1: \documentclass[preprint]{aastex}
2:
3: \begin{document}
4: \title{Measurement of the Europium Isotope Ratio for the Extremely Metal-Poor, $r$-Process-Enhanced Star
5: CS~31082-001\footnote{Based on data collected at the Subaru
6: Telescope, which is operated by the National Astronomical Observatory
7: of Japan.}}
8:
9: \author{Wako Aoki\altaffilmark{2}, Satoshi Honda\altaffilmark{2}, Timothy C. Beers\altaffilmark{3}, Christopher Sneden\altaffilmark{4}}
10: \altaffiltext{2}{National Astronomical Observatory, Mitaka, Tokyo, 181-8588 Japan; email: aoki.wako@nao.ac.jp, honda@optik.mtk.nao.ac.jp}
11: \altaffiltext{3}{Department of Physics and Astronomy, Michigan State University, East Lansing, MI 48824-1116; email: beers@pa.msu.edu}
12: \altaffiltext{4}{Department of Astronomy and McDonald Observatory, University of Texas, Austin, TX 78712; email: chris@verdi.as.utexas.edu}
13:
14: \begin{abstract}
15:
16: We report the first measurement of the isotope fraction of europium
17: ($^{151}$Eu and $^{153}$Eu) for the extremely metal-poor,
18: $r$-process-enhanced star CS~31082--001, based on high-resolution
19: spectra obtained with the Subaru Telescope High Dispersion
20: Spectrograph. We have also obtained new measurements of this ratio
21: for two similar stars with previous europium isotope measurements,
22: CS~22892--052 and HD~115444. The measurements were made using
23: observations of the Eu lines in these spectra that are most
24: significantly affected by isotope shifts and hyperfine splitting. The
25: fractions of $^{151}$Eu derived for CS~31082--001, CS~22892--052, and
26: HD~115444 are 0.44, 0.51, and 0.46, respectively, with uncertainties
27: of about $\pm 0.1$. CS~31082--001, the first star with a meaningful
28: measurement of U outside of the solar system, is known to exhibit
29: peculiar abundance ratios between the actinide and rare-earth elements
30: (e.g., Th/Eu), ratios that are significantly different from those for
31: other stars with large excesses of $r$-process elements, such as our
32: two comparison objects. Nevertheless, our analysis indicates that the
33: Eu isotope ratio of CS~31082--001 agrees, within the errors, with
34: those of other $r$-process-enhanced objects, and with that of
35: solar-system material.
36: %The possibility that there may exist a small
37: %dispersion in the Eu isotope ratios among $r$-process-enhanced stars
38: %is also discussed.
39:
40: \end{abstract}
41:
42: \keywords{Galaxy: abundances -- nuclear reactions, nucleosynthesis, abundances -- stars: abundances -- stars: Population II -- stars: individual (CS~31082--001)}
43:
44: \section{Introduction}\label{sec:intro}
45:
46: The chemical composition of very metal-poor ([Fe/H]$\lesssim -2.5$
47: \footnote{[A/B] = $\log(N_{\rm A}/N_{\rm B})- \log(N_{\rm A}/N_{\rm
48: B})_{\odot}$, and $\log \epsilon_{\rm A} = \log(N_{\rm A}/N_{\rm
49: H})+12$ for elements A and B.}) stars is expected to be determined by
50: a small number of nucleosynthesis events that preceded the formation
51: of these objects, while that of metal-rich stars like the Sun are the
52: result of the accumulated stellar yields throughout the long history
53: of the Galaxy. In recent years, abundance studies of heavy elements in
54: very metal-poor stars have provided quite important constraints on
55: models of the astrophysical neutron-capture processes, both the
56: $s$low-process and the $r$apid-process. Discoveries of
57: $r$-process-element-enhanced, very metal-poor stars, and their
58: subsequent abundance analyses, have shown that the abundance pattern
59: of neutron-capture elements with $Z \geq 56$ agrees very well with the
60: $r$-process component abundances in solar-system material \citep[e.g.,
61: ][]{sneden96, westin00}. These results suggest that the abundance
62: pattern of the $r$-process component in the solar system for $Z \geq
63: 56$ is not the result of a mixture of quite different abundance
64: patterns produced by individual processes, but rather, that the
65: patterns produced by individual processes are quite similar throughout
66: Galactic history.
67:
68: By way of contrast, the abundances of light {\it r}-process elements
69: ($Z<56$) of these objects are known to show a deviation from the
70: $r$-process component in the solar system. For instance,
71: \citet{sneden00} studied the abundances of neutron-capture elements,
72: including six elements with $40<Z<56$, of CS~22892--052 in detail, and
73: found a clear deviation from the scaled solar-system $r$-process
74: component, especially for Y, Rh, and Ag. They concluded that different
75: {\it r}-process sites may be responsible for the formation of the
76: lighter ($40<Z<56$) and heavier ($Z\geq 56$) neutron-capture elements.
77: A phenomenological model for these results was proposed by
78: \citet{wasserburg01}.
79:
80: In addition to the above studies of the {\it total} abundances of
81: neutron-capture elements, analyses of isotope fractions for individual
82: elements are also quite important for making detailed comparison with
83: predictions from theoretical models for the $r$-process. Europium is
84: one of the elements which have comparatively large isotope shifts
85: between the spectral lines of individual components ($^{151}$Eu and
86: $^{153}$Eu). \citet{sneden02} carried out an analysis of the Eu
87: isotope fractions for three $r$-process-element-enhanced metal-poor
88: stars, and showed that the fractions of $^{151}$Eu and $^{153}$Eu are
89: 0.5, agreeing with that of solar-system material. This demonstrates
90: that, even at the isotopic level, there is agreement between the
91: abundance pattern of {\it r}-process elements in very metal-poor stars
92: with the $r$-process component in the solar system.
93:
94: On the other hand, the extremely $r$-process-element-enhanced, very
95: metal-poor ([Fe/H] $= -2.9$) star CS~31082-001, first reported on by
96: \citet{cayrel01}, exhibits a somewhat different abundance pattern as
97: compared to otherwise similar stars. The abundances of heavy
98: ($Z\geq72$) neutron-capture elements in this star are {\it higher}
99: than would be predicted from the abundances of elements with $56\leq Z
100: \lesssim 70$, when the scaled solar-system $r$-process component
101: pattern is extended to the third $r$-process peak \citep{hill02}. This
102: suggests that the site and/or conditions under which $r$-process
103: nucleosynthesis of (at least) the heavy neutron-capture elements
104: occurs are not unique, but have some star-to-star variation. The
105: theoretical implications of CS~31082-001 have been considered in
106: detail by \citet{qian01}, Wanajo et al. (2002), and Schatz et
107: al. (2002).
108:
109: The agreement of the abundance pattern of CS~31082-001 with that of
110: the solar-system $r$-process component only appears to be valid in the
111: range $56 \leq Z \lesssim 70$. One open question at this stage is
112: whether the isotope ratios, such as $^{151}$Eu/$^{153}$Eu, agree with
113: those of other $r$-process-element-enhanced stars, as well as with
114: that of the solar-system material. In this paper, we report the
115: isotope fractions of Eu derived for three stars, CS~31082-001,
116: CS~22892-052 and HD~115444, based on high-resolution spectra obtained
117: with the Subaru Telescope and the High Dispersion Spectrograph (HDS,
118: Noguchi et al. 2002). The latter two objects have already been studied
119: by \citet{sneden02}, but for comparison purposes we have carried out
120: an independent analysis using new spectra with higher resolution than
121: were previously available.
122:
123: \section{Observations and Measurements}\label{sec:obs}
124:
125: Observations were carried out with the HDS of the 8.2m Subaru
126: Telescope. The detector is a mosaic of two 4k $\times$ 2k EEV-CCD's
127: with 13.5~$\mu$m pixels. Besides CS~31082-001, we observed two other
128: $r$-process-element-enhanced metal-poor stars, CS~22892-052
129: \citep{sneden96} and HD~115444 \citep{westin00}. These two objects
130: have already been studied by \citet{sneden02}, and are quite useful
131: for comparison purposes. Details of the observations are provided in
132: Table \ref{tab:obs}. While the spectra of CS~22892-052 and HD~115444
133: were taken with a resolving power $R=90,000$, CS~31082-001 was
134: observed with somewhat lower resolution, $R=60,000$, due to poor
135: weather conditions. Nevertheless, this resolving power is still
136: sufficient for the present analysis. The S/N ratios per 0.012 {\AA}
137: pixel at 4100 {\AA} are 110, 65, and 300 for CS~31082-001,
138: CS~22892-052 and HD~115444, respectively. Data reduction was performed
139: in the standard way within the IRAF\footnote{IRAF is distributed by
140: the National Optical Astronomy Observatories, which is operated by the
141: Association of Universities for Research in Astronomy, Inc. under
142: cooperative agreement with the National Science Foundation.}
143: environment, following procedures described by \citet{aoki02}.
144:
145: For future studies on the possible binarity of these objects, which
146: will be investigated on the basis of observed radial-velocity
147: variations, Table 1 also provides (heliocentric) radial velocities
148: measured for our spectra. The radial velocities were measured using
149: clean \ion{Fe}{1} lines. The radial velocity of CS~22892-052 obtained
150: in our present work is +13.2 km~s$^{-1}$, and shows no significant
151: changes from the results of \citet{preston01}, who suggested a small
152: variation of the radial velocity (with a period of 120 days) for this
153: object. CS~31082-001 was observed in July and October 2001. The
154: results of the two measurements agree very well, and no variation of
155: the radial velocity was found with respect to the results of
156: \citet{hill02}. Though there is no clear evidence for binarity of
157: these objects from the above measurements, further long-term
158: monitoring of their velocities is required to obtain any definitive
159: conclusions. We note that our abundance analysis for CS~31082-001 is
160: based only on the October 2001 spectrum, because the quality of the
161: data is superior to that obtained in July 2001.
162:
163: \section{Analysis of Europium isotope Ratios and Results}\label{sec:ana}
164:
165: The Eu isotope ratios were measured by fitting the observed spectra of
166: Eu lines with synthetic spectra calculated using model atmospheres
167: \citep{kurucz93} including the effect of the isotope shifts. We
168: adopted atmospheric parameters for our objects (effective temperature:
169: $T_{\rm eff}$, surface gravity: $\log g$, micro-turbulent velocity:
170: $v_{\rm micro}$, and metallicity, as represented by the iron
171: abundance, [Fe/H]) that are similar to those in previous work: $T_{\rm
172: eff}$(K) / $\log g$ /$v_{\rm micro}$(kms$^{-1}$) / [Fe/H] =
173: 4800/1.5/1.8/$-3.0$ for CS~31082-001 \citep{cayrel01},
174: 4750/1.3/2.3/$-3.0$ for CS~22892-052 \citep{sneden96}, and
175: 4650/1.5/2.1/$-3.0$ for HD~115444 \citep{westin00}. It is worth
176: stressing that, in contrast to studies of elemental abundances, the
177: analyses of isotope ratios are generally quite {\it insensitive} to
178: the adopted atmospheric parameters.
179:
180: The effects of isotope shifts and hyperfine splitting in Eu lines have
181: been discussed by \citet{lawler01}. The line list produced by
182: \citet{sneden02}, based on \citet{lawler01}, was applied in the
183: present work. \citet{sneden02} analyzed the \ion{Eu}{2} lines at 3819,
184: 3907, 4129, and 4205~{\AA}. Unfortunately, the 3907~{\AA} line was
185: not covered in our CS~31082-001 spectrum due to the gap between the
186: two CCD's of HDS. Instead, we analyzed the weaker \ion{Eu}{2} line at
187: 4435~{\AA} for CS~31082-001 and CS~22892-052. As shown in the
188: following analysis, this line is actually quite useful for the
189: measurement of the isotope ratios in these extremely Eu-enhanced
190: stars, because the lines at 3819, 4129, and 4205~{\AA} are
191: sufficiently strong that they compromise the accuracy of the derived
192: isotope fractions, as well as the total abundance. We note that the
193: 4435~{\AA} line in HD~115444 is too weak, hence it was not used in the
194: analysis of this object. The \ion{Eu}{2} 4522~{\AA} line has an
195: appropriate strength for isotopic and abundance analyses, but was not
196: used because of severe blending with other elemental lines.
197:
198: The instrumental profile of HDS, for spectra with resolving power
199: higher than $R \sim 90,000$, can be well-approximated by a Gaussian
200: shape, but this does not pertain to spectra obtained with lower
201: resolving powers \citep{noguchi02}. Accordingly, we have employed
202: Gaussian profiles for the instrumental line broadening for calculation
203: of the synthetic spectra of CS~22892-052 and HD~115444, which were
204: observed with $R=90,000$. For the analysis of CS~31082-001, which was
205: observed with $R=60,000$, we measured the instrumental profile from
206: the Th emission lines obtained for the wavelength calibration, and
207: applied the resulting profile to the spectrum synthesis calculations.
208:
209: The macro-turbulence for the individual stellar atmospheres was
210: estimated by fitting clean \ion{Fe}{1} lines with synthetic spectra
211: for each object, assuming Gaussian profiles. This approximation
212: should be valid since the axial rotation of these (very old) red
213: giants is expected to be quite small.
214:
215: In order to estimate the quality of the fit between observed and
216: synthetic spectra, we calculated the values of reduced $\chi^{2}$,
217: defined as:
218:
219: \begin{eqnarray}
220: \chi_{\rm r}^{2}= \frac{1}{\nu
221: -1}\sum_{i}\frac{(O_{i}-C_{i})^{2}}{\sigma_{i}^{2}} ,\nonumber
222: \end{eqnarray}
223:
224: \noindent
225: where $(O_{i}-C_{i})$ is the difference between the observed and
226: synthetic spectra at the $i$-th spectrum point \citep[e.g., ]
227: []{smith01}. The quantity $\sigma_{i}$ is defined as $\sigma_{i} =
228: (S/N \times \sqrt{f_{i}})^{-1}$, where $S/N$ is the signal-to-noise
229: ratio of the continuum level, and $f_{i}$ indicates the normalized
230: flux at the $i$-th point ($0\leq f_{i}\leq 1$ for absorption
231: profiles). In the above equation, $\nu$ is the number of degrees of
232: freedom in the fit, and is approximately the number of data points to
233: which the fit is applied (20-30 pixels). Then, $\chi_{\rm r}^{2}$ is
234: represented as:
235:
236: \begin{eqnarray}
237: \chi_{\rm r}^{2} \sim <\frac{(O_{i}-C_{i})^{2}}{\sigma_{i}^{2}}> . \nonumber
238: \end{eqnarray}
239:
240: \noindent
241: By dividing $O-C$ by $\sigma$, the dependence of the data quality on
242: the depth of the absorption is accounted for. In this definition,
243: $\chi_{\rm r}^{2}$ is expected to be unity in the best-fit case for
244: data of a given $S/N$, because the $\chi_{\rm r}^{2}$ value, taking
245: the $S/N$ ratio at the line into consideration, cancels out the
246: line-to-line and object-to-object differences in data quality. In the
247: following presentation, however, we show the result for a fixed value
248: of $S/N$, in order to compare directly the goodness of the fit between
249: individual lines. We fix the $S/N$ to be 100, which roughly represents
250: the $S/N$ ratio of the CS~31082-001 spectrum.
251:
252: We show examples of the observed and synthetic spectra in Figure
253: \ref{fig:4129A} and \ref{fig:4435A} for the \ion{Eu}{2} 4129~{\AA} and
254: 4435~{\AA} lines, respectively. We first assumed the fraction of
255: $^{151}$Eu to be 0.5, and derived the total Eu abundance for each
256: line. Then we fitted the synthetic spectra to the observed ones for
257: the longer wavelength (redder) part of the spectral line, which is
258: insensitive to the assumed isotope ratio, by shifting the observed
259: spectra. The shifts we applied are smaller than 0.01~{\AA}. These are
260: reasonable, as the uncertainties of the wavelength calibration of the
261: spectra and the absolute wavelength of the Eu lines are likely to be
262: in error at this level. We then calculated the $\chi_{\rm r}^{2}$ for
263: the red portion of the line to determine the wavelength shifts. We
264: estimated the uncertainties of the wavelength determination by taking
265: a value where $\chi_{\rm r}^{2}$ is twice larger than the best-fit
266: case.
267:
268: We next proceeded to the determination of the Eu isotope fractions, as
269: well as the total Eu abundances. We searched for the isotope
270: fractions that resulted in the smallest $\chi_{\rm r}^{2}$ for a given
271: Eu abundance. This analysis was carried out by changing the Eu
272: abundance in steps of 0.01~dex, and adopting the Eu abundances and
273: isotope fractions that gave the best $\chi_{\rm r}^{2}$. We estimated
274: the uncertainty of the derived Eu abundances by considering the range
275: in abundance over which the $\chi_{\rm r}^{2}$ is twice as large as
276: the best-fit case. The errors in the isotope fractions due to the
277: uncertainty of the total abundance of Eu were estimated from the range
278: of the isotope fractions that were allowed within the adopted
279: abundance uncertainty. As an example of this procedure,
280: Figure~\ref{fig:fit} shows the values of $\chi_{\rm r}^{2}$, as a
281: function of $^{151}$Eu/$^{153}$Eu, for the \ion{Eu}{2} 4129~{\AA}
282: line.
283:
284: Table~\ref{tab:results} lists the derived isotope fractions and the Eu
285: abundances determined by the above analysis for each line. The
286: $\chi_{\rm r}^{2}$ value and the fitting error ($\sigma_{\rm fit}$)
287: are also given. The fitting errors given here were simply estimated as
288: being those obtained when the $\chi_{\rm r}^{2}$ was assumed to be
289: twice larger than the best-fit case at the adopted Eu abundance. We
290: also estimated the errors due to the uncertainties of the adopted Eu
291: abundance, the macro-turbulent velocity, the continuum level, and the
292: wavelength calibration of the spectrum and the position of the Eu line
293: (see below). The total error ($\sigma_{\rm total}$) was estimated by
294: adding, in quadrature, these errors to $\sigma_{\rm fit}$, and is also
295: given in Table~\ref{tab:results}.
296:
297: \subsection{HD~115444}
298:
299: The values of $\chi_{\rm r}^{2}$ obtained from the fits of the three
300: Eu lines in HD~115444 are much smaller than those obtained for the
301: other two stars, which is expected due to the considerably higher S/N
302: ratio of the spectrum of this object. The small values of $\chi_{\rm
303: r}^{2}$, and the agreement of the $^{151}$Eu fractions derived from
304: the three lines consider, indicate that the Eu isotope fraction
305: derived for HD~115444 is quite reliable. The average of the
306: $^{151}$Eu fraction from the three lines is fr($^{151}$Eu)=0.46.
307:
308: We estimated the errors in our derived isotope fractions due to
309: uncertainties in (1) macro-turbulent velocity ($\Delta v_{\rm
310: macro}$), (2) wavelength calibration and line position ($\Delta
311: \lambda$), (3) continuum level of the spectrum ($\Delta$(cont)), and
312: (4) Eu abundance adopted ($\Delta log \epsilon$ (Eu)). The uncertainty
313: of the Eu abundance as well as the wavelength calibration (or absolute
314: line position) have been estimated as noted above, and are $\Delta
315: log\epsilon$ (Eu)=0.02~dex and $\Delta \lambda$=0.002~{\AA},
316: respectively. We assumed uncertainties for the macro-turbulent
317: velocity and the continuum level to be $\Delta v_{\rm
318: macro}=1$~km~s$^{-1}$ and $\Delta$(cont)=1\%, respectively, for this
319: star. These assumption are probably conservative for a spectrum with
320: $S/N$ = 200 $\sim 300$. The errors due to these uncertainties are
321: $\Delta$fr($^{151}$Eu)=$-0.07, -0.03$, and $-0.02$ for $\Delta v_{\rm
322: macro}$=+1.0~km~s$^{-1}$, $\Delta \lambda$=+0.002~{\AA}, and
323: $\Delta$(cont)=+0.01, respectively. The sense of the change in the
324: derived $^{151}$Eu fraction with respect to the change of the total Eu
325: abundance is dependent on the line under consideration. The
326: uncertainties in fr($^{151}$Eu) are about 0.02 for $\Delta log
327: \epsilon$ (Eu)=$\pm$ 0.02~dex.
328:
329: \subsection{CS~22892-052}
330:
331: A similar analysis was applied to the spectrum of CS~22892-052. The
332: lower $S/N$ of the spectrum of this object, as compared to that of
333: HD~115444, results in the larger values of $\chi_{\rm r}^{2}$, and the
334: larger fitting errors. In addition, there is a difficulty in the
335: analysis of the \ion{Eu}{2} 3819~{\AA} line. This line is quite strong
336: -- the central part of the absorption line is almost saturated. As a
337: result, although the line {\it depth} is insensitive to the increase
338: of the total Eu abundance, the line {\it width} increases along with
339: the Eu abundance, due to the growth of weak components in the bluer
340: part of the line to which weak $^{151}$Eu lines primarily
341: contribute. This behavior is similar to that for the increase of the
342: fraction of $^{151}$Eu for a given Eu abundance. As a result, the
343: effects of the changes of the total Eu abundance and the fraction of
344: $^{151}$Eu are degenerate. We attempted to estimate the uncertainty
345: of the $^{151}$Eu due to that of the Eu abundance by the same manner
346: as other lines, and found quite large uncertainties: $\Delta$
347: fr($^{151}$Eu)=0.15 and $\Delta log \epsilon$(Eu)=0.10~dex. For this
348: reason, we decided to exclude this line for determination of the
349: isotope ratio, and adopted the average of the values determined from
350: the other three lines, fr($^{151}$Eu)=0.51, as the final result. We
351: note that the uncertainty of the derived $^{151}$Eu fraction due to
352: the uncertainty in the total Eu abundance obtained from the
353: \ion{Eu}{2} 4129~{\AA} line is rather large (0.09). This is partly
354: due to the strength of this line, as in the case of the 3819~{\AA}
355: line, but the poor quality of the fit to this line also contributes.
356: The uncertainties arising from other factors are small ($\leq 0.04$),
357: even though a larger error in the continuum determination
358: ($\Delta$(cont)=0.02) was assumed for this object, taking the low
359: $S/N$ ratio of the spectrum into consideration.
360:
361: \subsection{CS~31082-001}
362:
363: The \ion{Eu}{2} lines in the spectrum of CS~31082-001 are even
364: stronger than those in the spectrum of CS~22892-052, due to the larger
365: excesses of the $r$-process elements, which are boosted by about
366: 0.4~dex relative to this star \citep{hill02}. We excluded the
367: 3819~{\AA} line from the analysis for this star, as we did in the
368: analysis of CS~22892-052. The other two lines, at 4129~{\AA} and
369: 4205~{\AA}, are also quite strong. The uncertainties of
370: fr($^{151}$Eu) due to that of the total Eu abundance are 0.07; the
371: most important factors arise from the total errors. In this sense, the
372: isotope ratio derived from the \ion{Eu}{2} 4435~{\AA} line is the most
373: reliable, because the line strength is appropriate for the analysis of
374: the total abundance and isotope fractions, even though the fitting
375: error for this line is larger than for other lines. The uncertainties
376: arising from other factors are minor.
377:
378: %\newpage
379: \section{Discussion and Concluding Remarks}
380:
381: Figure~\ref{fig:results} shows the results for our derived $^{151}$Eu
382: fractions based on each line considered; the error bars indicate the
383: uncertainty ($\sigma_{\rm total}$). The average of the results from
384: individual lines are 0.46, 0.51, and 0.44 for HD~115444, CS~22892-052
385: and CS~31082-001, respectively. The calculation, weighted by
386: $\sigma^{-1}$, alters the results by less than 0.01. The results for
387: HD~115444 and CS~22892-052 show excellent agreement with those by
388: \citet{sneden02}, who derived fr($^{151}$Eu)=0.5$\pm$0.1 for these
389: objects.
390:
391: The dotted line in Figure~\ref{fig:results} indicates the $^{151}$Eu
392: fraction in solar-system material (fr($^{151}$Eu)=0.478, Anders \&
393: Grevesse 1989). Since 95\% of the Eu in solar-system material is
394: expected to originate from the $r$-process \citep{arlandini99}, this
395: ratio well represents that of the $r$-process component in the solar
396: system\footnote{Though the $^{151}$Eu fraction of the $s$-process
397: component has not been constrained by observations,
398: \citet{arlandini99} predicted it to be 0.54, similar to that of the
399: $r$-process component (0.47), based on their stellar model. For this
400: reason, the contribution of the $s$-process nucleosynthesis is not
401: estimated from the Eu isotopes. However, it should be negligible in
402: the very metal-poor stars studied here, in particular for Eu.}. The
403: $^{151}$Eu fractions in our three objects, including CS~31082-001, are
404: {\it consistent with the solar-system value.}
405:
406: One clear difference between the abundance pattern of CS~22892-052 and
407: that of CS~31082-001 is the abundance ratios between the actinide and
408: the rare-earth elements produced by $r$-process nucleosynthesis (e.g.,
409: the Th/Eu ratio, Hill et al. 2002). This difference is speculated to
410: arise from the variety in the ratios of the neutron-to-seed-nuclei in
411: the $r$-process site which contributed to the abundances of heavy
412: elements in these objects. Theoretical studies of $r$-process
413: nucleosynthesis have shown that the neutron to seed-nuclei ratio is
414: strongly dependent on the entropy-per-baryon ratio and the electron
415: fraction ($Y_{e}$) in the nucleosyhthesis site, as well as on the
416: dynamic time-scale of the event \citep[e.g., ]
417: []{hoffman97,otsuki00,wanajo02}. These quantities are, however, quite
418: difficult to estimate from theoretical studies, hence numerical
419: simulations of the $r$-process often treat them as free
420: parameters.
421:
422: In contrast to the above, the abundance ratios of the nuclei produced by the
423: $r$-process surrounding Eu are expected to be insensitive to these parameters,
424: because the nucleosynthesis paths in this mass range (i.e., $A \sim$ 150) are
425: quite similar in the $r$-process models which predict the production of
426: actinide nuclei, even though the abundances of the actinides show a
427: significantly large dispersion \citep[e.g., ][]{otsuki02,wanajo02}. This
428: prediction is supported by the similarity in the abundance patterns of the
429: elements with $56 \leq Z \lesssim 70$ found in extremely metal-poor stars, as
430: mentioned in \S 1. The result of the present study, that similar
431: $^{151}$Eu/$^{153}$Eu ratios are found even in the
432: $r$-process-element-enhanced, extremely metal-poor stars with quite different
433: Th/Eu ratios, is naturally explained by the above theoretical expectation. In
434: conclusion, our analysis of the isotope shifts appearing in a few Eu
435: absorption lines shows that the Eu isotope fractions in very metal-poor,
436: r-process-enhanced stars exhibit the value expected from standard
437: nucleosynthesis models, and that they are independent of the global abundance
438: patterns from light to heavy r-process elements.
439:
440: Analyses of isotope fractions for other neutron-capture elements will
441: provide quite strong constraints on modeling of the {\it r}-process
442: nucleosynthesis. Ba and Pb are known to show rather large isotope
443: shifts in their absorption lines. Though the analysis for these
444: isotopes will prove more difficult than that of Eu, their
445: isotope ratios should be sensitive to related nuclear reactions.
446: Eu isotopes, which are (by comparison) rather easily measured from
447: high-resolution spectra of suitable quality, are not sensitive to these
448: processes. We would like to note, however, that small differences might be
449: expected if detailed nucleosynthesis processes are included. For instance,
450: processes that occur after freeze-out of the neutron-capture elements are
451: expected to affect the fine-structure of the abundance patterns of individual
452: nuclei \citep[e.g., ][]{qian97}. These effects will appear more clearly through
453: examination of abundances at the {\it isotopic} level, though they are likely
454: to be smoothed out at the {\it elemental} level. The typical error in modern
455: determinations of elemental abundances is 0.1--0.2~dex ($\sim$25--50\%). It
456: thus follows that, unavoidably, some dispersion ($\lesssim$ 25\%) {\it may}
457: exist even in the derived abundance patterns of the neutron-capture elements
458: with $56\leq Z
459: \lesssim 70$ in $r$-process-element-enhanced, metal-poor stars. We stress that
460: derived isotope ratios of $r$-process elements, such as Eu, are almost
461: completely free from these uncertainties, even though they are more difficult
462: to determine and higher quality spectra are required. The difference in the
463: mean $^{151}$Eu fractions between the stars CS~22892-052 and CS~31082-001 is
464: $\Delta$fr($^{151}$Eu)=0.07, smaller than the uncertainty of the analysis for
465: individual lines. However, the fact that the derived $^{151}$Eu fractions
466: obtained from the three Eu lines used in the analysis of CS~31082-001 are all
467: lower than those obtained for CS~22892-052 may perhaps indicate a small
468: difference of the $^{151}$Eu fractions between the two stars. Further
469: observational study of the Eu isotopes, based on higher quality spectra, for
470: these, and other $r$-process-enhanced, extremely metal-poor stars, is strongly
471: desired.
472:
473: \acknowledgments
474:
475: W.A. and S.H. are grateful for valuable discussions with
476: Drs. T. Kajino, K. Otsuki, S. Wanajo, K. Sumiyoshi, and M. Terasawa on
477: the modeling of {\it r}-process nucleosynthesis and interpretation of
478: our observational results. T.C.B acknowledges partial support of this
479: work from grants AST 00-98508 and AST 00-98549, awarded by the
480: U.S. National Science Foundation.
481:
482: \begin{thebibliography}{}
483: \bibitem[Anders \& Grevesse (1989)]{anders89} Anders, E. \& Grevesse, N. 1989, Geochem. Cosmochim. Acta, 53, 197
484: \bibitem[Aoki et al. (2002)]{aoki02} Aoki, W., et al. 2002, PASJ, 54, 427
485: \bibitem[Arlandini et al. (1999)]{arlandini99} Arlandini, C., K\"{a}ppeler, F., Wisshak, K., Gallino, R., Lugaro, M., Busso, M. \& Straniero, O. 1999, \apj, 525, 886
486: \bibitem[Cayrel et al. (2001)]{cayrel01} Cayrel, R. et al. 2001, Nature, 409, 691
487: \bibitem[Hill et al. (2002)]{hill02} Hill, V. et al. 2002, A\&A, 387, 560
488: \bibitem[Hoffman, Woosley \& Qian (1997)]{hoffman97} Hoffman, R.D., Woosley, S.E. \& Qian, Y.-Z. 1997, \apj, 482, 951
489: \bibitem[Kurucz (1993)]{kurucz93} Kurucz, R. L., 1993, CD-ROM 13, ATLAS9 Stellar Atmospheres Programs and 2km/s Grid (Cambridge: Smithsonian Astrophys. Obs.)
490: \bibitem[Lawler et al. (2001)]{lawler01} Lawler, J. E., Wickliffe, M. E., Den Hartog, E. A. \& Sneden, C. 2001, \apj, 563, 1075
491: \bibitem[Noguchi et al. (2002)]{noguchi02} Noguchi, K., Aoki, W., Kawanomoto, S., et al. 2002, PASJ, in press
492: \bibitem[Otsuki et al. (2000)]{otsuki00} Otsuki, K., Tagishi, H., Kajino, T. \& Wanajo, S. 2000, 533, 424
493: \bibitem[Otsuki et al. (2002)]{otsuki02} Otsuki, K., Mathews, G.J. \& Kajino, T. 2002, ApJ, submitted
494: \bibitem[Preston \& Sneden (2001)]{preston01} Preston G.W. \& Sneden C. 2001, AJ, 122, 1545
495: \bibitem[Qian \& Wasserburg (2001)]{qian01}Qian, Y. -Z \& Wasserburg, G. J. 2001, \apj, 552, L55
496: \bibitem[Qian et al. (1997)]{qian97} Qian, Y.-Z., Haxton, W.C., Langanke, K., \& Vogel, P. 1997, Phys.Rev C 55, 1532
497: \bibitem[Schatz et al. (2002)]{schatz02} Schatz, H., Toenjes, R., Pfeiffer, B., Beers, T.C., Cowan, J.J., Hill, V. \& Kratz, K.-L. 2002, \apj, 579, 626
498: \bibitem[Smith et al. (2001)]{smith01} Smith, V. V., Vargas-Ferro, O., Lambert, D. L. \& Olgin, J. G. 2001, \aj, 121, 453
499: \bibitem[Sneden et al. (2002)]{sneden02} Sneden, C., Cowan, J. J., Lawler, J. E., Burles, S., Beers, T. C. \& Fuller, G. M. 2002, \apj, 566, L25
500: \bibitem[Sneden et al. (2000)]{sneden00} Sneden, C., Cowan, J. J., Ivans, I. I., Fuller, G. M., Burles, S., Beers, T. C. \& Lawler, J. E. 2000, \apj, 533, L139
501: \bibitem[Sneden et al. (1996)]{sneden96} Sneden, C., McWilliam, A.,
502: Preston, G.W., Cowan, J.J., Burris, D. \& Armosky, B.J., 1996, ApJ 467, 819
503: \bibitem[Wanajo et al. (2002)]{wanajo02} Wanajo, S., Itoh, N., Ishimaru, Y., Nozawa, S. \& Beers, T.C. 2002, \apj, 577, 853
504: \bibitem[Wasserburg \& Qian (2001)]{wasserburg01}Wasserburg, G. J. \& Qian, Y. -Z. 2001, \apj, 529, L21
505: \bibitem[Westin et al. (2000)]{westin00} Westin, J., Sneden, C., Gustafsson,
506: B., \& Cowan, J.J. 2000, \apj, 530, 783
507: \end{thebibliography}
508:
509: \clearpage
510:
511: \begin{figure}
512: \caption[]{Comparison of the observed spectra (dots) and synthetic
513: spectra (lines) for the \ion{Eu}{2} 4129~{\AA} line. The name of the
514: object and the adopted fr($^{151}$Eu) value are given in each
515: panel. The solid line shows the synthetic spectra for the adopted
516: fr($^{151}$Eu); the dotted and dashed lines show those for ratios
517: which are smaller and larger by 0.15 in fr($^{151}$Eu),
518: respectively. The wavelengths and relative strength of the hyperfine
519: components for $^{151}$Eu and $^{153}$Eu are shown in the top panel.}
520: \label{fig:4129A}
521: \end{figure}
522:
523: \begin{figure}
524: \caption[]{The same as Figure~\ref{fig:4129A}, but for the \ion{Eu}{2}
525: 4435~{\AA} line.}
526: \label{fig:4435A}
527: \end{figure}
528:
529: \begin{figure}
530: \caption[]{Reduced $\chi^{2}_{\rm r}$ as a function of the fraction of
531: $^{151}$Eu determined from the 4129~{\AA} line. The asterisks, open
532: circles, and filled circles indicate the derived results for
533: HD~115444, CS~22892--052, and CS~31082--001, respectively. The values
534: of HD~115444 and of CS~22892-052 are multiplied by 2 and 1/2,
535: respectively, for clarity.}
536: \label{fig:fit}
537: \end{figure}
538:
539: \begin{figure}
540:
541: \caption[]{Results of the $^{151}$Eu fractions and uncertainties
542: ($\sigma_{\rm total}$). The results obtained by consideration of
543: \ion{Eu}{2} 3819~{\AA}, 4129~{\AA}, 4205~{\AA}, and 4435~{\AA} are
544: shown by squares, diamonds, open circles, and filled circles,
545: respectively. The dotted line is the $^{151}$Eu fraction in
546: solar-system material (0.478). Note that the isotopic fractions
547: derived for all three lines considered for CS~31082-001 are lower than
548: those obtained for CS~22892-052, although the error bars overlap.}
549:
550: \label{fig:results}
551: \end{figure}
552:
553: \clearpage
554: \begin{deluxetable}{lcccccccl}
555: \tablewidth{0pt}
556: \tablecaption{PROGRAM STARS AND OBSERVATIONS \label{tab:obs}}
557: \startdata
558: \tableline
559: \tableline
560: Star & Wavelength range (\AA) & Exp.$^{\rm a}$ & Obs. Date (JD) & Radial velocity (kms$^{-1}$) \\
561: \tableline
562: CS~31082-001 & 3080-3900, 3960-4780 & 200 (5) & 26 Oct., 2001 (2452208.9) & $+138.92\pm0.28$ \\
563: & 3540-4350, 4440-5250 & 20 (1) & 30 July, 2001 (2452121.1) & $+138.91\pm0.30$ \\
564: CS~22892-052 & 3540-4350, 4440-5250 & 120 (3) & 23 July, 2001 (2452114.0) & $+13.16\pm0.36$ \\
565: HD115444 & 3540-4350, 4440-5250 & 30 (2) & 5 July, 2000 (2451730.8) & $-27.12\pm0.34$ \\
566: & 3080-3900, 3960-4780 & 30 (2) & 16 Apr., 2001 (2452016.0) & $-27.11\pm0.28$ \\
567: \tableline
568: \enddata
569: ~ \\
570:
571: $^{\rm a}$ Total exposure time (minute) and number of exposures.
572:
573: \end{deluxetable}
574:
575:
576: \begin{deluxetable}{llccccc}
577: \tablewidth{0pt}
578: \tablecaption{RESULTS \label{tab:results}}
579: \startdata
580: \tableline
581: \tableline
582: star & line & fr($^{151}$Eu) & $log \epsilon$(Eu) & $\chi_{\rm r}^{2}$ & $\sigma_{\rm fit}$ & $\sigma_{\rm total}$ \\
583: \tableline
584: HD~115444 & 3819{\AA} & 0.42 & $-1.87$ & 0.59 & 0.07 & 0.11 \\
585: & 4129{\AA} & 0.48 & $-1.80$ & 0.49 & 0.09 & 0.12 \\
586: & 4205{\AA} & 0.49 & $-1.76$ & 0.10 & 0.04 & 0.08 \\
587: \tableline
588: CS~22892-052 & 4129{\AA} & 0.50 & $-1.01$ & 7.19 & 0.11 & 0.15 \\
589: & 4205{\AA} & 0.47 & $-0.95$ & 1.50 & 0.07 & 0.09 \\
590: & 4435{\AA} & 0.57 & $-0.90$ & 3.07 & 0.12 & 0.13 \\
591: \tableline
592: CS~31082-001 & 4129{\AA} & 0.43 & $-0.62$ & 2.22 & 0.06 & 0.10 \\
593: & 4205{\AA} & 0.36 & $-0.62$ & 2.50 & 0.05 & 0.09 \\
594: & 4435{\AA} & 0.52 & $-0.62$ & 2.51 & 0.10 & 0.11 \\
595: \tableline
596: \enddata
597:
598: \end{deluxetable}
599:
600: \clearpage
601: \plotone{f1.eps}
602: \clearpage
603: \plotone{f2.eps}
604: \clearpage
605: \plotone{f3.eps}
606: \clearpage
607: \plotone{f4.eps}
608:
609: \end{document}
610:
611:
612: