astro-ph0409383/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: %\documentstyle[12pt,aasms4]{aastex} % para submeter PcTEx
3: %\documentstyle[11pt, aaspp4]{aastex} % para 2 colunas UNIX
4: %\documentclass[aas2pp4]{aastex} % for 2 columns UNIX ddddddddddddddd
5: %\documentstyle[11pt, aaspp4]{aastex} % para preprint UNIX
6: %\documentstyle[11pt, aaspp4]{aastex} % para preprint PCtEx
7: %\documentstyle[apjpt4]{article} % para tables
8: 
9: %\documentclass{emulateapj}
10: 
11: 
12: \begin{document}
13: \title {Reappraising the Spite Lithium Plateau: Extremely Thin
14: and Marginally Consistent with WMAP
15: %\footnote{}
16: } 
17: 
18: \newcommand{\teff}{T$_{\rm eff}$ }
19: \newcommand{\tirfm}{T$_{\rm eff}^{\rm IRFM}$}
20: \newcommand{\tfe}{T$_{\rm eff}^{\rm FeI}$}
21: 
22: % Next 5 lines define \simless and \simgreat: "less than or approximately
23: % equal to" and "greater than or approximately equal to".
24: \newbox\grsign \setbox\grsign=\hbox{$>$} \newdimen\grdimen \grdimen=\ht\grsign
25: \newbox\simlessbox \newbox\simgreatbox
26: \setbox\simgreatbox=\hbox{\raise.5ex\hbox{$>$}\llap
27:      {\lower.5ex\hbox{$\sim$}}}\ht1=\grdimen\dp1=0pt
28: \setbox\simlessbox=\hbox{\raise.5ex\hbox{$<$}\llap
29:      {\lower.5ex\hbox{$\sim$}}}\ht2=\grdimen\dp2=0pt\def\simgreat{\mathrel{\copy\simgreatbox}}
30: \def\simless{\mathrel{\copy\simlessbox}}
31: % Next lines define "approximately proportional to"
32: \newbox\simppropto
33: \setbox\simppropto=\hbox{\raise.5ex\hbox{$\sim$}\llap
34:      {\lower.5ex\hbox{$\propto$}}}\ht2=\grdimen\dp2=0pt
35: \def\simpropto{\mathrel{\copy\simppropto}}
36: 
37: \author{Jorge Mel\'endez\altaffilmark{1} and Iv\'an Ram\'{\i}rez\altaffilmark{2}}
38: \altaffiltext{1}{Department of Astronomy, Caltech, M/C 105-24, 
39: 1200 E. California Blvd, Pasadena, CA 91125; jorge@astro.caltech.edu}
40: \altaffiltext{2}{Department of Astronomy, University of Texas at Austin, RLM 15.306, TX
41: 78712-1083; ivan@astro.as.utexas.edu}
42: 
43: \slugcomment{Submitted to the The Astrophysical Journal}
44: %\slugcomment{Send proofs to:  J. Mel\'endez}
45: 
46: \begin{abstract}
47: The lithium abundance in 62 halo dwarfs is determined 
48: from accurate equivalent widths reported in the literature
49: and an improved infrared flux method (IRFM) temperature scale. 
50: The Li abundance of 41 plateau stars (those with \teff $>$ 6000 K) 
51: is found to be independent of temperature and metallicity, 
52: with a star-to-star scatter of only 0.06 dex
53: over a broad range of temperatures (6000 K $<$ \teff $<$ 6800 K) and
54: metallicities ($-3.4 <$ [Fe/H] $< -1$), thus imposing
55: stringent constraints on depletion by mixing and
56: production by Galactic chemical evolution.
57: We find a mean Li plateau abundance of $A_{Li}$ = 2.37 dex
58: ($^7$Li/H = 2.34 $\times 10^{-10}$), 
59: which, considering errors of the order of 0.1 dex in the {\it absolute} 
60: abundance scale, is just in
61: borderline agreement with the constraints imposed by the 
62: theory of primordial 
63: nucleosynthesis and WMAP data (2.51 $< A_{Li}^{\rm WMAP}$ $<$ 2.66 dex).
64: \end{abstract}
65: 
66: \keywords{cosmology: observations - stars: abundances - stars: Population II}
67: 
68: \section{Introduction}
69: The Li plateau was discovered by Spite \& Spite (1982),
70: who showed that the $^7$Li abundance obtained from the
71: Li doublet at 6708 \AA\ in F and early G halo dwarfs is
72: independent of temperature and metallicity, suggesting
73: that the Li abundance determined in halo
74: stars represents the primordial abundance from
75: Big Bang nucleosynthesis (BBN). The standard
76: theory of BBN predicts the abundance of light elements,
77: in particular $^7$Li, as a function of the baryon-to-photon
78: ratio $\eta$. Thus, if the plateau is primordial, an inverse analysis 
79: could be used to determine $\eta$ (Spite \& Spite 1982).
80: 
81: On the other hand, Ryan and collaborators (Ryan et al. 1996, 1999, 2001; 
82: hereafter jointly referred as R01)
83: suggested that the Li abundance in halo stars is postprimordial, 
84: with a 0.3 dex increase from [Fe/H] = $-$3.5 to [Fe/H] = $-$1.\footnote{[A/B] 
85: $\equiv$ log ($N_A/N_B$) - log($N_A/N_B$)$_\odot$}
86: Those results rely on the 
87: temperature scale of Magain (1987), which was calibrated with only
88: 3 metal-poor ([Fe/H] $<$ -2) stars.
89: Considering the metallicity dependence of the Li
90: abundances, R01 extrapolated a primordial 
91: Li abundance of $A_{Li}$
92: $\approx$ 2.0 - 2.1 dex.\footnote{$A_{X}$ $\equiv$ log ($N_X$/$N_H$) + 12}
93: 
94: Recently, precise measurements of the cosmic background 
95: anisotropies by the Wilkinson Microwave Anisotropy Probe 
96: (WMAP, Spergel et al. 2003) have been used to constrain $\eta$, 
97: and employing the standard theory of BBN, the primordial Li abundance 
98: was predicted to be 
99: $A_{Li}^{\rm WMAP}$ = 2.58$_{-0.07}^{+0.08}$ dex (Cyburt et al. 2003),
100: a value that is much higher than the primordial abundance proposed by
101: R01. The large difference of about
102: 0.5 dex between the results of Ryan et al. and WMAP has
103: stimulated theoretical work on non-standard BBN
104: (e.g. Feng et al. 2003; Ichikawa \& Kawasaki 2004).
105: 
106: Bonifacio \& Molaro (1997, hereafter BM97) found that when the \teff scale based
107: on the infrared flux method (IRFM) is employed (that of Alonso et al. 1996), 
108: the Li abundance in metal-poor stars does not depend on metallicity, and 
109: $A_{Li}$ = 2.24 dex is obtained, higher than the primordial
110: abundance proposed by R01, but still
111: lower by a factor of 2 than the abundance suggested by the
112: WMAP data. Furthermore, Bonifacio et al. (2002) analyzed 5 metal-poor field stars
113: and 12 turn-off stars in the globular cluster NGC 6397,
114: obtaining a Li abundance 0.05 and 0.1 dex higher than in BM97,
115: respectively.
116: 
117: Ford et al. (2002) have determined the Li abundance in 
118: a few metal-poor stars employing the extremely weak 
119: ($\approx$ 2-4 m\AA) Li lines at 6104 \AA, 
120: resulting in $A_{Li}$ $\approx$ 2.5 dex, but note that the observational
121: errors are very large (about 30\%), not allowing them to draw
122: any firm conclusions. In fact, when the Li lines at 6708 \AA\ were used by
123: Ford et al. (2002), $A_{Li}$ $\approx$ 2.2 dex was found.
124: 
125: The temperature is the most important atmospheric parameter 
126: required to obtain the Li abundance,
127: and since metal-poor stars are too distant to measure their
128: angular diameters, indirect methods have to be used to
129: estimate their temperatures. The closest (and less model-dependent)
130: approach to the direct method is the IRFM,
131: where the temperature is found by a comparison between
132: the observed and theoretical ratio of
133: the bolometric flux to the infrared flux. In a pioneer work,
134: Magain (1987) calibrated the \teff scale of metal-poor
135: dwarfs (mostly \teff $<$ 6000 K and [Fe/H] $> -2$) employing 
136: IRFM temperatures of 11 dwarfs. The titanic work by 
137: Alonso et al. (1996) improved the \teff scale
138: including a larger number of metal-poor stars 
139: (down to [Fe/H] $\approx$ -3), employing spectroscopic
140: metallicities (when available) determined before 1992.
141: In the meantime, new spectroscopic and photometric
142: surveys have appeared in the literature 
143: (a milestone was the completion of 2MASS), 
144: providing the necessary input data for a new revision of 
145: the \teff scale (Ram\'{\i}rez \& Mel\'endez 2004b,c, hereafter RM04b,c). 
146: As an example, only from January 2001 to June 2003, 
147: more than 1500 new spectroscopic metallicity determinations
148: for field FGK stars have appeared in the literature, 
149: increasing significantly the number of calibrating stars with 
150: reliable [Fe/H] values, and hence allowing to better
151: define the metallicity dependence of the \teff scale.
152: We have determined the IRFM \teff of more 
153: than 10$^3$ stars (RM04b) using updated input data, 
154: and metallicity-dependent \teff vs. color calibrations 
155: were derived in the metallicity range $-3.5 <$ [Fe/H] $< +0.4$ (RM04c).
156: 
157: Employing our improved IRFM temperature scale, we show
158: that the Spite plateau has essentially zero scatter and
159: that the derived abundances are now consistent, although only
160: marginally, with WMAP data.
161: 
162: \section{Data and Atmospheric Parameters}
163: The sample was selected from relatively unevolved halo stars
164: (log $g \geq$ 3.5, [Fe/H] $<$ $-$1) with
165: previous Li observations and having either
166: an IRFM effective temperature determined by us (RM04b)
167: and/or with a previous determination of metallicity
168: as reported in the ``2003 updated'' version of the
169: Cayrel de Strobel et al. (2001, hereafter C03) 
170: catalog (see $\S$2 in Ram\'{\i}rez \& Mel\'endez 2004a). 
171: 
172: For most works previous to 1997, the equivalent
173: widths ($W_\lambda$) of the Li doublet at 6708 \AA\ 
174: were taken from the compilation of Ryan et al. (1996), 
175: adopting a minimum error of 1.5 m\AA\ for the individual
176: measurements. 
177: Other $W_\lambda$ considered are 
178: those by Molaro et al. (1995), Spite et al. (1996), 
179: Nissen \& Schuster (1997), Ryan \& Deliyannis (1998), Smith et al. (1998), 
180: Guti\'errez et al. (1999), Hobbs et al. (1999),
181:  King (1999),
182: Ryan et al. (1999, 2001), 
183: Bonifacio et al. (2002), and Ford et al. (2002).
184: The weighted averages of the equivalent widths and
185: errors ($W_\lambda$ and $\delta W_\lambda$) were computed,
186: as well as the standard errors ($\sigma_m$). The adopted error is
187: $\sigma_W \equiv$  max($\delta W_\lambda$, $\sigma_m$, 1 m\AA). 
188: A star was rejected if $\sigma_W \geq 0.1 W_\lambda$.
189: Stars with E(B-V) $>$ 0.05 dex were also discarded. 
190: Sixty two dwarfs satisfied our selection criteria.
191: 
192: The key parameter for a reliable Li abundance determination
193: in metal-poor stars is T$_{\rm eff}$; errors in the other atmospheric parameters 
194: (log $g$, [Fe/H], $v_t$) result in abundance errors of the order 
195: of 0.01 dex (e.g. Ryan et al. 1996).
196: Surface gravities and metallicities were taken
197: from the C03 catalog, and a microturbulence $v_t$ = 1 km s$^{-1}$ 
198: was adopted.
199: For 3 stars not included in the C03 catalog, we
200: adopted metallicities from a photometric calibration
201: (RM04b) that is based on C03 [Fe/H] values. 
202: %The metallicities of these 3 stars are consistent with previous works.
203: 
204: The effective temperatures are based on our IRFM \teff
205: scale. Individual IRFM temperatures were
206: taken from RM04b or recomputed employing new reddening corrections, 
207: as well as calculated employing the 
208: IRFM \teff calibrations of RM04c. We used up to 17 colors
209: in the {\it BVRI}(Johnson-Cousins), Tycho-Hipparcos, Str\"omgren, 
210: DDO, Vilnius, Geneva and 2MASS systems.
211: Most of the optical photometry was taken from the General Catalogue of 
212: Photometric Data (Mermilliod et al. 1997) and
213: the Tycho catalog, while the infrared photometry was taken from 2MASS.
214: E(B-V) values were computed employing several extinction maps, 
215: giving a higher weight to the maps by Schlegel et al. (1998) 
216: for high latitude stars (RM04b).
217: 
218: The use of several colors improves the determination 
219: of the effective temperature, since it alleviates the impact of 
220: photometric errors. Following RM04c, the adopted \teff 
221: was the mean of the IRFM determination and the 
222: temperature from the IRFM calibrations (T$_{\rm eff}^{\rm cal}$):
223: 
224: \begin{equation}
225: {\rm T_{\rm eff}} = \case{1}{n+m} (n {\rm T_{eff}^{IRFM}} + m {\rm T_{eff}^{cal}})
226: \end{equation}
227: 
228: \noindent where the weights we adopted were $n$ = 1  
229: and $m$ equal to the square root of the number of colors employed
230: (typically 10, thus $m \approx 3$).
231: T$_{\rm eff}^{\rm cal}$ was calculated taking
232: into account the errors of each IRFM color calibration.
233: The mean internal error of the \teff derived in
234: this work is $\delta$\teff $\approx$ 66 K.
235: It is important to note that the use of eq. (1)
236: results in a \teff that is only $\approx$ 22 K lower than
237: the temperature directly determined
238: from the IRFM. If only T$_{\rm eff}^{\rm IRFM}$ were
239: used, A$_{Li}$ would be about 0.015 dex higher.
240: 
241: \section{Li Abundances}
242: 
243: In order to determine the Li abundances, we computed a
244: grid of theoretical equivalent widths of the Li
245: doublet in the parameters space:
246: \teff = [5250 K, 5500 K, ..., 6750 K], log $g$ = [3.5, 4.0, 4.5],
247: [Fe/H] = [-1.0, -1.5, ..., -3.5], $A_{Li}$ = [1.7, 1.9, ..., 2.5].
248: We took into account the fine and hyperfine structure of the
249: doublet, with wavelengths and $gf$-values as given in
250: Andersen et al. (1984). Note that Smith et al. (1998) and
251: Hobbs et al. (1999) reported new $gf$-values based on
252: theoretical calculations, but the
253: difference with our adopted (laboratory) values is only 
254: $\approx$ 1\% (Smith et al. 1998). The collisional broadening 
255: constant was obtained from Barklem et al. (2000). 
256: A grid of 630 synthetic spectra was computed employing the 
257: 2002 version of the LTE spectral synthesis code MOOG (Sneden 1973) 
258: and Kurucz overshooting models.
259: The theoretical $W_\lambda$ were measured by integration.
260: 
261: Using the stellar parameters and observed $W_\lambda$
262: as input, the Li abundance was determined
263: from a full 4-dimensional interpolation in the synthetic grid
264: $A_{Li}$(\teff, log $g$, [Fe/H], log $W_\lambda$).
265: Note that in some previous studies (Thorburn 1994; Ryan et al. 1996, 1999)
266: a more restricted approach was adopted, 
267: considering only two fixed log $g$ values (no interpolation) and
268: metallicity interpolation only for models within $-$2 $<$ [Fe/H] $<$ $-$1 
269: (Ryan et al. 1996) or $-$3 $<$ [Fe/H] $<$ $-$2 (Thorburn 1994).
270: Our full interpolation results in lower internal 
271: abundance uncertainties, 
272: %and although the errors involved by the more
273: %restricted interpolations are only of the order of 0.01 dex, the full interpolation 
274: which is very important to determine 
275: the true scatter of the Li abundances.
276: 
277: NLTE effects are small for the Li plateau stars.
278: We applied the NLTE corrections ($\Delta x \equiv$ NLTE - LTE)
279: by Carlsson et al. (1994), which are
280: $\Delta x$ $\approx$ $-$0.04 and 0.0 dex for the hottest 
281: and coolest plateau dwarfs,
282: respectively. Although these NLTE corrections were not specifically 
283: calculated for our set of model atmospheres, their relative
284: dependence with \teff and [Fe/H] is still useful, and
285: should minimize the scatter of the derived abundances. 
286: Since the {\it absolute} NLTE corrections may be different, 
287: we prefer to preserve the mean $A_{Li}$(LTE) of the plateau stars
288: by adopting a NLTE correction of $\Delta x$ + 0.02 dex.
289: The final Li abundances are reported in Table 1.
290: 
291: In Fig. 1 are shown the Li abundances obtained in this work.
292: In the lower panel $A_{Li}$ vs. \teff is plotted, where it is
293: clearly seen that stars with \teff $>$ 6000 K have approximately
294: the same Li abundance (plateau stars), as no trend with
295: \teff (slope of $4 \pm 5 \times 10^{-3}$ per 100 K) is observed. 
296: In the upper panel are plotted
297: 41 plateau stars (\teff $>$ 6000 K) as a function of [Fe/H], 
298: showing an impressive constancy in their Li abundance from 
299: [Fe/H] $\approx$ $-$3.4 to $-$1,
300: with an amazingly small scatter and
301: essentially zero slope (0.013 $\pm$ 0.018 per dex).
302: This figure strongly supports
303: the primordial origin of the Li plateau (Spite \& Spite 1982),
304: and imposes stringent constraints on stellar mixing
305: and Galactic chemical evolution.
306: 
307: The very small star-to-star {\it observed} scatter ($\sigma_{obs}$ = 0.06 dex) 
308: is fully explained by errors in \teff 
309: and $W_\lambda$.
310: Adding in quadrature the typical errors 
311: of $\delta$T = 66 K ($\delta A_{Li} \approx$ 0.045 dex) 
312: and $\delta W_\lambda$ = 5.5\% ($\delta A_{Li} \approx$ 0.025 dex)
313: (as well as errors of 0.3 dex in log $g$ and [Fe/H]), 
314: a predicted error $\sigma_{pred}$ = 0.053 dex is obtained, 
315: in excellent agreement with the observed scatter ($\sigma_{obs}$ = 0.06 dex).
316: Note that if the 2 plateau stars with the highest Li abundances
317: are discarded, then $\sigma_{obs}$ = 0.05 dex.
318: 
319: \section{Discussion}
320: 
321: The abundances obtained in this work for 
322: plateau stars show a very small star-to-star-scatter.
323: This is a consequence of employing very accurate 
324: temperatures based on the IRFM and their photometric 
325: calibrations. 
326: RM04b,c have shown that the IRFM \teff scale is in almost 
327: perfect agreement (at the 10 K level)
328: with interferometric measurements of dwarfs and giants with [Fe/H] $> -0.6$.
329: Provided that the plateau is truly primordial,
330: the constancy of the Li abundance in our plateau stars
331: indicates that our IRFM \teff scale is also 
332: highly reliable when we approach low metallicities,
333: and this would be consistent with the small dependence of the
334: IRFM on the detailed assumptions of model atmospheres.
335: 
336: Although the Li abundances derived in this work have
337: very small internal errors, 
338: %as reflected in the almost zero scatter of the plateau Li abundances, 
339: systematic errors dominate the {\it absolute} abundances. 
340: For example, the use of Kurucz models with no
341: overshooting (NOVER, Castelli et al. 1997)
342: results in smaller $A_{Li}$ by about 0.08 dex. Note
343: that the Kurucz solar overshooting model reproduces
344: more solar observables than the NOVER model (Castelli et al. 1997).
345: Interestingly, the {\it (R-I)}$_{C}$ synthetic colors
346: computed from overshooting models
347: are in better agreement with the IRFM \teff scale 
348: than colors computed with NOVER models,
349: at least for solar metallicity dwarfs 
350: (6000 K $<$ \teff $<$ 6800 K, Mel\'endez \& Ram\'{\i}rez 2003)
351: and giants (4000 K $<$ \teff $<$ 6300 K, Ram\'{\i}rez \& Mel\'endez 2004a).
352: 
353: The mean Li abundance obtained for the plateau stars,
354: $A_{Li}$ = 2.37 dex, is considerably higher than the 
355: primordial Li abundance proposed by R01 ($A_{Li} \approx$ 2.0 - 2.1 dex).
356: As shown in Fig. 2, the reason for this
357: are the lower temperatures adopted by R01.
358: The temperature difference (between our \teff and those
359: adopted by R01) is metallicity dependent, 
360: and at very low metallicities ([Fe/H] $< -3)$ 
361: the \teff adopted by R01 is about 400 K lower than ours. 
362: When the temperatures adopted by R01 are used,
363: we also find a trend of Li abundance with metallicity.
364: If we use the same \teff (and $W_\lambda$) as those employed by
365: R01, $<A_{Li}>$ = 2.13 dex is found for stars with \teff $>$ 6000 K
366: and [Fe/H] $<$ -3, whereas R01 found $<A_{Li}>$ = 
367: 2.07 dex for the same stars
368: (the small difference is due to the different model atmospheres adopted). 
369: Note that when Magain (1987) calibration is used in its validity 
370: range (\teff $<$ 6000 K, [Fe/H] $> -2$), the agreement 
371: with our \teff scale is satisfactory (open circles, Fig. 2);
372: outside that range it may not be appropriate to use it.
373: 
374: Most (0.07 dex) of the 0.13 dex difference between the present
375: Li plateau abundance and that obtained by BM97 is due to different
376: model atmospheres,
377: and the remaining difference is because our \teff scale is 
378: hotter for metal-poor F dwarfs.
379: We have compared our IRFM temperatures with those given by
380: Alonso et al. (1996) (employing the same reddening adopted
381: by Alonso et al.) The agreement is good for metal-poor stars 
382: with \teff $<$ 6500 K, and our temperatures are 
383: just $\approx$ 50 K higher in the range 6000 $<$ \teff $<$ 6500.
384: For hotter stars the differences are larger. 
385: Since only 5 stars with \teff $>$ 6500 K are included 
386: in the present work, their exclusion 
387: does not alter our results (Fig. 1), 
388: but restricts our claim of the constancy of the plateau 
389: to $-3 <$ [Fe/H] $< -1$ and 6000 K $<$ \teff $<$ 6500 K.
390: 
391: Ryan et al. (1999) have criticized Alonso's 
392: calibrations for turn-off stars
393: because of their strong metallicity dependence.
394: Nevertheless, this effect is predicted by both
395: Kurucz and MARCS model atmospheres. In fact, 
396: for metal-poor ([Fe/H] = -3) 
397: F dwarfs, our ($V-K$) IRFM calibration is in good agreement 
398: with $V-K$ colors predicted from MARCS models (Houdashelt
399: et al. 2000). At [Fe/H] = -3, our calibration (RM04c) gives 
400: 6250 and 6500 K for $V-K_{2MASS}$ = 1.31 and 1.21, while Houdashelt
401: calibration results in \teff {\it higher} by +76 and +7 K,
402: respectively. The same exercise employing Kurucz models at $-2.5$
403: (M. S. Bessell 2004, private communication) gives a similar result.
404: 
405: Asplund et al. (2003) have shown that 3D hydrodynamical models with
406: NLTE computations give similar Li abundances than 1D + NLTE. 
407: However, that result was obtained neglecting collisions with H, 
408: and taking them into account, the Li abundance is
409: lowered (Barklem et al. 2003).
410: It is possible to compare 1D calculations 
411: (considering relative NLTE corrections) 
412: with 3D NLTE results by using the plateau star HD 84937, which has predicted
413: equivalent widths in 3D NLTE (Barklem et al. 2003). Employing
414: the same atmospheric parameters and Li abundance as given by 
415: Barklem et al. (2003), we obtain a 1D $W_\lambda$ = 11.4 m\AA,
416: which is very close to the 3D NLTE results, both without collisions
417: (11.1 m\AA) and when collisions are included (12.6 m\AA). Within
418: 0.05 dex, both 3D NLTE calculations are in (probably fortuitous) 
419: agreement with the 1D calculation. 
420: Nevertheless, this result is by any means conclusive, 
421: as the 3D results are based on only one snapshot (Barklem et al. 2003).
422: 
423: The observational star-to-star scatter found in this work 
424: ($\sigma_{obs}$ $\approx$ 0.05-0.06 dex) is considerably lower than those found 
425: in previous analyses of large samples: 0.13, 0.06-0.08, 0.089 and 0.094 dex 
426: according to Thorburn (1994), Spite et al. (1996), Ryan et al. (1996) and BM97, 
427: respectively. This extremely thin scatter of the plateau stars 
428: is fully explained by small errors in \teff and $W_\lambda$,
429: %($\sigma_{pred}$ = 0.05 dex), 
430: leaving little room (if any) 
431: for mixing or Galactic evolution.
432: Considering the constancy 
433: %(as a function of both \teff and [Fe/H])
434: of the Li abundance of the plateau stars, depletion is
435: constrained probably at a level of 0.05 dex, and if
436: rotational mixing plays a role, then our mean $A_{Li}$ 
437: should be slightly increased.
438: 
439: Within the {\it absolute} abundance uncertainties involved (of the
440: order of 0.1 dex)
441: in our simple 1D modeling, 
442: we conclude that our results are just in marginal agreement with the 
443: constraint imposed by WMAP,
444: and if new physics or non-standard Big Bang nucleosynthesis
445: are invoked (e.g. Feng et al. 2003; Ichikawa \& Kawasaki 2004), 
446: the present work constraints their effects to a level of 0.1-0.15 dex.
447: 
448: \acknowledgements
449: We thank the referee for his constructive comments,
450: and JG Cohen for her suggestions. 
451: JM acknowledges support from NSF grant AST-0205951 to JGC.
452: 
453: 
454: \begin{references}
455: 
456: \reference{} Alonso, A., Arribas, S. \&  Mart\'{\i}nez-Roger, C. 1996, 
457: A\&AS, 117, 227
458: 
459: \reference{} Andersen J., Gustafsson B., Lambert, D.L., 1984, A\&A, 136, 65
460: 
461: \reference{} Asplund, M., Carlsson, M., Botnen, A. V. 2003, A\&A, 399, L31
462: 
463: \reference{} Barklem, P. S., Piskunov, N. \& O'Mara, B. J. 2000, A\&AS, 142, 467
464: 
465: \reference{} Barklem, P. S., Belyaev, A. K., Asplund M. 2003, A\&A, 409, L1
466: 
467: \reference{} Bonifacio, P. \& Molaro, P. 1997, MNRAS, 285, 847 (BM97)
468: 
469: \reference{} Bonifacio, P., Pasquini, L., Spite, F. et al. 2002, A\&A, 390, 91
470: 
471: \reference{} Carlsson, M., Rutten, R. J., Bruls, J.H.M.J., Shchukina, N. G.
472: 1994, A\&A, 288, 860
473: 
474: \reference{} Castelli, F., Gratton, R. G. \& Kurucz, R. 1997, A\&A, 318, 841
475: 
476: \reference{} Cayrel de Strobel, G., Soubiran, C., Ralite, N. 2001, A\&A 273, 159
477: 
478: \reference{} Cyburt, R. H., Fields, B. D. \& Olive, K. A. 2003, Physics Letters B, 567, 227
479: 
480: \reference{} Feng, J. L., Rajaraman, A., Takayama, F. 2003, Phys. Rev. D 68, 063504
481: 
482: \reference{} Ford A., Jeffries, R. D., Smalley, B., et al. 2002, A\&A, 393, 617
483: 
484: \reference{} Guti\'errez, C. M., Garc\'{\i}a L\'opez, R. J., 
485: Rebolo, R., Mart\'{\i}n, E. L., Fran\c cois, P. 1999, A\&AS, 137, 93
486: 
487: \reference{} Hobbs, L. M., Thorburn, J. A., \& Rebull, L. M. 1999, ApJ, 523, 797
488: 
489: \reference{} Houdashelt, M. L., Bell, R. A., \& Sweigart, A. V. 2000, AJ, 119, 1448
490: 
491: \reference{} Ichikawa K. \& Kawasaki, M. 2004, Phys. Rev. D 69, 123506
492: 
493: \reference{} King, J. R. 1999, PASP, 111, 817
494: 
495: \reference{} Magain, P. 1987, A\&A, 181, 323
496: 
497: \reference{} Mel\'endez, J. \& Ram\'{\i}rez I. 2003, A\&A, 398, 705
498: 
499: \reference{} Mermilliod, J. C., Mermilliod, M. \& Hauck, B. 1997, A\&AS, 124, 349
500: 
501: \reference{} Molaro, P., Primas, F., Bonifacio, P. 1995, A\&A, 295, L47
502: 
503: \reference{} Nissen, P. E. \& Schuster, W. J. 1997, A\&A 326, 751
504: 
505: \reference{} Ram\'{\i}rez I. \& Mel\'endez J. 2004a, A\&A, 417, 301
506: 
507: \reference{} Ram\'{\i}rez I. \& Mel\'endez J. 2004b, ApJ, submitted (RM04b)
508: 
509: \reference{} Ram\'{\i}rez I. \& Mel\'endez J. 2004c, ApJ, submitted (RM04c)
510: 
511: \reference{} Ryan, S. G., Beers, T. C., Deliyannis, C. P. \& Thorburn J. A. 1996, ApJ, 458, 543 (R01)
512: 
513: \reference{} Ryan, S. G. \& Deliyannis, C. P. 1998, ApJ, 500, 398
514: 
515: \reference{} Ryan, S. G., Norris, J. E., Beers, T. C. 1999, ApJ, 523, 654 (R01)
516: 
517: \reference{} Ryan, S. G., Kajino, T., Beers, T. C. et al. 2001, ApJ, 549, 55 (R01)
518: 
519: \reference{} Schlegel, D. J., Finkbeiner, D. P. \& Davis, M. 1998, ApJ, 500, 525
520: 
521: \reference{} Smith, V. V., Lambert, D. L., Nissen, P. E. 1998, ApJ, 506, 405
522: 
523: \reference{} Sneden, C. 1973, Ph.D. thesis, University of Texas
524: 
525: \reference{} Spergel, D. N., Verde, L., Peiris, H. V. et al. 2003, ApJS, 148, 175
526: 
527: \reference{} Spite, F. \& Spite, M. 1982, A\&A, 115, 357
528: 
529: \reference{} Spite, M., Fran\c cois, P., Nissen, P. E., Spite, F. 
530: 1996, A\&A, 307, 172
531: 
532: \reference{} Thorburn, J. A. 1994, ApJ, 421, 318
533: 
534: \end{references}
535: 
536: \clearpage
537: 
538: \begin{figure}
539: \epsscale{}
540: \plotone{f1.eps}
541: \caption{Upper panel: 
542: $A_{Li}$ of the Spite plateau stars (\teff $>$ 6000 K)
543: as a function of [Fe/H]. 
544: Dotted lines indicate the mean
545: Li abundance of the plateau stars, and the solid line represents the
546: lower limit imposed by the WMAP constraint.
547: The error bars are the predicted error, 
548: $\sigma_{pred}$ (= 0.05 dex), and 3$\sigma_{pred}$ (= 0.15 dex).
549: Lower panel: $A_{Li}$ as a function of \teff
550: for stars with [Fe/H] $\leq -1.5$ (filled triangles) and 
551: [Fe/H] $> -1.5$ 
552: (open triangles). The star (HD 106038) with the highest 
553: Li abundance (open triangle inside the open circle) is a star
554: with peculiar abundances (Nissen \& Schuster 1997).
555: }
556: \end{figure}
557: 
558: 
559: \begin{figure}
560: \epsscale{}
561: \plotone{f2.eps}
562: \caption{Temperatures obtained in this work minus the temperatures 
563: from R01 (for stars in common with the present sample) 
564: as a function of the metallicities
565: adopted in the present work. Filled circles: plateau stars (\teff $>$ 6000 K);
566: open circles: stars with \teff $<$ 6000 K.}
567: \label{ryan}
568: \end{figure}
569: 
570: 
571: 
572: \begin{deluxetable}{lccccrcrrrrl}
573: %\tablenum{1}
574: \label{tab1}
575: \tablewidth{0pt}
576: \tabletypesize
577: \scriptsize
578: \tablecaption{Li abundances for the Spite plateau stars}
579: \tablehead{
580: \colhead{Star}  &
581: \colhead{${\rm T_{eff}}$} &
582: \colhead{log $g$} &
583: \colhead{[Fe/H]} & 
584: \colhead{$W_\lambda$} & 
585: \colhead{$A_{Li}$} \\
586: \colhead{}  &
587: \colhead{[K]} &
588: \colhead{[cm s$^{-2}$]} &
589: \colhead{[dex]} & 
590: \colhead{[m\AA]} & 
591: \colhead{[dex]}
592: }
593: \startdata
594: BD+002058   & 6092 &  4.19 & -1.16 & 41.9 & 2.49 \\
595: BD+023375   & 6124 &  3.97 & -2.29 & 32.5 & 2.34 \\
596: BD+030740   & 6443 &  3.76 & -2.68 & 19.5 & 2.31 \\
597: BD+090352   & 6075 &  4.38 & -2.03 & 34.0 & 2.34 \\
598: BD+092190   & 6487 &  4.11 & -2.05 & 17.3 & 2.28 \\
599: BD+174708   & 6154 &  3.93 & -1.64 & 25.8 & 2.28 \\
600: BD+203603   & 6248 &  4.15 & -2.12 & 26.2 & 2.32 \\
601: BD+241676   & 6510 &  3.91 & -2.32 & 25.3 & 2.46 \\
602: BD+262606   & 6157 &  4.20 & -2.37 & 30.5 & 2.33 \\
603: BD+282137   & 6229 &  3.50 & -2.20 & 27.2 & 2.33 \\
604: BD+342476   & 6433 &  3.92 & -2.10 & 22.6 & 2.37 \\
605: BD+422667   & 6086 &  4.13 & -1.51 & 30.6 & 2.32 \\
606: BD+592723   & 6121 &  4.00 & -1.83 & 30.5 & 2.33 \\
607: BD+710031   & 6331 &  4.05 & -1.96 & 31.0 & 2.44 \\
608: BD-043208   & 6404 &  3.77 & -2.36 & 25.4 & 2.40 \\
609: BD-100388   & 6240 &  3.61 & -2.37 & 27.0 & 2.33 \\
610: BD-133442   & 6484 &  3.98 & -2.87 & 20.5 & 2.35 \\
611: CD-2417504  & 6507 &  4.12 & -3.36 & 18.6 & 2.31 \\
612: CD-3018140  & 6336 &  4.08 & -1.95 & 30.3 & 2.44 \\
613: CD-3301173  & 6674 &  4.19 & -2.99 & 16.3 & 2.35 \\
614: CD-7101234  & 6375 &  4.43 & -2.35 & 26.0 & 2.38 \\
615: G059-024    & 6191 &  4.36 & -2.38 & 36.0 & 2.42 \\
616: G064-012    & 6682 &  4.10 & -3.31 & 23.4 & 2.53 \\
617: G064-037    & 6775 &  4.05 & -3.13 & 15.6 & 2.39 \\
618: G075-031    & 6036 &  4.07 & -1.04 & 48.0 & 2.54 \\
619: G192-043    & 6184 &  4.30 & -1.61 & 34.5 & 2.42 \\
620: G201-005    & 6370 &  3.79 & -2.51 & 25.0 & 2.37 \\
621: G206-034    & 6310 &  4.18 & -2.53 & 27.2 & 2.37 \\
622: HD016031    & 6216 &  4.07 & -1.82 & 27.0 & 2.33 \\
623: HD074000    & 6392 &  4.27 & -1.96 & 23.9 & 2.36 \\
624: HD084937    & 6345 &  3.96 & -2.13 & 24.8 & 2.35 \\
625: HD102200    & 6104 &  4.14 & -1.21 & 32.8 & 2.38 \\
626: HD108177    & 6099 &  4.30 & -1.64 & 31.4 & 2.33 \\
627: HD160617    & 6099 &  3.68 & -1.79 & 40.0 & 2.43 \\
628: HD166913    & 6096 &  4.08 & -1.58 & 37.0 & 2.40 \\
629: HD181743    & 6038 &  4.47 & -1.81 & 38.0 & 2.38 \\
630: HD218502    & 6222 &  3.94 & -1.79 & 27.1 & 2.34 \\
631: HD284248    & 6133 &  4.21 & -1.61 & 29.9 & 2.33 \\
632: HD338529    & 6504 &  3.90 & -2.27 & 23.9 & 2.43 \\
633: LP0056-0075 & 6197 &  4.36 & -2.67 & 25.5 & 2.27 \\
634: LP0831-0070 & 6521 &  4.23 & -3.07 & 23.8 & 2.42 \\
635: \enddata					  
636: \end{deluxetable}
637: 
638: 
639: 
640: \end{document}
641: 
642: