<25\%$ is permitted.
317: For $10< a < 20\Angstrom$, the DIRBE 25$\micron$
318: photometry allows up to $\fsi \simlt 30\%$,
319: and for $20<a< 25\Angstrom$, $\fsi\simlt 40\%$.
320:
321: Similarly, we can estimate $\fsi(a)$ for ultrasmall crystalline
322: silicate grains. As shown in Figure \ref{fig:emsn_acsi}b, crystalline
323: Mg$_{1.8}$Fe$_{0.2}$SiO$_{4}$ has pronounced features at
324: 10.0, 11.1, 16.3, 18.6, 23.0
325: and 33.6$\mu$m.
326: For $a < 10\Angstrom$,
327: nondetection of the 10.0$\micron$ feature implies
328: $\fsi \simlt 7\%$.
329: For $10 < a < 20\Angstrom$, an upper limit
330: $\fsi \simlt 15\%$
331: is imposed by the DIRBE 25$\mu$m data,
332: and for $20<a<25\Angstrom$, $\fsi\ltsim20\%$.
333: Interstellar crystalline silicates could presumably differ
334: from the
335: crystalline olivine Mg$_{1.8}$Fe$_{0.2}$SiO$_{4}$
336: considered here,
337: resulting in changes in the strengths and positions of the
338: infrared resonances (e.g., see J\"{a}ger et al.\ 1998 for
339: different crystalline silicate samples).
340: Observations of comet Hale-Bopp (see Hanner 1999) showed an emission
341: feature at 11.2 rather than 11.1$\micron$, and an emission feature at
342: $10.0\micron$, evident in some spectra, was not always apparent.
343: Our limits on $a < 10\Angstrom$ crystalline silicates relied on the
344: strength and position of the 10.0$\micron$ feature; since this feature in
345: interstellar crystalline silicates might
346: be either weaker or shifted in wavelength, it is appropriate to weaken
347: our upper limits for $a < 10\Angstrom$ to $\fsi \simlt 10\%$.
348: Below we will find that the 10$\micron$ absorption profile provides a
349: more restrictive limit on crystalline silicates.
350:
351: %The spectral signatures at longer wavelengths
352: %(16.3, 18.6, 22.9 and 33.5$\mu$m for Mg$_{1.8}$Fe$_{0.2}$SiO$_{4}$)
353: %can be used
354: %as direct measurements of the abundance of interstellar crystalline
355: %silicates. Spectroscopy of the diffuse emission (e.g., by SIRTF)
356: %in the wavelength range $\simgt 12\mu$m will either detect crystalline
357: %silicates in emission, or provide improved upper limits on the abundances
358: %of very small crystalline silicate grains.
359:
360: \begin{figure}[h]
361: \begin{center}
362: \epsfig{
363: % file=emsn_acsi.cps,
364: % file=emsn_acsi.ps,
365: file=f2.cps,
366: % file=f2.eps,
367: width=\figwidth}
368: \end{center}\vspace*{-1em}
369: \caption{
370: \footnotesize
371: \label{fig:emsn_acsi}
372: Contribution to the IR emission toward
373: ($44^{\rm o}\le l \le 44^{\rm o}40^\prime$,
374: $-0^{\rm o}40^\prime\le b \le 0^{\rm o}$) by:
375: (a) amorphous silicate grains with radii
376: $a$=3.5\AA\ ($\fsi=15\%$),
377: 5\AA\ ($\fsi=15\%$),
378: 7.5\AA\ ($\fsi=15\%$),
379: 10\AA\ ($\fsi=25\%$),
380: 12.5\AA\ ($\fsi=30\%$),
381: 15\AA\ ($\fsi=30\%$),
382: 20\AA\ ($\fsi=30\%$),
383: and 25\AA\ ($\fsi=40\%$);
384: (b) crystalline silicate grains with radii
385: $a$=3.5\AA\ ($\fsi=7\%$),
386: 5\AA\ ($\fsi=7\%$),
387: 7.5\AA\ ($\fsi=7\%$),
388: 10\AA\ ($\fsi=15\%$),
389: 12.5\AA\ ($\fsi=15\%$),
390: 15\AA\ ($\fsi=15\%$),
391: 20\AA\ ($\fsi=15\%$),
392: 25\AA\ ($\fsi=20\%$).
393: Diamonds: DIRBE photometry.
394: Heavy solid line:
395: 5--12$\micron$ spectrum observed by IRTS.
396: }
397: \end{figure}
398:
399: \subsection{Ultraviolet Extinction Curve\label{sec:upplimt_ext}}
400:
401: Ultrasmall silicate grains
402: make a steeply-rising contribution
403: to the ultraviolet (UV) extinction.
404: Weingartner \& Draine (2001) and Li \& Draine (2001) have obtained a
405: carbonaceous-silicate grain model which reproduces both
406: the observed extinction and the observed infrared emission.
407: This grain model has only a small abundance $\fsi\approx 1.3\%$ of
408: ultrasmall ($a<15\Angstrom$) silicate grains.
409: In Figure \ref{fig:ext_uv}
410: we show the extinction obtained by {\it adding} ultrasmall
411: silicate grains to this model.
412: Since the ultrasmall grains are
413: in the Rayleigh limit, the added extinction depends only on $\Delta\fsi$
414: (independent of grain size), and applies
415: to both crystalline and amorphous silicates since their
416: optical properties at $\lambda^{-1} \gtsim 5\micron^{-1}$
417: are taken to be similar (see \S\ref{sec:opct_si}).
418:
419: As expected, the
420: contribution to the $\lambda^{-1}\simgt 7\mu {\rm m}^{-1}$ region
421: would be considerable for $\Delta\fsi\gtsim 20$\%
422: (although
423: the effects on the optical and near-UV extinction are
424: negligible).
425: In view of (1) the uncertainties in the observed interstellar
426: extinction in this wavelength region (for example, the determinations
427: of Mathis [1990] and Fitzpatrick [1999] differ from one other by
428: $\sim 5\%$),
429: (2) the fact that the size distributions
430: of other components (carbonaceous grains and $a > 15\Angstrom$ silicates)
431: could be adjusted
432: to reduce their contribution to the far-UV rise,
433: and (3) the possibility that a mixture of ultrasmall silicates
434: (both neutral and charged) might have a more gradual onset of UV absorption
435: than our adopted silicate dielectric function
436: (which has a
437: relatively sharp increase in absorption near 6.5--7.0$\micron^{-1}$),
438: it seems clear that
439: the observed UV extinction does not rule out
440: $\Delta\fsi$ as large as
441: $\sim 20\%$ on lines of sight with $R_V=3.1$.
442:
443: \begin{figure}[h]
444: \begin{center}
445: \epsfig{
446: % file=csiext1.cps,
447: % file=csiext1.ps,
448: file=f3.cps,
449: % file=f3.eps,
450: width=\figwidth}
451: \end{center}\vspace*{-1em}
452: \caption{
453: \footnotesize
454: \label{fig:ext_uv}
455: Comparison of the average Galactic interstellar extinction
456: curve (open circles: Mathis 1990; dotted line: Fitzpatrick
457: 1999) with theoretical extinction curves (solid lines)
458: obtained by adding the contributions from ultrasmall silicate
459: grains with $\Delta\fsi$=5\%, 10\%, 15\%, 20\% and 25\% of the solar
460: silicon abundance (thin solid lines) to our best-fitting model
461: (with $\fsi=1.3\%$;
462: see Weingartner \& Draine 2001; Li \& Draine 2001).
463: Also plotted are the contributions (see Li \& Draine 2001) from
464: ``${\rm B_{sil}}$'' ($a \ge 250$\AA\ silicate);
465: ``${\rm S_{sil}}$'' ($a < 250$\AA\ silicate);
466: ``${\rm B_{carb}}$'' ($a\ge 250$\AA\ carbonaceous);
467: ``${\rm S_{carb}}$'' ($a < 250$\AA\ carbonaceous,
468: including PAHs).
469: }
470: \end{figure}
471:
472: \subsection{Silicate Absorption Profiles\label{sec:upplimt_silabs}}
473:
474: The 10$\mu$m silicate {\it absorption}
475: feature provides
476: another constraint on the abundance of
477: crystalline silicate grains of any size.
478: A recent high signal-to-noise spectroscopic study of
479: lines of sight towards Cyg OB2 \# 12 (diffuse ISM), Elias 16
480: (quiescent molecular cloud) and Elias 18 (embedded YSO) in the
481: Taurus dark cloud appears to rule out any narrow absorption features
482: between 8.4 and 12.5$\mu$m (Bowey et al.\ 1998).
483: Figure \ref{fig:ext_ir}
484: plots the 10$\mu$m silicate absorption profiles resulting from
485: addition of various amounts ($\Delta\fsi$) of ultrasmall crystalline
486: silicates to our best-fitting model (see Weingartner \& Draine 2001;
487: Li \& Draine 2001).
488: The absence of narrow
489: features near 10.0 and 11.1$\micron$
490: in the observed 8.4--12.5$\micron$ extinction
491: (Bowey et al.\ 1998)
492: suggests an upper limit $\Delta\fsi \simlt 3\%$
493: for crystalline silicates.
494: Note that this result is insensitive to size
495: as the grains are in the Rayleigh limit.
496:
497: As noted above, the precise wavelengths and strengths of IR resonance
498: features in crystalline silicates vary with composition, and there could be
499: more than one crystalline form present in
500: the interstellar medium.
501: Nevertheless, the lack of fine structure in the observed broad 10$\micron$
502: absorption feature makes it
503: unlikely that more than $\sim 5\%$ of the interstellar
504: Si could be in crystalline silicates.
505: This limit on crystalline silicates is more stringent than the
506: limit on ultrasmall crystalline silicates obtained in \S\ref{sec:upplimt_emsn}.
507: \begin{figure}[h]
508: \begin{center}
509: \epsfig{
510: % file=csiext2.cps,
511: % file=csiext2.ps,
512: file=f4.cps,
513: % file=f4.eps,
514: width=\figwidth}
515: \end{center}\vspace*{-1em}
516: \caption{
517: \label{fig:ext_ir}
518: \footnotesize
519: The 10$\mu$m silicate absorption profile
520: (thick solid line) for the carbonaceous-amorphous silicate
521: model of Li \& Draine (2001)
522: as well as profiles with an additional contribution from
523: crystalline silicate grains with $\Delta\fsi$=2\%, 4\%, 6\%, 8\%
524: and 10\% of the solar silicon abundance (thin solid lines).
525: The narrow feature at 11.5$\mu$m is due to a lattice
526: mode of crystalline graphite (Draine 1984).
527: Dot-dashed line is the absorption due to amorphous silicate
528: (i.e., with the contributions of other dust components subtracted).
529: Also shown is absorption which would be contributed by
530: crystalline silicate grains for various $\fsi$.
531: Also plotted is the observed extinction profile for dust toward
532: Cyg OB2 \#12 (Bowey et al.\ 1998) divided by
533: $N_{\rm H}=2\times10^{22}\cm^{-2}$ (Whittet et al.\ 1997).
534: }
535: \end{figure}
536:
537: \section{Discussion\label{sec:discussion}}
538:
539: The upper limits on the abundances of ultrasmall silicate
540: grains estimated from the IR emission spectra in \S\ref{sec:upplimt_emsn}
541: are for grains of a single size.
542: Adopting a log-normal size distribution
543: $dn/d\ln a \propto \exp\{-[{\rm ln}(a/a_0)]^2/(2\sigma^2)\}$,
544: we have calculated emission spectra for various $a_0$ and $\sigma$.
545: For this distribution, $a^3 dn/d\ln a$ peaks at $a_p = a_0e^{3\sigma^2}$.
546: For each distribution, we determined the maximum value of $\fsi$ such that
547: the infrared emission did not exceed either the IRTS 4.7--11.7$\micron$
548: spectrum or the DIRBE photometry.
549: Results are shown in
550: Table \ref{tab:si_abun}. It can be seen that $\fsi$ increases with
551: $a_p$:
552: for $3\simlt a_0\simlt6\Angstrom$ and $0.3\simlt \sigma \simlt0.6$,
553: $\fsi$ varies in the range of 20\%--55\%.
554: We take $\fsi\ltsim20\%$ to be the limit on $a<15\Angstrom$ silicate grains.
555:
556: \begin{table}[h]
557: \caption[]{Upper limits on the abundance $\fsi$ of ultrasmall amorphous
558: silicate grains placed by the absence of the 10$\mu$m silicate
559: emission feature (see \S\ref{sec:upplimt_emsn},
560: \S\ref{sec:upplimt_silabs})\label{tab:si_abun}.}
561: \begin{center}
562: \begin{tabular}{c c c c |
563: c c c c |
564: % c c c c |
565: c c c c |
566: c c c c}
567: \hline \hline
568: $a_0$ & $\sigma$ & $a_p$ & $\fsi$ &
569: % $a_0$ & $\sigma$ & $a_p$ & $\fsi$ &
570: $a_0$ & $\sigma$ & $a_p$ & $\fsi$ &
571: $a_0$ & $\sigma$ & $a_p$ & $\fsi$ &
572: $a_0$ & $\sigma$ & $a_p$ & $\fsi$ \\
573: \hline
574: 3\AA& 0.3& 3.9\AA& 20\%&
575: 4\AA& 0.3& 5.2\AA& 20\%&
576: 5\AA& 0.3& 6.5\AA& 20\%&
577: 6\AA& 0.3& 7.9\AA& 20\% \nl
578: 3\AA& 0.4 & 4.8\AA& 20\% &
579: 4\AA & 0.4 & 6.5\AA& 20\% &
580: 5\AA & 0.4 & 8.1\AA& 22\% &
581: 6\AA & 0.4 & 9.7\AA& 30\% \nl
582: 3\AA & 0.5 & 6.4\AA& 20\% &
583: 4\AA & 0.5 & 8.5\AA& 25\% &
584: 5\AA & 0.5 & 10.6\AA& 35\% &
585: 6\AA & 0.5 & 12.7\AA& 45\% \nl
586: 3\AA & 0.6 & 8.8\AA& 30\% &
587: 4\AA & 0.6& 11.8\AA& 40\% &
588: 5\AA & 0.6 & 14.7\AA& 50\% &
589: 6\AA & 0.6 & 17.7\AA& 55\%\nl
590: \hline
591: \end{tabular}
592: \end{center}
593: \end{table}
594:
595: One uncertainty in our upper limit on ultrasmall amorphous
596: silicates arises from uncertainty in the absorption cross section for
597: $1 \ltsim \lambda^{-1} \ltsim 4\micron^{-1}$ where
598: the interstellar starlight spectrum peaks.
599: We have assumed the ultrasmall silicate grains to be characterized by
600: the refractive index estimated for amorphous ``astronomical silicates''
601: (Draine \& Lee 1984; Weingartner \& Draine 2001).
602: We note that the visual absorptivity of crystalline olivine material
603: is sensitive to its Fe content.
604: If -- as suggested by the low Fe content observed in some circumstellar
605: environments (see Waelkens et al.\ 2000) and in comet Hale-Bopp
606: (Crovisier et al.\ 1997) -- {\it crystalline} astronomical
607: silicate material is less absorbing
608: than ``astronomical silicate'' (Draine \& Lee 1984), then
609: photon absorptions will be less frequent and the IR emission
610: correspondingly reduced, in which case the infrared emission limit
611: on ultrasmall silicates would be relaxed.
612: However,
613: for crystalline silicates, the strongest constraint comes from
614: lack of sharp absorption features near 10 and 11$\micron$, so our upper
615: limit is not affected by uncertainty
616: in the 1--4$\micron^{-1}$ absorption.
617:
618:
619: The observed 3.3, 6.2, 7.7, 8.6, and 11.3$\mu$m emission
620: bands (Mattila et al.\ 1996; Onaka et al.\ 1996)
621: require a significant fraction of carbon in polycyclic aromatic
622: hydrocarbon molecules with $a\ltsim15\Angstrom$
623: (the model of Li \& Draine 2001 requires $\approx 40$ carbon
624: atoms per $10^6$ H, or $\sim$10\% of the total C abundance).
625: In contrast, D\'esert et al.\ (1986) argued that not more than $\sim$1\% of
626: the Si could be in silicate particles with radii $a < 15\Angstrom$.
627: Such a paucity of ultrasmall silicate grains
628: would contrast sharply with the apparent large abundance of small
629: carbonaceous particles.
630: The present study, however, finds that as much as 20\% of the Si
631: could be locked up in ultrasmall silicate particles without violating
632: any observational constraints.
633:
634: We suspect that the analysis by
635: D\'esert et al.\ (based on 2--13$\micron$ observations of M82 and NGC 2023)
636: underestimated the fraction of the absorbed starlight energy which is
637: reradiated in the 6--12$\micron$ PAH features.
638: The present study is based on
639: a line of sight with photometry extending to the
640: far-infrared (see Figure \ref{fig:MIRmodel}).
641:
642: \section{Summary\label{sec:summary}}
643:
644: Upper limits on the abundances of ultrasmall amorphous and
645: crystalline silicate grains are estimated based on
646: observed IR emission spectra (\S\ref{sec:upplimt_emsn}),
647: the interstellar extinction curve (\S\ref{sec:upplimt_ext}),
648: and the 10$\mu$m silicate absorption profile (\S\ref{sec:upplimt_silabs}).
649: Our principal results are:
650:
651: \begin{enumerate}
652: \item For amorphous silicate grains,
653: non-detection of the 10$\mu$m silicate emission
654: feature imposes an upper limit
655: on the fraction $\fsi$ of interstellar Si in
656: very small silicate grains:
657: $\fsi \simlt 15\%$ for radii $a<10\Angstrom$, and
658: $\fsi \simlt 20\%$ for $a = 10\Angstrom$,
659: $\fsi \simlt 30\%$ for $10<a<20\Angstrom$,
660: and $\fsi \simlt 40\%$ for $20<a<25\Angstrom$.
661: For a distribution of sizes,
662: $\fsi \ltsim 20\%$ in grains with $a < 15\Angstrom$.
663: These limits are much higher than the earlier estimate by
664: D\'{e}sert et al.\ (1986) of $\fsi\simlt 1\%$ for $a < 15\Angstrom$.
665:
666: \item
667: The observed UV extinction curve in diffuse regions may allow as much as
668: $\fsi \simlt 20\%$ in ultrasmall silicate grains.
669:
670: \item
671: For crystalline silicate grains, the absence of fine structure in the
672: observed 10$\mu$m
673: silicate absorption profile provides the strongest constraint, limiting
674: the abundances of $a\ltsim 1\micron$ crystalline silicate grains
675: to $\fsi \simlt 5\%$.
676: \end{enumerate}
677:
678: Toward high galactic latitude ``cirrus'' both the radiation field
679: and the hydrogen column density are known, but mid-IR spectroscopy
680: is generally unavailable.
681: In the near future, the {\it Space Infrared Telescope
682: Facility} (SIRTF) may be able to obtain
683: 5--40$\micron$ spectra of cirrus regions to
684: either detect silicate emission or obtain
685: more rigorous upper limits on
686: the abundances of very small silicate grains.
687:
688: \acknowledgments
689: We thank L.J. Allamandola, J.M. Greenberg, and J.C. Weingartner
690: for helpful discussions.
691: We thank D.C.B. Whittet for providing us with the 10$\micron$ extinction
692: toward Cyg OB2 \#12.
693: We thank R.H. Lupton for the availability
694: of the SM plotting package. This research was supported in part by
695: NASA grant NAG5-7030 and NSF grants AST-9619429 and AST-9988126.
696:
697: \begin{thebibliography}{}
698: %\bibitem[]{}Allamandola, L.J., Tielens, A.G.G.M., \& Barker, J.R. 1985,
699: % ApJ, 290, L25
700: \bibitem[]{}Artymowicz, P. 2000, Space Sci. Rev., 92, 69
701: %\bibitem[]{}Bohren, C.F., \& Huffman, D.R. 1983, Absorption and
702: % Scattering of Light by Small Particles (New York: Wiley)
703: \bibitem[]{}Bowey, J.E., Adamson, A.J., \& Whittet, D.C.B.
704: 1998, MNRAS, 298, 131
705: \bibitem[]{}Bradley, J.P., Keller, L.P., Snow, T.P., et al.\
706: 1999, Science, 285, 1716
707: \bibitem[]{}Cesarsky, D., Jones, A.P., Lequeux, J., \& Verstraete, L.
708: 2000, A\&A, 358, 708
709: \bibitem[]{}Crovisier, J., Leech, K., Bockelee-Morvan, D.,
710: et al.\ 1997, Science, 275, 1904
711: \bibitem[]{}D\'{e}sert, F.X., Boulanger, F., L\'{e}ger, A., Puget, J.L.,
712: \& Sellgren, K. 1986, A\&A, 159, 328
713: \bibitem[]{}D\'{e}sert, F.X., Boulanger, F., \& Puget, J.L. 1990,
714: A\&A, 237, 215
715: \bibitem[]{}Draine, B.T. 1984, ApJ, 277, L71
716: %\bibitem[]{}Draine, B.T. 1990, in The Evolution of the Interstellar
717: % Medium, ed. L. Blitz (San Francisco: ASP), 193
718: \bibitem[]{}Draine, B.T., \& Anderson, N. 1985, ApJ, 292, 494
719: \bibitem[]{}Draine, B.T., \& Lee, H.M. 1984, ApJ, 285, 89
720: \bibitem[]{}Draine, B.T., \& Li, A. 2001, ApJ, 550, 000
721: (astro-ph/0011318)
722: \bibitem[]{}Duley, W.W., Jones, A.P., \& Williams, D.A.
723: 1989, MNRAS, 236, 709
724: \bibitem[]{}Dwek, E., Arendt, R.G., Fixsen, D.J., et al.\
725: 1997, ApJ, 475, 565
726: \bibitem[]{}Fitzpatrick, E.L. 1999, PASP, 111, 63
727: \bibitem[]{}Gillett, F.C., \& Forrest, W.J. 1973, ApJ, 179, L483
728: \bibitem[]{}Gilman, R.C. 1969, ApJ, 155, L185
729: \bibitem[]{}Greenberg, J.M. 1968, in Stars and Stellar Systems, Vol. VII,
730: ed. B.M. Middlehurst \& L.H. Aller (Chicago: Univ. of Chicago
731: Press), 221
732: %\bibitem[]{}Greenberg, J.M. 1978, in Cosmic Dust, ed. J.A.M. McDonnell
733: % (New York: Wiley), 187
734: \bibitem[]{}Grevesse, N., \& Sauval, A.J. 1998,
735: Space Sci. Rev., 85, 161
736: %\bibitem[]{}Habing, H.J. 1968, Bull. Astron. Inst. Netherlands, 19, 421
737: \bibitem[]{}Hackwell, J.A., Gehrz, R.D., \& Woolf, N.J. 1970,
738: Nature, 227, 822
739: \bibitem[]{}Hanner, M.S. 1999, Space Sci. Rev., 90, 99
740: \bibitem[]{}Hauser, M.G., Kelsall, T., Leisawitz, D., \& Weiland, J. 1998,
741: COBE Diffuse Infrared Background Experiment Explanatory Supplement
742: version 2.3,
743: COBE Ref. Pub. No. 98-A (Greenbelt, MD: NASA GSFC), available
744: in electronic form from the NSSDC
745: \bibitem[]{}Huffman, D. \& Stapp, J.L., 1973, in Interstellar Dust and
746: Related Topics (IAU Symp. 52), eds. J.M. Greenberg,
747: \& H.C. van de Hulst (Dordrecht: Kluwer), 297
748: \bibitem[]{}J\"{a}ger, C., Molster, F.J., Dorschner, J., Henning, Th.,
749: Mutschke, H., \& Waters, L.B.F.M. 1998, A\&A, 339, 904
750: \bibitem[]{}Kamijo, F. 1963, PASP, 15, 440
751: \bibitem[]{}Knacke, R.F., Gaustad, J.E., Gillett, F.C.,
752: \& Stein, W.A. 1969, ApJ, 155, L189
753: %\bibitem[]{}L\'{e}ger, A., \& Puget, J.L. 1984, A\&A, 137, L5
754: %\bibitem[]{}Lewis, R.S., Anders, E., \& Draine, B.T. 1989,
755: % Nature, 339, 117
756: \bibitem[]{}Li, A., \& Greenberg, J.M. 1997, A\&A, 323, 566
757: \bibitem[]{}Li, A., \& Draine, B.T. 2001, submitted to ApJ
758: (astro-ph/0011319)
759: %\bibitem[]{}Li, A., \& Draine, B.T. 2000b, to be submitted to ApJ
760: \bibitem[]{}Maas, R.W., Ney, E.P., \& Woolf, N.J. 1970, ApJ, 160, L101
761: \bibitem[]{}Mathis, J.S. 1990, ARA\&A, 28, 37
762: \bibitem[]{}Mathis, J.S. 1996, ApJ, 472, 643
763: \bibitem[]{}Mathis, J.S., Mezger, P.G., \& Panagia, N.
764: 1983, A\&A, 128, 212
765: %\bibitem[]{}Mathis, J.S., Rumpl, W., \& Nordsieck, K.H.
766: % 1977, ApJ, 217, 425
767: \bibitem[]{}Mattila, K., Lemke, D., Haikala, L.K., et al.\ 1996,
768: A\&A, 315, L353
769: \bibitem[]{}Mukai, T., \& Koike, C. 1990, Icarus, 87, 180
770: \bibitem[]{}Onaka, T., Yamamura, I., Tanabe, T., et al.\ 1996, PASJ, 48, L59
771: \bibitem[]{}Robie, R.A., Hemingway, B.S., \& Takei, H. 1982,
772: American Mineralogist, 67, 470
773: \bibitem[]{}Siebenmorgen, R., \& Kr\"{u}gel, E. 1992, A\&A, 259, 614
774: \bibitem[]{}Stein, W.A., \& Gillett, F.C. 1969, ApJ, 155, L197
775: \bibitem[]{}Waelkens, C., Malfait, K., \& Waters, L.B.F.M. 2000,
776: in Astrochemistry: From Molecular Clouds to Planetary
777: Systems (IAU Symp. 197), ed. E.F. van Dishoeck
778: (Dordrecht: Kluwer), in press
779: \bibitem[]{}Weingartner, J.C., \& Draine, B.T. 2001, ApJ, 548, 000
780: (astro-ph/0008146)
781: \bibitem[]{}Whittet, D.C.B., Boogert, A.C.A., Gerakines, P.A., et al.\ 1997,
782: ApJ, 490, 729
783: %\bibitem[]{}Whittet, D.C.B., Duley, W.W., \& Martin, P.G.
784: % 1990, MNRAS, 244, 427
785: \bibitem[]{}Woolf, N.J., \& Ney, E.P. 1969, ApJ, 155, L181
786: \end{thebibliography}
787: \end{document}