Tải bản đầy đủ - 0trang
VIII. Stellar Abundance, Galactic Chemical Evolution and Nucleo-Cosmochronology
This page intentionally left blank
MEASUREMENTS OF LI AND EU ISOTOPW
ABUNDANCES IN METAL-DEFICIENT STARS
National Astronomical Observatory of Japan
2-21-1, Osawa, Mitaku, Tokyo, 181-8588 Japan
Measurements of isotope abundances give quite strong constraints on nucleosynthesis models. High resolution spectrographs recently mounted on large telescopes
enable us t o measure isotope abundances for several elements in metal-deficient
stars. We report on the measurements of isotope abundances for Europium and
Lithium using the 8.2m Subaru Telescope.
1. Measurements of isotope abundances in stellar
Measurements of chemical abundances in st.ellar photospheres have been
giving valuable information t,o understand the nucleosynthesis processes in
the universe. The measuremenk are made by the detailed analysis of stellar spectra obtained with high resolution spectrograph using model stellar
photospheres. The analyses are generally made for ekmental abundances
rather than isotopic ones? because the spectral lines are basically determined by the chemical nature rather than the mass of the nuclei.
For this reason! the prediction by nucleosynthesis models are usually
reduced to the elemental abundances to compare with observed chemical
abundances measured for stars. An exception is the isotope composit.ion
in solar systeni material, most of which can be measured by the analysis
of meteorites. Measurenients of isotope abnundacnes for stars give quite
strong constraints on nucleosynthesis models.
Though t.he measurements of isotope abundances in stellar photospheres
are difficult, there are some opportunities (Table 1). One is to use niolecular
spectra, which are sonietinies significantly affect,edby the difference of mass
of the nuclei'. For instance, carbon isotope ratios (12C/13C) has been measured for a number of cool stars in field stars in our Galaxy as well as those
in clusters from the analyses of spectra of carbon-bearing molecules (e.g.,
Table 1. Measurements of isotope fractions in stellar photospheres
Li I6708 A
CH, CN, CO, C2
MgH 5130 8,
Ba 114554 8,
1 5 1 E ~ / 1 5 3 E ~ E u 114205 A etc.
' ~ i / ~ ~ i
1 6 0 / 1 7 0/18 0
CH, CN, CO). The carbon isotope ratio is a quite useful indicator of the
evolutionary stages of giant stars. Recent analysis of MgH lines have made
a rapid progress in the understanding of magnesium isotope abundances
(24Mg: 251\,Ig,and 26hi1g) in field and globular cluster stars (e.g., Yong et
al.'). These results have impact on the interpretation of the wavelength
shifts of Mg resonance lines found in some quasar absorption ~ y s t e r n ~ ' ~ ,
which is sonietimes interpreted as an evidence of the time variation of the
fine structure constant. Oxygen and silicon isotope abundances have been
measured for cool stars using molecular spectra in near infrared ranges (see
references in Table 1).
The atomic spectra of light elements are also affected by the difference of
the mass of the nuclei. The isotope shifts of hydrogen lines are well known?
and are measured in niany astronomical objects including quasar absorption
system (e.g.! Kirkman et a1.20). The third lightest, element lithium also
shows a rather large isotope shifts in the resonance line at 6708 A. The
measurements of Li isotope abundances are discussed in section 3.
The other possibility is to make use of the hyperfine splitting found in
heavy elements. The behavior of hyperfine splitting is different between
isotopes of an element in general, and the difference sometimes enables us
to estimate the isotope ratios by the detailed profile analysis of absorption
lines in stellar specbra. Magain & Zhao14 analyzed the absorption profile
of the Ba I1 4554 A resonance line to estimate the isotope comp0nent.s in
the metal-deficient ([F~/H]N-2.5) sOar HD 140283. Since Ba has 7 stable
isotopes, it is very difficult to determine the isotope fractions. However,
production of 138Ba dominates in the s-process, while isotopes with odd
mass number (L35Baand 137Ba) as well as 138Ba are yielded by r-process.
For this reason, the ratio of the contribut.ion by r- and s-processes to the
Ba in a star can be estimated from the analysis of the Ba line profile.
Their result suggested t.hat a significant part of the Ba in HD140283 is
produced by the s-process. This is a surprising result because only a small
contribut.ion of s-process is expected for stars with such low metallicity.
The Ba isotope fract'ions in this object is still in controversy (e.g., Lanibert
Recently, analyses of Eu lines were made for some metal-deficient,stars.
Our recent studies are reported in section 2.
Iv1easurement.sof the isotope abundances require very high quality spectra, because the analysis is usually based on t,he detailed profile fitting of
spectra calculated using niodel stellar phot.osphere to observed ones. The
spectral resolution of R 100,000 (3km s-') or higher is desirable to fully
resolve the stellar spectra which are intrinsically broadened by thernial motion and turbu1enc.e in the photosphere with several kni s-'. In order to
obtain high signal-to-noise spectra with such high spectral resolution, large
telescopes (e.g., the ESO Very Large Telescopes, the Subaru Telescope)
and high resolution spect,rographs are required. In this paper, we report
the recent isotope measurenients based on the high resolution spect,ra obtained with the High Dispersion Spectrograph (HDS2') of t.he 8.2m Subaru
2. Eu isotopes in very metal-deficient stars
Eu has two isotopes with odd niass number (15'Eu and 153E11). The effect
of hyperfine splitting is significant in both isotopes, but the degree of the
splitting is quite different. Hence, this element is an ideal case to nieasure isotope ratios. An accurate line list including hyperfine splitting was
recently provided by Lawler et a1.22,and, using this line list, Eu isotope
ratios were measured for three metal-deficient stars by Sneden et al.I7.
2.1. Eu isotopes produced by r-process
Figure 1 shows observed spectra of the Eu 11 4205 A line for three starslg.
HD 6268 (top panel) is known to have moderate enhancements of neutroncapture elements whose elemental abundance pattern is well explained by
the r-process nucleo~ynthesis~~.
The wavelengths and relative strength of
the hyperfine components for I5'Eu and 1 5 3 E are
~ shown in the top panel.
Since the hyperfine splitting of 151Eu is much more significant than that
of the other isotope, in particular in the bluer part of the absorption profile, the isotope ratios can be estimated from the profile analysis. The
dotted. solid, and dashed lines show the synthetic spectra calculated using
model photospheres for three different isotope fractions (fraction of 151Eu
(fr(15'Eu)) is set to be 0.38, 0.48, and 0.58). The x 2 fitting of these synthetic spectra to observed one derives fr(151Eu) to be 0.48 f 0.04. The
Figure 1. Comparison of the observed spectra (dots) and synthetic ones (lines) for the
Eu I1 4205 A line. The name of the object and the adopted fr(15'Eu) value are presented
in each panel. The solid line shows the synthetic spectra for the adopted fr(15'Eu); the
dotted and dashed lines show those for ratios which are smaller and larger by 0.10 in
fr(151Eu),respectively. The dot-dashed lines show the synthetic spectra for no Eu. The
wavelengths and relative strength of the hyperfine components for 151Eu and 1 5 3 E are
shown in the top panel.
uncertainty includes bhe 30 confidence level of the fitting and errors caused
by uncertainties of continuum level, line position, and Eu total abundance.
The result perfectly agrees with that of solar-system material24. Since
95% of the Eu in solar-system material is expected to originate from t.he
r - p r o c e ~,st,his
~ ~ ratio well represents that. of the r-process component in
the solar system. Similar analysis were also made for other three r-processelement-enhanced stars'' including CS 31082-001 in which uranium was
detected by Cayrel et a1.26 (see also Honda 27). These results and those
by Sneden et al.I7 show that the Eu isotope ratios in st,ars with excesses of
r-process elements are consistent u7it.h that in solar system material within
the errors. The remarkable agreement of elemental abundance patterns of
these objects with that of solar-system r-process component. was found by
previous studies (see references of H ~ n d a ~The
~ ) . analysis of Eu isot.opes
confirni for the first time this agreement in isotope level.
2.2. Eu isotopes produced by s-process: a new probe of
The other two stars shown in Figure 1 (LP 625-44 and CS 31062-050) are
very nietal-deficient ([Fe/H]N -2.5) , but have large excesses of s-process
e l e n ~ e n t s ~ ' ; These
~ ~ . excesses are explained by the mass t.ransfer across
binary syst.ems containing AGB stars which have already evolved to white
dwarfs. Indeed, variat.ionof radial velocity has been confirnied for these two
stars, indicating they have unseen companions which are presumably white
dwarEs. Even though about, 95 % of Eu in solar system material is believed
to originate from r-process, and this element is sonietiiiies refereed to as 'rprocess element', the majority of the Eu in these two nietal-deficient stars
are estimated to originate from the s-process from the abundance patterns
of elements between Ba and Eu 28,29.
The I5'Eu fractions derived for LP 625-44 and CS 31062450 are 0.60
and 0.55, respectively, with uncert,ainties of about *0.05. This is the first.
estimate of the Eu isot.opefractions produced by s-process, because t.hat can
not be estimated from solar system material in which r-process contribut.ion
is donlinant. These values are higher than found in solar-system material
(fr(151Eu)=0.478). However, they agree well with the predictions of recent
s-process models by Arlandini et al.25>who deduced fr(151Eu)=0.541 and
0.585 from their best-fit stellar and classical models, respectively.
The Eu isotope fractions are quite useful as a probe of 151Smbranching
of s-process nucleosynthesis (Figure 2). For 151Sm, whose half-life is about
90 years: the $decay rate is strongly dependent on temperature, while t.he
neutron capture rate is not. This niakes the 151Sm branching an excellent,
Figure 2. The s-process reaction path around the ls1Srn branching. Unstabb nuclei
are shown by boxes with dotted lines.
Previously; this branching has been analyzed using the
152Gdand 154Gdisotope ratios in solar-syst,emmat.erial, which are believed
to be significantly affected by this branching (e.g., Beer et al.31; Wisshak et
al.30). One dificu1t.yin this approach is that these Gd isotopes are affected
by a small amount of contamination from the pprocess: t.hough Obey are
shielded from the r-process.
We have made an analysis using t.he thermally pulsed s-process models, using updated reaction rates (see Aoki et al.lg). Figure 3 shows the
fr(151Eu) values calculated by our model. They are plotted as a function of
neutron density (Nn)
for four teniperatures (kT = 10, 15, 20, and 30 keV).
Also shown for coniparison by t,he hatched area is the fr(151Eu) range deduced for the s-process-element-enhanced st.ar LP 625-44 (the upper panel)
and CS 31062-050 (the lower panel). As can be seen in this figure, the
fr(151Eu) value is rather sensitive to t,he ambient temperature and neutron
density during the s-process.
The fr(151Eu) value is maximized in the range of neutxon density from
ATn = 5x lo7 to lo9 cmP3. For Arn > lo7 c ~ i i - ~the
? branching factor
at 151Sn1is higher than 0.9; and the nuclear flow bypasses 15'Eu. In this
Log N, (cm-9)
Figure 3. The fr(151Eu) values calculated for four temperatures (squares: kT = 30 keV,
triangles: 20 keV, circles: 15 keV, and asterisks: 10 keV) as a function of neutron density
(Nn). The upper and lower panels compare the calculations with the observed results
for LP 625-44 and CS 31062-050, respectively.
case; the fr(151Eu) value decreases wit.h increasing neutron density by the
effect of branching at 153Sm. In contrast.; in the range of low neutron
density (IV, < lo7 c111-~)~
the nuclear flow passes 151Euand 15’Gd, and the
elect.ron-capture on 153Gd is coniparable with, or faster than, the neutron
capture, which contributes to bhe production of 153Eu. This results in the
decrease of fr(151Eu)with decreasing neutron density in the low (Nn 5 lo7
c111r3) range (see Aoki et a1.l’ for more details).
The comparison of the calculations with the observational results indicates that s-processes wibh high neutron density (logN, 2 9) and low
temperat.ure (kT 5 20 keV), or those with quite low neut.ron density
(AT, < iO7c1r3), are not allowed for LP 625-44 (the upper panel of Figure
3). For CS 31062-050 (the lower panel), the process with high t.einperature
(kT 2 30 keV) and niediuni neutron density (10’ 5 ATn5 l O ’ ~ n - ~ is
allowed. These are new constraints on s-process nucleosynthesis provided
by the Eu isotope analysis.
Recent niodels of AGB stars show that t.he abundance patterns of nuclei
in branchings are affected by the s-process both during the t,hermal pulses
and between pulses (e.g., Arlandini et al.25). The reaction which provides
neut.ron in the former phase is 22Ne(a,n)25Mg, which produces a neutron
density as high as 108-101* c111r3. In the int,erpulse phase? I3C(ck,n)160 is
assumed to be t.he neutron source reaction, which leads to a lower neutron
lo7 C I ~ I - ~ ) .
Though the uncertaint,ies in measurements are still large, the comparison of observational results with model calculations in Figure 3 suggests
that the contribution of s-process with low temperature between the thermal pulses is large in CS 31062-050, while the process during thermal pulse
plays an iniport,ant role in LP 625-44. This suggests the s-processes contributed to these two stars occurred in quite different conditions. It should
be noted that their Pb/Ba abundance ratios are significantly ~lifferent~~>~’.
The low Pb/Ba ratio of LP 625-44 cannot be explained by standard AGB
models with low rnetallicity (e.g. Busso et a ~ ~and
a possible s-process during the t.herina1pulse was proposed by Iwanioto et
al.33. Further studies for Eu isotopes as well as elemental abundances with
higher accuracy are clearly desirable to understand the s-process nucleosynthesis in very metal-deficient AGB stars.
It should also be not.ed that, before deriving a clear conclusion, more
accurate reaction rates for isotopes around t.he branching point are required. Indeed, our first analysis using previous neutron-capt,ure rate of
151Srn which is about 50% higher than that used in the present work reN
sults in lower 151Eu fractions by 0.05-0.08. New experiments to determine
the neutron capture cross section of 151Sm are highly desirable to fix ratios
of 1 5 1 E ~ / 1 5 3 and
E ~ also 152Gd/154Gd.
3. Li isotopes
A number of studies have been made t.o accurately determine lithium abundances in nietal-deficient main-sequence stars to constrain the big-bang nucleosynthesis (see Coc et al.34). There are two stable isotopes of t’hiselement,
(‘Li and 7Li), and 7Li is believed t.o be dominant in most stars (t.he isotope ratio in solar system material is GLi/7Li= 0.08). The plateau found
in the abundance of this element in most metal-deficient stars has been
interpreted as the result of 7Li production in the big bang nucleosynthesis.
However, the value of the 7Li abundance plateau (log6 (7Li)
is by 0.4-0.5 dex lower than that. expected from the standard bin-bang
nucleosynthesis model constrained by the recent nieasureriients of cosmic
microwave background radiation by WMAP and the cosmic D/H ratios
34). An important uncertainty in observational data is a possible depletion
of Li in the stellar photosphere from its original value. Since ‘Li is more
easily destroyed during stellar evolution than 7Li>the depletion of 7Li will
be excluded if 6Li is det.ected in the same object. This is the cosmological
interest to search for ‘Li in metal-deficient main-sequence stars.
In addition, the origin of 6Li in very metal-deficient stars is still unclear:
while the major producer of their 7Li would be the big bang nucleosynthesis. The spallat.ion of heavier nuclei by cosniic-rays provided by supernova
explosions only explains the ‘Li in higher metallicity. Recent,ly; Suzuki
8c 1 n 0 u e ~proposed
shocks produced by t,he formation of the large scale
structure of the Galaxy as a new source of cosniic-ray. The nietallicity
dependence of Li isotope ratios is expecbed to be a key to understand the
origin of ‘Li (see references in Table 1).
‘Li has been first detected by Smith et aL5 for the met.al-deficient
([Fe/H]= -2.2) st,ar HD84937. Subsequent.studies determined the 6Li/7Li
isotope ratios to be about. 0.05 for a few objecbs wit.h similar nietallicity.
We have started an observat,ional program using the Subaru Telescope
to measure 6Li/7Li ratios to ext’end the study to lower nietallicity. As t.he
first sample of this program, we obt.ained a very high quality spectmni
( R = 95,000 and S / N = 1000) of the bright subgiant HD140283, whic.h
has the lowest metallicity ([Fe/H]= -2.5) among the objects for which Li
isotopes have been t,ried to measure to date. The detailed profile analysis
for the Li I 6708
line showed no evidence of 'Li in this object, and
determined a quite low upper liniit on the Li isotope ratio (6Li/7Li< 0.026).
The low upper-limit of 6Li/7Li may indicate a decrease of 6Li at the lowest
metallicity. However, unfortunately> this object may be sufficiently cool
t h a t 6Li was affected by internal processes of the star itself. even though
7Li and Be abundance5 show no depletion compared with other stars with
s i n d a r inetallicity.
In order to derive a clear conclusion, we also have obtained high quality spectra for several main-sequence stars with lower nietallicity. hIeasurenients of Li isotopes for larger sample of metal-deficient main-sequence
stars are ongoing using the ESO VLT. The behavior of Li isotopes in lowest
inetallicity will be soon revealed by these intensive studies.
Figure 3 was provided by N. Iwanioto who carried out the s-process model
1. Herzberg, G., 'Molecular Spectra and Modecular Structure. I.: Spectra of
Diatomic Molecules'; Van Nostrand Reinhold Company, New York (1950)
2. Yong, D., Lambert, D. L.? & Ivans, I. I., ApJ, 599, 1357 (2003)
3. Ashenfelter, T. P.? Mathews, G. J.: & Olive, K. A., ApJL? in press
4. Otsuki, K. et al. 2004, this volume
5. Smith, V. V., Lambert, D. L., & Nissen, P.; ApJ; 408, 262 (1993)
6. Smith, V. V., Lambert, D. L., & Nissen, P.: ApJ; 506, 405 (1998)
7. Nissen: P. E.? Lambert, D. L., Primas, F.; M Smith, V. V., AMA, 348, 211
8. Cayrel, R., Spite, &I.>Spite, F., Vangioni-Flam, E., CassB, M., & Audouze, J.;
A&A, 343, 923 (1999)
9. Sneden, C., Pilachowski, C. A.? & Vandenberg, D. A., ApJ, 311, 826 (1986)
10. Lambert: D. L., Gustafsson, B.? Eriksson! K., Hinkle? K. H., ApJS, 62, 373
11. Aoki?W.? & Tsuji, T.?A&A, 317, 845 (1997)
12. Harris, M. J., Lambert, D. L. & Smith, V. V., ApJ, 292, 620 (1985)
13. McWilliam, A.: Lambert, D. L.? MNRAS, 230, 573 (1988)
14. Magain, P. & Zhao, G.; A&A, 268, L27 (1993)
15. Tsuji, T., Ohnaka, K.; Hinkle: K. H., & Ridgeway, S. T., A&A, 289, 469
16. Lambert, D. L., Allende Prieto, C., &INUS, 335, 325 (2002)
17. Sneden, C.? Cowan, J. J.: Lawler, J. E., Burles, S.: Beers, T. C. & Fuller, G.
M., ApJt 566, L25 (2002)