1: \documentclass[12pt,preprint]{aastex}

2: \begin{document}

3: 

4: \title{The Mid-Infrared Colors of the ISM and Extended Sources at the Galactic Center}

5: 

6: \author{

7: R.~G.~Arendt\altaffilmark{1,2}, 

8: S.~R.~Stolovy\altaffilmark{3}, 

9: S.~V.~Ram\'{\i}rez\altaffilmark{4}, 

10: K.~Sellgren\altaffilmark{5}, 

11: A.~S.~Cotera\altaffilmark{6}, 

12: C.~J.~Law\altaffilmark{7,8}, 

13: F.~Yusef-Zadeh\altaffilmark{7},

14: H.~A.~Smith\altaffilmark{9},

15: D.~Y.~Gezari\altaffilmark{10}

16: }

17: 

18: %\email {richard.g.arendt@nasa.gov}

19: 

20: \altaffiltext{1}

21: {CRESST/UMBC/GSFC, Code 665, NASA/Goddard Space Flight Center,

22: 8800 Greenbelt Road, Greenbelt, MD 20771; richard.g.arendt@nasa.gov}

23: \altaffiltext{2}{Science Systems and Applications, Inc.}

24: \altaffiltext{3}

25: {{\it Spitzer} Science Center, California Institute of Technology, 

26: Mail Code 314-6, 1200 East California Boulevard, Pasadena, CA 91125}

27: \altaffiltext{4}

28: {IPAC, California Institute of Technology,

29: Mail Code 100-22, 1200 East California Boulevard, Pasadena, CA 91125}

30: \altaffiltext{5}

31: {Department of Astronomy, Ohio State University, 140 West 18th Av.,

32: Columbus, OH, 43210, USA}

33: \altaffiltext{6}

34: {SETI Institute, 515 North Whisman Road, Mountain View, CA 94043}

35: \altaffiltext{7}

36: {Department of Physics and Astronomy, Northwestern University,

37: Evanston, IL 60208}

38: \altaffiltext{8}

39: {Astronomical Institute ``Anton Pannekoek'', University of Amsterdam, 1098 SJ Amsterdam, Netherlands}

40: \altaffiltext{9}

41: {Harvard-Smithsonian Center for Astrophysics, 

42: 60 Garden Street, Cambridge, MA 02138}

43: \altaffiltext{10}

44: {NASA/Goddard Space Flight Center, Code 667, 8800 Greenbelt Road, Greenbelt, MD 20771}

45: 

46: 

47: \begin{abstract}

48: A mid-infrared (3.6 -- 8 $\micron$) survey of the Galactic Center has been carried out 

49: with the IRAC instrument on the {\it Spitzer Space Telescope}. This survey

50: covers the central $2\arcdeg \times1\fdg4$ ($\sim 280\times 200$ pc) of the Galaxy. 

51: At 3.6 and 4.5 $\micron$ the emission is dominated by stellar sources, the fainter 

52: ones merging into an unresolved background. At 5.8 and 8 $\micron$ the stellar 

53: sources are fainter, and large--scale diffuse emission from the ISM of the 

54: Galaxy's central molecular zone becomes prominent. 

55: The survey reveals that 

56: the 8 to 5.8 $\micron$ color of the ISM emission is highly uniform across

57: the surveyed region. This uniform color is 

58: consistent with a flat extinction law and 

59: emission from polycyclic aromatic hydrocarbons (PAHs). 

60: Models indicate that this broadband color should not be expected 

61: to change if the incident radiation field heating the dust and PAHs 

62: is $< 10^4$ times that of the solar neighborhood. The few regions with unusually 

63: red emission are areas where the PAHs are underabundant and 

64: the radiation field is locally strong enough 

65: to heat large dust grains to produce significant 8 $\micron$ emission.

66: These red regions include compact H II regions, Sgr B1, and wider 

67: regions around the Arches and Quintuplet Clusters. In these regions the radiation field 

68: is $\gtrsim 10^4$ times that of the solar neighborhood.

69: Other regions of very red emission indicate cases where thick dust clouds 

70: obscure deeply embedded objects or very early stages of star formation. 

71: \end{abstract}

72: 

73: \keywords{dust, extinction --- Galaxy: center --- infrared: ISM}

74: 

75: 

76: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

77: \section{Introduction}

78: 

79: The Galactic Center (GC) is naturally a region of great interest. 

80: The gravitational center of the Galaxy is marked by the supermassive

81: black hole at the radio source Sgr A*. 

82: Remarkably, two young, dense  stellar disks have been discovered 

83: within 0.5 pc orbiting Sgr A* (Paumard et al. 2006).

84: Further out is a molecular ring or the ``circumnuclear disk'' enclosing 

85: an ionized a mini--spiral structure (e.g. Christopher et al. 2005). 

86: These features are within a radius of $\sim7$ pc. 

87: A concentration of molecular material is observed out to a radius of $\sim200$

88: pc along the Galactic Plane; this region is known as the Central Molecular

89: Zone (CMZ). Aside from the stellar bulge of the Galaxy, the CMZ encompasses 

90: nearly all of the Galactic structures and objects that are 

91: characteristic of, or unique to, the GC.

92: 

93: The CMZ is estimated to contain about 10\% of the Galaxy's molecular gas 

94: content (e.g. Morris \& Serabyn 1996; Mezger 1978), or about 

95: $5-10\times10^7$ M$_{\sun}$ (e.g. Armstrong \& Barrett 1985; Stark et al. 1989). 

96: It has a two-component structure. The inner region 

97: consists of a set of low velocity clouds lying on or close to 

98: the Galactic plane, including the giant Sagittarius molecular 

99: clouds, which together comprise the ``disk population'' (Morris 

100: \& Serabyn 1996). The outer set of clouds form a comparatively 

101: high velocity (130--200 km s$^{-1}$) quasi--continuous ring structure 

102: that surrounds the first one at a radius of $\sim180$ pc (the 

103: so-called ``180-pc molecular ring''; Morris \& Serabyn 1996). 

104: Analysis of NH$_3$ and H$_3^+$ absorption lines in the CMZ has 

105: indicated that the molecular gas contains a widespread, relatively hot component

106: ($T \gtrsim 175K$, Wilson et al. 1982; Oka et al. 2005), 

107: which is not found apart from dense molecular cores elsewhere in the Galaxy.

108: Analysis by H\"uttemeister et al. (1993) indicates

109: the molecular gas has at least two phases: hot phase 

110: ($T \sim 200$K, $n_{H_2} \lesssim 10^3$ cm$^{-3}$) with a filling factor

111: of $\sim 25\%$, and a cooler component 

112: ($T \sim 25$K, $n_{H_2} \gtrsim 10^4$ cm$^{-3}$) which fills the remaining

113: 75\% of the volume of the molecular clouds.

114: 

115: The average dust temperature in the CMZ appears to be much lower than 

116: that of the gas, $T_{dust} \sim 25-30$ K (Odenwald \& Fazio 

117: 1984; Cox \& Laureijs 1989).

118: The much higher gas temperatures 

119: therefore indicate the presence of some efficient gas 

120: heating mechanisms, and supersonic winds and turbulence 

121: had been suspected; the large gas line widths, 15--50 km s$^{-1}$, 

122: are consistent with such ideas, although the details of 

123: the mechanisms are debated (SNRs, outflows, viscous 

124: heating effects, etc.).

125: Observations of widespread SiO emission indicates the recent or current 

126: presence of strong shocks in the CMZ (Martin-Pintado et al. 1997; 

127: H\"uttemeister et al. 1998).

128: The discrepancy between dust and gas temperatures has recently

129: been accounted for in terms of the impact of cosmic ray electrons 

130: heating molecular clouds directly and raising their 

131: temperatures (Yusef-Zadeh et al. 2007b).

132: 

133: In addition to the molecular gas, the CMZ is also infused with 

134: neutral and ionized components as revealed by enhanced infrared continuum and 

135: line emission (e.g.  Rodr\'iguez-Fern\'andez et al. 2004, 2005),

136: and by enhanced radio continuum emission (e.g. Altenhoff et al. 1979).

137: 

138: The launch of the {\it Spitzer Space Telescope} (Werner et al. 2004)

139: provides a new opportunity to investigate the GC at mid-- to far--infrared 

140: wavelengths with unprecedented sensitivity and spatial resolution.

141: {\it Spitzer's} Infrared Array Camera, IRAC, (Fazio et al. 2004) 

142: provides $5'$ field of view imaging capabilities with $<2''$ spatial resolution

143: at 4 broadband mid--infrared channels nominally at 3.6, 4.5, 5.8, and 8 $\micron$.

144: We used IRAC to carry out a $2\arcdeg \times1\fdg4$ 

145: ($\sim 280\times 200$ pc) survey of the GC region, which includes the 

146: entire CMZ. 

147: An initial summary of the results of this survey 

148: has been presented by Stolovy et al. (2006). A more extensive overview will be 

149: provided by S. Stolovy et al. (2008, in preparation). 

150: The measurement, cataloging, and characterization of point 

151: sources observed in our IRAC GC survey are described by Ram\'irez et al. (2008). 

152: In IRAC's 3.6 and 4.5 $\micron$ channels, the GC is dominated by stellar 

153: emission, both for resolved sources and in the unresolved background. 

154: At 5.8 and 8 $\micron$ wavelengths, most point

155: sources become significantly fainter, and a strong, structured background 

156: produced by the diffuse ISM is dominant. 

157: In this study, we focus on the diffuse ISM as revealed 

158: by IRAC. In particular we investigate what the broadband colors 

159: indicate about the nature of the dust producing the emission,

160: and the environment in which that dust is found.

161: 

162: Section 2 of this paper contains a brief summary of the IRAC observations.

163: Section 3 presents the colors observed for the diffuse emission

164: and for a few, more compact sources of special interest. Mostly 

165: this focuses on the 5.8 to 8 $\micron$ color of the ISM emission. 

166: The shorter wavelength colors are dominated by starlight and extinction

167: (reddening) effects. Section 4 discusses the implications of the 

168: uniformity of the color on large scales and the few exceptional 

169: regions on small scales. Though limited to the single 

170: 5.8 to 8 $\micron$ color, our analysis yields several general robust 

171: conclusions concerning the relative abundance of PAHs, 

172: the strength of the interstellar radiation field, and the 

173: shape of the extinction law in the GC. The last section (\S5)

174: contains a summary of this paper.

175: 

176: 

177: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

178: \section{Observations}

179: We observed the central region of the Galaxy using 12 Astronomical 

180: Observation Requests (AORs) which mapped the region in 

181: $\sim0\fdg25$ wide strips at roughly constant declination (a restriction due to {\it Spitzer's}

182: pointing constraints and the low ecliptic latitude of the field). Most strips were $\sim1\fdg75$ long,

183: except that the one containing Sgr~A was split into two parts (to mitigate the potential effects of 

184: latent images, with the brightest regions of the survey scheduled last). 

185: Additionally, those strips at the highest and lowest declinations were 

186: shorter, in order to limit the area of complete coverage to $|l| < 1\fdg0$ and $|b| < 0\fdg7$. 

187: The total area covered in all four channels was $\sim3.5$ square degrees, 

188: including the additional irregular edges of the region.

189: The medium scale 5--point Gaussian dither pattern was used. This pattern makes small offsets 

190: to the telescope pointing with an average offset of $\sim30''$. We did not use

191: the smaller scale pattern because it would have provided insufficient 

192: dithering for mitigating the stray light 

193: artifacts of bright stars, and would have hampered self--calibration of the 

194: data (see the Appendix). 

195: Frame times of 2 seconds were used throughout the survey. No regions of extended

196: emission were saturated except at Sgr A at 8 $\micron$, but more than $10^4$ bright point sources 

197: were saturated in at least one channel (Ram\'irez et al. 2008). 

198: Most saturated sources are isolated point sources, but 

199: high densities of saturated sources are found within the bright cluster at Sgr A and the Quintuplet Cluster.

200: These areas, and a sample of 12 other saturated sources, were re-observed in subarray mode with  

201: 0.02--second frame times to obtain unsaturated data. All analysis in this work was performed on 

202: mosaicked images that incorporated the subarray data for these sources.

203: 

204: We constructed our own mosaics from the total set of Basic Calibrated Data (BCD, the standard 

205: {\it Spitzer} data products intended for detailed scientific analysis) from all AORs.

206: The primary description of our processing is given by 

207: Stolovy et al. (2006, 2008 in preparation). In the appendix of the 

208: present paper, we provide additional detail

209: concerning aspects of the processing that are particularly relevant in 

210: studying the brightness and color of extended emission. 

211: 

212: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

213: \section{Analysis}

214: 

215: 

216: \subsection{The Diffuse ISM}

217: \subsubsection{The Large-Scale Distribution of 8 $\micron$ Intensity and 8/5.8 $\micron$ Color}

218: 

219: The 8 $\micron$ mosaic of the Galactic Center region is shown in Figure \ref{fig_ch4}. 

220: The general structure of the ISM emission in this region appears similar to other regions 

221: in the Galactic plane as revealed by the GLIMPSE survey (Benjamin et al. 2003).

222: Both in the GC and other regions, much of the ISM emission is resolved into

223: elongated arcs and filamentary structures (S. Stolovy et al. 2008, in preparation). 

224: Embedded among these are very bright compact regions

225: that contain specific local heating sources. Superimposed infrared dark clouds (IRDCs) have a 

226: similar structure to the emission, but are not well-correlated with the emission.

227: The ISM emission has a strong unresolved background component, which is often 

228: suppressed in depictions meant to highlight the structure of the emission 

229: (such as Fig. \ref{fig_ch4}).

230: The most fundamental difference between emission in the GC and that in other regions

231: of the Galactic plane, is that the GC emission is several times brighter. 

232: Figure \ref{fig_msx} shows a wider view of the Galactic 

233: Center region as seen by SPIRIT III on the MSX mission (Price et al. 2001),

234: revealing that the emission within our survey region is brighter than that 

235: typically found in the inner galaxy (at $|l| > 1\fdg5$). 

236: Figure \ref{fig_msxslice} provides a 

237: more quantitative illustration of the relative brightness of our survey region. 

238: The intensity enhancement is loosely correlated with 

239: the large--scale distribution of molecular emission in the galactic

240: center region (e.g. Martin et al. 2004). 

241: This indicates that much of the low--latitude ($|b| < 0\fdg2$)

242: structure in the galactic center mosaic really is in the central 

243: $\sim150$ pc radius of the Galaxy, but that the emission does not directly arise 

244: from the molecular component of the CMZ.   

245: 

246: In contrast, emission at higher 

247: latitudes ($|b| > 0\fdg2$) often appears to be more local features

248: superimposed along the line of sight. One indication of this is association 

249: with H$\alpha$ emission, which would be too heavily extincted to be 

250: visible at the distance to the GC, with the possible exception of 

251: objects seen at the extreme latitude limits of our survey 

252: (Stolovy et al. 2006, 2008 in preparation). Another is the association with 

253: very dark clouds, which must lie relatively close to the sun. Otherwise,

254: the general foreground emission along the line of sight would reduce the 

255: contrast of these features. The large angular scale and 

256: presence of fine detail in the higher latitude features also 

257: suggest they are relatively nearby clouds along the lines of sight.

258: Pauls \& Mezger (1975) use the velocity and line widths of 6 cm radio recombination 

259: lines to distinguish emission from the GC (large velocities and/or broad lines), 

260: from more local emission along the line of sight (velocities near zero, and narrow

261: line widths). All of the regions they examined at $|b| > 0\fdg16$ are

262: characterized as being along the line of sight, rather than at the GC.

263: 

264: Figure \ref{fig_color34} shows the ratio of 8 $\micron$ / 5.8 $\micron$ emission,

265: i.e. the $I(8 \micron) / I(5.8 \micron)$ color.

266: In this ratio virtually all of the diffuse emission features 

267: disappear, indicating that they have a very uniform color that 

268: is independent of brightness. Dark regions in the ratio map indicate

269: bluer colors associated with numerous individually resolved stars,

270: and with the unresolved stellar background which becomes comparatively 

271: bright in the innermost regions of the galaxy (central $\sim10'$). 

272: There are only a few unusually red point sources. Two of these are

273: asteroids (1567 Alikoski and 459 Signe), and two others are PN M1--26 (HD 316248; 

274: Sahai \& Trauger 1998) and the candidate PN 359.491+0.316 (IRAS 17399--2910; 

275: Becker et al. 1994). Brighter (i.e. redder) extended regions

276: in the ratio map are associated with a few special types of emission features 

277: which will be discussed below, and with the darkest of the IRDCs

278: [e.g. the complex near $(l,b) = (0\fdg24, -0\fdg45)$].

279: The apparent redness of the IRDCs may be an artifact of 

280: residual errors in the subtraction of the zodiacal light or 

281: instrumental offsets (see the Appendix and Table 1). 

282: Small offset errors would have the greatest 

283: impact on the faintest regions of the images, which are the IRDCs

284: and regions at the extreme latitudes along the edges of the mosaics.

285: 

286: \subsubsection{ISM Colors Corrected for the Effect of Unresolved Starlight}

287: To better isolate the ISM emission we have subtracted stellar emission 

288: in a manner similar to that applied to external galaxies (Pahre et al. 2004).

289: In effect, we decompose the 8 and 5.8 $\micron$ emission into ISM and stellar 

290: components using the 4.5 $\micron$ emission as strong tracer of the starlight. 

291: We assume that there exist fixed spatial templates of each component, such that 

292: \begin{equation}

293: I(\lambda) = \alpha_{\lambda}I_{ISM} + \beta_{\lambda}I_{stellar}

294: \end{equation}

295: where $I(\lambda)$ is the observed emission at $\lambda = 4.5$, 5.8 or 8 $\micron$,

296: $I_{ISM}$ and $I_{stellar}$ are fixed (but unknown) templates of the ISM and stellar 

297: emission, and $\alpha_{\lambda}$ and $\beta_{\lambda}$ are the scaling factors

298: for the respective templates. The ratios of the $\alpha_{\lambda}$ or 

299: $\beta_{\lambda}$ coefficients yield the colors of the ISM and stellar emission 

300: at the relevant wavelengths.

301: If the ISM emission were negligible at 4.5 $\micron$ and stellar emission 

302: were negligible at 8 $\micron$, then $\alpha_{4.5 \micron} = \beta_{8 \micron} = 0$

303: and a linear least squares fit of the form 

304: \begin{equation}

305: I(5.8 \micron) = A\ I(8 \micron) + B\ I (4.5 \micron) + C

306: \end{equation}

307: could be used to determine: $A = \alpha_{5.8\micron}/\alpha_{8\micron}$, 

308: the mean $5.8 / 8 \micron$ color of the ISM; 

309: $B = \beta_{5.8\micron}/\beta_{4.5\micron}$, 

310: the mean $5.8 / 4.5 \micron$ color of the starlight; 

311: and $C$, a constant term representing 

312: a linear combination of the mean systematic errors (including those of the detector offsets, 

313: zodiacal light, and the stellar model) in all three channels. 

314: However, because there is no {\it a priori} knowledge of the magnitude or sign

315: of the systematic errors in each channel, we attribute this term to only 

316: the 5.8 $\micron$ channel, which exhibits the greatest instability in detector offset.

317: 

318: In reality, the longer wavelength stellar emission and shorter 

319: wavelength ISM emission are not negligible. Therefore 

320: the interpretation of the $A$, $B$, and $C$ coefficients 

321: derived in Eq. 2 is slightly more complicated.

322: To derive both $\beta_{5.8\micron}/\beta_{4.5\micron}$ and 

323: $\beta_{8\micron}/\beta_{4.5\micron}$ from $A$ and $B$, 

324: we must assume the color of the starlight at one pair of wavelengths. 

325: We choose to set the same $[5.8]-[8.0] = +0.04$ mag 

326: theoretical color that was adopted by Pahre et al (2004), as these colors should 

327: be least affected by extinction. Then we can appropriately scale and subtract

328: $I (4.5 \micron)$ from the 5.8 and 8 $\micron$ emission to remove the mean stellar

329: emission at each of these wavelength. 

330: The derived $[4.5]-[5.8]$ and $[4.5]-[8.0]$ colors are about 0.1 mag redder than those

331: used by Pahre et al. (2004). 

332: Because $I (4.5 \micron)$ also contains a

333: weak $I_{ISM}$ component, after this subtraction of the starlight the ISM 

334: emission is slightly underestimated. The factor by which the ISM is reduced

335: can be calculated from $A$ and $B$, if we again assume the color of the ISM 

336: emission at one set of wavelengths. We use the ``BARE-GR-FG'' 

337: interstellar dust model of Zubko, Dwek, \& Arendt (2004) to calculate that

338: $\alpha_{4.5 \micron}/\alpha_{5.8 \micron} = 0.194$. From this we find that using 

339: $I (4.5 \micron)$ as a template of the starlight leads to underestimating the 

340: ISM emission by factors of 0.85 and 0.97 at 5.8 and 8 $\micron$ respectively. 

341: To summarize we construct:

342: \begin{equation}

343: I_{ISM}(8 \micron) = \left[I(8 \micron) - 0.449\ I(4.5 \micron)\right] / 0.97

344: \end{equation}

345: and

346: \begin{equation}

347: I_{ISM}(5.8 \micron) = \left[I(5.8 \micron) - 0.777\ I(4.5 \micron)\right]/0.85 + 1.68.

348: \end{equation}

349: 

350: The 8/5.8 $\micron$ ISM colors found after subtraction of the starlight (and offset error) 

351: are slightly redder and no longer differ at regions of high extinction, but otherwise appear 

352: quite uniform over a wide range of brightness.  This is depicted

353: in Figure \ref{fig_cmd34}, where the color is plotted as a function of the 8 $\micron$ 

354: brightness. The colors shown here are those of the integrated emission

355: along the line of sight. The shading of this figure is scaled to the log of the

356: number of pixels with a given color -- brightness combination. This logarithmic 

357: scaling enhances the visibility of the outliers in the color. 

358: Some of the unusually red extended regions that stand out in the $I(8\micron)/I(5.8\micron)$

359: colors (both before and after subtraction of starlight) are highlighted and

360: are discussed in \S3.2.

361: The median color ($\pm 1\sigma$ standard deviation) 

362: is $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron) = 2.71 \pm 0.19$,

363: independent of the 8 $\micron$ brightness. 

364: The locations of bright stellar sources 

365: have been excluded from this plot with an intensity threshold cut at 70 MJy sr$^{-1}$

366: at 4.5 $\micron$ and by excluding regions with $I(8 \micron)/I(5.8 \micron) < 2.3$

367: (before the subtraction of starlight), but remaining faint resolved stars 

368: and systematic errors in using the 4.5 $\micron$ emission as the template

369: cause much of the dispersion toward redder colors at low intensities ($<$50 MJy sr$^{-1}$). 

370: The observed $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ colors in the GC are

371: similar to (though slightly bluer than) 

372: the colors observed by Flagey et al. (2006) in 6 regions at longitudes

373: $l = 27\fdg5$, $105\fdg6$, and $254\fdg4$. (These latter colors are indicated by the arrows at

374: the left edge of Figure \ref{fig_cmd34}.) The difference is established using a two--tailed

375: Kolmogorov--Smirnov test, which yields a maximum deviation between the cumulative 

376: distribution function of $D = 0.58$. For the given number of data points, if the two 

377: distributions are intrinsically the same, the probability of $D$ exceeding this value

378: is 0.02. Therefore, this test finds only a 2\% chance that the  

379: Flagey et al. (2006) results are consistent with a random sampling of 

380: the colors that we measure near the GC. However, while this statistic indicates the colors 

381: are likely to be different, we cannot rule out that procedural differences or systematic 

382: errors are the cause of the apparent differences.

383: 

384: \subsubsection{Shorter Wavelength Colors}

385: We are unable to make an analogous comparison to 

386: $I(8 \micron) / I(3.6 \micron)$ and $I(8 \micron) / I(4.5 \micron)$ colors 

387: measured by Flagey et al. (2006) because the very bright and confusing unresolved

388: stellar background toward the GC overwhelms nearly all diffuse ISM emission at 

389: the shorter wavelengths. Point source--subtracted images (see Ram\'irez et al. 2008)

390: provide some aid in distinguishing non--stellar 3.6 and 4.5 $\micron$ emission,

391: but all such regions appear to be H II regions and PDRs associated with specific

392: heating sources, rather than diffuse ISM heated by the 

393: interstellar radiation field (ISRF).

394: 

395: \subsection{Unusual Emission Regions}

396: 

397: Along the Galactic plane, there are a few large ($\gtrsim 4'$) emission 

398: features that stand out in the $I(8 \micron) / I(5.8 \micron)$ color map with unusually red colors.

399: The largest of these is an oblong shell--like structure 

400: centered on the Arches cluster which 

401: contains $\sim$200 stars with masses $>20$ M$_{\sun}$ (Figer et al. 2002). 

402: The shell is much more clearly delineated by the color of its emission, than 

403: by the 8 $\micron$ surface brightness (Figure \ref{red_regions}, top row). 

404: On the north, this shell seems bounded 

405: by the innermost of the thermal radio arches surrounding the cluster. 

406: The shell is only faintly discernible in the shorter wavelength colors.

407: A similar but smaller region appears near

408: the Quintuplet cluster (Figure \ref{red_regions}, middle row), which is comprised of a cluster of hot 

409: O/B and WR stars responsible for ionizing the nearby ISM (Cotera et al. 1996). 

410: However, the five cocoon 

411: stars (Moneti at al. 2001) that give this cluster its name 

412: are so bright at IRAC wavelengths that artifacts from

413: these sources obscure much of this excess. 

414: 

415: There are weaker indications

416: of unusually red emission in the Sgr B1 region as well (Figure \ref{red_regions}, 

417: bottom row). The redder infrared emission is 

418: associated with 5.8 and 8 $\micron$ counterparts of the ``ionized bar'' 

419: and ``ionized rim'' features identified in 3.6, 6 and 20 cm radio observations 

420: by Mehringer et al. (1992). Unlike the Arches and Quintuplet regions, however, the 

421: most luminous stars here are not grouped in a tight easily-identified cluster, but 

422: are apparently dispersed throughout the volume. 

423: 

424: All these extended red regions (the vicinities of the Arches 

425: and Quintuplet Clusters and Sgr B1)

426: can be distinguished in Figure \ref{fig_cmd34} as the cluster of outlier points at 

427: $250 < I(8 $\micron$) < 1000$ MJy sr$^{-1}$ 

428: and $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron) > 3.2$.

429: We expect that as with the compact H II regions, the

430: excessively red emission is caused by thermal emission from very warm dust in strong 

431: radiation fields. However, in these cases the excess emission appears over 

432: a wider region because of the larger number of luminous stars in these 

433: clusters and lower column densities around these more evolved systems allowing

434: a strong radiation field to propagate over a larger volume.

435: 

436: The \object{Pistol Nebula} on the edge of the Quintuplet Cluster is a different type 

437: of red emission region. The 8 $\micron$ IRAC data reveal an 

438: entire shell or bubble around the Pistol Star, with a diameter of 

439: $\sim22''$ (almost 1 pc at the distance of the GC). The brighter portion of this bubble 

440: is likely being ionized by the hot stars in Quintuplet cluster (Figer et al. 1999). 

441: The Pistol Nebula exhibits the most extremely red 

442: $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ color in the 

443: region, as it is only marginally distinguishable at 5.8 $\micron$. 

444: (The 5.8 $\micron$ emission is very difficult to quantify because of strong

445: banding artifacts from the extremely bright stars of the Quintuplet Cluster.)

446: Physically, the emission from this 

447: region is associated with the ejecta from single star (the Pistol Star) 

448: which is an LBV (Figer et al. 1998). In this case, ISO 

449: spectroscopy of the nebula already suggests that the 8 $\micron$ excess here is due to 

450: a relatively red continuum (with deep silicate extinction at 10 $\micron$) and 

451: lack of PAH emission (Moneti et al. 1999). The [Ar II] 6.99 $\micron$ is the strongest line

452: between 5 and 10 $\micron$, but it it not bright enough to be the dominant cause for the 

453: red color. There are no other similar (bright, sharply defined, and very red) 

454: circumstellar bubbles evident in our galactic 

455: center survey. Two partial shells around massive

456: stars near the GC have been identified by Mauerhan et al. (2007). Neither of

457: these shells exhibit especially red $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$

458: color, although in one case red emission is present interior to the shell.

459: 

460: There are only two extended ($\sim20"$) emission sources $>0.1^{\circ}$ away from 

461: the Galactic plane that show strong 8 $\micron$ excesses (Figure \ref{red_regions2}). 

462: Unlike the bright compact H II regions, 

463: these sources are not very bright at 8 $\micron$, and thus even fainter or absent 

464: at 5.8 $\micron$. 

465: However, both brighten rapidly at longer wavelengths

466: and were detected as IRAS point sources at 25 $\micron$ (\object{IRAS 17456-2850} and 

467: \object{IRAS 17425-2920}) and are visible in the MSX

468: band E (21.3 $\micron$) images. Both are located near (in projection) 

469: extended star forming regions (RCW 137 in the case of IRAS 17425--2920) 

470: with strong cirrus emission and IRDCs. Both are slightly 

471: extended and bowed or arced. Neither source contains an obvious point source, nor has a

472: known energizing star in the crowded field. IRAS 17425--2920 does not appear

473: in the 1616 GHz radio maps of the Sgr C region (Liszt \& Spiker 1995).

474: The characteristics and the locations of these 

475: sources on the outskirts of active star forming regions suggest 

476: that these sources represent a very early stage of star formation.

477: Alternately, they could represent local heating by a fully-formed star 

478: that is interacting with an otherwise quiescent 

479: molecular or neutral cloud. A third option is that these objects indicate

480: some sort of shock phenomenon and the unusual colors are the result of strong

481: line emission. For example, the candidate supernova remnant SSTGFLS J222557+601148 

482: exhibits only line emission with no continuum (Morris et al. 2006). 

483: This particular object is bright in the MIPS 24 $\micron$ band by virtue of

484: a strong [O IV] 26 $\micron$ line, but invisible to IRAC. In the Cas A 

485: SNR, [Ar II] at 6.99 $\micron$ can contribute 40\% of the emission 

486: in the IRAC 8 $\micron$ band (Ennis et al. 2006). 

487: Outside of our survey area at $l = 7\arcdeg$, Rho et al. (2006) detect a similar source in 

488: the IRAC and MIPS observations of the Trifid Nebula (Rho et al. 2006). 

489: The region they call ``Trifid Junior'' lies the north of the H II

490: region and the reflection nebula. It is a small arc of emission at 24 $\micron$ which 

491: can only barely be seen at 8 $\micron$ and is absent at 5.8 $\micron$. The general 

492: appearance of the local ($\sim5'$) vicinity is very similar to that of

493: IRAS 17425--2920. 

494: 

495: \subsection{Molecular Emission: A Dearth of 4.5 $\micron$ Excess Sources}

496: Molecular outflows from young stellar objects are often dramatically highlighted

497: in IRAC images because of CO and H$_2$ emission lines that 

498: are especially strong in the 4.5 $\micron$ IRAC band (e.g. Smith et al. 2006). 

499: This wavelength is usually depicted as green in multi--color images, and 

500: such outflows are therefore recognized by their green colors in such images. 

501: While some star forming regions can 

502: contain dozens of outflow sources (e.g. Orion, DR21, NGC 1333),

503: there is only a single small source that is strong at 4.5 $\micron$ {\it and} 

504: has a jet--like morphology within our GC survey (Figure \ref{fig_jet}).

505: There are dozens of other sources that also have 4.5 

506: $\micron$ excesses, but these are unresolved or only slightly diffuse,

507: and usually within or near more extended regions of bright 8 $\micron$ emission.

508: Many of these may be outflows from young stellar objects that 

509: are too distant or intrinsically small to be resolved. Further discussion

510: of these sources can be found in Yusef-Zadeh et al. (2007a).

511: 

512: In several of the large, more local star--forming regions at higher

513: Galactic latitudes there are broad diffuse regions of relatively 

514: enhanced 4.5 $\micron$ emission [e.g. near $(l, b) = (0\fdg48, -0\fdg67)$]. 

515: Rather than being an indication of molecular emission, these may be cases

516: where we spatially resolve an H II region from its surrounding PDR. The 

517: recent study of M 17 by Povich et al. (2007) shows that the absence of PAH emission

518: within the H II region where a strong EUV radiation field is present is well 

519: correlated with an apparent excess at 4.5 $\micron$. Thus the excess is really 

520: only a decrease in the strength of the PAH features that dominate the 5.8 and 8 

521: $\micron$ emission (and to a lesser degree 3.6 $\micron$ emission) 

522: in the surrounding PDR.

523:  

524: 

525: \subsection{Unseen Objects}

526: There are 6 known SNRs 

527: (Sgr A East, G0.3+0.0, G0.9+0.1, G1.0--0.1, G359.0--0.9 and G359.1--0.5; 

528: Green 2006) that lie at least partially within

529: our survey regions, but at 3.6 -- 8 $\micron$ none of these are visible by either their intensity 

530: or color. This is not unexpected given the high degree of confusion and the fact that SNRs

531: are often difficult to distinguish in the infrared even with $2''$ spatial resolution.

532: A thorough examination of the IRAC GLIMPSE survey ($10\arcdeg < |l| < 65\arcdeg$) 

533: by Reach et al. (2006) found IRAC counterparts 

534: for only 18 of 95 SNRs. Those that were detected are undergoing strong interactions

535: with dense, often molecular, clouds. The lack of detections in our survey means that none of 

536: the 6 SNRs toward the GC is undergoing a strong enough interaction to be visible above the 

537: brightest background confusion in the Galaxy. The southern portion of IC 443 is a 

538: well--studied example of a SNR / molecular cloud interaction (e.g. Cesarsky et al. 1999).

539: Emission with the same surface brightness as IC 443 (AORID = 4422912 and 4423168) 

540: could be distinguished at higher latitudes, but would be lost in the confusion

541: at $|b| \lesssim 0\fdg2$. Fainter examples of shocked molecular emission 

542: such as W44 (Reach et al. 2006) would likely be lost in the confusion at any location

543: within our survey.

544: 

545: The enigmatic infrared source known as AFGL 5376 does not show any distinct emission at 

546: any IRAC wavelength, despite being a bright extended source at $\sim24\ \micron$ 

547: wavelengths (Uchida et al. 1990, 1994; Law 2007). We do note that there are several 

548: small IRDCs and filaments (in the vicinity of 

549: $(l,b) = (0\fdg7244,-0\fdg4759)$) that can been seen in extinction 

550: against the 24 $\micron$ emission of AFGL 5376 as well 

551: as the general 3.6 -- 8 $\micron$ background.

552: 

553: There is also no particular correlation between the diffuse infrared emission and 

554: the nonthermal radio filament and threads that are present

555: in the Galactic center region (Yusef-Zadeh et al. 2004; 

556: Stolovy et al. 2006, 2008 in preparation).

557: 

558: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

559: \section{Discussion} 

560: 

561: Overall, the GC exhibits homogenous IRAC colors that are 

562: quite similar to those of other regions of the Galactic plane.

563: Such a uniform $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ color for the 

564: diffuse emission indicates that there is little reddening 

565: associated with extinction at these wavelengths, 

566: despite evidence for highly variable extinction in the form of IRDCs

567: and extinction measurements made at shorter wavelengths (e.g. 

568: Gosling et al. 2006). 

569: This is expected given the flattening of Galactic extinction curve 

570: (Lutz et al. 1996; Indebetouw et al. 2005; Flaherty et al. 2007).

571: Even if the extinction did produce reddening, it would be extremely unlikely 

572: for the line of sight reddening to be correlated with variation in 

573: the intrinsic color of the emission to produce such an overall uniform color.

574: 

575: The uniform colors are also evidence that the emission spectrum is largely independent 

576: of the local and global locations and environments. This strongly suggests that

577: emission from polycyclic aromatic hydrocarbons (PAHs) dominates the diffuse

578: ISM emission (5.8-8 $\micron$), all the way to the Galactic Center. The relative strengths 

579: of PAH emission bands that fall within the IRAC 5.8 and 8 $\micron$ 

580: filter bandpasses are unaffected by the strength of the radiation field heating

581: the grains over a wide range of intensities (Li \& Draine 2001). Even as 

582: the interstellar radiation field strength increases toward the GC, there is no 

583: evident change in the color of the PAH emission. Figure \ref{fig_ZDAmodels} shows model 

584: spectra of the ISM emissivity normalized to the strength of 

585: the interstellar radiation field (Mathis et al. 1983), 

586: as characterized by a scale factor $\chi$. The dust model used 

587: here is composed of PAHs and bare graphite and silicate grains

588: (the ``BARE-GR-FG'' model of Zubko et al. 2004). 

589: This model is similar to those of Li \& Draine (2001), but is derived 

590: via additional constraints provided by elemental abundances. 

591: 

592: For the standard ISRF ($\chi = 1$), most emission at $\lambda < 60\ \micron$

593: is from very small grains ($\lesssim 0.01\ \micron$) and PAHs. 

594: These grains have a low heat capacity 

595: and are heated to high temperatures by absorption of a single photon. 

596: They radiate that energy quickly and cool below equilibrium temperatures

597: before the next energetic photon hits. Thus, for these stochastically heated

598: particles, the grains do not get any warmer as the radiation field increases; 

599: they only go through the same temperature spikes more frequently.

600: Therefore, the spectrum normalized by $\chi$ remains roughly constant.

601: This is especially evident at $\lambda < 10\ \micron$.

602: Larger grains ($\gtrsim 0.01\ \micron$) 

603: have higher heat capacities and absorb photons frequently 

604: enough that they radiate at their equilibrium temperatures.

605: For the local ISRF ($\chi = 1$), this thermal emission peaks at

606: $\sim 150$ $\micron$. As $\chi$ increases, the larger grains get warmer and 

607: the thermal emission moves to shorter wavelengths. Eventually, when 

608: $\chi \gtrsim 10^4$, the thermal emission begins to enter the 8 $\micron$

609: IRAC band. At this point the IRAC colors can be affected.

610: 

611: In Figure \ref{fig_ZDA_ratios} we plot $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ predicted

612: by the models of a function of radiation field strength. 

613: The figure shows the ``BARE-GR-FG'' model is consistent with the observed

614: $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ mean color as long as $\chi \lesssim 10^4$.

615: It is also evident that while the color initially reddens as

616: the thermal emission from the hot dust enters the 8 $\micron$ band,

617: the color begins getting bluer when the thermal emission pushes into the 

618: 5.8 $\micron$ band as well. However, the reddest colors attained by this 

619: model (at $\chi \sim 10^5$) are not nearly red enough to explain the 

620: unusually red regions that are we identify in the GC region (\S 3.2).

621: Alternatively, if the PAH component is removed from the ``BARE-GR-FG'' model,

622: then at $\chi \sim 10^5$ the color of the emission can be sufficiently 

623: red to match the observations. 

624: At $\chi \lesssim 10^4$, the colors of the emission would be much bluer

625: than observed if PAHs were generally absent from the ISM in the GC.

626: Qualitatively similar results are obtained 

627: from the Draine \& Li (2007) dust models, which include a 

628: parameter ($q_{PAH}$) that characterizes the relative abundance of PAHs.

629: 

630: The red shell around the Arches Cluster provides a nice example of

631: a region with a locally strong radiation field. Figer et al. (2002)

632: have estimated that luminosity of the Arches Cluster as $10^{7.8} L_{\sun}$.

633: The extended red shell has a radius of $\sim 100"$ or 3.9 pc.

634: Thus the energy density at the outer edge of this shell is 

635: $\sim 4.4 \times 10^{-9}$ erg cm$^{-3}$. Compared to the nominal 

636: strength of the local ISRF, $7 \times 10^{-13}$ erg cm$^{-3}$ 

637: (Spitzer 1978), we find that $\chi \sim 6300$ at the edge of the shell.

638: This value of $\chi$ indicates that we should indeed expect to see 

639: emission with red $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ colors in the vicinity 

640: of the Arches Cluster, and that the region of red emission cannot 

641: be much larger than what is observed. 

642: 

643: The red emission region near the Quintuplet Cluster is related to a 

644: similarly enhanced radiation field. This region corresponds to the 

645: portion of the ``Sickle'' that is nearest to the Quintuplet Cluster.

646: Spectroscopy of this region by Simpson et al. (1997, 2007) reveal 

647: gas in a state of high excitation, consistent with radiation fields 

648: of $\chi > 10^3$. Calculations by Rodr\'iguez-Fern\'andez et al.

649: (2001) indicate that the Quintuplet Cluster will fill a smaller volume 

650: with a strong radiation field than the Arches Cluster. 

651: This is in agreement with the smaller extent of the red emission 

652: in the vicinity of the Quintuplet Cluster

653: compared to the Arches Cluster (Fig. \ref{red_regions}). 

654: In particular, the extent of the radiation field strong enough to cause 

655: redder $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ colors does not reach to the 

656: (Galactic) southern end of the ``Radio Arc Bubble''. This is 

657: consistent with the fact that while $20-25 \micron$ observations 

658: (IRAS, MSX, MIPS) show a complete bubble, the portions more distant 

659: from the Quintuplet cluster do not exhibit unusually red 

660: $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ colors. Comparison of the 

661: spectroscopic results of Simpson et al. (2007) indicates that the 

662: $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ color is fairly well correlated 

663: (correlation coefficient $r > 0.6$) with 

664: the continuum intensity at $\lambda > 10\ \micron$, and with 

665: H 7-6 12.37 $\micron$, [Ne II] 12.81 $\micron$, [P III] 17.89 $\micron$, 

666: [S III] 18.71, 33.48 $\micron$, and [S IV] 10.51 $\micron$. 

667: The correlation of the $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ color with the 

668: $\lambda > 10\ \micron$ continuum color, or with line ratios 

669: is weaker. An exception is that correlation with 

670: [Si II] 34.83 $\micron$ / [Fe II] 26 $\micron$ 

671: (or 17.94 $\micron$) is quite strong, $r > 0.75$.

672: This is rather surprising as both species have similar ionization potentials

673: and would be similarly affected by depletion into (or destruction of) dust grains.

674: 

675: 

676: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

677: \section{Summary}

678: 

679: Our {\it Spitzer}/IRAC survey of the Galactic Center 

680: (Stolovy et al. 2006, 2008 in preparation) 

681: reveals prominent 5.8 and 8 $\micron$ emission from the ISM in the 

682: central $\sim$ 200 pc of the Galaxy in unprecedented detail. 

683: The infrared emission is loosely associated with molecular

684: emission from the CMZ, and several times brighter than 

685: infrared emission elsewhere in the Galactic plane. In general, the 

686: $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ color of the emission is very uniform 

687: across the CMZ, and similar to that seen in other parts of the Galactic 

688: plane (e.g. Flagey et al. 2006). 

689: Clearly we cannot infer too much from a single color, given the 

690: large number of parameters that can be adjusted for the dust models

691: used to interpret the emission. However, we can generally conclude 

692: the predominant uniformity of the observed $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ 

693: color in the GC implies 

694: (1) the emission is dominated by PAHs, 

695: (2) the general ISRF in the central 200 pc of the Galaxy does not reach extreme 

696: values ($\chi \lesssim 10^4$), and 

697: (3) the interstellar extinction law at these wavelengths is 

698: flat (e.g. Indebetouw et al. 2005; Flaherty et al. 2007) leading to no significant

699: reddening despite obvious cases of locally high extinction.

700: 

701: There are several extended regions that exhibit much redder than average

702: $I_{ISM}(8 \micron)/I_{ISM}(5.8 \micron)$ colors despite not being especially bright.

703: Red regions around the Arches Cluster, the Quintuplet Cluster, and Sgr B1 

704: are apparently produced by a combination of locally strong radiation 

705: fields ($\chi \gtrsim 10^4$), and a lack of PAHs (perhaps a consequence

706: of the strong radiation field). Other very red regions that are more compact 

707: and further from the Galactic plane appear to be associated with 

708: the onset of star formation, although they do not appear to contain any

709: detectable point sources.

710: 

711: \acknowledgments

712: This work is based on observations made with the Spitzer Space 

713: Telescope, which is operated by the Jet Propulsion Laboratory, California 

714: Institute of Technology under a contract with NASA. Support for this 

715: work was provided by NASA through an award issued by JPL/Caltech.

716: This research has made use of the NASA/ IPAC Infrared Science Archive, which 

717: is operated by the Jet Propulsion Laboratory, California Institute of Technology, 

718: under contract with the National Aeronautics and Space Administration. 

719: We thank the referee for pointing out the non-negligible influence

720: of the 4.5 $\micron$ ISM emission in our analysis.

721: 

722: {\it Facilities:} \facility{Spitzer (IRAC)}

723: 

724: \appendix

725: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

726: \section{Data Reduction Details}

727: \subsection{Post--BCD Processing}

728: 

729: The BCD are single frames (exposures) calibrated with the standard calibration data 

730: applied by the Spitzer Science Center (SSC) pipeline. While on average this calibration is

731: quite good, in some cases the state of the IRAC detectors may be different from what the 

732: standard calibration assumes. Thus the BCD can still contain some types of instrumental 

733: artifacts, and the individual frames may not blend smoothly into a large mosaicked image,

734: especially when that image incorporates multiple AORs.

735: 

736: One significant artifact in the 3.6 and 4.5 $\micron$ channels is ``column pulldown''.

737: This is an electronic artifact that induces low intensities in detector 

738: columns coincident with bright (usually saturated) sources. 

739: A standard correction was applied to individual BCD frames for the 3.6 and 4.5 $\micron$ channels. 

740: This greatly mitigates the effect, but does not reduce it to undetectable levels. 

741: A similar correction for banding was applied to 8 $\micron$ BCD frames 

742: to eliminate excess intensity in detector rows containing very bright sources. The 5.8

743: $\micron$ data could have benefitted from such corrections in both detector 

744: rows and columns. However the character of the artifact is sufficiently different 

745: from the behavior at 8 $\micron$,

746: that attempts at correcting 5.8 $\micron$ banding were not successful enough to 

747: apply them in an automated way. Thus these artifacts remain, and are a source of 

748: systematic error in particular regions where saturated sources are clustered. 

749: 

750: The BCD frames for each AOR were self-calibrated using the procedure 

751: described by Fixsen et al. (2000). This procedure uses the bright infrared 

752: background emission (combined zodiacal light and Galactic emission) as a calibration source.

753: The background is not uniform, but it is stable. Thus, when 

754: the dithering and mapping place different detector pixels on the same sky location, the 

755: resulting values for {\it properly} calibrated data should be the same, apart from a random 

756: noise variation from gaussian detector noise and Poisson photon shot noise. Not all

757: detector pixels observe a single sky location, but with a large data set such as

758: ours, if two different detector pixels did not observe a common sky location then there

759: is an intermediate pixel (or a chain of several such pixels) that has observed at least

760: one sky location in common with each of the original two pixels. Through such inter-comparisons,

761: the relative calibrations of each detector pixel can be determined. This is solved

762: using a linear least--squares algorithm. In the framework of this procedure 

763: the data ($D^i$, i.e. each pixel of each BCD frame) are modeled 

764: as the sum of the true sky intensity ($S^{\alpha}$ at sky coordinates $\alpha$),

765: a fixed detector offset ($F^p$ as a function of detector pixel $p$), 

766: and a variable offset ($F^q$ a single value for all pixels of frame $q$):

767: \begin{equation}

768: D^i = S^{\alpha} + F^p + F^q .

769: \end{equation}

770: The least--squares self--calibration allows us to determine the offset terms in the

771: model (i.e. the proper detector calibration), and then invert Eq. A1 to derived the 

772: true sky intensity, $S^{\alpha}$ from the data, $D^i$. 

773: The images of the fixed detector offsets ($F^p$) are shown in Figure \ref{fig_fp} for each 

774: IRAC channel and each AOR. For the 3.6 and 4.5 $\micron$ data, the obvious feature is the 

775: diffuse stray light ``butterfly wing'' pattern at the top of the detector. This feature 

776: represents the integrated stray light of all the sources of the bright unresolved 

777: background, which is not effectively removed through the standard BCD calibration.

778: The amplitude of this artifact is small and is not evident directly in the single BCD frames or the mosaics, 

779: but if left uncorrected it would induce artifacts in the ratio of 3.6 / 4.5 $\micron$ emission as

780: shown in Figure \ref{fig_rat12}. This figure also illustrates that the column pulldown correction is 

781: not entirely effective at both 3.6 and 4.5 $\micron$ for the very brightest sources, but the 

782: pulldown correction usually does improve the data for weaker, more frequent cases. 

783: The fixed detector offsets for the 5.8 and 8 $\micron$ channels also show (weakly)

784: the diffuse stray light patterns for these channels (Fig. \ref{fig_fp}). 

785: Additionally the 8 $\micron$ results clearly show the accumulation of latent 

786: images from bright sources, 

787: the clearing of these after an anneal, and the further accumulation of new latent images.

788: The strong latent images acquired at 8 $\micron$ during AOR 0s, 

789: are from the central cluster at Sgr A.

790: The relative strength and extent of the latent image indicates that the Sgr A cluster

791: is clearly the highest density of bright sources in our survey.

792: 

793: The variable detector offset ($F^q$) is illustrated in Figures \ref{fig_fqplot} and \ref{fig_fqmap}.

794: The time sequence 

795: of these shows a regular variation each time one of the AOR strips crosses the brightest

796: part of the Galactic plane. For the InSb detectors used at 3.6 and 4.5 $\micron$,

797: $|F^q| \lesssim 0.5$ MJy/sr and is usually much smaller than the observed 

798: sky intensities (which do not include zodiacal light) as shown in Table~1.

799: The variable offsets can be much larger for the 5.8 and 8 $\micron$ 

800: detectors, as seen in Fig. \ref{fig_fqplot}. Here the variable 

801: offsets are correcting for any residual errors in the ``first--frame'' effect (a change in 

802: detector offset as a function of time since the preceding frame), and more noticeably

803: for detector droop (where the detector offset changes as a function of the total 

804: illumination on the detector). 

805: Figure \ref{fig_rat34} shows the ratio of our mosaic compared to a 

806: standard post-BCD pipeline mosaic for one strip of our survey (AORID = 13368320) 

807: at 5.8 and 8 $\micron$. 

808: This ratio clearly shows artifacts at the frame boundaries, indicating offset errors

809: in one (or both) of the mosaics. These errors are largest at the bright emission 

810: through the center of the strip at $b = 0$, especially for the 8 $\micron$ channel.

811: The facts that (a) the frame seams are faintly visible in the 

812: post-BCD mosaic and not our self--calibrated versions, and (b) the subsequent science analysis reveals 

813: no similar discrepancies in the ratios of 5.8 to 8 $\micron$ emission despite independently

814: derived corrections for each channel, indicate that the variable offsets derived 

815: from the self--calibration are appropriate. In other words, the self--calibration procedure

816: does remove artifacts from the images which would otherwise cause systematic errors 

817: when comparing intensities at different wavelengths.

818: 

819: 

820: \subsection{Extended Source Calibration and Zodiacal Light}

821: 

822: The absolute calibration of IRAC is based on the photometry of point sources

823: in a fixed size aperture. In the present study, we are interested in the surface

824: brightness of extended emission. Therefore, we have applied extended source corrections

825: to the data used in this analysis. Specifically, the factors used here are:

826: [0.944,0.937,0.772,0.737] at 3.6, 4.5, 5.8, and 8 $\micron$ respectively

827: (Reach et al. 2005; Cohen et al. 2007).

828: 

829: Accurate determination of the colors of diffuse emission also relies on 

830: having properly subtracted any instrumental offsets. 

831: The major difficulty in measuring the correct instrumental offset

832: is that IRAC does not use its shutter to take dark frames. 

833: Instead, ``skydarks'' (observations at a dark location near 

834: the north ecliptic pole) are used to in the pipeline 

835: to remove instrumental offsets. This works well as long as the instrumental offsets 

836: remain stable between the time that the skydarks are measured and the science data 

837: are observed. IRAC's 5.8 $\micron$ channel show the greatest instabilities in its

838: offsets. The lack of constraints that would be provide by true dark frames also

839: leads to an offset degeneracy in the self--calibration of these data. 

840: Inspection of Eq. 1 reveals that the procedure is 

841: insensitive to an arbitrary offset to the derived sky intensity $S^{\alpha}$ as long as

842: the same offset is subtracted from the derived instrumental offset $F^p$ (or $F^q$).

843: 

844: Finally, for measurement of the intensity and color of diffuse Galactic emission 

845: the zodiacal light foreground also must be subtracted. The zodiacal light 

846: must be estimated via a model. The model used for {\it Spitzer} 

847: is derived from the Kelsall et al. (1998) model of the zodiacal light emission

848: as observed by the DIRBE instrument on the {\it COBE} spacecraft. The model is 

849: estimated to be accurate to $\sim 2\%$, however larger unrecognized systematic 

850: errors may be present. Its use for IRAC data requires interpolation

851: to IRAC's wavelengths, and, more significantly, extrapolation from the view 

852: in Earth orbit to the view from {\it Spitzer}'s location several tenths of

853: an astronomical unit away from Earth. Additionally, because of the skydark subtraction,

854: the zodiacal light must be known along the line of sight to the north ecliptic

855: pole as well as toward the Galactic center. Thus the zodiacal light intensity we 

856: subtract from the mosaic images is the difference between that estimated for 

857: the Galactic Center observations and that of the skydark observations. 

858: This is calculated from the mean values of the ZODY$\_$EST and SKYDRKZB keywords 

859: in the BCD headers, except for the 5.8 $\micron$ channel where these keyword 

860: values are in error and are inconsistent with the model spectrum; thus the 

861: estimates from SPOT\footnote{http://ssc.spitzer.caltech.edu/propkit/spot/} are 

862: used instead for the 5.8 $\micron$ data. 

863: We do not account for the small gradient in the zodiacal light intensity. 

864: 

865: Table 1 lists the median and mean intensities of the final GC mosaics, along with the 

866: self--calibration offsets, and the estimated intensities of the zodiacal light.

867: The zodiacal light is essentially negligible for the 3.6 and 4.5 $\micron$ data,

868: but could potentially influence 5.8 and 8 $\micron$ results where intensities are low. 

869: The variable offset per frame ($F^q$) can reach large values, but 

870: as revealed in Figure \ref{fig_fqmap},

871: the largest values are closely confined to Galactic plane where intensities are also 

872: much higher than the mean or median values.

873: 

874: 

875: \begin{thebibliography}{}

876: 

877: \bibitem[Altenhoff et al.(1979)]{1979A&AS...35...23A} Altenhoff, W.~J., 

878: Downes, D., Pauls, T., \& Schraml, J.\ 1979, \aaps, 35, 23 

879: 

880: \bibitem[Armstrong \& Barrett(1985)]{1985ApJS...57..535A} Armstrong, J.~T., 

881: \& Barrett, A.~H.\ 1985, \apjs, 57, 535 

882: 

883: \bibitem[Becker et al.(1994)]{1994ApJS...91..347B} Becker, R.~H., White, 

884: R.~L., Helfand, D.~J., \& Zoonematkermani, S.\ 1994, \apjs, 91, 347 

885: 

886: \bibitem[Benjamin et al.(2003)]{2003PASP..115..953B} Benjamin, R.~A., et 

887: al.\ 2003, \pasp, 115, 953 

888: 

889: \bibitem[Cesarsky et al.(1999)]{1999A&A...348..945C} Cesarsky, D., Cox, P., 

890: Pineau des For{\^e}ts, G., van Dishoeck, E.~F., Boulanger, F., \& Wright, 

891: C.~M.\ 1999, \aap, 348, 945 

892: 

893: \bibitem[Christopher et al.(2005)]{2005ApJ...622..346C} Christopher, M.~H., 

894: Scoville, N.~Z., Stolovy, S.~R., \& Yun, M.~S.\ 2005, \apj, 622, 346 

895: 

896: \bibitem[Cohen et al.(2007)]{2007MNRAS.374..979C} Cohen, M., et al.\ 2007, 

897: \mnras, 374, 979 

898: 

899: \bibitem[Cotera et al.(1996)]{1996ApJ...461..750C} Cotera, A.~S., Erickson, 

900: E.~F., Colgan, S.~W.~J., Simpson, J.~P., Allen, D.~A., \& Burton, M.~G.\ 

901: 1996, \apj, 461, 750 

902: 

903: \bibitem[Cox \& Laureijs(1989)]{1989IAUS..136..121C} Cox, P., \& Laureijs, 

904: R.\ 1989, The Center of the Galaxy, 136, 121 

905: 

906: \bibitem[Draine \& Li(2007)]{2007ApJ...657..810D} Draine, B.~T., \& Li, A.\ 

907: 2007, \apj, 657, 810 

908: 

909: \bibitem[Ennis et al.(2006)]{2006ApJ...652..376E} Ennis, J.~A., Rudnick, 

910: L., Reach, W.~T., Smith, J.~D., Rho, J., DeLaney, T., Gomez, H., \& Kozasa, 

911: T.\ 2006, \apj, 652, 376 

912: 

913: \bibitem[Fazio et al.(2004)]{2004ApJS..154...10F} Fazio, G.~G., et al.\ 

914: 2004, \apjs, 154, 10 

915: 

916: \bibitem[Figer et al.(1999)]{1999ApJ...525..759F} Figer, D.~F., Morris, M., 

917: Geballe, T.~R., Rich, R.~M., Serabyn, E., McLean, I.~S., Puetter, R.~C., \& 

918: Yahil, A.\ 1999, \apj, 525, 759 

919: 

920: \bibitem[Figer et al.(1998)]{1998ApJ...506..384F} Figer, D.~F., Najarro, 

921: F., Morris, M., McLean, I.~S., Geballe, T.~R., Ghez, A.~M., \& Langer, N.\ 

922: 1998, \apj, 506, 384 

923: 

924: \bibitem[Figer et al.(2002)]{2002ApJ...581..258F} Figer, D.~F., et al.\ 

925: 2002, \apj, 581, 258 

926: 

927: \bibitem[Fixsen et al.(2000)]{2000ApJS..128..651F} Fixsen, D.~J., Moseley, 

928: S.~H., \& Arendt, R.~G.\ 2000, \apjs, 128, 651 

929: 

930: \bibitem[Flagey et al.(2006)]{2006A&A...453..969F} Flagey, N., Boulanger, 

931: F., Verstraete, L., Miville Desch{\^e}nes, M.~A., Noriega Crespo, A., \& 

932: Reach, W.~T.\ 2006, \aap, 453, 969 

933: 

934: \bibitem[Flaherty et al.(2007)]{2007ApJ...663.1069F} Flaherty, K.~M., 

935: Pipher, J.~L., Megeath, S.~T., Winston, E.~M., Gutermuth, R.~A., Muzerolle, 

936: J., Allen, L.~E., \& Fazio, G.~G.\ 2007, \apj, 663, 1069 

937: 

938: \bibitem[Gosling et al.(2006)]{2006ApJ...640L.171G} Gosling, A.~J., 

939: Blundell, K.~M., \& Bandyopadhyay, R.\ 2006, \apjl, 640, L171 

940: 

941: \bibitem[Green (2006)]{2006web} Green D. A., 2006, `A Catalogue of 

942: Galactic Supernova Remnants (2006 April version)', Astrophysics Group, 

943: Cavendish Laboratory, Cambridge, United Kingdom 

944: (available at ``http://www.mrao.cam.ac.uk/surveys/snrs/'')

945: 

946: \bibitem[Huettemeister et al.(1998)]{1998A&A...334..646H} H\"uttemeister, 

947: S., Dahmen, G., Mauersberger, R., Henkel, C., Wilson, T.~L., \& 

948: Martin-Pintado, J.\ 1998, \aap, 334, 646 

949: 

950: \bibitem[Huettemeister et al.(1993)]{1993A&A...280..255H} H\"uttemeister, 

951: S., Wilson, T.~L., Bania, T.~M., \& Martin-Pintado, J.\ 1993, \aap, 280, 

952: 255 

953: 

954: \bibitem[Indebetouw et al.(2005)]{2005ApJ...619..931I} Indebetouw, R., et 

955: al.\ 2005, \apj, 619, 931 

956: 

957: \bibitem[Kelsall et al.(1998)]{1998ApJ...508...44K} Kelsall, T., et al.\ 

958: 1998, \apj, 508, 44 

959: 

960: \bibitem[Law(2007)]{2007arXiv0705.0114L} Law, C.~J.\ 2007, Ph. D. thesis, ArXiv e-prints, 

961: 705, arXiv:0705.0114 

962: 

963: \bibitem[Li \& Draine(2001)]{2001ApJ...554..778L} Li, A., \& Draine, B.~T.\ 

964: 2001, \apj, 554, 778 

965: 

966: \bibitem[Liszt \& Spiker(1995)]{1995ApJS...98..259L} Liszt, H.~S., \& 

967: Spiker, R.~W.\ 1995, \apjs, 98, 259 

968: 

969: \bibitem[Lutz et al.(1996)]{1996A&A...315L.269L} Lutz, D., et al.\ 1996, 

970: \aap, 315, L269 

971: 

972: \bibitem[Martin et al.(2004)]{2004ApJS..150..239M} Martin, C.~L., Walsh, 

973: W.~M., Xiao, K., Lane, A.~P., Walker, C.~K., \& Stark, A.~A.\ 2004, \apjs, 

974: 150, 239 

975: 

976: \bibitem[Martin-Pintado et al.(1997)]{1997ApJ...482L..45M} Martin-Pintado, 

977: J., de Vicente, P., Fuente, A., \& Planesas, P.\ 1997, \apjl, 482, L45 

978: 

979: \bibitem[Mathis et al.(1983)]{1983A&A...128..212M} Mathis, J.~S., Mezger, 

980: P.~G., \& Panagia, N.\ 1983, \aap, 128, 212 

981: 

982: \bibitem[Mauerhan et al.(2007)]{2007ApJ...662..574M} Mauerhan, J.~C., Muno, 

983: M.~P., \& Morris, M.\ 2007, \apj, 662, 574 

984: 

985: \bibitem[Mehringer et al.(1992)]{1992ApJ...401..168M} Mehringer, D.~M., 

986: Yusef-Zadeh, F., Palmer, P., \& Goss, W.~M.\ 1992, \apj, 401, 168 

987: 

988: \bibitem[Mezger(1978)]{1978A&A....70..565M} Mezger, P.~G.\ 1978, \aap, 70, 565 

989: 

990: \bibitem[Moneti et al.(1999)]{1999usis.conf..723M} Moneti, A., Blommaert, 

991: J.~A.~D.~L., Najarro, F., Figer, D., \& Stolovy, S.\ 1999, ESA SP-427: The 

992: Universe as Seen by ISO, 723 

993: 

994: \bibitem[Moneti et al.(2001)]{2001A&A...366..106M} Moneti, A., Stolovy, S., 

995: Blommaert, J.~A.~D.~L., Figer, D.~F., \& Najarro, F.\ 2001, \aap, 366, 106 

996: 

997: \bibitem[Morris \& Serabyn(1996)]{1996ARA&A..34..645M} Morris, M., \& 

998: Serabyn, E.\ 1996, \araa, 34, 645 

999: 

1000: \bibitem[Morris et al.(2006)]{2006ApJ...640L.179M} Morris, P.~W., Stolovy, 

1001: S., Wachter, S., Noriega-Crespo, A., Pannuti, T.~G., \& Hoard, D.~W.\ 2006, 

1002: \apjl, 640, L179 

1003: 

1004: \bibitem[Odenwald \& Fazio(1984)]{1984ApJ...283..601O} Odenwald, S.~F., \& 

1005: Fazio, G.~G.\ 1984, \apj, 283, 601 

1006: 

1007: \bibitem[Oka et al.(2005)]{2005ApJ...632..882O} Oka, T., Geballe, T.~R., 

1008: Goto, M., Usuda, T., \& McCall, B.~J.\ 2005, \apj, 632, 882 

1009: 

1010: \bibitem[Pahre et al.(2004)]{2004ApJS..154..229P} Pahre, M.~A., Ashby, 

1011: M.~L.~N., Fazio, G.~G., \& Willner, S.~P.\ 2004, \apjs, 154, 229 

1012: 

1013: \bibitem[Pauls \& Mezger(1975)]{1975A&A....44..259P} Pauls, T., \& Mezger, 

1014: P.~G.\ 1975, \aap, 44, 259 

1015: 

1016: \bibitem[Paumard et al.(2006)]{2006ApJ...643.1011P} Paumard, T., et al.\ 

1017: 2006, \apj, 643, 1011 

1018: 

1019: \bibitem[Povich et al.(2007)]{2007ApJ...660..346P} Povich, M.~S., et al.\ 

1020: 2007, \apj, 660, 346 

1021: 

1022: \bibitem[Price et al.(2001)]{2001AJ....121.2819P} Price, S.~D., Egan, 

1023: M.~P., Carey, S.~J., Mizuno, D.~R., \& Kuchar, T.~A.\ 2001, \aj, 121, 2819 

1024: 

1025: \bibitem[Ramirez et al.(2008)]{2008} Ram\'{\i}rez, S., et al. 2008, \apjs, 175, 147

1026: 

1027: \bibitem[Reach et al.(2005)]{2005PASP..117..978R} Reach, W.~T., et al.\ 

1028: 2005, \pasp, 117, 978 

1029: 

1030: \bibitem[Reach et al.(2006)]{2006AJ....131.1479R} Reach, W.~T., et al.\ 

1031: 2006, \aj, 131, 1479 

1032: 

1033: \bibitem[Rho et al.(2006)]{2006ApJ...643..965R} Rho, J., Reach, W.~T., 

1034: Lefloch, B., \& Fazio, G.~G.\ 2006, \apj, 643, 965 

1035: 

1036: \bibitem[Rodr{\'{\i}}guez-Fern{\'a}ndez \& 

1037: Mart{\'{\i}}n-Pintado(2005)]{2005A&A...429..923R} 

1038: Rodr{\'{\i}}guez-Fern{\'a}ndez, N.~J., \& Mart{\'{\i}}n-Pintado, J.\ 2005, 

1039: \aap, 429, 923 

1040: 

1041: \bibitem[Rodr{\'{\i}}guez-Fern{\'a}ndez et al.(2001)]{2001A&A...377..631R} 

1042: Rodr{\'{\i}}guez-Fern{\'a}ndez, N.~J., Mart{\'{\i}}n-Pintado, J., \& de 

1043: Vicente, P.\ 2001, \aap, 377, 631 

1044: 

1045: \bibitem[Rodr{\'{\i}}guez-Fern{\'a}ndez et al.(2004)]{2004A&A...427..217R} 

1046: Rodr{\'{\i}}guez-Fern{\'a}ndez, N.~J., Mart{\'{\i}}n-Pintado, J., Fuente, 

1047: A., \& Wilson, T.~L.\ 2004, \aap, 427, 217 

1048: 

1049: \bibitem[Sahai \& Trauger(1998)]{1998AJ....116.1357S} Sahai, R., \& 

1050: Trauger, J.~T.\ 1998, \aj, 116, 1357 

1051: 

1052: \bibitem[Simpson et al.(1997)]{1997ApJ...487..689S} Simpson, J.~P., Colgan, 

1053: S.~W.~J., Cotera, A.~S., Erickson, E.~F., Haas, M.~R., Morris, M., \& 

1054: Rubin, R.~H.\ 1997, \apj, 487, 689 

1055: 

1056: \bibitem[Simpson et al.(2007)]{2007arXiv0708.2103S} Simpson, J.~P., Colgan, 

1057: S.~W.~J., Cotera, A.~S., Erickson, E.~F., Hollenbach, D.~J., Kaufman, 

1058: M.~J., \& Rubin, R.~H.\ 2007, ApJ, 670, 1115 

1059: 

1060: \bibitem[Smith et al.(2006)]{2006ApJ...645.1264S} Smith, H.~A., Hora, 

1061: J.~L., Marengo, M., \& Pipher, J.~L.\ 2006, \apj, 645, 1264 

1062: 

1063: \bibitem[Spitzer(1978)]{1978ppim.book.....S} Spitzer, L.\ 1978, New York 

1064: Wiley-Interscience, 1978.~333 p.  

1065: 

1066: \bibitem[Stark et al.(1989)]{1989IAUS..136..129S} Stark, A.~A., Bally, J., 

1067: Wilson, R.~W., \& Pound, M.~W.\ 1989, The Center of the Galaxy, 136, 129 

1068: 

1069: \bibitem[Stolovy et al.(2006)]{2006JPhCS..54..176S} Stolovy, S., et al.\ 

1070: 2006, Journal of Physics Conference Series, 54, 176 

1071: 

1072: \bibitem[Uchida et al.(1990)]{1990ApJ...351..443U} Uchida, K., Morris, M., 

1073: \& Serabyn, E.\ 1990, \apj, 351, 443 

1074: 

1075: \bibitem[Uchida et al.(1994)]{1994ApJ...421..505U} Uchida, K.~I., Morris, 

1076: M.~R., Serabyn, E., \& Bally, J.\ 1994, \apj, 421, 505 

1077: 

1078: \bibitem[Werner et al.(2004)]{2004ApJS..154....1W} Werner, M.~W., et al.\ 

1079: 2004, \apjs, 154, 1 

1080: 

1081: \bibitem[Wilson et al.(1982)]{1982A&A...115..185W} Wilson, T.~L., Ruf, K., 

1082: Walmsley, C.~M., Martin, R.~N., Batrla, W., \& Pauls, T.~A.\ 1982, \aap, 

1083: 115, 185 

1084: 

1085: \bibitem[Yusef-Zadeh et al.(2004)]{2004ApJS..155..421Y} Yusef-Zadeh, F., 

1086: Hewitt, J.~W., \& Cotton, W.\ 2004, \apjs, 155, 421 

1087: 

1088: \bibitem[Yusef-Zadeh et al.(2007b)]{2007ApJ...665L.123Y} Yusef-Zadeh, F., 

1089: Wardle, M., \& Roy, S.\ 2007b, \apjl, 665, L123 

1090: 

1091: \bibitem[Yusef-Zadeh et al.(2007a)]{2007arXiv0705.3001Y} Yusef-Zadeh, F., et 

1092: al.\ 2007a, ArXiv e-prints, 705, arXiv:0705.3001 

1093: 

1094: \bibitem[Zubko et al.(2004)]{2004ApJS..152..211Z} Zubko, V., Dwek, E., \& 

1095: Arendt, R.~G.\ 2004, \apjs, 152, 211 

1096: 

1097: \end{thebibliography}

1098: 

1099: 

1100: \clearpage

1101: 

1102: \begin{deluxetable}{lccccccccccc}

1103: \tabletypesize{\scriptsize}

1104: \tablewidth{0pt}

1105: \tablecaption{Comparison of the Observed Sky Intensities and Offset Terms}

1106: \tablehead{

1107: \colhead{} &

1108: \multicolumn{2}{c}{$I$ (MJy sr$^{-1}$)} & & 

1109: \colhead{$Z$ (MJy sr$^{-1}$)} & & 

1110: \multicolumn{2}{c}{$F^p$ (MJy sr$^{-1}$)} & &

1111: \multicolumn{3}{c}{$F^q$ (MJy sr$^{-1}$)} \\

1112: \cline{2-3}\cline{5-5}\cline{7-8}\cline{10-12}

1113: \colhead{Wavelength ($\micron$)} &

1114: \colhead{median} & 

1115: \colhead{mean} & &

1116: \colhead{mean} & &

1117: \colhead{median} & 

1118: \colhead{mean} & &

1119: \colhead{$\sigma$} & 

1120: \colhead{median} &

1121: \colhead{max}

1122: }

1123: \startdata

1124: 3.6 & 14.1 &  23.9 & & 0.097 & & -0.0367 & 0.0173 & & 0.186 & 0.0139  & 1.92  \\

1125: 4.5 & 11.0 &  18.5 & & 0.193 & & -0.0613 & 0.0110 & & 0.201 & 0.00906 & 0.874 \\

1126: 5.8 & 26.7 &  33.9 & & 1.298 & &  0.0331 & 0.0686 & & 0.930 & 0.0839  & 8.49  \\

1127: 8   & 66.1 &  77.1 & & 6.181 & & -0.0663 & -0.315 & & 9.14  & 0.659   & 99.3

1128: \enddata

1129: \tablecomments{$I$ = the observed intensity in the final image. $Z$ = the difference between 

1130: the zodiacal light toward the GC and that toward the skydark calibration field, 

1131: $F^p$ = self--calibration detector offset per detector pixel $p$, $F^q$ = 

1132: self--calibration detector offset per frame $q$.

1133: }

1134: \end{deluxetable}

1135: 

1136: \clearpage

1137: 

1138: \begin{figure}

1139: \includegraphics[scale=0.80,angle=180]{f1.eps}

1140: \caption{The mosaic of the {\it Spitzer} IRAC Galactic Center survey at 8 $\micron$. 

1141: Intensities are scaled logarithmically from 10 to 10$^3$ MJy sr$^{-1}$.

1142: The brightest diffuse emission is correlated with the Central Molecular Zone (CMZ).

1143: \label{fig_ch4}}

1144: \end{figure}

1145: 

1146: \begin{figure}

1147: \plotone{f2.eps}

1148: \caption{Galactic Center region at 8.28 $\micron$ as seen by the MSX

1149: mission (logarithmic intensity scale). The white contour indicated the 

1150: extent of our Spitzer IRAC survey. The bright Central Molecular 

1151: Zone at a galactocentric radius of $r < 100$ pc is clearly distinct 

1152: from the wider span of the inner galaxy $3\arcdeg > l > -3\arcdeg$

1153: ($r < 400$ pc).

1154: \label{fig_msx}}

1155: \end{figure}

1156: 

1157: \begin{figure}

1158: \plotone{f3.eps}

1159: \caption{Intensity slices through the MSX 8.28 $\micron$ data at several latitudes.

1160: The data have been binned to $24"$ pixels and medians are taken over a $6\farcm8$

1161: swath in latitude. The slice at $b = 0$ shows that the structures of the 

1162: Central Molecular Zone are 2 -- 10 times brighter than the emission in 

1163: other parts of the Galactic plane.\label{fig_msxslice}}

1164: \end{figure}

1165: 

1166: \begin{figure}

1167: \includegraphics[scale=0.80,angle=180]{f4.eps}

1168: \caption{Galactic Center region depicted in the $I (8 \micron) / I(5.8 \micron)$ color

1169: (linear grayscale).

1170: This color shows little variation or correlation with the 8 $\micron$ intensity 

1171: (Fig. \ref{fig_ch4}). Stellar sources have $I (8 \micron) / I(5.8 \micron) < 1$ and 

1172: appear black here. Only a few extended regions have redder than average colors. Several 

1173: of these are shown in detail in Figs. \ref{red_regions} and \ref{red_regions2}. 

1174: The distinct complex near $(l,b) = (0\fdg24,-0\fdg45)$ 

1175: is at the darkest of the IRDCs, and thus may be affected by small 

1176: errors in subtraction of the zodiacal light emission or detector offsets. 

1177: \label{fig_color34}}

1178: \end{figure}

1179: 

1180: \begin{figure}

1181: \plotone{f5.eps}

1182: \caption{The 8/5.8 $\micron$ color vs. brightness in the Galactic Center region.

1183: The shading represents the diffuse emission. Contours indicate the 

1184: regions containing 50, 90, and 99\% of the data.

1185: The solid line shows the median color (2.71) of the diffuse emission, 

1186: and the arrows at left indicate the colors reported for 6 other 

1187: regions of the Galactic plane by Flagey et al. (2006). 

1188: The cluster of points shaded in green indicate those found in the shell around

1189: the Arches Cluster (see Fig. \ref{red_regions}), and the cluster of points shaded in magenta

1190: indicate those found in the more compact regions at higher latitudes 

1191: (see Fig. \ref{red_regions2}). The large dispersion in the 

1192: color at $I_{ISM}(8\ \micron) < 35$ MJy 

1193: sr$^{-1}$ is from faint locations at $|b| > 0.5$ where errors in the 

1194: subtraction of stellar and zodiacal backgrounds can have large effects on the colors.

1195: \label{fig_cmd34}}

1196: \end{figure}

1197: 

1198: \begin{figure}

1199: \centering

1200: \begin{tabular}{ccc}

1201: \includegraphics[width=2in]{f6a.eps}&

1202: \includegraphics[width=2in]{f6b.eps}&

1203: \includegraphics[width=2in]{f6c.eps}\\

1204: \includegraphics[width=2in]{f6d.eps}&

1205: \includegraphics[width=2in]{f6e.eps}&

1206: \includegraphics[width=2in]{f6f.eps}\\

1207: \includegraphics[width=2in]{f6g.eps}&

1208: \includegraphics[width=2in]{f6h.eps}&

1209: \includegraphics[width=2in]{f6i.eps}

1210: \end{tabular}

1211: \caption{Unusually red regions at low latitudes. The Arches Cluster region (top row),

1212: Quintuplet Cluster region (middle row) and Sgr B1 region (bottom row). 

1213: Left column shows 4.5, 5.8, 8 $\micron$ emission

1214: logarithmically scaled as blue, green, and red respectively. The center column shows 

1215: $I_{ISM}(8\micron)/I_{ISM}(5.8\micron)$ linearly scaled in the range [2.0, 4.0]. The 

1216: right column shows $I(4.5\micron)/I(3.6\micron)$ linearly 

1217: scaled in the range [0.5, 1.5]. 

1218: (In the middle and right columns, lighter shades indicate redder colors.)

1219: All these regions contains extended emission that, 

1220: despite being relatively faint, is very red in $I_{ISM}(8\micron)/I_{ISM}(5.8\micron)$ 

1221: colors but not in $I(4.5\micron)/I(3.6\micron)$ colors. 

1222: \label{red_regions}} 

1223: \end{figure}

1224: 

1225: \begin{figure}

1226: \centering

1227: \begin{tabular}{ccc}

1228: \includegraphics[width=2in]{f7a.eps}&

1229: \includegraphics[width=2in]{f7b.eps}&

1230: \includegraphics[width=2in]{f7c.eps}\\

1231: \includegraphics[width=2in]{f7d.eps}&

1232: \includegraphics[width=2in]{f7e.eps}&

1233: \includegraphics[width=2in]{f7f.eps}\\

1234: \end{tabular}

1235: \caption{Unusually red regions at high latitudes. IRAS 17456--2850 (top row) 

1236: and IRAS 17425--2920 (bottom row). Left column shows 4.5, 5.8, 8 $\micron$ emission

1237: logarithmically scaled as blue, green, and red respectively. The center column shows 

1238: $I_{ISM}(8\micron)/I_{ISM}(5.8\micron)$ linearly scaled in the range [2.0, 4.0]. 

1239: The right column shows $I(4.5\micron)/I(3.6\micron)$ linearly 

1240: scaled in the range [0.5, 1.5]. 

1241: (In the middle and right columns, lighter shades indicate redder colors.)

1242: Both these IRAS sources are faint and extended, but 

1243: very red in $I_{ISM}(8\micron)/I_{ISM}(5.8\micron)$ colors.

1244: \label{red_regions2}} 

1245: \end{figure}

1246: 

1247: \begin{figure}

1248: \plottwo{f8a.eps}{f8b.eps}

1249: \caption{The left panel shows 3.6 (blue), 4.5 (green), and 5.8 (red) $\micron$ emission

1250: centered on 2MASS 17482532-2911590. 

1251: There is a slightly extended diffuse feature just $\sim10"$

1252: south of the point source. This extended emission is relatively strong at 4.5 $\micron$ 

1253: as indicated by the greenish color, and a high $I(4.5\micron)/I(3.6\micron)$ ratio

1254: (shown in the right panel on a linear range of [0,3]). This color is often an indication

1255: of molecular emission in outflows or Herbig-Haro objects associated with very young stars.

1256: \label{fig_jet}} 

1257: \end{figure}

1258: 

1259: \begin{figure}

1260: \plotone{f9.eps}

1261: \caption{The colored curves show the emission spectra for a selected dust grain

1262: model (``BARE-GR-FG'' from Zubko et al. 2004) subjected

1263: to radiation fields of 

1264: $\chi = 10^{[0,1,2,3,4,4.5,5,5.5,6,7]}$ times the local ISRF.

1265: The black lines indicate the filter responses (arbitrarily normalized) 

1266: for the 4 IRAC channels, and for the MIPS 24 and 70 $\micron$ channels.

1267: The PAH spectrum is unchanged with radiation field strength, 

1268: and hence the IRAC flux ratios remain unaltered until the larger, 

1269: roughly equilibrium temperature grains get so warm that they encroach 

1270: on the 8 $\micron$ IRAC band.

1271: Note that the continuum, not PAH emission bands, 

1272: dominates the 3.6 and 4.5 $\micron$ IRAC filters. The 3.3 $\micron$ PAH feature

1273: contributes $<30\%$ of the total flux in the IRAC 3.6 $\micron$ channel for these models.

1274: \label{fig_ZDAmodels}} 

1275: \end{figure}

1276: 

1277: \begin{figure}

1278: \plottwo{f10a.eps}{f10b.eps}

1279: \caption{(left) This figure quantifies the $I(8\ \micron) /I(5.8\ \micron)$ ratio

1280: for the dust models shown in Figure \ref{fig_ZDAmodels}. 

1281: The dashed line is the median ratio over the

1282: whole GC region.

1283: (right) This plot shows the corresponding the $I(24\ \micron) /I(8\ \micron)$ ratio. 

1284: Involving longer wavelengths, this color is more sensitive to changes 

1285: in the cooler, larger dust grains and lower values of the radiation

1286: field $\chi$.

1287: \label{fig_ZDA_ratios}} 

1288: \end{figure}

1289: 

1290: \begin{figure}

1291: \plottwo{f11a.eps}{f11b.eps}

1292: ~~\\

1293: \plottwo{f11c.eps}{f11d.eps}

1294: \caption{(a-d) Images of the offsets ($F^p$) derived for each AOR in a time ordered

1295: sequence. These show the diffuse stray light artifacts that remained in 

1296: the BCD, and the accumulation of long--term latent images at 8 $\micron$. 

1297: The last three AORs (0n, 0s, 1w) were separated from the first by a downlink and 

1298: anneal, which cleared previous long term latent images. The Sgr A region creates 

1299: strong latent images during AOR 0s, which then remain throughout AOR 1w.

1300: Display ranges are [-2,2], [-2,2], [-5,5], [-4,4] MJy sr$^{-1}$ for 3.6 

1301: through 8 $\micron$ respectively.

1302: \label{fig_fp}

1303: }

1304: \end{figure}

1305: 

1306: \begin{figure}

1307: \plotone{f12.eps}

1308: \caption{Ratios of standard post-BCD mosaics to our self--calibrated BCD mosaics 

1309: for one AOR strip of the survey (AOR key = 13371648) 

1310: at 3.6 $\micron$ (left) and 4.5 $\micron$ (right). Display range = [0.9,1.2].

1311: The light horizontal bands indicate the regions that are affected by the 

1312: diffuse stray light in the post--BCD mosaics. 

1313: These artifacts are present in the ratio of 3.6/4.5 $\micron$ post-BCD

1314: mosaics, but not in ratios of our self--calibrated mosaics.

1315: The dark vertical lines illustrate the changes due to the application of 

1316: the column pull--down correction.\label{fig_rat12}}

1317: \end{figure}

1318: 

1319: \begin{figure}

1320: \plotone{f13.eps}

1321: \caption{Variable offsets per frame ($F^q$) for each channel are plotted 

1322: in a time ordered sequence. The 5.8 and 8 $\micron$ channel offsets are relatively 

1323: large and well--correlated with the local sky intensity, as revealed by

1324: mapping the offsets (Figure \ref{fig_fqmap}).\label{fig_fqplot}}

1325: \end{figure}

1326: 

1327: \begin{figure}

1328: \plotone{f14.eps}

1329: \caption{Variable offsets per frame ($F^q$) for each channel are mosaicked into images 

1330: in Galactic coordinates. The linear grey scale ranges are the same as shown in 

1331: Figure \ref{fig_fqplot}. 

1332: The 5.8 and 8 $\micron$ channel offsets are relatively large and well--correlated 

1333: with the local sky intensity, and thus these maps resemble very low resolution images

1334: of the Galactic center region.\label{fig_fqmap}}

1335: \end{figure}

1336: 

1337: \begin{figure}

1338: \plotone{f15.eps}

1339: \caption{Ratio of the standard post-BCD mosaic to our self--calibrated BCD mosaic 

1340: for one AOR strip of the survey (AOR key = 13368320) 

1341: at 8 $\micron$. Display range = [0.8,1.3] MJy sr$^{-1}$.

1342: This ratio highlights the corrections made by the fixed 

1343: and variable offsets derived via self--calibration. Some of 

1344: the sharpest frame boundaries evident here (in the expanded view) are visible in 

1345: the standard post--BCD mosaics, but not in the self--calibrated mosaics. The bright region

1346: around Sgr B1 is at the center of this $0\fdg25 \times 1\fdg75$ strip. 

1347: The lack of banding correction in the post-BCD mosaic causes the 

1348: horizontal streaks at the locations of bright stars.

1349: \label{fig_rat34}}

1350: \end{figure}

1351: 

1352: \end{document}