0705.0905/pre.tex
1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3: %\usepackage{apjfonts}
4: %\usepackage{graphicx}
5: 
6: % last modified 5 Apr 07 by GCS
7: % ms09.tex
8: 
9: % to convert from preprint to emulateapj
10: % uncomment \rotate for Table 4-5, remove "*" after "deluxetable" (Tab. 2, 4-5)
11: 
12: \newcommand{\vdag}{(v)^\dagger}
13: \newcommand{\mum}{\ifmmode{\rm \mu m}\else{$\mu$m}\fi}
14: \newcommand{\iras}{{\it IRAS}}
15: \newcommand{\iso}{{\it ISO}}
16: \newcommand{\msx}{{\it MSX}}
17: 
18: \begin{document}
19: 
20: \title{The unusual hydrocarbon emission from the early carbon star 
21: HD 100764:  The connection between aromatics and aliphatics}
22: 
23: \author{
24: G.~C.~Sloan\altaffilmark{1},
25: M.~Jura\altaffilmark{2},
26: W.~W.~Duley\altaffilmark{3},
27: K.~E.~Kraemer\altaffilmark{4},
28: J.~Bernard-Salas\altaffilmark{1},
29: W.~J.~Forrest\altaffilmark{5},
30: B.~Sargent\altaffilmark{5},
31: A.~Li\altaffilmark{6},
32: D.~J.~Barry\altaffilmark{1},
33: C.~J.~Bohac\altaffilmark{5},
34: D.~M.~Watson\altaffilmark{5},
35: \& J.~R.~Houck\altaffilmark{1}
36: %A. Scott\altaffilmark{3,7},
37: }
38: \altaffiltext{1}{Cornell University, Astronomy Department, 
39:   Ithaca, NY 14853-6801; sloan@isc.astro.cornell.edu, 
40:   jbs@isc.astro.cornell.edu, don@isc.astro.cornell.edu, jrh13@cornell.edu}
41: \altaffiltext{2}{Department of Physics and Astronomy and Center 
42:   for Astrobiology, University of California, Los Angeles, CA 
43:   90095-1562; jura@astro.ucla.edu}
44: \altaffiltext{3}{Department of Physics and Astronomy, University of Waterloo, 
45:   200 University Avenue West, Waterloo, ON N2L 3G1, Canada; 
46:   wwduley@uwaterloo.ca}
47: \altaffiltext{4}{Air Force Research Laboratory, Space Vehicles
48:   Directorate, 29 Randolph Rd., Hanscom AFB, MA 01731;
49:   kathleen.kraemer@hanscom.af.mil}
50: \altaffiltext{5}{Department of Physics and Astronomy, University 
51:   of Rochester, Rochester, NY 14627-0171; forrest@pas.rochester.edu,
52:   bsargent@pas.rochester.edu, cbohac@pas.rochester.edu, dmw@pas.rochester.edu}
53: \altaffiltext{6}{Department of Physics and Astronomy, University of 
54:   Missouri Columbia, MO 65211, lia@missouri.edu}
55: %\altaffiltext{7}{COM DEV International, 1725 Woodward Drive, 
56: %  Ottawa, ON K2C 0P9, Canada}
57: 
58: \slugcomment{Accepted in ApJ}
59: 
60: %\slugcomment{PREPRINT}
61: 
62: \begin{abstract}
63: 
64: We have used the Infrared Spectrograph (IRS) on the {\it 
65: Spitzer Space Telescope} to obtain spectra of HD 100764, an 
66: apparently single carbon star with a circumstellar disk.  The 
67: spectrum shows emission features from polycyclic aromatic
68: hydrocarbons (PAHs) that are shifted to longer wavelengths 
69: than normally seen, as characteristic of ``class C'' systems 
70: in the classification scheme of Peeters et al.  All seven of 
71: the known class C PAH sources are illuminated by radiation 
72: fields that are cooler than those which typically excite PAH 
73: emission features.  The observed wavelength shifts are 
74: consistent with hydrocarbon mixtures containing both aromatic 
75: and aliphatic bonds.  We propose that the class C PAH spectra 
76: are distinctive because the carbonaceous material has not 
77: been subjected to a strong ultraviolet radiation field,
78: allowing relatively fragile aliphatic materials to survive.
79: 
80: \end{abstract}
81: 
82: \keywords{ circumstellar matter --- stars:  carbon }
83: 
84: 
85: \section{Introduction} % Sec. 1.0
86: 
87: Complex hydrocarbons are found in a wide variety of 
88: astrophysical environments.  \cite{gfm73} first detected
89: what became known as the unidentified infrared (UIR) 
90: emission features, the strongest of which are at 3.3, 6.2, 
91: 7.7--7.9, 8.6, 11.3, and 12.7~\mum.  \cite{lp84} and 
92: \cite{atb85} identified the carrier of these features as
93: polycyclic aromatic hydrocarbons (PAHs).  The debate over 
94: PAHs as the carrier of the UIR features raged for many 
95: years, and while the identification now seems firm 
96: \cite[e.g.][]{ahs99}, many questions still remain over 
97: the origin, evolution, and exact composition of astrophysical 
98: hydrocarbons.
99: 
100: \cite{pee02} used data from the Short Wavelength Spectrometer 
101: (SWS) on the {\it Infrared Space Observatory} (\iso) to group 
102: PAH spectra into three classes based on the positions of peak 
103: emission and profiles of the C--C skeletal deformation modes 
104: within the PAHs at 6.2 and 7.7--7.9~\mum.  The ``class A'' 
105: sources display features at 6.22 and 7.65~\mum\ and generally 
106: include H II regions, reflection nebulae, galaxies, and 
107: Herbig AeBe (HAeBe) stars still embedded in the regions where
108: they formed.  The ``class B'' sources, which typically 
109: include planetary nebulae and isolated HAeBe stars, show 
110: these same features, but shifted to 6.26 and 7.85~\mum.  
111: \cite{pee02} also identified two ``class C'' sources 
112: including the well-known Egg Nebula (AFGL 2688) which are in 
113: transition from the asymptotic giant branch (AGB) to the 
114: planetary nebula phase.  These two sources show emission at 
115: 6.26~\mum, no emission near 7.65~\mum, and a broad feature 
116: centered at $\sim$8.2~\mum.  It remains unclear how these 
117: different classes of PAHs arise in different astrophysical 
118: environments.
119: 
120: Observations with the Infrared Spectrograph 
121: \citep[IRS;][]{hou04} on the {\it Spitzer Space Telescope} 
122: \citep{wer04} have added to the small sample of class C PAH 
123: spectra.  \cite{kra06}  reported the detection of a class C 
124: PAH spectrum from MSX SMC 029, a transition object in the 
125: Small Magellanic Cloud (SMC).  \cite{jur06} detected a class 
126: C PAH spectrum from HD 233517, an apparently single red giant 
127: (K2 III) with a substantial infrared excess which most likely 
128: arises from a circumstellar disk \citep{jur03}.  This system 
129: is remarkable because some of the hydrocarbons were probably 
130: synthesized from the oxygen-rich disk.  
131: 
132: % \section{HD 100764}
133: 
134: HD 100764 is one of a handful of red giants (luminosity class 
135: III) with large infrared excesses characteristic of orbiting 
136: disks.  While most of these stars are oxygen-rich like 
137: HD 233517, HD 100764 is carbon-rich.  As part of a 
138: program to better understand these stars, we obtained IRS 
139: data for HD 100764.  We have found that its IRS spectrum 
140: reveals low-contrast class C PAH features not detectable in 
141: the previous ground-based spectroscopy.  In this paper, we 
142: describe the PAH spectrum of HD 100764 in detail and relate 
143: it to other class C PAH sources.
144: % and present a model to explain the properties of these systems.
145: 
146: HD 100764, also known as CCS 1886 and CGCS 3066, is 
147: classified as an early carbon star, R2 \citep{san44} or 
148: C1,1 \citep{yam72}.  Using optical spectroscopy, \cite{dom84} 
149: found that the central star has a temperature of 4850 K, is 
150: iron poor, and has an enhanced $^{13}$C/$^{12}$C ratio, 
151: making this carbon star one of the uncommon J stars.  The 
152: Hipparcos parallax of 2.78 ${\pm}$ 1.18 mas yields an 
153: uncertain distance of $\sim$360 pc and thus a total 
154: luminosity of 73 L$_{\odot}$ \citep{par91}, indicating that 
155: the star is a first-ascent red giant.
156: 
157: \cite{le90} identified HD 100764 as a potential 
158: silicate/carbon star because of its red \iras\ [12]$-$[25] 
159: color as measured by the {\it Infrared Astronomy Satellite} 
160: (\iras).  As a consequence, it was observed twice using 
161: ground-based mid-infrared spectrometers over a decade ago.  
162: \cite{lev92} obtained an 8--14~\mum\ spectrum at the
163: Wyoming Infrared Observatory (WIRO) with a resolution 
164: ($\lambda/\Delta\lambda$) of $\sim$50 and a signal/noise 
165: (S/N) ratio of $\sim$5.  While they could rule out the 
166: presence of silicate dust emission at 10~\mum\ in the 
167: spectrum, they could say little more.  \cite{ski94} obtained 
168: a higher quality spectrum with CGS3 at the United Kingdom 
169: Infrared Telescope (UKIRT), also saw no spectral features 
170: from dust, and concluded that the infrared emission was 
171: dominated by amorphous carbon dust in a circumstellar disk.
172: The disk explains why the strong infrared excess does not 
173: appreciably redden the optical spectrum.  
174: 
175: Although uncertain, it is plausible that this disk formed 
176: when HD 100764 became a red giant and engulfed a companion.
177: The luminosity of HD 100764 is consistent with other R
178: stars with measured parallaxes, which \cite{alk98} have
179: determined are first-ascent red giants, not AGB stars. 
180: \cite{mcc97} observed that R stars do not have companions
181: and suggested that their carbon-rich nature and the lack of 
182: a companion could both be explained by coalescence of an
183: expanding giant and its companion.
184: 
185: 
186: \section{The infrared spectra} % Sec. 2.0
187: 
188: \subsection{Observations and analysis} % Sec. 2.1
189: 
190: We observed HD 100764 with both the Short-Low (SL) and 
191: Long-Low (LL) modules of the IRS in standard staring mode on 
192: 2006 January 14.  Because the source is bright for the IRS, 
193: we used the shortest possible ramps in both modules, 6-second 
194: ramps which provide only four samples.  This source was 
195: observed independently with the Short-High (SH) and Long-High 
196: (LH) modules in {\it Spitzer} program 278 (AOR key 19481344), 
197: and we include those data here with permission of the P.I., 
198: M. Werner.  % 3235
199: 
200: We followed the standard calibration method used at Cornell.  
201: We started with the S14 pipeline output from the {\it 
202: Spitzer} Science Center (SSC) and subtracted a sky image from
203: each image before extracting a spectrum.  The high-resolution 
204: observations include separate sky images obtained close to 
205: the target.  For SL, the sky image for one spectral aperture 
206: is obtained when the source is in the other aperture (i.e. 
207: aperture differerences:  SL2$-$SL1 and SL1$-$SL2, where 
208: ``SL2'' means ``SL Order 2'').  For LL, we simply differenced 
209: the images with the source in the same aperture, but in the 
210: other nod position.  For a source as bright as HD 100764, the 
211: primary benefit of background removal is the subtraction of 
212: most of the rogue pixels.  These pixels have temporarily 
213: increased dark currents and are the primary source of noise 
214: in images with bright sources.  We applied {\rm 
215: imclean.pro}\footnote{Available from the SSC as {\rm 
216: irsclean.pro}.} to replace remaining rogue and bad pixels.  
217: 
218: Spectra were extracted from the individual images using 
219: the SSC-defined apertures and the routines {\rm profile}, 
220: {\rm ridge}, and {\rm extract}, which are available with the 
221: {\rm SPICE} software package ({\it Spitzer} IRS Custom
222: Extraction).   To calibrate to flux-density
223: units, we used a spectral correction generated from IRS 
224: observations of standard stars, HR 6348 (K0 III) for SL, and 
225: HR 6348, HD 166780 (K4 III) and HD 173511 (K5 III) for LL 
226: \citep[see ][ for more details]{slo07}.  The high-resolution 
227: data were calibrated using $\xi$ Dra (K2 III).  Error bars 
228: are based on the standard deviation of the spectrum in the 
229: two separate nods in each aperture.  The resulting spectrum 
230: has a S/N ratio which is typically between 200 and 400.
231: 
232: We applied scalar multiplicative corrections to the segments 
233: to remove discontinuities and align them upwards to the 
234: presumably best-centered spectral segment, in this case LL2.  
235: This step required corrections of 9\% to SL2, 2\% to SL1, and 
236: 1\% to LL1.  These corrections indicate the accuracy of the 
237: pointing of the telescope and are consistent with typical
238: observations.  Finally, we combined the bonus order with 
239: valid data in the overlapping orders and removed all data 
240: outside the ranges of trustworthy signal.  Table 1 gives the 
241: wavelength ranges used for each low-resolution order.
242: 
243: % Table 1 goes here
244: 
245: \begin{deluxetable}{ccc} % Table 1
246: \tablecolumns{3}
247: \tablewidth{0pt}
248: \tablenum{1}
249: \tablecaption{Low-resolution wavelength ranges}
250: \tablehead{
251:   \colhead{order} & \colhead{SL (\mum)} & \colhead{LL (\mum)} }
252: \startdata
253:  second & 5.10--7.59  & 14.20--21.23 \\
254:  bonus  & 7.73--8.39  & 19.28--21.23 \\
255:  first  & 7.59--14.33 & 20.46--37.00 \\
256: \enddata
257: \end{deluxetable}
258: 
259: \begin{figure} % Fig. 1
260: \includegraphics[width=3.5in]{f1c.eps}
261: \caption{The spectrum of HD 100764 with the low-resolution
262: modules (blue or black) and with the high-resolution 
263: modules (red or gray).}
264: \end{figure}
265: 
266: The flux density from HD 100764 in the 5--37~\mum\ spectral 
267: region ranges between 5 and 11 Jy, and we have therefore 
268: assessed the  saturation and non-linearities in the response 
269: function, especially in LL.  The S14 pipeline has a greatly 
270: improved correction for these problems, and it produces much 
271: better spectra compared to previous pipeline versions.  
272: Figure 1 presents the full spectrum of HD 100764, both our 
273: low-resolution data and also the high-resolution data 
274: obtained by M. Werner et al.  A comparison of the two spectra 
275: reveals no evidence for any residual artifacts from the
276: saturation and linearity corrections in our low-resolution 
277: spectrum.  One artifact that does remain is the fringing 
278: from the LL1 filter apparent between 21 and 25~\mum.  We
279: do not attempt a correction because the analysis in this
280: paper concentrates on the features at shorter wavelengths.
281: For this paper, we focus on the
282: low-resolution spectrum from 5 to 14~\mum.
283: 
284: Superimposed on the continuum produced by the circumstellar 
285: disk are emission features at wavelengths near those 
286: associated with PAHs and a broad emission feature centered 
287: at 10.6~\mum.  To isolate the PAH-related features we 
288: subtracted a spline-fitted continuum, as Figure 2 shows.  
289: This spline passes through and includes the broad feature 
290: peaking at 10.63~\mum.  We examine this feature separately 
291: in \S 2.5.  The spectrum of HD 100764 also shows a narrow 
292: absorption band from acetylene (C$_2$H$_2$) at 13.7~\mum.  
293: When this absorption band appears with PAH emission features, 
294: it is only in the rare class C PAH spectra \citep{kra06}.
295: 
296: The low-contrast features in the 8--14~\mum\ range escaped 
297: previous detection.  The 8~\mum\ feature has a contrast of 
298: 5\% but lies at the edge of the ground-based window, and the 
299: feature near 11.3~\mum\ has a contrast of only 3\%, too small 
300: to be seen even in the spectrum by \cite{ski94}, which has a 
301: S/N ratio of $\sim$15 at those wavelengths.  
302: 
303: Figure 2 ({\it top}) plots the spectra from the two nod 
304: positions separately.  Systematic errors in the spectra have 
305: their greatest impact on low-contrast features, and they 
306: often show up as differences between the nods.  Some detailed 
307: differences in the PAH features are apparent, most notably at 
308: $\sim$ 8.5~\mum, but these differences are not large enough 
309: to have a significant quantitative effect on our conclusions.  
310: Figure 2 ({\it bottom}) presents the continuum-subtracted 
311: spectrum from HD 100764, along with line segments used to 
312: extract the strength and position of each PAH emission 
313: feature, as described in \S 2.3.
314: 
315: % KK - can we quantify here, e.g. lamba_c?
316: 
317: \begin{figure} % Fig. 2
318: \includegraphics[width=3.5in]{f2c.eps}
319: \caption{The spectrum of HD 100764.  The top panel shows
320: the coadded spectrum and the spectra from the individual nod 
321: positions (offset by 1 and 2 Jy, respectively).  The thin 
322: solid curve is a spline fit to the data at wavelengths marked 
323: with diamonds.  These points lie outside of the PAH features 
324: and C$_2$H$_2$ absorption band at 13.7~\mum.  The bottom panel 
325: shows the resulting continuum-subtracted spectrum, along with 
326: the error bars and the line segments fit to extract the 
327: strengths and centers of the spectroscopic features.  The 
328: line segments are thick in the wavelength regions used to fit 
329: the continuum and thin where the feature is integrated.  The 
330: equivalent width of the 13.7~\mum\ acetylene band is measured 
331: before any continuum subtraction.}
332: \end{figure}
333: 
334: \subsection{The comparison sample} % Sec. 2.2
335: 
336: \begin{deluxetable*}{llrllll} % Table 2
337: \tablecolumns{7}
338: \tablewidth{0pt}
339: \tablenum{2}
340: \tablecaption{HD 100764 and the comparison sample}
341: \tablehead{
342:   \colhead{ } & \colhead{ } & \colhead{ } & \colhead{ } & 
343:   \colhead{Spectral} & \colhead{$T_{eff}$} & \colhead{PAH} \\
344:   \colhead{Target} & \colhead{Instrument} & \colhead{AOR key or TDT} & 
345:   \colhead{Object type\tablenotemark{a}} & \colhead{type\tablenotemark{a}} & 
346:   \colhead{(K)\tablenotemark{a}} & \colhead{class} }
347: \startdata
348: HD 100764       & IRS & 16262656          & red giant       & 
349:   C1,1    &  4850   & C \\
350: IRAS 13416-6243 & SWS & 62803904          & post-AGB        & 
351:   G1 I    &  5440   & C \\
352: AFGL 2688       & SWS & 33800604 35102563 & post-AGB        & 
353:   F5 Ia   &  6520   & C \\
354: MSX SMC 029     & IRS & 10656256          & post-AGB        & 
355:   \nodata & \nodata & C \\
356: HD 233517       & IRS &  3586048          & red giant       & 
357:   K2 III  &  4475   & C \\
358: SU Aur          & IRS &  3533824          & T Tauri         & 
359:   G1      &  5945   & C \\
360: HD 135344       & IRS &  3580672          & Herbig AeBe     & 
361:   F4 Ve   &  6590   & B/C \\
362: HD 44179        & SWS & 70201801          & post-AGB        & 
363:   A0 III  &  9520   & B \\
364: NGC 1333 SVS 3  & SWS & 65902719          & reflection neb. & 
365:   B6      & 14000   & A \\
366: \enddata
367: \tablenotetext{a}{See \S 3.1 for details and references.}
368: \end{deluxetable*}
369: 
370: \begin{figure} % Fig. 3
371: \includegraphics[width=3.5in] {f3c.eps}
372: \caption{Infrared spectra of several sources showing
373: class C PAH features, along with spectra of HD 44179
374: (class B) and NGC 1333 SVS 3 (class A).  The spectra of 
375: IRAS 13416, AFGL 2688, NGC 1333 SVS 3, and HD 44179 are 
376: from the \iso/SWS database \citep[as processed 
377: by][]{slo03}; the remaining spectra are from the IRS.  
378: The thin lines show the continua fit to the spectra, based 
379: on a spline through the wavelengths marked with diamonds.
380: The resulting continuum-subtracted PAH spectra appear 
381: in Fig. 4.  
382: %The letters in parentheses give the class
383: %of PAH emission, as defined by \cite{pee02}
384: }
385: \end{figure}
386: 
387: \begin{figure} % Fig. 4
388: \includegraphics[width=3.5in]{f4c.eps}
389: \caption{The continuum-subtracted spectrum of HD 100764 
390: along with several comparison spectra (also
391: continuum-subtracted).  The name of each source is followed 
392: by the classification of its PAH spectrum.  The vertical 
393: lines are at the centers of the PAH features for a class B 
394: spectrum:  6.26, 7.90, 11.26, and 12.67~\mum.  The majority
395: of the class C spectra show features shifted to the red of
396: these lines.  The dashed lines at 6.85 and 7.25 mark the 
397: positions of aliphatic emission features.  The bottom two 
398: spectra are the prototypical class A and class B spectra from 
399: the \iso/SWS database.  The remainder are the sample of known 
400: class C PAH spectra (or in the case of HD 135344, nearly 
401: class C PAH spectrum).  The extractions of the PAH features 
402: at 12.0 and 12.7~\mum\ in AFGL 2688 are particularly 
403: untrustworthy due to their extremely low contrast and 
404: resultant sensitivity to the location of the spline points 
405: for the continuum.  In the spectrum of AFGL 2688, the bottom 
406: of the 13.7~\mum\ feature has been truncated to prevent it 
407: from overrunning the spectra below.}
408: \end{figure}
409: 
410: To place the class C PAH spectrum from HD 100764 in context,
411: we have also examined several other PAH spectra.  These 
412: comparison spectra include the two class C PAH sources from 
413: the SWS database, AFGL 2688 and IRAS 13416$-$6243, along with 
414: SWS spectra of NGC 1333 SVS 3 and HD 44179 (the Red 
415: Rectangle).  These sources are prototypical examples of class 
416: A and B PAH spectra, respectively.  All four spectra are from 
417: the atlas of SWS spectra produced by \cite{slo03}, and they 
418: have been regridded to a lower spectral resolution.  In the 
419: case of AFGL 2688, the spectrum is an average of two 
420: observations.  The comparison sample also includes the four 
421: sources observed by the IRS which are currently known to have 
422: PAH spectra either in or near class C, including the K giant 
423: HD 233517 \citep{jur06}, the post-AGB source MSX SMC 029 
424: \citep{kra06}, the HAeBe source HD 135344 \citep{slo05}, 
425: and the intermediate-mass T Tauri star SU Aur \citep{fur06}.  
426: 
427: Table 2 presents the comparison sample, including their 
428: spectral classes and effective temperatures, which are
429: described in \S 3.1.
430: 
431: % Table 2 goes here.
432: 
433: \subsection{Extracting the PAH features} % Sec. 2.3
434: 
435: \begin{deluxetable}{ccccc} % Table 3
436: \tablecolumns{5}
437: \tablewidth{0pt}
438: \tablenum{3}
439: \tablecaption{Wavelength intervals used to extract spectral features}
440: \tablehead{
441:   \colhead{} & \multicolumn{2}{c}{For class C sources} &
442:                \multicolumn{2}{c}{For class A and B sources} \\
443:   \colhead{Feature (\mum)} & 
444:   \colhead{$\lambda_{blue}$ (\mum)} & \colhead{$\lambda_{red}$ (\mum)} &
445:   \colhead{$\lambda_{blue}$ (\mum)} & \colhead{$\lambda_{red}$ (\mum)} }
446: \startdata
447: %6.0    &  5.89--5.95  &  6.07--6.13   & \\
448:  6.2    &  5.80--5.95  &  6.65--6.80   & as class C   & 6.54--6.69   \\
449:  7.9    &  7.30--7.60  &  9.00--9.30   & 7.10--7.40   & 8.92--9.17   \\
450:  8.6    &  8.32--8.44  &  8.86--8.98   & as class C   & as class C   \\
451: 11.3    & 10.80--11.00 &  11.85--11.95 & 10.68--10.86 & 11.72--11.90 \\
452: 12.0    & 11.85--11.95 &  12.35--12.60 & 11.72--11.90 & 12.13--12.25 \\
453: 12.7    & 12.35--12.60 &  13.25--13.40 & 12.07--12.19 & 12.98--13.10 \\ 
454: 13.7    & 13.25--13.55 &  13.83--14.13 & as class C   & as class C   \\
455: \enddata
456: \end{deluxetable}
457: 
458: \begin{center}
459: \begin{deluxetable*}{lccccccc} % Table 4
460: %\rotate
461: \tablecolumns{8}
462: \tablewidth{0pt}
463: \tablenum{4}
464: \tablecaption{Central wavelengths of the PAH and C$_2$H$_2$ features}
465: \tablehead{
466:   \colhead{ } & \multicolumn{6}{c}{$\lambda_{c}$ (\mum)} \\
467:   \colhead{Target} & \colhead{6.2~\mum\ PAH} & \colhead{7.7--8.2~\mum\ PAH} &
468:   \colhead{8.6~\mum\ PAH} & \colhead{11.3~\mum\ PAH} & 
469:   \colhead{12.0~\mum\ PAH} & \colhead{12.7~\mum\ PAH} & 
470:   \colhead{13.7~\mum\ C$_2$H$_2$} }
471: \startdata
472: HD 100764      &  6.33$\pm$0.01 &  8.15$\pm$0.06 & \nodata & 
473:                  11.47$\pm$0.07 & 12.14$\pm$0.15\tablenotemark{a} &
474:                  12.83$\pm$0.27\tablenotemark{a} & 13.70$\pm$0.02 \\
475: IRAS 13416     &  6.31$\pm$0.01 &  8.22$\pm$0.02 &  8.71$\pm$0.02 &
476:                  11.42$\pm$0.03 & 12.25$\pm$0.07 & 12.87$\pm$0.10 & 
477:                  13.70$\pm$0.00 \\
478: AFGL 2688      &  6.29$\pm$0.01 &  8.20$\pm$0.01 &  8.68$\pm$0.03 &
479:                  11.36$\pm$0.03 & 12.13$\pm$0.03 & 12.93$\pm$0.03 & 
480:                  13.71$\pm$0.01 \\
481: MSX SMC 029    &  6.25$\pm$0.04 &  8.12$\pm$0.10 &  8.68$\pm$0.05 &
482:                  11.32$\pm$0.07 & \nodata        & \nodata        & 
483:                  13.67$\pm$0.02 \\
484: HD 233517      &  6.28$\pm$0.03 &  8.26$\pm$0.03 &  8.72$\pm$0.07 &
485:                  11.35$\pm$0.02 & \nodata        & 12.75$\pm$0.09 & \nodata \\
486: SU Aur         &  6.26$\pm$0.01 &  8.19$\pm$0.03 &
487:                   8.67$\pm$0.18\tablenotemark{a} & 11.33$\pm$0.01 & 
488:                  12.02$\pm$0.03\tablenotemark{b} & 12.78$\pm$0.08 & \nodata \\
489: HD 135344      &  6.28$\pm$0.03 &  8.08$\pm$0.04 &  8.64$\pm$0.07 &
490:                  11.35$\pm$0.01 & \nodata        & 12.75$\pm$0.08 & \nodata \\
491: HD 44179       &  6.26$\pm$0.00 &  7.84$\pm$0.06 &  8.66$\pm$0.01 &
492:                  11.27$\pm$0.00 & 12.00$\pm$0.04 & 12.71$\pm$0.02 & \nodata\\
493: NGC 1333 SVS 3 &  6.23$\pm$0.00 &  7.68$\pm$0.01 &  8.61$\pm$0.01 &
494:                  11.26$\pm$0.01 & 12.00$\pm$0.02 & 12.70$\pm$0.01 & \nodata\\
495: \enddata
496: \tablenotetext{a}{S/N of extracted feature $<$ 2.}
497: \tablenotetext{b}{S/N of extracted feature $<$ 3.}
498: \end{deluxetable*}
499: \end{center}
500: 
501: \begin{center}
502: \begin{deluxetable*}{lccccccc} % Table 5
503: %\rotate
504: \tablecolumns{8}
505: \tablewidth{0pt}
506: \tablenum{5}
507: \tablecaption{Extracted strengths of the PAH and C$_2$H$_2$ features}
508: \tablehead{
509:   \colhead{ } & \multicolumn{6}{c}{$F$ (10$^{-15}$ W m$^2$)} &
510:   \colhead{$W_{\lambda}$ (10$^{-3}$ \mum) \tablenotemark{a}} \\
511:   \colhead{Target} & \colhead{6.2~\mum} & \colhead{7.7--8.2~\mum} &
512:   \colhead{8.6~~\mum} & \colhead{11.3~\mum} & \colhead{12.0~\mum} & 
513:   \colhead{12.7~\mum} & \colhead{13.7~\mum} } 
514: \startdata
515: HD 100764      &   4.61$\pm$0.24  &   9.15$\pm$0.82  &   \nodata       &
516:                    2.27$\pm$0.28  &   0.32$\pm$0.17  &   0.16$\pm$0.13 &
517:                    3.18$\pm$0.64 \\
518: IRAS 13416     &   57.3$\pm$2.4   &    225$\pm$5     &   1.73$\pm$0.51 &
519:                   22.9 $\pm$1.5   &   7.39$\pm$1.66  &   7.71$\pm$1.54 &
520:                    5.10$\pm$0.48 \\
521: AFGL 2688      &    111$\pm$4     &    860$\pm$10    &   18.6$\pm$1.6  &
522:                     147$\pm$15    &   76.5$\pm$7.4   &    204$\pm$14   &
523:                    6.23$\pm$0.57 \\
524: MSX SMC 029    &   0.16$\pm$0.03  &   0.29$\pm$0.03  & 0.051$\pm$0.007 &
525:                   0.068$\pm$0.012 &  0.005$\pm$0.021 & 0.001$\pm$0.015 &
526:                   14.84$\pm$2.96 \\
527: HD 233517      &   1.46$\pm$0.15  &   1.75$\pm$0.08  &  0.11$\pm$0.03  &
528:                    0.66$\pm$0.04  &   \nodata        & 0.091$\pm$0.037 &
529:                    1.22$\pm$3.09 \\
530: SU Aur         &   6.08$\pm$0.35  &  6.64$\pm$0.35   &   0.14$\pm$0.12 &
531:                    3.77$\pm$0.13  & 0.080$\pm$0.037  &   0.31$\pm$0.08 &
532:                    \nodata \\
533: HD 135344      &   5.44$\pm$0.54  &  9.42$\pm$0.46   &   0.43$\pm$0.12 &
534:                    2.34$\pm$0.06  &  \nodata         &   0.14$\pm$0.04 &
535:                    \nodata \\
536: HD 44179       &   2969$\pm$23    &   5731$\pm$677   &    960$\pm$19   &
537:                    1182$\pm$19    &   48.3$\pm$10.6  &    239$\pm$14   &
538:                    \nodata \\
539: NGC 1333 SVS 3 &    241$\pm$3     &    280$\pm$7     &   69.9$\pm$1.8  &
540:                     172$\pm$5     &   3.52$\pm$0.63  &   57.2$\pm$1.2  &  
541:                    \nodata \\
542: \enddata
543: \tablenotetext{a}{Equivalent width of the acetylene absorption band.}
544: \end{deluxetable*}
545: \end{center}
546: 
547: For each source, we follow the algorithm described for
548: HD 100764 by spline-fitting and removing a continuum,
549: as Figure 3 illustrates.  It has proven necessary to shift 
550: the spline points from one spectrum to the next to avoid 
551: some of the spectral structure which only appears in 
552: individual spectra.  Care must be taken in the case of 
553: IRAS 13416 and AFGL 2688 because the PAH features have such 
554: low contrast with respect to the continuum.  In the case of 
555: AFGL 2688, the features at 12.0 and 12.7~\mum\ must be 
556: treated with caution because their apparent strength and 
557: position shift substantially with the choice of spline 
558: points.  The spline fit to SU Aur removes the silicate 
559: emission feature; this method has been tested for several 
560: similar sources showing both PAH and silicate emission by 
561: \cite{kel07}.  Two of the spectra have been smoothed:
562: MSX SMC 029 with a 3-pixel boxcar, and HD 233517 with a 
563: 5-pixel boxcar.  Figure 4 presents the continuum-subtracted 
564: PAH spectra for HD 100764 and the comparison sample.  
565: 
566: Fitting and removing a spline removes most, but not all, of 
567: the continuum spectrum.  Because we have not forced the fitted 
568: spline to pass precisely through the continuum on either side 
569: of each feature, small residual tilts remain.  While these
570: would have little impact on the extracted strengths of the
571: features, they would shift the measured center.  We therefore
572: continue to use the method applied to a sample of HAeBe
573: stars showing PAH emission by \cite{slo05}.  As they did, we 
574: fit line segments to the continuum on either side of each PAH 
575: feature and integrate the flux above this estimated continuum 
576: to measure the strength of the feature.  We determine the 
577: wavelength with half the emitted flux to either side, which
578: for the remainder of this paper, we call the central 
579: wavelength.
580: 
581: Other authors have analyzed the position of the peak 
582: wavelength, but this result differs from the central 
583: wavelength for two reasons.  First, PAH features are usually
584: asymmetric, so the peak will lie slightly to the blue of the
585: center.  Second, most of the features are blends of multiple
586: transitions.  The feature peaking at $\sim$7.7--7.9~\mum\ 
587: can be analyzed as a combination of features at 7.65 and 
588: 7.85~\mum\ \citep[e.g.][]{pee02,bt05}.  As the 7.85~\mum\ 
589: component increases with respect to the 7.65~\mum\ component, 
590: the central wavelength will move steadily to the red, but the 
591: peak wavelength will not move significantly until the red 
592: component outshines the blue component.  Analyzing the 
593: central wavelength provides better insight about the relative 
594: strengths of the various components of each PAH feature.
595: 
596: % Table 3 goes here.
597: 
598: Table 3 gives the wavelength ranges used to fit the line 
599: segments to each feature; these have changed some from the 
600: ranges used by \cite{slo05}.  We use the class C 
601: wavelengths, for all of the sources except HD 44179 and 
602: NGC 1333 SVS 3, which require shifted wavelengths to extract 
603: the features.  All of the spectra except HD 100764 and SU Aur
604: have a feature at 8.6~\mum\ from the C--H in-plane bending
605: mode on the red shoulder of the C--C feature peaking between 
606: 7.7 and 8.2~\mum.  To eliminate the influence of the 
607: 8.6~\mum\ feature on the position and strength of the broader 
608: feature, we fit a line segment underneath it and include only
609: the flux below that line in our measurement of the 
610: 7.7--8.2~\mum\ feature.
611: 
612: % Table 4 goes here.  
613: % Table 5 goes here.
614: 
615: Table 4 presents the extracted central wavelengths, while
616: Table 5 gives the equivalent fluxes (i.e., integrated flux 
617: densities), and for the acetylene band at 13.7~\mum, the 
618: equivalent width.  In cases where no feature is apparent 
619: or the strength of the extracted feature has a S/N ratio 
620: less than one, Table 4 does not quote a central wavelength.  
621: Table 4 includes notes indicating those wavelengths for 
622: features with S/N ratios less than 2 or 3.  Results for 
623: some of the comparison spectra have been published 
624: recently, but the results presented here represent newer 
625: S14 pipeline output and reflect modifications to the 
626: continuum fitting and extraction wavelengths.
627: 
628: 
629: \subsection{The PAH spectra} % Sec. 2.4
630: 
631: Figures 3 and 4 include the spectra of our class A and B
632: prototypes, NGC 1333 SVS 3 and HD 44179.  Using these and 
633: other SWS data, we find that the C--C modes produce features 
634: in class A spectra with typical central wavelengths of 6.23 
635: and 7.7~\mum, while class B spectra have features centered at 
636: 6.26 and 7.9~\mum\ \citep[in essential agreement 
637: with][]{pee02}.  In the two class C spectra in the SWS 
638: database, these features are shifted to 6.30 and 8.2~\mum, 
639: respectively.  In HD 100764, they are centered at 6.33 and 
640: 8.15~\mum.  In the other class C IRS sources, these features 
641: appear in the 6.25--6.28 and 8.09--8.29~\mum\ ranges.
642: 
643: The features beyond 10~\mum\ arise from out-of-plane C--H
644: bending modes.  The wavelength of these modes depends on the 
645: number of adjacent hydrogen atoms on a single aromatic ring
646: \citep[e.g.][]{atb89}.  Single H atoms produce the solo mode, 
647: which class A and B spectra show between 11.21 and 
648: 11.30~\mum\ with a mean at 11.26~\mum.  Rings with two 
649: adjacent H atoms produce the duo mode, which normally appears 
650: at 12.00~\mum.  Three adjacent H atoms produce the trio mode 
651: which appears at 12.70~\mum\ in the class A and B spectra.  
652: 
653: In all of the class C sources, the solo out-of-plane bending
654: mode has shifted to longer wavelengths.  In the comparison
655: class C PAH spectra observed by the IRS, the features have
656: shifted from the nominal 11.26~\mum\ position to between 11.32 
657: and 11.36~\mum.  In the two class C spectra in the SWS database,
658: the shifts are difficult to measure due to possible effects from 
659: molecular absorption and artifacts.  Our best estimate places the 
660: band centers in the 11.36--11.42~\mum\ range.  In HD 100764 this 
661: feature is centered further to the red, at 11.47~\mum. 
662: 
663: The duo and trio modes also appear to shift in the class C 
664: spectra, but they are fainter and more difficult to measure.  
665: A shift is apparent in both features in the SWS spectra of 
666: AFGL 2688 and IRAS 13416, from 12.0 to $\sim$12.1--12.2~\mum\ 
667: for the duo mode and from 12.7 to $\sim$12.9~\mum\ for the 
668: trio mode, but these results should be treated with 
669: reservation.  The problem is that the contrast of these 
670: features is small, and their apparent strength and position 
671: depend on the details of how we fit a spline to the 
672: continuum.  Of the IRS comparison spectra, SU Aur shows the 
673: clearest duo-mode feature, but the S/N ratio is only $\sim$2 
674: and the shift is minimal.  The spectrum of HD 100764 shows a 
675: more significant shift, but again, the S/N ratio is only 
676: $\sim$2.  The situation with the trio mode is similar in the
677: IRS spectra.  A shift is apparent, but the uncertainties in 
678: central wavelength are large.
679: 
680: In summary, the positions of the PAH features in the spectrum
681: of HD 100764 are fully consistent with a class C PAH spectrum.
682: Two of the features, at 6.33 and 11.47~\mum, represent 
683: extreme positions for our sample.  The spectrum is definitely 
684: not class A or B.
685: 
686: \subsection{Other spectral features}  % Sec. 2.5
687: 
688: In Figures 2 and 4, the spectrum of HD 100764 shows structure
689: between the PAH features at 6.3 and 8.1~\mum\ reminiscent 
690: of the features observed in emission at 6.85 and 7.25~\mum\ 
691: in the spectra of the HAeBe stars HD 34282 and HD 169142 by 
692: \cite{slo05}.  These features are most obvious in the 
693: spectrum of AFGL 2688 (where they can be seen in both scan 
694: directions).  The carriers of the 6.85 and 7.25~\mum\ 
695: features have been identified as aliphatic hydrocarbons 
696: \citep{chi00} and/or hydrogenated amorphous carbon 
697: \citep[HAC;][]{fur99}.  In HD 100764, the features are faint 
698: and close to our uncertainty level.  The 7.25~\mum\ feature 
699: is stronger, which is a little unusual.  As a consequence, 
700: our identification of these features in HD 100764 is only 
701: tentative.  Other class C spectra also show possible spectral
702: structure in this area but with some variations.  For
703: example, IRAS 13416 shows an extra feature (all three appear 
704: in both scan directions).  The class A and B prototypes do 
705: not show the 6.85 and 7.25~\mum\ features.
706: 
707: Excluding the emission features from PAHs and related 
708: molecules, the remaining spectrum from HD 100764 consists of 
709: a cool dust continuum, absorption due to acetylene at 
710: 13.7~\mum, and a broad emission feature centered at 
711: $\sim$10.6~\mum.  
712: 
713: The acetylene band is centered at 13.7~\mum, which is 
714: typical for a carbon star in the Galactic sample, and it has
715: an equivalent width of only 0.0032 $\pm$ 0.0006~\mum, about
716: a factor of 10--20 times weaker than seen in the SWS sample
717: of Galactic carbon stars \citep{slo06}.   The acetylene bands 
718: are apparent in sources with optically thick dust shells, 
719: indicating that the absorbing molecules are not in the 
720: stellar photosphere, but instead intermixed with or above the 
721: emitting dust \citep[][ and references within]{slo06}.  In the 
722: case of HD 100764, it is similarly likely that the acetylene 
723: absorption is not from the stellar photosphere.  Instead,
724: the acetylene is probably distributed in and/or above the 
725: disk.
726: 
727: The broad emission feature to the blue of the 11.47~\mum\
728: PAH feature can be fit reasonably well with a Gaussian
729: centered at 10.63~\mum\ with a full width at half maximum
730: of 1.37~\mum.  As \cite{spe05} have explained, typical SiC 
731: dust grains emit at $\sim$11.3~\mum, but larger SiC grains 
732: emit predominantly at 10.8~\mum\ because the contribution 
733: from the longitudinal optic at 10.8~\mum\ increases with 
734: respect to the transverse optic at 11.5~\mum\ as the grain 
735: size grows.  \cite{slo06} may have observed this effect in 
736: two carbon stars in the SMC, but they were trying to explain 
737: a shift from 11.3 to 11.0~\mum.  A shift past 10.8~\mum\ to 
738: 10.6~\mum\ seems far less likely.  We conclude that this 
739: broad emission feature at 10.6~\mum\ is not likely to arise 
740: from SiC; it remains unidentified.
741: 
742: 
743: \section{Discussion} % Sec. 3.0
744: 
745: \subsection{Dependence of the PAH emission on excitation temperature} % Sec. 3.1
746: 
747: \begin{figure} % Fig. 5
748: \includegraphics[width=3.5in]{f5c.eps}
749: \caption{The correlation between the central wavelength 
750: (defined in \S 2.3) of the two strongest PAH features and the 
751: effective temperature of the exciting star.  The top panel 
752: plots the position of the 7.7--8.2~\mum\ PAH complex, and the 
753: bottom panel plots the position of the 11.3~\mum\ feature.  
754: This figure contains three Herbig AeBe stars not included
755: in Tables 2, 4 or 5.  The IRS data for these sources, 
756: HD 34282, HD 141569, and HD 169142, appear elsewhere 
757: \citep{slo05,kel07}.  HD 135344 is the right-most square.  
758: The two prototypes are NGC 1333 SVS 3 (far left) and HD 44179 
759: (second to left).}
760: \end{figure}
761: 
762: Table 2 gives the spectral types and effective temperatures
763: of our PAH sample.  HD 100764 has been classified as R2 or 
764: C1,1 \citep{san44,yam72} and has an effective temperature of 
765: 4850 K \citep{dom84}.  The two class C sources in the SWS 
766: database appear to be post-AGB objects evolving into 
767: planetary nebulae (PNe).  The central star in AFGL 2688 is 
768: classified as F5~Ia, and the surrounding nebulosity shows no 
769: strong emission lines \citep{cra75}.   IRAS 13416$-$6243 has 
770: a spectral type of G1~I \citep{hu93} and also shows no 
771: ionized emission \citep{kau93}\footnote{\cite{vds00} actually 
772: detected Br$\gamma$ {\it absorption} at 2.17~\mum.}.  
773: \cite{van89} describe it as in transition from the AGB to a 
774: PN.  We have determined temperatures from the spectral 
775: classes for these two objects using the relation calibrated 
776: by \cite{mcw91}.
777: 
778: The PAH spectrum of MSX SMC 029 is best described as class C.
779: While the 6.2~\mum\ feature has not shifted much from the 
780: position seen in class B, the 7--9~\mum\ PAH complex and 
781: 11.3~\mum\ feature are shifted to the red.  Like the post-AGB
782: objects AFGL 2688 and IRAS 13416, MSX SMC 029 shows an 
783: absorption band from acetylene at 13.7~\mum.  \cite{kra06}
784: argue that it is one of the few known post-AGB objects in the 
785: SMC.  It is unfortunate that there are no optical data to help
786: classify the spectrum of the central source.
787: 
788: \cite{slo05} presented the spectra of four HAeBe stars, three 
789: classified as early to mid A dwarfs, and one, HD 135344, as 
790: F4 Ve \citep{dmr97}, which corresponds to an effective 
791: temperature of 6590 K \citep[using the calibration of][Table 
792: A5]{kh95}.  The PAHs around the A dwarfs resemble class B 
793: spectra, except that the 7.85~\mum\ feature is shifted 
794: to 7.9--8.0~\mum, and in the F dwarf HD 135344, this feature 
795: is shifted even further, to 8.08~\mum, making it intermediate
796: between class B and C.  \cite{kel07} have re-examined these 
797: four spectra as part of a larger sample of 18 HAeBe stars and 
798: four related stars, and they confirm these previous results.  
799: As a group, the HAeBe stars exhibit PAH spectra which differ 
800: from the typical class B PAH spectra, with more redshifted 
801: features.  The coolest stars in the sample have spectra which 
802: are even closer to class C.  One source examined by 
803: \cite{kel07} is the intermediate-mass T Tauri star SU Aur, 
804: which has a spectral class of G1 and an effective temperature
805: of 5945 K \citep{cal04}.  \cite{fur06} noted that while 
806: silicate emission dominates its infrared spectrum, emission 
807: from PAHs is apparent at 6.2~\mum\ and on the wings of the
808: 10~\mum\ silicate feature at 8 and 11.3~\mum.  This source is 
809: the coolest star in the sample of \cite{kel07}, and it is 
810: their one source with an unambigious class C PAH spectrum.
811: 
812: HD 233517 has a spectral type of K2~III \citep{fek96}, an 
813: effective temperature of 4475 K \citep{bal00}, and in the
814: infrared, a class C PAH spectrum.  It is the coolest object 
815: in our comparison sample, and the center of its 7--9~\mum\ 
816: PAH complex is also at the longest wavelength observed.  
817: 
818: By comparison, the class A and B prototypes are warmer.
819: NGC 1333 SVS 3 has a spectral class of B6 \citep{hwj84},
820: which corresponds to an effective temperature of 14000 K
821: \citep{kh95}.  HD 44179 has a spectral type of A0 III
822: \citep{coh75}, corresponding to 9520 K \citep{kh95}.  While
823: a range of temperatures have been used to model the Red
824: Rectangle, we adopt 9520 K as most consistent with the
825: optical spectral class. %\cite{vih06} use a value of 8250 K.
826: 
827: Figure 5 plots the positions of the PAH features as a function
828: of the effective temperature of the exciting source.  In 
829: addition to the comparison sample in Tables 2, 4, and 5, 
830: the figure also includes the HAeBe stars HD 34282, HD 141569, 
831: and HD 169142.  \cite{slo05} first published the IRS spectra 
832: of these sources, and the results presented here are from the 
833: more recent analysis by \cite{kel07}.  The correlation of 
834: the position of the 7.7--8.2~\mum\ PAH complex with 
835: effective temperature is impressive.  While the 11.3~\mum\
836: PAH feature also tends to shift to longer wavelength as the
837: temperature decreases, the trend is not so clear, with little
838: difference between the class A and B sources and more
839: scatter in the class C sources.
840: 
841: \subsection{Spatial dependence of the PAHs in extended sources} % Sec. 3.2
842: 
843: In the original PAH model, ultraviolet radiation was required 
844: to excite the UIR emission features \citep[e.g.][]{atb89}.
845: Observations by \citet[][ 2000]{uch98} first observed PAH 
846: emission in cooler radiation fields, with effective 
847: temperatures as low as 6800 K in the case of the reflection
848: nebula vdB 133.  While these observations initially cast some 
849: doubt on the PAH model, \cite{ld02} showed that the absorption 
850: of visible photons by large or ionized PAHs can produce the 
851: observed UIR features.
852: 
853: Since PAH emission often arises in a photo-dissociation 
854: region (PDR), where a front is eating its way into a denser 
855: region and destroying the PAHs in the process, one might 
856: expect a change in the nature of the PAH spectrum as the 
857: temperature of the exciting radiation field varies.  
858: \cite{uch00} did not observe variation in the relative 
859: strengths of the features in their sample, but they did 
860: observe a change in the shape of the 7--9~\mum\ PAH complex.
861: 
862: \cite{bt05} analyzed the shape of the 7--9~\mum\ complex as 
863: a function of position within the reflection nebulae vdB 133, 
864: NGC 1333 SVS 3, and vdB 17 (also within NGC 1333).  They 
865: found that closer to the exciting star, the strength of the 
866: 7.65~\mum\ component increases relative to the 7.85~\mum\ 
867: component, thus shifting the measured center of the 
868: 7.7--7.9~\mum\ PAH feature to the blue.  They also found that 
869: this complex grows stronger relative to the C--H modes at 
870: longer wavelengths near the exciting sources, and suggested 
871: that the wavelength shift at 7.7--7.9~\mum\ might be related 
872: either to the ionization fraction of the PAHs or to PAH 
873: processing by the radiation field, as previously proposed 
874: by \cite{hon01}.  
875: 
876: %KK - raise distinction between UV hardness and UV intensity
877: 
878: \subsection{The possible role of aliphatics} % Sec. 3.3
879: 
880: Hydrocarbons can exist in both aromatic or aliphatic
881: forms.  Aromatic hydrocarbons are dominated by sp$^2$ bonds
882: which produce the polycyclic ring structure that define
883: PAHs.  Aliphatics include sp$^1$ bonds as seen in acetylene
884: and sp$^3$ bonds as seen in methane and longer alkane chains.
885: The sp$^2$ bonds efficiently spread the energy absorbed from
886: a photon over the entire molecule before it can disrupt a
887: single bond, making the aromatic carbon structure in PAHs 
888: more stable than aliphatic bonding structures.  Thus, when a 
889: mixture of aliphatic and amorphous hydrocarbons is exposed to 
890: a harsh radiation environment, we expect the aliphatic bonds 
891: to be broken first.  We propose that the class C spectra 
892: arise from mixtures of aromatic and aliphatic material that 
893: have not yet been subjected to intense ultraviolet radiation, 
894: allowing the more fragile aliphatic bonds to survive.
895: 
896: Consider the class C spectrum of MSX SMC 029.  This source
897: also shows a strong absorption band from acetylene 
898: (C$_2$H$_2$) at 13.7~\mum\ centered within a broader 
899: absorption band in the 12--16~\mum\ region \citep[see][Fig. 
900: 4]{kra06}.  \cite{kra06} argue that this broader band 
901: is due to other aliphatic molecules seen in a similar 
902: spectrum at higher spectral resolution in SMP LMC 11, a 
903: source in Large Magellanic Cloud in transition to the early 
904: planetary nebula stage \citep[][Fig. 1]{jbs06}.  Acetylene 
905: appears in other class C PAH spectra, including AFGL 2688, 
906: IRAS 13146, and our spectrum of HD 100764\footnote{It is 
907: missing from HD 233517, which is an oxygen-rich source, and 
908: the disks around HD 135344 and SU Aur.}.  The association of
909: the acetylene absorption band at 13.7~\mum\ and the broader 
910: 12--16~\mum\ absorption with class C PAH emission is 
911: consistent with our proposal that the class C spectra arise 
912: from mixtures of PAHs and aliphatics.
913: 
914: The ground-based mid-infrared spectrum of HD 56126 
915: (IRAS 07134+1005) shows red-shifted PAH features similar to 
916: our IRS spectrum of HD 100764, with emission both in a 
917: plateau at 8.0--8.2~\mum\ and from out-of-plane bending modes 
918: at 11.4 and 12.2~\mum\ \citep{jus96}.  Although not as 
919: pronounced, the infrared spectra of IRAS 04296+3429 and 
920: IRAS 05341+0852 display similar unusual characteristics.  
921: All three of these sources are post-AGB objects.  HD 56126 
922: is classified as F5 I \citep{nas65}, while IRAS 05341 and 
923: IRAS 04296 are F4 Iab: and G0 Ia, respectively.  These 
924: sources also show unusual spectra in the 3~\mum\ range, which 
925: is dominated by C--H stretching modes.  IRAS 05341 shows a 
926: feature at 3.40~\mum\ that dominates the normally stronger 
927: feature at 3.29~\mum\ \citep{gv90}, while IRAS 04296 also 
928: shows an unusually strong 3.40~\mum\ component 
929: \citep{geb92}.  The 3.29~\mum\ feature arises from an 
930: aromatic C--H stretching mode \citep[e.g.][]{dw81,atb89}, 
931: while the 3.4~\mum\ feature arises from a similar mode in 
932: aliphatic C--H bonds \citep{dw81,jm90,geb94}.  
933: 
934: In the Orion Bar, \cite{geb89} observed an increase in the 
935: ratio of the aliphatic 3.4~\mum\ feature to the aromatic 
936: 3.29~\mum\ feature in the shielded molecular zone compared 
937: with the ionized zone.  \cite{slo97} verified these 
938: observations with long-slit 3~\mum\ spectroscopy and showed 
939: that the 3.4~\mum\ feature actually arises from multiple 
940: aliphatic components.  \cite{job96} obtained spatially 
941: resolved 3~\mum\ spectroscopy in the reflection nebulae 
942: NGC 1333 and NGC 2023 and found that where the ultraviolet 
943: radiation field was less intense, the 3.4~\mum\ feature is 
944: stronger compared to the 3.29~\mum\ feature.  All of these 
945: authors argued that the radiation field is altering an 
946: initial mix of aliphatic and aromatic bonds by destroying the 
947: more fragile aliphatic bonds.  
948: 
949: %(HD 56126 and IRAS 04296 also show a strong 21~\mum\ feature.)
950: 
951: \subsection{Hydrogenated amorphous carbon} % Sec. 3.4
952: 
953: \begin{figure} % Fig. 6
954: \includegraphics[width=3.5in]{f6c.eps}
955: \caption{A laboratory spectrum of hydrogenated amorphous 
956: carbon heated to 800 K\citep{sco97}.  The two portions of 
957: the spectrum separated by the vertical hatched line at 
958: 7.5~\mum\ were obtained separately and have been arbitrarily 
959: normalized here to the same vertical range.  The apparent 
960: relative strengths of the two halves are not meaningful.  The
961: positions of the bands, however, are a good indication of
962: how the wavelengths of the various modes responsible
963: for the observed PAH features may shift as aliphatic bonds
964: are attached to the aromatic hydrocarbons that form the
965: basis of free PAH molecules.  Specifically, the two strongest
966: features at 6.32 and 11.45~\mum\ have shifted from 6.23 and
967: 11.26~\mum\ in more normal PAH spectra.  The vertical lines
968: mark the positions of the 6.85 and 7.25~\mum\ features 
969: seen in spectra of cooler HAC samples.}
970: \end{figure}
971: 
972: Hydrogenated amorphous carbon (HAC) is a generic name for a 
973: mixture of aliphatic and aromatic carbon, consisting of 
974: PAH clusters embedded within a matrix of aliphatically bonded
975: material \citep[see ][ and references therein]{jdw90}.  This
976: description is similar to the material we are proposing to
977: explain the class C PAH spectra.
978: 
979: Compared to previously obtained class C PAH spectra, 
980: HD 100764 represents an extreme case because its features are 
981: more consistently red-shifted and its exciting radiation field 
982: is one of the coolest known.  The PAHs around this source must 
983: be relatively unprocessed by energetic photons, making this 
984: source an excellent candidate for the presence of HAC-like 
985: structures.
986: 
987: Figure 6 shows two representative laboratory spectra of a HAC 
988: sample at $\sim$ 800 K \citep{sco97}.  HAC films with a 
989: thickness of $\sim$1~\mum\ were prepared by ablating graphite 
990: in a hydrogen atmosphere with an excimer laser as reported 
991: previously \citep{sd96,sdj97}.  Films were deposited on a 
992: stainless steel substrate, placed in a vacuum chamber, and 
993: heated by conduction to $\sim$800 K.  Infrared emission 
994: spectra were obtained with a % Bomem MB-100 
995: Fourier-transform spectrometer at a resolution of 4 cm$^{-1}$ 
996: over a spectra range of 500--6000 cm$^{-1}$ (1.7--20~\mum).  
997: The spectra were calibrated by taking the ratio of the 
998: emission from HAC to that of an uncoated stainless steel 
999: substrate at the same temperatures and making a cubic 
1000: polynomial baseline correction.  The two spectral segments 
1001: from 5.5 to 8.0 and 7.5 to 14~\mum\ in Figure 6 were obtained 
1002: separately and their relative intensities are unknown.  We 
1003: have simply scaled both segments to a maximum of 1.0 
1004: arbitrarily in the figure.  
1005: 
1006: While the relative band strengths in Figure 6 are not well
1007: determined, the positions are more clear.  The emission 
1008: features have shifted to the red, with the nominal 6.23 and 
1009: 11.26~\mum\ features appearing at 6.32 and 11.45~\mum, 
1010: respectively.  If the features in the 12--13~\mum\ range can 
1011: be attributed to the duo and trio C--H out-of-plane bending 
1012: modes, then they have shifted from 11.99 and 12.67~\mum\ to 
1013: 12.39 and 13.43~\mum.  The feature centered at 8.53~\mum\ 
1014: could be a red-shifted C--C mode, or it could simply be the 
1015: C--H in-plane bending mode.  This laboratory HAC spectrum
1016: is consistent with the wavelength shifts we have observed in
1017: class C PAH sources.
1018: 
1019: \subsection{PAHs, HAC, and the class C spectrum} % Sec. 3.5.
1020: 
1021: \begin{figure} % Fig. 7
1022: \includegraphics[width=3.5in]{f7c.eps}
1023: \caption{Laboratory spectra of two simple PAHs:  naphthalene 
1024: and benzene with two adjacent methyl side groups 
1025: \citep[from][]{cob05}.  The subsitution of aliphatic 
1026: sidegroups on benzene adds a second C--H stretching mode 
1027: at 3.4~\mum, produces a strong absorption at 6.8~\mum, 
1028: shifts the quartet out-of-plane C--H bending mode from 
1029: 12.8 to 13.5~\mum, and produces more spectral structure 
1030: in the vicinity of the observed 6.85 and 7.25~\mum\
1031: aliphatic features (indicated with solid vertical lines).}  
1032: \end{figure}
1033: 
1034: Bonds in aromatic and aliphatic species have resonances at 
1035: different wavelengths \citep[see][ and the many references 
1036: therein]{bel78}.  In the case of the C--H stretching 
1037: fundamental, the shift from aromatic to aliphatic is always 
1038: to the red, from 3.29 to 3.40~\mum, but for other vibrational 
1039: modes, the shifts can be in either direction, or both, 
1040: depending on the molecules in question.  We hypothesize that 
1041: at longer wavelengths, the substitution of aliphatic 
1042: sidegroups on the edges of large PAH molecules shifts the 
1043: out-of-plane aromatic bending modes to the red.  While 
1044: laboratory data of large PAHs with aliphatic substitutions 
1045: are lacking, data for some of the simplest PAHs are 
1046: available, and as Figure 7 shows, are consistent with our 
1047: hypothesis.
1048: 
1049: Using data from the NIST Chemistry WebBook, maintained by the
1050: National Institute for Standards and Technology \footnote{At 
1051: http://webbook.nist.gov/chemistry/}, we show in Figure 7 the 
1052: spectra of naphthalene and 1,2-dimethylbenzene, which is 
1053: benzene with two adjacent subsitutions of methyl sidegroups.  
1054: The two spectra in Figure 7 were obtained from vapor samples 
1055: using a Fourier-transform infrared spectrometer at a 
1056: resolution of 4 cm$^{-1}$ \citep{cob05}.  The napthalene 
1057: was at a temperature of 518 K, while the 1,2-dimethylbenzene 
1058: was at 423 K.  Since both molecules have four adjacent H 
1059: atoms on each ring, the only out-of-plane C--H bending mode
1060: they show in the 11--14~\mum\ region is the quartet mode.
1061: This mode has shifted from 12.8~\mum\ in naphthalene to 
1062: 13.5~\mum\ in the simpler benzene molecule with the methyl 
1063: sidegroups, consistent with the anticipated effect of the  %%% REFERENCE?
1064: substitution of a methyl sidegroup.  However, this spectral 
1065: difference may also result from the absence of the second 
1066: ring in naphthalene, making further analysis of larger 
1067: molecules highly desirable.  
1068: 
1069: We also note that the presence of aliphatic C--H stretching 
1070: modes in the methyl sidegroups on the benzene molecule has 
1071: added emission at 3.4~\mum\ not present in the naphthalene.  
1072: Furthermore, in the 6--8~\mum\ region, the strongest emisson 
1073: feature in naphthalene is at 6.6~\mum, but a stronger feature 
1074: at 6.8~\mum\ is apparent in the methyl-substituted benzene.  
1075: As described by \cite{bel78} the carbon skeletal breathing 
1076: modes in aromatics cluster in the 6.2 to 6.9~\mum\ range 
1077: while similar modes from methyl (CH$_3$) and methylene 
1078: (CH$_2$) groups cluster in the 6.8--7.5~\mum\ region.  As
1079: noted in \S2.4, we may have detected these methyl and 
1080: methylene features in emission in some of the class C 
1081: PAH spectra, including HD 100764.
1082: 
1083: Laboratory HAC spectra by \cite{fur99} clearly show emission
1084: features at 6.85 and 7.25~\mum.  The HAC spectrum in Figure 6
1085: only shows a single emission feature at 7.0~\mum\ at about
1086: half the strength of the 6.3~\mum\ feature.  The differences
1087: in the spectra probably arise in how the samples were treated
1088: and measured.  \cite{fur99} obtained absorption spectra from
1089: samples at 300 K, while the HAC spectrum in Figure 6 was
1090: observed in emission at $\sim$800 K.  This higher temperature 
1091: has probably modified the simple groups responsible for the 
1092: 6.85 and 7.25~\mum\ features, and it has also enhanced the 
1093: C$-$C modes at 6.2~\mum. 
1094: 
1095: % AL suggests we give the H/C ratio
1096: 
1097: HAC samples have a number of degrees of freedom, including 
1098: the ratio of aliphatic and aromatic bonds and the size 
1099: distribution and structure of the embedded PAHs among others, 
1100: so it is not surprising that no HAC sample matches the 
1101: observed astrophysical spectra in detail.  It is worth 
1102: recalling that no laboratory PAH spectrum provides a precise 
1103: match to astrophysical PAH spectra either.  Available data 
1104: are consistent with the hypothesis that increasing the ratio 
1105: of aliphatic to aromatic bonds in a hydrocarbon mixture 
1106: explains the differences between class B and class C PAH 
1107: spectra.  
1108: 
1109: We propose that complex astrophysical hydrocarbons are 
1110: originally synthesized as some form of HAC.  When this 
1111: material is exposed to harsh radiation fields, the aliphatic 
1112: bonds are destroyed and only free PAHs remain.   \cite{sdp97}
1113: provide evidence that this process could occur.  They exposed 
1114: HAC mixtures to UV radiation and found that the dissolution 
1115: of the HAC led to the production of large PAH molecules with 
1116: attached aliphatic sidegroups.  \cite{hu06} have recently 
1117: shown that these mixtures of aromatic and aliphatic bonds 
1118: have the potential for reproducing the observed features 
1119: in the 6--9~\mum\ range.  These modified PAHs have less than 
1120: a few hundred atoms and still respond to the absorption of 
1121: single photons like more typical free PAHs do.
1122: 
1123: Two separate processes are responsible for the PAH spectra
1124: we observe.  First, the PAHs must be freed from the aliphatic
1125: matrix in which they are initially embedded.  Second, the 
1126: features must be excited.  This latter step requires less 
1127: energy than that needed to completely strip the PAHs of all 
1128: aliphatic side-chains.  The positions of the observed PAH 
1129: features depend on how completely the PAHs have been freed 
1130: from the underlying matrix.  In the class C PAH sources, the 
1131: aliphatic side groups have only been partially destroyed.
1132: 
1133: The sample of isolated HAeBe stars represent a stage in the
1134: evolution of PAHs between class C and class B.  The 
1135: 7.85~\mum\ feature is shifted measurably to 7.9--8.0~\mum\ or 
1136: beyond \citep{slo05,kel07}.  Here again, the PAHs may not be 
1137: entirely stripped from the underlying aliphatic matrix, since 
1138: the A and F stars exciting the PAH spectrum may not produce 
1139: enough photons with sufficient energy to break all of the 
1140: aliphatic bonds.  
1141: 
1142: \subsection{The problem of the reflection nebulae} % Sec. 3.6
1143: 
1144: RNe typically do not show class C PAH spectra, even though
1145: the exciting source can be as cool as 6800 K \citep{uch98}.
1146: The two RNe in the sample examined by \cite{pee02} are both 
1147: class A, as are the three spectra presented by \cite{uch00}.  
1148: The objects studied by \cite{bt05} include one from 
1149: Uchida et al., and all three showed variations in the 
1150: position of the 7.7--7.9~\mum\ feature with radiation 
1151: field, but not enough to move them even as far as a true 
1152: class B spectrum.
1153: 
1154: We suggest that while the PAHs in these complex, evolving 
1155: environments are now subject only to cool radiation fields, 
1156: they may have been exposed in the recent past to a large 
1157: dose of more intense ultraviolet radiation.  RNe are 
1158: typically diffuse and are associated with complicated
1159: star-forming environments.  This diffuse nature means that
1160: the PAHs are poorly shielded from both current and past 
1161: radiation fields.  Furthermore, PAHs in a diffuse environment 
1162: are not able to regenerate any broken aliphatic bands. 
1163: 
1164: %While reflection nebulae may be the strongest example of
1165: %the interplay of the processing and excitation mechanisms 
1166: %necessary for PAH emission, it is unlikely that they are the
1167: %only example.  
1168: The typical PAH emission source is a PDR in a star formation 
1169: region, and the emitting materials in these environments must 
1170: have a complicated history of UV irradiation.  This history 
1171: blurs the link between the current PAH composition and the 
1172: current radiation field, which is probably why the 
1173: correlation illustrated in Fig. 5 has not been noticed 
1174: before.  Only in more isolated sources such as those examined 
1175: here is the link clear.
1176: 
1177: 
1178: \section{Conclusions} % Sec. 4.0
1179: 
1180: The PAH spectrum of HD 100764 adds to the small, but growing
1181: sample of class C PAH spectra, as defined by \cite{pee02}.  
1182: It has one of the coolest exciting stars and some of the
1183: most redshifted features in the class.  Since all 
1184: seven of the known class C spectra are excited by relatively 
1185: cool stars of spectral class F or later, we argue that the
1186: hydrocarbons in these systems are unusual because they have 
1187: not been exposed to much ultraviolet radiation.  
1188: 
1189: We hypothesize that carbonaceous materials are synthesized as 
1190: large HAC conglomerates, and that aliphatic bonds are 
1191: subsequently broken in harsh radiation fields.  We suggest 
1192: that PAH spectra evolve over time; the hydrocarbons in 
1193: class C objects are relatively protected and unprocessed 
1194: while class B and A PAHs have been exposed to more energetic 
1195: photons and are more processed.
1196: 
1197: 
1198: \acknowledgments
1199: 
1200: We are grateful to the referee, whose careful reading and
1201: thoughtful comments led to considerable improvements in the
1202: manuscript.  We would also like to thank T. Geballe and 
1203: P. Roche for useful comments and suggestions and M. Werner 
1204: for generously providing us with his high-resolution spectra 
1205: of HD 100764.  These observations were made with the {\it 
1206: Spitzer Space Telescope}, which is operated by JPL, California 
1207: Institute of Technology, under NASA contract 1407 and 
1208: supported by NASA through JPL (contract 1257184).  Support for 
1209: this work was provided by NASA through contract 960803 issued 
1210: by JPL/Caltech.  WMD is supported by a grant from the NSERC 
1211: of Canada.  AL is supported in part by the NASA/{\it Spitzer} 
1212: theory program.  This research has made use of the SIMBAD 
1213: database and VizieR tool for catalog access, operated at the 
1214: Centre de Données astronomiques de Strasbourg.
1215: 
1216: 
1217: \begin{thebibliography}{}
1218: 
1219: \bibitem[Alksnis et al.(1998)]{alk98} Alksnis, A., Balklavs, A.,
1220:   Dzervitis, U., \& Eglitis, I. 1998, \aap, 338, 209
1221: \bibitem[Allamandola et al.(1999)]{ahs99} Allamandola, L.~J., Hudgins, 
1222:   D.~M., \& Sandford, S.~A. 1999, \apj, L115
1223: \bibitem[Allamandola et al.(1985)]{atb85} Allamandola, L.~J., Tielens,
1224:   A.~G.~G.~M., \& Barker, J.~R. 1985, \apj, 290, L25
1225: \bibitem[Allamandola et al.(1989)]{atb89} Allamandola, L.~J., Tielens, 
1226:   A.~G.~G.~M., \& Barker, J.~R. 1989, \apjs, 71, 733
1227: \bibitem[Balachandran et al.(2000)]{bal00} Balachandran, S.~C., Fekel, F.~C., 
1228:   Henry, G.~W., \& Uitenbroek, H. 2000, ApJ, 542, 978
1229: \bibitem[Bellamy(1978)]{bel78} Bellamy, L.~J. 1978, The Infra-Red Spectra
1230:   of Complex Molecules (London:  Chapman \& Hall)
1231: \bibitem[Bernard-Salas et al.(2006)]{jbs06} Bernard-Salas, J., Peeters, E.,
1232:   Sloan, G.~C., Cami, J., \& Houck, J.~R. 2006, \apj, 652, L29
1233: \bibitem[Bouwman et al.(2001)]{bou01} Bouwman, J., Meeus, G., de Koter, A.,
1234:   Hony, S., Dominik, C., \& Waters, L.~B.~F.~M. 2001, \aap, 375, 950
1235: \bibitem[Bregman \& Temi(2005)]{bt05} Bregman, J. \& Temi, P. 2005,
1236:   \apj, 621, 831
1237: \bibitem[Calvet et al.(2004)]{cal04} Calvet, N., Muzerolle, J.; Brice\~{n}o, 
1238:   C., Hern\'{a}ndez, J., Hartmann, L., Saucedo, J.~L., \& Gordon, K.~D.
1239:   2004, \aj, 128, 1294
1240: \bibitem[Chiar et al.(2000)]{chi00} Chiar, J.~E., Tielens, A.~G.~G.~M., 
1241:   Whittet, D.~C.~B., Schutte, W.~A., Boogert, A.~C.~A., Lutz, D., van 
1242:   Dishoeck, E.~F., \& Bernstein, M.~P. 2000, \apj, 537, 749
1243: \bibitem[Coblentz Society, Inc.(2005)]{cob05} Coblentz Society, Inc. 2005,
1244:   in NIST Chemistry WebBook, NIST Standard Reference Database Number 69, 
1245:   ed. P.~J. Linstrom \& W.~G.  Mallard, (Gaithersburg, MD:  National 
1246:   Institute of Standards and Technology, http://webbook.nist.gov)
1247: \bibitem[Cohen et al.(1975)]{coh75} Cohen, M., et al. 1975, \apj, 196, 179
1248: \bibitem[Crampton et al.(1975)]{cra75} Crampton, D., Cowley, A.~P., 
1249:   \& Humphreys, R.~M. 1975, \apj, 198, L135
1250: \bibitem[Dominy(1984)]{dom84} Dominy, J.~F. 1984, \apjs, 55, 27
1251: \bibitem[Duley \& Williams(1981)]{dw81} Duley, W.~W. \& Williams, D.~A.
1252:   1981, \mnras, 196, 269
1253: \bibitem[Dunkin et al.(1997)]{dmr97} Dunkin, S.~K., Barlow, M.~J., \&
1254:   Ryan, S.~G. 1997, \mnras, 286, 604
1255: %\bibitem[Engelke(1992)]{eng92} Engelke, C.~W. 1992, \aj, 104, 1248
1256: \bibitem[Fekel et al.(1996)]{fek96} Fekel, F.~C., Webb, R.~A., White,
1257:   R.~J., \& Zuckerman, B. 1996, \apj, 462, L95
1258: \bibitem[Furlan et al.(2006)]{fur06} Furlan, E., Hartmann, L., Calvet, N., 
1259:   D'Alessio, P., Franco-Hern\'{a}ndez, R., Forrest, W.~J., Watson, D.~M., 
1260:   Uchida, K.~I.; Sargent, B., Green, J.~D., Keller, L.~D., \& Herter, T.~L.
1261:   2006, \apjs, 165, 568
1262: \bibitem[Furton et al.(1999)]{fur99} Furton, D.~G., Laiho, J.~W., \& 
1263:   Witt, A.~N. 1999, \apj, 526, 752
1264: \bibitem[Geballe et al.(1994)]{geb94} Geballe, T.~R., Joblin, C., 
1265:   d'Hendecourt, L.~B., Jourdain de Muizon, M., Tielens, A.~G.~G.~M., 
1266:   \& Leger, A. 1994, \apj, 434, L15
1267: \bibitem[Geballe et al.(1989)]{geb89} Geballe, T.~R., Tielens, A.~G.~G.~M., 
1268:   Allamandola, L.~J., Moorhouse, A., \& Brand, P.~W.~J.~L. 1989, \apj, 341, 278
1269: \bibitem[Geballe et al.(1992)]{geb92} Geballe, T.~R., Tielens, A.~G.~G.~M., 
1270:   Kwok, S., \& Hrivnak, B.~J.  1992, ApJ, 387, L89
1271: \bibitem[Geballe \& van der Veen(1990)]{gv90} Geballe, T.~R. \& van der Veen, 
1272:   W.~E.~C.~J. 1990, \aap, 235, 9
1273: \bibitem[Gillett et al.(1973)]{gfm73} Gillett, F.~C., Forrest, W.~J., \& 
1274:   Merrill, K.~M. 1973, \apj, 183, 87
1275: \bibitem[Harvey et al.(1984)]{hwj84} Harvey, P.~M., Wilking, B.~A., \& Joy, M.
1276:   1984, \apj, 278, 156
1277: \bibitem[Hony et al.(2001)]{hon01} Hony, S., Van Kerckhoven, C.,
1278:   Peeters, E., Tielens, A.~G.~G.~M., \& Allamandola, L.~J. 2001, \aap,
1279:   370, 1030
1280: \bibitem[Houck et al.(2004)]{hou04} Houck, J.~R., et al. 2004, \apjs,
1281:   154, 18
1282: \bibitem[Hu et al.(1993)]{hu93} Hu, J.~Y., Slijkhuis, S., de Jong, T., \&
1283:   Jiang, B.~W. 1993, \aaps, 100, 413
1284: \bibitem[Hu et al.(2006)]{hu06} Hu, A., Alkhesho, I., \& Duley, W.~W. 2006,
1285:   \apj, 653, L157
1286: %\bibitem[Jaschek \& Jaschek(1987)]{jj87} Jaschek, C. \& Jaschek, M.
1287: %  1987, The Classification of Stars (Cambridge:  Univ. of Cambridge Press)
1288: \bibitem[Joblin et al.(1996)]{job96} Joblin, C., Tielens, A.~G.~G.~M., 
1289:   Allamandola, L.~J., \& Geballe, T.~R. 1996, \apj, 458, 610
1290: \bibitem[Jones et al.(1990)]{jdw90} Jones, A.~P., Duley, W.~W., \&
1291:   Williams, D.~A. 1990, QJRAS, 31, 567
1292: \bibitem[Jourdain de Muizon et al.(1990)]{jm90} Jourdain de Muizon, M.,
1293:   d'Hendecourt, L.~B., \& Geballe, T.~R. 1990, \aap, 235, 367
1294: \bibitem[Jura(2003)]{jur03} Jura, M. 2003, \apj, 582, 1032
1295: %\bibitem[Jura(2003b)]{jur03b} Jura, M. 2003b, \apj, 584, L91
1296: \bibitem[Jura et al.(2006)]{jur06} Jura, M., Bohac, C.~J., Forrest, W.~J.,
1297:   Green, J., Watson, D.~M., Sloan, G.~C., Markwick-Kemper, F., \&
1298:   Najita, J. 2006, \apj, 637, L45
1299: %\bibitem[Jura et al.(2002)]{jur02} Jura, M., Chen, C., \& Plavchan, P.  
1300: %  2002, \apj, 574, 963
1301: \bibitem[Justtanont et al.(1996)]{jus96} Justtanont, K., Barlow, M.~J., 
1302:   Skinner, C.~J., Roche, P.~F, Aitken, D.~K., \& Smith, C.H. 1996, \aap, 
1303:   309, 612
1304: \bibitem[K\"{a}ufl et al.(1993)]{kau93} K\"{a}ufl, H.~U., Renzini, A., \&
1305:   Stanghellini, L. 1993, \apj, 410, 251
1306: \bibitem[Keller et al.(2007)]{kel07} Keller, L.~D., et al. 2007, \apj,
1307:   almost submitted (really)
1308: \bibitem[Kenyon \& Hartmann(1995)]{kh95} Kenyon, S.~J. \& Hartmann, L. 1995, 
1309:   \apjs, 101, 117
1310: \bibitem[Kraemer et al.(2006)]{kra06} Kraemer, K.~E., Sloan, G.~C.,
1311:   Bernard-Salas, J., Price, S.~D., Egan, M.~P., \& Wood, P.~R. 2006,
1312:   \apj, 652, L25
1313: \bibitem[Leger \& Puget(1984)]{lp84} Leger, A. \& Puget, J.~L. 1984, \aap,
1314:   137, L5
1315: \bibitem[Li \& Draine(2002)]{ld02} Li, A. \& Draine, B.~T. 2002, \apj, 572, 232
1316: \bibitem[Lloyd Evans(1990)]{le90} Lloyd Evans, T. 1990, \mnras, 342, 336
1317: \bibitem[LeVan et al.(1992)]{lev92} LeVan, P.~D., Sloan, G.~C.,
1318:   Little-Marenin, I.~R., \& Grasdalen, G.~L. 1992, \apj, 392, 702
1319: \bibitem[McClure(1997)]{mcc97} McClure, R.~D. 1997, \pasp, 109, 256
1320: \bibitem[McWilliam(1991)]{mcw91} McWilliam, A. 1991, \aj, 101, 1065
1321: \bibitem[Nassau et al.(1965)]{nas65} Nassau, J.~J., Stephenson, C.~B.,
1322:   \& MacConnell, D.~J. 1965, in Luminous Stars in the Northern Milky Way 6
1323:   (Hamburg:  Hamburg Sternwarte and Warner and Swasey Observatory)
1324: %\bibitem[NIST(2005)]{nist05} National Institute of Standards and Technology
1325: %  2005, NIST Chemistry WebBook:  NIST Standard Reference Database No. 69
1326: %  (available at http://webbook.nist.gov/chemistry/)
1327: \bibitem[Parthasarathy(1991)]{par91} Parthasarathy, M. 1991, \aap, 247, 429
1328: \bibitem[Peeters et al.(2002)]{pee02} Peeters, E., Hony, S., Van
1329:   Kerckhoven, C., Tielens, A.~G.~G.~M., Allamandola, L.~J.; Hudgins,
1330:   D.~M., \& Bauschlicher, C.~W. 2002, \aap, 390, 1089
1331: %\bibitem[Pendleton \& Allamandola(2002)]{pa02} Pendleton, Y.~J. \& 
1332: %  Allamandola, L.~J. 2002, \apjs, 138, 75
1333: %\bibitem[Perryman et al.(1997)]{hip97} Perryman, M.~A.~C. et al. 1997, 
1334: %  \aap, 323, L49
1335: %\bibitem[Ridgway et al.(1980)]{ri80} Ridgway, S.~T., Joyce, R.~R., White,
1336: %  N.~M., \& Wing, R.~F. 1980, \apj, 235, 126
1337: \bibitem[Sanford(1944)]{san44} Sanford, R.~F. 1944, \apj, 99, 145
1338: \bibitem[Scott(1997)]{sco97} Scott, A. 1997, Ph.D. Thesis, Univ. Waterloo,
1339:   Walterloo, Ontario
1340: \bibitem[Scott \& Duley(1996)]{sd96} Scott, A. \& Duley, W.~W. 1996, \apj,
1341:   472, L123
1342: \bibitem[Scott et al.(1997a)]{sdp97} Scott, A., Duley, W.~W., \& Pinho, 
1343:   G.~P. 1997a, \apj, 489, L189
1344: \bibitem[Scott et al.(1997b)]{sdj97} Scott, A., Duley, W.~W., \& Jahani,
1345:   H.~R. 1997b, \apj, 490, L175
1346: %\bibitem[Seahra \& Duley(2000)]{sd00} Seahra, S.~S. \& Duley, W.~W. 2000, 
1347: %  \apj, 542, 898
1348: \bibitem[Skinner(1994)]{ski94} Skinner, C.~J. 1994, \mnras, 271, 300
1349: \bibitem[Sloan et al.(1997)]{slo97} Sloan, G.~C., Bregman, J.~D., Geballe, 
1350:   T.~R., Allamandola, L.~J., \& Woodward, C.~E. 1997, \apj, 474, 735
1351: \bibitem[Sloan et al.(2003)]{slo03} Sloan, G.~C., Kraemer, K.~E., 
1352:   Price, S.~D., \& Shipman, R.~F. 2003, \apjs, 147, 379
1353: \bibitem[Sloan et al.(2005)]{slo05} Sloan, G.~C., et al. 2005,
1354:   \apj, 632, 956
1355: \bibitem[Sloan et al.(2006)]{slo06} Sloan, G.~C., Kraemer, K.~E., Matsuura, 
1356:   M., Wood, P.~R., Price, S.~D., \& Egan, M.~P. 2006, \apj, 645, 1118
1357: \bibitem[Sloan et al.(2007)]{slo07} Sloan, G.~C., et al. 2007,
1358:   almost submitted
1359: \bibitem[Speck et al.(2005)]{spe05} Speck, A.~K., Thompson, G.~D., \&
1360:   Hofmeister, A.~M. 2005, \apj, 634, 426
1361: \bibitem[Uchida et al.(1998)]{uch98} Uchida, K.~I., Sellgren, K., \& Werner,
1362:   M. 1998, \apj, 493, 109
1363: \bibitem[Uchida et al.(2000)]{uch00} Uchida, K.~I., Sellgren, K., 
1364:   Werner, M.~W., \& Houdashelt, M.~L. 2000, \apj, 530, 817
1365: \bibitem[van der Veen et al.(1989)]{van89} van der Veen, W.~E.~C.~J.,
1366:   Habing, H.~J., \& Geballe, T.~R. 1989, \aap, 226, 108
1367: \bibitem[Van de Steene et al.(2000)]{vds00} Van de Steene, G.~C., van Hoof,
1368:   P.~A.~M., \& Wood, P.~R. 2000, \aap, 362, 984
1369: %\bibitem[Vijh et al.(2006)]{vij06} Vijh, U.~P., Witt, A.~N., York, D.~G.,
1370: %  Dwarkadas, V.~V., Woodgate, B.~E., \& Palunas, P. 2006, \apj, 653, 1336
1371: \bibitem[Werner et al.(2004)]{wer04} Werner, M.~W., et al. 2004,
1372:   \apjs, 154, 1
1373: \bibitem[Yamashita(1972)]{yam72} Yamashita, Y. 1972, Ann. Tokyo. Astr. Obs.,
1374:   13, 169 % or 1975, 15, 47
1375: \end{thebibliography}
1376: 
1377: \end{document}
1378: