1: \documentclass[12pt,preprint]{aastex}
2:
3: \shorttitle{AM-1 and Pal~14}
4: \shortauthors{Dotter, Sarajedini, \& Yang}
5: \slugcomment{Accepted by the Astronomical Journal}
6: \newcommand{\rgc}{\mathrm{R_{GC}}}
7:
8: \begin{document}
9:
10: \title{Globular Clusters in the Outer Galactic Halo: AM-1 and Palomar~14
11: \thanks{Based on observations with the NASA/ESA {\it Hubble Space Telescope},
12: obtained at the Space Telescope Science Institute, which is operated
13: by AURA, Inc., under NASA contract NAS 5-26555, under program
14: GO-6512.}}
15:
16: \author{Aaron Dotter}
17: \affil{Department of Physics and Astronomy, Dartmouth College, 6127 Wilder Laboratory,
18: Hanover, NH 03755}
19:
20: \author{Ata Sarajedini and Soung-Chul Yang}
21: \affil{Department of Astronomy, University of Florida, 211 Bryant Space Science Center, Gainesville, FL 32611}
22:
23: \begin{abstract}
24: AM-1, at $\sim$120 kpc, and Pal~14, at $\sim$70 kpc, are two of the most distant
25: Galactic globular clusters known. We present {\it Hubble Space Telescope} WFPC2
26: photometry of AM-1 and Pal~14 that reveals unprecedented depth and detail in the
27: color-magnitude diagrams of these two clusters. Absolute and relative age
28: measurements confirm that both are younger than the inner halo globular cluster
29: M~3 by 1.5--2 Gyr assuming all three clusters have similar compositions.
30: Thus AM-1 and Pal~14 join Pal~3, Pal~4, and Eridanus (studied by Stetson
31: et al.) as distant Galactic globular clusters with red horizontal branches
32: and young ages relative to the inner halo. Within the context of the
33: entire body of research on the ages of second parameter
34: globular clusters, the observed correlation between age and horizontal branch
35: morphology suggests that age is the best candidate for the second parameter.
36: However, this conclusion is tempered by the lack of precise
37: chemical abundance determinations for a significant fraction of second parameter
38: globular clusters.
39: \end{abstract}
40:
41: \keywords{globular clusters: individual --- AM-1, Palomar~14, M~3}
42:
43: \section{Introduction}
44: The second parameter phenomenon, a long-standing problem in Galactic astronomy,
45: recognizes that the horizontal branch (HB) morphology of Galactic globular
46: clusters (GCs) cannot be explained by the first parameter, metallicity, alone:
47: at least one more factor, a `second parameter,' must be involved.
48:
49: In a landmark study, \citet{sz78} demonstrated that GCs inside the solar
50: circle (which they deemed the inner halo) exhibit a tight correlation
51: between [Fe/H] and HB morphology while those that lie outside the solar circle
52: (the outer halo) show no such well-defined trend. This in turn led \citet{sz78}
53: to theorize that the formation of the Galaxy was a prolonged, complex process.
54:
55: While the second parameter phenomenon has informed our understanding of galaxy
56: formation, the identity of the second parameter remains a topic of some debate.
57: A number of candidates have been suggested including age, helium abundance, CNO
58: abundances,
59: cluster central density, and stellar rotation (see \citet{fp96} for a review).
60: It is clear that its influence grows with distance from the Galactic center
61: \citep{sz78,zi80,ldz94}. To gain a better understanding of the second parameter it is
62: logical to carefully examine the most distant GCs. Only six known GCs inhabit the
63: Galaxy at distances greater than 50 kpc \citep{ha96}
64: \footnote{References to the Harris catalog actually refer to the February 2003 revision
65: that can be found at http://www.physics.mcmaster.ca/$\sim$harris/mwgc.dat.}
66: and five of the six are prime
67: examples of second parameter GCs, i.e. metal poor GCs with red HBs.
68:
69: All six of the GCs with $\rgc >$ 50 kpc: Pal~14, NGC~2419,
70: Eridanus, Pal~3, Pal~4, and AM-1 (listed in order of increasing $\rgc$) have been
71: observed by the Hubble Space Telescope (HST) using the Wide Field Planetary
72: Camera 2 (WFPC2). \citet{ha97} presented a HST/WFPC2 color-magnitude diagram
73: (CMD) of NGC~2419---the only one with $\rgc >$ 50 kpc and a blue HB---and
74: concluded that it is roughly coeval with the old, metal poor cluster M~92.
75: In contrast, \citet[hereafter S99]{st99} presented HST/WFPC2 CMDs of Pal~3,
76: Pal~4, and Eridanus and concluded that these clusters are younger than relatively
77: nearby clusters of similar metallicities (M~3 and M~5) by $\sim$1.5--2 Gyr assuming
78: the inner and outer halo GCs have similar abundances.
79:
80: The two remaining GCs, AM-1 and Pal~14, are the subject of this paper.
81: AM-1 was classified as a GC by \citet{am79} though it had
82: been observed previously and named ESO 201-10 by \citet{ho75}. \citet{am79}
83: estimated the distance to AM-1 at 300 kpc and, although it has since been
84: revised downward to $\sim$120 kpc \citep{aa84,or85,ma89,hi06}, it is still the most
85: distant Galactic GC known. Age estimates for AM-1 have proved uncertain to date
86: because none of the published photometry has reached the main sequence.
87: Pal~14, at $\rgc \sim$70 kpc, was discovered by van den Bergh in 1958 and later
88: classified as a GC by \citet{ar60}; it is also known as AvdB.
89: \citet{sa97} published the deepest CMD of Pal~14 to date, reaching more than a
90: magnitude fainter than the main sequence turn off (MSTO), and concluded that Pal~14 is
91: 3--4 Gyr younger than inner halo GCs with similar metallicities.
92:
93: We describe the observations and data reduction in $\S$2 and present the WFPC2 CMDs
94: of AM-1 and Pal~14 in section $\S$3. Available information regarding the reddening
95: and metallicity of both clusters is discussed in $\S$4 followed by relative and absolute
96: age analyses in $\S$5. The results are discussed and put into context in $\S$6 and
97: the paper concludes with a summary in $\S$7.
98:
99: \section{Observations and Data Reduction}
100:
101: The observations used in the present study were obtained with HST/WFPC2 as part
102: of program number GO-6512 (PI: Hesser) during cycle 6. Table \ref{obslog} shows
103: the observing log. The 36 AM-1 and 14 Pal~14 images were retrieved from the HST archive
104: and calibrated using the pipeline bias and flat-field procedures.
105:
106: The stellar photometry was performed with the HSTPhot software \citep{dolphin} which
107: is designed especially for WFPC2 observations. A number of preprocessing steps were
108: performed before applying HSTPhot. First, image defects such as bad and saturated
109: pixels were masked using the data quality files available for each WFPC2 image. Cosmic
110: rays and hot pixels were flagged using tasks within HSTPhot. Next, the positional
111: offsets between each image and a reference image were measured using the IMEXAMINE
112: task in IRAF. These were input into the multiphotometry mode of HSTPhot in order
113: to measure the magnitudes of all detected profiles on each image. The measurements
114: were done using point spread functions (PSFs) that accompany the HSTPhot software
115: employing the `weighted PSF fitting mode' which places more weight on the
116: central pixels in a profile and less on the outer pixels.
117:
118: The instrumental magnitudes were transformed to the native WFPC2 VEGAmag system as
119: well as the ground-based Johnson-Cousins VI system using equations included in HSTPhot.
120: The final set of standardized photometry was extracted from the results file by
121: selecting only those stars with 'Object Codes' that suggest a well-measured star-like
122: profile. The object codes from HSTPhot are defined and described in the work of
123: \citet{dolphin}. These come from a comparison of the $\chi$ values of the fits using a
124: stellar profile, a single pixel without a background, and a flat profile representing
125: an extended object. For detections where the stellar profile fits best, one of three star
126: classes is assigned to the object that characerize the probability that the object is
127: either: a single star (object code = 1), an unresolved binary (object code = 2), or a
128: line-of-sight binary (object code = 3). In this work, we have selected only those
129: detections with object code equal to one for further analysis.
130:
131: Figure \ref{photerr} shows how the photometric errors in both bands behave as a function
132: of apparent magnitude. In the plot, F555W and F814W are referred to as V and I, respectively.
133: For AM-1 the average photometric errors are 0.002 in both bands at the level of the
134: HB and rise to 0.011 mag in F555W and 0.012 mag in F814W at the level of the MSTO.
135: For Pal~14 the average photometric errors are 0.002 mag in both bands at the level of the HB
136: and rise to 0.007 mag in F555W and 0.008 mag in F814W at the level of the MSTO. As Figure
137: \ref{photerr} demonstrates, the scatter in photometric errors is somewhat greater in Pal~14
138: than in AM-1.
139:
140:
141: \section{Color-Magnitude Diagrams}
142:
143: The WFPC2 F555W--F814W CMDs of AM-1 and Pal~14 are presented in Figure \ref{CMD}.
144: Both CMDs are remarkably free of field star contamination. Figure \ref{CMD} also indicates
145: the average photometric errors by showing error bars at F555W=22, 24, 26, and 28 towards
146: the left side of each panel.
147: The CMD of AM-1 (left panel of Figure \ref{CMD}) includes over 4,000 stars and extends
148: from the brightest giant at F555W $\sim$ 18.5 down to approximately F555W = 29 though
149: the main sequence becomes washed out by F555W $\sim$ 27. The figure shows a
150: well-populated MS and lower red giant branch (RGB), a tight red clump at F555W $\sim$ 21,
151: and a significant population of blue stragglers.
152: The right panel of Figure \ref{CMD} shows over 2,000 stars in Pal~14. It exhibits a narrow MS
153: that extends about 5 mags below the turnoff. Stars brighter than the turnoff are
154: relatively rare in Pal~14 and thus the red clump and RGB are not as well-defined as in AM-1.
155: Pal~14 contains a handful of blue stragglers.
156:
157: Figure \ref{LF} presents F555W luminosity functions (LFs) of AM-1 and Pal~14 but we stress
158: that these results are preliminary and not corrected for incompleteness. The error bars
159: are $\sqrt{N}$. Plotted along with
160: the observed LFs are model LFs based on the isochrone fitting described in $\S$5.2 and
161: assuming power-law mass functions (MFs). In this paper, the power-law MF takes the form
162: dN/dM$\propto$M$^x$ where x is referred to as the power-law slope.
163: The model LFs were normalized to the clusters by matching the number of stars within
164: $\pm$1 mag of the MSTO. In the top panel, the AM-1 LF is fit well by models with a
165: flat (x=0) or positive power-law MF. In the bottom panel, the LF of Pal 14 indicates a
166: negatively sloped MF but here the model fits are worse than for AM-1. This is
167: related to the flat appearance of the Pal 14 LF (in the logarithmic plot) between
168: V=23 and ~28. The models, regardless of power-law slope, suggest more curvature in this
169: region. The unusual LF of Pal 14 deserves further attention and will likely be improved
170: by data covering more of the cluster.
171:
172:
173: %Preliminary analyses suggest that, at least in the core, the AM-1 luminosity function (LF) is
174: %similar to those of Pal~3, Pal~4, and Eridanus (based on the S99 data), each with a roughly
175: %flat mass function. In contrast, Pal~14 is best fit by a mass function slope near the
176: %\citet{salp} value. From the \citet{ha96} catalog we find that Pal~14 has approximately
177: %twice the radial extent (in parsecs) of the largest of the remaining four GCs and a central
178: %surface brightness $\sim$1.5--2 mag fainter than the others.
179:
180:
181: %The CMD, mass function, radial
182: %extent, and surface brightness comparisons show that while Pal~14 is a metal poor GC with
183: %a red HB, in other ways it has little in common with the other second
184: %parameter clusters in the outer reaches of the Galactic halo. The only major exception,
185: %age, is discussed in $\S$5.
186:
187:
188: \section{Reddening and Metallicity}
189:
190: Before we can measure relative cluster ages we need to adopt a reddening and metallicity
191: for each cluster in our sample, including our comparison cluster, for which we
192: have selected the well-known Galactic globular M~3.
193: No single spectroscopic study encompasses AM-1, Pal~14, and M~3. However, with M~3 measured
194: by \citet[hereafter ZW84]{zw84}, AM-1 measured by \citet{su85} using the same spectral
195: features as ZW84, and Pal~14 measured by \citet{ar92}, there is something
196: approacing a common scale. \citet{ar92} found good agreement between their [Fe/H] values and
197: ZW84 for clusters that appeared in both studies.
198:
199: Table \ref{param} shows the distance moduli from \citet{ha96}, reddening estimates from the
200: \citet{sc98} dust maps, metallicities from the sources cited in the previous paragraph, and
201: the iron and $\alpha$-element abundances for M~3 from \citet[hereafter PVI]{pvi}.
202: PVI determined abundances by averaging high resolution spectroscopic measurements from
203: 36 stars in M~3. The PVI abundances are likely the most accurate available for M~3
204: and were adopted for the purpose of isochrone fitting in $\S$5.
205: Since the metallicities listed in column three of Table \ref{param} for each cluster agree
206: to within 0.1 dex, and certainly to within the quoted errors, it is reasonable to
207: differentially compare these three GCs.
208: %AM-1 has a very small reddening: E(B--V) = 0.008 \citep{sc98}. In contrast, the
209: %metallicity is rather poorly constrained. \citet{su85} measured Ca II H and K lines in
210: %two giants and arrived at [Fe/H] = --1.7$\pm$0.2.
211: It should also be noted that \citet{hi06} determined that a Yonsei--Yale isochrone \citep{ki02}
212: with [Fe/H] = --1.4 and [$\alpha$/Fe] = +0.3 at 11 Gyr provided the best fit to the RGB.
213: Since the photometry presented by \citet{hi06} for AM-1 did not reach the MSTO
214: we do not consider this result---at least regarding the age---to be well--constrained.
215:
216: %Pal~14 has a low but nonzero reddening of E(B--V)=0.035 \citep{sc98}. \citet{ar92} provided
217: %a metallicity estimate of [Fe/H]=--1.60$\pm$0.18 based on Ca triplet spectra of four red
218: %giants and we have adopted this value.
219:
220: Since the observational data considered in this paper are in either the ground-based V--I or
221: WFPC2 F555W--F814W color and the reddening estimates are given in E(B--V) it is necessary to
222: convert them. \citet{bcp98} determined E(V--I)/E(B--V) = 1.32 + 0.06(V--I) for stars with
223: spectral types from B to K. Since we are interested in colors lying between 0.5 $<$ V--I
224: $<$ 1.5, we chose (V--I) = 1 as the average color and thus adopted E(V--I) = 1.38~E(B--V).
225: In the native HST/WFPC2 photometric system, we take E(F555W--F814W) = 1.2~E(B--V) from
226: Table 13 of \citet{ho95}.
227:
228:
229: \section{Cluster Age Determinations}
230:
231: \subsection{The Horizontal Method}
232:
233: We begin by employing the relative age-dating technique known as the
234: horizontal method pioneered by \citet{sd90} and \citet[hereafter VBS]{vbs90}.
235: Adopting the approach advocated by VBS (also used by S99), we register the
236: photometry of each program
237: cluster and the fiducial sequence of a comparison cluster with similar metal
238: abundance using the color of the MSTO and the magnitude
239: at a point on the main sequence that is 0.05 mag redder than the MSTO. Once
240: registered in this manner, younger clusters will have redder RGBs than older
241: ones.
242:
243: For our comparison cluster, M~3,
244: the CMD used to construct a fiducial sequence was taken from the ACS Survey of
245: Galactic GCs \citep{sa07}. For direct comparison with the data presented in this paper,
246: the ACS data were converted from F606W and F814W to ground-based V and I using
247: the transformation equations in \citet{si05}. A fiducial was constructed using only stars
248: with photometric errors less than 0.05 mag in both bandpasses. The high quality CMD
249: and the fiducial sequence of M~3 are shown in the left panel of Figure \ref{M3fid}.
250: We now proceed to compare this sequence to the photometric data of our
251: program clusters.
252:
253: The left panel of Figure \ref{vbs} shows our photometry for AM-1 and the
254: fiducial sequence of M~3 matched using the prescription of VBS while
255: the right panel of the same figure illustrates the comparison of Pal~14 with
256: M~3 matched in the same manner. Note that the evolved stars in Pal~14
257: are relatively few in number (see Figure \ref{CMD}). As a result, we have
258: augmented the WFPC2 photometry (circles) with ground-based photometric standard
259: stars reported by \citet[triangles]{st00}. We have offset the WFPC2 data in color
260: and magnitude to match the photometric zero point of the ground-based data.
261: Specifically,
262:
263: \begin{eqnarray}
264: <V(HST) - V(Stetson)> = 0.068 \pm 0.007; N = 31\\
265: <I(HST) - I(Stetson)> = 0.007 \pm 0.007; N = 17
266: \end{eqnarray}
267:
268:
269: The lines parallel to the fiducial RGBs represent the locations of clusters that are
270: 2 Gyr older (left) and 2 Gyr younger (right) than the comparison cluster.
271: The $\pm$2 Gyr lines were derived by applying the VBS method to [Fe/H]=--1.5,
272: [$\alpha$/Fe]=+0.2 isochrones from \citet{d08} with ages ranging from 10 to 14 Gyr.
273: From the isochrones we found that V--I decreases by 0.0197 mag per Gyr and thus
274: the lines were plotted at $\sim$0.04 mag on either side of the M~3 fiducial.
275:
276: The RGBs of AM-1 and Pal~14 suggest that these clusters are 1--2 Gyr younger
277: than M~3. Figure \ref{vbs} shows that the RGB stars in AM-1 lie to the red of the M~3
278: fiduical sequence. The case of Pal~14 is less concrete due to the lack of a well-defined
279: sequence of stars. The trend in Pal~14 is more obvious if
280: one considers the range from V=20 to 22. These relative age statements depend
281: on the assumption that the program clusters and M~3 share similar chemical compositions.
282:
283:
284: \subsection{Isochrone Fitting}
285: Before the isochrone fitting results are presented, two points regarding the CMDs
286: of AM-1 and Pal~14 need to be made. First, we have elected to present the isochrone
287: analysis of the AM-1 CMD in the native HST/WFPC2 system
288: because no published ground-based VI photometry exists for this cluster, which
289: would allow us to check our photometric zero point.
290: Second, as we did in the application of the horizontal method in the previous section,
291: we will use the ground-based Pal~14 data to better define the location of the
292: evolved-star sequences.
293:
294: Theoretical isochrones from the Dartmouth Stellar Evolution Database
295: \footnote{http://stellar.dartmouth.edu/$\sim$models/}\citep{d07,d08}
296: were used in the analysis. These isochrones employ state of the art input physics and
297: include partially inhibited diffusion and gravitational settling of He and metals
298: \citep{c01} and a model atmosphere based surface boundary condition.
299: The reader is referred to \citet{d07,d08} for a full discussion of their properties.
300: These models have been tabulated in 0.5 Gyr intervals but, for the purpose of
301: fitting isochrones to fiducials and CMDs, the isochrones have been interpolated to
302: produce a finer grid spacing of 0.1 Gyr in order to allow for more
303: precise age constraints.
304:
305: Adopting the PVI abundances of M~3 and starting with the distance and reddening estimates
306: in Table \ref{param}, we then vary the age, distance, and reddening until a satisfactory
307: correspondence is achieved between the main sequence of the cluster and the isochrone,
308: making sure that the point 2 mag below the MSTO (i.e., $M_V \sim$ 6) matches the data.
309: This is the same procedure followed by \citet{sa07}.
310: The results of isochrone fitting are collected in Table \ref{ages} and shown in
311: the right panel of Figure \ref{M3fid}, Figure \ref{AM1iso}, and Figure \ref{Pal14iso}.
312: In short, we find that Pal~14 is younger than AM-1 by $\sim$0.5 Gyr if both clusters
313: have similar abundances.
314:
315: The distance and reddening estimates
316: derived from the isochrone fits provides a check on the fitting method. Comparing the
317: reddening estimates derived from isochrone fitting in Table \ref{ages} to the values
318: listed in Table \ref{param}, we find that differences in the reddening values are
319: within the expected uncertainty in a given reddening of $\pm$0.02 mag in E(B--V).
320: As with the reddenings, the isochrone-based distances listed in Table \ref{ages}
321: generally agree with those in Table \ref{param} to within the expected error
322: of $\pm$0.1 mag in distance modulus.
323:
324: \subsection{Consistency Check on the Ages}
325:
326: Thus far we have presented two types of age analyses, the horizontal VBS method and
327: isochrone fitting, to probe the ages of AM-1 and Pal~14 relative to M~3. Both
328: methods suggest that the program clusters, which exhibit HBs that are
329: predominantly redward of the RR Lyrae instability strip, are 1--2 Gyr younger
330: than the comparison cluster, M~3, which exhibits stars on the red and
331: blue portions of the HB. In this section, our aim is to provide a consistency
332: check of those results using the method employed by \citet{sll}. This method
333: compares the relative locations of the MSTO in two clusters whose CMDs have
334: been properly registered as described in more detail below.
335:
336: Since all three clusters have similar metallicities, we expect them to have
337: comparable HB absolute magnitudes and RGB colors at the level of the HB.
338: This is slightly modulated by the fact that cluster age also affects the level
339: of the red clump and the color of the RGB. The former has been calibrated by
340: \citet{cr01} to be 0.028 mag/Gyr for ages older than $\sim$5 Gyr and
341: metallicities below [Fe/H] $\sim$ --1.0. The latter can be assessed using our
342: isochrones from which we find a variation of 0.003 mag/Gyr in the V--I
343: color of the RGB at the level of the HB.
344:
345: With these guidelines in mind, we proceed to use the measured HB
346: magnitudes and RGB colors of AM-1, Pal~14, and M~3 along with their
347: relatives ages (Table \ref{ages}) to shift them to the same distance and reddening
348: scale. This allows us to compare their MSTO/SGB locations
349: with each other as shown in Fig. \ref{mstocomp}. We see that the turnoff
350: regions of both AM-1 and Pal~14 are demonstrably brighter than M~3.
351: Also included in Fig. \ref{mstocomp} are \citet{d08} isochrones with [Fe/H]=--1.5
352: and [$\alpha$/Fe]=+0.2 at 10.6 and 12.6 Gyr (the best-fit age to the M~3 fiducial and
353: 2 Gyr younger). The MSTO regions
354: of AM-1 and Pal~14 are consistent with ages that are 1--2 Gyr younger than M~3, in
355: agreement with the horizontal method and isochrone fitting.
356:
357: As we close out this section, it is important to reiterate that the age estimates
358: presented in this paper are accurate only to the extent that AM-1, Pal~14, and M~3
359: have similar compositions. In order to gauge the effect of changing either [Fe/H] or
360: [$\alpha$/Fe] we use an isochrone with [Fe/H]=--1.5 and [$\alpha$/Fe]=+0.2 at 12 Gyr
361: as the baseline and search for an isochrone with the same separation between the MSTO
362: and the base of the RGB at different compositions. An increase (decrease) in [$\alpha$/Fe]
363: of 0.2 dex causes the age to decrease (increase) by 0.5 Gyr. If [Fe/H] increases by 0.2
364: dex, then the age decreases by 0.5 Gyr but if [Fe/H] decreases by the same amount, then
365: the age increases by $\sim$1 Gyr. Our age measurements must be viewed with some caution
366: until more conclusive abundances can be derived for the sample clusters.
367:
368:
369: \section{Discussion}
370:
371: We provide strong evidence that AM-1 and Pal~14 join the S99 sample of Pal~3, Pal~4, and
372: Eridanus in having younger ages than comparable inner halo GCs. Hence all five GCs
373: with $\rgc >$ 50 kpc and red HBs are younger than their counterparts with bluer HBs in
374: the inner halo.
375: These five GCs are second parameter clusters (in the sense that they have redder HBs
376: at a given metallicity than the inner halo trend would suggest) and all have been
377: demonstrated to have relatively young ages. This makes a strong case for age as the
378: second parameter in the outer halo. Can we make a case for age as the second parameter
379: throughout the Galaxy - i.e. among the entire sample of Galactic globular clusters?
380:
381: \citet[see their Table 1]{m05} presented a compilation of parameters (e.g. $\rgc$,
382: HB type, [Fe/H]) for the entire Galactic GC population. In addition, each is assigned
383: to a sub-population based on the categories introduced by \citet{z93} along with the
384: addition of two categories not available in 1993: GCs associated with the Sagittarius
385: dwarf galaxy and GCs containing multiple stellar populations. According to the
386: \citet{z93} classification, GCs that follow the inner halo trend of HB type vs. [Fe/H]
387: belong to the ``Old Halo'' (OH) while those that diverge from this trend by a certain
388: amount (or more) belong to the ``Younger Halo'' (YH).
389:
390: The YH clusters are synonymous with the second parameter phenomenon and while
391: they dominate the OH population at large $\rgc$, the vast majority of them have
392: $\rgc <$ 50 kpc. Thus the outer halo clusters may not be a representative sample of
393: the entire YH population but there is evidence for age as the second parameter among GCs
394: with $\rgc <$ 50 kpc as well. Among these, GCs such as Ruprecht~106 \citep{buon93}, IC~4499
395: \citep{fer95}, Palomar~12 \citep{stet89}, and Pyxis \citep{sarageis96} have been shown to be
396: younger than inner halo GCs with similar metallicities and bluer HB morphologies. This trend
397: is also seen through the detailed examination of `second parameter pairs' such as NGC~288 and
398: NGC~362 \citep[VBS]{sd90} or M~3 and M~13 \citep{rey2001}. At the same time, accurate
399: abundance determinations of second parameter GCs (particularly the more distant
400: ones) are desperately needed. Accurate abundances for the YH population, coupled with
401: a systematic age analysis of these clusters, will provide deeper insight into the
402: Galactic GC population and the nature of the second parameter.
403:
404: We note in passing that \citet{m05} do not consider Pal~12 to be a member of the YH population
405: because of its supposed association with the Sgr dwarf spheroidal galaxy. However, Pal~12 is
406: spatially distinct from the main body of Sgr so that we can consider it to be a member of the
407: Milky Way halo. In any case, if the hypothesis of \citet{sz78} and the implications of the
408: analysis performed by \citet{m05} are correct---that all of the YH clusters originated in
409: (now disrupted) dwarf spheroidal galaxies---then we have no choice but to consider clusters
410: like Pal~12 part of the Milky Way's YH population and thus part of the Milky Way halo.
411:
412: \section{Conclusion}
413: Color magnitude diagrams of AM-1 and Pal~14 from HST/WFPC2 F555W--F814W photometry were presented.
414: The CMDs reveal unprecedented depth, reaching $\sim$4 mag below the MSTO in both clusters.
415: AM-1 and Pal~14 have very similar CMD morphologies. AM-1 shows a more densely populated RGB and HB,
416: albeit with very few bright red giants, and an impressive display of blue stragglers. Pal~14 has a
417: sparsely populated RGB and HB which is likely due to its large size (compared to the area covered by
418: WFPC2) and low density.
419:
420: Three different age measurement techniques reveal that AM-1 and Pal~14 are $\sim$1.5--2
421: Gyr younger than M~3 assuming all three GCs have similar compositions. Isochrone fitting
422: suggests that if AM-1 and Pal~14 have similar abundances, then Pal~14 is about 0.5 Gyr
423: younger than AM-1. S99 found that
424: Pal~3, Pal~4 and Eridanus are also $\sim$1.5--2 Gyr younger than inner halo GCs of
425: comparable metal abundance. Thus all five of the GCs with $\rgc >$ 50 kpc and red HBs
426: are 1.5--2 Gyr younger than GCs with similar metallicities in the inner halo (to the
427: extent that assumptions regarding similar abundances are valid).
428:
429: The variation of the second parameter phenomenon with $\rgc$ has been known for some time, but
430: the identity of the second parameter itself remains a topic of active debate in the literature,
431: even some 40 years after it was discovered. When combined with previously published results on
432: the ages of clusters that exhibit the second parameter effect, the present study provides further
433: evidence that age is a strong candidate for the second parameter. However, there are a significant
434: number of globular clusters important to identifying the second parameter that have not been
435: accurately age-dated. Of equal importance is the precise determination of abundances in these
436: clusters. Improvements to both the total metal abundance and the abundance ratios in these clusters
437: will lead to more appropriate comparisons with well-studied, nearby GCs and theoretical
438: models and, ultimately, to a better understanding of the second parameter phenomenon and
439: the formation history of the Galactic halo.
440:
441: \acknowledgments
442: We thank Peter Stetson for sharing the HST photometry of Pal~3, Pal~4, and Eridanus.
443: We thank the anonymous referee for insightful comments that improved the presentation and
444: focus of the paper.
445:
446: \begin{thebibliography}{}
447: \bibitem[Aaronson et al.(1984)]{aa84}Aaronson, M., Schommer, R. A., Olszewski, E. W. 1984, \apj, 276, 221
448: \bibitem[Armandroff et al.(1992)]{ar92}Armandroff, T. E., Da Costa, G. S., \& Zinn, R. J. 1992, \aj, 104, 164
449: \bibitem[Arp \& van den Bergh(1960)]{ar60} Arp, H.~C. \& van den Bergh, S. 1960, \pasp, 72, 48
450: \bibitem[Bessell et al.(1998)]{bcp98}Bessell, M. S., Castelli, F. \& Plez, B. 1998, \aap, 333, 231
451: \bibitem[Buonanno et al.(1993)]{buon93}Buonanno, R., Corsi, C. E., Fusi Pecci, F., Richer, H. B., \& Fahlman, G. G. 1993, \aj, 105, 184
452: \bibitem[Chaboyer et al.(2001)]{c01}
453: \bibitem[Crowl et al.(2001)]{cr01}Crowl, H. H., Sarajedini, A., Piatti, A. E., Geisler, D., Bica, E., Claria, J. J., Santos, J. F. C., Jr. 2001, \aj, 122, 220
454: \bibitem[Dolphin(2000)]{dolphin}Dolphin, A. E. 2000, \pasp, 112, 1383
455: \bibitem[Dotter et al.(2007)]{d07}Dotter, A., Chaboyer, B., Jevremovi\'c, D., Baron, E., \& Ferguson, J. W., Sarajedini, A., \& Anderson, J. 2007, \aj, 134, 376
456: \bibitem[Dotter et al.(2008)]{d08}Dotter, A., Chaboyer, B., Jevremovi\'c, D., Kostov, V., Baron, E., \& Ferguson, J. W. 2008, \apjs, in press
457: \bibitem[Ferraro et al.(1995)]{fer95}Ferraro, I., Ferraro, F. R., Fusi Pecci, F., Corsi, C. E., \& Buonanno, R. 1995, \mnras, 275, 1057
458: \bibitem[Fusi Pecci et al.(1996)]{fp96}Fusi Pecci, F., Bellazzini, M., Ferraro, F. R., Buonanno, R., \& Corsi, C. E. 1996, ASPC, 92, 221
459: \bibitem[Girardi et al.(2000)]{gir2000} Girardi, L., Bressan, A., Bertelli, G., \& Chiosi, C. 2000, \aaps, 141, 371
460: \bibitem[Harris(1996)]{ha96}Harris, W. E. 1996, \aj, 112, 1487
461: \bibitem[Harris et al.(1997)]{ha97}Harris, W.~E., Bell, R. A., VandenBerg, D. A., Bolte, M., Stetson, P. B., Hesser, J. E., van den Bergh, S., Bond, H. E., Fahlman, G. G., \& Richer, H. B. 1997, \aj, 114, 1030
462: \bibitem[Hilker(2006)]{hi06}Hilker, M. 2006, \aap, 448, 171
463: \bibitem[Holmberg et al.(1975)]{ho75}Holmberg, E.~B., Lauberts, A., Schuster, H.~E. \& West, R.~M. 1975, \aap, 22, 327
464: \bibitem[Holtzman et al.(1995)]{ho95}Holtzman, J. A., Burrows, C. J., Casertano, S., Hester, J. J., Trauger, J. T., Watson, A. M., \& Worthey, G. 1995, \pasp, 107, 1065
465: \bibitem[Kim et al.(2002)]{ki02}Kim, Y., Demarque, P., Yi, S.~K., \& Alexander, D.~R. 2002, \apjs, 143, 499
466: \bibitem[Lee, Demarque, \& Zinn(1994)]{ldz94}Lee, Y. W., Demarque, P., Zinn, R. J. 1994, \apj, 423, 248
467: \bibitem[Mackey \& van den Bergh(2005)]{m05}Mackey, A.~D. \& van den Bergh, S. 2005, \mnras, 360, 631
468: \bibitem[Madore \& Arp(1979)]{am79}Madore, B.~F. \& Arp, H.~C. 1979, \apj, 227, 103
469: \bibitem[Madore \& Freedman(1989)]{ma89}Madore, B.~F. \& Freedman, W.~L. 1989, \apj, 340, 812
470: \bibitem[Ortolani(1985)]{or85}Ortolani, S. 1985, \aap, 137, 269
471: \bibitem[Pritzl, Venn, \& Irwin(2005)]{pvi}Pritzl, P. J., Venn, K. A., \& Irwin, M. 2005, \aj, 130, 2140 (PVI)
472: \bibitem[Rey et al.(2001)]{rey2001}Rey, S. -C>, Yoon, S. -J, Lee, Y. -W.,
473: Chaboyer, B., \& Sarajedini, A. 2001, \aj, 122, 3219
474: \bibitem[Salpeter(1955)]{salp}Salpeter, E. E. 1955, \apj, 121, 161
475: \bibitem[Sarajedini(1997)]{sa97}Sarajedini, A. 1997, \aj, 113, 682
476: \bibitem[Sarajedini \& Demarque(1990)]{sd90} Sarajedini, A. \& Demarque, P. 1990, ApJ, 365, 219
477: \bibitem[Sarajedini \& Geisler(1996)]{sarageis96}Sarajedini, A. \& Geisler, D. 1996, \aj, 112, 2013
478: \bibitem[Sarajedini et al.(2007)]{sa07}Sarajedini, A. et al. 2007, \aj, 133, 1658
479: \bibitem[Sarajedini, Lee, \& Lee(1995)]{sll}Sarajedini, A., Lee, Y. W., \& Lee, D. H. 1995, \apj, 450, 712
480: \bibitem[Schlegel et al.(1998)]{sc98}Schlegel, D. J., Finkbeiner, D. P., \& Davis, M. 1998, \apj, 500, 525
481: \bibitem[Searle \& Zinn(1978)]{sz78}Searle, L. \& Zinn, R. 1978, \apj, 225, 357
482: \bibitem[Sirianni et al.(2005)]{si05}Sirianni, M. et al. 2005, \pasp, 117, 1049
483: \bibitem[Stetson(2000)]{st00}Stetson, P. B. 2000, \pasp, 112, 925
484: \bibitem[Stetson et al.(1989)]{stet89}Stetson, P. B., Hesser, J. E., Smith, G. H.,
485: VandenBerg, D. A., \& Bolte, M. 1989, \aj, 97, 1360
486: \bibitem[Stetson et al.(1999)]{st99}Stetson, P.~B., Bolte, M., Harris, W. E., Hesser, J. E., van den Bergh, S., VandenBerg, D. A., Bell, R. A., Johnson, J. A., Bond, H. E., Fullton, L. E., Fahlman, G. G., \& Richer, H. B. 1999, \aj, 117, 247 (S99)
487: \bibitem[Suntzeff et al.(1985)]{su85}Suntzeff, N., Olszewski, E., Stetson, P.~B. 1985, \aj, 90, 1481
488: \bibitem[VandenBerg et al.(1990)]{vbs90} VandenBerg, D. A., Bolte, M., \& Stetson, P. B. 1990,
489: AJ, 100, 445 (VBS)
490: \bibitem[Zinn(1980)]{zi80}Zinn, R. 1980, \apj, 241, 602
491: \bibitem[Zinn(1993)]{z93}Zinn, R. 1993, ASPC, 48, 38
492: \bibitem[Zinn \& West(1984)]{zw84}Zinn, R. \& West, M. J. 1984, \apjs, 55, 45 (ZW84)
493: \end{thebibliography}
494:
495:
496: \begin{deluxetable}{lccc}
497: \tablecolumns{5}
498: \tablewidth{0pc}
499: \tablecaption{Observing Log\label{obslog}}
500: \tablehead{\colhead{Cluster}&\colhead{Date}&\colhead{Filter}&\colhead{Exposure Time}
501: }
502: \startdata
503: AM--1 & 1999 June & F555W & 4 x 180s \\
504: & & & 1 x 900s \\
505: & & & 8 x 1000s \\
506: & & & 5 x 1300s \\
507: & & F814W & 4 x 230s \\
508: & & & 5 x 1200s \\
509: & & & 4 x 1300s \\
510: & & & 5 x 1400s \\
511: Palomar 14 & 1999 July & F555W & 4 x 160s \\
512: & & & 1 x 900s \\
513: & & & 7 x 1000s \\
514: & & F814W & 4 x 230s \\
515: & & & 4 x 1100s \\
516: & & & 2 x 1200s \\
517: & & & 2 x 1300s \\
518: \enddata
519: \end{deluxetable}
520:
521: \begin{deluxetable}{cccccc}
522: \tablecolumns{6}
523: \tablewidth{0pc}
524: \tablecaption{Globular Cluster Parameters\label{param}}
525: \tablehead{\colhead{Name}&\colhead{$DM_V$}&\colhead{E(B--V)}&\colhead{[Fe/H]}&
526: \colhead{[Fe/H]$_{PVI}$}&\colhead{[$\alpha$/Fe]$_{PVI}$}}
527: \startdata
528: M~3 & 15.12 & 0.013 & --1.66$\pm$0.06 & --1.50 & +0.22 \\
529: AM-1 & 20.43 & 0.008 & --1.7$\pm$0.2 & \nodata& \nodata \\
530: Pal~14 & 19.47 & 0.035 & --1.60$\pm$0.18 & \nodata& \nodata \\
531: \enddata
532: \end{deluxetable}
533:
534:
535: \begin{deluxetable}{ccccc}
536: \tablecolumns{5}
537: \tablewidth{0pc}
538: \tablecaption{Isochrone-based Parameters\label{ages}}
539: \tablehead{\colhead{Name}&\colhead{DM$_V$}&\colhead{E(B--V)}&\colhead{Age (Gyr)}&\colhead{Rel. Age (Gyr)}}
540: \startdata
541: M~3 & 15.05 & 0.011 & 12.6 & \nodata \\
542: AM-1 & 20.41 & 0.016 & 11.1 & --1.5 \\
543: Pal~14 & 19.51 & 0.033 & 10.5 & --2.1 \\
544: \enddata
545: \end{deluxetable}
546:
547: \begin{figure}
548: \plotone{f1}
549: \caption{Photometric errors as a function of apparent magnitude in AM-1 and Pal~14.
550: For the sake of brevity, we have used V=F555W and I=F814W in the plot.\label{photerr}}
551: \end{figure}
552:
553: \begin{figure}
554: %\includegraphics[scale=0.6]{plots/CMD}
555: \plotone{f2}
556: \caption{HST/WFPC2 CMDs of AM-1 (left) and Pal~14 (right). AM-1 reveals a tight red
557: clump, a well-defined RGB below the level of the red clump, a MS that extends $\sim$4
558: mag below the turnoff, and an abundant collection of $\sim$20 blue stragglers. Pal~14
559: shows a sparsely populated red clump and RGB, a narrow MS that extends $\sim$5 mag
560: below the turnoff, and a handful of blue stragglers. Characteristic photometric errors
561: are plotted at intervals in both panels.\label{CMD}}
562: \end{figure}
563:
564: \begin{figure}
565: \epsscale{0.9}
566: \plotone{f3}
567: \caption{Luminosity functions in F555W of AM-1 (top) and Pal~14 (bottom) with $\sqrt{N}$ error bars but without correcting for incompleteness. Model LFs are plotted based on the isochrone fits described in $\S$5.2 and assuming power-law mass functions (which we define as dN/dM$\propto$M$^x$). AM-1 is fit well down to V=27 by a flat or positively-sloped MF but Pal~14 is rather poorly fit by the models, see text for discussion.\label{LF}}
568: \end{figure}
569:
570: \begin{figure}
571: %\includegraphics[scale=0.6]{plots/M3fid}
572: \plotone{f4}
573: \caption{The left panel shows the HST/ACS CMD of M~3 with the fiducial sequence (solid line) over-plotted. The CMD only includes stars with photometric errors less than 0.05 mag in both bands. The right panel compares the best fit isochrone (dashed line) to the fiducial sequence (solid line).\label{M3fid}}
574: \end{figure}
575:
576:
577: \begin{figure}
578: %\includegraphics[scale=0.9]{plots/VBS}
579: \plotone{f5}
580: \caption{The horizontal (VBS) method applied to our photometry of AM-1 (left panel) and
581: Pal~14 (right panel). The fiducial sequence (solid line) is that of the globular cluster M~3.
582: The filled triangles in the right panel represent the ground-based
583: Pal~14 photometry obtained from \citet{st00} to which the WFPC2 photometric zero point has been offset. The parallel lines along the lower RGB represent age differences of $\pm$2 Gyr.\label{vbs}}
584: \end{figure}
585:
586: \begin{figure}
587: %\includegraphics[scale=0.75]{plots/AM1iso}
588: \plotone{f6}
589: \caption{Results of isochrone fitting to AM-1. The relevant parameters used in the fit are listed on the figure.\label{AM1iso}}
590: \end{figure}
591:
592: \begin{figure}
593: %\includegraphics[scale=0.75]{plots/Pal14iso}
594: \plotone{f7}
595: \caption{Same as Figure \ref{AM1iso} but for Pal~14. The open circles are WFPC2 photometry and the filled triangles are from \citet{st00}.\label{Pal14iso}}
596: \end{figure}
597:
598: \begin{figure}
599: %\includegraphics[scale=0.58]{plots/SLL}
600: \epsscale{1.0}
601: \plotone{f8}
602: \caption{The \citet{sll} method applied to AM-1 (left panel) and Pal~14 (right panel).
603: The M~3 fiducial sequence (solid lines) and isochrones with [Fe/H]=--1.5, [$\alpha$/Fe]=+0.2, and ages of 10.6 and 12.6 Gyr (dashed lines) are plotted. The filled triangles in the right panel represent the ground-based Pal~14 photometry obtained from \citet{st00} to which the WFPC2 photometry (circles) has been matched.\label{mstocomp}}
604: \end{figure}
605:
606: \end{document}
607: