1:
2:
3: %\documentclass[12pt,preprint]{aastex}
4: \documentclass{emulateapj}
5: %% preprint2 produces a double-column, single-spaced document:
6:
7: %\documentclass[preprint2]{aastex}
8: %\documentclass[preprint2]{aastex}
9:
10: \newcommand{\vdag}{(v)^\dagger}
11: \newcommand{\myemail}{skywalker@galaxy.far.far.away}
12:
13: \shorttitle{Evidence of a Metal Rich Galactic Bar}
14: \shortauthors{Soto et al.}
15:
16: \begin{document}
17:
18: \title{Evidence of a Metal Rich Galactic Bar from the Vertex Deviation of the
19: Velocity Ellipsoid }
20:
21: \author{M. Soto }
22: %\altaffilmark{1}}
23: \affil{Leiden Observatory, Leiden University, P.O. Box 9513, 2300RA, Leiden, The Netherlands}
24: \email{soto@strw.leidenuniv.nl}
25:
26: \author{R.M. Rich }
27: %\altaffilmark{2}}
28: \affil{Department of Physics and Astronomy, UCLA, Los Angeles, CA, 90095-1547}
29: \email{rmr@astro.ucla.edu}
30:
31: \author{K. Kuijken }
32: %\altaffilmark{1}}
33: \affil{Leiden Observatory, Leiden University, P.O. Box 9513, 2300RA, Leiden, The Netherlands}
34: \email{kuijken@strw.leidenuniv.nl}
35:
36:
37: \begin{abstract}
38: We combine radial velocities, proper motions, and low resolution abundances
39: for a sample of $315$ K and M giants in the Baade's Window $(l,b)=(0.9^\circ,-4^\circ)$
40: Galactic bulge field. The velocity ellipsoid of stars with $\rm [Fe/H]>-0.5$ dex
41: shows a vertex deviation in the plot of
42: radial versus transverse velocity, consistent with that expected
43: from a population with orbits supporting a bar. We demonstrate that
44: the significance of this vertex deviation using non-parametric
45: rank correlation statistic is $>99$\%. The velocity
46: ellipsoid for the metal poor ($\rm [FeH]<-0.5$) part of the population shows no
47: vertex deviation and is consistent with an isotropic, oblate rotating
48: population. We find no evidence for kinematic subgroups, but there
49: is a mild tendency for the vertical velocity dispersion $\sigma_b$ to
50: decrease with increasing metallicity.
51:
52:
53: \end{abstract}
54:
55:
56: \keywords{Galaxy: bulge --- Galaxy: kinematics and dynamics -- Galaxy: abundances }
57:
58: \section{Introduction}
59:
60: The galactic bulge is the nearest example of a bulge/spheroidal
61: population. A factor of 100 closer than M31, the proximity of the
62: bulge not only permits detailed measurement of stellar abundances and
63: radial velocities, but also of stellar proper motions. The first such
64: study undertaken (at the suggestion RMR) was that of \citet{sp92}.
65: The proper motions were measured from plates obtained at the 200-inch
66: prime focus in 1950 and 1983; these data produced measurable proper
67: motions on the order of $0.5 \rm \ mas yr^{-1}$.
68:
69: The \citet{sp92} study was the source of an input catalog for a
70: subsequent study of bulge radial velocity and low resolution
71: abundances \citep{terndrup95,sadler96}. An early study of the
72: abundances and kinematics of these stars, using published abundances
73: from \citet{rich88} and radial velocities from \citet{rich90} was
74: undertaken by \citet{zhao94}. That study suffered from having a
75: relatively small sample size, and the lack of a definitive bulge iron
76: abundance scale; nonetheless the most metal rich subset of stars
77: exhibited a vertex deviation of the velocity ellipsoid in the plot of
78: radial versus transverse velocity. Modeling by \citet{zhao94} showed
79: that such a vertex deviation, or rotation of the major axis of the
80: velocity ellipsoid, is a feature expected from the long axis orbits
81: that would support a bar structure. Such a bar had been claimed by
82: \citet{BlitzSpergel} based on asymmetries in surface photometry of the
83: bulge region, and by \citet{binney} from the kinematics of molecular gas.
84:
85: The possibility of a dominant stellar bar population was strengthened
86: by the spectacular results from the {\sl COBE} satellite
87: \citep{dwek94}. All-sky maps of the Milky Way showed an asymmetry
88: that was consistent with a bar, with major axis in the first
89: quandrant. Many subsequent studies confirm the presence of a bar
90: structure based on deprojecting the spatial distribution of the
91: stellar population \citep{bin97} and the Schwarzschild method has been
92: used to build a self consistent rapidly rotating bar \citep{zhao96,
93: hafner00}.
94:
95: Here we combine the proper motions of \citet{sp92}, the abundances of
96: \citet{terndrup95}, and the radial velocities of \citet{sadler96} to
97: explore the correlation of abundances and kinematics in the bulge
98: stellar population. Surveys now in progress \citep{sumi04} have
99: achieved total numbers of order $5\times 10^6$ stars. However, the
100: sample of 315 stars on which we report here is at present the largest
101: that has proper motions, radial velocities, and spectroscopically
102: estimated abundances.
103:
104: In exploring correlations of abundances and kinematics, our aim is to
105: constrain models for the origin of the bulge and bar. When the
106: composition of bulge stars \citep{mr94,fmr06b} is considered, it can
107: be argued that the bulge formed early and rapidly \citep{matt99,
108: ferr03}. \citet{kr02} isolated the old bulge population by proper
109: motion, and \citet{zoc03} demonstrated a comparison with a 13 Gyr old
110: main sequence turnoff population; these studies find little room for a
111: detectable trace young population in the bulge. However, there are
112: complicating issues in the chemical evolution. Mg is found to be elevated in
113: bulge stars, while O is only mildly so \citep{mr94,fmr06b,lecu06} and the
114: "explosive alphas" Ca, Si, and Ti are also less enhanced than Mg,
115: tracking the thick disk trends with [Fe/H]. Finally, the bar poses a
116: problem for early formation and enrichment, as the longevity of bars
117: is an issue. If the metals in the bulge were built early and rapidly,
118: how and when did the bar appear and why has it survived?
119:
120: Correlating abundances and kinematics gives another viewpoint on
121: chemical evolution. Even if the abundance distribution appears to fit
122: the closed box model and the population is old, the correlation of
123: abundances and kinematics may reveal distinct kinematic subgroups.
124: Adding two proper motion vectors to the radial velocity increases the
125: power of this analysis. There have been indications of
126: abundance/kinematics correlations in the bulge. \citet{rich90} found
127: that the most metal rich stars have a smaller velocity dispersion.
128: \citet{sp92} found a similar trend in their small sample.
129: \citet{zhao94} found evidence for vertex deviation in the most metal
130: rich stars. In a radial velocity study of a large number of bulge M
131: giants, \citet{scw90} found indications for a smaller velocity
132: dispersion for a metal rich subgroup of stars. \citet{min96} obtained
133: spectroscopy in a bulge field at (l,b)=(8,7) and found differing
134: kinematics (more rotation support and a colder velocity dispersion)
135: for stars with $\rm [Fe/H]>-1$. In this paper, we explore
136: abundance/kinematics correlations using both the radial velocity and
137: proper motion data, and abundances from low resolution spectroscopy.
138: We use the new iron abundance scale of \citet{fmr06a} to recalibrate
139: the \citet{sadler96} low resolution abundances (those being based on
140: the average of the Fe 5270 and 5335 features in the spectra). While
141: the derived abundance for any individual star is uncertain to roughly
142: $+0.3$ dex, the calibration ensures that the kinematic transitions we
143: report here are tied to the best currently available iron abundance
144: scale. The most metal rich stars in \citet{sadler96} have formal
145: metallicities in excess of 1 dex; we do not take those values
146: seriously, as they are transformed from outliers.
147:
148: In this {\it Letter} we describe our kinematic analysis of the sample
149: of 315 red giants.
150:
151:
152: \section{Analysis}
153:
154: In an axisymmetric potential, one of the axes of the velocity
155: dispersion tensor of an equilibrium population must lie in the
156: meridional plane. Any deviation from this situation indicates a
157: non-axisymmetry, or a time variation. We will use the space velocities
158: of our stars to measure the orientation of the axis of the velocity
159: dispersion tensor in the radial-longitudinal velocity plane (known as
160: the {\em vertex deviation $l_v$}) and determine whether it is
161: consistent with zero. The angle $l_v$ is defined as
162: \begin{equation}
163: \tan(2l_v) = \frac{2\sigma_{lr}^2}{|\sigma_r^2 - \sigma_l^2|}
164: \end{equation}
165: where $\sigma_{lr}$, $\sigma_r$ and $\sigma_l$
166: are the covariance and standard deviation of the measured velocity
167: components along the $r$ (line of sight) and $l$ (galactic longitude)
168: directions. We measure the dispersions with an iterative clipping
169: algorithm, by repeatedly rejecting stars that lie outside the ellipse
170: with principal axes of 2.5 times the corresponding measured dispersion.
171: Table~\ref{tab:veldisp} and Fig.~\ref{fig:veldisp} summarize the
172: velocity ellipsoid parameters found, both for the full sample and for
173: metal poor and metal rich subsets. For each subsample we have
174: calculated the Spearman rank correlation coefficient $r_s$ and its
175: significance.
176:
177: \begin{table}
178: \begin{center}
179: \caption{Velocity ellipsoid parameters.\label{tab:veldisp}.
180: }
181: \begin{tabular}{ l c c c c c c c c }
182: \tableline\tableline
183: %Star & Height & $d_{x}$ & $d_{y}$ & $n$ & $\chi^2$ & $R_{maj}$ & $R_{min}$ &
184: %\multicolumn{1}{c}{$P$\tablenotemark{a}} & $P R_{maj}$ & $P R_{min}$ &
185: %\multicolumn{1}{c}{$\Theta$\tablenotemark{b}} \\
186: [Fe/H] & $\sigma_r$ & $\sigma_l$ & $\sigma_b$ & $l_v$ &
187: $r_S$ & N (N$_{\rm rej}$) & Prob $r_S$ \\
188: & \multicolumn{3}{c}{km/sec} & deg & & & \% \\
189: \tableline
190: All & 109 & 94 & 86 & -25 & -0.15 & 315(21) & 0.85 \\
191: %probrs=0.00854614
192: $<$-0.5 & 108 & 85 & 94 & 28 & +0.09 & 77(11) & 39 \\
193: %probrs=0.389356
194: $>$-0.5 & 107 & 96 & 83 & -30 & -0.23 & 238(11) & 0.02 \\
195: %probrs=0.000191179
196:
197: \tableline
198: \end{tabular}
199: \end{center}
200: \end{table}
201:
202:
203: \begin{figure*}
204: %\epsscale{0.3}
205: \includegraphics[width=0.32\hsize]{f1a.eps}
206: \includegraphics[width=0.32\hsize]{f1b.eps}
207: \includegraphics[width=0.32\hsize]{f1c.eps}
208: \caption{Velocity ellipsoids for metal rich and metal poor populations in
209: Baade's Window. An iterative clipping algorithm to reject stars outside the
210: velocity ellipsoid at 2.5 $\sigma$ has been applied.
211: \emph{Left to right}, full sample, stars with $\rm [Fe/H]<-0.5$, and stars
212: with $\rm [Fe/H]>-0.5$. In each panel the vertex deviation $l_v$, the
213: Spearman rank correlation coefficient $r_S$, and the number of stars $N$,
214: is listed.
215: \label{fig:veldisp}}
216: \end{figure*}
217:
218:
219: \begin{figure*}
220: %\epsscale{0.3}
221: \includegraphics[width=0.32\hsize]{f2a.eps}
222: \includegraphics[width=0.32\hsize]{f2b.eps}
223: \includegraphics[width=0.32\hsize]{f2c.eps}
224: \includegraphics[width=\hsize]{f2d.eps}
225: \caption{The kinematics as a function of metallicity. \emph{Top row: Left},
226: Vertex angle $l_v$ vs. $\rm [Fe/H]$ using a running box car of width 0.4
227: dex.
228: \emph{Center}, the corresponding significance of the vertex
229: deviation, using a non-parametric Spearman rank correlation. The
230: combined significance for all stars with $\rm [Fe/H]>-0.5$ is greater
231: than 99\%.
232: \emph{Right}, histogram showing the number of stars in 0.2 dex
233: $\rm [Fe/H]$ bins. \emph{Bottom row}, velocity dispersions in $b$, $l$ and
234: $r$, for the same running metallicity bins. Note the break in the
235: kinematics in $\sigma_b$ at $\rm [Fe/H]=-0.5$.
236: \label{fig:abunkin}}
237: \end{figure*}
238:
239:
240: The most significant vertex deviation is seen in
241: Fig.~\ref{fig:veldisp} for the metal rich sample ($\rm [Fe/H]>-0.5$),
242: whereas metal poor stars show no significant deviation. This result is
243: consistent with the results from \citet{zhao94}, based on
244: $\sim20\%$ of the present sample, and proves the reliability of their results
245: (Zhao et al.\ also saw a vertex deviation in the metal-poor sample, of
246: opposite sign to the metal-rich set: with the addition of the new
247: stars in this study the significance of this deviation has dropped).
248: This difference between metal rich and metal poor stars is a strong
249: stellar-dynamical demonstration of the existence of a stellar bar in
250: the galactic center \citep{binney,zhao94}, and suggests that much of
251: the disk takes part in it.
252:
253: Fig.~\ref{fig:abunkin} shows the velocity dispersion components and
254: vertex deviation as a function of the metallicity, using a running
255: mean over metallicity bins of width 0.4dex. The Figure clearly shows a
256: rather sharp transition near $\rm [Fe/H]=-0.5$: more metal-poor stars
257: show little or no deviation, whereas the more metal-rich stars'
258: velocity ellipsoid is misaligned by about $30^\circ$, with a rank
259: correlation between radial and longitudinal velocity components for
260: these stars that is significant at the 99\% level. A further
261: indication of this change in kinematic properties is the dispersion in
262: the (completely independent) vertical velocity component, $\sigma_b$,
263: which shows a clear break at $\rm [Fe/H]=-0.5$ with the more
264: metal-poor stars' $b$ velocity dispersion being higher.
265:
266: We have also examined the correlations between the $b$ velocity and
267: the other components, but found no significant deviation of the
268: velocity ellipsoids out of the plane.
269:
270: The giant stars thus consist of a metal-rich population with a
271: significant vertex deviation of $30$ degrees, and a metal-poor
272: population that has no significant deviation, and hotter kinematics.
273:
274: The general trend of the bulge to have this vertex deviation is
275: also seen in three minor axis fields with proper motions from
276: HST imaging with WFPC2 (Soto et al. 2007 in prep.).
277: These fields were observed
278: for radial velocities
279: using the integral field spectrograph on VIMOS at VLT.
280: The combination of radial velocities and $l$ proper motions
281: yields an observable vertex deviation in the fields (in samples
282: not segregated by metallicity).
283:
284: \section{Discussion}
285:
286: Our observations support the presence of a striking transition in the
287: kinematics of the bulge, from an apparently isotropic oblate spheroid
288: to a bar, at $\rm [Fe/H]=-0.5$. We have used a non-parametric rank
289: correlation statistic to determine that the significance of the vertex
290: deviation for the metal-rich stars is $>99$\%. There is a
291: corresponding but shallower trend toward lower vertical velocity
292: dispersion $(\sigma_b)$ at higher abundance. It is difficult to
293: imagine how a system with such a kinematic transition could be
294: considered to have evolved in a one-zone chemical evolution scenario.
295:
296: Do abundance trends of any other elements show a transition at $\rm
297: [Fe/H]=-0.5$? Examining the trends given in \citet{fmr06b} and
298: \citet{lecu06}, we find no break in the trends of Mg, O, Ca, Si, Ti,
299: Na, and [Al/Fe] with [Fe/H] that would correspond to $-0.5$ dex.
300: \citet{fmr06b} consider the trend of the explosive alpha elements $\rm
301: \langle Ca+Si+Ti\rangle=\alpha_{exp}$ with [Fe/H]. They find that
302: even to the lower limit of bulge [Fe/H], at $\rm [Fe/H]<-0.5$, the
303: bulge trend of $[\alpha_{exp}/\rm Fe]$ lies above the locus of halo
304: and thick disk stars. Hence, the stars with $\rm [Fe/H]<-0.5$ cannot
305: be considered simply to belong to the halo or thick disk. Similarly,
306: virtually all stars with $\rm [Fe/H]>-0.5$ have strongly enhanced Mg,
307: a strong indication that the chemistry reflects a history of rapid
308: enrichment that extends through the boundary of this kinematic
309: transition.
310:
311: In order to form the bar via a stellar-dynamical instability, which
312: then thickens into a bulge \citep{raha91}, the stars would already
313: have to have been in place. The kinematics thus suggest a scenario in
314: which all stars with $\rm [Fe/H]>-0.5$ stars were formed in a disk
315: {\em before} the bar instability occurred; the more metal-poor stars
316: are then a population unrelated to the disk, bar and bulge, consistent
317: with their hotter kinematics.
318: But the continuity in bulge abundance trends across $\rm [Fe/H]=-0.5$
319: argues against the existence of two entirely separate populations with
320: different chemical evolution histories. Evidently the situation is
321: more complicated than a simple few-component model can describe. It is
322: clearly important to explore the existence of these trends in larger
323: samples and in additional fields in the bulge.
324:
325:
326: \acknowledgements
327:
328: We thank the Lorentz Center in Leiden for an enjoyable workshop during
329: which this paper was completed. RMR acknowledges support from GO-9816
330: from the Space Telescope Science Institute. KK acknowledges travel
331: support from the Leids Kerkhoven-Bosscha Fonds.
332:
333:
334: \newpage
335:
336:
337: \begin{thebibliography}{}
338:
339: \bibitem[Beers and Sommer-Larsen(1995)]{beers95} Beers, T., Sommer-Larsen, J. 1995, \apjs, 96, 175
340:
341: \bibitem[Binney et al.(1991)]{binney} Binney, J., Gerhard, O.E., Stark,
342: A.A., Bally, J., Uchida, K. 1991, \mnras, 252, 210
343:
344: \bibitem[Binney et al.(1997)]{bin97} Binney, J., Gerhard, O., \& Spergel, D.\ 1997, \mnras, 288, 365
345:
346: \bibitem[Blitz \& Spergel(1991)]{BlitzSpergel}
347: Blitz, L. \& Spergel, D.N.\ 1991, \apj, 379, 631
348:
349: \bibitem[Burstein et al.(1984)]{burstein84} Burstein, D., Faber, S.M., Gaskell, C,M., Krumm, N. 1984, \apj, 287, 586
350:
351: \bibitem[Cunha \& Smith(2006)]{cunha06} Cunha, K., Smith, V.V. 2006, preprint astro-ph/0607393
352:
353: \bibitem[Dwek et al.(1994)]{dwek94} Dwek, E., et al. 1994, ApJ, 445, 716
354:
355: \bibitem[Englmaier and Gerhard(1999)]{englmaier99} Englmaier, P., and Gerhard, O. 1999, MNRAS, 304, 512
356:
357: \bibitem[Faber et al.(1985)]{faber85} Faber, S.M., Friel, E.D., Burstein, D., Gaskell, C.M. 1985, \apjs, 57, 711
358:
359: \bibitem[Feltzing et al.(2003)]{feltzing03} Feltzing, S., Bensby, T., \& Lundstr\"om, I., 2003, \aap, 397, L1
360:
361: \bibitem[Ferreras et al.(2003)]{ferr03} Ferreras, I., Wyse, R.~F.~G., \& Silk, J.\ 2003, \mnras, 345, 1381
362:
363: \bibitem[Fulbright et al.(2006a)]{fmr06a} Fulbright, J.~P., McWilliam, A., \& Rich, R.~M.\ 2006a, \apj, 636, 821
364:
365: \bibitem[Fulbright et al.(2006b)]{fmr06b} Fulbright, J.~P., McWilliam, A. \& Rich, R.~M.\ 2006b, astro-ph/0609087
366:
367: \bibitem[Gratton et al.(2000)]{gratton00} Gratton, R.G., Carretta, E., Matteucci, F., \& Sneden, C. 2000, \aap, 358, 671
368:
369: \bibitem[Fuhrmann(1998)]{fuhrmann98} Fuhrmann, K. 1998, \aap, 338, 161
370:
371: \bibitem[H{\"a}fner et al.(2000)]{hafner00} H{\"a}fner, R., Evans, N.~W., Dehnen, W., \& Binney, J.\ 2000, \mnras, 314, 433
372:
373: \bibitem[Kuijken \& Rich(2002)]{kr02} Kuijken, K., \& Rich, R.~M.\ 2002, \aj, 124, 2054
374:
375: \bibitem[Lecureur et al.(2006)]{lecu06}Lecureur,A. et al. (2006), astro-ph/0610346
376:
377: \bibitem[Majewski(1992)]{majewski92} Majewski, S. 1992, \apjs, 78, 87
378:
379: \bibitem[Matteucci et al.(1999)]{matt99}
380: Matteucci, F., Romano, D., \& Molaro, P.\ 1999, \aap, 341, 458
381:
382: \bibitem[McWilliam \& Rich(1994)]{mr94}
383: McWilliam, A., \& Rich, R.~M.\ 1994, \apjs, 91, 749
384:
385: \bibitem[Minniti(1996)]{min96} Minniti, D.\ 1996, \apj, 459, 175
386:
387: \bibitem[Raha et al.(1991)]{raha91}
388: Raha, N., Sellwood, J.~A., James, R.~A., \& Kahn, F.~D.\ 1991, \nat, 352, 411
389:
390: \bibitem[Rich(1988)]{rich88} Rich, R.M. 1988, \aj, 95, 828
391:
392: \bibitem[Rich(1990)]{rich90} Rich, R.M. 1990. \apj, 362, 604
393:
394: \bibitem[Sadler et al.(1996)]{sadler96}
395: Sadler, E.M., Rich, R.M., Terndrup, D.M. 1996, \aj, 112, 171 (Paper II)
396:
397: \bibitem[Sharples et al.(1990)]{scw90}
398: Sharples, R., Walker, A., \& Cropper, M.\ 1990, \mnras, 246, 54
399:
400: \bibitem[Spaenhauer et al.(1992)]{sp92}
401: Spaenhauer, A., Jones, B.F., Whitford, A.E. 1992, \aj, 103, 297
402:
403: \bibitem[Sumi et al.(2004)]{sumi04} Sumi, T., et al.\ 2004, \mnras, 348, 1439
404:
405: \bibitem[Terndrup et al.(1995)]{terndrup95}
406: Terndrup, D.M., Sadler, E.M., Rich, R.M. 1995, \aj, 110, 1174
407:
408: \bibitem[Thomas et al.(2003)]{thomas03} Thomas, D., Maraston, C., Bender, R. 2003, \mnras, 339, 897
409:
410: \bibitem[Zhao et al.(1994)]{zhao94} Zhao, H., Spergel, D.N., Rich, R.M. 1994, \aj, 108, 2154
411:
412: %\bibitem[Zhao, Spergel \& Rich(1996)]{zhao96} Zhao, H., Rich, R.M. \& Spergel, %D.N. 1996, \mnras, 282, 175
413:
414: \bibitem[Zhao(1996)]{zhao96} Zhao, H.S. 1996, \mnras, 283, 149
415:
416:
417:
418: \bibitem[Zoccali et al.(2003)]{zoc03} Zoccali, M., et al.\ 2003, \aap, 399, 931
419:
420: %\bibitem[Auri\`ere(1982)]{aur82} Auri\`ere, M. 1982, \aap,
421: % 109, 301
422: %\bibitem[Canizares et al.(1978)]{can78} Canizares, C. R.,
423: % Grindlay, J. E., Hiltner, W. A., Liller, W., \&
424: % McClintock, J. E. 1978, \apj, 224, 39
425: %\bibitem[Peterson(1976)]{pet76} Peterson, C. J. 1976, \aj, 81, 617
426: %\bibitem[Rudnick et al.(2003)]{rudnick03} Rudnick, G. et al., 2003, \apj, 599, 847
427: %\bibitem[Spitzer(1985)]{spi85} Spitzer, L. 1985, Dynamics of
428: % Star Clusters, J. Goodman \& P. Hut, Dordrecht: Reidel, 109
429: %\bibitem[Treu et al.(2003)]{treu03} Treu, T. et al., 2003, \apj, 591, 53
430:
431:
432: \end{thebibliography}
433:
434:
435:
436: \end{document}
437:
438: