1: % SAMPLE2.TEX -- AASTeX macro package tutorial paper.
2:
3: % The first item in a LaTeX file must be a \documentstyle command to
4: % declare the overall style of the paper. The two \documentstyle lines
5: % that are relevant for the AASTeX macros are shown; one is commented out
6: % so that the file can be processed.
7:
8: \documentclass[preprint]{aastex}
9: %\documentstyle[12pt,aasms]{article}
10: %\documentstyle[11pt,aaspp]{article}
11: %\documentstyle[aaspptwo]{article}
12: \usepackage{natbib}
13: \bibpunct[;]{(}{)}{;}{a}{}{,}
14:
15: % There are two optional preamble declarations that enable to user to
16: % control certain formatting options. \tighten is used with the
17: % aasms substyle to turn off double-spacing; don't do this for
18: % actual manuscripts intended for editorial review, only for your friends.
19: %
20: % \eqsecnum changes the way equations are numbered. Normally,
21: % equations are just numbered sequentially through the entire paper.
22: % If \eqsecnum appears in the preamble, equation numbers will
23: % be sequential through each section, and will be formatted "(sec-eqn)",
24: % where sec is the current section number and eqn is the number of the
25: % equation within that section. \eqsecnum can be used with
26: % either substyle.
27:
28: %\tighten
29: %\eqsecnum
30:
31: % Here's some slug-line data. They're never printed out by these
32: % substyles because they're only relevant to the actual publication
33: % process, and these styles aren't used in publication (yet).
34: % The receipt and acceptance dates would be filled in by the editorial
35: % staff on the appropriate dates; they are commented out in this sample
36: % so that the abstract environment prints out rules so that the dates
37: % can be typed onto the manuscript according to current practice.
38:
39: %\received{4 August 1988}
40: %\accepted{23 September 1988}
41: \journalid{337}{15 January 1989}
42: \articleid{11}{14}
43:
44: % This is the end of the "preamble". Now we wish to start with the
45: % real material for the paper, which we indicate with \begin{document}.
46: % Following the \begin{document} command is the front matter for the
47: % paper, viz., the title, author and address data, the abstract, and
48: % any keywords or subject headings that are relevant.
49:
50: %\slugcomment{Maybe ApJ Letter, maybe not}
51:
52: \begin{document}
53:
54: \title{Evidence for a Young Stellar Population in NGC~5018}
55:
56: \author{Andrew J. Leonardi\altaffilmark{1} }
57: \affil{CB \#3255, Department of Physics \& Astronomy, University of North
58: Carolina, Chapel Hill, NC 27599-3255}
59:
60: \and
61:
62: \author{Guy Worthey}
63: \affil{Department of Physics \& Astronomy, St. Ambrose University,
64: 518 W. Locust St., Davenport, IA 52803-2829}
65:
66: % Notice that each of these authors has alternate affiliations, which
67: % are identified by the \altaffilmark after each name. The actual alternate
68: % affiliation information is typeset in footnotes at the bottom of the
69: % first page, and the text itself is specified in \altaffiltext commands.
70: % There is a separate \altaffiltext for each alternate affiliation
71: % indicated above.
72:
73: \altaffiltext{1}{Visiting Astronomer, Cerro Tololo Inter-American Observatory.
74: CTIO is operated by AURA, Inc.\ under contract to the National Science
75: Foundation.}
76:
77: % The abstract environment prints out the receipt and acceptance dates
78: % if they are relevant for the journal style. For the aasms style, they
79: % will print out as horizontal rules for the editorial staff to type
80: % on, so long as the author does not include \received and \accepted
81: % commands. This should not be done, since \received and \accepted dates
82: % are not known to the author.
83:
84: \begin{abstract}
85: Two absorption line indices, Ca II and H$\delta$/$\lambda$4045, measured
86: from high-resolution spectra are used with evolutionary synthesis
87: models to verify the presence of a young stellar population
88: in NGC~5018. The derived age of this population is $\sim$ 2.8 Gyr with a
89: metallicity roughly solar and it completely
90: dominates the integrated light of the galaxy near 4000 \AA.
91: \end{abstract}
92:
93: \keywords{line: profiles --- galaxies: abundances --- galaxies: elliptical
94: and lenticular, cD --- galaxies: individual (NGC 5018) --- galaxies:
95: starburst --- galaxies: stellar content}
96:
97: % That's it for the front matter. On to the main body of the paper.
98: % We'll only put in tutorial remarks at the beginning of each section
99: % so you can see entire sections together.
100: %
101: % In the first two sections, you should notice the use of the LaTeX \cite
102: % command to identify citations. The citations are tied to the
103: % reference list via symbolic tags. We have chosen the first three
104: % characters of the first author's name plus the last two numeral of the
105: % year of publication. The corresponding reference has a \bibitem
106: % command in the reference list below.
107: %
108: % Please go to the LaTeX manual for a complete description of the
109: % \cite-\bibitem mechanism.
110:
111: \section{Introduction}
112:
113: Morphological peculiarities in the optical images of galaxies are now almost
114: invariably taken to be signs of a past tidal interaction or merger. Though
115: theoretical models have had success reproducing tidal tails, shells, and
116: other structures \citep[e.g.,][]{tt72,q84,hq88,mbr93},
117: in the absence of explicit details of the interaction such as the Hubble
118: types of the progenitors, relative progenitor sizes, and impact parameter, a
119: unique solution is difficult to come by.
120:
121: Dynamical friction during a collision almost certainly leads to a merger of the
122: stellar systems \citep{s83} and simulations have shown that any
123: accompanying gas will rapidly dissipate to the center during minor mergers involving
124: either ellipticals \citep{wh93} or disk galaxies
125: \citep{mh94a}. If the conditions are right, it is
126: reasonable to expect that this gas inflow will result in star formation. The
127: duration, intensity, and even the starting time of the star formation, however,
128: depend on the morphology of the progenitor and the details of the
129: interaction \citep{mh94b}. As much post-merger
130: information as possible is needed to help constrain the merger possibilities,
131: including analysis of the resulting young stellar population (YSP).
132:
133: Unfortunately, unless the merger remnant is presently forming stars at a
134: reasonably vigorous rate (hence producing readily apparent emission lines) or has
135: experienced a very recent and relatively strong episode of star formation placing
136: it in the starburst regime, a YSP can be difficult to detect. Broadband colors
137: and other YSP indicators return to pre-star formation levels very quickly
138: \citep{bas90,cs94} and
139: many indicators that imply the presence of a YSP can also be explained by
140: intrinsic metallicity differences in the final population of the merger
141: remnants \citep[hereafter BBB]{bbb93}. Ambiguity
142: about whether a YSP even exists or not complicates the details of the merger
143: considerably.
144:
145: In this paper, we use spectral indices in conjunction with evolutionary synthesis models to
146: detect and determine the age and metallicity of a YSP in the particular case of
147: a possible merger remnant,
148: NGC~5018. It represents an update to the age-dating technique introduced in
149: \citet[hereafter LR]{lr96}. NGC~5018 is appropriate to the
150: present discussion because there is ongoing uncertainty concerning the presence of
151: a YSP. Certain observations imply the existence of a YSP while others seem to be
152: inconsistent with the presence of a YSP. NGC~5018 and the roots of the controversy
153: are described
154: in \S2. In \S3, a review of the age-dating technique and its
155: refinements are given. The results of the technique applied to NGC~5018 are
156: presented in \S4 and \S5 contains the conclusions.
157:
158: \section{NGC~5018}
159:
160: NGC~5018 is a member of the \citet{mc83} catalog of
161: shell elliptical galaxies and is considered a probable merger remnant
162: \citep{fpcmv86}. As noted by \citet{ssfbdg90}
163: and BBB, NGC~5018 has an abnormally weak $\mathrm{Mg_{2}}$ index
164: for its luminosity: Its measured Mg$_2$ is 0.209 \citep{twfbg98}
165: even though the mean Mg$_2$-$\sigma$ relation suggests Mg$_2 = 0.301$
166: for an elliptical galaxy with NGC 5018's measured velocity dispersion
167: of $\sigma = 223$ km s$^{-1}$ \citep{bbf93}, about 6 standard
168: deviations away from the mean.
169: Although deviations from the line-strength-luminosity
170: relation correlate well with the amount of morphological disturbance in
171: elliptical galaxies \citep{ssfbdg90}, NGC~5018 has
172: abnormally weak line strengths even when this correlation is accounted for.
173: For a class of objects, these authors ruled out metallicity
174: variations in the galaxies as the cause for the correlation due to the
175: physical implausibility
176: of stronger mergers leading to more metal-poor stellar populations in the
177: remnants. \citeauthor{ssfbdg90} concluded that mergers produce a YSP which is
178: observed in the decreased line strengths. On a galaxy-by-galaxy basis,
179: however, an intrinsic metallicity variation cannot be ruled out by
180: a low $\mathrm{Mg_{2}}$ index alone. In NGC~5018, a low $\mathrm{Mg_{2}}$
181: index coupled with the lack of an upturn
182: in its UV spectral energy distribution (SED), led BBB to conclude that
183: NGC~5018 consisted of a metal-poor old stellar population, in stark
184: contrast to other ellipticals of the same luminosity.
185: BBB were unable to match both the UV SED observations and the $\mathrm{Mg_{2}}$
186: index with composite populations created by mixing spectral templates of
187: metal-rich elliptical galaxies and a contaminating YSP template. Only
188: templates containing metal-poor populations approached both observations.
189:
190: Indirect observational evidence that a YSP does in fact exist in NGC~5018 is
191: extensive. The shells present in its optical image are photometrically
192: bluer than the surrounding parts of the galaxy \citep{fpcmv86}
193: suggesting a younger age for the shells.
194: The detection of an HI gas bridge connecting NGC~5018 with the nearby spiral
195: NGC~5022 \citep{kgvjk88} is evidence of an ongoing interaction
196: while a possible past interaction is implied by a stellar
197: bridge connecting the two and the embedded dust lane in NGC~5018
198: \citep{mh97}.
199: Possible young globular cluster candidates, formed during a past interaction
200: and perhaps only several hundred Myr old, have been observed
201: \citep{hk96}.
202: Furthermore, \citet{ghjn94} measured extended H$\alpha$+[N II]
203: emission in the central region coinciding with the embedded dust lane which
204: they associated with star forming regions.
205: Also, IR emission has been detected in the same area
206: \citep{jkkg87}.
207: \citet{tb87} showed in an IR two-color diagram that
208: NGC~5018 lies in a region quite different
209: from that occupied by infrared ``cirrus'', which is emission from diffuse
210: dust in the interstellar medium of a galaxy. Instead, it is closer to
211: the region where the IR emission from warmer dust associated with HII
212: regions dominates \citep{h86,blw88}
213: , suggesting a YSP source for the IR emission.
214:
215: While the observations are compelling, they are not conclusive. The
216: direct detection of the YSP is needed to resolve the issue. BBB chose
217: to observe in the far-UV for exactly that reason, since in principle,
218: this region of the spectrum is dominated by young stars
219: \citep{o88} and the low UV flux level in NGC~5018 led
220: them to the metal-poor scenario. They discounted dust obscuration as
221: the cause because the best available photometry at that time
222: \citep{fpcmv86} showed that the dust lane in NGC~5018
223: does not extend into the region where their IUE spectrum was taken.
224: Subsequent observations, however \citep{cd94,ghjn94}
225: , indicate that not only is dust
226: present throughout the central region but also is patchy in nature,
227: making the reddening effects difficult to ascertain.
228: \citet{cd94} showed that reasonable
229: expectations of dust obscuration and a YSP can explain both the low
230: $\mathrm{Mg_{2}}$ index and the UV flux depletion in the central
231: region of NGC~5018. Both sets of authors concluded that a YSP was a
232: more probable explanation for the observations.
233:
234: The conflict is
235: illustrated in Figure~\ref{fig:lick} where we have plotted data from
236: the Lick group \citep{twfbg98} for NGC~5018 and other
237: systems. The top panel shows a $\lambda 2750-V$ color plotted against
238: the $\mathrm{C_{2}}$4668 Lick index. The $\mathrm{C_{2}}$4668 index is
239: more metal-sensitive than Mg$_2$ and is less subject, but not immune,
240: to abundance ratio effects. BBB remarked that NGC~5018's
241: $\mathrm{Mg_{2}}$ index and UV spectrum resembles M32's, even though
242: M32 is a much less luminous galaxy, which led them to surmise a system
243: with an abundance like that of a dwarf galaxy rather than a giant
244: elliptical. The UV data here supports this view. As can be seen in
245: the top panel of Figure~\ref{fig:lick}, NGC~5018 has approximately the
246: same UV color but a slightly weaker $\mathrm{C_{2}}$$\lambda$4668 index than M32
247: indicating that NGC~5018 is about twice as old and more metal-poor
248: than M32. The bottom panel of Figure~\ref{fig:lick} plots H$\beta$
249: (uncorrected for any emission fill-in) against $\mathrm{C_{2}}$4668
250: and shows NGC~5018 nearly on the same model grid line as M32 implying
251: similar age, but still $\sim$ 0.15 dex more metal-poor than M32.
252: Emission corrections for H$\beta$ would push NGC~5018 vertically
253: upward to younger ages and higher metallicity, worsening the
254: discrepancy in age between the two panels. On the other hand, a
255: correction for UV extinction in the upper panel would make the two
256: panels agree better.
257:
258: To help resolve the still uncertain nature of NGC~5018, we utilize
259: spectroscopic observations, along with an updated age-dating method to
260: unambiguously show that a YSP is present in NGC~5018.
261: The modeling technique is discussed in \S3. The
262: long-slit spectra of NGC~5018 were acquired at the KPNO 4m telescope in June~1995 by
263: Lewis Jones and kindly provided to us. Four 30-minute exposures were acquired
264: with the R-C spectrograph with grating KPC-22B at second order and the T2KB
265: 2048x2048 CCD. The slit width was 2 arcsec. A 14.5 pixel aperture was extracted
266: from the raw spectrum and with a CCD spatial scale of $\sim$ 0.69 arcsec/pixel,
267: the aperture size on NGC~5018 is 2''x10''. The dispersion of the spectra is
268: 0.7 \AA/pxl and the resolution
269: is FWHM $\sim$ 1.8 \AA. Data reductions were done in IRAF. For details see
270: \citet{j99}. A representative spectrum, emphasizing the wavelength
271: region of interest, is shown in Figure~\ref{fig:spec}. It has been normalized
272: to unity at 4040 \AA.
273:
274: \section{The Age-Dating Technique}
275:
276: The age-dating technique as described in LR uses two spectral indices developed
277: in \citet{r84, r85}. Each index is defined by taking the ratio
278: of counts in the bottoms of two neighboring absorption lines without reference
279: to the continuum levels. The specific absorption lines used are identified on the
280: NGC~5018 spectrum in Figure~\ref{fig:spec}. The first index, H$\delta$/$\lambda$4045,
281: measures
282: the integrated spectral type of a galactic stellar population and is produced
283: from the ratio of the central intensity in H$\delta$ relative to the central
284: intensity in the neighboring
285: Fe I $\lambda$4045 line. Note that the way the index is defined,
286: H$\delta$/$\lambda$4045 \emph{decreases} as H$\delta$ gets stronger. The second
287: index, Ca II, formed from the ratio of
288: the central intensity of Ca II H+H$\epsilon$ relative to Ca II K, is constant in
289: stars with a spectral type later than F2 but then decreases dramatically for
290: earlier type stars, reaching a minimum at spectral type A0. It provides an
291: unambiguous signature for stars hotter than F2 in the integrated light of a
292: stellar population.
293:
294: The indices are computed for single-age theoretical populations from evolutionary
295: synthesis models. When plotted together in the two-dimensional index space, the
296: indices resolve the well-known degeneracy between a YSP's age and light
297: contribution to a composite stellar population \citep[e.g.,][LR]{cs87,bas90,cs94}.
298: A valuable property of these indices particularly applicable to
299: NGC~5018 is their virtual insensitivity to reddening. Since only neighboring
300: spectral features are used, reddening does not affect their values and the
301: embedded dust in the center of NGC~5018 will not obscure the evidence of a
302: YSP.
303:
304: In LR, the evolutionary synthesis models of \citet{bc93}
305: were used with the spectral library updated to include the higher resolution
306: stellar library of \citet{jhc84}. The models
307: were restricted to solar abundance populations only, thus metallicity effects on the
308: indices could only be explored crudely. To remedy this, the technique now employs
309: the evolutionary synthesis models of \citet{w94}.
310:
311: The Worthey models work as follows: For a given age and metallicity, a theoretical
312: isochrone \citep{bbcfn94} is consulted, with each point on
313: the isochrone representing a parcel of stars of known luminosity, temperature,
314: and gravity. The spectral indices for that isochrone point are interpolated
315: from empirical fitting functions of the indices from a high-resolution spectral
316: library \citep[see also \citealt{lea96}]{j99} that has been
317: smoothed to the resolution of the
318: NGC~5018 spectra. The indices are weighted by luminosity and
319: number and added up along the isochrone to get the spectral index values for the
320: entire integrated population. The population is formed from an instantaneous
321: burst of star formation. A finite burst, however, is more realistic and will become
322: a future feature of the models. These models were originally designed to disentangle
323: age and metallicity effects in the integrated light of old stellar populations. To
324: extend the age coverage to include very young ages ($<$ 1 Gyr) and also extend spectral
325: coverage to the blue, the empirical library was augmented with 2103
326: theoretical stellar spectra computed with the \citet{k95} SYNTHE
327: program. The indices were computed for each synthetic spectrum and the values were
328: used as a lookup table with specific isochrone points being found by interpolating
329: between the synthetic grid points \citep[for more details see][]{l00}.
330:
331: Figure~\ref{fig:burst} shows the Ca II index plotted against the
332: H$\delta$/$\lambda$4045 index
333: for two stellar populations with [Fe/H] = -0.7 and [Fe/H] =
334: 0.0 respectively.
335: In the figure, the solid squares represent the index values for an instantaneous
336: burst of star formation that has evolved to the labeled age in Gyr, so each curve
337: follows the evolution of the indices for a stellar population of
338: the given metallicity. Both indices initially decrease for young systems as they
339: age, reaching a minimum at about 0.25--0.5 Gyr as the O and B stars die
340: out and A stars begin to dominate the integrated light, thus generating strong
341: Balmer lines. Subsequently, as the Balmer lines weaken, the indices increase
342: again. Also plotted in Figure~\ref{fig:burst} are
343: the index values observed
344: for a select Galactic globular cluster, 47 Tuc ([Fe/H] = -0.7).
345: The long-slit observation was acquired at the CTIO 1.5m telescope in November
346: 1995. A 10 minute exposure of 47~Tuc was acquired using the Cassegrain spectrograph
347: with the Loral 1200x800 CCD and the B\&L grating \#58 at second order. During the
348: exposure, the slit was trailed across the core diameter of the cluster to obtain
349: a true integrated light
350: measurement. The dispersion of the spectrum is 1.12 \AA/pixel and the resolution is
351: FWHM $\sim$ 2.6 \AA. Data reductions were done in IRAF. For details see
352: \citet{l00}. Although the Ca II index loses much of its
353: age discriminating power at older ages, we still determine a reasonable
354: globular cluster age of approximately~15 Gyr for 47~Tuc at the appropriate
355: metallicity. Simultaneous age and metallicity discrimination with high S/N spectra is
356: most effective between
357: the ages of 0.25 Gyr and 4 Gyr. Since the Ca II and H$\delta$/$\lambda$4045
358: indices are insensitive to metallicity for ages less than $\sim$ 0.25 Gyr, the ability to
359: uniquely determine a metallicity is lost whereas for ages greater than $\sim$ 4.0 Gyr, the
360: Ca II index approaches the constant value for late type stars and its evolution essentially
361: halts.
362:
363: Unfortunately, the isochrones used here only model the horizontal branch as a red
364: clump.
365: To illustrate where such a population with a blue horizontal branch would fall on the
366: Ca II--H$\delta$/$\lambda$4045 diagram, also included
367: on Figure~\ref{fig:burst} are the index values for a 20 minute
368: exposure of the Galactic globular cluster M15 taken during the same observing
369: run as 47 Tuc and for a very metal-poor ([Fe/H] = -1.7) 15.1 Gyr model with
370: a red clump (note that both the models in
371: Figure~\ref{fig:burst} and the globular cluster spectra
372: have been smoothed out to the resolution and intrinsic velocity dispersion of
373: the NGC~5018 spectra described
374: above). While~47 Tuc has a predominantly red horizontal branch, M15 has a blue
375: horizontal branch \citep{l89}. The agreement between the [Fe/H] = -1.7
376: model and M15
377: is poor. Blue horizontal branch stars are luminous enough to contribute significantly
378: the Ca II
379: index and drive it below what the models predict for this age and metallicity.
380:
381: \section{Results}
382:
383: \subsection{Composite Populations}
384:
385: The question we must answer for NGC~5018 is whether a YSP is contaminating the
386: light of an old, underlying population or the light is originating from a solely
387: old, metal-poor population. To create a composite population, we assume that the
388: old population has an age of 15 Gyr but the metallicity can vary. We then
389: interpolate in the index space between the old population point and a YSP point in
390: constant increments to represent different levels of contamination by the YSP.
391: The flux
392: contributions for the two populations have been normalized at 4040 \AA.
393: The computed indices for the theoretical composite populations are illustrated
394: in Figures~\ref{fig:yspm17}--\ref{fig:yspp00}. In each figure, the full
395: evolution of the YSP is plotted in a similar manner to Figure~\ref{fig:burst},
396: one YSP metallicity per figure. A solid, colored circle represents the index
397: values for the~15 Gyr population denoting the old underlying stellar
398: system, each color a different old population metallicity which may or may
399: not be the same as the YSP's metallicity. The lines connecting the old
400: population points with select ages along the YSP evolution curve represent
401: composite stellar populations. The crosses along each line are interpolations
402: between these two populations in 25\% increments
403: of the contribution of the YSP to the
404: integrated light near 4000 \AA. The YSP points used for interpolation were chosen
405: so that the resulting composite population would come as closely as possible to
406: NGC~5018.
407:
408: The mean indices for the four spectra of NGC~5018 are plotted as an open triangle
409: with error bars in Figures~\ref{fig:yspm17}--\ref{fig:yspp00}. Error bars for the
410: NGC~5018 data points were calculated by computing the rms scatter among the four
411: observations. The model trajectories were computed after both the empirical and the
412: synthetic spectral libraries
413: had been smoothed with gaussians to match the resolution
414: and intrinsic velocity dispersion of the NGC~5018 spectra which allows comparison between
415: the theoretical integrated indices and those of NGC~5018.
416: The observed indices for~47 Tuc and a spectrum of M32, obtained during the same observing
417: run as the globular cluster spectra, are also plotted in
418: Figures~\ref{fig:yspm17}--\ref{fig:yspp00} for comparison purposes.
419:
420: \subsection{Index Plots}
421:
422: Figure \ref{fig:yspm17} shows the index values for a very metal-poor
423: YSP ([Fe/H] = -1.7) mixed with two possible old, underlying
424: populations. Figure~\ref{fig:yspm07} does the same for a moderately
425: metal-poor YSP ([Fe/H] = -0.7). If we also allow the old population
426: to be metal-poor (red symbols), we have an extreme version of the case
427: put forth by BBB of a strictly metal-poor population. If NGC~5018 were to
428: have such a
429: population and yet still contain the large amount of structure
430: attributed to the galaxy by \citet{ssfbdg90}, BBB
431: postulated that NGC~5018 would either have to be the result of the merging of many
432: metal-poor components or a large metal-poor elliptical
433: that experienced a recent merger. The figures show quite definitively
434: that an old stellar population with a globular cluster-like metallicity of -1.7 combined
435: with either the very metal-poor YSP or the moderately metal-poor YSP is
436: disallowed. Even by choosing
437: an age of 13.2 Gyr for the ``young" population to approach NGC~5018
438: as close as possible, the set of allowed indices
439: defined by the possible composite populations are not near the location of
440: NGC~5018 in the figures. No combination
441: of metal-poor young and old populations nor a single coeval metal-poor
442: population can reproduce the observed indices.
443:
444: If a metal-poor YSP is mixed with a metal-rich old population, we have the
445: situation depicted with the blue symbols in Figures~\ref{fig:yspm17}
446: and~\ref{fig:yspm07}. In this scenario, NGC~5018 could
447: have evolved as a normal elliptical galaxy but then interacted with a
448: young, metal-poor disk galaxy. Dust obscuration as suggested by
449: \citet{cd94} would still be needed to explain the
450: lack of an upturn in the UV SED. In any case, this population mixture
451: is disallowed as well, in agreement with BBB. If a large percentage
452: of the light is originating in the old population, the composite has a
453: similar Ca II index value as NGC~5018 but its H$\delta$/$\lambda$4045
454: index is too weak compared to the dominant population in NGC~5018.
455:
456: It is only when we allow the YSP to be metal-rich, i.e., solar or greater,
457: that the model indices match the observed values for NGC~5018.
458: Figure~\ref{fig:yspp00full} shows the solar YSP curve from Figure~\ref{fig:burst}
459: again along with four old population points, two metal-poor (green and red
460: symbols) and two metal-rich (blue and magenta symbols). For clarity, the
461: interpolation curves between the old population points and the YSP curve have
462: been omitted from Figure~\ref{fig:yspp00full} and are instead shown on
463: Figure~\ref{fig:yspp00} which is identical to Figure~\ref{fig:yspp00full} but
464: on an expanded scale. We derive an age of $\sim$ 2.8 Gyr for the YSP in
465: NGC~5018. Also it is interesting to note in Figure~\ref{fig:yspp00} that the
466: light at 4000 \AA\ is completely dominated by the YSP with virtually 100\% of the
467: light originating from the YSP regardless of the assumed old population metallicity.
468: In fact, for an old population with [Fe/H] = -1.7, \emph{any} contribution
469: to the light by the old population will drive the model indices away from
470: NGC~5018. Even in the most likely scenario, a metal-rich old population with a
471: metal-rich YSP, the old population does not contribute to the integrated light.
472: In this picture, NGC~5018 can evolve as a normal elliptical galaxy and then
473: interact with a metal-rich companion requiring no unusual events to create the stellar
474: population of NGC~5018.
475: \emph{Thus if an underlying old metal-poor population is present
476: in NGC~5018, it cannot be contributing significantly to the integrated blue
477: light.}
478:
479: We conclude on the basis of these results that there is a young stellar
480: population in the central regions of NGC~5018. We infer that the age of
481: this YSP is on the order of 2.8 Gyr, the metallicity is near
482: solar, and it is providing virtually all of the light at 4000 \AA. These
483: determinations are reddening independent due to the nature of the spectral
484: indices used.
485:
486: \section{Discussion}
487:
488: The two observations of BBB which led them to conclude that a YSP is
489: not present in NGC~5018 are a low $\mathrm{Mg_{2}}$ index coupled with
490: the lack of an upturn in the UV SED both comparable to that found in M32. As seen
491: in Figures~\ref{fig:yspm17}--\ref{fig:yspp00}, however, NGC~5018 lies in a
492: different region of the H$\delta$/$\lambda$4045--Ca II diagram than M32 hence
493: it cannot be just a high luminosity version of M32. With the patchiness of
494: the dust in the central regions of NGC~5018, there may be as much as two
495: magnitudes of extinction in the far-UV part of the spectrum which
496: could explain the lack of an upturn
497: \citep{cd94}. The results
498: derived here, however, indicate that large amounts of dust obscuration
499: need not be invoked, although a modest amount of extinction would
500: bring Figure \ref{fig:lick} into better harmony with the spectral
501: index results. The YSP age of~2.8 Gyr derived here is reddening
502: insensitive and thus robust regardless of the structure of the dust
503: distribution. Also, as Figure~\ref{fig:yspp00} shows, the YSP is
504: completely dominating the visible part of the spectrum of NGC~5018. A
505: YSP of this age does not have an upturn in the UV part of the spectrum.
506: Hence the small far-UV upturn in NGC~5018 may simply reflect the
507: characteristics of the 2.8-Gyr-old stellar population
508: dominating the light.
509:
510: The models predict that a YSP of this age is about 6 times brighter (at
511: 4000\AA ) than an ancient population of the same metallicity and IMF. Hence if
512: the relative contributions from the old and young populations were only half
513: and half, then the galaxy must be 6 parts old to 1 part
514: young in the central region. This would qualify as a gas-rich
515: merger event with a rapid gas inflow to the center \citep{wh93}
516: of NGC 5018. If the light contribution is
517: weighted even more heavily toward the young population as we suggest, this
518: implies a more violent event, perhaps a merger of two spiral
519: galaxies or some other event involving large amounts of gas consumed in
520: star formation, with an extremely high inflow to the nucleus. Note that,
521: if nebular emission is filling in the Balmer lines, the derived age
522: decreases, the YSP is brighter, and so the size of the merging event
523: decreases.
524:
525: A solar-[Fe/H] population of age 2.8 Gyr has an Mg$_2 \approx 0.20 $ mag,
526: as is observed in NGC~5018. Aging this population to 15 Gyr raises the
527: Mg$_2$ to 0.27 mag, still about 2 standard deviations from the mean
528: relation for elliptical galaxies, though a dust-hidden YSP could be diluting
529: the Mg$_2$ index \citep{cd94}. We note, however,
530: that compared to
531: M32, NGC~5018's C$_2$4668 index is weaker but its Mg$_2$ index is
532: stronger, implying that NGC~5018 may also participate in the general
533: trend for large ellipticals to have enhanced Mg abundance
534: \citep[e.g.,][]{w98}. This will further increase NGC~5018's Mg$_2$ line
535: strength as it ages so that it may one day fall among other
536: ellipticals in the Mg$_2$-$\sigma$ relation.
537:
538: As mentioned previously, significant emission has been observed in the
539: center of NGC~5018 \citep{ghjn94}. Since both of
540: the spectral indices we used depend on the intensity in the center of
541: a Balmer line, contamination by emission could affect the results
542: significantly even though the contamination decreases as one moves
543: toward higher order lines in the Balmer sequence. Specifically,
544: emission will weaken H$\delta$ relative to Fe I $\lambda$4045 and to a
545: lesser degree weaken Ca II H+H$\epsilon$ relative to Ca II K. Each
546: index, if contaminated, will therefore have a higher value than if
547: there were no emission. From
548: Figures~\ref{fig:yspm17}--\ref{fig:yspp00}, it can be seen that if
549: emission contamination were removed, the data points for NGC~5018
550: would shift toward younger ages. The determined YSP age of 2.8
551: Gyr represents therefore an \emph{upper limit on the YSP age}. The limited
552: spectral coverage does not permit a more detailed analysis of the
553: emission contamination, but the fact that H$\beta$ from Figure
554: \ref{fig:lick} gives the same age indicates that the emission must be
555: relatively modest.
556:
557: The power of the method used in this paper is its ability to discriminate
558: between different, plausible stellar populations. The verification of the
559: presence of a YSP in
560: NGC~5018 and its age of~2.8 Gyr are quite unambiguous, irrespective of the nature
561: of the old, underlying population. Also, this type of determination is not
562: unique to NGC~5018. A similar procedure can be applied to any galaxy with
563: morphological peculiarities suspected of harboring more than one coeval stellar
564: population. With this tool, the effects of a dynamical interaction on the
565: integrated light of a galaxy can be analyzed more fully.
566:
567: \acknowledgments
568:
569: AJL would like to thank Dr.\ Jim Rose for a lot of guidance and many helpful
570: discussions, Lewis Jones for providing the spectra of NGC 5018 and the referee,
571: Dr.\ Scott Trager, for many helpful suggestions which improved the paper. This
572: research was partially supported by NSF grant AST-9320723 to the University of
573: North Carolina and by NASA through grant GO-06664.01-95A awarded by the Space
574: Telescope Science Institute which is operated by the Association of
575: Universities for Research in Astronomy, Inc., for NASA under Contract
576: NAS5-26555.
577:
578: \clearpage
579:
580: % Now comes the reference list. In this document, we used \cite to call
581: % out citations, so we must use \bibitem in the reference list, which
582: % means we use the LaTeX thebibliography environment. Please note that
583: % \begin{thebibliography} is followed by a null argument. If you forget
584: % this, mayhem ensues, and LaTeX will say "Perhaps a missing item?" when
585: % you run it. Do not call us, do not send mail when this happens. Put
586: % the silly {} after the \begin{thebibliography}.
587: %
588: % Each reference has a \bibitem command to define the citation format
589: % and the symbolic tag, as well as a \bibitem[]{} command which sets up
590: % formatting parameters for the reference list itself.
591: %
592: % If we had not bothered with the \cite-\bibitem business, calling out
593: % the references outselves, the reference list could be enclosed in
594: % a references environment (\begin{references} has no null argument),
595: % and there would be no need for the leading \bibitem's.
596:
597: \begin{thebibliography}{}
598: \bibitem[Bender, Burstein, \& Faber(1993)]{bbf93} Bender, R., Burstein, D., \& Faber, S. M. 1993, \apj, 411, 153
599: \bibitem[Bertelli et al.(1994)]{bbcfn94} Bertelli, G., Bressan, A., Chiosi, C., Fagotto, F., \&
600: Nasi, E., 1994, \aaps, 106, 275
601: \bibitem[Bertola, Burstein, \& Buson(1993)]{bbb93} Bertola, F., Burstein, D., \& Buson, L. M., 1993, \apj,
602: 403, 573 (BBB)
603: \bibitem[Bica, Alloin, \& Schmidt(1990)]{bas90} Bica, E., Alloin, D., \& Schmidt, A., 1990, \aap, 228, 23
604: \bibitem[Bruzual \& Charlot(1993)]{bc93} Bruzual A., G., \& Charlot, S., 1993, \apj, 405, 538
605: \bibitem[Burstein et al.(1988)]{bbbfl88} Burstein, D., Bertola, F., Buson, L. M., Faber, S. M., \& Lauer, T. R.
606: 1988, \apj, 328, 440
607: \bibitem[Bushouse, Lamb, \& Werner(1988)]{blw88} Bushouse, H. A., Lamb, S. A., \& Werner, M. W., 1988, \apj, 335, 74
608: \bibitem[Carollo \& Danziger(1994)]{cd94} Carollo, C. M., \& Danziger, I. J., 1994, \mnras, 270, 743
609: \bibitem[Charlot \& Silk(1994)]{cs94} Charlot, S., \& Silk, J., 1994, \apj, 432, 453
610: \bibitem[Couch \& Sharples(1987)]{cs87} Couch, W. J., \& Sharples, R. M., 1987, \mnras, 229, 423
611: \bibitem[Fort et al.(1986)]{fpcmv86} Fort, B. P., Prieur, J.-L., Carter, D., Meatheringham, S. J.,
612: \& Vigroux, L., 1986, \apj, 306, 110
613: \bibitem[Goudfrooij et al.(1994)]{ghjn94} Goudfrooij, P., Hansen, L., J\mbox{\o}rgensen, H. E., \&
614: N\mbox{\o}rgaard-Nielsen, H. U., 1994, \aaps, 105, 341
615: \bibitem[Helou(1986)]{h86} Helou, G., 1986, \apj, 311, L33
616: \bibitem[Hernquist \& Quinn(1988)]{hq88} Hernquist, L., \& Quinn, P. J., 1988, \apj, 331, 682
617: \bibitem[Hilker \& Kissler-Patig(1996)]{hk96} Hilker, M., \& Kissler-Patig, M., 1996, \aap, 314, 357
618: \bibitem[Jacoby, Hunter, \& Christian(1984)]{jhc84} Jacoby, G. H., Hunter, D. A., \& Christian, C. A., 1984, \apjs,
619: 56, 257
620: \bibitem[Jones(1999)]{j99} Jones, L. A., 1999, Ph.D. thesis Univ. of North Carolina
621: \bibitem[Jura et al.(1987)]{jkkg87} Jura, M., Kim, D.-W., Knapp, G. P., \& Guhathakurta, P., 1987, \apj,
622: 312, L11
623: \bibitem[Kim et al.(1988)]{kgvjk88} Kim, D.-W., Guhathakurta, P., van Gorkom, J. H., Jura, M.,
624: \& Knapp, G. R., 1988, \apj, 330, 684
625: \bibitem[Kurucz(1995)]{k95} Kurucz, R. L., 1995, private communication
626: \bibitem[Lee(1989)]{l89} Lee, Y.-W., 1989, Ph.D. thesis Yale University
627: \bibitem[Leitherer et al.(1996)]{lea96} Leitherer, C., et al., 1996, \pasp, 108, 996
628: \bibitem[Leonardi(2000)]{l00} Leonardi, A. J., 2000, in preparation
629: \bibitem[Leonardi \& Rose(1996)]{lr96} Leonardi, A. J., \& Rose, J. A., 1996, \aj, 111, 182 (LR)
630: \bibitem[Malin \& Carter(1983)]{mc83} Malin, D. F., \& Carter, D., 1983, \apj, 274, 534
631: \bibitem[Malin \& Hadley(1997)]{mh97} Malin, D. F., \& Hadley, B., 1997, Pub. Astro. Soc. of Aus., 14, 52
632: \bibitem[Mihos, Bothun, \& Richstone(1993)]{mbr93} Mihos, J. C., Bothun, G. D., \& Richstone, D. O., 1993, \apj,
633: 418, 82
634: \bibitem[Mihos \& Hernquist(1994a)]{mh94a} Mihos, J. C., \& Hernquist, L., 1994a, \apj, 425, L13
635: \bibitem[Mihos \& Hernquist(1994b)]{mh94b} Mihos, J. C., \& Hernquist, L., 1994b, \apj, 431, L9
636: \bibitem[O'Connell(1988)]{o88} O'Connell, R. W., 1988, in Towards Understanding Galaxies at High
637: Redshift, ed. R. G. Kron \& A. Renzini (Dordrecht: Kluwer), 177
638: \bibitem[Quinn(1984)]{q84} Quinn, P. J., 1984, \apj, 279, 596
639: \bibitem[Rose(1984)]{r84} Rose, J. A., 1984, \aj, 89, 1238
640: \bibitem[Rose(1985)]{r85} Rose, J. A., 1985, \aj, 90, 1927
641: \bibitem[Schweizer(1983)]{s83} Schweizer, F., 1983, in Proc. IAU Symp. 100, Internal Kinematics
642: and Dynamics of Galaxies, ed. E. Athanassoula (Dordrecht: Reidel), 319
643: \bibitem[Schweizer et al.(1990)]{ssfbdg90} Schweizer, F., Seitzer, P., Faber, S. M., Burstein, D.,
644: Dalle Ore, C. M., \& Gonzalez, J. J., 1990, \apj, 364, L33
645: \bibitem[Thronson \& Bally(1987)]{tb87} Thronson, H. A., \& Bally, J., 1987, \apj, 319, L63
646: \bibitem[Toomre \& Toomre(1972)]{tt72} Toomre, A., \& Toomre, J., 1972, \apj, 178, 623
647: \bibitem[Trager et al.(1998)]{twfbg98} Trager, S. C., Worthey, G., Faber, S. M., Burstein, D.,
648: \& Gonz\'alez, J. J. 1998, \apjs, 116, 1
649: \bibitem[Weil \& Hernquist(1993)]{wh93} Weil, M. L., \& Hernquist, L., 1993, \apj, 405, 142
650: \bibitem[Worthey(1994)]{w94} Worthey, G., 1994, \apjs, 95, 107
651: \bibitem[Worthey(1998)]{w98} Worthey, G., 1998, \pasp, 110, 888
652: \bibitem[Worthey, Dorman, \& Jones(1996)]{wdj96} Worthey, G., Dorman, B., \& Jones, L. A. 1996, \aj, 112, 948
653:
654: \end{thebibliography}
655:
656: % Finally, we have figure captions. Usually these must be on a separate
657: % page, although unlike table, it is often permissible to have several
658: % figure captions on the same page. We force the page break between
659: % the reference list and the figure captions.
660: %
661: % The \caption command in the figure environment works like the one in the
662: % table environment (it's the same one, actually), except that this one
663: % produces identification text that reads "Figure N."
664:
665: \clearpage
666:
667: \begin{figure}
668: \caption{Two age-sensitive indices, a UV-visual color and an
669: H$\beta$ index, are plotted versus metal-sensitive spectral index
670: C$_2$4668. The color $\lambda 2750 - V = {\rm log} F_{\lambda 2750} -
671: {\rm log} F_V$ following Burstein et al. 1988 and BBB, and has units of
672: dex rather than magnitudes. The marked dust screen extinction also
673: refers to a straight logarithm, that is, $A_V = 0.1$ dex, or 0.25 mag.
674: The $\lambda 2750$ passband is an average of BBB and Burstein et al.
675: data from $\lambda 2650$ to $\lambda 2850$ with their extinction
676: corrections applied. Optical spectral indices on the Lick system come
677: from Trager et al. (1998). The optical data was taken from a
678: significantly smaller aperture than the IUE UV data, so the C$_2$4668
679: index should be weakened somewhat (a few tenths of \AA ) to compensate
680: for this mismatch in the upper panel. Error bars in bold refer to M32
681: and M31 data, wider bars to the rest of the sample. Models from Worthey
682: (1994) appear as a grid labeled by scaled solar metallicity and age in
683: Gyr. Models are scaled solar and thus do not track element
684: overenhancements. Such an effect probably exists for the C$_2$4668
685: index; the larger galaxies have a C$_2$4668 index stronger than their
686: average abundance would dictate. In the upper panel, an arrow attached
687: to M31 is an estimate of the effect of subtracting the hot star
688: component (which is not included in the models) from the spectrum, as
689: modeled by Worthey et al. (1996). The indicated effect is an
690: underestimate because Worthey et al. maximized the contribution from
691: very hot PAGB stars. Similar corrections should be made for NGC 3115 and
692: NGC 3379, but we lack appropriate spectrophotometric data. NGC 3115 has
693: a UV upturn similar to M31, but NGC 3379 has considerably more UV flux,
694: so its correction is likely to be quite large. The other galaxies have
695: very weak UV emission so that their corrections are negligible. Large
696: corrections for emission fill-in in the H$\beta$ index are unlikely for
697: the galaxies plotted. If corrections are applied, galaxies move
698: vertically to younger ages and somewhat higher abundance.}
699: \label{fig:lick}
700: \end{figure}
701:
702: \begin{figure}
703: \caption{One of the long-slit spectra of NGC~5018 used to calculate the mean
704: spectral indices.
705: Each spectra was smoothed with a gaussian of 0.5 pixels. The absorption features
706: used in the age-dating technique are identified.}
707: \label{fig:spec}
708: \end{figure}
709:
710: \begin{figure}
711: \caption{The CaII index is plotted vs the H$\delta$/$\lambda$4045 index for an
712: instantaneous burst of star formation with [Fe/H] = -0.7 (red solid and
713: short-dashed line) and [Fe/H] = 0.0 (blue long-dashed line). The colored
714: lines represent the evolution of the index values as the population ages. Various ages
715: (in Gyr) have been marked by solid squares. For clarity, the path from
716: 0.004 Gyr to 0.5 Gyr on the [Fe/H] = -0.7 curve has been marked with a
717: short-dashed line while the evolution subsequent to 0.5 Gyr has been marked
718: with a solid line. An old, metal-poor model (green circle) is plotted to show the
719: effects of horizontal branch morphology on the indices (see text). The error bars
720: for the labeled globular
721: cluster points (open triangles) are smaller than the plotting symbol.} \label{fig:burst}
722: \end{figure}
723:
724: \begin{figure}
725: \caption{Ca II plotted vs H$\delta$/$\lambda$4045 for a composite population
726: consisting of a metal-poor young population ([Fe/H] = -1.7; evolution
727: marked by the solid black curve, with specific ages labeled by the solid squares)
728: and one of two OSP (age = 15.1 Gyr) populations ([Fe/H] = -1.7, marked by the
729: solid red circle
730: and [Fe/H] = 0.0, marked by the solid blue circle).
731: Dotted lines between the old population point and the 13.2 Gyr YSP point denote
732: interpolations between these two populations with each cross along the
733: OSP [Fe/H] = 0.0 curve marked with the fractional contribution of the YSP
734: to the total integrated light at 4000 \AA. The crosses along the OSP [Fe/H] = -1.7
735: curve are similar but have been left off for clarity. The indices for NGC~5018,
736: M32, and 47 Tuc are plotted as open triangles.} \label{fig:yspm17}
737: \end{figure}
738:
739: \begin{figure}
740: \caption{Same as Fig. \ref{fig:yspm17} but with the YSP [Fe/H] = -0.7.}
741: \label{fig:yspm07}
742: \end{figure}
743: \begin{figure}
744: \caption{Same as Fig. \ref{fig:yspm17} but with the YSP [Fe/H] = 0.0. Also,
745: two other old populations have been added: [Fe/H] = -0.7 (green solid circle)
746: and [Fe/H] = +0.4 (magenta solid circle). For clarity, the interpolations between
747: the OSPs and the YSP have been omitted and are shown in Fig. \ref{fig:yspp00} on
748: an expanded scale.} \label{fig:yspp00full}
749: \end{figure}
750:
751: \begin{figure}
752: \caption{Identical to Fig. \ref{fig:yspp00full} but with an expanded scale on the
753: Ca II axis. Also, the interpolations between the OSP points and the YSP points
754: nearest NGC~5018 are shown.
755: The increment crosses along the
756: [Fe/H] = -0.7 and [Fe/H] = +0.0 curves have been omitted for clarity.}
757: \label{fig:yspp00}
758: \end{figure}
759:
760:
761: % That's all, folks.
762: %
763: % The technique of segregating major semantic components of the document
764: % within "environments" is a very good one, but you as an author have to
765: % come up with a way of making sure each \begin{whatzit} has a corresponding
766: % \end{whatzit}. If you miss one, LaTeX will probably complain a great
767: % deal during the composition of the document. Occasionally, you get away
768: % with it right up to the \end{document}, in which case, you will see
769: % "\begin{whatzit} ended by \end{document}".
770:
771: \end{document}
772:
773: