1: % /scr2/jlc/m71/paper_abund_accepted1/ms.tex May 2001
2: %
3: %
4: %\documentstyle[12pt,aasms4]{article}
5: \documentclass[preprint]{aastex}
6: %\documentstyle[11pt,aaspp4,flushrt]{article}
7: %\documentstyle[aas2pp4]{article}
8: %
9: % aas2pp4 is dual column, aaspp4 is what we've been using (11pt),
10: % aasms4 (12pt) is what is submitted.
11: %
12:
13: \slugcomment{AJ, in press (Sep. 2001)}
14:
15: \newcommand{\kms}{km~s$^{-1}$}
16: \newcommand{\subsun}{\mbox{$_{\odot}$}}
17: \newcommand{\etal}{{\it et al.\/}}
18: \newcommand{\teff}{$T_{eff}$}
19: \newcommand{\grav}{log($g$)}
20:
21:
22: \begin{document}
23:
24: \title{Abundances in Stars from the Red Giant Branch Tip to the
25: Near the Main Sequence Turnoff
26: in M71: I. Sample Selection, Observing Strategy and Stellar Parameters
27: \altaffilmark{1}}
28:
29: \author{Judith G. Cohen\altaffilmark{2},
30: Bradford B.Behr\altaffilmark{2,3} and Michael M. Briley\altaffilmark{4}}
31:
32: \altaffiltext{1}{Based on observations obtained at the
33: W.M. Keck Observatory, which is operated jointly by the California
34: Institute of Technology and the University of California}
35:
36: \altaffiltext{2}{Palomar Observatory, Mail Stop 105-24,
37: California Institute of Technology}
38:
39:
40: \altaffiltext{3}{Current address: Department of Astronomy, University of Texas,
41: Austin, Texas 78712}
42:
43: \altaffiltext{4}{Department of Physics, University of Wisconsin,
44: Oshkosh, Wisconsin}
45:
46:
47: \begin{abstract}
48:
49: We present the sample for an abundance analysis of 25
50: members of M71 with luminosities ranging from the red giant branch tip to
51: the upper main sequence. The spectra are of high dispersion and
52: of high precision. We describe the observing strategy and
53: determine the stellar parameters for the sample stars using both
54: broad band colors and fits of H$\alpha$ profiles.
55: The derived stellar parameters agree with those from the
56: Yale$^2$ stellar evolutionary tracks to within 50 -- 100K for a fixed
57: \grav, which is within the level of the uncertainties.
58:
59:
60: \end{abstract}
61:
62: \keywords{globular clusters: general ---
63: globular clusters: individual (M71) --- stars: evolution -- stars: abundances}
64:
65:
66: \section{Introduction}
67:
68: By virtue of their large populations of coeval stars, the Galactic
69: globular clusters present us with a unique laboratory for the study of
70: the evolution of low mass stars. The combination of their extreme
71: ages, compositions and dynamics also allows us a glimpse at the early
72: history of the Milky Way and the processes operating during its
73: formation. These aspects become even more significant in the context of
74: the star-to-star light element inhomogeneities found among red giants
75: in every globular cluster studied to date. The large differences in the surface
76: abundances of C, N, O, and often Na, Mg, and Al have defied a
77: comprehensive explanation in the three decades since their discovery.
78:
79: Proposed origins of the inhomogeneities typically break down into two
80: scenarios: 1) As C, N, O, Na, Mg, and Al are related to proton capture
81: processes at CN and CNO-burning temperatures, material cycled through a
82: region above the H-burning shell in evolving cluster giants may be
83: brought to the surface with accompanying changes in composition. While
84: standard models of low mass stars do not predict this ``deep mixing,''
85: several theoretical mechanisms have been proposed (e.g., the meridional
86: mixing of Sweigart \& Mengel 1979, and turbulent diffusion, Charbonnel
87: 1994, 1995) with varying degrees of success. Moreover, there is ample
88: observational evidence that deep mixing does take place during the red
89: giant branch (RGB) ascent of metal-poor cluster stars
90: (see the reviews of Kraft 1994 and
91: Pinsonneault 1997 and references therein). 2) It has also become
92: apparent that at least some component of these abundance variations
93: must be in place before some cluster stars reach the giant branch.
94: Spectroscopic observations of main sequence turn-off stars in 47 Tuc
95: (Briley \etal\ 1996, \&
96: Cannon \etal\ 1998 and references therein), NGC 6752 (Suntzeff \& Smith 1991,
97: Gratton \etal\ 2001) and most recently
98: in M71 (Cohen 1999, Briley \& Cohen 2001)
99: have shown variations in CN
100: and CH-band and in some cases Na and O line strengths as well
101: consistent with patterns found among the evolved giants of these clusters.
102:
103: All we know about stellar evolution strongly suggests that
104: these low mass main sequence
105: globular cluster stars are incapable of producing significant
106: amounts of C, N or O while on the main sequence and also incapable
107: of deep dredge-up.
108: Both would be required to reproduce the observed abundance variations.
109: This leads directly to the possibility
110: that the early cluster material was at least partially inhomogeneous in
111: these elements or that some form of modification of
112: the relative abundances of these elements has
113: taken place within the cluster since the currently observed
114: cluster stars were formed. Some suggested culprits include mass-loss
115: from intermediate mass asymptotic giant branch stars and supernovae
116: ejecta (see Cannon \etal\ 1998 for an excellent discussion of these
117: possibilities).
118:
119: In addition, King \etal\ (1998) have added
120: another complication
121: to the issue of abundance variations within globular clusters. Their
122: analysis of a small number of sub-giants in M92 yields
123: [Fe/H] = $-$2.52 dex
124: \footnote{The standard nomenclature is adopted; the abundance of
125: element X is given by [X/H] = log10[N(X)/N(H)] $-$ log10[N(X)/N(H)]\subsun.},
126: a value smaller by about a factor of two than that derived
127: from the spectra of giants in M92
128: by many authors including Cohen (1978) and
129: Sneden \etal\ (1991). If this is not due to some error in the
130: analysis or a variation in non-LTE corrections that has not been
131: properly included, this result is quite puzzling
132: since the Fe abundance
133: in the photosphere of these stars could not possibly
134: be altered by mixing.
135:
136: In an effort to unveil the source of the star-to-star element variations
137: seen among the light elements within
138: globular clusters, as well as to determine the constancy, or lack thereof,
139: of the abundances of the heavy elements such as Fe, where no foreseeable mixing can
140: be expected to alter its abundance, we have in initiated the present program
141: to study at high dispersion stars over a large range in luminosity within
142: the nearer galactic globular clusters. We begin with the nearest globular cluster
143: easily reached from a northern hemisphere site, M71.
144:
145: \section{The Selection of Stars}
146:
147: Stars were chosen to span the range from the tip of the red giant branch
148: to the main sequence turnoff of M71. Membership considerations at this stage
149: involved location on the existing $B,V$
150: photometric sequences of Arp \& Hartwick (1971) and, for the more luminous
151: stars, assignment of a high probability of membership by Cudworth (1985) in
152: his proper motion survey of this globular cluster. When possible,
153: stars were chosen which had known
154: CH and CN band strengths from the survey by Briley, Smith \& Claver (2001) of the
155: red giant branch or the work of Cohen (1999) for the main sequence
156: region. Unpublished spectra of these bands from Cohen were
157: available for some of the subgiants as well.
158: Within luminosity ranges of 1 mag, an effort
159: was made to select stars that spanned the full range of observed CH
160: and CN band strengths, i.e. CN weak and CN strong stars.
161: Only reasonably isolated stars were selected.
162:
163: Because this cluster lies at low galactic latitude, field star
164: contamination is a serious issue. It was not possible to define
165: the cluster sequence clearly in the subgiant regime. There the
166: evolution is rapid, hence the stellar density along the isochrone is low,
167: while the number of field stars is rising rapidly towards fainter magnitudes.
168: Cudworth's (1985) proper motion survey in M71 does not
169: reach faint enough to include subgiants, and even with the very recent study of
170: Geffert \& Maintz (2000), not available at the time our sample
171: was selected, isolation of a clean sample of subgiants in M71
172: would be quite difficult.
173:
174: Throughout this paper, the star names are from Arp \& Hartwick (1971),
175: or, when not included in the former, are created from the
176: object's J2000 coordinates.
177:
178: \section{The HIRES Observations}
179:
180: All spectra were obtained with HIRES (Vogt \etal\ 1994) at
181: the Keck Observatory. A maximum
182: slit length of 14 arc sec can be used with our instrumental configuration
183: without orders overlapping. Since an image rotator for HIRES is available
184: (built under the leadership of David Tytler), if we can find pairs of program stars
185: with separations less than 8 arcsec, they can be observed together on
186: a single exposure. Ideally pairs consisted of two members
187: of the M71 sample, but when that was not feasible, pairs with a random
188: star of suitable brightness (i.e. brighter than the sample star) were chosen
189: in the hope that the second star would also turn out to be a member of M71.
190:
191: The desired minimum SNR was 75 over a 4 pixel resolution element for
192: a wavelength near the center of echelle order 56 ($\sim$6400 \AA).
193: This is calculated strictly from the counts in the object spectrum, and
194: excludes noise from cosmic ray hits, sky subtraction, flattening problems, etc.
195: Since the nights were dark, sky subtraction is not an issue except at
196: the specific wavelengths corresponding to strong night sky emission lines,
197: such as the Na D doublet. This SNR goal was
198: achieved, at considerable cost in observing time,
199: for all but the faintest star. The faintest star
200: was not a program star, but rather an object that fell within the slit
201: for a program star setup. Its SNR is only 50 per resolution element.
202: %
203: % 2.4 e/DN, 600 DN total achieves 75/1
204: %
205:
206: The fainter program stars required
207: integration times of several hours.
208: If there was more than one potential brighter second star, then both such
209: could be observed during the course of the exposures for the
210: fainter star by changing the position angle of the instrument's slit
211: at some point during the exposure sequence while still keeping the faint M71 star
212: in the slit.
213:
214: Approximate measurements of the radial velocity were made on line,
215: and if a star was determined to be a non-member, the observations were terminated.
216: If the probable non-member was the second component in a pair, an attempt
217: was made to switch to another position angle to pick up a different
218: second star, when a possible candidate that was bright enough
219: was available within the limits of the 8 arcsec maximum separation.
220: Through creative use of close pairs, a sample of 29 stars was observed in M71
221: with HIRES.
222:
223: To avoid crowding of spectral lines, the observations were centered
224: at about 6500\AA. A 1.15 arcsec slit was used, which provides a spectral
225: resolution of 34,000. All long integrations were broken
226: up into separate exposures, each 1200 sec long, to optimize cosmic
227: ray removal.
228:
229: Because the HIRES detector is undersized, our spectra do not cover the
230: full length of each echelle order without gaps in the wavelength coverage.
231: We wanted to include key lines of critical elements,
232: specifically the 6300, 6363 [OI] lines, the 7770 O triplet, the
233: Na doublet at 6154, 6160\AA, and the 6696, 6698\AA\ Al I lines.
234: Two instrumental configurations were used for the brightest stars,
235: as it was impossible to create a single one which included all
236: the desired spectral features in the wavelength range 6000 to 8000 \AA.
237: In particular, although the 6696, 6698\AA\ Al I doublet is the most useful
238: feature of that element in this spectral region, we could not get it
239: to fit
240: into a single instrumental configuration together with the O lines.
241: For the faintest stars, only a single configuration was used, which
242: included the O lines but did not include the Al I doublet.
243:
244: The spectra were reduced by BBB using Figaro (Shortridge 1993) scripts
245: with commands written by McCarthy and Tomaney (McCarthy 1988)
246: specifically for echelle data reduction.
247:
248: Table 1 gives details of the HIRES exposures for each star, with the
249: total exposure time for the primary and for the Al configuration.
250: The signal level per pixel in the continuum at 6150 \AA\ is also given, from
251: which the SNR can be calculated assuming Poisson statistics and ignoring
252: issues of cosmic ray removal, flattening etc. The latter become non-negligible
253: for the very long HIRES integrations necessary as faint as possible
254: in M71.
255: Also listed is the radial velocity for each star measured from the HIRES
256: spectra and
257: the probability of membership assigned by Cudworth (1985) based on
258: his proper motion study, which included only the brighter stars in
259: the sample. A montage of spectra of a single echelle
260: order for selected stars covering the luminosity range of the sample
261: is shown in Figure~1.
262:
263: It should be noted that to acquire this set of relatively high precision
264: and high dispersion spectra took a total of five nights of assigned time
265: at the Keck Observatory. One of the assigned nights was used for a
266: backup program due to poor seeing conditions.
267:
268:
269:
270:
271: \section{Radial Velocities}
272:
273: Radial velocities were measured from all the HIRES spectra by
274: cross correlating orders 56 and 57 against the spectrum
275: of a bright template star (in practice the brightest observed M71 star)
276: from each run. The radial velocity
277: of the template star from each run was determined by fitting Gaussians
278: to 20 strong isolated features in these two orders
279: using wavelengths from the NIST Atomic Spectra Database Version 2.0 (NIST
280: Standard Reference Database \#78).
281: Heliocentric corrections appropriate for each exposure were then applied.
282:
283: The internal radial velocity errors were calculated following
284: the precepts of Davis \& Tonry (1979) using the relation
285: $\sigma(v_r) = \alpha/[1+R(TD)]$, where the parameter $R(TD)$
286: is a measure of the ratio of the
287: height of the peak of the cross correlation to the noise in the
288: cross correlation function away from the peak.
289: The constant $\alpha$ was set at 15 \kms, which represents a value
290: at the upper end of those found in other recent HIRES programs using similar
291: instrumental configurations by
292: Mateo \etal\ (1998), Cook \etal\ (1999), and C\^ot\'e \etal\ (1999).
293: Every star except the faintest one observed,
294: G53414\_4435,
295: % o51291_3655ft RIGHT NAME
296: has an internal error
297: in $v_r$ under 1 \kms, while this star only has a somewhat larger
298: internal error of 1.5 \kms.
299:
300: Based on their measured radial velocities, four of the
301: 29 stars observed with HIRES are not members
302: of M71 (stars Y, G53425\_4608, G53475\_4547 and G53394\_4624).
303: All of these four stars were chosen not as members
304: of the primary M71 sample but as bright(er) stars to complete
305: a pair, i.e. chosen primarily on the basis of their location
306: on the sky without as careful
307: scrutiny of their colors as was done for the primary sample.
308:
309: The radial velocities for the 25 members of M71 observed with
310: HIRES are listed in Table 1.
311: They have a mean $v_r$ of $-$21.7 \kms. After removing in quadrature
312: an internal uncertainty of 1.0 \kms, we find
313: $\sigma$ = 2.6 \kms. This is in excellent agreement
314: with the value of Peterson \& Latham (1986) of
315: $-$22.5 \kms\ ($\sigma$ = 2.4 \kms) determined from a sample of 17
316: bright giants in M71.
317:
318: The only star which shows lines obviously broader than those expected from
319: the instrumental resolution is M71 I-80, a RHB star. The profiles
320: of weak lines of this star suggest that it is rotating at a
321: projected velocity of $\sim$8 \kms, an issue which will be
322: discussed by B.Behr in a future publication.
323:
324:
325: \section{The Stellar Parameters}
326:
327: We seek to determine for the M71 sample of stars the stellar
328: atmosphere parameters necessary
329: to carry out an abundance analysis from our HIRES data.
330: We adopt for M71 the distance (3900 pc) and reddening (E(B-V) = 0.25 mag)
331: from the on-line compilation of Harris (1996).
332: \footnote{The extinction maps of
333: Schlegel, Finkbeiner \& Davis (1998)
334: from their analysis of the COBE/DIRBE database are not reliable this close to
335: the galactic plane. For M71, with $b = -4.6^{\circ}$,
336: they deduce $E(B-V) = 0.32$ mag.}
337: The relative extinction in various passbands is taken from
338: Cohen \etal\ (1981) (see also Schlegel, Finkbeiner \& Davis 1998).
339: Based on the high dispersion analysis of Cohen (1983) for
340: several red giants in M71, since confirmed by
341: additional high resolution studies of giants in M71 by
342: Leep, Oke \& Wallerstein (1987)
343: and by Sneden \etal\ (1994),
344: we adopt as an initial guess a metallicity for the
345: cluster of [Fe/H] = $-$0.7 dex.
346:
347:
348: \subsection{\teff\ From Broad Band Colors - Predictions of the Model Atmospheres}
349:
350: We utilize here the grid of predicted broad band colors and
351: bolometric corrections of
352: Houdashelt, Bell \& Sweigart (2000) based on the
353: MARCS stellar atmosphere code (Gustafsson \etal\ 1975). Before proceeding
354: we demonstrate that the Kurucz and MARCS predicted colors are essentially
355: identical, at least for the specific colors used here.
356:
357: We compare the colors predicted from the MARCS code from Houdshelt \etal\
358: with those from
359: the Kurucz ATLAS code (Kurucz 1992). We take the predicted $V-K$ color
360: from each model in the MARCS grid with [Fe/H] = $-0.5$, and interpolate
361: within the Kurucz color grid at the same abundance and at the \grav\
362: of the MARCS model
363: to find the \teff\ that would be deduced.
364:
365: A contour plot of the difference $\Delta$\teff(Kurucz - MARCS)
366: that
367: results when the $V-K$ color is used
368: is shown in Figure 2. The three contour levels shown
369: correspond to $\Delta$\teff = 0, 30 and 60 K. Also shown
370: in this figure as the thick curve is a 12 Gyr isochrone for M71
371: from the very recently completed Yale$^2$ isochrones of
372: Yi \etal\ (2001).
373: Along this isochrone, $\Delta$\teff = 0 to 25 K
374: for the subgiants and main sequence turnoff, and $\Delta$\teff =
375: 0 to 15 K for the red giant branch. Throughout the entire displayed
376: range in \teff,\grav, $\delta$\teff\ ranges from 0 to 50 K. We thus demonstrate that
377: to within a tolerance of 25 K, the Kurucz and MARCS temperature scales
378: from broad band $V-K$ colors are identical.
379:
380:
381: \subsection{\teff\ From Broad Band Colors}
382:
383: Broad band $B,V$ colors are available from the work of Arp \& Hartwick (1971).
384: With such high reddening and metallicity, we chose to ignore the $B$ measurements.
385: Stetson (2000) provides $V,I$ photometry for about 25\% of the cluster area,
386: specifically the NE quadrant. To supplement this, JGC carried out
387: $V, R, I$ photometry using short exposure images of M71 taken
388: with LRIS (Oke \etal\ 1995) at the Keck Observatory
389: for slitmask alignment purposes. These frames were
390: calibrated by observations of standard fields from Landolt (1992).
391: The brightest M71 giants were saturated on all the LRIS images.
392:
393: For a smaller sample of $\sim$235
394: stars in M71, we have obtained infrared photometry at $K$
395: using the infrared acquisition and guiding
396: camera on NIRSPEC (McLean \etal\ 1998, 2000) at the Keck Observatory.
397: These were taken during a night dedicated to infrared spectroscopy
398: in M71, and are basically setup images for the spectroscopy.
399: The data is calibrated to the standard stars of Persson \etal\ (1998).
400: A 256 x 256 pixel NICMOS detector is used with a scale of
401: 0.18 arcsec/pixel. Hence the fields are very small and
402: main sequence stars dominate the sample. The
403: frames were reduced in a standard manner.
404: This is supplemented by the infrared photometry of Frogel, Persson
405: \& Cohen (1979) for the upper giant branch of M71. For the
406: single star in common between the two infrared samples
407: there is reasonable agreement (see below). Infrared photometry was obtained
408: for a few additional stars using the camera of Murphy \etal\ (1995)
409: at the 1.5m telescope at Palomar Mountain. With a scale of 0.6
410: arcsec/pixel and a 256x256 Nicmos array,
411: only a small portion of M71 can be covered at once. Exposures of more than
412: 2000 sec (broken up into many spatial positions and repeats) are
413: required to reach the fainter stars in our sample, and the crowding
414: is severe with these relatively large pixels. However, $K$
415: magnitudes for five of the stars in the HIRES sample were
416: obtained in this way.
417:
418: The $V,V-K$ color magnitude diagram for M71 for the HIRES sample is shown
419: in Figure 3. The different sources of $K$ photometry are indicated by
420: different symbols. The stars observed by Frogel, Cohen \& Persson (1979)
421: are also indicated; three stars from their sample,
422: whose membership probabilities are below 15\% in the proper motion
423: study of Cudworth (1985), have been excluded as it is unlikely
424: that they are members of M71. There are two stars in the HIRES sample
425: with multiple measurements for $K$. These are indicated by horizontal lines
426: connecting the relevant points in this figure. (M71 1-45, the
427: brightest star in $V$ in the HIRES sample,
428: near the tip of the RGB
429: has two measurements of $K$ differing by only 0.09 mag, difficult
430: to see on the figure.)
431:
432: The photometry of Frogel \etal\ (1979) reaches to RHB. While the number
433: of stars in common with the previously published photometry is only one,
434: the consistency of the M71 RGB and RHB delineated by the
435: published photometry and by our photometry shown in Figure~3 indicates
436: that our mixing of several different sources for $K$ has been done
437: in a valid manner.
438:
439: The observed broad band colors for
440: each program star ($V-I$ and, when available, $V-K$), corrected for extinction, are used to determine
441: \teff. The set of models with metallicity of $-$0.5 dex, nearest to our
442: initial estimate of [Fe/H], is used.
443: Table~3 lists the \teff\ deduced from each of $V-I$ and $V-K$, when
444: available.
445:
446: The calibration of our photometric data, as distinct from that of Stetson (2000),
447: could be better. We assume an uncertainty of 0.02 mag applies to
448: $V-I$ from Stetson (2000), an uncertainty of 0.03 mag applies to
449: colors from the LRIS images, and an uncertainty of 0.05 mag
450: applies to all $V-K$ colors. Given the relatively high
451: reddening of M71, there is an additional uncertainty
452: due to possible spatial variations in reddening across the field of the cluster.
453: We assume this occurs for $E(B-V$) at a level of 10\%, which is the
454: fractional variation in $E(B-V)$ detected across much more heavily reddened
455: globular clusters by Cohen \& Sleeper (1995).
456: This translates into a total uncertainty in \teff\ of 75 K for giants rising
457: to 150 K for main sequence stars using $V-I$, divided about equally between
458: the two contributions (uncertainty in reddening and photometric
459: colors), and 40 K from $V-K$ for giants rising to 70 K
460: for dwarfs, with the dominant contribution arising from the
461: photometric uncertainties. We adopt the larger of these uncertainties
462: (those from $V-I$)
463: as applicable for our \teff\ determinations.
464:
465:
466:
467: \subsection{Computation of \grav \label{grav}}
468:
469: Once an initial guess at \teff\ has been established from a broad
470: band color, it is possible with minimal assumptions
471: to evaluate \grav\ using observational data.
472: The adopted distance modulus, initial
473: guess at \teff, and an assumed stellar mass (we adopt 0.8 $M$\subsun\
474: for the upper main sequence stars, and 0.6 $M$\subsun\ for
475: the RHB stars) are combined with
476: the known interstellar absorption, the predictions of the
477: model atmosphere grid
478: for bolometric corrections as well as a broad band observed $V$ mag to
479: calculate \grav.
480:
481: An iterative scheme is used to correct for the small
482: dependence of the predictions of the model atmosphere grid on
483: \grav\ itself. Rapid convergence is achieved.
484:
485: It is important to note that because of the constraint of
486: a known distance to M71, the
487: uncertainty in \grav\ is small, $\le0.1$ dex when comparing
488: two members of M71. Propagating an uncertainty of 15\% in the cluster
489: distance, 5\% in the stellar mass, and 3\% in \teff\ from a
490: reddening uncertainty of 0.04 mag in $E(B-V)$, and ignoring
491: any covariance, leads to
492: a potential systematic error of $\pm$0.2 dex for \grav.
493:
494: \subsection {\teff\ and \grav\ from H$\alpha$ Profiles}
495:
496: The profiles of the Balmer lines can also be used in principle to determine
497: \teff\ and \grav\ in the temperature range characteristic of the M71 stars.
498: The estimates of stellar parameters obtained in this way are
499: to first order independent of reddening and of any photometric data.
500: There is little sensitivity to surface gravity or overall abundance.
501: Given the constraints on \grav\ and known approximate metallicity
502: imposed by the globular cluster membership of the sample stars,
503: the primary dependence of the Balmer line profiles in this regime of
504: \teff\ and \grav\ for the HIRES sample of members of M71 is on \teff.
505:
506: We attempt to use the H$\alpha$ profiles for this purpose.
507: The HIRES spectra themselves are not suitable for this purpose,
508: as the large scale continuum
509: determination, particularly in these relatively metal rich cool stars,
510: is suspect at the level of 1 to 2\% due to imperfect removal of the variation
511: of the instrumental response across each echelle order. Instead
512: the Balmer line profiles were obtained from observations
513: with LRIS (Oke \etal\ 1995)
514: at the Keck Observatory. A 1200 g/mm grating with 0.7 arcsec
515: wide slits was centered at 6500 \AA\ to yield 0.63\AA/pixel or a
516: spectral resolution of $\sim$1.7\AA.
517: The same slitmasks that were designed, fabricated and utilized for the
518: CH and CN observations of 79 main sequence stars
519: in M71 described in Cohen (1999)
520: were used again in M71 for these observations. An additional slitmask of
521: subgiants was also designed and used for this purpose.
522:
523: The spectra were reduced in the usual fashion using Figaro (Shortridge 1993).
524: Continuum bandpasses were defined based on examination of the much
525: higher dispersion HIRES spectra of M71 stars. The median
526: value within each of the continuum regions was chosen as a
527: representative value for the bandpass. A second order polynomial
528: fit to the signal for each of the ``line free'' regions was used to
529: define the continuum. To improve the SNR still further
530: for the main sequence stars, the H$\alpha$ spectra of
531: three to five stars
532: of similar luminosity along the
533: main sequence of M71 were summed, then the resulting profile was normalized.
534: Figure~4 shows H$\alpha$ profiles for three stars summed
535: near the bright end of Cohen's (1999) main sequence sample, which
536: corresponds in the color-magnitude diagram of M71 to the main
537: sequence stars in the HIRES sample, as
538: well as for a subgiant.
539:
540: These profiles were compared against the spectral flux calculated by
541: Hauschildt \etal\ (1999) for [Fe/H] = $-$0.7 dex.
542: \footnote {As has already been pointed out by
543: van't Veer-Menneret \& Megessier (1996), the predicted Balmer line profiles
544: released with the ATLAS 9 models
545: of Kurucz (1992) are not valid and fail to reproduce the solar profile.
546: They found, as do we, that these H$\alpha$ profiles over-estimate
547: \teff\ by several hundred degrees.}
548: The grid spacing of the spectral synthesis is 2 \AA.
549: We fit the predicted flux of Hauschildt \etal\ in the region of
550: H$\alpha$ using the same procedure as had been applied to the
551: stellar spectra to generate a set of model Balmer line profiles
552: with a normalized continuum level.
553:
554: Even with the use of
555: LRIS spectra instead of
556: echelle spectra, the continuum determination across the
557: H$\alpha$ profile is still uncertain
558: by 1\%. The sensitivity to \teff\ of the wings of the predicted H$\alpha$ profiles
559: is not large compared to this potential uncertainty.
560: To minimize its effect, we compared the observed
561: and predicted H$\alpha$ profiles only over
562: the region within 3 \AA\ of
563: the line center. The resulting values of \teff\ for our M71 stars
564: are still considerably
565: higher than those derived from the broad band colors.
566: Because of the problems mentioned above as well as the potential
567: impact of continued
568: small improvements to the broadening theory for Balmer lines
569: (see, for example, Barklem, Piskunov \& O'Mara 2000), we
570: decided to use the H$\alpha$ profiles only to estimate relative
571: values for \teff\ from star to star within the M71 sample, forcing
572: agreement with the \teff\ deduced from the colors at the main sequence.
573:
574: The H$\alpha$ profiles provide measurements of \teff\
575: which are in agreement with those derived from the stellar colors
576: to within the uncertainties of each method.
577: One might hope to determine the stellar mass at the turnoff
578: directly from the observations through the gravity dependence of
579: these Balmer line profiles. However,
580: the required precision in the observed Balmer line
581: profiles of better than 1\% is not easily achieved, nor is it
582: clear that the theoretical profiles are sufficiently accurate.
583: Furthermore, the dependence of the Balmer line profiles on
584: \teff\ is much larger. Thus
585: determining \teff\ itself with sufficient precision to
586: then extract a precise value for \grav\
587: would be extremely difficult.
588:
589: \section{Comparison of Stellar Parameters with Isochrones}
590:
591: Table 3 provides a summary of the stellar parameters for the 25 members
592: of our M71 sample determined both from broad band photometry and from
593: H$\alpha$ fits. In addition to the values of \teff\ from
594: $V-I$, from $V-K$ and from H$\alpha$ (when appropriate)
595: \footnote{H$\alpha$ is not included in the determination of \teff\
596: for the three hottest stars
597: near the M71 main sequence turnoff.},
598: a mean temperature is listed. The weight of the
599: H$\alpha$ value, when used, is half that
600: of the values from $V-I$ and from $V-K$.
601: With the adopted zero point
602: for assignment of \teff\ from the H$\alpha$ profiles,
603: the good agreement between the three values, consistent with the
604: expected observational errors, is gratifying.
605:
606: Given that these stars
607: sample the population of a globular cluster, \teff\ should
608: decrease monotonically as the luminosity of the star increases.
609: Furthermore stars in the
610: same region of the cluster isochrone ideally should have very similar stellar
611: parameters. The weighted values of \teff\ given in Table~3 do not
612: quite achieve this. We therefore slightly adjusted the
613: weighted \teff\ by not more than 100 K (150 K for star G53392\_4624)
614: (values typical of our
615: observational uncertainties)
616: while retaining the mean
617: relationship unaltered to try to achieve this.
618: The adopted \teff\ for each star in the M71 sample, listed
619: in the final column of table~3, is the value
620: used in the abundance analyses presented in subsequent papers in this series.
621:
622: For the non-members,
623: since their distances are unknown, no value of \grav\ can be obtained
624: and the derived \teff\
625: will be incorrect if the reddening is different from the value adopted for M71.
626:
627: Figure~5 compares the adopted \teff\ and \grav\ for our HIRES sample of
628: members of M71 with the isochrone predicted for a stellar system
629: with an age of 12 Gyr with
630: [Fe/H] = $-0.7$ dex from
631: the very recently completed Yale$^2$ tracks of Yi \etal\ (2001).
632: Scaled solar mixture abundances are used in the Yale$^2$ calculations
633: for all elements heavier than He.
634:
635: First we note that the set of M71 stars observed with HIRES provides
636: a reasonable sample of the cluster isochrone from the RGB tip to
637: the upper main sequence, with the exception of the lack of
638: subgiants.
639:
640: Comparing theory and observation using the set of parameters shown
641: in Figure~5, quite different from the usual
642: color-magnitude diagram, is a very stringent test.
643: The agreement with
644: with the new Yale$^2$ isochrone is quite good. The \teff\ of the
645: theoretical giant branch for the metallicity of M71, which is well
646: known from past work and determined yet again in Paper II,
647: is only 50 -- 100 K cooler at a fixed \grav\ than are the observed stars.
648:
649: We already know from comparison with the infrared photometry of
650: Frogel \etal\ (1979)
651: that one cannot ascribe this systematic discrepancy to uncertainties in the
652: measurements. So we now consider the various types of systematic
653: errors that might have occurred.
654: There are two known systematic errors in the handling of the observational
655: data described above. The first is a systematic underestimate of
656: \grav\ by 0.04 dex as the theoretically predicted mass
657: along the upper RGB is
658: 0.88 M\subsun, while a mass of 0.80 M\subsun was used to calculate the
659: surface gravities for the cluster stars (excluding the RHB stars)
660: from the observed magnitudes
661: and colors. The second is an underestimate
662: in \teff\ of less than 20 K because the model grid used to define
663: the predicted broad band colors had [Fe/H] = $-$0.5 dex, not the
664: nominal metallicity of M71 of $-$0.7 dex.
665:
666: As discussed in \S\ref{grav}, the internal errors from star to star
667: in \grav\ are small, while the systematic error is dominated
668: by the uncertainty in the distance, and is indicated in Figure~5
669: by arrows. The errors indicated in \teff\ are dominated by
670: the uncertainty in the reddening. An overestimate of the
671: reddening $E(B-V)$ by $\sim$0.04 mag, which seems unlikely,
672: could reproduce most of the discrepancy shown in
673: Figure~5 through the resulting underestimates of \teff.
674:
675: Another area of concern is the validity of the
676: relationships we have adopted between color, \teff, and \grav.
677: As discussed earlier, we
678: have carefully checked the consistency of
679: the predicted colors from Houdashelt \etal\ (2000) with those from
680: Kurucz (1992) computed using the ATLAS code, and have also examined
681: the the comparison with the
682: empirical color--\teff--[Fe/H] relations for dwarfs and for
683: giants established by Alonso \etal\ (1996, 1999). For $V-K$, the
684: MARCS and Kurucz predictions are in very close agreement, while the
685: empirical fits to the angular diameter measurements using the infrared
686: flux method carried out by Alonso \etal\ yield a \teff\ about 50 K cooler for a fixed
687: $V-K$ color in the relevant range.
688:
689: In addition, the theoretical tracks utilized thus far
690: do not include enhancement of the
691: $\alpha$-elements, which is common in metal poor globular cluster
692: giants. However, the O-enhanced tracks of
693: Bergbusch \& VandenBerg (1992) do not fit any better for the
694: nominal metallicity of M71. This is not surprising as
695: Bergbusch \& VandenBerg show that to first order the effects of enhancing
696: O are equivalent to using a model with scaled solar abundances
697: with an appropriately calculated higher global
698: metallicity. This would make
699: the predicted RGB cooler, making the discrepancy slightly
700: worse. Their latest $\alpha$-enhanced models given in
701: VandenBerg \etal\ (2000) retain this behavior.
702:
703: We know from Paper II the correct [Fe/H] for M71, and will
704: shortly know from Paper III the $\alpha$-element enhancements.
705: With that information plus the stellar parameters of Table~3,
706: once any small remaining discrepancies between the predicted
707: and observed stellar
708: parameters is understood, one can check for consistency with the new
709: $\alpha$-enhanced tracks of VandenBerg \etal\ (2000).
710:
711:
712: The total effect under consideration (i.e. the discrepancy between
713: the theoretical stellar isochrones and the behavior of the observed
714: cluster sample in M71 shown in Figure~3)
715: is only $\sim$50 -- 100 K in \teff. There are several possible
716: contributions on the observational side
717: which may be large enough to explain it,
718: including an error in the adopted reddening for the cluster
719: and uncertainties in the relation utilized between color and stellar
720: atmospheric parameters.
721: Hence we have chosen to wait for
722: more such comparisons to be carried out in the domain of \teff, \grav\
723: for
724: additional clusters in future papers before speculating further on this issue.
725:
726: \section{Looking Forward}
727:
728: With this information in hand, we are ready
729: to carry out an abundance analysis based on measurements
730: of equivalent widths from the HIRES spectra of the M71 sample.
731: An analysis of the Fe abundances for this
732: sample of M71 stars is presented in the next paper
733: in this series (Ram\'{\i}rez \etal\ 2001).
734:
735:
736:
737:
738: \acknowledgements
739:
740: The entire Keck/HIRES and LRIS user communities owes a huge debt to
741: Jerry Nelson, Gerry Smith, Steve Vogt, Bev Oke, and many other
742: people who have worked to make the
743: Keck Telescope and HIRES and LRIS a reality and to operate and
744: maintain the Keck Observatory. We are grateful to the
745: W. M. Keck Foundation for the vision to fund
746: the construction of the W. M. Keck Observatory.
747: We thank Peter Stetson for supplying his M71 photometry in easily
748: accessible form and Peter Hauschildt for calculating a grid
749: of H$\alpha$ profiles for us. Partial support
750: was provided to MMB by a Theodore Dunham, Jr. grant
751: for Research in Astronomy and by the National Science Foundation under
752: grants AST-9819614 to JGC and AST-9624680 to MMB.
753:
754:
755: \begin{references}
756:
757: \reference{} Alonso, A., Arribas, S. \& Martinez-Roger, C., 1996, A\&A, 313, 873
758:
759: \reference{} Alonso, A., Arribas, S. \& Martinez-Roger, C., 1999, A\&AS, 140, 261
760:
761: \reference{} Arp, H.~C. \& Hartwick, F.~D.~A., 1971, \apj, 167, 499
762:
763: \reference{} Barklem, P.S., Piskunov, N. \& O'Mara, B.~J., 2000, A\&A, 363, 1091
764:
765: \reference{} Bergbusch, P.~A. \& VandenBerg, D.~A., 1992, \apjs, 81, 163
766:
767: \reference{} Briley, M.~M. \& Cohen, J.~G., 2001, \aj, in
768: press (Astro-ph/0104099)
769:
770: \reference{} Briley, M. M., Smith, V. V., Suntzeff, N. B., Lambert, D. L., Bell,
771: R. A., \& Hesser, J. E. 1996, Nature, 383, 604
772:
773: \reference{} Briley, M.~M., Smith, G.~H. \& Claver, C.~F., 2001, \aj, submitted
774:
775: \reference{} Cannon, R.D., Croke, B.F.W., Bell, R.A., Hesser, J.E. \&
776: Stathakis, R.A., 1998, \mnras, 298, 601
777:
778: \reference{} Charbonnel, C., 1994, A\&A, 282, 811
779:
780: \reference{} Charbonnel, C., 1995, \apjl, 453, L4
781:
782: % \reference{} Charbonnel, C., Meynet, G., Maeder, A., Schaller, G. \& Schaerer, D.,
783: % 1993, A\&AS, 101, 415
784: % grid of models
785:
786: % \reference{} Cohen, J.~G., 1979, \apj, 223, 487
787: % M3 and M13
788:
789: \reference{} Cohen, J.~G., 1978, \apj, 231, 751
790: % M92 (-2.35 Fe/H) and M15
791:
792: % \reference{} Cohen, J.~G., 1980, \apj, 241, 981
793: % M71, M67 and NGC 2420 (4 stars, -1.27, 4100, 0.7)
794:
795: \reference{} Cohen, J.~G., 1983, \apj, 270, 654
796: % (metal rich globs) (-0.7 dex)
797:
798: % \reference{} Cohen, J.~G., 1999a, \aj, 117, 2428
799: % M13 paper
800:
801: \reference{} Cohen, J.~G., 1999, \aj, 117, 2434
802: % M71 paper
803:
804: \reference{} Cohen, J.~G., Frogel, J.~A., Persson, S.~E. \& Elias, J.~H.,
805: \apj, 249, 481, 1981
806:
807: \reference{} Cohen, J.~G. \& Sleeper, E.~C., 1995, \aj, 109, 242
808:
809: \reference{} Cook, K., Mateo, M., Olszewski, E.W., Vost, S.S.,
810: Stubbs, C. \& Diercks, A., 1999, \pasp, 111, 306
811:
812: \reference {} C\^ot\'e, P., Mateo, M., Olszewski, E.W. \& Cook, K.H.,
813: 1999, \apj, 526, 147
814:
815: \reference{} Cudworth, K.~M., 1985m \aj, 90, 65
816:
817: \reference{} Davis, M. \& Tonry, J.L., 1977, \aj, 84, 1511
818:
819: \reference{} Frogel, J.~A., Persson, S.~E. \& Cohen, J.~G., 1979, \apj, 227, 499
820:
821: \reference{} Geffert, M. \& Maintz, G., 2000, A\&AS, 144, 227
822:
823: \reference{} Gratton, R.~G. \etal\, 2001, A\&A, in press
824: (see Astro-ph/0012457)
825:
826: \reference{} Gustafsson, B., Bell, R.A., Eriksson, K. \& Nordlund, \AA.,
827: 1975, A\&A, 42, 407
828:
829: \reference{} Harris, W.~E., 1996, \aj, 112, 1487
830:
831: \reference{} Hauschildt, P.~H., Allard, F., Ferguson, J., Baron, E. \&
832: Alexander, D.~R., 1999, \apj, 525, 871
833:
834: \reference{} Houdashelt, M.~L., Bell, R.~A. \& Sweigart, A.~V., 2000, \aj, 119, 1448
835:
836: \reference{} King, J.~R., Stephens, A., Boesgaard, A.~M. \& Deliyannis, C.~P.,
837: 1998, \aj, 115, 666
838:
839: \reference{} Kraft, R.~P. 1994, \pasp , 106, 553
840:
841: \reference{} Kurucz, R.~L., 1992, CD-ROM 13
842:
843: \reference{} Leep, E.M., Oke, J.B. \& Wallerstein, G., 1987, \aj, 93, 338
844:
845: % \reference{} Lejeune, T. \& Schaerer, D.~S., 2000, A\&A, 366, 538
846:
847: % \reference{} Lebreton, Y., Perrin, M.-N.,
848: % Fernandes, J., Cayrel, R., Baglin, A. \& Cayrel de Strobel, G., 1997,
849: % in {\it HIPPARCOS: Venice '97}, ed. B.Battrick, pg. 379
850:
851: \reference{} Landolt, A.R., 1992, \aj, 104, 340
852:
853: \reference{} Mateo, M., Olszewski, E.W., Vogt, S.S. \& Keane, M.J.,
854: 1998, \aj, 116, 2315
855:
856: \reference{} McCarthy, J.~K., 1988, PhD thesis, California Institute of Technology
857:
858: \reference{} McClean, I.~S. \etal\, 1998, SPIE, 3354, 566
859:
860: \reference{} McClean, I.~S. \etal\, 2000, \pasp\ (in press)
861:
862: \reference{} Murphy, D.~C., Persson, S.~E., Pahre, M.~A.,
863: Sivaramakrishnan, A. \& Djorgovski, S.~G., 1995, \pasp, 107, 1234
864:
865: \reference{} Oke, J.~B., Cohen, J.~G., Carr, M., Cromer, J.,
866: Dingizian, A., ,Harris F.~H., Labrecque, S., Lucinio, R., Schaal, W.,
867: Epps, H., \& Miller, J. 1995, \pasp, 107, 307
868:
869: \reference{} Persson, S.~E., Murphy, D.~C., Krzeminsky, W., Roth, M.
870: \& Rieke, M.~J., 1998, \aj, 116, 2475
871:
872: \reference{} Peterson, R.~C. \& Latham, D.~W., 1986, \apj, 305, 645
873:
874: \reference{} Pinsonneault, M., 1997, \araa, 35, 557
875:
876: \reference{} Ram\'{\i}rez, S., Cohen, J.~G., Buss, R. \& Briley, M., 2001,
877: \aj, submitted
878:
879: \reference{} Schlegel, D.~J., Finkbeiner, D.~P. \& Davis, M., 1998, \apj, 500, 525
880:
881: \reference{} Shortridge, K., 1993, {\it {The Figaro 2.4 Manual} }
882:
883: \reference{} Sneden, C., Kraft, R.~P., Langer, G.~E., Prosser, C.~F. \&
884: Shetrone, M.~D., 1994, \aj, 107, 1773
885:
886: \reference{} Sneden, C., Kraft, R.~P., Prosser, C.~F. \& Langer, G.~E.,
887: 1991, \aj, 102, 2001
888:
889: \reference{} Stetson, P.~B., 2000, \pasp, 112, 925
890:
891: \reference{} Suntzeff, N.~B. \& Smith, V.~V., 1991, \apj, 381, 160
892:
893: \reference{} Sweigart, A.~V. \& Mengel, J.~G., 1979, \apj, 229, 624
894:
895: \reference{} VandenBerg, D.~A., Swenson, F.~J., Rogers, F.~J.,
896: Iglesias, C.~A. \& Alexander, D.~R., 2000, \apj, 532, 430
897:
898: \reference{} van't Veer-Menneret, C. \& Megessier, C., 1996, A\&A, 309, 879
899:
900: \reference{} Vogt, S.~E. \etal\, 1994, SPIE, 2198, 362
901:
902: \reference{} Yi, S., Demarque, P., Kim, Y.-C., Lee, Y.-W., Ree, C.
903: Lejeune, Th. \& Barnes, S., 2001, \apj\ (submitted) (Astro-ph/0104292)
904:
905: \end{references}
906:
907: \clearpage
908:
909:
910: %
911: % Table 1
912: %
913: \begin{deluxetable}{lrrrrrrrr}
914: \tablenum{1}
915: \tablewidth{0pt}
916: %\scriptsize
917: \tablecaption{The Sample of Stars in M71}
918: \label{tab1}
919: \tablehead{\colhead{ID\tablenotemark{a}} & \colhead{$V$} &
920: \colhead{Date Obs.} & \colhead{Primary} & \colhead{Al.}
921: & \colhead{Signal/pixel}
922: & \colhead{$v_r$} & \colhead{$\mu$ Prob.\tablenotemark{b}}
923: & \colhead{Notes} \nl
924: \colhead{} & \colhead{(mag)} & \colhead{} & \colhead{(sec)} & \colhead{(sec)}
925: & \colhead{(DN)\tablenotemark{c}} &
926: \colhead{(km s$^{-1}$)} & \colhead{(\%)} \nl
927: }
928: \startdata
929: 1-45 & 12.36 & Aug.1999 & 400 & 300 & 2425 & $-$19.0 & 99 \nl
930: I & 12.42 & Aug.1999 & 600 & 400 & 2495 & $-$13.9 & 99 \nl
931: 1-66 & 13.07 & Aug.1999 & 1000 & 400 & 3325 & $-$23.6 & 96 \nl
932: 1-64 & 13.12 & Aug.1999 & 700 & 1200 & 1535 & $-$17.1 & 99 \nl
933: 1-56 & 13.21 & Aug.1999 & 700 & ... & 770 & $-$21.2 & 99 \nl
934: 1-95 & 13.35 & Aug.1999 & 1800 & 900 & 3565 & $-$20.3 & 99 \nl
935: 1-81 & 13.68 & Aug.1999 & 1000 & 400 & 1750 & $-$24.3 & 99 \nl
936: Y & 13.95 & Aug.1999 & 3600 & ... & 8800 & $-$2.1~ & 0 & d \nl
937: 1-1 & 14.14 & Aug.1999 & 600 & 300 & 675 & $-$23.5 & 99 \nl
938: 1-80 & 14.45 & June 2000 & 900 & ... & 4690 & $-$20.7 & 99 & HB\nl
939: 1-87 & 14.47 & June 2000 & 2100 & 900 & 2940 & $-$22.8 & ...& HB \nl
940: 1-94 & 14.58 & Aug.1999 & 1800 & 900 & 975 & $-$24.6 & 98 & HB \nl
941: 1-60 & 14.55 & Aug.1999 & 1800 & ... & 1015 & $-$20.8 & 99 \nl
942: 1-59 & 14.71 & Aug.1999 & 1800 & ... & 875 & $-$24.1 & ... \nl
943: G53476\_4543 & 15.07 & Aug.1999 & 7200 & ... & 4300 & $-$22.5 & ... \nl
944: 2-160 & 15.14 & June 2000 & 2100 & 900 & 1600 & $-$25.2 & 5 \nl
945: G53447\_4707 & 15.16 & Aug.1999 & 7200 & 1200 & 3575 & $-$19.8 & 28 \nl
946: G53425\_4608 & 15.47 & Aug.1999 & 1200 & 1200 & 780 & +16.5 & 0 & d \nl
947: G53445\_4647 & 15.59 & June 2000 & 3600 & ... & 865 & $-$19.2 & 88 \nl
948: G53447\_4703 & 16.03 & Aug.1999 & 7200 & 1200 & 1515 & $-$27.2 \nl
949: G53425\_4612 & 16.32 & Aug.1999 & 1200 & 1200 & 335 & $-$21.0 \nl
950: G53477\_4539 & 16.33 & Aug.1999 & 7200 & ... & 1255 & $-$11.4 \nl
951: G53475\_4547 & 16.63 & Aug.1999 & 3600 & ... & 600 & +36.3 & & d \nl
952: G53457\_4709 & 16.75 & June 2000 & 4500 & ... & 2960 & $-$21.1 \nl
953: G53391\_4628 & 16.86 & Aug.1999 & 7200 & ... & 1210 & $-$21.3 \nl
954: G53394\_4624 & 16.95 & Aug.1999 & 4800 & ... & 700 & +3.4 & & d \nl
955: G53417\_4431 & 17.60 & Aug.1999 & 12000 & ... & 875 & $-$19.8 \nl
956: G53392\_4624 & 17.72 & Aug.1999 & 12000 & ... & 820 & $-$22.1 \nl
957: G53414\_4435 & 17.97 & Aug.1999 & 12000 & ... & 425 & $-$21.4 \nl
958: %
959: % exposures with secondary configuration included as part of primary set.
960: %
961: \enddata
962: \tablenotetext{a}{Identifications are from Arp \& Hartwick (1971)
963: or are assigned based on the J2000 coordinates, rh rm rs.s dd dm dd becoming
964: Grmrss\_dmdd.}
965: \tablenotetext{b}{This is the probability of membership assigned
966: by Cudworth (1985) on the basis of his proper motion survey.}
967: \tablenotetext{c}{The CCD gain is 2.4 e/DN. The signal is measured in
968: the continuum of the spectra taken with the primary HIRES configuration
969: near 6150\AA.}
970: \tablenotetext{d}{This star is presumed to not be a member of M71.}
971: \end{deluxetable}
972:
973:
974:
975: %
976: % Table 2
977: %
978: \begin{deluxetable}{llllrrr}
979: \tablenum{2}
980: \tablewidth{0pt}
981: %\scriptsize
982: \tablecaption{Photometry for the M71 Sample}
983: \label{tab2}
984: \tablehead{\colhead{ID\tablenotemark{a}} & \colhead{$V$} &
985: \colhead{$R$} & \colhead{$I$} & \colhead{$K$}
986: & \colhead{RA} & \colhead{Dec} \nl
987: \colhead{} & \colhead{(mag)} & \colhead{(mag)} & \colhead{(mag)} & \colhead{(mag)}
988: & \colhead{(J2000)} \nl
989: }
990: \startdata
991: \multispan{3}{Probable Members:} \nl
992: 1--45 & 12.36 & ... & ... & 8.12 & 19 53 48.37 & +18 48 00.3 \nl
993: I & 12.423 & ... & 10.76 & 8.56 & 19 53 44.74 & +18 46 35.1 \nl
994: 1--66 & 13.071 & ... & 11.49 & ... & 19 53 45.22 & +18 46 55.5 \nl
995: 1--64 & 13.122 & ... & 11.49 & 9.32 & 19 53 46.12 & +18 47 26.2 \nl
996: 1--56 & 13.21 & ... & 11.77 & 9.79 & 19 53 48.40 & +18 48 23.5 \nl
997: 1-95 & 13.35 & ... & 11.94 & ... & 19 53 41.01 & +18 46 04.8 \nl
998: 1--81 & 13.68 & ... & 12.184
999: & 10.22 & 19 53 45.48 & +18 46 49.7 \nl
1000: 1--1 & 14.14 & ... & 12.76 & 10.92 & 19 53 52.32 & +18 44 52.9 \nl
1001: 1--80 & 14.45 & 13.85 & 13.29 & ... & 19 53 44.20 & +18 46 48.0 \nl
1002: 1--87 & 14.47 & 13.83 & 13.28 & ... & 19 53 45.58 & +18 45 48.9 \nl
1003: 1--94 & 14.58 & 13.94 & 13.42 & 12.01 & 19 53 40.85 & +18 46 00.9 \nl
1004: 1--60 & 14.55 & 13.83 & 13.23 & 11.69 & 19 53 41.78 & +18 48 41.5 \nl
1005: 1--59 & 14.71 & 13.92 & 13.22 & 11.39 & 19 53 42.06 & +18 48 37.3 \nl
1006: G53476\_4543 & 15.07 & 14.38 & 13.77 & ... & 19 53 47.62 & +18 45 43.2 \nl
1007: 2--160 & 15.14 & 14.46 & 13.89 & ... & 19 53 45.19 & +18 48 33.2 \nl
1008: G53447\_4707 & 15.16 & 14.48 & 13.94 & 12.50 & 19 53 44.65 & +18 47 07.4 \nl
1009: G53445\_4647 & 15.59 & 14.88 & 14.31\tablenotemark{b} & ... & 19 53 44.50 & +18 46 47.0 \nl
1010: G53447\_4703 & 16.03 & 15.27 & 14.74 & 13.20 & 19 53 44.65 & +18 47 03.3 \nl
1011: G53425\_4612 & 16.32 & 15.66 & 15.13 & ... & 19 53 42.45 & +18 46 11.7 \nl
1012: G53477\_4539 & 16.33 & 15.66 & 15.10 & ... & 19 53 47.72 & +18 45 39.2 \nl
1013: G53457\_4709 & 16.75 & 16.02 & 15.57 & ... & 19 53 45.69 & +18 47 08.8 \nl
1014: G53391\_4628 & 16.86 & 16.19 & 15.60 & ... & 19 53 39.05 & +18 46 28.2 \nl
1015: G53417\_4431 & 17.60 & 17.05 & 16.52 & 15.49 & 19 53 41.72 & +18 44 31.2 \nl
1016: G53392\_4624 & 17.72 & 17.13 & 16.66 & ... & 19 53 39.18 & +18 46 23.9 \nl
1017: G53414\_4435 & 17.97 & 17.44 & ... & 15.78 & 19 53 41.37 & +18 44 34.8 \nl
1018: ~ \nl
1019: \multispan{3}{Probable Non-Members:} \nl
1020: Y & 13.95 & ... & 12.62 & 11.13 & 19 53 41.31 & +18 44 27.4 \nl
1021: G53425\_4608 & 15.47 & 14.68 & 14.08 & ... & 19 53 42.48 & +18 46 07.7 \nl
1022: G53475\_4547 & 16.63 & 16.13 & 15.80 & ... & 19 53 47.53 & +18 45 47.3 \nl
1023: G53394\_4624 & 16.95 & 16.35 & 15.91 & ... & 19 53 39.41 & +18 46 23.8 \nl
1024: \enddata
1025: \tablenotetext{a}{Identifications are from Arp \& Hartwick (1971)
1026: or are assigned based on the J2000 coordinates, rh rm rs.s dd dm dd becoming
1027: Grmrss\_dmdd.}
1028: \tablenotetext{b}{This star has $V, I$ from Stetson as well as from the
1029: short LRIS images, but Stetson's
1030: $I$ is very discrepant from that from the LRIS short images. This is
1031: the only such case found so far.}
1032: \end{deluxetable}
1033:
1034:
1035:
1036: %
1037: % Table 3
1038: %
1039: \begin{deluxetable}{lllllcl}
1040: \tablenum{3}
1041: \tablewidth{0pt}
1042: %\scriptsize
1043: \tablecaption{Stellar Parameters for the M71 Sample}
1044: \label{tab3}
1045: \tablehead{\colhead{ID\tablenotemark{a}} &
1046: \colhead{\teff\ (K)} & \colhead{\teff\ (K)} & \colhead{\grav} &
1047: \colhead{\teff\ (K)} &
1048: \colhead{\teff (K)} & \colhead{\teff(K)} \nl
1049: \colhead{} & \colhead{($V-K$)} & \colhead{($V-I$)} & \colhead{} &
1050: \colhead{(H$\alpha$)} &
1051: \colhead{(Weighted)\tablenotemark{b}} & \colhead{(Adopt)} \nl
1052: }
1053: \startdata
1054: 1--45 & 3950 & ... & 0.9 & ... & 3950 & 3950 \nl
1055: I & 4120 & 4175 & 1.0 & ... & 4150 & 4150 \nl
1056: 1--66 & ... & 4305 & 1.35 & ... & 4310 & 4250 \nl
1057: 1--64 & 4170 & 4225 & 1.35 & ... & 4200 & 4200 \nl
1058: 1--56 & 4475 & 4560 & 1.6 & ... & 4525 & 4525 \nl
1059: 1--95 & ... & 4630 & 1.65 & ... & 4630 & 4550 \nl
1060: 1--81 & 4450 & 4460 & 1.75 & ... & 4455 & 4550 \nl
1061: 1--1 & 4625 & 4710 & 2.05 & ... & 4670 & 4700 \nl
1062: 1--80\tablenotemark{c} & ... & 5290 & 2.45 & ... &
1063: 5290 & 5300 \nl
1064: 1--87\tablenotemark{c} & ... & 5205 & 2.45 & ... & 5205 & 5300 \nl
1065: 1--94\tablenotemark{c} & 5320 & 5290 & 2.45 & ... & 5315 & 5300 \nl
1066: 1--60 & 4980 & 4845 & 2.3 & ... & 4910 & 4900 \nl
1067: 1--59 & 4560 & 4480 & 2.3 & ... & 4520 & 4600 \nl
1068: % fixed typo, no change in Teff (5/6/2001)
1069: % 1--59 & 4560 & 4480 & 2.3 & ... & 4680 & 4600 \nl
1070: G53476\_4543 & ... & 4890 & 2.65 & ... & 4890 & 4900 \nl
1071: 2--160 & ... & 5175 & 2.7 & ... & 5175 & 5100 \nl
1072: G53447\_4707 & 5225 & 5130 & 2.75 & 5300 & 5200 & 5175 \nl
1073: G53445\_4647 & ... & 4960 & 2.85 & ... & 4960 & 5050 \nl
1074: G53447\_4703 & 5035 & 4920 & 3.0 & ... & 4985 & 5000 \nl
1075: G53425\_4612 & 5090 & 5205 & 3.15 & 5100 & 5140 & 5150 \nl
1076: G53477\_4539 & ... & 5090 & 3.15 & 5300 & 5160 & 5150 \nl
1077: G53457\_4709 & ... & 5240 & 3.35 & 5400 & 5290 & 5200 \nl
1078: G53391\_4628 & ... & 5010 & 3.35 & ... & 5010 & 5100 \nl
1079: % original table
1080: % G53417\_4431 & 6010 & 5580 & 4.05 & 5800 & 5800 & 5800 \nl
1081: % G53392\_4624 & ... & 5650 & 4.05 & 5800 & 5700 & 5800 \nl
1082: % G53414\_4435 & 5895 & ... & 4.15 & 5800 & 5860 & 5900 \nl
1083: % original table
1084: G53417\_4431 & 6010 & 5580 & 4.05 & 5800\tablenotemark{d} & 5800 & 5800 \nl
1085: G53392\_4624 & ... & 5650 & 4.05 & 5800\tablenotemark{d} & 5650 & 5800 \nl
1086: G53414\_4435 & 5895 & ... & 4.15 & 5800\tablenotemark{d} & 5895 & 5900 \nl
1087: \enddata
1088: \tablenotetext{a}{Identifications are from Arp \& Hartwick (1971)
1089: or are assigned based on the J2000 coordinates, rh rm rs.s dd dm dd becoming
1090: Grmrss\_dmdd.}
1091: \tablenotetext{b}{\teff\ from the H$\alpha$ profile has half weight.}
1092: \tablenotetext{c}{These are RHB stars in M71.}
1093: \tablenotetext{d}{The H$\alpha$ profiles are not used to derive \teff\
1094: for these stars.}
1095: \end{deluxetable}
1096:
1097: % log(g) lowered from 2.55 to 2.45 for RHB stars to take into
1098: % account mass 0.6 Msun, not 0.8 Msun.
1099:
1100:
1101:
1102: \clearpage
1103:
1104:
1105: \begin{figure}
1106: \epsscale{0.7}
1107: % Comment in the following line to embed the postscript figure into the manuscript
1108: \plotone{figure1.ps}
1109: \caption[figure1.ps]{A section of order 58 is shown for the brightest (at $V$)
1110: M71 star in our sample at the top of the figure and the faintest at the
1111: bottom. Starting with the brightest star, we display stars in
1112: the sample in increments of five in order of decreasing luminosity,
1113: omitting the RHB stars.
1114: \label{fig1}}
1115: \end{figure}
1116:
1117: \begin{figure}
1118: \epsscale{0.7}
1119: % Comment in the following line to embed the postscript figure into the manuscript
1120: \plotone{figure2.ps}
1121: \caption[figure2.ps]{To demonstrate that the Kurucz and MARCS
1122: grids of predicted colors are essentially identical,
1123: contours of $\Delta$\teff\ computed for $V-K$ are
1124: displayed. $\Delta$\teff\ is the difference between the \teff\ predicted
1125: from the ATLAS models of Kurucz (1992) and the MARCS models of Houdshelt,
1126: Bell \& Sweigart (2000) for a fixed abundance ([Fe/H] = $-$0.5 dex) and
1127: a $V-K$ color taken from the MARCS grid. The contour levels are
1128: 0, 30, and 60 K. The thick curve is a 12 Gyr isochrone for M71 taken
1129: from the Yale$^2$ tracks of Yi \etal\ (2001). See the text for details.
1130: \label{fig2}}
1131: \end{figure}
1132:
1133:
1134:
1135: \begin{figure}
1136: \epsscale{0.7}
1137: % Comment in the following line to embed the postscript figure into the manuscript
1138: \plotone{figure3.ps}
1139: \caption[figure3.ps]{The $V,~V-K$ color-magnitude diagram for M71 for the
1140: HIRES sample. Filled circles denote measurements
1141: from SCAM/NIRSPEC, open circles denote measurements from the P60 IR camera,
1142: and crosses denote the set of RGB and RHB stars from Frogel, Persson \&
1143: Cohen (1979). For each of the two stars with more than one independent
1144: observation at $K$, horizontal lines connect the pair of points.
1145: \label{fig3}}
1146: \end{figure}
1147:
1148:
1149:
1150:
1151: \begin{figure}
1152: \epsscale{1.0}
1153: % Comment in the following line to embed the postscript figure into the manuscript
1154: \plotone{figure4.ps}
1155: \caption[figure4.ps]{Profiles of H$\alpha$ from LRIS spectra
1156: are shown for a subgiant
1157: and for the average of three main sequence stars near the bright end of
1158: the sample of Cohen (1999).
1159: \label{fig4}}
1160: \end{figure}
1161:
1162:
1163: \begin{figure}
1164: \epsscale{0.8}
1165: % Comment in the following line to embed the postscript figure into the manuscript
1166: \plotone{figure5.ps}
1167: \caption[figure5.ps]{The \teff\ and \grav\ deduced here for the sample
1168: of M71 members with HIRES spectra is shown as is the
1169: 12 Gyr Yale$^2$ isochrone of Yi \etal\ (2001) for
1170: [Fe/H = $-0.7$ dex (solid curve). A distance of 3900 pc with a reddening
1171: $E(B-V) = 0.25$ mag has been adopted. The arrows in \grav\
1172: indicate the systematic error which is dominated by the contribution
1173: from the distance uncertainty; the internal error from star to
1174: star within M71 is considerably smaller. The error bars
1175: in \teff\ shown for
1176: the most and least luminous M71 stars in the HIRES sample are dominated by uncertainties in the reddening and are typical of the sample.
1177:
1178: \label{fig5}}
1179: \end{figure}
1180:
1181:
1182: \end{document}
1183:
1184:
1185: % Castilho, Pasquini, Allen, Barbuy \& Molaro, 2000, A\&A, 361, 92 (NGC 6397)
1186: