1: %\documentclass[manuscript]{aastex}
2: \documentclass[12pt,preprint]{aastex}
3: %\documentclass[preprint2]{aastex}
4:
5: %\usepackage{natbib}
6: %\usepackage{txfonts}
7:
8: \shorttitle{Dusty Sources at the GC}
9: \shortauthors{Viehmann et al.}
10:
11: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
12:
13: \begin{document}
14:
15: \title{Dusty Sources at the Galactic Center:\\
16: The $N$- and $Q$-band view with VISIR}
17:
18: \author{
19: T.~Viehmann\altaffilmark{1} \and A.~Eckart\altaffilmark{1} \and
20: R.~Sch\"odel\altaffilmark{1} \and J.-U.~Pott\altaffilmark{1,2} \and
21: J.~Moultaka\altaffilmark{1}
22: }
23: \altaffiltext{1}{I.~Physikalisches Institut, Universit\"at zu K\"oln,
24: Z\"ulpicher~Str.~77, D-50937 K\"oln, Germany \email{viehmann@ph1.uni-koeln.de}}
25: \altaffiltext{2}{ESO, Karl-Schwarzschild-Str.~2, D-85748 Garching, Germany}
26:
27: \begin{abstract}
28: We present mid-infrared $N$- and $Q$-band photometry of the Galactic Center from
29: images obtained with the mid-infrared camera VISIR at the ESO VLT in May 2004.
30: The high resolution and sensitivity possible with VISIR enables us to
31: investigate a total of over 60 point-like sources, an unprecedented number for
32: the Galactic Center at these wavelengths.
33: Combining these data with previous results at shorter wavelengths
34: \citep{viehmann2005} enables us to construct SEDs covering the $H$- to $Q$-band
35: regions of the spectrum, i.e. 1.6 to 19.5~$\mu$m.
36: We find that the SEDs of certain types of Galactic Center sources show
37: characteristic features.
38: We can clearly distinguish between luminous Northern Arm bow-shock sources,
39: lower luminosity bow-shock sources, hot stars, and cool stars.
40: This characterization may help clarify the status of presently unclassified
41: sources.
42: \end{abstract}
43:
44: \keywords{Galactic Center --- Mid-infrared --- Stellar classification}
45:
46: %\maketitle
47:
48: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
49:
50: \section{Introduction}
51: \label{INTRO}
52:
53: The Galactic Center is known as a bright source of near- and mid-infrared
54: radiation since the late 1960s \citep{becklin1968, becklin1969, low1969}.
55: The main source of near-infrared radiation is photospheric emission from a
56: dense stellar cluster, i.e.\ a crowded field of point sources,
57: while almost all of the mid-infrared radiation originates from extended gas and
58: dust features as well as dust emission from the circumstellar regions of a dozen
59: individual sources interacting with the more extended Galactic Center
60: interstellar medium.
61:
62: The nature of the brightest mid-infrared sources is more or less clear --
63: IRS~1W, 2, 5, 10W, and 21 are bow-shock sources, caused by bright stars with
64: strong winds plowing through the ambient gas and dust of the Northern Arm
65: \citep{tanner2002, tanner2003, tanner2005, rigaut2003, eckart2004, geballe2004}.
66: Of the fainter sources, IRS~7 is an M2 supergiant, however, most of the others
67: remain enigmatic.
68:
69: Combined photometry at near- and mid-infrared wavelengths is an ideal tool
70: to investigate the nature of the bright mid-infrared sources, since it gives
71: information on both the mid-infrared sources themselves and the stellar sources
72: that could be powering them.
73: We therefore present $N$- and $Q$-band photometry of the Galactic Center,
74: obtained with the ESO VLT, in order to further investigate the nature of
75: the mid-infrared sources.
76: Additional data at shorter wavelengths is taken from \citet{viehmann2005} to
77: provide a more complete picture.
78:
79: Previous studies at mid-infrared wavelengths have been severely limited in
80: resolution (\citealt{becklin1978} $\sim$2\arcsec, \citealt{gezari1985}
81: $\ga$1\arcsec, \citealt{stolovy1996} $\sim$0.7\arcsec), however, since the
82: availability of the mid-infrared camera VISIR at the ESO VLT, the Galactic
83: Center can be studied in great detail in the $N$- and $Q$-bands (8-13~$\mu$m
84: and 17-24~$\mu$m), with a resolution of 0.3\arcsec to 0.6\arcsec (the
85: diffraction limit of an 8~m telescope), provided that the seeing is good enough.
86: This corresponds to a linear resolution of 2500 to 5000~AU at the distance of
87: 8~kpc to the Galactic Center.
88:
89:
90: \section{Observations and data reduction}
91: \label{obs}
92:
93: \subsection{VISIR observations}
94: \label{VISOBS}
95: During the commissioning of the VISIR mid-infrared imager/spectrograph at the
96: ESO VLT (UT3, Melipal), the Galactic Center was observed on May 7th -- 11th
97: 2004.
98: Many narrowband filters spanning the whole $N$-band (8-13~$\mu$m), as well as
99: the 18.7~$\mu$m and 19.5~$\mu$m ($\Delta\lambda=0.88$ and 0.40~$\mu$m) filters
100: in the $Q$-band were used.
101: In this paper we concentrate on the 8.6, 11.3, and 12.8~$\mu$m
102: ($\Delta\lambda=0.42$, 0.59, and 0.21~$\mu$m) filters in the $N$-band, since the
103: highest number of images (and thus the longest integration times) were produced
104: with these filters.
105: The images cover a field of view of approximately 32\arcsec$\times$32\arcsec at
106: a pixel scale of 0.127\arcsec per pixel \citetext{see e.g.\ \citealp{lagage2003}
107: for further details on VISIR}.
108: The resulting angular resolution (due to seeing and the diffraction limit)
109: varies between 0.3\arcsec at 8.6~$\mu$m and approximately 0.6\arcsec in the
110: $Q$-band.
111: Each of the images supplied by the VISIR pipeline consists of the average of
112: (typically) seven chopped (chopper frequency 0.25~Hz, amplitude 12.5\arcsec,
113: position angle 0\degr) and flat-fielded exposures, with a typical integration
114: time of 14~s per averaged image.
115: \footnote{It should be noted that the unfortunate choice of position angle for
116: chopping results in a negative artifact in the images, since the bright source
117: IRS~8 lies in the corresponding field of view.}
118:
119: \subsubsection{Additional data}
120: For comparison with shorter wavelengths, $H$-, $K_\mathrm{s}$-, $L$- and
121: $M$-band (1.6, 2.1, 3.8 and 4.7~$\mu$m) data were taken from
122: \citet{viehmann2005}.
123:
124: \subsection{Data reduction}
125: The chopped and flat-fielded exposures supplied by the pipeline were selected
126: based on image quality (FWHM and signal-to-noise ratio) and coadded,
127: resulting in total integration times of 168, 182, 140, 70 and 140 seconds at
128: 8.6, 11.3, 12.8, 18.7 and 19.5~$\mu$m, respectively.
129: The final 19.5~$\mu$m image is shown in Fig.~\ref{Q3full}, while a three-color
130: composite of the $N$-band images is shown in Fig.~\ref{rgb}.
131:
132: \subsubsection{Photometry}
133: \label{calib}
134: Flux densities for the MIR sources were obtained using aperture photometry with
135: an aperture of approximately 1\arcsec.
136: Additional flux densities were obtained with apertures of 2\arcsec and
137: 0.5\arcsec in order to correct for confusion effects for closely separated
138: sources, and to optimise the background subtraction.
139: For the purpose of background subtraction, the sky contribution is fitted to an
140: annulus situated between radii of 1.6\arcsec and 2.1\arcsec (from the center of
141: the aperture) in the case of the 1\arcsec aperture, and between 2.6\arcsec and
142: 3.1\arcsec or between 1\arcsec and 1.5\arcsec for the 2\arcsec and 0.5\arcsec
143: apertures, respectively.
144:
145: Since no suitable calibration stars were observed, the flux density calibration
146: proved difficult.
147: Very few $N$- or $Q$-band magnitudes or flux densities of Galactic Center
148: sources have been published at present.
149: We used IRS~21 as calibration source, since it is the $N$-band source studied in
150: greatest detail \citep[e.g.][]{gezari1985, tanner2002}.
151: We used the flux densities for IRS~21 at 8.8, 12.5, 20.8, and 24.5~$\mu$m
152: given by \citet{tanner2002}, assuming that the emission from IRS~21 follows the
153: general shape of the integrated Galactic Center spectrum as measured by ISO
154: \citep{lutz1996}.
155: For 8.6~$\mu$m, we also took into account the flux density at 8.7~$\mu$m
156: published by \citet{stolovy1996}.
157: Corrections for the different aperture sizes and background subtraction were
158: applied.
159: For extinction correction, we assumed A$_\mathrm{V}$=25 \citep{scoville2003,
160: viehmann2005}, adopting the extinction law of \citet{moneti2001} for the $N$-
161: and $Q$-bands.
162:
163: We estimate the total uncertainty (i.e. statistical and systematic) of the
164: photometry results as $\pm$30\%, rising to $\pm$50\% for the faintest sources
165: ($\lesssim$0.5~Jy).
166: The dominating uncertainty here is that of the flux density calibration, and
167: also the correct background subtraction for the faint sources.
168:
169:
170: \section{Results}
171: \label{results}
172:
173: \subsection{New detections}
174: \label{newdet}
175: We have obtained photometry for a total of 35 sources at 19.5~$\mu$m rising to
176: 64 sources at 8.6~$\mu$m.
177: This is probably the largest number of Galactic Center sources investigated in
178: the $N$- and $Q$-bands up to now and includes the first detection of the
179: Galactic Center Helium-emission-line stars IRS~16NE, IRS~16NW and AF/AHH, as
180: well as the red giants IRS~12N, IRS~14NE and IRS~14SW in the $N$-band, as shown
181: in Fig.~\ref{findthem}. Additionally, it is now possible to separate the WC9
182: star IRS~6E \citep{blum1996} from the bright extended emission associated with
183: IRS~6W. The corresponding data can be found in table~\ref{lotsofnumbers}.
184:
185: \subsection{SEDs}
186: Combining the VISIR data with shorter wavelength data enables us to construct an
187: unprecedented (for the Galactic Center) number of spectral energy distributions
188: (SEDs) across the whole 1.6 to 19.5~$\mu$m range. These are shown in
189: Fig.~\ref{SEDs}.
190:
191: \subsubsection{Classification by SED?}
192: \label{sedtype}
193: Preliminary analysis of a few typical cases indicates that these SEDs can
194: provide interesting information on the source characteristics, since different
195: source types have SEDs of considerably different shape.
196: This could possibly allow for preliminary classification of presently
197: unclassified sources.
198: The different types of SED found in our sample are as follows:
199:
200: \paragraph{I.\ Northern Arm embedded sources\\}
201: These are luminous sources that are embedded in the ``Northern Arm'' stream of
202: gas and dust, and also in the bar area of the minispiral.
203: IRS~1W, 2L, 5, 10W and 21 belong to this group, which shows very red and
204: fairly featureless SEDs, with their maximum in the $N$-band, indicating fairly
205: low temperatures \citep[approximately 300~K according to][]{gezari1985}.
206: These objects are successfully described as bow shock sources
207: \citep{tanner2002, tanner2003, tanner2005, rigaut2003, eckart2004,
208: geballe2004}.
209: IRS~13E, which is situated in the bar area of the minispiral, shows similar
210: characteristics, although it is clearly hotter \citep{gezari1996}, with the
211: embedded source identified as a cluster of massive young stars
212: \citep[e.g.][]{krabbe1995, najarro1997, eckart2004, maillard2004}.
213: \paragraph{II.\ Lower luminosity bow shock sources\\}
214: These sources are less luminous and often situated close to the edge of the
215: minispiral.
216: In almost all cases, they are identified as bow-shock sources by their
217: morphology, the bow-shock structure being visible in high-resolution $L$- or
218: $M$-band images obtained with adaptive optics \citep[see e.g.][for such
219: images]{genzel2003, clenet2004m}.
220: A striking example (source 41 in our sample) is presented by
221: \citet{clenet2004}.
222: These sources show very red SEDs with their maxima longwards of 12~$\mu$m.
223: Their characteristics at shorter wavelengths are diverse: some show a fairly
224: flat SED between 1.6 and 11.3~$\mu$m, while others exhibit a minimum around
225: 8.6~$\mu$m.
226: These differences are probably due to differences between the type of stars
227: powering the bow-shocks (see below), and due to the presence and density of
228: the dusty medium that surrounds them and with which they interact.
229: We propose that IRS~2S, which has been classified as a cool red giant by
230: \citet{blum1996}, belongs to this type of source, since it shows a flat SED
231: from 1.6 to 11.3~$\mu$m and brightens considerably at longer wavelengths.
232: \paragraph{III.\ Cool stars\\}
233: The SEDs of these sources all have their maximum in the 2 to 5~$\mu$m region.
234: The SEDs then fall off to the $N$-band, where they appear to flatten out. In
235: those cases which are still detectable at wavelengths longer than 8.6~$\mu$m
236: (despite decreasing resolution and sensitivity), the $N$- and $Q$-band flux
237: densities of this type of source appear to be more or less constant for each
238: source.
239: This effect is probably due to circumstellar dust emission, which is to be
240: expected, since the brightest cool red giants at the Galactic Center have been
241: identified as AGB stars \citep{ott1999, clenet2001} and are thus situated in a
242: mass-loss phase, creating dust envelopes.
243: A \emph{very} weak bow-shock type interaction with the surrounding medium is
244: also possible in the case of AGB stars, since they have fairly strong stellar
245: winds.
246: \paragraph{IV.\ Hot stars\\}
247: These sources exhibit SEDs that fall off steeply (steeper than those of cool
248: stars) between 5 and 8~$\mu$m.
249: Therefore, the $M-N$ color (more particularly $M-m_{8.6\mu\mathrm{m}}$)
250: appears to be a criterion for distinguishing hot and cool stars, if
251: spectroscopic information is not available.
252: The clearly identified hot stars in our sample fall into two distinctive
253: subgroups:
254: The Galactic Center He-stars IRS~16NE, IRS~16NW and AF/AHH, which are
255: frequently classified as LBV or Ofpe/WN9 \citep[e.g.][]{clenet2001}, are
256: brightest at short wavelengths with SEDs that fall monotonically towards
257: longer wavelengths.
258: The dominant source of infrared emission is thus probably photospheric, with a
259: wind (or faint bow shock) contribution in the $L$- and $M$-band
260: \citep[see][]{viehmann2005}.
261: On the other hand, IRS~29 and IRS~6E, which have been classified as WC9
262: Wolf-Rayet stars \citep[e.g.][]{blum1996}, exhibit SEDs which peak in the $L$-
263: and $M$-bands, while their $H$- and $K$-band flux densities appear to be
264: considerably lower.
265: The corresponding interpretation is that these stars are surrounded by a
266: self-generated dust shell which partly extincts the shorter wavelength
267: emission and itself emits strongly at 3-5~$\mu$m
268: \citep[see also][]{moultaka2004}.
269: \paragraph{V.\ IRS~3\\}
270: This source clearly stands apart from all other Galactic Center mid-infrared
271: sources.
272: While \citet{horrobin2004} classified IRS~3 as a WC5/6 star, \citet{pott2005}
273: and R.~Genzel (private communication) have pointed out that this
274: identification probably applies to a different, fainter source situated
275: approximately 0.12\arcsec to the east of IRS~3.
276: This does not rule out that IRS~3 may be a massive Wolf-Rayet star, but
277: \citet{pott2005} have also shown that it may be an extreme AGB star.
278: The true nature of IRS~3 therefore remains unclear.
279: Among the bright $N$-band sources, it is the most compact \citep[see new
280: interferometric MIDI VLTI results by][]{pott2005}, apart from IRS~7
281: \citep{gezari1985}.
282: It is also the brightest source in the $M$-band and one of the brightest over
283: the whole mid-infrared region covered here ($L$- to $Q$-band), while it
284: appears unremarkable in the $K$-band and very faint at shorter wavelengths.
285: The shape of the SED of IRS~3 is thus unlike any of the typical cases
286: discussed at present:
287: IRS~3 is much fainter than the luminous bow-shock sources in the
288: near-infrared, while clearly hotter \citep[approximately 400 to 600~K,
289: see][]{gezari1985, gezari1996} and thus showing maximum emission at a shorter
290: wavelength ($M$-band vs.\ 10~$\mu$m).
291: Compared to luminous cool stars (IRS~7, IRS~10E*), IRS~3 is clearly fainter at
292: near-infrared wavelengths and much brighter in the $N$- and $Q$-bands.
293: The same applies to the comparison of IRS~3 with the hot stars in our sample.
294: However, a certain similarity to the SED of IRS~29 (WC9 type) is discernible:
295: The source is faint and obviously extincted in the $H$-band, has a maximum at
296: short midinfrared wavelengths ($L$- and $M$-band) and shows a fairly clear
297: absorption feature at approximately 10~$\mu$m (see section \ref{silicate}).
298: The reason for this similarity is that the emission from both objects comes
299: from thick, self-generated dust shells \citep[compare][]{viehmann2005}.
300:
301: \subsubsection{Silicate dust}
302: \label{silicate}
303: The spectrum between 8.6~$\mu$m and 12~$\mu$m is especially interesting, since
304: there is a broad interstellar silicate absorption feature at 9.7~$\mu$m
305: \citep{lutz1996}.
306: Our data at 11.3~$\mu$m lie on the flank of this absorption; consequently,
307: sources that are underluminous at this wavelength appear to show additional, and
308: therefore \emph{intrinsic} silicate absorption, i.e.\ absorption in excess of
309: that found in the overall ISO spectrum \citep{lutz1996}.
310: IRS~3 and IRS~6W clearly show this effect, which is also visible in
311: Fig.~\ref{rgb} (lack of green), as do IRS~34 (both components) and source 24.
312: The most likely explanation is that these stars are surrounded by (possibly
313: self-generated) dust shells containing additional silicate dust.
314: The cool star IRS~10E* and the hot stars IRS~6E and IRS~29, which have been
315: identified as WC9 Wolf-Rayet stars \citep[e.g.][]{blum1996}, might also show
316: signs of such absorption, although the steepening of their SED is slight and may
317: therefore be insignificant.
318: In the case of the WC9 stars, intrinsic silicate absorption is difficult to
319: explain, however, one possibility is that the absorption is due to collected
320: material not originally produced by the stars themselves.
321: Enhanced, patchy foreground extinction over the whole IRS~3 to IRS~6W area is a
322: possible alternative explanation.
323:
324: In contrast, none of the luminous, bow-shock dominated sources shows intrinsic
325: silicate absorption.
326: Therefore they appear yellow in the color image shown in Fig.~\ref{rgb}.
327: It seems that the lower luminosity bow-shock sources also do not show such an
328: absorption effect.
329:
330: \subsubsection{Excess emission at 12.8~$\mu$m}
331: \label{high12}
332: Due to the shape of the Galactic Center spectrum \citep[e.g.][]{lutz1996},
333: slightly higher flux densities are to be expected at 12.8~$\mu$m.
334: This is clearly visible in the SEDs obtained from our observations, however,
335: some sources show a larger excess, which appears to indicate additional
336: intrinsic emission at 12.8~$\mu$m and must be explained.
337:
338: While \citet{willner1978} state that the 12.8~$\mu$m \ion{Ne}{2} line emission
339: at the Galactic Center is extended and not associated with compact mid-infrared
340: sources, \citet{lacy1979, lacy1980} find that part of the \ion{Ne}{2} emission
341: comes from compact regions of ionized gas which are indeed associated with
342: mid-infrared continuum sources. Consequently, \ion{Ne}{2} line emission is a
343: possible explanation for at least some of these excesses.
344:
345: In the case of IRS~10E* (and IRS~9 if the small excess of this source is
346: significant), we attribute the 12.8~$\mu$m excess to a narrow streamer of
347: enhanced emission running approximately southwards from IRS~5, which probably
348: resulted in an insufficient background subtraction (see section~\ref{calib}).
349: The excess 12.8$\mu$m emission of IRS~2, 6W and 13 could also be the result of
350: insufficient background subtraction due to the clumpy nature of the extended
351: emission from this part of the minispiral, which shows quite a complex
352: structure.
353:
354: The excess 12.8~$\mu$m emission clearly visible in the SEDs of IRS~3, 7 and
355: possibly 29 (also visible in Fig.~\ref{rgb}), however, cannot be explained in
356: this fashion.
357: Since \citet{lacy1980} state that IRS~3 and IRS~7 do not have associated gas
358: clouds, it appears unlikely that the excess of these sources is due to
359: \ion{Ne}{2} line emission.
360: We therefore suggest that the excess of these sources at 12.8~$\mu$m is due to
361: continuum emission caused by a faint bowshock-like interaction of the stellar
362: winds of these stars with the surrounding medium.
363: Since this is a small effect, the corresponding excess that is to be expected at
364: 11.3~$\mu$m could be masked by intrinsic silicate absorption (see above).
365: The absence of a clear excess at 18.7~$\mu$m in these sources does not rule out
366: such an explanation, since the known bow-shock sources (types I. and II. in
367: section~\ref{sedtype}) show diverse characteristics in the $Q$-band (the
368: fainter sources have their maxima there, the brighter ones have their maxima in
369: the $N$-band).
370:
371: \subsection{Unusual sources east of IRS~5}
372: \label{irs5x}
373: Among the most intriguing mid-infrared sources at the Galactic Center are four
374: bright pointlike sources located to the east of the bright Northern Arm source
375: IRS~5.
376: The nature of these sources, which we refer to as IRS~5NE, IRS~5E, IRS~5S, and
377: IRS~5SE (see Fig.~\ref{findthem}), is currently unclear.
378: They are remarkable in that they are almost as bright as IRS~7 in the $N$-band,
379: while they appear much less prominent at shorter wavelengths, although they
380: remain bright in the $L$- and $M$-bands.
381: Detailed inspection reveals that IRS~5SE, which appears \emph{slightly} extended
382: at longer wavelengths, is in fact double, consisting of a blue point source to
383: the east (IRS~5SE2, dominant component in $K$-band images) and a fainter source
384: to the west, which shows a tail like structure and is the main source of
385: emission at longer wavelengths (IRS~5SE1, dominant component in $N$- and
386: $Q$-band).
387: This may represent an intriguing case of possible interaction, either of the two
388: sources with each other -- if they lie close together along the line of sight --
389: or of the western component with the surrounding medium. A further possibility
390: is that a bow shock generated by the eastern component is superimposed on the
391: western one, however, their separation makes this unlikely.
392:
393: All of these sources except IRS~5SE2 show SEDs that are either flat between 3.8
394: and 20~$\mu$m (IRS~5NE) or increase dramatically towards longer wavelengths.
395: Therefore, these sources show similar characteristics to identified
396: lower-luminosity bow-shock sources (e.g. IRS~9N, see section~\ref{sedtype}),
397: however, their appearance on high-resolution NAOS/CONICA images does not
398: indicate such a structure, which is normally visible for this type of source
399: (IRS~5SE1 is the obvious exception here).
400: It therefore appears that a reliable classification and/or understanding of
401: these sources can only be achieved with spectroscopic methods or possibly
402: interferometry using VLTI.
403:
404:
405: \section{Conclusion}
406: \label{conc}
407:
408: We have presented mid-infrared photometry of the Galactic Center from images
409: obtained with the VISIR camera at the ESO VLT.
410: In combination with shorter wavelength data \citep{viehmann2005}, we were able
411: to construct SEDs for an unprecedented number of sources (for the Galactic
412: Center) in this wavelength regime.
413: We find that different source types show characteristic differences in the shape
414: of their SEDs, which should help in the preliminary classification of compact
415: mid-infrared sources.
416: In particular, we find that additional, i.e. intrinsic, absorption at
417: approximately 10~$\mu$m is a sign of dust formation.
418: Many of the brightest Galactic Center mid-infrared sources, however, appear to
419: be bow-shock powered and do not show dust formation features in their SEDs.
420: In addition, we find that hot and cool stars show different
421: $M-m_{8.6\mu\mathrm{m}}$ colors, with the cool stars showing redder colors, as
422: is to be expected.
423:
424: We have reported on four fairly bright $N$-band point sources located to the
425: east of the bright Northern Arm source IRS~5.
426: The most likely explanation for their unusual appearance at mid-infrared
427: wavelengths is that these sources are bow-shock sources of lower luminosity than
428: the typical Northern Arm sources.
429:
430: The main uncertainty in our results is due to calibration uncertainties as a
431: result of the lack of suitable reference observations.
432: Therefore, mid-infrared observations of the Galactic Center remain necessary in
433: order to obtain more complete information on the dusty mid-infrared sources in
434: this intriguing environment.
435:
436:
437: \acknowledgements
438:
439: This work was supported in part by the Deutsche Forschungsgemeinschaft
440: (DFG) via grant SFB 494.
441:
442: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
443:
444: \bibliographystyle{apj}
445: \bibliography{ms}
446:
447: \clearpage
448:
449: \input{tab1}
450:
451: \clearpage
452:
453: \begin{figure}
454: % \plottwo{f1.eps}{f1_color.eps}
455: \plotone{f1reduced.eps}
456: \caption{VISIR 19.5~$\mu$m image of the Galactic Center. The field of view is
457: 32\arcsec$\times$32\arcsec. East is to the left, and north is up.}
458: \label{Q3full}
459: \end{figure}
460: % \footnote{The negative artifact slightly above the center of the image is due
461: % to the bright source IRS~8, which lies in the field of view used for chopping.
462: % (See section~\ref{VISOBS}.)}
463:
464: \begin{figure}
465: \plotone{f2reduced.eps}
466: \caption{VISIR $N$-band three-color composite view of the Galactic Center. The
467: blue channel is 8.6~$\mu$m, green is 11.3~$\mu$m and red is 12.8~$\mu$m.}
468: \label{rgb}
469: \end{figure}
470:
471: \begin{figure}
472: % \plottwo{f3.eps}{f3_color.eps}
473: \plotone{f3_color.eps}
474: \caption{VISIR 8.6~$\mu$m view of the Galactic Center. The field of view is
475: approximately 24\arcsec$\times$24\arcsec. The insets show enlarged views of the
476: newly detected stars AF/AHH, IRS~6E, IRS~12N, IRS~14NE, IRS~14SW, IRS~16NE and
477: IRS~16NW (see section~\ref{newdet}), as well as the intriguing group of bright
478: sources located to the east of IRS~5 (upper left, see section~\ref{irs5x}).}
479: \label{findthem}
480: \end{figure}
481:
482: \begin{figure}
483: \plotone{f4.eps}
484: \caption{Spectral energy distributions of Galactic Center mid-infrared
485: sources. The SEDs have been grouped by type, as described in
486: section~\ref{sedtype}: (a) shows the typical luminous bow-shock sources, (b)
487: shows lower luminosity bow-shock sources, (c) cool stars, (d) hot stars, while
488: (e) and (f) show the SEDs of unclassified sources. In (e) and (f), IRS~3 and
489: the Sgr~A* clump are distinguished by a dash-dotted line.}
490: \label{SEDs}
491: \end{figure}
492:
493: \end{document}
494: