1: \documentclass[12pt,preprint]{aastex}
2: \usepackage{epsfig}
3: \shorttitle{Oxygen abundance of open cluster dwarfs}
4: \shortauthors{Shen, Liu, Zhang, Jones \& Lin}
5:
6:
7: \begin{document}
8: \title{ Oxygen abundance of open cluster dwarfs}
9:
10: \author{Z.-X. Shen}
11: \affil{Department of Astronomy, Peking University, Beijing 100871,
12: P. R. China;
13: {shenzx@bac.pku.edu.cn}}
14: \and
15:
16: \author{X.-W. Liu}
17: \affil{Department of Astronomy, Peking University, Beijing 100871,
18: P. R. China; liuxw@bac.pku.edu.cn}
19: \and
20:
21: \author{H.-W. Zhang}
22: \affil{Department of Astronomy, Peking University, Beijing 100871,
23: P. R. China;}
24: \and
25:
26: \author{B. Jones}
27: \affil{ UCO/Lick Observatory, Departments of Astronomy and Astrophysics, University
28: of California, Santa Cruz CA 95064, U.S.A.}
29: \and
30:
31: \author{D. N. C. Lin}
32: \affil{ UCO/Lick Observatory, Departments of Astronomy and Astrophysics, University
33: of California, Santa Cruz CA 95064, U.S.A.}
34:
35: \begin{abstract} We present oxygen abundances of dwarfs in the young open
36: cluster IC\,4665 deduced from the \ion{O}{1} $\lambda$7774 triplet lines and of
37: dwarfs in the open cluster Pleiades derived from the [\ion{O}{1}] $\lambda$6300
38: forbidden line. Stellar parameters and oxygen abundances were derived using
39: the spectroscopic synthesis tool SME (Spectroscopy Made Easy). We find a
40: dramatic increase in the upper boundary of the \ion{O}{1} triplet abundances
41: with decreasing temperature in the dwarfs of IC\,4665, consistent with the
42: trend found by Schuler et al. in the open clusters Pleiades and M\,34, and to a
43: less extent in the cool dwarfs of Hyades (Schuler et al. 2006a) and UMa (King
44: \& Schuler 2005). By contrast, oxygen abundances derived from the [\ion{O}{1}]
45: $\lambda$6300 forbidden line for stars in Pleiades and Hyades (Schuler et al.
46: 2006b) are constant within the errors. Possible mechanisms that may lead a
47: varying oxygen triplet line abundance are examined, including systematic errors
48: in the stellar parameter determinations, the NLTE effects, surface activities
49: and granulation. The age-related effects stellar surface activities
50: (especially the chromospheric activities) are suggested by our analysis to
51: blame for the large spreads of oxygen triplet line abundances.
52:
53: \end{abstract}
54:
55: \keywords{Open cluster and associations: individual (IC\,4665, Pleiades) --
56: stars: abundances -- stars: activities}
57:
58: \section{Introduction}
59:
60: Oxygen is an abundant element of particular relevance for the Galactic chemical
61: evolution and its formation history. Oxygen in long-lived low-mass stars
62: represents the chemical composition of the gas at the time those stars are
63: formed and carries information of the chemical history of stars in different
64: populations in our Galaxy. It is a bona fide primary element formed exclusively
65: in the interiors of massive stars and then released to the interstellar medium
66: (ISM) via Type\,II supernova explosions. By analyzing oxygen abundances
67: relative to the iron group elements in the atmosphere of different type stars,
68: one can trace the history of SN\,II feedback and the occurrence rates of
69: Type\,II and Type\,Ia supernova explosions.
70:
71: Recent reviews on stellar oxygen abundance analyses are given by King (1993),
72: Israelian et al. (1998), Nissen et al. (2002) and Takeda (2003). Oxygen has a
73: limited number of lines in the visual part of stellar spectra. Apart from
74: molecular OH lines in the ultraviolet (UV) and in the infrared (IR), oxygen
75: abundances are traditionally determined from the high-excitation \ion{O}{1}
76: 3s\,$^5$S$^{\rm o}$ -- 3p\,$^5$P $\lambda\lambda$7771.94, 7774.16, 7775.39
77: triplet permitted lines or from the much weaker [\ion{O}{1}] 2p$^4$\,$^3$P --
78: 2p$^4$\,$^1$D $\lambda\lambda$6300.30, 6363.78 forbidden lines. Both methods
79: have their advantages and disadvantages. The triplet lines are strong and
80: therefore easy to measure, and are free from blending effects. However, their
81: strengths can be significantly affected by effects such as deviations from the
82: local thermal equilibrium (LTE) (e.g., Kiselman 1993). On the other hand,
83: although the $\lambda\lambda$6300,6363 forbidden lines are believed to be free
84: from the non-LTE (NLTE) effects (e.g., Kiselman 1991; Takeda 2003), they are
85: much weaker than the triplet lines and blended with lines from other species.
86: The stronger component $\lambda$6300 of the doublet is blended with a
87: \ion{Ni}{1} line at 6300.34\,\AA\,(Lambert 1978; Johansson et al. 2003),
88: whereas the weaker $\lambda$6363 line can be contaminated by a weak CN red
89: system line (Lambert 1978). Discrepancies between oxygen abundances deduced
90: from the two indicators for field stars have long been observed and the possible causes have
91: been much discussed. In particular, in metal-poor stars, a distinct trend has
92: been reported in the sense that the permitted lines tend to yield
93: systematically higher oxygen abundances than the forbidden lines (e.g., Cavallo
94: et al. 1997). An inspection of the parameter-dependence of the discordance
95: indicates that the extent of the discrepancy tends to be comparatively lessened
96: for stars of high $T_{\rm eff}$ and/or ${\rm log}\,g$ (Takeda 2003).
97:
98: Many efforts have been attempted to resolve the discrepancy. King (1993)
99: suggested that the temperature scale for metal-poor dwarfs is probably
100: 150-200\,K too low and revised the scale upward to bring the oxygen abundances
101: derived from the two indicators into better agreement. However, the
102: assumptions made by King (1993) in determining the theoretical colors were
103: challenged by Balachandran \& Carney (1996). A controversial factor
104: in oxygen abundance determinations has been the role of the NLTE corrections.
105: Kiselman \& Nordlund (1995) re-examined the treatment of the NLTE effects in
106: earlier studies and concluded that the cross sections for inelastic collisions
107: with neutral hydrogen may have been overestimated. They argued that while
108: adjusting stellar parameters cannot resolve the oxygen abundance discrepancy,
109: it can be by resorting to more realistic 3D models. NLTE corrections of oxygen
110: abundances deduced from the triplet lines have been performed by many
111: investigators (e.g., Tomkin et al. 1992; Mishenina et al. 2000; Carretta et
112: al. 2000; Nissen et al. 2002; Takeda 2003). It is shown that the formation of
113: the \ion{O}{1} triplet is quite simple in the sense that only the transition
114: related quantities are important, but not the details of the atomic model. It
115: is also found that the NLTE effects of those particular lines can be well
116: described by a classic two-level-atom model and that the metallicity cannot be
117: an essential factor (c.f. Takeda 2003 and the references therein).
118:
119: While the physical cause of the discrepancy between the oxygen abundances
120: derived from the two types of line remains an open question, a modest view is
121: that the abundances from the forbidden lines are probably less problematic
122: given their weak $T_{\rm eff}$-sensitivity and negligible NLTE effects, and should
123: therefore be more reliable than those derived from the permitted lines. On the
124: other hand, given the weakness of the forbidden lines, the permitted lines
125: remain indispensable, in particular for the analysis of warm stars and for
126: stars of low metallicities.
127:
128: Members of an open cluster are assumed to form from a chemically homogeneous
129: cloud in a short time scale, such that they should all have the same chemical
130: composition. Stars in a cluster are distinguished only by mass and thus provide
131: good test beds to probe the underlying physical cause of the discrepant oxygen
132: abundances determined from the permitted lines and from the forbidden lines.
133: Schuler et al. (2004) reported their analysis of oxygen abundances for a
134: sample of late F, G and K dwarfs of the open clusters Pleiades (100 Myr) and
135: M\,34 (250 Myr). They find a dramatic increase of the \ion{O}{1} triplet
136: abundance with decreasing effective temperature in both clusters. Later on,
137: similar but to a less extent trends are found in UMa (King \& Schuler 2005) and
138: Hyades (Schuler et al. 2006a; for stars with $T_{\rm eff} \la$ 6000\,K). In
139: Hyades, oxygen triplet abundances of stars with $T_{\rm eff} \ga$ 6000\,K
140: increase with increasing temperature due to NLTE effect. The phenomenon were
141: not found in the other open clusters probably because of their limited star
142: samples. In contrast to the trend in oxygen triplet line abundances,
143: [\ion{O}{1}] $\lambda$6300 line abundances of three Pleiades stars (Schuler et
144: al. 2004) and 8 Hyades stars (Schuler et al. 2006b)
145: are nearly constant and yield average values much lower compared to the triplet
146: line results. Schuler et al. (2004, 2006a) suggest that surface inhomogeneities
147: rather than chromospheric activities are possibly the main cause of the
148: anomalous oxygen triplet line abundances. On the other hand, Morel \& Micela
149: (2004) compare oxygen triplet line and forbidden line abundances for stars
150: spanning a wide range of activity level and find that the magnitude of the
151: abundance discrepancy increases with increasing level of chromospheric/coronal
152: activities.
153:
154: Here we present oxygen triplet line abundances for a sample of dwarfs
155: of the open cluster IC\,4665 (35 Myr, Mermilliod 1981). In our
156: previous analysis (Shen et al. 2005), we showed that within the
157: measurement uncertainties the iron abundance is uniform, with a
158: standard deviation of 0.04 dex. This upper limit in the dispersion of
159: [Fe/H] among the IC 4665 member stars was used to infer that the total
160: reservoir of heavy elements retained by the nascent disks is limited.
161: Nevertheless, gas giant planets can form through core-accretion in
162: protostellar disks around these stars near snow lines where the
163: surface density of water ice may be significantly enhanced (Ida \& Lin
164: 2004, Ciesla \& Cuzzi 2006). In this case, the emergence of gas giants
165: may introduce a dispersion in [O/H] among the cluster member stars.
166:
167: A first step in this difficult task is to calibrate the Oxygen
168: abundance as a function of stellar parameters. In addition to our
169: sample of the IC 4665 member stars, forbidden line oxygen abundances
170: for several Pleiades dwarfs are also obtained. Section~2 describes the
171: observations and procedures of data reduction. Oxygen abundances
172: determined from the two types of line are presented in Section~3.
173: Possible explanations for the abundance determination discrepancy are
174: discussed in Section~4, followed by a brief summary in Section~5.
175:
176: \section{Observations and data reduction}
177:
178: Observations of both clusters were carried out in October 1999 and October 2000
179: using the HiRes spectrograph (Vogt 1992)
180: mounted on the Keck I 10\,m telescope. The spectra were recorded on
181: a Tektronix $2048\times 2048$ CCD of $24\times 24$\,$\mu{\rm m}$ pixel size.
182: For stars of IC\,4665, the spectra spanned a wavelength range from $\sim 6300$
183: to 8730\,\AA, split into 16 orders, with small interorder gaps amongst
184: them. The integration time ranged from 10 minutes to half an hour, yielding
185: signal-to-noise ratios ranging from $\sim 30$ to 150 per resolution element.
186: For Pleiades stars, the spectra covered from $\sim 4500$ to 6900\,\AA\,
187: with signal-to-noise ratios of $\sim$ 200 to 400. The Pleiades spectra have
188: previously been used by Wilden et al. (2002) to study the metallicity
189: dispersion amongst the member stars. A detailed description of those spectra
190: can be found there. All spectra have a resolving power of about 60,000.
191:
192: All spectra were reduced using {\sc iraf} in a similar manner as described in
193: Soderblom et al. (1993). The noao.imred.echelle package was used for
194: flat-fielding, scattered light removal, order extraction and wavelength
195: calibration. The latter was achieved using exposures of a Th-Ar lamp.
196:
197: We did not include all 18 IC\,4665 stars which have been analysis in our
198: first paper devoted to IC\,4665
199: (Shen et al. 2005; Paper~I thereafter). The three cool stars P\,332, P\,349 \& P\,352 are found to have
200: abnormal stronger ${\rm H}\alpha$ absorption than that of normal dwarfs at the
201: same temperatures. Given that the membership of the three stars are only based
202: on the color-magnitude diagram, it is likely to us that they are
203: background subgiants but not members in IC\,4665. We thus exclude them from our
204: sample so that they would not affect our oxygen abundance analysis results.
205: Further proper motion detection is need to clarify the membership of them.
206: The existence of these subgiants in our sample would not affect our conclusions in Paper~I.
207:
208: By performing on our spectra the cross-correlation analysis with solar
209: spectrum, we found that P\,19, our hottest star is a binary. The separation of
210: the two members on the spectrum is 1.5\,\AA. Based on the relative flux at
211: ${\rm H}\alpha$, we suggest that the companion is a hot A type star. Given that
212: A stars have much weaker flux than the late F stars at the red part of the
213: spectrum and the spectrum line separation of them are large, the existence of a
214: companion would not affect our analysis results either on stellar parameters or
215: on element abundances. P\,19 is thus kept in our sample.
216:
217: \section{Abundance determinations}
218:
219:
220: Spectral synthesis analyses were carried out using the software package SME
221: (Spectroscopy Made Easy) developed by Valenti \& Piskonov (1996). SME can be
222: used to determine stellar and atomic parameters by matching the synthetic
223: spectrum to the observed one. It consists of a spectral synthesis code written
224: in C++ and a parameter optimization routine written in IDL. It uses Kurucz
225: stellar atmospheric models. The input parameters include $T_{\rm eff}$, ${\rm
226: log}\,g$, radial and rotational velocities, micro- and macro-turbulence
227: velocities, element abundances, and a list of spectral line atomic data (${\rm
228: log}\,gf$ and the van der Waals damping constants). The overall
229: metallicity, quantified by parameter [M/H], is used to interpolate a grid of
230: model atmospheres and to scale the solar abundance pattern (except for helium
231: and the elements with individually solved abundances) when calculating the
232: opacities. [M/H] is an independent model parameter, rather than a quantity
233: constructed from the abundances of individual elements. Effects on stellar
234: atmosphere caused by deviations of individual element abundances from the solar
235: abundance pattern are neglected. This is appropriate as the current work only
236: deals with stars of approximately solar metallicity. We assume the [M/H] of
237: stars in IC\,4665 and Pleiades to be zero. SME solves the radiative transfer to generate
238: a synthetic spectrum. A nonlinear least squares algorithm is then used to solve
239: for any subset of the aforementioned input parameters. The radiative transfer
240: routine assumes LTE, no opacity from molecular lines and negligible magnetic
241: field.
242:
243: A 10\,\AA\ wide spectral segment centered at 7774\,\AA\, was selected to solve
244: for the \ion{O}{1} triple line abundance. Fig.~\ref{otrip} compares the
245: observed (histogram line) and SME synthesized (smooth curve) profiles of the
246: oxygen triplet lines in four stars, spanning a wide range of effective
247: temperatures and rotational velocities. The deduced oxygen abundances are
248: presented in Table~\ref{triplet} [assuming a solar oxygen abundance O = 8.87 on
249: a logarithmic scale where H = 12.00 (Grevesse, Noels \& Sauval 1996)]. In
250: Table~\ref{triplet}, the stellar parameters $T_{\rm eff}$, $\log\,g$ and
251: micro-turbulent velocities are listed in Columns (2)--(4), respectively.
252: Columns (5)--(6) give, respectively, the ${\rm H}\alpha$ emission fluxes taken
253: from Mart\'{i}n \& Montes (1997) and the logarithmic X-ray luminosities divided
254: by the bolometric luminosity $L_{\rm bol}$, log $(L_{\rm X}/L_{\rm bol})$,
255: taken from Giampapa, Prosser \& Fleming (1998). Column (7) gives the amplitude
256: of modulation in the V-band from Allain et al. (1996). Columns (8)--(9) give,
257: respectively, the oxygen triplet line abundances and their uncertainties for
258: three stars, selected to represent different temperature regimes spanned by our
259: sample stars. Column (10) gives the oxygen triplet abundances deduced from the
260: equivalent width measurements. A detailed description of the procedures used
261: to determine the stellar parameters and element abundances has been presented
262: in our Paper~I.
263: Full results, including stellar parameters and abundances of other individual
264: elements are presented there along with a detailed error analysis.
265:
266: To investigate the discrepancies between the oxygen abundances derived
267: from the permitted triplet lines on the one hand and those deduced from the
268: forbidden lines on the other, the simplest way is to directly compare the two
269: set of abundances. However, our current set of spectra of IC\,4665 were
270: hampered by limited wavelength coverage (the stronger component $\lambda$6300
271: of the [O~{\sc i}] $\lambda\lambda$6300,6363 fell on the edge of the spectra)
272: and by relatively low S/N ratios, and we were unable to derive oxygen
273: abundances from the weak $\lambda\lambda$6300,6363 forbidden lines. Instead, we
274: have used the [\ion{O}{1}] $\lambda$6300 abundances deduced for dwarfs of the
275: Pleiades. This is reasonable, given that the two clusters have very similar
276: metallicities and ages. The Pleiades spectra of Wilden et al. (2002) did not
277: cover the oxygen triplet lines, yet the much higher S/N ratios of those spectra
278: allowed us to determine oxygen abundances using the [\ion{O}{1}] $\lambda$6300
279: forbidden line for six of them, i.e. H\,II\,129, 250, 522, 1039, 1298, and
280: 2462. In the current analysis, we have adopted the stellar parameters determined by
281: Wilden et al. (2002).
282:
283: The [\ion{O}{1}] $\lambda$6300 line is known to be contaminated by a weak
284: \ion{Ni}{1} line (Lambert 1978; Johansson et al. 2003). To be consistent with
285: the previous work of Schuler et al. (2004), we modeled the blending
286: \ion{Ni}{1} line assuming a ${\rm log}\,gf = -2.31$ (Allende Prieto et al.
287: 2001). Apart from the \ion{Ni}{1} line, between the narrow wavelength range of
288: 6300.1--6300.5\,{\AA} where the [O~{\sc i}] $\lambda$6300 line falls, Kurucz
289: lists dozens CN lines on his web
290: site\footnote{http://kurucz.harvard.edu/LINELISTS/LINESMOL/}. However these
291: weak CN lines are believed to be only important for the continuum opacities and
292: so far no cases of significant contribution of CN lines to the observed
293: [\ion{O}{1}] $\lambda$6300 line strengths have been reported for solar type
294: stars (see, e.g., Allende Prieto et al. 2001; Schuler et al. 2004). We have
295: also considered possible contamination of the [\ion{O}{1}] $\lambda$6300 line
296: by telluric absorption. From the line list of the solar spectrum of Moore et al.
297: (1966), the nearest telluric absorption line to the [\ion{O}{1}] line is at
298: 6299.23\,\AA, in addition to two unidentified absorption lines at
299: 6299.41 and 6299.59\,\AA. Pleiades has a heliocentric radial velocity of
300: $5.41\pm 0.37$\,km\,s$^{-1}$ (Kharchenko et al., 2005), leading to a blue-shift
301: of $-16.6$\,km\,s$^{-1}$, or $-0.35$\,{\AA} in wavelength at 6300\,{\AA} at the
302: time the spectra were taken, in good agreement with the average value of
303: 0.3\,\AA\ determined from nearby strong lines in our spectra. We thus conclude
304: that it is unlikely that our measurements of the [\ion{O}{1}]
305: $\lambda$6300.30 line were affected by the nearby telluric absorption line and the two nearby lines of unknown
306: origin.
307:
308: The $\lambda$6300 feature was well fitted using the atomic data for the
309: [\ion{O}{1}] $\lambda$6300 line retrieved from the VALD (Vienna Atomic Line
310: Database; Piskunov et al. 1995; Ryabchikova et al. 1999; Kupka et al. 1999),
311: after taking into account contribution from the blending \ion{Ni}{1} line. As
312: such, we have not solved the atomic data for the [\ion{O}{1}] $\lambda$6300
313: line, but instead adopted those retrieved from the VALD. In order to reduce the
314: systematic errors, we re-measured the Ni abundances in the six stars using
315: \ion{Ni}{1} lines in the wavelength segments 5575--5595, 5636--5657, and
316: 6107--6132\,\AA. Atomic data for Ni lines in those wavelength segments were
317: solved using the National Solar Observatory solar spectrum (Kurucz et al.
318: 1984). The Ni abundances ${\rm log}\,N({\rm Ni})$ derived by us (H\,II\,129:
319: 6.11; H\,II\,250: 6.11; H\,II\,522: 6.15; H\,II\,1039: 6.13; H\,II\,1298: 6.15;
320: H\,II\,2462: 6.12) agree with those obtained by Wilden el al. (2002) within
321: 0.04\,dex. Contributions from the [\ion{O}{1}] forbidden line and from the
322: \ion{Ni}{1} line to the observed 6300\,\AA\, feature were synthesized
323: separately. The resultant synthetic spectra as well as their sum are plotted in
324: Fig.~\ref{pspectrum} and compared to the observed ones. The oxygen abundances
325: thus derived from the $\lambda$6300 forbidden line are listed in
326: Table\,\ref{forbid}.
327:
328: Stellar elemental abundances are traditionally derived using the
329: equivalent width (EW) method. For comparison, we have also determined the
330: oxygen triplet abundances from the measured EWs of the \ion{O}{1} triplet
331: lines. The results are listed in Table~\ref{triplet}. The EWs were measured using multiple Gaussian fits. The oxygen
332: abundances were then derived using ABONTEST8, a program developed by P. Magain.
333: ABONTEST8 uses LTE plane-parallel model atmospheres given by Kurucz (1993). We
334: have adopted stellar parameters deduced from SME. Uncertainties of the results
335: were dominated by those in the measured EWs, which were estimated using the
336: equation given by Norris et al. (2001). For cool stars in our sample, the
337: spectra had typical S/N's between 30--50 and uncertainties in the derived
338: oxygen abundances were estimated to be about 0.15--0.25\,dex. For warm stars,
339: the spectra had better S/N's and the uncertainties are smaller. Compared to SME
340: values, oxygen abundances derived from the EW method are found to be 0.15\,dex
341: lower on average, with a standard deviation of 0.11\,dex and a maximum
342: difference of 0.4\,dex. Larger deviations are obtained for stars with lower
343: temperatures.
344:
345: We have also measured the EWs of the [\ion{O}{1}] $\lambda$6300 line and
346: calculated oxygen abundances. Contributions from the blending \ion{Ni}{1} line
347: to the observed EWs of the [\ion{O}{1}] $\lambda$6300 line were estimated and
348: corrected for using the EWs of the \ion{Ni}{1} $\lambda$6300.34 line measured
349: on SME model spectra. The results are listed in Table~\ref{forbid}. Within the
350: uncertainties, they are consistent with the SME values. The mean value of the
351: EW-based abundances is $-$0.05\,dex, only 0.04\,dex lower than the average of
352: the SME-based abundances. Note that the $\lambda$6300 absorption line falls in
353: the extended wings of two nearby strong \ion{Si}{1} $\lambda$6299.60 and
354: \ion{Fe}{1} $\lambda$6301.50 lines. The presence of these strong lines makes the
355: placement of local continuum rather difficult, leading to large uncertainties
356: in EW measurements, in particular for weak lines such as the [\ion{O}{1}]
357: $\lambda$6300 line. By contrast, these strong lines are well modeled
358: in SME and therefore abundances deduced using the SME method are less
359: affected by their presence.
360:
361:
362: \section{Discussion}
363:
364: Oxygen abundances relative to the solar value for a sample of 15 IC\,4665
365: stars, derived from the triplet permitted lines, are plotted in
366: Fig.~\ref{ic4665} against effective temperature $T_{\rm eff}$. Typical error
367: bars for three selected stars of representative values of $T_{\rm eff}$ are
368: indicated in the diagram. The upper boundary of the abundance show a clear
369: trend with $T_{\rm eff}$, increasing dramatically by about 1.3\,dex as $T_{\rm
370: eff}$ decreases from $\sim 6400$ to 4900~K. For stars of intermediate $T_{\rm
371: eff}$ ($\sim 5400$~K), the abundances exhibit a large scatter. A similar trend
372: is also found for EW-based triplet line oxygen abundances, which show
373: variations over a range of 1.0\,dex, only slightly smaller than the spread of
374: 1.3\,dex for the SME-based values. This demonstrates that the observed scatter
375: is not an artifact caused by our SME analysis method. The upper boundary trend
376: is similar to what found previously for dwarfs in Pleiades and M\,34 by Schuler
377: et al. (2004), and to a less extent for the cool dwarfs ($\la$6000\,K) in
378: Hyades by Schuler et al. (2006a) and UMa by King \& Schuler (2005) based on the
379: method of EW analysis. However, no large spread within small $T_{\rm eff}$
380: intervals has been reported in the other open clusters as in IC\,4665.
381:
382: In contrast to the triplet abundances, no such trend is found for oxygen
383: abundances deduced from the $\lambda$6300 forbidden line. In
384: Fig.~\ref{pleiades} we plot the [\ion{O}{1}] abundances of six Pleiades stars
385: in our sample against $T_{\rm eff}$ and another three Pleiades stars previously
386: analyzed by Schuler et al. (2004). Also included in the figure are oxygen
387: forbidden line abundances of 8 Hyades stars taken from Schuler et al. (2006b)
388: based on stellar atmosphere models with overshoot. Fig.~\ref{pleiades} shows
389: clearly that oxygen abundances of Pleiades stars derived from the [\ion{O}{1}]
390: forbidden line have small scatter over the entire range of $T_{\rm eff}$, from
391: $\sim 4800$ to 5800~K. The three stars of Schuler et al. (2004) yield an
392: average [\ion{O}{1}] abundance of 0.14\,dex, which is 0.15\,dex higher than our
393: mean abundance of $-$0.01\,dex. The measured EWs of the [\ion{O}{1}]
394: $\lambda$6300 line in the sample of Schuler et al. (2004) are generally
395: 1--2\,m\AA\ stronger than ours. However, given the estimated observational
396: error of $\sim$3\,m\AA\ from Schuler et al. (2004), the EW differences seem not
397: so significant. I prefer that the abundance difference between the two works
398: are due to the systematic errors. Although the three coolest dwarfs
399: (4573$\leq$ $T_{\rm eff}$ $\leq$ 4834\,K) seem to increase with decreasing
400: $T_{\rm eff}$, [\ion{O}{1}] abundances of stars in Hyades do not show spreads
401: within the stated uncertainties. The Hyades [\ion{O}{1}] abundances are still
402: almost constant with a mean value of 0.19\,dex which is consistent with the
403: metallicity of Hyades ([Fe/H]=0.13; Paulson et al. 2003).
404:
405: Under the assumption that member stars of a given open cluster all have the
406: same chemical composition, an assumption that is indeed supported by
407: measurements of oxygen abundances using the [\ion{O}{1}] forbidden line, it is
408: probably reasonable to assume that the dramatic increase of the oxygen triplet
409: line abundances with decreasing $T_{\rm eff}$ as shown in Fig.~\ref{ic4665} is
410: an artifact and therefore the triplet line abundances do not represent the true
411: stellar surface oxygen abundances. Rather, it seems that they must have been
412: affected by some systematic effects which are a strong function of\, $T_{\rm
413: eff}$. Possible mechanisms that might be responsible for the abnormal triplet
414: line oxygen abundances include systematic uncertainties in the abundance
415: determinations, NLTE effects, granulation, surface activities (e.g.,
416: chromospheric activities, stellar spots). We examine these possibilities one by
417: one in the following subsections.
418:
419: \subsection{Abundance determination uncertainties}
420:
421: In Paper~I, we discuss our method of effective temperature determination and
422: compare the results with those deduced from the photometric method using color
423: indices. In our analysis based on SME, we use about one hundred iron lines of
424: different excitation energies when determining the effective temperature and
425: require that they yield self-consistent element abundances. SME however
426: does not directly force a balance in the excitation potential or ionization
427: states. Instead, it uses a Levenberg-Marquardt algorithm to minimize $\chi^2$
428: in a synthetic spectral model of the observed spectrum. During this process,
429: each line gives its own independent verdict regarding the free parameters,
430: regardless whether SME is used to solve stellar parameters such as $T_{\rm
431: eff}$, $\log\,g$ or element abundances. The standard deviations of the ensemble
432: of hundreds of lines provides estimates of the uncertainties in the derived
433: parameter. On the other hand, as previously found by Valenti \& Fischer
434: (2005), this algorithm generally grossly underestimates the true uncertainties.
435: The lack of a trend in [Fe/H] with $T_{\rm eff}$ (see Paper~I) gives a
436: post-check that the analysis has been properly carried out, in other words, we
437: have $T_{\rm eff}$ properly determined. The method has the advantage that the
438: results are insensitive to potential uncertainties in the stellar atmospheric
439: models and in the photometric measurements of the color indices. In the
440: spectroscopic method, the effective temperature and element abundances are
441: determined from the same set of observations, thus avoiding any potential
442: uncertainties introduced by combining observations obtained with different
443: instruments and telescopes in order to calibrate the color indices.
444:
445: In Paper~I, we estimate that SME temperatures are accurate to $\pm 100$~K.
446: The estimate was based on the fact that the differences between temperatures
447: derived from SME and from $(B-V)$ colors yield a standard deviation of only
448: 141~K and the fact that we believe that spectroscopic temperatures are more
449: accurate than color temperatures. By adding in quadrature the measurement error
450: in a single observation, and the numerical error yielded by the analysis
451: algorithm (the standard deviation of final parameter values about the mean for
452: each observation, due only to differences in initial values), Valenti \&
453: Fischer (2005) estimated the uncertainties for temperatures obtained by fitting
454: a single observation of a generic star to be 44~K. Their work was based on an
455: improved version of SME. Given that our spectra generally have lower S/N's than
456: theirs, it is probably realistic to assume that our estimates of $T_{\rm eff}$
457: are accurate to approximately 100~K. Values of $\log\,g$ were calculated using
458: Eq.\,(16.2) of Gray (1992) and assigned a nominal uncertainty of 0.1\,dex. Our
459: estimates of typical uncertainties of 100~K for $T_{\rm eff}$, 0.1\,dex for
460: $\log\,g$ and 30\% for $v_{\rm mic}$ are in line with what assumed by Wilden et
461: al. (2002) who also adopt SME in their analysis.
462:
463: The effects of uncertainties in individual stellar parameters on the derived
464: element abundances have been analyzed in Paper~I. The oxygen triplet permitted
465: lines are sensitive to effective temperature, surface gravity, and
466: microturbulent velocity determinations. In order to obtain an estimate of the
467: relative error of the triplet line oxygen abundance, we recalculated the
468: abundance by varying the effective temperature by amounts of $\pm 100$\,K,
469: ${\rm log}\,g$ by $\pm 0.1$\,dex and the microturbulent velocity by $\pm 30$\%.
470: Calculations were carried for three stars (here we changed P\,332 in Paper~I to
471: P\,71 as the former has been excluded from the sample. The calculated
472: uncertainty difference between P\,332 and P\,71 would not affect our results in
473: Paper~I.), selected to represent the full range of effective temperature of
474: stars in the current sample. Uncertainties introduced in the process of line
475: profile fitting were also estimated and incorporated into the final error
476: budget by varying the elemental abundance and then analyzing its effects on the
477: residuals of the fit. The results, tabulated in Table~\ref{triplet}, show that
478: even though the total formal errors of oxygen abundances derived from the
479: triplet lines can reach as much as 0.26\,dex for stars of effective
480: temperatures lower than 5500\,K, they are insufficient to account for the
481: observed systematic variations of the oxygen triplet line abundances of more
482: than 1\,dex as shown in Fig.~\ref{ic4665}.
483:
484: Valenti \& Fischer (2005) have used SME Version 2 to derive stellar parameters
485: and elemental abundances for a sample of over 1000 Galactic field stars, mainly
486: of dwarfs of spectral types F, G and K, using high quality spectra obtained as
487: a by-product of the California \& Carnegie Planet Search program. A thorough
488: error analysis was also carried out and some trends of the deduced elemental
489: abundances as a function of $T_{\rm eff}$ were found. They interpreted such
490: trends as spurious and used polynomial fits to remove them. In all cases, the
491: corrections amount to less than 0.2\,dex for the temperature range from 4900 to
492: 6400~K. Such trends are not unique to SME but also exist in other methods of
493: abundance analysis, such as those based on EWs (e.g. Schuler et al. 2003).
494: But more importantly, the range of variation of the triplet line oxygen
495: abundances observed in Fig.~\ref{ic4665} is so large such that we deem it
496: highly unlikely that the variations are of similar origins of the spurious
497: trends found in Valenti \& Fischer (2005).
498:
499: Uncertainties in the derived [O/H]$_{\rm forbid}$ abundances are dominated by
500: photon shooting errors as a result of the limited S/N ratios of the weak
501: diagnostic line. For oxygen abundances derived from the forbidden line,
502: uncertainties caused by errors in $T_{\rm eff}$ and in ${\rm log}\,g$ are
503: essentially identical for all the sample stars and amount to 0.02 and 0.04\,dex,
504: respectively, for stars in our sample. The [\ion{O}{1}] $\lambda6300$ line is very
505: weak and therefore not affected by microturbulent velocity determinations.
506: The total errors in the deduced
507: abundances are obtained by adding photon shooting errors and those introduced
508: by uncertainties in $T_{\rm eff}$ and in ${\rm log}\,g$ in quadrature. The
509: results are listed in the last column of Table\,\ref{forbid}.
510: Fig.~\ref{pleiades} shows that the abundances deduced from the [\ion{O}{1}]
511: forbidden line have a much smaller scatter compared to those derived from the
512: \ion{O}{1} triplet lines. The values vary over a range of 0.28\,dex with a
513: standard deviation of 0.1\,dex, comparable to the uncertainties of
514: 0.07--0.11\,dex for individual measurements. The sample yields an average [O I]
515: abundance of $-$0.01\,dex, i.e. nearly identical to the solar value.
516:
517: \subsection{NLTE effects}
518:
519: The \ion{O}{1} triplet lines are known to suffer from the NLTE effects. For a
520: wide range of stellar atmosphere parameters, the NLTE effects will lead to
521: overestimated oxygen abundances (e.g. Gratton et al. 1999; Takeda 2003). The
522: effects are however small, less than 0.1\,dex for cool ($T_{\rm eff} < 6000$~K)
523: dwarfs of approximately solar metallicity. For stars of higher temperatures,
524: the NLTE effect are expected to be larger (Takeda 2003). Thus the trend caused
525: by the NLTE effects is exactly opposite to what observed. It seems highly
526: unlikely that the nominal NLTE effects are responsible for the trend observed
527: in Fig.~\ref{ic4665}.
528:
529: Observations (e.g., King \& Boesgaard 1995) show that for stars with $T_{\rm
530: eff} \ga 6200$~K triplet-based oxygen abundances are enhanced compared to
531: forbidden line-based values, in agreement with the theoretical predictions of
532: Takeda (2003) for main-sequence stars of $T_{\rm eff} \sim 6500$~K. On the
533: other hand, the enhancement is small (about 0.18\,dex) and is likely to be lost
534: in the huge variations depicted in Fig.~\ref{ic4665} even if it exists in
535: IC\,4665. We note that P\,19 ($T_{\rm eff} = 6370$~K), the warmest star in our
536: sample, has the lowest oxygen abundance. Its spectrum is of high quality, with
537: a S/N of 150. SME and EW analyses give very similar abundances, [O/H] $=
538: -0.31$ and $-0.26$\,dex, respectively. According to the theoretical
539: calculations of Takeda (2003), the EWs of the oxygen triplet lines measured for
540: P\,19 implies an NLTE correction of $\sim -0.2$\,dex. Applying this correction
541: will lead to very low oxygen abundance of $\sim$ $-$0.5\,dex for P\,19. Even
542: the existence of a binary companion is not likely to affect its oxygen
543: abundance. It appears to us that P\,19 does seem to have an abnormally low
544: oxygen abundance compared to other member stars of the cluster.
545:
546: In their analysis of the open cluster M\,34, Schuler et al. (2003) find that Si
547: abundances derived from lines of high excitation potentials ($\chi \sim 6
548: $\,eV) increase with decreasing temperature, whereas abundances of Fe, Ti, Cr,
549: Ca, Al and Mg, all derived from lines of relatively low excitation potentials
550: ($\chi \approx 2$--4\,eV), decrease. They suggest that
551: overexcitation/ionization, caused by deviations from the LTE in cool stars, may
552: be responsible for the observed abundance trends. Similar trends are also
553: observed in our results of abundance analysis for IC\,4665 (Paper~I) and those
554: for the Pleiades by Wilden et al. (2002), where it is found that as $T_{\rm
555: eff}$ decreases, Si abundance, deduced from lines with $\chi$ $\sim$ 6\,eV,
556: increases, whereas Cr abundance, derived from lines of $\chi$ $\sim$
557: 0--4\,eV, decreases. There are however exceptions. For example, in IC\,4665,
558: Ti abundance determined from lines of $\chi \sim 0$--2\,eV increases as $T_{\rm
559: eff}$ decreases. Given the relatively high excitation potential of the oxygen
560: triplet lines, it is difficult to rule out the possibility that the observed
561: strong increase of the oxygen triplet line abundances with decreasing $T_{\rm
562: eff}$ is caused by overexcitation/ionization as proposed by Schuler et al.
563: (2003). Further quantitative investigations of this scenario, i.e. deviations
564: from the LTE can indeed result in overexcitation/ionization in cool stars, and
565: if so, its effects on elemental abundances derived from lines of different
566: excitation potentials, are highly desirable.
567:
568: \subsection{Surface activities}
569:
570: \subsubsection{Chromospheric activities}
571:
572: Takeda (1995) suggests that the oxygen triplet line formation could be affected
573: by chromospheric activities. In their analysis of 14 single-lined RS CVn
574: binaries, Morel \& Micela (2004) find suspiciously high \ion{O}{1} triplet line
575: abundances, by as much as 1.8\,dex, in chromospherically active systems. They
576: plot oxygen abundances deduced from the triplet lines and from the forbidden
577: line against chromospheric activity indicators, including strength of the
578: \ion{Ca}{2} H$+$K lines and X-ray flux. They find that for both RS CVn binaries
579: and stars of young open clusters, as the level of chromospheric activity
580: increases, the \ion{O}{1} triplet line abundance increases, whereas those
581: derived from the forbidden lines remains nearly constant around the solar
582: value. The results of Morel \& Micela strongly suggest that the steep
583: rise of the \ion{O}{1} triplet line abundance with decreasing $T_{\rm eff}$ is
584: probably caused by chromospheric activities.
585: However, one difficulty of the study was that all stars in their RS CVn binary
586: sample have effective temperatures near or below 5000\,K. As a result, Morel \&
587: Micela were unable to rule out the possibility that the observed large
588: variations of the oxygen triplet line abundances with $T_{\rm eff}$ was not
589: caused by inadequacy in the Kurucz LTE atmospheric models for late K-type stars.
590: On the other hand, a recent analysis of six K dwarfs and subgiants by
591: Affer et al. (2005) shows that the usage of LTE Kurucz model
592: atmospheres does not lead to large variations of oxygen triplet line
593: abundances in those old K stars which have low surface activities.
594:
595: From an analysis of more than 200 active stars in open clusters and in the
596: field, Messina et al. (2001) concluded that the strength of activities
597: increases from zero-age, reaching maxima around the age of Pleiades ($\sim$70
598: Myr), and then declines. If surface activities account for the observed
599: behavior of the triplet line abundance, then one would expect a correlation
600: between the linear fitting slope of the oxygen triplet line abundances as a
601: function of effective temperature and the age of the open clusters. We
602: calculated the slopes for stars with $T_{\rm eff}$ below 6000~K in each
603: cluster. Stars with $T_{\rm eff} >$ 6000~K are excluded because they are
604: believed to be affected by NLTE effect (e.g., Schuler et al. 2006a). From the
605: currently rather restricted amount of data, we find almost identical absolute
606: values of the slopes for the two younger open clusters Pleiades ($\sim$70\,Myr)
607: and M\,34 ($\sim$250\,Myr) of ($-8.2\pm2.1$) and ($-8.8\pm3.0)\times10^{-4}$
608: respectively, compared to the much smaller slopes of ($-3.0\pm0.7$) and
609: ($-5.1\pm0.4)\times10^{-4}$ for the older clusters UMa ($\sim$600\,Myr) and
610: Hyades ($\sim$600\,Myr). The slope of IC\,4665 ($\sim$35\,Myr) is also much
611: small ($(-4.4\pm2.0)\times10^{-4}$), comparable to those of UMa and Hyades.
612: The result seems to be consistent with the prediction of Messina et al. (2001).
613: We should notice that the small slope of IC\,4665 is due to the large scatter
614: around 5400\,K. Comparing the triplet abundances variation of different
615: clusters within the same $T_{\rm eff}$ range in Fig.~\ref{ic4665}, IC\,4665
616: stars show a much higher oxygen abundance upper boundary than the older
617: clusters UMa and Hyades. Although inconclusive, these results seem to suggest
618: that age-related effects (e.g., surface activities including chromospheric
619: activity, stellar spots and flares) is influencing triplet abundance
620: determinations for young open clusters.
621:
622: In order to clarify the role of chromospheric activity in the behavior of
623: oxygen triplet abundance in young open clusters, we calculated a factor R based on the emission of chromospheric activity sensitive lines ${\rm H}\alpha$ and
624: \ion {Ca}{2} IRT $\lambda\lambda$8498,8662. Montes \& Mart\'{i}n (1998)
625: provide a set of standard quiet stars observed at high-resolution thus we can
626: use them in the application of the spectral subtraction technique to obtain the
627: active-chromosphere contribution to these activity-sensitive lines.
628: We calculated R from the fluxes (F) of ${\rm H}\alpha$ and
629: \ion {Ca}{2} IRT $\lambda\lambda$8498,8662 by
630: %After
631: %measuring the emission fluxes of the ${\rm H}\alpha$ and
632: %\ion {Ca}{2} IRT lines $\lambda\lambda$8498,8662 lines by subtracting the spectra of the standards to the IC\,4665 spectra with similar $(B-V)$ color, we
633: %scaled the flux to the standard absorption line level.
634: \[
635: R = {F_{\rm IC\,4665}-F_{\rm standard} \over F_{\rm standard}}
636: \]
637: The standard quiet stars are chosen from the sample of Montes \& Mart\'{i}n (1998) to have similar $(B-V)$ as the IC\,4665 stars.
638: The results are listed in Table~\ref{R}. Observational errors are also estimated and listed in columns (5)-(7).
639: Column (8) of Table~\ref{R} gives the errors of the \ion{O}{1} triplet abundance for each star, calculated from
640: a linear fitting equation of $T_{\rm eff}$ and the uncertainties of the three
641: representative stars.
642:
643: The oxygen triplet abundance is plotted against the three activity indicators
644: R(H$\alpha$), R(8498), and R(8662) in Fig.~\ref{Hairt}. The linear correlation
645: coefficients of the triplet line abundances with the activity indicators are
646: calculated and listed in Table~\ref{active}. Given the correlation coefficients
647: of 0.69, 0.55 and 0.57 of the oxygen triplet abundances with the three
648: indicators respectively, only marginal correlations could be suggested for
649: them. However, from Fig.~\ref{Hairt}, we find that for stars with high R (high
650: activity), there does exist a clear correlation between triplet abundances and
651: the activity indicators, whereas the stars with low R show totally no
652: correlation.
653: Given that the stars which show near-zero or even nagative R meaning that they
654: should have stable chromospheric activity also have oxygen abundances
655: which are consistent with the [\ion{O}{1}] abundances within uncertainties,
656: it is likely that the oxygen triplet abundance variation of the four stars with
657: R smaller than 0.1 is just due to errors other than activities. If we
658: excluded these 4 stars whose triplet abundances are not affected by
659: chromospheric activities, the linear correlation coefficients of oxygen triplet
660: abundances with the activity indicators would increase to 0.76, 0.73 and 0.81
661: for the remaining stars, which are significant at the confidence levels of
662: 99.99\%, 99.98\% and 99.99\%, respectively. These are much convinced values to
663: support that chromospheric activity is responsible for the oxygen triplet
664: abundance behaviors. The only point which prevents the correlation coefficients
665: to increase over 0.90 is P\,107, a fast rotator whose emission level may be
666: underestimated.
667:
668: Besides our work, X-ray luminosities, normalized to unit bolometric luminosity,
669: are available for three of our stars, P\,60, P\,71 and P\,100 (Giampapa,
670: Prosser \& Fleming 1998). ${\rm H}\alpha$ emission fluxes, derived using the
671: technique of spectral subtraction, are available for seven of them (Mart\'{i}n
672: \& Montes 1997). The two chromospheric activity indicators are listed in
673: Table~\ref{triplet}.
674: %In Fig.~\ref{activ}, we plot the oxygen triplet line abundance against the
675: %normalized X-ray luminosity, log $(L_{\rm X}/L_{bol})$, and ${\rm H}\alpha$
676: %emission flux, $F({\rm H}\alpha)$.
677: Although the X-ray data are scarce, they do point to a positive correlation
678: with oxygen abundances determined from the triplet lines. The linear
679: correlation coefficient is 0.98 with a statistic significance of 87\%. Whereas
680: the ${\rm H}\alpha$ emission data show a large scatter without a clear
681: correlation between the oxygen triplet abundances and the ${\rm H}\alpha$
682: emission fluxes (the linear correlation coefficient is 0.34). Among the stars,
683: P\,94 has a low oxygen triplet line abundance and a low H$\alpha$ emission
684: flux. If we exclude the star with the lowest ${\rm H}\alpha$ emission flux
685: (P\,107), then the remaining six stars do show a positive correlation with a
686: linear correlation coefficient of 0.78 with a significance of 93\%.
687:
688: In addition to IC\,4665, we have also calculated the linear correlation
689: coefficients of oxygen triplet line abundances with different surface activity
690: indicators (H$\alpha$, X-ray, \ion{Ca}{2} H+K lines and infrared triplet lines,
691: and the amplitude of modulation in the V-band) for other open clusters. The
692: results are listed in Table.~\ref{active}. In the table, the last column gives
693: the references for the data used. Values in parentheses are numbers of stars
694: used in the calculations. For Pleiades stars, the oxygen triplet abundances
695: are found to be mildly correlated with the two chromospheric activity
696: indicators, X-ray luminosity and the \ion{Ca}{2} infrared triplet (IRT) line
697: strength, but not with H$\alpha$ emission. For M\,34 stars, the triplet
698: abundances are mildly correlated with the \ion{Ca}{2} IRT line strength, but
699: not with H$\alpha$ emission. For UMa stars, the oxygen triplet abundances are
700: mildly correlated with the \ion{Ca}{2} H+K line strength. For Hyades stars, no
701: correlations are found between oxygen triplet abundances and the X-ray emission
702: or the \ion{Ca}{2} H+K line strength. Except for IC\,4665 where the oxygen
703: triplet abundances are found to be correlated with the X-ray luminosities in
704: three stars, no obvious correlations are found for other open clusters. The
705: correlations between oxygen abundances in different clusters and different
706: chromospheric activity indicators are thus too ambiguous to draw a firm
707: conclusion. One should remember that in all cases but our work, the stellar
708: spectra and the chromospheric activity indicators were not measured at the same
709: epoch. Therefore, we really do not know the activity levels of the stars at the
710: time the spectra were taken. Simultaneous observations of oxygen triplet lines
711: and chromospheric activity indicators are essential to clarify the situation.
712:
713: In the previous works, one fact that made the situation further complicated is
714: that both oxygen triplet abundances and chromospheric activity indicators are
715: correlated with $T_{\rm eff}$. Thus the triplet abundance-activity correlation
716: may be a consequence of the $T_{\rm eff}$ correlations, prohibiting a strong
717: conclusion from being made about a causal relationship between activity
718: indicators and anomalous triplet abundances. However, in our sample of
719: IC\,4665, the correlation between oxygen triplet abundance and $T_{\rm eff}$
720: are not so monotonous (the linear correlation coefficient is 0.53) as there is
721: a large spread around 5400\,K. At the same time, the correlation of the
722: abundance and chromospheric activity indicators are much better for stars which
723: show chromospheric activity emissions (the coefficient $\sim$ 0.75). This
724: would be a strong evidence to solve the degeneracy between the triplet
725: abundance/activity correlation and the abundance/$T_{\rm eff}$ correlation.
726: The results support chromospheric activity as the main reason for the abnormal
727: oxygen triplet abundance observed in young open clusters. Furthermore,
728: theoretical models investigating the possible effects of stellar activities on
729: line formation and element abundance determinations are still needed to totally
730: disentangle the problem.
731:
732: \subsubsection{Stellar spots}
733:
734: Atmospheric perturbations (e.g., stellar spots, flares) have been invoked to
735: explain at least partially the large Li abundance spread in cool members of
736: young open clusters (Barrado y Navascu\'{e}s et al. 2001; Ford et al. 2002;
737: Xiong \& Deng 2005). Surface activities such as variations in spot filling
738: factor may affect element abundance determinations in two ways. Firstly, they
739: can cause color anomalies (Stauffer et al. 2003) and thus introduce
740: uncertainties in effective temperature determinations. In SME, effective
741: temperatures are determined directly from the absorption line spectrum, rather
742: than from the stellar energy distribution, i.e. color indices. Thus to the
743: first approximation, color anomalies caused by atmospheric perturbations are
744: irrelevant to our analyses based on the technique of SME. Alternatively,
745: atmospheric perturbations can influence directly the line formation process.
746: Using a simple model consisting of arbitrarily chosen line flux contributions
747: from cool and hot spots, Schuler et al. (2006a) found that stars with spots of
748: different temperatures is capable to reproduce the observed equivalent widths
749: for 3 Hyades stars with different $T_{\rm eff}$ and triplet abundances.
750: However, how such effects can lead to a correlation between oxygen triplet
751: abundances and $T_{\rm eff}$, as depicted in Fig.~\ref{ic4665}, is still not
752: clear.
753:
754: Based on the model of Bouvier et al. (1993), the observed amplitude of
755: modulation in the V-band could be a crude estimation of the area of the spot
756: coverage on the surface. The amplitudes of the V-band variation for 5 of our
757: sample stars given by Allain et al. (1996) are listed in Table ~\ref{triplet}.
758: The linear correlation coefficient between them and the triplet abundances is
759: 0.46 with a statistic significance of 56\%, which leaves the correlation
760: between spots and the triplet abundances still unrecognized. However, we should
761: notice that again the photometric and spectroscopic observations are not
762: carried out at the same time. Simultaneous observations are still needed to
763: draw a clear conclusion.
764:
765: Various studies have indicate that there are correlations between the
766: activities of different stellar atmosphere layers (e.g. Messina et al.
767: 2003). Therefore even if we obtained the correlation between the
768: oxygen triplet abundances and chromospheric activity levels, we cannot
769: tell if the correlation are due to chromospheric activity or
770: activities of the other layers (such as spots).
771:
772:
773: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
774: \subsection{Granulation corrections}
775:
776: Photospheric temperature fluctuation is another controversial factor that may
777: affect spectroscopic abundance determinations. Detailed calculations of
778: granulation abundance corrections have been carried out in recent years. The
779: 3D abundance corrections to be applied to the 1D solutions for individual
780: oxygen lines have been given by several works (e.g., Kiselman \& Nordlund
781: 1995; Asplund 2001; Allende Prieto et al. 2001; Asplund \& Garc\'{i}a P\'{e}rez
782: 2001; Nissen et al. 2002). Kiselman \& Nordlund (1995) report negligible 3D
783: effects for the high-excitation \ion{O}{1} triplet lines, which form deep in
784: the photosphere. Similar results are found by Nissen, Primas \& Asplund
785: (2001). Nissen et al. (2002) show that 3D correction for the [\ion{O}{1}]
786: forbidden line $\lambda$6300 is negligible in most cases, except in low
787: gravity/metallicity atmospheres where it can increase to about 0.2\,dex.
788: Steffen \& Holweger (2002) study the problem of LTE line formation in the
789: inhomogeneous solar photosphere based on detailed 2-dimensional radiation
790: hydrodynamics simulations of the convective surface layer of the Sun. By means
791: of a strictly differential 1D/2D comparison of the emergent equivalent widths,
792: they obtain the granulation abundance corrections for individual lines to be
793: applied to the standard abundances determined based on a homogeneous 1D model
794: atmosphere. For the oxygen triplet lines, the correction amounts to about
795: 0.01\,dex. By constructing a realistic time-dependent, 3D, hydrodynamic model
796: of the solar atmosphere, Asplund et al. (2004) determine the solar
797: photospheric oxygen abundance from a variety of diagnostic lines, including the
798: [\ion{O}{1}] forbidden lines, the \ion{O}{1} triplet permitted lines as well as
799: OH vibration-rotation lines and the OH pure rotation lines. In the case of the
800: \ion{O}{1} triplet permitted lines, 3D NLTE calculations have been performed,
801: revealing significant departures from the LTE as a result of photon losses in
802: the lines. The NLTE effects due to this process yield corrections that amount
803: to 0.2--0.3\,dex. The differences between the 1D and 3D models are however
804: found to be small, less than 0.1\,dex. From these detailed calculations, it
805: seems that granulation corrections are probably unlikely to be able to account
806: for the large variations in the oxygen triplet line abundance discuss in the
807: current paper, at least in solar type stars. Calculations of granulation
808: corrections for other types of star have not been reported and are preferred,
809: especially for late type stars.
810:
811: \section{Summary}
812:
813: We have found a dramatic increase in the oxygen triplet abundance upper
814: boundary with decreasing effective temperature in the cool dwarfs of young open
815: cluster IC\,4665, similar to what previously observed in Pleiades and M\,34 by
816: Schuler et al. (2004), in Hyades by Schuler et al. (2006a) and in UMa by King
817: \& Schuler (2005). By contrast, oxygen abundances derived from the
818: [\ion{O}{1}] $\lambda$6300 forbidden line are found to be constant in Pleiades
819: and Hyades. It seems that the [\ion{O}{1}] $\lambda$6300 forbidden line is
820: relatively free from the various processes that may have affected abundances
821: determined from the \ion{O}{1} triplet lines and is therefore a better
822: abundance indicator.
823:
824: At the present moment, the uncertainties in the measured values of
825: [O/H] are too large to place any meaningful planet-formation
826: constraint based on the oxygen abundance of cluster member stars.
827: Under the assumption that oxygen abundance is homogeneous in a given
828: cluster, an assumption that is supported by abundance determinations
829: using the [\ion{O}{1}] forbidden line, we have investigated various
830: possible mechanisms that may be responsible for the observed trend and
831: scatter of oxygen triplet line abundances. Although the \ion{O}{1}
832: triplet lines are sensitive to stellar parameters, we show that the
833: possible uncertainties in our parameter determinations are unlikely to
834: be sufficient to explain the more than 1\,dex variations in oxygen
835: triplet line abundances. The possible effects of canonical NLTE as
836: the dominant cause of the problem is ruled out as they predict a trend
837: of oxygen abundance as $T_{\rm eff}$ varies that is exactly opposite
838: to what observed. Available calculations of granulation corrections
839: show that the effects are generally small, at least for solar type
840: stars, and therefore may contribute little to the observed spreads of
841: oxygen triplet line abundances. The variation of oxygen triplet
842: abundances as a function of effective temperatures are found to be
843: larger in younger open clusters and smaller in older clusters.
844: Age-related effects stellar surface activities are then suggested to
845: blame for the large spreads of oxygen triplet abundances. This
846: assumption are supported by the correlation analysis between the
847: triplet abundance and the simultaneous observation of H$\alpha$ and
848: \ion{Ca}{2} IRT emissions which, are indicators of stellar
849: chromospheric activity levels. The lack of a monotonous correlation
850: between $T_{\rm eff}$ and the triplet abundance implies that the the
851: triplet abundance/activity correlation is not a consequence of the
852: abundance/$T_{\rm eff}$ correlations. We could not exclude the
853: possibility that stellar spots play a role in the abnormal behavior of
854: the detected oxygen triplet abundances.
855:
856: \begin{acknowledgements} The authors wish to thank Dr. Debra Fischer for her
857: expert assistance in the usage of SME. We thank Dr. Frank Grupp for his valuable suggestions on this paper. ZXS and XWL acknowledge Chinese NSFC
858: Grant 10373015.
859:
860: \end{acknowledgements}
861:
862: \begin{thebibliography}{}
863: \bibitem[2005]{aff} Affer, L., Micela, G., Morel, T., Sanz-Forcada, J.
864: \& Favata, F. 2005, A\&A, 433, 647
865: \bibitem[1996]{all} Allain, S., Bouvier, J., Prosser, C., Marschall, L. A., \&
866: Laaksonen, B. D. 1996, A\&A, 305, 498
867: \bibitem[2001]{allende} Allende Prieto, C., Lambert, D. L., \& Asplund, M. 2001, ApJ, 556, L63
868: \bibitem[2001]{asp1} Asplund, M. 2001, NewA Rev., 45, 565
869: \bibitem[2001]{asp2} Asplund, M., \& Garc\'{i}a P\'{e}rez, A. E. 2001, A\&A, 372, 601
870: \bibitem[2004]{asp} Asplund, M., Grevesse, N., Sauval, A. J., Allende Prieto, C., \& Kiselman, D. 2004, A\&A, 417, 751
871: \bibitem[1996]{bal} Balachandran, S. C., \& Carney, B. W. 1996, AJ, 111, 946
872: \bibitem[2001]{bar} Barrado y Navascu\'{e}s, D., Garc\'{i}a L\'{o}pez, R. J., Severino, G., \& Gomez, M. T. 2001, A\&A, 371, 652
873: \bibitem[1993]{bou} Bouvier, J., Cabrit, S., Fernandez, M., Martin, E. L., \&
874: Matthews, J. M. 1993, A\&A, 272, 167
875: \bibitem[2000]{car} Carretta, E., Gratton, R. G., \& Sneden, C. 2000, A\&A, 356, 238
876: \bibitem[1997]{cav} Cavallo, R. M., Pilachowski, C. A., \& Rebolo, R. 1997, PASP, 109, 226
877: \bibitem[2006]{ciesla} Ciesla, F. J. \& Cuzzi, J. N. 2006, Icarus, 181, 178
878: \bibitem[2002]{ford} Ford, A., Jeffries, R. D., \& Smalley, B. 2002, A\&A, 391, 253
879: \bibitem[1998]{gia} Giampapa, M. S., Prosser, C. F., \& Fleming, T. A. 1998, ApJ, 501, 624
880: \bibitem[1999]{gratton} Gratton, R. G., Carretta, E., Eriksson, K., \& Gustafsson, B. 1999, A\&A, 350, 955
881: \bibitem[1992]{gray} Gray, D. F., 1992, The observation and analysis of stellar photospheres, P. 356, equation (16.2)
882: \bibitem[1996]{gre} Grevesse, N., Noels, A., \& Sauval, A. J. 1996, Standard Abundances, in {\it Cosmic Abundances}, eds S. S. Holt and G. Sonneborn,
883: ASP Conference Series {\bf 99}, 117-126
884: \bibitem[2004]{ida} Ida, S. \& Lin, D. N. C. 2004, ApJ, 604, 388
885: \bibitem[1998]{isra} Israelian, G., Garc\'{i}a L\'{o}pez, R. J., \& Rebolo, R. 1998, ApJ, 507, 805
886: \bibitem[2003]{joh} Johansson, S., Litz\'{e}n, U., Lundberg, H., \& Zhang, Z. 2003, ApJ, 584, L107
887: \bibitem[2005]{kharchenko} Kharchenko, N. V., Piskunov, A. E., R$\ddot{o}$ser, S.,
888: Schilbach, E., Scholz, R.-D., 2005, A\&A, 438, 1163
889: \bibitem[1993]{king} King, J. R. 1993, AJ, 106, 1206
890: \bibitem[1995]{k2} King, J. R. \& Boesgaard, A. M. 1995, AJ, 109, 383
891: \bibitem[2005]{k3} King, J. R. \& Schuler, S. C. 2005, PASP, 117, 911
892: \bibitem[1991]{kis} Kiselman, D. 1991, A\&A, 245, L9
893: \bibitem[1993]{kis2} Kiselman, D. 1993, A\&A, 275, 269
894: \bibitem[1995]{kn} Kiselman, D., \& Nordlund, \AA. 1995, A\&A, 302, 578
895: \bibitem[1999]{kup} Kupka, F., Piskunov, N. E., Ryabchikova, T. A., Stempels, H. C., \& Weiss, W. W. 1999, A\&AS, 138, 119 (VALD-2)
896: \bibitem[1984]{kur} Kurucz, R. L., Furenlid, I., Brault, J., \& Testerman, L. 1984, National Solar Observatory Atlas No. 1: Solar Flux Atlas from 296 to 1300 nm (Sunspot: Natl. Solar Obs.)
897: \bibitem[1993]{kuru} Kurucz R., 1993, Atlas Stellar Atmosphere Program and 2 km/s Grid CDROM vol 13, Smithsonian Astrophysical Observatory, Cambridge
898: \bibitem[1978]{lam} Lambert, D. L. 1978, MNRAS, 182, 249
899: \bibitem[1997]{mar} Mart\'{i}n, E. L., \& Montes, D. 1997, A\&A, 318, 805
900: \bibitem[1981]{mer} Mermilliod, J. C. 1981, A\&AS, 44, 467
901: \bibitem[2001]{mess} Messina, S., Rodon$\grave{\rm o}$, M. \& Guinan, E. F. 2001, A\&A, 366, 215
902: \bibitem[2001]{mess2} Messina, S., Pizzolato, N., Guinan, E. F., \& Rodon$\grave{\rm o}$, M. 2003, A\&A, 410, 671
903: \bibitem[2000]{mis} Mishenina, T. V., Korotin, S. A., Klochkova, V. G., \& Panchuk, V. E. 2000, A\&A, 353, 978
904: \bibitem[1998]{mon} Montes, D. \& Mart\'{i}n, E. L. 1998, A\&AS, 128, 485
905: \bibitem[1966]{moo} Moore, E., Minnaert, M. G. J., Houtgast, J. 1966, The Solar Spectrum 2935\,\AA\, to 8770\,\AA, National Bureau of Standards Monograph 61
906: \bibitem[2004]{mor} Morel, T., \& Micela, G. 2004, A\&A, 423, 677
907: \bibitem[2001]{nissen01} Nissen, P. E., Primas, F., \& Asplund, M. 2001, NewA Rev., 45, 545
908: \bibitem[2002]{nissen} Nissen, P. E., Primas, F., Asplund, M., \& Lambert, D. L. 2002, A\&A, 390, 235
909: \bibitem[2001]{no} Norris, J. E., Ryan, S. G., \& Beers, T. C. 2001, ApJ, 561, 1034
910: \bibitem[2002]{pa} Paulson, D. B., Saar, S. H., Cochran, W. D. \& Hatzes, A. P.
911: 2002, AJ, 124, 572
912: \bibitem[2003] {pa2} Paulson, D. B., Sneden, C., \& Cochran, W. D. 2003, AJ,
913: 125, 318
914: \bibitem[1995]{pis} Piskunov, N. E., Kupka, F., Ryabchikova, T. A., Weiss, W. W., \& Jeffery, C. S. 1995, A\&AS, 112, 525 (VALD-1)
915: \bibitem[1993]{pro} Prosser, C. F. 1993, AJ, 105, 1441
916: \bibitem[1999]{rya} Ryabchikova T. A. Piskunov N. E., Stempels H. C., Kupka F., \& Weiss W. W. 1999, Physica Scripta T83, 162-173 (VALD-2)
917: \bibitem[2003]{schu} Schuler, S. C., King, J. R., Fischer, D. A., Soderblom, D. R., \& Jones, B. F. 2003, AJ, 125, 2085
918: \bibitem[2004]{schu2} Schuler, S. C., King, J. R., Hobbs, L.M., \& Pinsonneault, M. H. 2004, ApJ, 602, L117
919: \bibitem[2006]{sc} Schuler, S. C., King, J. R., Terndrup, D. M., et al. 2006a,
920: ApJ, 636, 432
921: \bibitem[2006]{sc2} Schuler, S. C., Hatzes, A. P., King, J. R., et al. 2006b,
922: AJ, 131, 1057
923: \bibitem[2004]{shen} Shen Z.-X., Jones, B. F., Lin, D. N. C., Liu, X.-W., Li, S.-L. 2005, ApJ, 635, 608 (Paper~I)
924: \bibitem[1993]{sod} Soderblom, D. R., Stauffer, J. R., Hudon, J. D., \& Jones, B. F. 1993, ApJS, 85, 315
925: \bibitem[2001]{sod2} Soderblom, D. R., Jones, B. F. \& Fischer, D. 2001, ApJ,
926: 563, 334
927: \bibitem[2003]{stau} Stauffer, J. R., Jones, B. F., Backman, D., Hartmann, L. W., Barrado y Navascu\'{e}s, D., Pinsonneault, M. H., Terndrup, D. M., \& Muench, A. A. 2003, AJ, 126, 833
928: \bibitem[2002]{ste} Steffen, M., \& Holweger, H. 2002, A\&A, 387,258
929: \bibitem[1995]{tak} Takeda, Y. 1995, PASJ, 47, 463
930: \bibitem[2003]{tak2}Takeda, Y. 2003, A\&A, 402, 343
931: \bibitem[1992]{tom} Tomkin, J., Lemke, M., Lambert, D. L., \& Sneden, C. 1992, AJ, 104, 1568
932: \bibitem[1996]{val} Valenti, J. A., \& Piskunov, N. 1996, A\&AS, 118, 595
933: \bibitem[2005]{vf} Valenti, J. A., \& Fischer, D. A. 2005, ApJS, 159, 141
934: \bibitem[1992]{voi} Vogt, S. S. 1992, in ESO Workshop on High Resolution Spectroscopy with the VLT, ed. M. H. Ulrich (Garching: ESO), P. 223
935: \bibitem[2002]{wil} Wilden, B. S., Jones, B. F., Lin, D. N. C., \& Soderblom, D. R. 2002, AJ, 124, 2799
936: \bibitem[2005]{xd} Xiong, D.-R., \& Deng, L. 2005, ApJ, 622, 620
937: \end{thebibliography}
938:
939: \begin{table}
940: \tabcolsep 3pt
941: \centering
942: \caption{Stellar parameters and oxygen triplet line abundances of IC\,4665 dwarfs}
943: \label{triplet}
944: \begin{tabular}{lccccccrcr}
945: \hline
946: \hline
947: \noalign{\smallskip}
948: Star & $T_{\rm eff}$(SME)& $\log\,g$ & $v_{\rm mic}$ & $F({\rm H\alpha})^{a}$ &
949: log & A(V)$^{c}$& [O/H]$_{\rm SME}$ & Error$_{\rm SME}$ & [O/H]$_{\rm EW}$\\
950: & (K) &(cm\,s$^{-2}$) & (km\,s$^{-1}$)& & ($L_{\rm X}/L_{\rm bol}$)$^{b}$ &(mag)& &\\
951: \noalign{\smallskip}
952: \hline
953: \noalign{\smallskip}
954: P\,19 & 6370 & 4.44 & 0.26 & * & * & & $-$0.26 & 0.10 & $-$0.31\\
955: P\,147 & 6189 & 4.49 & 1.02 & * & * & & 0.08 & * & 0.00\\
956: P\,39 & 5867 & 4.50 & 1.25 & * & * &0.04& $-$0.03 & * & $-$0.09\\
957: P\,107 & 5626 & 4.56 & 0.68 & 5.5 & * & & 0.51 & * & 0.45\\
958: P\,150 & 5535 & 4.57 & 1.62 & 6.6 & * &0.06& 0.35 & * & 0.15\\
959: P\,151 & 5494 & 4.58 & 1.46 & * & * & & 0.17 & 0.24 & 0.07\\
960: P\,60 & 5483 & 4.58 & 1.27 & * & $-$3.44 & & 0.11 & * & 0.01\\
961: P\,75 & 5347 & 4.47 & 1.60 & 6.5 & * &0.05& 0.46 & * & 0.19\\
962: P\,165 & 5292 & 4.59 & 1.56 & 6.6 & * & & 0.23 & * & 0.19\\
963: P\,267 & 5286 & 4.65 & 0.43 & * & * & & $-$0.03 & * & $-$0.24\\
964: P\,64 & 5267 & 4.62 & 0.95 & * & * & & 0.15 & * & 0.03\\
965: P\,71 & 5251 & 4.60 & 1.55 & 6.5 & $-$3.11 &0.04& 0.64 & 0.26 & 0.33\\
966: P\,199 & 5168 & 4.64 & 0.57 & * & * & & 0.22 & * & ---$^{d}$\\
967: P\,94 & 5168 & 4.64 & 0.87 & 6.0 & * & & $-$0.20 & * & $-$0.33\\
968: P\,100 & 4913 & 4.65 & 1.46 & 6.7 & $-$3.00 &0.10& 1.00 & * & 0.60\\
969: \noalign{\smallskip}
970: \hline
971: \end{tabular}
972: \begin{description}
973: \item $^{a}$ From Mart\'{i}n \& Montes (1997); in units of erg\,cm$^{-2}$\,s$^{-1}$\,\AA$^{-1}$.
974: \item $^{b}$ From Giampapa, Prosser \& Fleming (1998);
975: \item $^{c}$ From Allain et al. (1996)
976: \item $^{d}$ Very low S/N ($<$ 30). No reliable EW measurements could be given for
977: the shallow oxygen triplet lines.
978: \end{description}
979: \end{table}
980:
981: \clearpage
982:
983: \begin{table}
984: \begin{center}
985: \caption{Oxygen forbidden line abundances of Pleiades dwarfs}
986: \label{forbid}
987: \begin{tabular}{lcrcrc}
988: \hline
989: \hline
990: \noalign{\smallskip}
991: Star & $T_{\rm eff}$ (K) & [O/H]$_{\rm SME}$ & Error$_{\rm SME}$& [O/H]$_{\rm EW}$ & Error$_{\rm EW}$\\
992: \noalign{\smallskip}
993: \hline
994: \noalign{\smallskip}
995: H\,II\,250 & 5731 & 0.02 & 0.08 & $-$0.03 & 0.14\\
996: H\,II\,129 & 5369 & $-$0.05 & 0.07 & $-$0.02 & 0.14\\
997: H\,II\,2462 & 5256 & $-$0.10 & 0.11 & $-$0.14 & 0.17\\
998: H\,II\,522 & 4966 & $-$0.14 & 0.11 & $-$0.17 & 0.18\\
999: H\,II\,1039 & 4908 & 0.14 & 0.09 & 0.07 & 0.14\\
1000: H\,II\,1298 & 4854 & 0.05 & 0.09 & $-$0.03 & 0.15\\
1001: \noalign{\smallskip}
1002: \hline
1003: \end{tabular}
1004: \end{center}
1005: \end{table}
1006:
1007: \begin{table}
1008: \begin{center}
1009: \caption{The surface activity indicators of H$\alpha$ and two \ion{Ca}{2} IRT lines for IC\,4665 stars}
1010: \label{R}
1011: \begin{tabular}{lrrrrrrr}
1012: \hline
1013: \hline
1014: \noalign{\smallskip}
1015: Star & R(H$\alpha$) & R(8498) & R(8662) & Err(H$\alpha$) & Err(8498) &Err(8662) & Err(O)\\
1016: \noalign{\smallskip}
1017: \hline
1018: \noalign{\smallskip}
1019: P19 & 0.210 & 0.318 & 0.225 & 0.062 & 0.054 & 0.048 & 0.098\\
1020: P147 & 0.000 & 0.038 & 0.063 & 0.061 & 0.048 & 0.045 & 0.123\\
1021: P39 & 0.274 & 0.350 & 0.303 & 0.045 & 0.040 & 0.039 & 0.168\\
1022: P107 & 0.189 & 0.247 & 0.288 & 0.052 & 0.052 & 0.047 & 0.201\\
1023: P150 & 0.676 & 0.659 & 0.531 & 0.058 & 0.053 & 0.058 & 0.214\\
1024: P151 & 0.341 & 0.535 & 0.482 & 0.062 & 0.056 & 0.051 & 0.219\\
1025: P60 & 0.417 & 0.498 & 0.486 & 0.094 & 0.082 & 0.076 & 0.221\\
1026: P75 & 0.601 & 0.647 & 0.610 & 0.104 & 0.092 & 0.092 & 0.240\\
1027: P165 & 0.851 & 0.637 & 0.511 & 0.090 & 0.079 & 0.090 & 0.247\\
1028: P267 &$-$0.016 &$-$0.103 &$-$0.130 & 0.118 & 0.108 & 0.108 & 0.248\\
1029: P64 &$-$0.086 &$-$0.173 &$-$0.132 & 0.097 & 0.076 & 0.084 & 0.251\\
1030: P71 & 0.731 & 0.666 & 0.613 & 0.160 & 0.143 & 0.127 & 0.253\\
1031: P199 &$-$0.071 &$-$0.109 &$-$0.134 & 0.132 & 0.149 & 0.128 & 0.264\\
1032: P94 & 0.204 & 0.377 & 0.275 & 0.090 & 0.067 & 0.070 & 0.264\\
1033: P100 & 1.186 & 0.951 & 0.733 & 0.173 & 0.127 & 0.132 & 0.300\\
1034:
1035: \noalign{\smallskip}
1036: \hline
1037:
1038: \end{tabular}
1039: \end{center}
1040: \end{table}
1041:
1042:
1043: \begin{table}
1044: \begin{center}
1045: \caption{The linear correlation coefficients of oxygen triplet line abundances
1046: with different surface activity indicators}
1047: \label{active}
1048: \begin{tabular}{clllllc}
1049: \hline
1050: \hline
1051: \noalign{\smallskip}
1052: Cluster & H$\alpha$ & X-ray & \ion{Ca}{2} H+K & \ion{Ca}{2} IRT & A(V)&Ref$^a$\\
1053: \noalign{\smallskip}
1054: \hline
1055: \noalign{\smallskip}
1056: IC\,4665 & 0.34 (7)$^b$ &0.98 (3) & * & * &0.46 (5)& 1, 2, 11\\
1057: & 0.69 (15) & & & 0.55/0.57$^c$(15)&* &3\\
1058: Pleiades & 0.04 (15) &0.72 (11) & * & 0.64 (15) &* &4, 5, 6\\
1059: M\,34 & 0.15 (8) & * & * & 0.71 (7) &* &4, 7\\
1060: UMa & * & * &0.77 (6) & * &* &8\\
1061: Hyades & * &0.16 (28) &0.21 (40)& * &* &6, 9, 10\\
1062: \noalign{\smallskip}
1063: \hline
1064: \end{tabular}
1065: \begin{description}
1066: \item $^{a}$ (1) Mart\'{i}n \& Montes (1997); (2) Giampapa, Prosser \& Fleming
1067: (1998); (3) This work; (4) Schuler et al. (2004); (5) Soderblom et al. (1993); (6) Morel \&
1068: Micela (2004); (7) Soderblom et al. (2001); (8) King \& Schuler (2005), and
1069: references therein; (9) Schuler et al. (2006a); (10) Paulson et al. (2002);
1070: (11) Allain et al. (1996)
1071: \item $^{b}$ Values in parentheses are numbers of stars analyzed.
1072: \item $^{c}$ For $\lambda$8498 and $\lambda$8662 respectively.
1073: \end{description}
1074: \end{center}
1075: \end{table}
1076:
1077:
1078:
1079: \begin{figure}
1080: \plotone{fig1.eps}
1081: \caption{Sample spectra of four IC\,4665 stars centered on the \ion{O}{1}
1082: triplet lines. The spectra were selected to illustrate the full range of S/N
1083: ratios achieved amongst the sample stars. Also plotted are the best fit
1084: synthetic spectra (smooth curves) obtained for individual stars.}
1085: \label{otrip}
1086: \end{figure}
1087:
1088: \begin{figure}
1089: \plotone{fig2.eps}
1090: \caption{Spectra of six Pleiades stars centered on the \ion{O}{1} $\lambda$6300
1091: forbidden line. In each panel, also overplotted are three synthetic spectra
1092: (smooth curves) showing, from top to bottom, the contributions from the
1093: \ion{Ni}{1} $\lambda$6300.34 line, from the [\ion{O}{1}] $\lambda$6300.30
1094: forbidden line and their sum, respectively.}
1095: \label{pspectrum}
1096: \end{figure}
1097:
1098: \begin{figure}
1099: \plotone{fig3.eps}
1100: \caption{Oxygen triplet line abundances of IC\,4665 dwarfs (filled triangles)
1101: plotted against effective temperature. Typical errors are shown for three
1102: selected stars of representative effective temperatures. Also plotted are
1103: oxygen triplet line abundances of Pleiades (open triangles) and M\,34
1104: dwarfs (crosses) taken from Schuler et al. (2004), Hyades dwarfs (open
1105: squares) from Schuler et al. (2006a) and UMa dwarfs (filled squares) from King
1106: \& Schuler (2005)}
1107: \label{ic4665}
1108: \end{figure}
1109:
1110:
1111: \begin{figure}
1112: \plotone{fig4.eps}
1113:
1114: \caption{[\ion{O}{1}] forbidden line abundances of Pleiades dwarfs plotted
1115: against effective temperature. Filled triangles are our results and open
1116: triangles are values taken from Schuler et al. (2004). Also plotted are
1117: [\ion{O}{1}] forbidden line abundances of Hyades (filled squares) taken from
1118: Schuler et al. (2006b).}
1119:
1120: \label{pleiades}
1121: \end{figure}
1122:
1123: %\begin{figure}
1124: %\plotone{fig5.eps}
1125: %\caption{Oxygen triplet line abundance as a function of a). normalized X-ray
1126: %luminosity; b). H$\alpha$ emission flux for IC\,4665 sample stars.}
1127: %\label{activ}
1128: %\end{figure}
1129: %\clearpage
1130:
1131: \begin{figure}
1132: %\resizebox{17cm}{13cm}
1133: \plotone{fig5.eps}
1134: \caption{Oxygen triplet line abundance as a function of H$\alpha$ and \ion{Ca}{2} IRT line $\lambda\lambda$8498,8662 emission factor R for IC\,4665 sample stars.}
1135: \label{Hairt}
1136: \end{figure}
1137: \clearpage
1138:
1139:
1140: \end{document}
1141:
1142: