Ogle-2009-Blg-076s — the Most Metal-Poor Dwarf Star in the Galactic Bulge11affiliation: Based on observations made with the European Southern Observatory telescopes, Program ID 082.B-0453(B).
Measurements based on a large number of red giant stars suggest a broad metallicity distribution function (MDF) for the Galactic bulge, centered on . However, recently, a new opportunity emerged to utilize temporary flux amplification (by factors of or more) of faint dwarf stars in the Bulge that are gravitationally lensed, making them observable with high-resolution spectrographs during a short observational window. Surprisingly, of the first 6 stars measured, 5 have , suggesting a highly skewed MDF, inconsistent with observations of giant stars. Here we present a detailed elemental abundance analysis of OGLE-2009-BLG-076S, based on a high-resolution spectrum obtained with the UVES spectrograph at the ESO Very Large Telescope. Our results indicate it is the most metal-poor dwarf star in the Bulge yet observed, with . Our results argue against a strong selection effect disfavoring metal-poor microlensed stars. It is possible that small number statistics is responsible for the giant/dwarf Bulge MDF discrepancy. Should this discrepancy survive when larger numbers of Bulge dwarf stars (soon to be available) are analyzed, it may require modification of our understanding of either Bulge formation models, or the behavior of metal-rich giant stars.
Subject headings:gravitational lensing — Galaxy: bulge — Galaxy: formation — Galaxy: evolution — stars: abundances — stars: fundamental parameters
The metallicity distribution function (MDF) of the Galactic bulge (hereafter, the Bulge) as measured by red giant stars has a long controversial history. In the 1980s and 1990s studies of red giant stars in Baade’s window using low-dispersion spectroscopy, showed that the Bulge MDF was quite metal-rich at (e.g., Whitford & Rich, 1983; Rich, 1988; Terndrup et al., 1990). Later, the mean metallicity was revised downward to by the first high-resolution study of K giant stars in Baade’s window by McWilliam & Rich (1994). Since then, results based on high-resolution spectroscopic studies of several hundred giant stars in or around Baade’s window, show that the Bulge MDF peaks slightly below solar metallicity at (Fulbright et al., 2006, 2007; Cunha & Smith, 2006; Cunha et al., 2007, 2008; Rich & Origlia, 2005; Rich et al., 2007; Lecureur et al., 2007; Zoccali et al., 2003, 2008). However, the recent progress in observing low-luminosity dwarf stars in the Bulge while they are being optically magnified during gravitational microlensing events appears to give markedly different results and have re-ignited the debate over the Bulge MDF. So far, detailed elemental abundances based on high-resolution spectroscopy for a total of six microlensed dwarf and subgiant stars have been published (Johnson et al., 2007, 2008; Cohen et al., 2008, 2009; Bensby et al., 2009). Five of these six stars have high super-solar metallicities in the range and it is only OGLE-2008-BLG-209S that is somewhat metal-poor at (Bensby et al., 2009). The average metallicity from the six microlensed Bulge stars is , which is 0.4 dex higher than the MDF obtained from giant stars.
OGLE-2008-BLG-209S is not just the only one of the six microlensed stars that turned out to be metal-poor but also the only one that turned out to be slightly evolved, with stellar parameters placing it on the subgiant branch. This led Cohen et al. (2009) to speculate that OGLE-2008-BLG-209S is an exception to the rule, and that it should perhaps be excluded from the microlensed stellar sample. The average metallicity would then rise to , i.e., even more discrepant when comparing to the giant star MDF.
As microlensing events have no preference of picking source stars with a particular metallicity, observing enough events should provide an unbiased estimate of the Bulge MDF. So far, the microlensing events that have been observed in the Bulge lie at approximately the same angular distance from the Galactic center as the red giant stars in Baade’s window. Hence, there is no reason to expect that their MDFs should differ. By randomly picking 6 stars, 40 000 times, from the Zoccali et al. (2008) sample of giant stars, Cohen et al. (2009) found that in only 0.4 % of the cases, the randomly drawn red giant sample was as metal-rich as the sample of microlensed dwarf stars. The question was then which of the two tracers that give the correct picture: giant stars or the microlensed dwarf stars?
In this Letter we present first results from a detailed abundance analysis of OGLE-2009-BLG-076S, the source star of another microlensing event toward the Bulge. We focus here on the metallicity, the -elements Mg, Si, and Ti, the stellar age, distance, and the star’s kinematic properties. The full analysis and results for other -elements, light elements, iron-peak elements, -, and -process elements, will be presented in an upcoming paper.
2. Observations and data reduction
On 2009 March 21, the OGLE early warning system111http://ogle.astrouw.edu.pl/ogle3/ews/ews.html (Udalski, 2003) identified OGLE-2009-BLG-076S as a possible high-magnification microlensing event toward the Bulge at deg. Since the intrinsic source flux (inferred from the microlensing model) indicated that the star was a dwarf star we triggered our ToO observations, without knowing the color of the star, with the ESO Very Large Telescope (VLT) on Paranal. Due to the limited visibility of the Bulge in March, the target had to be observed toward the end of the night. Hence, OGLE-2009-BLG-076S was observed during the early morning hours beginning March 26 (MJD: 4916.29099), a few hours after peak brightness (see Fig. 1). At maximum the light from the source star was amplified by a factor of 68. Using the UVES spectrograph (Dekker et al., 2000), located at VLT UT2, configured with dichroic number 2, the target was observed for a total of two hours, split into four 30 minute exposures. The resulting spectrum was recorded on three CCDs with wavelength coverages between 3760-4980 Å (blue CCD), 5680-7500 Å (lower red CCD), and 7660-9460 Å (upper red CCD). A slitwidth yielded a spectral resolution of .
The data were reduced with the most recent version of the UVES pipeline. The typical signal-to-noise () ratio per pixel at 6000 Å in the lower red CCD is (compare Fig. 2). In the spectrum from the blue CCD, the was, however, too low to allow either secure identification of spectral line features or accurate measurement of equivalent widths. Hence, the spectral region blue-ward of 5680 Å was not used in the analysis.
Before the observation of the main target, we observed HR 6141, a rapidly rotating B2V star, at an airmass similar to what was expected for the Bulge star, to divide out telluric lines in the spectrum. We also obtained a solar spectrum, by observing the asteroid Pallas, that was used to normalize the elemental abundances in OGLE-2009-BLG-076S to those in the Sun. The per pixel at 6000 Å in the solar spectrum and the B star spectrum is and , respectively.
Examples of the Bulge star spectrum and the solar spectrum are shown in Fig. 2.
3.1. Stellar parameters from spectroscopy
The determination of stellar parameters and calculation of elemental abundances were done in the very same way as in Method 1 of Bensby et al. (2009), wherein full details of the analysis can be found, which is based on equivalent width measurements and one-dimensional LTE model stellar atmospheres calculated with the Uppsala MARCS code (Gustafsson et al., 1975; Edvardsson et al., 1993; Asplund et al., 1997). The linelist is an expanded version of the linelist of Bensby et al. (2003, 2005) and is fully given in Bensby et al. (2010, in prep.). The effective temperature () is determined by requiring excitation balance of abundances from Fe i lines, the microturbulence parameter () by requiring balance of abundances from Fe i lines versus the reduced equivalent width (), and the surface gravity () from ionization balance between abundances from Fe i and Fe ii lines.
Due to shorter wavelength coverage and lower only a fraction of the Fe i and Fe ii lines available in the above mentioned linelist could be measured and utilized in the analysis. Hence, based on 57 Fe i and 7 Fe ii lines we find that OGLE-2009-BLG-076S has K, , , and an absolute Fe abundance of , or .
The errors in the stellar parameters are of standard nature, and we estimate them to be of the same magnitude as in Bensby et al. (2009), i.e., K, and . These errors will introduce a random error of 0.12 dex in [Fe/H] (see Table 6 in Bensby et al., 2009). Further, the standard deviation around the mean Fe i abundance is 0.10 dex. The corresponding formal error () in the mean Fe abundance is 0.013 dex, which is negligible in comparison to the effects that the uncertainties in the stellar parameters have on [Fe/H]. Hence, we estimate that the total error in [Fe/H] is 0.12 dex.
3.3. Microlensing techniques
Using microlensing techniques (Yoo et al., 2004), based on and data taken with the SMARTS 1.3 m telescope at CTIO, OGLE-2009-BLG-076S is estimated to have an intrinsic color of , i.e., similar to the Sun, and an absolute magnitude of . The last term originates from the fact that the microlensing technique assumes that the source star (S) lies at the same distance as the clump stars, in which case this term would vanish. The color- calibration by Ramírez & Meléndez (2005) then implies that the effective temperature should be 5670 K. This is 200 K lower than the “spectroscopic” temperature that we derive. However, if the true temperature of OGLE-2009-BLG-076S is lower, the abundances derived from the Fe i lines will also be lower, making the star even more metal-poor.
4. Properties of OGLE-2009-BLG-076S
4.1. Evolutionary status
The values of the effective temperature and surface gravity demonstrates that OGLE-2009-BLG-076S is a true dwarf star. Utilizing the fundamental relationship between surface gravity, effective temperature, bolometric magnitude (see, e.g., Eq. 4 in Bensby et al. 2003), we find an absolute magnitude of , in reasonable agreement with the value based on microlensing techniques. Plotting the star on top of the Yonsei-Yale isochrones (Yi et al., 2001; Kim et al., 2002; Demarque et al., 2004) we find an age of Gyr, for OGLE-2009-BLG-076S, (see Fig. 3). This old age is consistent with Bulge stars being an old population (e.g., Feltzing & Gilmore, 2000), and is similar to what is found for thick disk stars at the same metallicity (e.g. Bensby et al., 2005, 2007). From the evolutionary tracks of Yi et al. (2003) we find that OGLE-2009-BLG-076S has a mass of M.
4.2. Radial velocity
4.3. Bulge membership
Microlensed sources are not random Bulge sources, they are heavily biased to come from more distant parts of the Bulge. This is because the lens must be in front of the source, and the probability of lensing is roughly proportional to the square-root of the lens-source distance. Most lenses are themselves in the Bulge, particularly toward this direction, where the integrated density of Bulge sources is extremely high.
Furthermore, the event is at deg, i.e., 350 pc below the Galactic center. In this direction, the Bulge (which is flattened by a factor of ) has an effective width of 600 pc. Beyond this range, its density falls off very rapidly (see, e.g., the model by Han & Gould, 1995, 2003). Hence, a random source would have a very low probability of lying in front of the red clump centroid.
Given the low probability of the source star being a foreground disk star, that the measured radial velocity of OGLE-2009-BLG-076S is consistent with the bulk of Bulge stars, that it has an old age consistent with the Bulge being an old stellar population, it is very likely that OGLE-2009-BLG-076S is a Bulge dwarf star.
4.4. Elemental abundance ratios
The abundances of OGLE-2009-BLG-076S were normalized to those of the Sun (as determined from the Pallas spectrum) on a line-by-line basis and then averaged for each element. The absolute iron abundance measured in Sect. 3.1 is then equal to .
Figure 4 shows three abundance ratios in OGLE-2009-BLG-076S compared to the average abundance ratios for thick disk stars in the metallicity range and to Bulge giant stars in the metallicity range . Compared to the Bulge giant stars, OGLE-2009-BLG-076S shows the same degree of Mg and Ti enhancements, while it has a lower enhancement in Si. Compared to the thick disk dwarf stars, OGLE-2009-BLG-076S has similar enhancements in all three elements, with a hint of being slightly more enhanced than the median thick disk. These -to-iron over-abundances at indicate a period of intense star formation in the Bulge where chemical enrichment mainly is due to massive stars, exploding as core-collapse supernovae. Furthermore, the solar-type abundance ratios for the other five microlensed dwarf stars at super-solar metallicities (Bensby et al., 2009; Johnson et al., 2007, 2008; Cohen et al., 2008, 2009) is consistent with the drop in -enhancement above solar metallicity observed in giant stars (compare Fig. 10 in Bensby et al., 2009).
5. Discussion and conclusion
OGLE-2009-BLG-076S is the currently most metal-poor dwarf star in the Bulge for which a detailed elemental abundance analysis based on high-resolution spectroscopy has been performed. It has a metallicity of , abundance ratios consistent with in situ Bulge giant stars, and an old age of 11.5 Gyr. This result demonstrates that the Bulge does indeed contain metal-poor dwarf stars, as predicted by the presence of metal-poor giant stars, and that it is possible to observe them while they are being gravitationally microlensed.
The now seven microlensed dwarf and subgiant stars in the Bulge have an average metallicity of (using [Fe/H] values from Bensby et al. 2009 for the first four events). A two-sided Kolmogorov-Smirnow test with the giant stars from Zoccali et al. (2008) gives a very low probability that the MDFs from the dwarfs and the giants are the same (see Fig. 5).
There are essentially three possibilities to explain the giant/dwarf Bulge MDF discrepancy: (i) Due to systematic errors in either giant or dwarf metallicity measurements, the identical underlying MDF appears different. Cohen et al. (2009) discuss this, including the size of the required errors, and find it very unlikely that the analysis methods are faulty; (ii) The dwarf and the giant MDFs are different and have been correctly measured to be different. The discrepancy is caused by mass loss (as discussed by Cohen et al., 2009), or because of different spatial distributions and chemical inhomogeneities in the Bulge; (iii) The dwarf and giant MDFs are identical, but small number statistics for the dwarf stars happen to cause them look different; (iv) or some other unknown mechanism.
The new results for OGLE-2009-BLG-076S favor a combination of (ii) and (iii). While the giant star MDF is well-defined, it could also be, as suggested by Cohen et al. (2008) that the most metal-rich giant stars lose their entire envelope before reaching the RGB tip. Hence the most metal-rich giants are missing, and the MDF is somewhat shifted to lower [Fe/H]. This in combination with the small sample, perhaps too small to be representative of the Bulge MDF, of now seven microlensed stars: five very metal-rich dwarf stars; one subgiant star just below solar metallicity; and one metal-poor dwarf star, makes the difference between the giant star MDF and the MDF from the sample of microlensed stars too large. However, with OGLE-2009-BLG-076S, we have shown that metal-poor dwarf stars exist in the Bulge, and that, if enough microlensing events are observed, the differences between the giant and dwarf MDFs could in principle diminish to a level where the differences only are due to the effects of mass-loss of the most metal-rich giant stars.
- Asplund et al. (1997) Asplund, M., Gustafsson, B., Kiselman, D., & Eriksson, K. 1997, A&A, 318, 521
- Bensby et al. (2003) Bensby, T., Feltzing, S., & Lundström, I. 2003, A&A, 410, 527
- Bensby et al. (2005) Bensby, T., Feltzing, S., Lundström, I., & Ilyin, I. 2005, A&A, 433, 185
- Bensby et al. (2009) Bensby, T., Johnson, J., Cohen, J., Feltzing, S., Udalski, A., Gould, A., Huang, W., Thompson, I., Simmerer, J., & Adén, D. 2009, A&A, 499, 737
- Bensby et al. (2007) Bensby, T., Zenn, A. R., Oey, M. S., & Feltzing, S. 2007, ApJ, 663, L13
- Cohen et al. (2008) Cohen, J. G., Huang, W., Udalski, A., Gould, A., & Johnson, J. A. 2008, ApJ, 682, 1029
- Cohen et al. (2009) Cohen, J. G., Thompson, I. B., Sumi, T., Bond, I., Gould, A., Johnson, J. A., Huang, W., & Burley, G. 2009, arXiv:0904.2020v1 [astro-ph.SR]
- Cunha et al. (2007) Cunha, K., Sellgren, K., Smith, V. V., Ramirez, S. V., Blum, R. D., & Terndrup, D. M. 2007, ApJ, 669, 1011
- Cunha & Smith (2006) Cunha, K., & Smith, V. V. 2006, ApJ, 651, 491
- Cunha et al. (2008) Cunha, K., Smith, V. V., & Gibson, B. K. 2008, ApJ, 679, L17
- Dekker et al. (2000) Dekker, H., D’Odorico, S., Kaufer, A., Delabre, B., & Kotzlowski, H. 2000, in Proc. SPIE Vol. 4008, 534–545
- Demarque et al. (2004) Demarque, P., Woo, J.-H., Kim, Y.-C., & Yi, S. K. 2004, ApJS, 155, 667
- Edvardsson et al. (1993) Edvardsson, B., Andersen, J., Gustafsson, B., Lambert, D. L., Nissen, P. E., & Tomkin, J. 1993, A&A, 275, 101
- Feltzing & Gilmore (2000) Feltzing, S., & Gilmore, G. 2000, A&A, 355, 949
- Fulbright et al. (2006) Fulbright, J. P., McWilliam, A., & Rich, R. M. 2006, ApJ, 636, 821
- Fulbright et al. (2007) —. 2007, ApJ, 661, 1152
- Gustafsson et al. (1975) Gustafsson, B., Bell, R. A., Eriksson, K., & Nordlund, A. 1975, A&A, 42, 407
- Han & Gould (1995) Han, C., & Gould, A. 1995, ApJ, 447, 53
- Han & Gould (2003) —. 2003, ApJ, 592, 172
- Howard et al. (2008) Howard, C. D., Rich, R. M., Reitzel, D. B., Koch, A., De Propris, R., & Zhao, H. 2008, ApJ, 688, 1060
- Johnson et al. (2007) Johnson, J. A., Gal-Yam, A., Leonard, D. C., Simon, J. D., Udalski, A., & Gould, A. 2007, ApJ, 655, L33
- Johnson et al. (2008) Johnson, J. A., Gaudi, B. S., Sumi, T., Bond, I. A., & Gould, A. 2008, ApJ, 685, 508
- Kim et al. (2002) Kim, Y., Demarque, P., Yi, S. K., & Alexander, D. R. 2002, ApJS, 143, 499
- Lecureur et al. (2007) Lecureur, A., Hill, V., Zoccali, M., Barbuy, B., Gómez, A., Minniti, D., Ortolani, S., & Renzini, A. 2007, A&A, 465, 799
- McWilliam & Rich (1994) McWilliam, A., & Rich, R. M. 1994, ApJS, 91, 749
- Ramírez & Meléndez (2005) Ramírez, I., & Meléndez, J. 2005, ApJ, 626, 465
- Rich (1988) Rich, R. M. 1988, AJ, 95, 828
- Rich & Origlia (2005) Rich, R. M., & Origlia, L. 2005, ApJ, 634, 1293
- Rich et al. (2007) Rich, R. M., Origlia, L., & Valenti, E. 2007, ApJ, 665, L119
- Terndrup et al. (1990) Terndrup, D. M., Frogel, J. A., & Whitford, A. E. 1990, ApJ, 357, 453
- Udalski (2003) Udalski, A. 2003, Acta Astronomica, 53, 291
- Whitford & Rich (1983) Whitford, A. E., & Rich, R. M. 1983, ApJ, 274, 723
- Yi et al. (2001) Yi, S., Demarque, P., Kim, Y., Lee, Y., Ree, C. H., Lejeune, T., & Barnes, S. 2001, ApJS, 136, 417
- Yi et al. (2003) Yi, S. K., Kim, Y.-C., & Demarque, P. 2003, ApJS, 144, 259
- Yoo et al. (2004) Yoo, J., DePoy, D. L., Gal-Yam, A., & et al. 2004, ApJ, 603, 139
- Zhao (1996) Zhao, H. S. 1996, MNRAS, 283, 149
- Zoccali et al. (2008) Zoccali, M., Hill, V., Lecureur, A., Barbuy, B., Renzini, A., Minniti, D., Gómez, A., & Ortolani, S. 2008, A&A, 486, 177
- Zoccali et al. (2003) 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