1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: % Chiar et al 2006
3: % 29 June 2006 - revised manuscript
4: % Spectropolarimetry Proceedings
5: % dvips -Ppdf -t letter spectropol.dvi
6: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
7: \documentclass[12pt,preprint]{aastex}
8:
9: %\markboth{Chiar, Adamson, Mason, Wright, Whittet, Chrysostomou,
10: %Hough,Roche}{Spectropolarimetry of the 3.4 \micron\ Band in the Quintuplet}
11:
12: %\setcounter{equation}{0}
13: %\setcounter{figure}{0}
14: %\setcounter{footnote}{0}
15: %\setcounter{section}{0}
16: %\setcounter{table}{0}
17:
18: \slugcomment{Accepted to ApJ, July 3, 2006}
19: \newcommand{\etal}{et\,al.}
20:
21: \begin{document}
22:
23: \title{Spectropolarimetry of the 3.4 \micron\ Feature in the Diffuse ISM toward
24: the Galactic Center Quintuplet Cluster}
25:
26: \author{J.\,E. Chiar\altaffilmark{1}, A.\,J. Adamson\altaffilmark{2}, D.\,C.\,B. Whittet\altaffilmark{3},
27: A. Chrysostomou\altaffilmark{4}, J.\,H. Hough\altaffilmark{4}, \\ T.\,H. Kerr\altaffilmark{2},
28: R.\,E. Mason\altaffilmark{5}, P.\,F. Roche\altaffilmark{6}, and G. Wright\altaffilmark{7}}
29:
30: \altaffiltext{1}{SETI Institute, 515 N. Whisman Drive, Mountain View, CA
31: 94043, and NASA-Ames Research Center, Mail Stop 245-3, Moffett
32: Field, CA 94035}
33:
34: \altaffiltext{2}{Joint Astronomy Centre, 660 N. A'ohoku Place, University Park,
35: Hilo, Hawaii 96720}
36:
37: \altaffiltext{3}{Rensselaer Polytechnic Institute, Department of Physics,
38: Applied Physics and Astronomy, Troy, New York, 12180}
39:
40: \altaffiltext{4}{Department of Physical Sciences, University of Hertfordshire,
41: Hatfield, AL10 9AB, UK}
42:
43: \altaffiltext{5}{Gemini Observatory, 670 N. A'ohoku Place, University Park,
44: Hilo, Hawaii 96720}
45:
46: \altaffiltext{6}{Department of Astrophysics, Oxford University, Denys Wilkinson
47: Building, Keble Road, Oxford OX1 3RH, UK}
48:
49: \altaffiltext{7}{UK Astronomy Technology Centre, Royal Observatory Edinburgh,
50: Blackford Hill, Edinburgh EH9 3HJ, UK}
51:
52:
53: \begin{abstract}
54: Aliphatic hydrocarbons exhibit an absorption feature at 3.4~\micron\ observed
55: toward sources that sample diffuse regions of the interstellar medium. The
56: absorbers responsible for this feature are assumed to reside in some component
57: of interstellar dust, but the physical nature of the particles (size, shape,
58: structure, etc.) is uncertain. Observations of interstellar polarization
59: provide discrimination. Since the grains that carry the silicate absorption
60: feature are known to be aligned, polarization across the 3.4 \micron\
61: hydrocarbon feature can be used to test the silicate core-organic refractory
62: mantle grain theory. Although the 3.4 \micron\ feature has been observed to
63: be devoid of polarization for one line of sight toward the Galactic center, a
64: corresponding silicate polarization measurement for the same line of sight was
65: not available. Here, we present spectropolarimetric observations toward
66: GCS~3-II and GCS~3-IV toward the Galactic center, where the 9.7 \micron\
67: silicate polarization has been previously observed. We show that polarization
68: is not detected across the 3.4 \micron\ feature to a limit of $0.06\pm 0.13$\%
69: (GCS~3-II) and $0.15\pm 0.31$\% (GCS~3-IV), well below the lowest available
70: prediction of polarization on the basis of the core-mantle model. We conclude
71: that the hydrocarbons in the diffuse ISM do not reside on the same grains as
72: the silicates and likely form a separate population of small grains.
73: \end{abstract}
74:
75: \keywords{ISM: clouds --- dust; extinction --- ISM: general --- ISM: individual (GCS 3-II, GCS 3-IV) ---
76: ISM: lines and bands --- ISM: molecules}
77:
78: \section{Introduction}
79:
80: Dust grains in the diffuse interstellar medium are composed of both oxygen-rich
81: and carbon-rich material \nocite{mathis96,draine03} (e.g.\ Mathis 1996; Draine
82: 2003). While the O-rich component is attributed to amorphous silicates, the
83: composition of the C-rich component remains controversial, proposed candidates
84: including graphite, diamond, amorphous carbon, aromatic and aliphatic
85: hydrocarbons, and organic-refractory residues. Of these, aliphatic
86: hydrocarbons are confirmed to be present in the interstellar media in our own
87: and other galaxies through observations of vibration absorption features
88: centered at 3.4 \micron\ (C--H stretch), 6.85 and 7.25 \micron\ (deformation
89: modes). In the Milky Way, hydrocarbons are detected along lines of sight that
90: probe the local diffuse ISM within 3~kpc of the Sun
91: \nocite{adamson_whittet_duley90,sandford_etal91,pendleton_etal94,whittet_etal97}
92: (Adamson \etal\ 1990; Sandford \etal\ 1991; Pendleton \etal\ 1994; Whittet
93: \etal\ 1997) and along the line of sight toward the Galactic Center
94: \nocite{mcfadzean_etal89,sandford_etal91,pendleton_etal94,chiar_etal00,chiar_etal02}
95: (GC; McFadzean \etal\ 1989; Sandford \etal\ 1991; Pendleton \etal\ 1994; Chiar
96: \etal\ 2000, 2002). These same hydrocarbon features are also observed in dusty
97: Seyfert and ultraluminous infrared galaxies
98: \nocite{mason_etal04,imanishi_dudley_maloney06,imanishi00agn,spoon_etal04}
99: (Imanishi 2000; Mason \etal\ 2004; Spoon \etal\ 2004; Imanishi, Dudley, \&
100: Maloney 2006). Within the profile of the 3.4 \micron\ feature, the relative
101: depths of subfeatures at 3.39, 3.42 and 3.49 \micron\ are indicative of
102: short-chained hydrocarbons with other perturbing chemical groups (Sandford
103: \etal\ 1991). Based on the depth of the corresponding deformation features at
104: 6.85 and 7.25 \micron, the best candidate material is likely to be hydrogenated
105: amorphous carbon \nocite{chiar_etal00,pendleton_allamandola02,mennella_etal02}
106: (HAC; Chiar \etal\ 2000; Pendleton \& Allamandola 2002; Mennella \etal\ 2002).
107:
108: Although these aliphatic hydrocarbons seem to be ubiquitous constituents of
109: galactic interstellar media, and much has been learned about their molecular
110: structure, their formation pathways and physical location within grain
111: populations are still debated. Two production routes have been proposed. One
112: postulates that organic refractory mantles (ORM) are formed in the diffuse ISM
113: by UV processing of ice mantles originating in denser clouds
114: \nocite{greenberg89,greenberg_li96,li_greenberg97} (Greenberg 1989; Greenberg
115: \& Li 1996; Li \& Greenberg 1997). Dust grains are, indeed, expected to cycle
116: between dense and diffuse phases of the ISM before being consumed by star
117: formation \nocite{mckee89} (McKee 1989). Spectroscopically, carbonaceous
118: materials produced in this manner provide a reasonable match to the 3.4
119: \micron\ absorption feature, though they do not reproduce the observed
120: deformation modes at longer wavelengths (see Pendleton \& Allamandola 2002).
121: This model predicts that the carrier of the 3.4~\micron\ feature resides in a
122: mantle on a silicate core. In the alternative scenario, production occurs in
123: situ in the diffuse ISM via hydrogen-bombardment of pre-existing carbon
124: particles \nocite{mennella_etal02} (e.g.\ Mennella et al.\ 2002). This model
125: predicts hydrocarbons and silicates to be separate grain populations, at least
126: in low-density environments where grain-grain coagulation is less important.
127: This difference in physical location predicted for the absorber provides a
128: means of discrimination.
129:
130: Much can be learned about the physical properties of dust grains through
131: polarimetric measurements of their solid-state absorption features. Excess
132: polarization across a particular feature is a sign that the grains in which the
133: absorber resides are aligned by the magnetic field. Furthermore, lack of excess
134: polarization across a particular feature implies poorly aligned grains or
135: grains with optical properties, including shape, that induce little or no
136: polarization. Note that the alignment efficiency is a strong function of
137: particle size (e.g.\ Draine 2003), with very small ($<0.01~\micron$) grains
138: being generally much less well aligned compared with larger ``classical''
139: grains. The aligned grain population includes silicates, which exhibit excess
140: polarization across the 9.7 \micron\ and 18.5 \micron\ absorption features in
141: lines of sight that probe both the diffuse ISM and stars embedded in their
142: nascent molecular clouds
143: \nocite{smith_etal00,aitken_etal88_spectropol_gl,aitken_etal89,wright_etal02}
144: (e.g.\ Aitken \etal\ 1988; Aitken \etal\ 1989; Smith \etal\ 2000; Wright \etal\
145: 2002). In dense molecular clouds where conditions are sufficiently cold for
146: ice mantles to form, polarization excesses across the 3 \micron\ H$_2$O-ice
147: feature \nocite{hough_etal88,kobayashi_etal80,holloway_etal02} (e.g.\ Kobayashi
148: \etal\ 1980; Hough \etal\ 1988; Holloway \etal\ 2002) and the solid-CO feature
149: \nocite{chrysostomou_etal96} (Chrysostomou \etal\ 1996) are also detected,
150: giving firm support for the existence of aligned grains that have silicate
151: cores surrounded by ice mantles. It is reasonable to suppose that such grains
152: will continue to produce polarization as they cycle into the diffuse ISM, where
153: conditions are generally more conducive to efficient alignment
154: \nocite{lazarian_goodman_myers97} (Lazarian,
155: Goodman, \& Myers 1997), irrespective of modification to the mantles.
156:
157: Since the silicate feature toward the GC is known to be polarized, a good test
158: of the ORM theory is measurement of polarization across the 3.4 \micron\
159: feature. If the hydrocarbons reside in mantles on silicate cores, then clearly
160: the feature should be polarized at a similar level. Nagata \etal\ (1994) found
161: no significant peak around 3.4 \micron\ in their low-resolution
162: spectropolarimetric data of several Galactic center sources (including GC-IRS7
163: and GCS 3-II). Higher resolution spectropolarimetric observations of GC-IRS7
164: were carried out by \nocite{adamson_etal99} Adamson et al.\ (1999), who showed
165: that the 3.4 \micron\ feature in GC-IRS7 is not polarized. This challenged the
166: ORM theory. However, interpretation of this result was complicated by the fact
167: that the silicate polarization toward GC-IRS7 was inferred from a measurement
168: toward the nearby line of sight GC-IRS3. Although GC-IRS3 is located only
169: $\sim3\arcsec$ away from GC-IRS7, since the polarization toward GC-IRS7 was
170: not measured directly, the possibility remains that mantled silicate grains
171: along this line of sight either fail to align or produce no net polarization
172: for some other reason (Li \& Greenberg 2002). To remove this ambiguity, we
173: have undertaken spectropolarimetry of the 3.4~\micron\ hydrocarbon feature
174: toward two lines of sight in the Quintuplet cluster for which silicate
175: spectropolarimetry already exists \nocite{smith_etal00} (Smith et al.\ 2000).
176:
177: In \S2, we describe the observations and the data reduction methods used to
178: produce the polarization spectra. In \S3, we discuss the available
179: spectroscopic and polarimetric observations of the Quintuplet sources and in
180: \S4 we present the new spectropolarimetric data for the 3.4 \micron\ feature.
181: Finally, in \S5 we discuss the implications for grain models of the
182: non-detection of 3.4 \micron\ polarization.
183:
184:
185: \section{Observations and Data Reduction}
186: Using the United Kingdom Infrared Telescope's (UKIRT's) spectrometer CGS4 and
187: polarimeter, IRPOL2, we have obtained spectropolarimetry from 3.2 to 3.8
188: \micron\ of the lines of sight toward GCS 3-II and GCS 3-IV. The observations
189: were carried out on the night of 11 August 2003 (UT). IRPOL2 combines a
190: Wollaston prism and rotating half-wave plate. Each Stokes Q and U spectrum
191: resulted from object and sky exposures at each of four position angles of the
192: half-wave plate.
193:
194: Two major issues with these data should be borne in mind. Firstly, the CGS4
195: spectropolarimetry arrangement is not optimal --- the dispersion due to the
196: Wollaston prism precedes the slit, and so at least two precise alignments are
197: necessary to avoid considerable instrumental polarization: (i) alignment
198: between the slit position angle and Wollaston dispersion direction, (ii)
199: alignment between the telescope nod and slit position angle. The latter is
200: straightforward and catered for very precisely by the telescope and instrument
201: control system. The former requires on-sky measurements in stable conditions to
202: ensure the slit position angle calibration is adjusted appropriately. In the
203: case of these observations, conditions were not stable and the target was at a
204: large negative declination, so the alignment was known to be less than
205: perfect. In the resulting data there is indeed detectable excess instrumental
206: polarization compared to that normally expected. There is no evidence that it
207: contains spectral features, consistent with its origin in throughput losses as
208: a function of dispersed beam position on the slit. The second issue is the
209: presence of ripples in the polarization spectrum, in this case having at least
210: two ``frequencies''. Removal of this ripple was achieved as described by
211: \nocite{adamson_etal99} Adamson \etal\ (1999), and this appears to have been
212: successful.
213:
214: Because the polarization in any wavelength bin is a vector quantity, all
215: binning processes were carried out on the Q and U spectra and the binned
216: results combined in the usual way to reproduce a binned polarization spectrum.
217: The final binned polarization spectra are shown in Fig.~\ref{fig:gcs3spec}.
218: The gaps in the observed spectra at 3.29--3.34 and 3.37--3.41 \micron\ occur
219: where data have been removed because of severe absorption in the
220: hydrocarbon-containing optical cement in the IRPOL2 Wollaston prism.
221:
222:
223:
224: \section{The Line of Sight toward the Quintuplet Members GCS 3-II, IV}
225:
226: The Quintuplet cluster lies about $14\arcmin$ northeast of the Galactic Center
227: \nocite{nagata_etal90,okuda_etal90} (Nagata et al.\ 1990; Okuda et al.\ 1990).
228: The visual extinction along the line of sight is estimated to be $\sim 29$~mag
229: \nocite{figer_mclean_morris99} (Figer, McLean, \& Morris 1999). The line of
230: sight toward the Quintuplet sources includes both diffuse ISM and dense cloud
231: material \nocite{schutte_etal98,okuda_etal90,chiar_etal00} (Schutte \etal\
232: 1998; Okuda \etal\ 1990; Chiar et al. 2000). However, the majority of the
233: absorption arises in the diffuse ISM as evidenced by the observation of weak
234: 3.0 \micron\ H$_2$O-ice features toward GCS 3-I and GCS 3-IV (Chiar \etal\
235: 2000)\footnote{ Based on observations with the Infrared Space Observatory's
236: (ISO's) Short Wavelength Spectrometer (SWS). The $14\arcsec\times20\arcsec$
237: ISO-SWS beam was centered on GCS~3-I and included all four GCS~3 objects.}.
238: Assuming a Taurus-like correlation between $A_V$ and $\tau_{3.0}$ (see Whittet
239: \etal\ 2001), \nocite{whittet_etal01taurusext} we estimate that approximately 6
240: magnitudes of $A_V$ can be attributed to dense clouds.
241:
242: Near-infrared observations of continuum polarization toward the Quintuplet
243: sources show vectors that lie along the Galactic plane, indicating that it is
244: interstellar in origin and induced by magnetically aligned dust grains
245: \nocite{kobayashi_etal83,okuda_etal90} (Kobayashi \etal\ 1983; Okuda \etal\
246: 1990). Silicate feature absorption and corresponding polarization at 9.7
247: \micron\ is reported toward GCS 3-II and GCS 3-IV by \nocite{smith_etal00}
248: Smith \etal\ (2000). The $p(\lambda)$ profiles are consistent with pure
249: absorptive polarization (Smith et al.\ 2000). The position angle of the
250: silicate polarization observed toward GCS 3-II and GCS 3-IV is essentially
251: constant across the feature and similar to the interstellar value, and there is
252: no underlying emission component to complicate the analysis. Therefore, it is
253: unlikely that the silicate feature has a circumstellar component: it is safe to
254: assume that it is carried by general-ISM dust along the line of sight. The
255: 9.7~\micron\ features have peak optical depths $\tau=2.9$ and 3.2 in GCS 3-II
256: and GCS 3-IV, respectively. The corresponding peak polarizations ($p$) are
257: $9\pm 0.3$ (\%) and $10.2\pm 1.5$ (\%). Based on the observed polarization
258: spectra, Smith \etal\ (2000) deduce the polarization excess ($\Delta p$) above
259: the continuum to be 8 (\%) and 8.2 (\%) for GCS 3-II and GCS 3-IV,
260: respectively. A measure of the polarization efficiency of the feature is then
261: given by the ratio $\Delta p/\tau$, which is comparable (2.8 for GCS 3-II and
262: 2.6 for GCS 3-IV) between the two sources.
263:
264: The 3.4 \micron\ hydrocarbon absorption feature has been observed toward
265: several Quintuplet Cluster sources. Chiar \etal\ (2000) deduce $\tau_{3.4}=0.16$
266: for GCS 3-I based on spectroscopy with the Short Wavelength Spectrometer of the Infrared Space Observatory.
267: Chiar, Adamson \& Pendleton
268: (in prep.), using higher spatial resolution UKIRT-CSG4 data, find that the
269: depth of the 3.4 \micron\ feature is spatially invariant. They measure the
270: optical depth to be $\tau_{3.4}=0.17$ toward five lines of sight toward the
271: Quintuplet (including GCS 3-II and GCS 3-IV). Thus, like the silicates, the
272: hydrocarbons reside in the diffuse ISM along the line of sight and are unlikely
273: to be local to the sources themselves.
274:
275: \section{Spectropolarimetry of the 3.4 \micron\ Hydrocarbon Feature}
276:
277: Based on the observed optical depth and polarization of the silicate feature,
278: it is straightforward to predict that if the 3.4 \micron\ feature is the result
279: of ORM attached to silicate cores, the 3.4 \micron\ feature should be similarly
280: polarized with $\Delta p \approx 0.47$\% (GCS 3-II) and 0.44\% (GCS 3-IV).
281: Figure 1 shows that the filtered, reconstituted polarization spectra (points
282: with error bars) of both GCS 3-II and GCS 3-IV are to be essentially
283: featureless. To estimate the maximum allowable polarization signal in the 3.4
284: \micron\ feature, a plausible continuum polarization was traced that fits the
285: observed $p(\lambda)$ spectrum at 3.2 \micron\ and 3.7--3.8 \micron\ (solid
286: line). We created a model polarization spectrum using the average hydrocarbon
287: spectrum calculated by Chiar \etal\ (2002; their Fig.~6) and assuming
288: $\Delta$p$/\tau$ for the hydrocarbon feature to be equal to that of the
289: silicate feature. We then carried out a least squares fit in order to minimize
290: the residuals between the binned polarization data and model spectrum. The free
291: parameter was $\Delta$p, the polarization above continuum at the peak of the
292: modeled band. The maximum allowable polarization above the continuum is shown
293: by the dashed line in Fig.~1. The best fits are 0.06$\pm$0.13\% for GCS3-II
294: and 0.15$\pm$0.31\% for GCS3-IV respectively, consistent with zero in both
295: cases and well below the predicted level (shown by the dotted line in Fig.~1).
296: The uncertainties quoted are formal 99\% confidence limits, but ignore possible
297: systematic errors in continuum placement. The largest uncertainty is probably
298: in the continuum placement at the short-wavelength end of the spectrum; but
299: plausible placements of the continuum level result in differences no greater
300: than 0.02\% in the fitted feature strength and do not therefore substantially
301: alter the conclusions.
302:
303: \section{Implications for the Nature of the Hydrocarbon Dust}
304:
305: We have extended the search for polarization associated with the 3.4 \micron\
306: absorption feature (Adamson \etal\ 1999) to two additional lines of sight, with
307: consistent results: the feature is essentially unpolarized. The Adamson \etal\
308: study made a plausible comparison between their observed limit toward GC IRS7
309: and the predicted polarization excess for absorbers located on aligned
310: silicate grains, but the conclusions were subject to an assumed lack of spatial
311: variation in the silicate feature, as described in \S1. This ambiguity has now
312: been removed. Based on models of the polarization properties of silicate and
313: carbonaceous analog material, it has been suggested \nocite{li_greenberg02} (Li
314: \& Greenberg 2002) that polarization across the 3.4~\micron\ feature could be
315: weaker per unit optical depth compared to the silicate feature. Li \&
316: Greenberg (2002) consider mantle/core volume ratios (mantle thicknesses) of
317: $V_{\rm carb}/V_{\rm sil} = 0.25$, 1, and 2 for spheroids over a range of
318: elongations. For the most extreme case of a thick mantle ($V_{\rm carb}/V_{\rm
319: sil} = 2$), the polarization of the 3.4 \micron\ feature could be up to 25\%
320: weaker per unit optical depth compared to the silicate feature without
321: violating the ORM model. Our data show that the observed polarization across
322: the 3.4 \micron\ feature is well below even this prediction.
323:
324: We conclude that the agent responsible for the hydrocarbon feature in the
325: diffuse ISM is located in a grain population that is both physically separate
326: from the silicates and far less efficient as a producer of polarization. A good
327: candidate is a population of very small carbonaceous grains that are optically
328: isotropic and/or unresponsive to the alignment mechanism. A likely source of
329: such grains is indicated by the observations of \nocite{chiar_etal98crl} Chiar
330: \etal\ (1998), who find a 3.4~\micron\ feature essentially identical to the
331: interstellar feature in the outflow of a C-rich evolved star. In contrast,
332: molecular clouds exposed to significant UV flux, presumed to be the formation
333: sites of ORM, show a conspicuous absence of the 3.4~\micron\ feature
334: \nocite{shenoy_etal03} (Shenoy \etal\ 2003). Amorphous carbon grains entering
335: the ISM from the winds of evolved stars will readily become hydrogenated. An
336: excellent spectroscopic match is obtained to the observed deformation modes, as
337: well as to the C--H stretch, with laboratory analogs produced by hydrogenation
338: of nano-sized carbon grains \nocite{chiar_etal00,mennella_etal02} (Chiar \etal\
339: 2000; Mennella \etal\ 2002). Hydrogenated carbon grains are also good
340: candidates for the 217.5~nm extinction bump, a ubiquitous feature of the
341: extinction curve in the diffuse ISM \nocite{schnaiter_etal99} (Schnaiter \etal\
342: 1999), which is also observed to be devoid of polarization in all but two (out
343: of 30) lines of sight studied
344: \nocite{clayton_etal92,somerville_etal94,clayton_etal95uvpol,wolff_etal97}
345: (Clayton \etal\ 1992, 1995; Somerville \etal\ 1994; Wolff \etal\ 1997).
346:
347: %%%%%%%%%% Something about small grains a la Schnaiter, Furton, etc, Fix up
348: %%%%%%%%%% this last paragraph
349:
350:
351:
352: \acknowledgments
353: The United Kingdom Infrared Telescope (UKIRT) is operated by the Joint
354: Astronomy Centre on behalf of the U.K.\ Particle Physics and Astronomy Research
355: Council. We thank the Department of Physical Sciences, University of
356: Hertfordshire for providing IRPOL2 for the UKIRT. J.E.C. was supported by
357: NASA's Long Term Space Astrophysics program under grant 399-20-61-02.
358: D.C.B.W.\ acknowledges financial support from NASA grant NAG5-13469.
359:
360:
361: \bibliographystyle{apj}
362: \bibliography{references}
363:
364:
365: \begin{figure}
366: \plotone{f1.eps}
367: \caption{Polarization spectra of GCS 3-II [top] and GCS 3-IV [bottom].
368: Observed polarization is shown by the filled circles with error bars. The
369: estimated continuum polarization is shown by the solid line. The maximum
370: allowed polarization as determined from a $\chi^2$ fit, in the 3.4 \micron\
371: feature is shown by the dashed line. The maximum polarization predicted for
372: grains with organic mantles, assuming that $\Delta$p/$\tau$ is the same
373: as for the silicates, is shown by the dotted line.\label{fig:gcs3spec}}
374: \end{figure}
375:
376:
377: \end{document}
378: