1: % manuscript produces a one-column, double-spaced document:
2: %%\documentclass[manuscript]{aastex}
3: %% preprint[2] produces a [double-column], single-spaced document:
4: %\documentclass[12pt,preprint]{aastex}
5: \documentclass{emulateapj}
6: %% \documentclass[preprint2]{aastex}
7:
8: \newcommand{\vdag}{(v)^\dagger}
9: \newcommand{\kms}{km\,s$^{-1}$}
10: \newcommand{\ionCaT}{\ion{Ca}{2}}
11:
12: \shorttitle{Towards a Solution for the Calcium Puzzle}
13: \shortauthors{Michielsen et al.}
14:
15: \begin{document}
16:
17: \title{Towards a Solution for the Ca\,II Triplet Puzzle~:~Results
18: from Dwarf Elliptical Galaxies}
19:
20: \author{Dolf Michielsen} \affil{School of Physics and Astronomy, University of
21: Nottingham, University Park, NG7~2RD Nottingham, UK}
22: %\email{dolf.michielsen@nottingham.ac.uk}
23:
24: \author{Mina Koleva\altaffilmark{1} and Philippe Prugniel\altaffilmark{2}}
25: \affil{Universit\'e de Lyon, Lyon, France ; Universit\'e Lyon~1, Villeurbanne,
26: F-69622, France ; Centre de Recherche Astronomique de Lyon, Observatoire de
27: Lyon, F-69561, France ; CNRS, UMR 5574 ; Ecole Normale Sup\'erieure de Lyon,
28: Lyon, France}
29: %\email{mina@phys.uni-sofia.bg}
30: %\email{prugniel@obs.univ-lyon1.fr}
31:
32: \author{Werner~W. Zeilinger} \affil{Institut f\"{u}r Astronomie,
33: Universit\"{a}t Wien, T\"{u}rkenschanzstrasse 17, A-1180 Wien, Austria}
34: %\email{zeilinger@astro.univie.ac.at}
35:
36: \author{Sven De Rijcke\altaffilmark{3} and Herwig Dejonghe}
37: \affil{Sterrenkundig Observatorium, Ghent University, Krijgslaan 281, S9,
38: B-9000 Ghent, Belgium}
39: %\email{sven.derijcke@ugent.be}
40: %\email{herwig.dejonghe@ugent.be}
41:
42: \author{Anna Pasquali} \affil{Max-Planck-Institut f\"ur Astronomie,
43: Koenigstuhl 17, D-69117 Heidelberg, Germany}
44: %\email{pasquali@mpia-hd.mpg.de}
45:
46: \author{Ignacio Ferreras} \affil{Physics Department, King's College London,
47: Strand, London WC2R~2LS, UK}
48: %\email{ferreras@star.ucl.ac.uk}
49:
50: \and
51:
52: \author{Victor~P. Debattista\altaffilmark{4}} \affil{Centre for Astrophysics,
53: University of Central Lancashire, Preston PR1~2HE, UK}
54: %\email{vpdebattista@uclan.ac.uk}
55:
56: \altaffiltext{1}{Department of Astronomy, St.~Kl.~Ohridski University of
57: Sofia, BG-1164 Sofia, Bulgaria}
58:
59: \altaffiltext{2}{GEPI Observatoire de Paris-Meudon, 5 place Jules Janssen,
60: Meudon, F-92195, France}
61:
62: \altaffiltext{3}{Research Postdoctoral Fellow of the Fund for Scientific
63: Research, Flanders, Belgium (FWO).}
64:
65: \altaffiltext{4}{RCUK Academic Fellow}
66:
67: \begin{abstract}
68: We present new estimates of ages and metallicities, based on
69: FORS/VLT optical (4400--5500\,\AA) spectroscopy, of 16 dwarf
70: elliptical galaxies (dE's) in the Fornax Cluster and in Southern
71: Groups. These dE's are more metal-rich and younger than previous
72: estimates based on narrow-band photometry and low-resolution
73: spectroscopy. For our sample we find a mean metallicity ${\rm [Z/H]}
74: = -0.33$\,dex and mean age 3.5\,Gyr, consistent with similar samples
75: of dE's in other environments (Local Group, Virgo). Three dE's in
76: our sample show emission lines and very young ages. This suggests
77: that some dE's formed stars until a very recent epoch and were self-enriched
78: by a long star formation history. Previous observations of large
79: near-infrared ($\sim 8500$\,\AA) \ionCaT\ absorption strengths in
80: these dE's are in good agreement with the new metallicity estimates,
81: solving part of the so-called Calcium puzzle.
82: \end{abstract}
83:
84: \keywords{galaxies: dwarf --- galaxies: fundamental parameters --- galaxies:
85: stellar populations}
86:
87: \section{Introduction}
88:
89: The observed strength of the near-infrared \ionCaT\ triplet absorption
90: lines in early-type galaxies has presented astronomers with an
91: interesting puzzle over the past couple of years. \citet{cenarro01}
92: defined a new CaT* index, carefully correcting for the underlying H
93: Paschen absorption. Whereas other metallicity tracers, such as Mg$_2$,
94: correlate with velocity dispersion $\sigma$, it was found that CaT*
95: \emph{anti-correlates} with $\sigma$ in elliptical galaxies (E's) and
96: in bulges of spiral galaxies \citep{saglia02, cenarro03,
97: falcon03}. Population synthesis model predictions also show that, for
98: sub-solar metallicities, CaT* should be sensitive to metallicity but
99: virtually independent of age, while at super-solar metallicity, the
100: CaT* saturates \citep{vazdekis03}. However, taking metallicities
101: estimated from optical spectra, \citet{saglia02} reported that the
102: measured CaT* values in E's are smaller by 0.5\,{\AA} than those
103: predicted by population synthesis models. \citet{michielsen03}
104: (hereafter Paper~1) showed that the anti-correlation of CaT* with
105: $\sigma$ extends into the dwarf elliptical (dE) regime. These dE's
106: were expected to have metallicities of the order of [Z/H]$\,\sim -1$,
107: and ages of the order of 10\,Gyr \citep{heldmould94, rakos01}. The
108: measured CaT* values were significantly larger than those expected for
109: such old, metal-poor stellar systems.
110:
111: All of the proposed solutions to this conundrum, such as variations of
112: the initial-mass function or the calcium yield as a function of
113: metallicity or velocity dispersion, require considerable fine-tuning.
114: None of them satisfactorily explains both the small CaT* in E's and
115: the large CaT* in dE's without creating other difficulties, e.g.\ with
116: the FeH~$\lambda$9916 index values observed in bright ellipticals (FeH
117: is strong in dwarf stars but nearly absent in giants) and with stellar
118: mass-to-light ratios \citep{saglia02, cenarro03}. However, the stellar
119: populations of E's and dE's are most likely not single-age,
120: single-metallicity populations, or SSPs, as was implicity assumed in
121: essentially all age and metallicity estimates \citep[see
122: e.g.][]{pasquali05}. Moreover, the stars that dominate the blue
123: spectral range (mostly hot dwarf stars) do not necessarily have the
124: same mean ages/metallicities as the stars producing the red light
125: (mostly cool giants). These issues, together with systematic
126: uncertainties inherent to population synthesis tools, potentially
127: contribute to the CaT puzzle.
128:
129: Because of their low surface brightness, accurate estimates of the
130: ages and metallicities of dE's are still scarce. Recent studies of
131: dE's in the Virgo cluster \citep{g03, v04} report younger ages and
132: higher metallicities than found in Fornax dE's. In Paper~1, the ages
133: and metallicities of the dE's were taken from the literature, where
134: low-resolution, modest S/N spectroscopic \citep{heldmould94} or
135: narrow-band photometric \citep{rakos01} techniques were used. As a
136: sanity check, we have now acquired high-resolution, high S/N optical
137: spectra with FORS/VLT of all the dE's for which we presented CaT*
138: measurements in Paper~1. This puts us in a position where we can for
139: the first time compare the CaT* measurements with model predictions
140: based on robust age and metallicity estimates.
141:
142: \section{Observations and data reduction} \label{obs}
143:
144: The sample consists of 7 dE's in the Southern NGC\,5044, NGC\,5898 and Antlia
145: Groups, and 9 dE's in the Fornax Cluster. All the observations were carried
146: out in Service Mode at the ESO-VLT, in seeing conditions between $0.6
147: -0.9$\arcsec~FWHM. Integration times varied between 2 and 7 hours per galaxy.
148: The Group dE's were observed in April and May 2005 using FORS2 with
149: \texttt{GRISM\_600B} and a slit width of 0.5\arcsec. This results in a
150: wavelength range of $3300 - 6200$\,\AA\ at a 3.0\,\AA\ (FWHM) resolution. The
151: Fornax dE's were observed in December 2005 and January 2006 using FORS1 with
152: the holographic \texttt{GRISM\_1200g} and a slit width of 1.0\arcsec. This
153: results in a wavelength range of $4350 - 5530$\,\AA\ at a 2.6\,\AA\ (FWHM)
154: resolution. The data reduction was carried out using the ESO-MIDAS
155: package\footnote{The image processing package ESO-MIDAS is developed and
156: maintained by the European Southern Observatory.}. The spectra for each
157: galaxy were bias-subtracted, flat-fielded, corrected for cosmic-ray events,
158: wavelength calibrated and co-added. Then the sky was subtracted and the
159: spectra were extinction-corrected and flux-calibrated using spectrophotometric
160: standard stars. For the purpose of this Letter, we extracted 1D spectra over an
161: aperture of radius $R_e/8$ \citep[or 1\arcsec\ for galaxies with $R_e <
162: 8$\arcsec;][]{derijcke05}, to be consistent with the region in which the CaT*
163: index was measured in Paper~1. These spectra have a S/N $> 75$\,\AA$^{-1}$ in
164: the wavelength region 4700$-$5500\,\AA. A full description of the data
165: analysis will be presented in a forthcoming paper (Koleva et al., in
166: preparation).
167:
168: \section{Results} \label{res}
169:
170: \subsection{Ages and metallicities} \label{agemet}
171:
172: To measure the ages and metallicities of the dE's, we compared our
173: observations to single-age, single-metallicity population (SSP)
174: synthesis models. We used two different techniques and models:
175: \emph{(i)} \citet[TMB03]{TMB03} models in the Lick/IDS index system,
176: and \emph{(ii)} Pegase-HR models \citep{leborgne04} with the
177: ELODIE.3.1 stellar library \citep{prugniel07} to which we perform a
178: full spectrum fit \citep{koleva07}. To transform our spectra to the
179: Lick/IDS system, we smoothed our spectra to the Lick/IDS resolution
180: ($\sim 8.4$\,\AA\ FWHM) and measured the H$\beta$, Mg$b$, Fe5270, and
181: Fe5335 indices \citep{wortheyetal94}. In each setup we have 3 stars in
182: common with the original Lick/IDS library. This number is too small to
183: determine systematic offsets, so we applied the offsets to the Jones
184: library \citep{wo97}. These corrected indices agree within the error
185: bars with those given in \citet{wo97}. We then compared the indices of
186: the dE's with TMB03 models with varying [$\alpha$/Fe] abundance
187: ratios, applying a quadratic interpolation over the 9 nearest SSP
188: model grid points. The 1$\sigma$ error contours on the derived
189: quantities are calculated by Monte Carlo simulations taking into
190: account the measured errors in the indices \citep[see][for a detailed
191: description]{cardieletal03}. Note that systematic errors introduced by
192: the conversion to the Lick/IDS system are not taken into account. This
193: approach yields an estimate for the ages, metallicities and abundance
194: ratios of the targeted dE's. We find that the dE's have solar
195: abundance ratios and metallicities in the range $-0.7 < {\rm [Z/H]} <
196: +0.2$, which justifies the use of the ELODIE.3.1 library to perform
197: the full-spectrum fit. This yields a second, independent, age and
198: metallicity estimate for the program dE's. The 1$\sigma$ errors are
199: computed in a similar way as with the indices. As is obvious from
200: Figure~\ref{fig_compare_agemet}, both techniques agree, with an rms
201: difference of 1.63\,Gyr in age and 0.09\,dex in [Z/H]. For a
202: discussion of possible systematic differences between the model
203: predictions, we refer interested readers to \citet{leborgne04}. This
204: makes us confident that our age and metallicity estimates are
205: robust. The advantage of the full spectrum fitting method is that we
206: do not have to convert the data to the (low-resolution) Lick/IDS
207: system and that we can carefully take into account filling by emission
208: lines. Emission is present in three dE's in our sample and we did not
209: study them using the index method. Very low-level [O\,\textsc{III}]
210: (5007\,\AA) emission was detected in FCC\,150, that would not have
211: been detected if the resolution or the S/N of the observations were
212: only slightly worse. The first 5 columns of Table~\ref{tab_agemet}
213: give, for our sample of dE's, the ages, metallicities and
214: [$\alpha$/Fe] ratio derived using TMB03, and the age and metallicity
215: derived using Pegase-HR. In the following we will use Pegase-HR
216: results. Because the full fit uses all available information in the
217: spectrum, the errors are typically a factor 2-3 smaller than with
218: inversions of bi-index grids.
219:
220: \begin{figure}
221: \vspace{5cm}
222: \special{hscale=75 vscale=75 hsize=570 vsize=150
223: hoffset=-58 voffset=-50 angle=0 psfile="f1.eps"}
224: \caption{
225: Comparison of SSP-equivalent age and metallicity estimates using
226: Lick/IDS indices in combination with TMB03 models and a full spectrum
227: fit with Pegase-HR models. The error bars are 1$\sigma$ errors. Both
228: approaches clearly agree.\label{fig_compare_agemet}}
229: \end{figure}
230:
231: %One has to keep in mind that we are treating the stellar population as
232: %if it were formed in a single instantaneous burst, which is of course
233: %an oversimplification of the real star formation histories of these
234: %objects \citep[see e.g.][]{pasquali05}. The solar abundance ratios
235: %point to an extended star formation history, while the SSP-equivalent
236: %age is rather dating the end of the star formation. Moreover, we will
237: %combine the ages and metallicities derived in the blue ($\sim$4700 --
238: %5500\,\AA) part of the spectrum with measurements obtained in the NIR
239: %($\sim$8500\,\AA). While hot dwarf stars dominate the blue light, the
240: %red part of the spectrum is dominated by cool giants. Therefore, if
241: %the star formation history is more complex than an instantaneous
242: %burst, there will be a discrepancy in the age/metallicity of stars
243: %dominating the light at different wavelengths. Especially galaxies
244: %with signs of recent star formation activity should be considered
245: %carefully. The analysis of more complex star formation histories
246: %is however beyond the scope of this Letter and will be explored in
247: %more detail in a forthcoming paper.
248:
249: \begin{table*}
250: \scriptsize
251: \begin{center}
252: \caption{Ages, metallicities and [$\alpha$/Fe] ratios\label{tab_agemet}}
253: \begin{tabular}{l*{7}{r@{\,$\pm$\,}l}}
254: \tableline
255: \tableline
256: & \multicolumn{6}{c}{Lick/IDS Indices - TMB03} & \multicolumn{4}{c}{Spectrum Fit - Pegase-HR} & \multicolumn{2}{c}{HM94} & \multicolumn{2}{c}{R01}\\
257: Galaxy Name& \multicolumn{2}{c}{Age} & \multicolumn{2}{c}{[Z/H]} & \multicolumn{2}{c}{[$\alpha$/Fe]} & \multicolumn{2}{c}{Age} & \multicolumn{2}{c}{[Z/H]} & \multicolumn{2}{c}{[Fe/H]} & \multicolumn{2}{c}{[Fe/H]} \\
258: & \multicolumn{2}{c}{(Gyr)}& \multicolumn{2}{c}{(dex)} & \multicolumn{2}{c}{(dex)} & \multicolumn{2}{c}{(Gyr)}& \multicolumn{2}{c}{(dex)} & \multicolumn{2}{c}{(dex)} & \multicolumn{2}{c}{(dex)} \\
259: \tableline
260: FCC\,043 & 1.82 & 0.17 & $-$0.05 & 0.04 & 0.05 & 0.03 & 2.28 & 0.04 & $-$0.08 & 0.01 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} \\
261: FCC\,046 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} & 1.15 & 0.02 & $-$1.07 & 0.02 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} \\
262: FCC\,136 & 6.56 & 1.48 & $-$0.18 & 0.05 & 0.08 & 0.05 & 5.37 & 0.23 & $-$0.26 & 0.01 & \multicolumn{2}{c}{\dots} & $-$0.41 & 0.12 \\
263: FCC\,150 & 6.50 & 0.80 & $-$0.41 & 0.04 & 0.00 & 0.05 & 3.87 & 0.12 & $-$0.38 & 0.01 & $-$0.89 & 0.03 & $-$0.75 & 0.11 \\
264: FCC\,204 & 1.77 & 0.17 & 0.09 & 0.04 & $-$0.06 & 0.03 & 2.46 & 0.05 & 0.07 & 0.01 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} \\
265: FCC\,207 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} & 1.50 & 0.05 & $-$0.72 & 0.03 & $-$1.19 & 0.05 & $-$1.51 & 0.09 \\
266: FCC\,245 & 6.75 & 1.23 & $-$0.66 & 0.09 & 0.13 & 0.08 & 5.04 & 0.34 & $-$0.59 & 0.03 & $-$1.00 & 0.07 & $-$1.17 & 0.10 \\
267: FCC\,266 & 6.56 & 0.97 & $-$0.56 & 0.06 & $-$0.01 & 0.05 & 3.95 & 0.17 & $-$0.47 & 0.01 & $-$0.84 & 0.07 & $-$0.83 & 0.11 \\
268: FCC\,288 & 5.56 & 0.76 & $-$0.56 & 0.05 & $-$0.04 & 0.06 & 4.20 & 0.24 & $-$0.57 & 0.02 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} \\
269: DW\,1 & 5.23 & 1.42 & $-$0.19 & 0.06 & 0.03 & 0.06 & 4.35 & 0.21 & $-$0.26 & 0.02 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} \\
270: DW\,2 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} & 1.06 & 0.02 & $-$0.56 & 0.03 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} \\
271: FS\,29 & 2.25 & 0.51 & 0.10 & 0.06 & $-$0.09 & 0.03 & 3.57 & 0.15 & $-$0.06 & 0.01 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} \\
272: FS\,75 & 2.45 & 0.58 & 0.16 & 0.07 & 0.02 & 0.03 & 4.01 & 0.17 & 0.04 & 0.01 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} \\
273: FS\,76 & 4.91 & 1.29 & 0.29 & 0.08 & 0.03 & 0.03 & 6.59 & 0.19 & 0.09 & 0.01 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} \\
274: FS\,131 & 6.88 & 2.35 & $-$0.12 & 0.07 & 0.04 & 0.06 & 4.64 & 0.33 & $-$0.04 & 0.02 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} \\
275: FS\,373 & 0.87 & 0.08 & 0.01 & 0.06 & 0.01 & 0.03 & 1.82 & 0.04 & 0.00 & 0.01 & \multicolumn{2}{c}{\dots} & \multicolumn{2}{c}{\dots} \\
276: \tableline
277: \end{tabular}
278: \end{center}
279: \end{table*}
280:
281: \subsection{Comparison with literature data}
282:
283: There is clearly a large disparity between the new age and metallicity
284: estimates presented in this Letter and the existing estimates from the
285: literature that needs to be explained. \citet[hereafter HM94]{heldmould94} and
286: \citet[hereafter R01]{rakos01} studied Fornax dE's and have respectively 5 and
287: 6 galaxies in common with our sample. Their metallicity estimates are listed
288: in the last two columns of Table~\ref{tab_agemet}. First of all, one of the
289: problems that HM94 faced at the time was the lack of age and metallicity
290: dependent population synthesis models. They assumed that the ages of the dE's
291: are comparable to those of Galactic globular clusters in order to justify the
292: use of a metallicity scale calibrated on globular clusters. However, most
293: model bi-index grids are not perpendicular in age and metallicity. At a given
294: measured metal index, a younger age generally implies a higher metallicity.
295: HM94 were aware of this shortcoming at the time and cautioned that their
296: metallicities were probably underestimates. One of the galaxies we have in
297: common with HM94, FCC\,207, shows emission lines, indicative of on-going star
298: formation. We find a very young SSP-equivalent age of 1.5\,Gyr. From H$\alpha$
299: imaging, we know that the emission in FCC\,207 is concentrated in the central
300: 1\arcsec\ \citep{derijcke03}. HM94 did not report emission in FCC\,207 (see
301: their Figure~1), probably because their technique of nodding the telescope
302: perpendicular to the slit to sample the whole nuclear region diluted the
303: emission lines. They did note that the H$\delta$ absorption of FCC\,207 was
304: considerably stronger than the (emission-filled) H$\gamma$ and H$\beta$
305: absorption, but attributed this to measurement errors on H$\delta$.
306:
307: Secondly, R01 calibrated their narrow-band photometry metallicity scale on the
308: observations of HM94, re-observing all the HM94 dE's. In fact, R01 noted that
309: the metallicity derived from the narrow-band colours underestimates the true
310: metallicity, and that the off-set is larger if the population is younger.
311: However they only show 10 and 13\,Gyr models, using the latter to compute
312: corrections for their metallicity scale. Using corrections derived for
313: younger ages may reconcile their results with ours.
314:
315: \subsection{CaT measurements and predictions} \label{cat}
316:
317: \begin{figure}
318: \vspace{9cm}
319: \special{hscale=65 vscale=65 hsize=570 vsize=250
320: hoffset=-35 voffset=-30 angle=0 psfile="f2.eps"}
321: % \includegraphics[width=0.45\textwidth]{f2}
322: % \plotone{f2.eps}
323: \caption{ CaT, PaT and CaT* versus age and metallicity for the
324: Fornax cluster and Southern group dE's. Overlaid are
325: \citet{vazdekis03} SSP models with Salpeter IMF. FCC\,046
326: (square), FCC\,207 (diamond) and DW2 (star) are the three dE's
327: with emission lines. FCC\,245 is the most significant
328: outlier.\label{fig_models}}
329: % \end{center}
330: \end{figure}
331:
332: \citet{cenarro01} defined a set of new indices that quantify the
333: strengths of \emph{(i)} the \ionCaT\ lines (CaT), \emph{(ii)} the
334: Paschen P12, P14, and P17 lines (PaT), and \emph{(iii)} the \ionCaT\
335: corrected for the absorption contributed by the Paschen P13, P15, and
336: P16 lines (${\rm CaT*} = {\rm CaT} - {\rm 0.93} \times {\rm PaT}$). In
337: Figure~\ref{fig_models}, we use our new age and metallicity estimates
338: to compare the observed CaT, PaT and CaT* strengths to the SSP
339: predictions of \citet{vazdekis03} with Salpeter initial mass function
340: (IMF). FS\,75 was not observed in the \ionCaT. For most of the dE's,
341: the values are now consistent with the model predictions. FCC\,046 has
342: a high CaT for its age and metallicity. However, this galaxy has
343: emission lines and shows evidence for recent star formation
344: \citep{derijcke03}. This is reflected in the high PaT absorption
345: strength, and the corrected CaT* agrees very well with the model
346: prediction based on FCC\,046's metallicity. the values of the other
347: Fornax dE with emission, FCC\,207, agree perfectly well with the model
348: predictions. Surprisingly, in DW\,2 we also found emission lines.
349: However, the CaT of DW\,2 is already too low compared to the models,
350: and a high PaT would aggravate the situation for CaT*. Finally,
351: FCC\,245 appears to be quite normal, morphologically and
352: photometrically speaking, but its CaT is also lower than expected for
353: its age and metallicity. Still, 13 out of 15 of the observed dE's now
354: have measured ages, metallicities, and CaT* values that agree with
355: model predictions. The fact that CaT* values predicted using
356: age/metallicity estimates that were derived from spectral features in
357: the blue part of the spectrum ($\sim$4700 -- 5500\,\AA) agree quite
358: reasonably with the observed CaT* values in the NIR ($\sim$8500\,\AA)
359: indicates that the SSP assumption is not the cause of the CaT puzzle
360: for dE's.
361:
362: \section{Discussion and conclusions} \label{disc}
363:
364: At least in the dE regime, the \ionCaT\ Triplet puzzle seems to be
365: solved. With the new age and metallicity estimates presented in this
366: Letter, the predicted and observed CaT* indices are in good agreement
367: for all sample galaxies but two. The fact that CaT* values predicted
368: using age/metallicity estimates that were derived from spectral
369: features in the blue part of the spectrum agree with the observed CaT*
370: values in the NIR indicates that the SSP assumption is not the cause
371: of the CaT puzzle for dE's. Rather, the CaT puzzle in the dE regime
372: was caused by the spuriously low metallicities and high ages, derived
373: from lower resolution spectra using less sophisticated theoretical
374: models, assigned to Fornax dE's. This shows that the CaT* index, and,
375: in old stellar systems in which the PaT index is small, the CaT index
376: as well, is indeed a good tracer of metallicity.
377:
378: This also solves the apparent dichotomy between dE's on the one hand
379: and globular clusters, Local Group dwarf spheroidals (dSph), and
380: ultra-compact dwarfs (UCD) on the other hand. CaT line-strengths
381: measured in individual stars of Local Group dSphs have been shown to
382: be a very accurate tracer of metallicity \citep{b06,to03} and
383: [Fe/H]-values derived from CaT measurements have been used extensively
384: to construct metallicity distributions of the stars in dSphs. Also,
385: for UCDs \citep{ev07} and globular clusters \citep{saglia02} the CaT
386: index has proved to be an excellent tracer of metallicity. Here, we
387: have shown that in dE's as well, the CaT* index measured from
388: integrated-light spectra can be used as a tracer of metallicity.
389:
390: %In the high-metallicity regime,
391: %possible explanations for the low CaT* values can now be
392: %re-examined. A good candidate is a true underabundance of calcium in
393: %massive systems, because also the blue Ca4227 index is found to be low
394: %in E's \citep[e.g.][]{vazdekis97, thomas03Ca}. Variations of
395: %supernova yields with respect to metallicity, in particular a
396: %reduction in the Ca yields at high metallicity could explain the Ca
397: %underabundance in giant E's, whilst dE's are unaffected
398: %\citep{thomas03Ca}. However, \citet{prochaskaetal05} argue that
399: %strong CN absorption may induce the low Ca4227 values in E's. Another
400: %proposed explanation for the low Ca absorption is a variation in the
401: %dwarf/giant ratio through a varying IMF \citep{cenarro03}. However
402: %this fails to reconcile at the same time the behaviour of both the
403: %optical and the near-infrared Ca features \citep{cenarro04}. Finally,
404: %the dependence of the star formation history on galaxy mass and
405: %environment may also play a important role \citep{cenarro04}.
406: %Therefore, a better understanding of \emph{(i)} calcium supernova
407: %yields, \emph{(ii)} the exact dependence of the CaT* on calcium and
408: %other elements and \emph{(iii)} the behaviour of calcium-sensitive
409: %indices with more complex star formation histories, is needed before
410: %we can completely solve the calcium puzzle,
411:
412: To summarise, we derive new age and metallicity estimates for 16 dE's
413: in the Fornax Cluster and in Southern Groups using high S/N optical
414: VLT/FORS1+2 spectra. We have measured the H$\beta$, Mg$b$, Fe5270, and
415: Fe5335 indices in the Lick/IDS system and applied the TMB03 models to
416: them. We find that these dE's have solar [$\alpha$/Fe] abundance
417: ratios. A full-spectrum fit using Pegase-HR with the ELODIE.3.1
418: stellar library provides us with a second, independent age and
419: metallicity estimate for these galaxies. We find both approaches to be
420: in excellent agreement. With mean metallicity ${\rm [Z/H]} =
421: -0.35$\,dex and ages younger than $\approx 7$\,Gyr, these dE's are
422: more metal-rich and younger than previously thought. Some even show
423: strong emission lines, an indication of on-going star formation, in
424: agreement with previous H$\alpha$ imaging of dE's
425: \citep{derijcke03,michielsen04}. The ages and metallicities we derive
426: for the dE's in the Fornax cluster and in Southern groups fall in
427: roughly the same range as those derived by \citet{g03} and \citet{v04}
428: for dE's in the Virgo cluster. This is at variance with previous
429: estimates for Fornax dEs which yielded lower metallicities and higher
430: ages \citep{heldmould94, rakos01}, based on lower resolution spectra
431: and less sophisticated theoretical models. The new age and metallicity
432: estimates are in good agreement with the observed \ionCaT\ triplet
433: absorption strengths, solving the Calcium puzzle for low-mass systems.
434:
435: \acknowledgments Based on observations collected at the European Southern
436: Observatory, Paranal, Chile (Programs 075.B-0179 and 076.B-0196). DM
437: acknowledges the EU MAGPOP RTN for financial support.
438:
439: \begin{thebibliography}{}
440:
441: % CaT in the Fornax dSph
442: \bibitem[Battaglia et al.(2006)]{b06} Battaglia, G., et al.\ 2006, \aap, 459,
443: 423
444:
445: % estimating age and metalliticy using bivariate grids
446: \bibitem[Cardiel et al.(2003)]{cardieletal03} Cardiel, N., Gorgas, J.,
447: S{\'a}nchez-Bl{\'a}zquez, P., Cenarro, A.~J., Pedraz, S., Bruzual,
448: G., \& Klement, J.\ 2003, \aap, 409, 511
449:
450: % differences in optical Ca in Coma and Field
451: \bibitem[Cenarro et al.(2004)]{cenarro04} Cenarro, A.~J.,
452: S{\'a}nchez-Bl{\'a}zquez, P., Cardiel, N., \& Gorgas, J.\ 2004, \apjl, 614,
453: L101
454:
455: % sigma - CaT anti-correlation in E's : IMF variations?
456: \bibitem[Cenarro et al.(2003)]{cenarro03} Cenarro, A.~J., Gorgas, J.,
457: Vazdekis, A., Cardiel, N., \& Peletier, R.~F.\ 2003, \mnras, 339, L12
458:
459: % definition of CaT, PaT and CaT*
460: \bibitem[Cenarro et al.(2001)]{cenarro01} Cenarro, A.~J., Cardiel, N., Gorgas,
461: J., Peletier, R.~F., Vazdekis, A., \& Prada, F.\ 2001, \mnras, 326, 959
462:
463: % structural paramters of sample dEs (R_eff)
464: \bibitem[de Rijcke et al.(2005)]{derijcke05} de Rijcke, S., Michielsen, D.,
465: Dejonghe, H., Zeilinger, W.~W., \& Hau, G.~K.~T.\ 2005, \aap, 438, 491
466:
467: % Halpha imaging of FCC\,046 and FCC\,207
468: \bibitem[De Rijcke et al.(2003)]{derijcke03} De Rijcke, S., Zeilinger, W.~W.,
469: Dejonghe, H., \& Hau, G.~K.~T.\ 2003, \mnras, 339, 225
470:
471: % CaT in UCDs
472: \bibitem[Evstigneeva et al.(2007)]{ev07} Evstigneeva, E.~A., Gregg, M.~D.,
473: Drinkwater, M.~J., \& Hilker, M.\ 2007, \aj, 133, 1722
474:
475: % sigma - CaT anti-correlation in bulges of spirals
476: \bibitem[Falc{\'o}n-Barroso et al.(2003)]{falcon03} Falc{\'o}n-Barroso, J.,
477: Peletier, R.~F., Vazdekis, A., \& Balcells, M.\ 2003, \apjl, 588, L17
478:
479: % ages/metallicities of Virgo dE's
480: \bibitem[Geha et al.(2003)]{g03} Geha, M., Guhathakurta, P., \& van der Marel,
481: R.~P.\ 2003, \aj, 126, 1794
482:
483: %spectroscoy of dE's in the Fornax Cluster: central ages and metallicities
484: \bibitem[Held \& Mould(1994)]{heldmould94} Held, E.~V., \& Mould, J.~R.\ 1994,
485: \aj, 107, 1307 (HM94)
486:
487: % full spectrum fitting
488: \bibitem[Koleva et al.(2007)]{koleva07} Koleva, M., Prugniel, P., Ocvirk, P.,
489: \& Le Borgne, D.\ 2007, in the proceedings of IAUS241, "Stellar Populations
490: as Building Blocks of Galaxies", eds. A. Vazdekis and R. Peletier
491:
492: % population synthesis with pegase-HR
493: \bibitem[Le Borgne et al.(2004)]{leborgne04} Le Borgne, D., Rocca-Volmerange,
494: B., Prugniel, P., Lan{\c c}on, A., Fioc, M., \& Soubiran, C.\ 2004, \aap,
495: 425, 881
496:
497: % Halpha in Fornax dE's
498: \bibitem[Michielsen et al.(2004)]{michielsen04} Michielsen, D., De Rijcke, S.,
499: Zeilinger, W. W., Prugniel, P., Dejonghe, H., Roberts, S., 2004, MNRAS, 353,
500: 1293
501:
502: % sigma - CaT anti-correlation in dE's
503: \bibitem[Michielsen et al.(2003)]{michielsen03} Michielsen, D., De Rijcke, S.,
504: Dejonghe, H., Zeilinger, W.~W., \& Hau, G.~K.~T.\ 2003, \apjl, 597, L21
505: (Paper~1)
506:
507: % Apples dE
508: \bibitem[Pasquali et al.(2005)]{pasquali05} Pasquali, A., Larsen, S.,
509: Ferreras, I., Gnedin, O.~Y., Malhotra, S., Rhoads, J.~E., Pirzkal, N., \&
510: Walsh, J.~R.\ 2005, \aj, 129, 148
511:
512: % low Ca4227 by CN absorption
513: \bibitem[Prochaska et al.(2005)]{prochaskaetal05} Prochaska L.~C., Rose J.~A.,
514: \& Schiavon R.~P., 2005, \aj, 130, 2666
515:
516: % ELODIE.3.1 stellar library
517: \bibitem[Prugniel et al.(2007)]{prugniel07} Prugniel, P., Soubiran, C.,
518: Koleva, M., \& Le Borgne, D.\ 2007, ArXiv Astrophysics e-prints,
519: arXiv:astro-ph/0703658
520:
521: % Stroemgren photometry for ages and Metallicities of Fornax dE's
522: \bibitem[Rakos et al.(2001)]{rakos01} Rakos, K., Schombert, J., Maitzen,
523: H.~M., Prugovecki, S., \& Odell, A. \ 2001, \aj, 121, 1974 (R01)
524:
525: % sigma - CaT anti-correlation in Es
526: \bibitem[Saglia et al.(2002)]{saglia02} Saglia, R.~P., Maraston, C., Thomas,
527: D., Bender, R., \& Colless, M.\ 2002, \apjl, 579, L13
528:
529: \bibitem[Tolstoy et al.(2003)]{to03} Tolstoy, E. Irwin, M. J., Cole,
530: A. A., Pasquini, L., Gilmozzi, R., Gallagher, J. S., 2001, \mnras,
531: 327, 918
532:
533: % population synthesis models with variable [alpha/Fe]
534: \bibitem[Thomas et al.(2003a)]{TMB03} Thomas, D., Maraston, C., \& Bender, R.\
535: 2003a, \mnras, 339, 897 (TMB03)
536:
537: % new clues on the Ca underabundance in Es
538: \bibitem[Thomas et al.(2003b)]{thomas03Ca} Thomas, D., Maraston, C., \&
539: Bender, R.\ 2003b, \mnras, 343, 279
540:
541: % ages/metallicities in Virgo dEs
542: \bibitem[van Zee et al.(2004)]{v04} van Zee, L., Barton, E. J., Skillman, E.
543: D., 2004, \aj, 128, 2797
544:
545: % stellar synthesis model predictions of CaT, PaT and CaT*
546: \bibitem[Vazdekis et al.(2003)]{vazdekis03} Vazdekis, A., Cenarro, A.~J.,
547: Gorgas, J., Cardiel, N., \& Peletier, R.~F.\ 2003, \mnras, 340, 1317
548:
549: % Ca4227 lower than model predictions
550: \bibitem[Vazdekis et al.(1997)]{vazdekis97} Vazdekis, A., Peletier, R.~F.,
551: Beckman, J.~E., \& Casuso, E.\ 1997, \apjs, 111, 203
552:
553: % Lick/IDS indices and resolution
554: \bibitem[\protect\citeauthoryear{Worthey et al.}{1994}]{wortheyetal94} Worthey
555: G., Faber S.~M., Gonzalez J.~J., Burstein D., 1994, ApJS, 94, 687
556:
557: % transfer to the Lick/IDS system
558: \bibitem[Worthey \& Ottaviani(1997)]{wo97} Worthey, G., \& Ottaviani, D.~L.\
559: 1997, \apjs, 111, 377
560:
561: \end{thebibliography}
562:
563: \end{document}
564:
565: