1: \documentclass[11pt,twoside]{article}
2:
3: %%% PREAMBLE MATTER
4: \usepackage[dvips]{graphicx}
5: \usepackage{asp2006}
6: \usepackage{epsf}
7: \usepackage{psfig}
8: \usepackage{lscape}
9:
10: \markboth{Busso and Moehler}{The UV Spectrum of the Galactic Bulge} %%% Fill in authors' names and short running title
11:
12: \pagestyle{myheadings}
13: \setcounter{equation}{0}
14: \setcounter{figure}{0}
15: \setcounter{footnote}{0}
16: \setcounter{section}{0}
17: \setcounter{table}{0}
18:
19: %%% MAIN PART OF DOCUMENT
20:
21: \begin{document}
22:
23: \vspace{1.5cm}
24:
25: {\flushleft To be published in the Proceedings of the Third Meeting on Hot Subdwarfs and Related Objects, Bamberg 2007}
26:
27: \title{The UV Spectrum of the Galactic Bulge}
28: \author{G.~Busso}
29: \affil{INAF-Osservatorio di Teramo, Via M.~Maggini s.n.c., 64100 Teramo, Italy}
30: \author{S.~Moehler}
31: \affil{European Southern Observatory, Karl-Schwarzschild-Strasse 2, 85748 Garching bei M\"unchen, Germany}
32:
33:
34: \begin{abstract} %%% Abstract to run on from here.
35: The UV excess shown by elliptical galaxies in their spectra is
36: believed to be caused by evolved low-mass stars, in particular sdB
37: stars. The stellar system most similar to the ellipticals for
38: age and metallicity, in which it is possible to resolve these stars,
39: is the bulge of our Galaxy. sdB star candidates were observed in the
40: color magnitude diagram of a bulge region by Zoccali et al.\ (2003).
41: The follow-up spectroscopic analysis of these stars confirmed that
42: most of these stars are bulge sdBs, while some candidates turned out
43: to be disk sdBs or cool stars. Both spectroscopic and photometric
44: data and a spectral library are used to construct the integrated
45: spectrum of the observed bulge region from the UV to the
46: optical: the stars in the color magnitude diagram are associated to
47: the library spectra, on the basis of their evolutionary status and
48: temperature. The total integrated spectrum is obtained as the sum
49: of the spectra associated to the color magnitude diagram.
50: The comparison of the obtained integrated spectrum with old single stellar
51: population synthetic spectra calculated by Bruzual \& Charlot~(2003)
52: agrees with age and metallicity of the bulge found by previous work. The bulge integrated
53: spectrum shows only a very weak UV excess, but a too strict selection
54: of the sample of the sdB star candidates in the
55: color magnitude diagram and the exclusion of post-Asymptotic Giant Branch stars could
56: have influenced the result.
57: \end{abstract}
58:
59:
60: \section{Introduction}\label{sec:intro}
61:
62: The UV excess that elliptical galaxies and bulge of spiral galaxies
63: show in their spectra at $\lambda$ shorter than 2300~\AA~ was one of
64: the most puzzling discoveries in the last 30 years, since these
65: stellar systems are believed to be old and metal rich, without
66: young and massive stars emitting most of their flux at short
67: wavelength. It is now widely accepted that this UV emission is caused
68: by evolved low mass stars, in particular Extreme Horizontal Branch
69: stars (EHB), called also sdB stars from their spectral classification.
70: These stars are faint in the optical wavelength range and with the current instrumentation
71: it is impossible to resolve them in the nearest
72: galaxies.
73: The stellar system most similar
74: to the ellipticals for age and metallicity in which it is possible to
75: resolve sdB stars is the bulge of our Galaxy. A sample of sdBs star
76: candidates was observed in the Galactic bulge by Zoccali et
77: al.~(2003)\nocite{zoccali03}, by means of $V$ and $I$ photometry
78: of the region MW05 from the ESO Imaging Survey (EIS\footnote{\tt
79: http://www.eso.org/science/eis/}, the observations were taken with the Wide Field Imager, WFI@2.2m).
80: These stars could be either
81: highly reddened sdBs or cooler stars affected by lower
82: reddening. A follow-up spectroscopic analysis of these stars has been
83: necessary and observations at the Very Large Telescope (VLT)
84: telescope were obtained. The data reduction and the comparison of the
85: obtained spectra with models of hot evolved stars confirmed indeed
86: that most of these stars are bulge sdBs, while some candidates turned
87: out to be disk sdBs or cool stars (for more details, see Busso et
88: al.~2005\nocite{busso05}). To be sure that the observed bulge region
89: was not peculiar, other bulge fields were searched for sdB
90: candidates: EIS photometric data of the bulge fields MW07 and MW08
91: were reduced and analyzed, finding that sdB star candidates are
92: present also in these fields.
93:
94: This work presents the procedure adopted (following the recipe as in
95: Santos et al.~1995) to construct the integrated spectrum of the bulge
96: region MW05.
97:
98: \section{Correction for Reddening and for other Contaminating Stellar
99: Populations}
100: In order to construct the integrated spectrum of the bulge, it is
101: necessary to correct for the extinction caused by the interstellar
102: medium and to take in account that along the line of sight we are
103: observing not only the bulge but also the Galactic disk. Moreover also
104: a globular cluster (NGC~6558) is present in the observed field.
105:
106: To correct for reddening the Schlegel et al.~(1998)\nocite{schlegel98}
107: dust maps were used. Their resolution is roughly 5 arcmin, that is
108: about half the size of a WFI chip:
109: the reddening was calculated then for the 16 regions resulting by
110: splitting each of the 8 WFI chips in half, correcting in this
111: way also for differential reddening.
112:
113: If two stellar populations have different kinematics, the more precise
114: method to distinguish them is to compare proper motions of the stars,
115: which can be measured only by comparing the positions of the same
116: stars at two different epochs. In this case though, the observations
117: were taken in one epoch only. To subtract the
118: contaminating stars of the globular cluster NGC~6558 in the MW05
119: bulge field, a region centered on the center of the globular cluster,
120: with a radius equal to the tidal radius of the cluster, was
121: considered. To validate this approach, also radial star counts on the
122: image were taken in account: in the external regions, roughly beyond
123: 2000 pixels ($\sim$8 arcmin) from the center, the star counts start to flatten,
124: indicating that there is no significant contribution from the
125: globular cluster beyond that range. Only stars in this external
126: region were then taken into account.
127:
128: To decontaminate the bulge color magnitude diagram (CMD) from the
129: foreground disk population, a statistical approach was adopted, using
130: synthetic CMDs of the disk in the direction of the observed bulge
131: field. The disk simulations (one for each WFI chip) were provided by
132: S.~Ragaini (priv. comm., PhD Thesis at the University of
133: Padua, for a detailed description of the simulations, see Vallenari et
134: al.~2000\nocite{vallenari00}, 2006\nocite{vallenari06}).
135:
136: For each disk star in the disk CMD (Fig.~\ref{fig:disk_sub}, top right
137: panel) the closest star in the bulge CMD was picked up and
138: subtracted. The ``photometric'' distance between two stars in the CMD
139: was defined as:
140: \begin{eqnarray*}d=\sqrt{[7\times\Delta(V\!-\!I)]^2 + \Delta I^2}. \end{eqnarray*}
141: and the bulge star with the smallest distance from the disk star was
142: subtracted. The resulting, cleaned CMD of the bulge is shown in the
143: left bottom panel of Fig.~\ref{fig:disk_sub}, while the CMD of the
144: stars statistically removed from the bulge CMD is shown in the right
145: bottom panel.
146:
147: %%%%%%%%%% FIG 1 %%%%%%%%%%%%%%%%%%%
148: \begin{figure}[htbp]
149: \begin{center}
150: \includegraphics[width=3.5in,height=3.5in]{fig1bw.eps}
151: \caption{Example of the statistical subtraction of the simulated disk
152: field from the observed bulge field MW05, in this case for the WFI
153: chip \#1. Top left panel: observed bulge field CMD; top right panel:
154: simulated disk field CMD, where also the number of simulated disk
155: stars not present in the empirical bulge CMD is labeled;
156: bottom left panel: bulge CMD after the disk subtraction; bottom
157: right panel: subtracted disk CMD}
158: \label{fig:disk_sub}
159: \end{center}
160: \end{figure}
161: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
162:
163: The remaining bulge stars were combined to obtain the final CMD for
164: the bulge field MW05 shown in Fig.~\ref{fig:mw05final}. In this figure
165: the typical bulge sequences, as the Main Sequence (MS), the
166: red giant branch (RGB), the red horizontal branch (HB) clump (at
167: $V-I\sim$0.6 and $I\sim$14) are evident. The grey
168: dots are stars that remain after the disk subtraction, because the
169: simulations do not manage to reproduce well the photometric errors.
170: The filled grey circles could be Blue Stragglers but previous studies as
171: Kuijken \& Rich 2001 ruled out their presence in the bulge, therefore they are assumed to be caused by a not precise subtraction. The
172: stars marked as filled squares seem to form an extended HB sequence. Blue HB
173: stars have been previously observed in the bulge by Peterson et
174: al.~(2001)\nocite{peterson01} and moreover an extended HB is not typical
175: for the disk population. The black triangles are probably a combination of
176: real post-HB stars and remaining disk blue MS stars. Since there is
177: no way to disentangle the two populations in this case and since
178: post-HB stars should play only a marginal role in the UV excess (Brown
179: et al.~1997\nocite{brown97}), they were left out from the construction
180: of the integrated spectrum. The stars marked with asterisks represent the sdB star
181: candidates and they are a mixture of real sdB stars and cooler stars
182: with lower reddening. For the construction of the integrated spectrum
183: therefore only the black (MS, RGB, Red HB, Blue HB and Extended HB) stars were considered.
184:
185: %%%%%%%%%% FIG 2 %%%%%%%%%%%%%%%%%%%
186: \begin{figure}[!tbp]
187: \begin{center}
188: \includegraphics[width=3in]{fig2bw.eps}
189: \caption{Final color magnitude diagram for the bulge field MW05. See text for
190: the stars shown with different symbols. Only stars in black were used to construct the integrated spectrum}
191: \label{fig:mw05final}
192: \end{center}
193: \end{figure}
194: %%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%
195:
196: \section{The Bulge integrated spectrum}
197:
198: To finally construct the integrated spectrum of the Galactic
199: bulge from the UV to the optical the method of Santos et al.~(1995) was
200: adopted, using both spectroscopic and photometric data and spectral
201: libraries (Pickles 1998, Lejeune et al.~1997) and the Bamberg archive of
202: optical and UV-spectra of hot subdwarfs
203: (Heber, priv.comm.) to extend the sdB spectra to the UV.
204:
205: To construct the sdB spectrum, first of all it was necessary to decide
206: which of the sdB star candidates in the CMD
207: to take into account. Only stars with a good photometry were selected
208: (error in the color magnitude diagram smaller than 0.1), obtaining 112 sdB star
209: candidates. The disk simulation, while taking into account the
210: presence of Horizontal Branch stars, does not include sdB stars which
211: therefore remain in the corrected disk CMD. However the analysis of the spectra (see
212: Sec.~\ref{sec:intro}) showed that only $\sim$80\% of the
213: candidates are really sdB stars and not all these stars have turned
214: out to belong to the Galactic bulge. The percentages found from the
215: spectroscopic analysis were applied to the total number of sdB
216: stars candidates, since there is no reason to assume that peculiar
217: stars were picked up during the selection of the spectroscopic
218: targets: of 116 candidates therefore, only 55 \% have been considered to be true
219: sdB stars of the bulge (Busso et al.\ 2005), the remaining
220: ones being cool stars or hot stars belonging to the disk or
221: unclear membership, thus narrowing the sdBs sample to 62 candidates.
222:
223: %%%%%%%%%% FIG 3 %%%%%%%%%%%%%%%%%%%
224: \begin{figure}[htbp]
225: \begin{center}
226: \includegraphics[width=3.5in,height=3.5in]{fig3bw.eps}
227: \caption{Association between library (in this case Pickles 1998) spectra and CMD: the spectra library
228: points are marked with different symbols (grey filled squares: dwarfs(V); grey filled triangles:
229: subgiant(IV)); grey and black filled pentagons: giants (III); black filled circles : horizontal branch stars; empty star (sdB stars). For each kind of stars, each point has different
230: temperature. Every CMD star has been associated to the library
231: spectrum corresponding to the box where the star belongs.}
232: \label{fig:box_on_cmd}
233: \end{center}
234: \end{figure}
235: %%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%
236:
237:
238: In order to associate the appropriate spectrum to the other CMD stars,
239: the Pickles (1998) spectral library was plotted on the cleaned and
240: reddening corrected CMD of the bulge region, as shown in
241: Fig.~\ref{fig:box_on_cmd}. In this figure, the library points are
242: marked with different symbols, depending on the luminosity of class
243: (filled squares: dwarfs~(V); filled triangles: subgiant~(IV)); filled pentagons: giants~(III)) while HB stars are marked with filled circles and sdB stars with an empty star. Each
244: different point corresponds to a library spectrum with a different
245: temperature. The boxes on the CMD were chosen so that every box
246: contains at least one library point. For each box the
247: following parameters were calculated: the mean absolute
248: magnitude $<M_{V}>_j$
249: \footnote{assuming an average distance of 8.15 Kpc, calculated as the most probable distance.}
250: of the stars in the $j$-th box; the corresponding
251: library spectrum ``magnitude'' $m_{V_j}$ obtained from the
252: convolution with the $V$ filter; the weighting factor $C_j$ of the stars
253: inside a certain box contributing to the total spectrum ($n_j$ is the
254: number of stars inside the $j^{th}$ box):
255: \begin{center}
256: $C_j = n_j 10^{-0.4(<M_V>_j - m_{Vj})}$
257: \end{center}
258: Finally, the total integrated spectrum of this bulge region was calculated as sum of
259: all spectra ($f_j$) associated to the CMD boxes, taking in account their weights $C_j$:
260: \begin{center}
261: $ \mathcal{F}_{TOT} = \Sigma^N_{n=1} C_{j} f_{j}$
262: \end{center}
263:
264: The same procedure was applied also using the BaSeL (Lejeune et al.\
265: 1997) syntethic spectral library, where the spectra at solar
266: metallicity, with temperature and luminosity class corresponding to
267: those of the Pickles (1998) stars were selected. The result is shown
268: in Fig.~\ref{fig:bulge_spek}, where also the comparison with the
269: integrated spectra (from Bruzual \& Charlot~2003) of two simple
270: stellar populations (SSPs) with an age of 11 Gyr and metallicity Z=0.008 and
271: Z=0.02 is shown.
272:
273: The two integrated spectra obtained using the Pickles and BaSeL
274: library are very similar, both in the optical and UV range (see
275: bottom panel). Neither the SSP
276: at Z=0.008 ([Fe/H]$\sim -$0.5) nor the one at Z=0.02
277: ([Fe/H]$\sim-$0.09) agree perfectly with the integrated spectra,
278: particularly in the region between 3300 and 4000~\AA, where the two
279: synthetic spectra bracket the integrated one. Probably a population
280: with an intermediate metallicity would fit better. It is necessary to
281: keep in mind though that the bulge is not a single stellar population
282: with only one value for the metallicity, but it has a metallicity
283: distribution from metal-poor (down to [Fe/H]$\sim -$2) to metal-rich (up
284: to [Fe/H]$\sim$0.5), with a peak at [Fe/H]$\sim -$0.2 (Zoccali et
285: al.~2003, Rich \& Origlia~2005),
286: intermediate between the metallicity of the two single stellar
287: populations adopted. The expected age for the bulge is 10-13 Gyr
288: while the adopted single stellar populations are 11 Gyr old. We note
289: in passing that the fact that the spectrum characteristic of a single
290: stellar population, hence formed in an "instantaneous" burst, is in
291: agreement with the recent work of Zoccali et al.~(2006), that, by
292: means of $\alpha$-element analysis, found that probably the bulge
293: formed very quickly.
294:
295: Integrated and synthetic spectra do not match in the UV instead,
296: with the synthetic spectra showing a larger UV flux. This could be
297: explained with two reasons. Firstly, in their stellar population
298: synthesis, Bruzual \& Charlot~(2003) take into account also post-HB
299: and post-AGB, which have a strong flux in the UV range. These stars
300: were not accounted for here since it was not possible to
301: disentangle them from the disk population. Secondly, probably the
302: selection of sdBs stars was too strict.
303: On the base of the
304: spectroscopic analysis, were selected, as bulge sdBs, on the color magnitude diagram only 55\% of the candidates to construct the
305: integrated spectrum. Taking into account also the stars with unknown
306: membership the percentage would raise to 65\%, increasing the UV flux
307: by about 20\%, which is insufficient to achieve agreement with
308: the SSP predictions.
309:
310: Thus the Galactic bulge probably shows only a very weak UV excess,
311: in contrast to what is observed in the bulge of M31. This is
312: consistent with UV observations of the closest extragalactic
313: systems, which show that this UV excess can vary strongly from object
314: to object (Rich et al.~2005). In addition, the same observations indicate
315: also that the UV excess shows no correlation with
316: physical parameters as metallicity or velocity dispersion. A
317: correlation with age appears more probable but it would be visible
318: only at high redshift (Brown et al.~2003).
319:
320: %%%%%%%%%% FIG 4 %%%%%%%%%%%%%%%%%%%
321: \begin{figure}[htbp]
322: \begin{center}
323: %\includegraphics[width=3in,height=2in]{fig4abw.eps}
324: %\includegraphics[width=3in,height=2in]{fig4bbw.eps}
325: \includegraphics[width=80mm,height=59mm]{fig4abw.eps}
326: \includegraphics[width=80mm,height=59mm]{fig4bbw.eps}
327: \caption{Comparison between integrated spectra and single stellar
328: population SED. Top panel: The integrated bulge spectrum obtained with
329: the Pickles 1998 and BaSeL (Lejeune et al.~1997) library are shown with a solid and dashed black line respectively. The Bruzual \& Charlot~(2003) spectra
330: for a single stellar population 11~Gyr old and with metallicity z=0.008
331: and 0.02 are shown in grey (see label). To allow the
332: comparison the spectra were scaled at $\lambda=$5870~\AA. Bottom
333: panel: as top panel, but in the UV range.}
334: \label{fig:bulge_spek}
335: \end{center}
336: \end{figure}
337: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
338:
339: To verify the procedure, we constructed the integrated spectrum also
340: for the Galactic Globular Clusters NGC~6388 and NGC~6441, which show
341: an unexpected population of sdB stars. The procedure was the
342: same as for the bulge fields, with the same choice of not including
343: the post-HB and post-AGB stars. The integrated spectra were compared
344: with the observed integrated spectrum of the two clusters in the
345: optical (Schiavon et al.~2005) and in the UV (Rich et al.~1993). Also
346: in these cases, while in the optical the calculated and observed
347: integrated spectra match, in the UV there seems to be no
348: agreement. This could be a hint that sdB stars are not the
349: dominant cause for the UV excess in these systems, while a more
350: important role could be played by post-HB and post-AGB stars.
351:
352: In conclusion, the bulge of our Galaxy probably shows only a weak UV
353: excess. We note that this result is neither in contradiction with
354: nearby observations, as in the case of M31, nor with observations of
355: more distant galaxies. In fact the UV excess can vary considerably
356: from object to object and shows no dependence on physical parameters
357: like metallicity and velocity dispersion, as explained by Rich et
358: al.\ (2005).
359:
360: \acknowledgments %%%text of acknowledgments runs on after this command.
361: We warmly thanks Manuela Zoccali for providing the photometric data, Uli Heber for the Bamberg spectroscopic archive of sdBs and Sukyoung Yi for the theoretical tracks.
362: GB gratefully acknowledges support from the Deutsche Forschungsgemeinschaft through grant Mo 602/8.
363: \begin{thebibliography}{}
364:
365: \bibitem[]{brown97} Brown, T.~M., Ferguson, H.~C., Davidsen, A.~F., \& Dorman, B., 1997, ApJ, 482, 685
366: \bibitem[]{brown03} Brown, T.~M., Ferguson, H.~C., Smith, E., Bowers, C.~W., Kimble, R.~A., Renzini, A., \& Rich, R.~M., 2003, ApJ, 584, L69
367: \bibitem[]{bruzual03} Bruzual, G., \& Charlot, S., 2003, MNRAS, 344, 1000
368: \bibitem[Busso et al.~ 2005]{busso05} Busso, G., Moehler, S., Zoccali, M., Heber, U., \& Yi, S. K., 2005, ApJ, 633, L29
369: \bibitem[]{kuijken01} Kuijken, K., \& Rich, R.~M., 2001, AAS, 199, 9113
370: \bibitem[]{lejeune97} Lejeune, Th., Cuisinier, F., \& Buser, R., 1997, A\&AS, 125, 229
371: \bibitem[]{peterson01} Peterson, R.~C., Terndrup, D.~M., Sadler, E.~M., \& Walker, A.~R., 2001 ApJ, 547, 240
372: \bibitem[]{pickles98} Pickles, A.~J., 1998, PASP, 110, 863
373: \bibitem[]{rich93} Rich, R.~M., Minniti, D., \& Liebert, J.~W, 1993, ASPC, 50, 231
374: \bibitem[]{rich05} Rich, R.~M., Salim, S., Brinchmann, J., Charlot, S., and 24 coauthors, 2005, ApJ, 619, L107
375: \bibitem[]{rich05b} Rich, R.~M., \& Origlia, L., 2005, ApJ, 634, 1293
376: \bibitem[Santos et al.~1995]{santos95} Santos, J.~F.~C.~Jr., Bica, E., Dottori, H., Ortolani, S., \& Barbuy, B., 1995, A\&A, 303, 753
377: \bibitem[]{schiavon05} Schiavon, R.~P., Rose, J.~A., Courteau, S., \& MacArthur, L.~A., 2005, ApJS, 160, 163
378: \bibitem[Schlegel et al.~1998]{schlegel98} Schlegel, D.~J., Finkbeiner, \& D.~P., Davis, M., 1998, ApJ, 500, 525
379: \bibitem[Vallenari et al.~2000]{vallenari00} Vallenari, A., Bertelli, G., \& Schmidtobreick, L., 2000, A\&A, 361, 73
380: \bibitem[Vallenari et al.~2006]{vallenari06} Vallenari, A., Pasetto, S., Bertelli, G., Chiosi, C., Spagna, A., \& Lattanzi, M., 2006, A\&A, 451, 125
381: \bibitem[Zoccali et al.~2003]{zoccali03} Zoccali, M., Renzini, A., Ortolani, S., Greggio, L., Saviane, I., Cassisi, S., Rejkuba, M., Barbuy, B., Rich, R. M., \& Bica, E., 2003, A\&A, 399, 931
382: \bibitem[]{zoccali06} Zoccali, M., Lecureur, A., Barbuy, B., Hill, V., Renzini, A., Minniti, D., Momany, Y., Gomez, A., \& Ortolani, S., 2006, A\&A, 457, L1
383:
384: \end{thebibliography}
385:
386: \end{document}
387: