1997MNRAS.288..108P
Mon. Not. R. Astron. Soc. 288, 108-116 (1997)
Galactic chemical evolution of primary elements in the solar
neighbourhood - II. Elements affected by the s-process
B. E. J. Pagel1* and G. Tautvaisiene*
'NORDITA, Blegdamsvej 17, Dk-2100 Copenhagen 0, Denmark
2Institute of Theoretical Physics and Astronomy, Gostauto 12, Vilnius 2600, Lithuania
Accepted 1997 January 23. Received 1997 January 23; in original form 1996 May 13
ABSTRACT
We apply the analytical model developed previously to the Galactic chemical
evolution of elements that in the Solar system come primarily from the s-process,
but at low metallicities are believed to come primarily from the r-process. We find
that a 'primary' production model is indicated for the s-process, but with two
separate time delays that are treated in the 'delayed production' approximation and
tum out to be of the order of 37 Myr and 2.7 Gyr, corresponding to progenitor
masses of about 8 and 1.5 M0 respectively. While the latter component fits in well
with current models of s-process synthesis of material in the Solar system, the early
component raises some difficulties because asymptotic giant branch (AGB) stars of
high mass and low metallicity are observed to generate a high ratio of heavy to light
s-process elements, whereas the field stars presumed to represent the effects of
Galactic chemical evolution still have Solar system-like heavy to light ratios.
Key words: nuclear reactions, nucleosynthesis, abundances - stars: abundances stars: Population II - solar neighbourhood.
1 INTRODUCTION
In a previous paper (Pagel & Tautvaisiene 1995, hereafter
Paper I), we developed an analytical model for chemical
evolution in the Galactic disc (including metal-deficient
stars, some of which are also believed to belong to the thick
disc) and applied it to the development of a simple understanding of the trends of oxygen, a-particle and r-process
elements relative to iron, finding that the assumption of
constant yields with a single time delay for iron from SNIa,
and a shorter single time delay for r-process elements (Eu
and Th), gave a satisfactory fit to the abundances deduced
from observation. A time delay for iron from SNIa very
similar to the one estimated by us has since been found on
the basis of numerical calculations of Galactic chemical
evolution (GeE) by Yoshii, Tsujimoto & Nomoto (1996).
In the present paper we address the problem of the chemical evolution of elements affected by the s-process, which is
more difficult because their absolute and relative yields
could in principle be a function of metallicity driven by the
abundances of neutron sources and poisons and of iron
*E-mail: [email protected](BEIP);[email protected] (GT)
seeds, although we shall argue that no direct effects of
metallicity are apparent in stars that display the effects of
'normal' GeE.
Abundances of elements above the iron group in the
Solar system are thought to have contributions (in differing
proportions) from two main processes, the r- and s-processes of neutron capture, with in some cases a small contribution from the p-process (Burbidge et al. 1957). The
s-process has been further subdivided into two distinct components: the main s-process, which is fitted by an exponential distribution of exposures and dominates the s-process
contribution to Rb and heavier elements; and the weak sprocess, which may be better fitted by a single exposure and
dominates the s-process contribution to lighter elements
such as eu and Zn (Kappeler, Beer & Wisshak 1989). The
weak process is generally believed to result from the
22Ne(a,n)25Mg reaction in massive stars and to be ineffective at low metallicities ([Fe/H] < -1) because of the shortage of both 22Ne (resulting from the original eNO content)
and iron seeds; this leads to 'secondary' production at some
stage, aggravated at the lowest metallicities by the abundance of various neutron 'poisons' in the He-burning zone
(Prantzos, Hashimoto & Nomoto 1990; Raiteri, Gallino &
©1997 RAS
© Royal Astronomical Society • Provided by the NASA Astrophysics Data System
1997MNRAS.288..108P
Galactic chemical evolution of primary elements - II
Busso 1992; Baraffe & Takahashi 1993). The elements
mainly affected are largely produced in explosive synthesis
in supernovae, however, and will not be considered in this
paper.
The main s-process is nicely explained as being a consequence of thermal pulses in the AGB phase of a low-mass
( ~ 1.5 M o ), moderately low-metallicity (Z ~ Zo/3) star, in
which a small 13C pocket is assumed to be ingested by the
rising intershell convective zone generated in each pulse
and provides neutrons by the 13C (ex, n) 160 reaction, and this
is followed by a brief neutron burst from 22Ne at the peak
temperature reached by the convection zone during the
pulse which takes care of a few minor features in the abundance distribution (Hollowell & Then 1989; Kappeler et al.
1990). In this model any eventual dependence of the yield
on metallicity is hard to predict, since it depends on the
mass of the 13C pocket, which has to be assumed arbitrarily.
Observed stellar abundances in the disc, at least, show little
variation in the s-process abundances relative to iron, which
is probably a good standard to judge by (as opposed to
oxygen for instance) because the low mass of the s-process
progenitors will lead to a similar time delay. Clayton (1988)
argued that s-process production from the 13C source could
well be 'primary' because the neutron density (and hence
the mean exposure or 'flue nee' 'to) is inversely proportional
to the density of targets, i.e. in first approximation to Z, thus
compensating the reduced abundance of iron seeds. Malaney & Fowler (1989) tested this effect with numerical
models and found that it actually resulted in overkill, i.e. it
led to an s-process yield that increases with decreasing
metallicity (again in conflict with a constant s/Fe ratio) and
suggested that a constant yield might be restored if a larger
amount of the neutron poison 14N was produced in the 13C
pocket at lower metallicities; this itself depends on the arbitrary amount of protons ingested by semiconvection or by
some other mixing process into the 12C-rich intershell
region. Be that as it may, the increase in 'to towards low
metallicities implies an increase in the abundance of heavy
main s-process products (Ba, La) relative to lighter ones (Sr,
Y, Zr), which is indeed observed in low-metallicity giant CH
stars (Wallerstein & Greenstein 1964; Luck & Bond 1991;
Vanture 1992; Kipper & J0rgensen 1994; McWilliam et al.
1995) and in massive AGB stars (7 Mo) in the Small Magellanic Cloud (SMC) (Plez, Smith & Lambert 1993), although
in these cases a single exposure seems to be indicated,
rather than an exponential distribution. On the other hand,
there is no trace of such an enhancement of heavy s-process
elements relative to lighter ones in Galactic field stars of low
metallicity (see Fig. 3); a partial explanation may be that,
owing to the large scatter that exists in stellar metallicities at
any given time, there is but a poor correlation between the
metallicities of stars observed and those of their s-process
progenitors, so that any abundance effects on the yields
tend to be washed out. In any case, we shall treat the sprocess as 'primary' and discuss it in the same way as we did
the undisputedly primary elements studied in Paper 1.
2 R- AND S-PROCESS CONTRIBUTIONS TO
INDIVIDUAL ELEMENTS
Early work on heavy-element abundances in metal-deficient
stars, going back to investigations by Pagel (1965) and Spite
109
& Spite (1978) and reviewed by Wheeler, Sneden & Truran
(1989), showed that Sr, Y, Zr and especially Ba begin to
decline in abundance relative to Fe as [Fe/H] decreases
below - 2.3 or so, and Truran (1981) postulated that, at low
metallicities, only the r-process was effective in producing
them, a suggestion that has since been supported by many
element:element ratios measured in metal-deficient stars,
e.g. Sneden & Parthasarathy (1983), Gilroy et al. (1988),
Lambert (1989). However, an s-process contribution is
noticeable in many metal-deficient stars, even when [Fe/
H] < - 2.5; an interesting case is that of the subgiant HD
140273 in which Magain (1995) has argued on the basis of
absence of hyperfine structure broadening of the Ba II line
..1.4554 that mainly even-A isotopes (134, 136 and 138),
formed by the s-process, are present, although Ba is overdeficient relative to Eu (Gilroy et al. 1988), a fact that had
previously been attributed to an r-process origin. There are
two factors that may contribute to the solution of this paradox. Possibly the BalEu ratio was underestimated by Gilroy
et al., leaving room for a dominant s-process contribution,
and possibly the hyperfine structure broadening of the odd
barium isotopes is diluted by the existence of a significant rprocess contribution to 138Ba; Kappeler et al. (1989) give
only an upper limit for the latter, but the existence of a
smooth relation between r-process abundance and mass
number (fig. 6.7 of Kappeler et al. 1989) does imply that this
contribution is at least not negligible.
Taking the results of Gilroy et al. (1988) at face value,
Pagel (1989) attempted to explain the chemical evolution of
Ba by postulating an instantaneous r-process component
corresponding to the r-process contribution of 12 per cent
to the Solar system abundance (Cameron 1982) combined
with a single time-delayed contribution from a primary sprocess, but this model was quite unsatisfactory because it
predicted large positive values of [BalFe] at moderate metal
deficiencies (Pagel 1991). A number of developments since
then make a fresh discussion of interest.
(i) Mathews & Cowan (1990) have pointed out that Eu,
virtually a pure r-process element, itself displays significant
underabundance relative to Fe at the lowest metallicities,
consistent with a small time delay for the r-process (cf.
Mathews, Bazan & Cowan 1992). We have adopted this
idea in Paper 1.
(ii) The discovery by Sneden et al. (1994) of the very
metal-weak ([Fe/H] ~ - 3.1), r-process-rich «[r/Fe]) = 1.7)
star CS 22892 - 052, which acts as a Rosetta stone for the rprocess. A very careful analysis of this star (Sneden et al.
1996) reveals that the r-process pattern there is identical
within the errors to that in the Solar system (after allowing
for Th decay), which is a highly significant result as it provides a firm basis for the analysis of r- and s-process contributions in other stars. The r-process contribution to each
element can now be judged either according to Solar system
data (Cameron 1982; Kappeler et al. 1989) or according to
the deficiency found relative to Eu or Dy in that star, with
reasonably good agreement. Relevant numbers are shown
in Table 1, where we list relative contributions from the
weak and main s-processes and from the r-process according to Solar system data after Kappeler et al. (1989) updated
by Raiteri et al. (1992), the resulting deficiency relative to
europium in a pure r-process and the corresponding defi-
© 1997 RAS, MNRAS 288,108-116
© Royal Astronomical Society • Provided by the NASA Astrophysics Data System
.....
.~
\D
\D
-.:]
@
[IJ
.....
~
'JJ.
.....
.....
0\
I
C<J
0
.....
N
QC
.QC
I
9
~
.....
~
~
[IJ
[IJ
....
n
~
=-
'e
0
'"1
[IJ
>
'JJ.
>
>
.....
Z
~
=-
.....
~
0'"
Q.,
~
Q.,
-<
....
0
'"1
""C
•
~
~
.....
n
'JJ.
0
~
-....
....
n
0
=
9
0
[IJ
.....
'"1
->
r£I
•
·
.•_
. .. .
._ttI
•
•
•
•
......
•
••
••
. ..
0
..<It o
~....
..
;.
•
~.~) ......
,
,
I
,
,
*
0
'
I
-3
*
0
.....................
-4
I
,
I
• •
•
•
I
,
-2
[Fe/H]
"
•
•
I
I
,
I.·
••
,
I
I
I
-1
•
•
I
00
o
• •• ¢
,.
,
,
0
;
:
···········································i,········· ....•.•........... ! ....•........... f:?.t!' ... ~ ..... .
*
o
~
•
.............................................................................•...........•...•.........
.
I!I.
• •
.~
:
C1
°l
-0.5
,
,
L
I
-4
,.,
l
0
I
-3
•
•
' "".
:.:..,••.• : ,.,:..•..
'i
•
•••• .:
L,
L!
!
,
-2
I
[Fe/H]
!
!
r
-1
L " " "
•
L!
,
!
r
,
o
:! ,
i
w •. ,.: .... }
" .• """" ......................................... ~ ... : .... ..
~~ OO-m*~
~
j
! ~:.:~. •. ,.•.:.•.•.•.•. •.•.•.•.~.: .•.•. ,. . :.•.;'.:', .J.:.•.•. •. ','•..:. •.•.',":. ~.· i·. · '1
I ~:t.:. ~: :m •.:.• " ~:;:;\ :,; ,:.' ~~.': j
! _:.: t·.·.•. :.;. . . . •.•.','. . •. .'~'." "',.:'•.. ,. . .•.":'.:. :~.:. ~.
~
L:t~;~m;~;\;;\'j
Magalhiies (1990: filled inverted triangles) and Tautvai§iene (1997: open rhombs). The black five-pointed star shows the results of Sneden et al. (1996) for CS 22892 - 052 and the large empty
square is the prediction for a pure r-process based on the Solar system data in Table 1.
~.
§.
E;<
~
~
~
o
I:l...
$:l
~
~
i :;;t', .: .: :",'.:. :',":'.'~':':",. :. . .'. . :.:. '.-'•.:. . :; .;'.':'.'" .•:', :'. '1 -
Figure 1. Element ratios to europium as a function of [Fe/H], after Gratton & Sneden (1994: filled circles), Fran<;ois, Spite & Spite (1993: circled 'plus' signs), da Silva, de la Reza & de
~
";'
r£I
"<j
~
"h
r£I
";'
~
";'
r£I
"-
~
"h
r£I
";'
~
".a
•
~
~
~
";'
:::0
0
~
~
@
~
"
o"""'
"""'"
1997MNRAS.288..108P
1997MNRAS.288..108P
Galactic chemical evolution of primary elements - II
stars show r-process enhancements (which also lead to relatively high abundances of Sr and Ba), which suggests that
they have been affected by a nearby supernova, while stars
in OJ Centauri curiously show a shortage of r-process and an
excess of s-process (Norris & da Costa 1995; Smith, Cunha
& Lambert 1995). Inspection of the data, with a reasonable
allowance for uncertainties (e.g. figs 14 and 15 of McWilliam et al.) suggests, however, that only a minority (albeit a
substantial one) of field stars display such anomalies, while
the majority form a well-defined pattern in the lower part of
the ([m/Fe], [Fe/HD plane that can be clearly seen in the
data of Gratton & Sneden (1994), which also seem to be .
among the most accurate. We shall therefore take these
results as representing what we choose to regard as 'normal'
GCE, to which it is of interest to apply our model from
Paper I.
ciencies actually found in CS 22892 - 052 by Sneden et al.
(1996).
(iii) Several more abundance determinations have been
made in recent years, especially those by Edvardsson et al.
(1993) for disc stars and by Gratton & Sneden (1994) for
metal-deficient stars, which we emphasize the most in our
fits. The recent determination by Woolf, Tomkin & Lambert (1995) of europium abundances in a subset of the disc
sample of Edvardsson et al. (1993) gives results that are in
excellent agreement with the model we developed in Paper
I, inspired by the work of Mathews et al. (1992). Other
investigations (Ryan, Norris & Bessell 1991; McWilliam et
al. 1995) have demonstrated the existence of a vast amount
of scatter in the heavy-metal: iron ratio at given metallicity
for [Fe/H] < - 2, not all attributable to the Ba II or CH star
phenomenon (cf. Griffin et al. 1982). In particular, several
Table 1. Relative r- and s-process contributions to some elements in the Solar System.
Element
log N
s-weak
s-main
r
[el./Eu]pure r
[el./Eu]cs
37
38
39
40
Rb
Sr
Y
Zr
2.40
2.93
2.22
2.61
.14
.09
.04
.02
.39
.77
.85
.78
.47 ± .10
.14 ± .07
.11 ± .06:
.20 ± .03
-.33
-.85
-.96
-.70
-1.0
-1.25
-.90
56
57
58
59
60
62
63
66
Ba
La
Ce
Pr
Nd
Sm
Eu
Dy
2.21
1.20
1.61
0.71
1.47
0.97
0.54
1.15
.01
.01
.01
.01
.00
.00
.00
.00
.88
.75
.77
.45
.46
.30
.03
.12
.11 ± .02
.25 ± .08
.23 ± .01
.54 ± .09
.53 ± .03
.70 ± .03
.97 ± .06
.88 ± .15
-.96
-.60
-.64
-.27
-.34
-.15
-.79
-.61
-.65
-.31
-.43
-.23
-.06
+.10
22892
Numbers in the first four rows (apart from the error estimates) are taken from Raiteri,
Gallino & Busso (1992). The others are based on data for individual isotopes given by
Kiippeler, Beer & Wisshak (1989). Numbers in the second column are meteoritic abundances relative to log H = 12.
Table 2. Yields (in units of Zo) and time delays.
Alternative model
Basic model
r
w~
.00
~
D..
Mprogen.
Rb
Sr
Y
Zr
Ba
La
Ce
Pr
Nd
Sm
Eu
Dy
~
.055
.013
.010
.022
.0lD
.013
.020
.055
.040
.055
.080
.080
r
5
.007
23 Myr
11M0
.46
.11
.08
.17
.08
.11
.16
.46
.33
.46
.64
.64
~
.011
37 Myr.
~
~
8.5M0
'" 1. 5M 0
.11
.29
.30
.28
.30
.29
.26
.11
.17
.11
.01
.01
111
.80
2.7 Gyr
.11
.29
.30
.28
.30
.29
.26
.00
.040
.010
.008
.015
5
'" 11M0
.009
~ 25 Myr
'" 10M0
.32
.08
.064
.12
.23
.39
.40
.37
.007
'" 23 Myr
.80
2.7Gyr
'" 1. 5M0
~
.14
.23
.24
.22
.11
.17
.11
.01
.01
© 1997 RAS, MNRAS 288, 108-116
© Royal Astronomical Society • Provided by the NASA Astrophysics Data System
.....
'D
@
t::.
>
>
>
.....
~
~
t-.>
I
.....
.....
0.,
0
00
>-'
~QC)
QC)
[IJ
9
[IJ
.....
~
~
Z
a:::
'JJ.
~
.....
~
Y'
'D
-.l
~
[IJ
~
....=-
'e
0
'"1
[IJ
'JJ.
-1.0
0.5
-0.5
0.0
~
I; - - -
..
•
J
,..
" I
..
•
• •
••
•. ..
.o
-4
-3
*
*
•
___ J
,/
[Fe/H]
-2
---.
-1
••
00
o
- ........ =;;.:::-'"
. ........................ ' ............... '~ii~ .~.~..~..~.~.~ .. " , .... ~... '."' ..... -...<?.
.. ..
~~.~.~ ~ ~.~.~.~.~.~~ ~.~.~.~.~ ~.~.~.~.~:~ ~ ~·~·~·~~;J·~··:-·:.:;i
•
o
~
0.0
~ , 'I
,,""
I '
i.
r. · · · · · · · · · · · . ··~. . ·. ·...~ . ~~......~·~·~· . ····.~. . ~.~~ . =:.;' :'
'~I
-0.5'
'I'
]
1
-4
-3
[Fe/H]
-2
-1
o
i -:: t••::::::::::~. ::>:,:·::·:. ::::•;:; •:•~.:.: ,::j
I ~~~t~b::::~:,:,',:,~,:,:,:":,~.:.i.;~::.:~:;.:~:~.:,1
!
0.5
!~1::':'~"';~:''';'''::TI
! _::t,,:~,':,,':::";:;'m::~,
i •~
r'
! ~~t:~:,,:c,'~;"':TI
Figure 2. Element ratios to europium as a function of [Fe/H], fitted with model parameters from Table 2. The continuous line shows the output from our basic model, while the dashed line
shows that from the alternative model for Rb, Sr, Y and Zr. Data sources as in Fig. 1.
';;- -0.5
~
-1.0
r£I
';'
"....
Z
~
r£I
';'
~
tJ
"-
r£I
';'
ES
",D
r£I
';'
=-
.....
~
0'"
Q.,
~
Q.,
-<
....
0
'"1
""C
•
~
~
.....
n
'JJ.
0
~
-....
....9n
0
=
0
[IJ
.....
'"1
->
~
~
0
:::0
@
~
.
§.
~
~
::::
~
o
~
§
~
~
-.
~
;.....
?::l
N
f-'
f-'
1997MNRAS.288..108P
1997MNRAS.288..108P
Galactic chemical evolution of primary elements - II
r-process contribution, and is readily explained in the context of our model and chemical evolution equations from
Paper I with the yields and time delays given in Table 2, as
shown in Fig. 2. (Our value of wA for the r-process has been
reduced from 0.008 in Paper I to 0.007 in the light of the
information from additional elements.) The real situation is
probably more complicated, but the data hardly justify a
more elaborate treatment.
The yields in Table 2 have been chosen ad hoc with
certain constraints. The prompt and slightly delayed yields
for the r-process are in the same ratio and have the same
time delays for all elements, in accordance with the
evidence that the r-process has a unique outcome wherever
it occurs. In the case of the s-process, we assume that the
two contributions should be in the same ratio and involve
the same time delays for the elements near a given 'magic
number' peak, but possibly different ratios and time delays
between the N=50 peak (Rb to Zr) and the N=82 peak
(Ba to Sm). (This could be an oversimplification because
3 SEPARATING R- AND S-PROCESS
CONTRIBUTIONS
To visualize the development of r- and s-process contributions to the various elements of interest, we show in Fig. 1
the data for element ratios relative to europium, taken as
representative of a virtually pure r-process. As expected,
most elements with a large s-process contribution according
to Table 1 show a progressive increase from somewhere
near the pure r-process ratio to the Solar system ratio at [Fel
H] =0. Furthermore, the typical 's-process' elements Sr, Y,
Ba, La and Ce seem to define a plateau between [Fel
H] = -2.1 and -0.5 (cf. fig. 11 of Gratton & Sneden 1994),
whereafter they move up to solar ratios, although Rb (for
which the abundances are rather uncertain, ± 0.2 dex in the
average according to Gratton & Sneden 1994) and Zr do
not conform to this pattern. The pattern, if real, is precisely
what one might expect for a primary element involving two
contributions with different time delays, over and above the
0.0
r- ................................................ ~ .~.':: ..::::::':.:':.:': .'::~..:.: :,:.:,:.::.:..: .!..~
..L!!->.--7::-"""'I
':j;'
0::
~ -0.5
~
••
•
•
/
r- - - - - - - - - - -El- __ ,
-
•
••
•
-LO I-
-
o
0.5
r:'
113
o
-- ... ..
_._. ___ ._._.~._._:._._._., .......... 9
Ul
~
o
\
~ -0.5
•
------- --•
•
o
••
0
•
-4
o
-1
-2
-3
[Fe/H]
>:'
o
" ode
§IIJ 01
. . tP
*
--------------,
"OJ
~ -0.5
11\ '"
-tb' -
~f3-
Il..
.<!
,,-0- ....- ......
t; ;
0
"
"•
-....••••.•....•.............
- - -- - -- -- '1':........................................••.........•.
- - -,\, ;' --.- - •;--*' -- -••- -
;.t~
•
••
•
-4
-3
-2
-1
o
[Fe/H]
Figure 3. Ratio of Ba abundance to those of other elements. The solid line shows the values calculated from our basic model and the dashed
line those from the alternative model. Data points, apart from those with sources identified in Fig. 1, are from Primas, Malara & Castelli
(1994: open five-pointed stars), Gratton & Sneden (1988: open circles), Hartmann & Gehren (1988: six-pointed stars), Zhao & Magain
(1990: open squares), Tautvaisiene & StraiZys (1989: open triangles), Tautvaisiene (in preparation: open rhombs), Edvardsson et al. (1993:
'plus' signs) and Mishenina, Klochkova & Panchuk (1995: asterisks). The large open square shows the r-process ratio predicted from Solar
system data according to Table 1 and the filled five-pointed star the corresponding ratio from CS 22892 - 052.
© 1997 RAS, MNRAS 288,108-116
© Royal Astronomical Society • Provided by the NASA Astrophysics Data System
1997MNRAS.288..108P
114 B. E. J. Pagel and G. Tautvaisiene
the Rb/Sr ratio, for example, is sensitive to neutron density:
Tomkin & Lambert 1983; Malaney 1986; Lambert et al.
1995.) In our basic model, we have the same ratios (which
happen to be 1: 1) and time delays for both sets of elements,
which gives a tolerably good fit in most cases, except for Sr,
Y and Zr in one star, HD 122563 ([Fe/H] = - 2.66). The
remaining discrepancies do not seem especially significant
in the context of Fig. 2, but they appear somewhat more
significant in Fig. 3 in which we plot ratios of Ba to elements
on the N =50 magic number peak, for which considerably
more data are available. The agreement is somewhat
improved if we postulate an alternative model that assumes
a slightly smaller time delay for the first s-process component affecting the elements Rb to Zr, leaving that of the
other component unchanged. Some numerical results from
our models are given in Table 3.
Fig. 4 shows the evolution of heavy-metal to iron ratios,
which we consider to be well fitted by either of our models,
bearing in mind that we aim to account for the lower envelope in the region below [Fe/H] = - 2.3 where there is a
large scatter. We believe that our model, postulating a low
plateau for [Fe/H] < - 2.7, followed by a sudden rise to [m/
Fe] ~ 0 between [Fe/H] = - 2.7 and - 2.3, is a real feature
of the data which is naturally explained by a primary model
with time delays, and that any dependence of yields on
metallicity is obscured by other factors. The near constancy
of s-process to iron ratios for [Fe/H] > - 2 results from the
similarity in time delays between SNIa affecting iron
(wA=O.4 according to Paper I) and low-mass AGB stars
(wA = 0.8) affecting s-process elements, but the slight difference between them gives a good fit to the observed tendency
of elements with s-process contributions to have positive [m/
Fe] values (smaller than those of the mainly r-process elements) in the range - 2 < [Fe/H] < -1, followed by slightly
negative values in the range - 0.8 < [Fe/H] < - 0.3.
low-mass, moderately low-metallicity star (Kappeler et al.
1990), but this seems only to account for roughly about half
of the abundance of s-process nuclides in the present-day
Galaxy. The rest begins to appear already at [Fe/H] = - 2.5
or so, in stars that are almost as old as the Galaxy itself, and
is surprisingly similar in the heavy/light s-process element
ratio which is a signature of the neutron exposure history.
Abundance data (Norris & Da Costa 1995; Smith et al.
1995) for the globular cluster w Cen, where s-process elements are somewhat overabundant relative to field stars and
other clusters for [Fe/H] > - 1.4 or so, also suggest that the
heavy/light ratio is enhanced relative to Solar system values
in just a few extreme cases, whereas for most of the stars it
is the same. Norris & Da Costa accordingly suggest that the
self-enrichment of w Cen involved low-mass AGB stars (1.5
to 3 Mo) evolving over time-scales of 0.5 to 2 Gyr or so,
whereas Smith et al. favour more massive intermediatemass stars (up to 8 Mo) with a high single neutron exposure
from the 13C source capable of explaining the low but nonzero Eu abundance by a pure s-process. We see some difficulties in both hypotheses, since the long time-scale in
w Cen, if relevant, should (as has also been suggested by
Gratton & Sneden 1994) apply equally to the field stars for
which we are trying to model GCE and for which we find an
[Fe/H] threshold for the early s-process differing only a little
from that for the r-process to which a short time-scale most
probably applies; and the massive AGB stars, while fitting
our time-scale very nicely, would only be relevant to the
extreme stars with large heavy/light s-process ratios. Furthermore, it seems that even low-mass AGB stars (the progenitors of CH stars) produce large heavy/light ratios at
such low metallicities. Thus, both in our case and in that of
the less extreme stars in w Cen, we are forced back on the
hypothesis that an unknown class of fairly massive stars
supplied s-process elements in Solar system-like proportions, presumably from the 13C source, at an early stage in
the evolution of the Galaxy.
Our alternative model allows for the possibility of an
additional class of massive stars producing s-process nuclei
on the first magic number peak only; this could arise from a
low single neutron exposure, or from an exponential distribution with a low value of '0' The 'secondary' nature of the
22Ne source makes the 13C source a more likely candidate in
this case also, however.
4 DISCUSSION
We have found that at least two separate contributions to
the s-process are required to explain the GCE of the
relevant elements. One of these, shown in columns 5 and 9
of Table 2, is exactly what one might have expected from
current models of s-process synthesis of the nuclides in the
Solar system, i.e. thermal pulses in the AGB phase of the
Table 3. Model abundances.
u
.007
.009
.011
.040
.40
.80
4.1
[Fe/HJ
-2.71
-2.60
-2.51
-1.95
-1.09
-0.65
0.00
[Sr/HJ
[Y/HJ
[Zr/HJ
a
b
a
b
a
b
-4.04
-3.47
-3.23
-1.90
-0.94
-0.74
0.00
-4.16
-3.60
-2.92
-1.82
-0.87
-0.67
0.00
-4.16
-3.59
-3.35
-1.93
-0.96
-0.76
0.00
-4.38
-3.82
-2.97
-1.84
-0.89
-0.69
0.00
-3.85
-3.30
-3.06
-1.86
-0.90
-0.71
0.00
-3.98
-3.42
-2.86
-1.79
-0.84
-0.64
0.00
[Ba/HJ
[La/H]
[Nd/H]
[Eu/HJ
-4.16
-3.59
-3.35
-1.93
-0.96
-0.76
0.00
-4.04
-3.47
-3.23
-1.90
-0.94
-0.74
0.00
-3.55
-2.99
-2.75
-1.75
-0.81
-0.61
0.00
-3.25
-2.70
-2.46
-1.60
-0.68
-0.48
0.00
aBasic model.
b Alternative model.
© 1997 RAS, MNRAS 288, 108-116
© Royal Astronomical Society • Provided by the NASA Astrophysics Data System
9
.....
~
~
'JJ.
en
~
.....
~
~
en
en
~
....=-n
'e
0
'"1
'JJ.
>
>
>
en
.....
Z
~
=-
.....
~
0'"
Q.,
~
Q.,
-<
....
0
'"1
""C
q
•
~
n
'JJ.
0
~
-....
....9n
0
=
0
.....
'"1
en
->
~
~
0
:::0
@
@
0\
......
I
......
0
00
......
QC
~QC
N
~
s:::
~en
~
-..l
\0
\0
......
rz.
_1
n
0
"*
0
0.5
. tI
-15
-1.5
-1.0 l::...
"*
II
-4
I I I
II
-3
I I I I I I I I
II
.
I I I I I I I I
II
-1
. •*
.
•
o
:
II
.-
•
I I I I I I I I
..-
"*
.'"
.... --; .........•... ..-.........,. .
[Fe/H]
-2
I I I I I I I I
#
'J
n /
.J
fO
r
o
....
. ...
~ ...-..~.:;.~.-:-..~....... ~~.-..
I Ij
]
~
"'...
OJ
rz.
0.0
0.5
-1.0
-4
-3
~ -0.5 FF--------~
Q
'>;
.
t:..
"';:l
rz.
W
"
.!"
•
.
" •
'"
"
[Fe/H]
-2
•
,,"
-1
. ..
~
<>
".
•
...................................~.. ~.~
,.
•
o
1
z.. ·...... ..::-:·.-,.~ ...... ~..~.+~ ~ilI\i,:,.;' tit.,.=!
o.o~ .......................................~~.;.~: ......... ~ .. ~
~ -0.5 [
S
OJ
rz.
~
"'-0
OJ
rz.
~
w
rz.
2..
W
rz.
"'OJ
Figure 4. Heavy-element to iron ratios as a function of [Fe/H]. Symbols as in Figs 1 and 3 with the addition of data from Morell, Kallander & Butcher (1992: filled rhombs) and Zhao & Magain
(1991: filled six-pointed stars).
~
"*
0
-1.0~ _ _ _ _ _ _ _ _ _ _ _ _ _
';;- -0.5
W
rz.
~
';::;- -05
W
'---'
! -~.~[
ES
"'.0
OJ
rz.
VI
,.....
,.....
~
1;j
~
~
~
~
~.
~
..Q.,
;:::
s·
12'
~
"-
2'
~
g.
......
(=:;.
~
S'
(")
1997MNRAS.288..108P
1997MNRAS.288..108P
116 B. E.1. Pagel and G. Tautvaisiene
ACKNOWLEDGMENTS
We thank Poul-Erik Nissen for helpful comments. GT
thanks NORDITA for hospitality while the work for this
paper was being carried out.
REFERENCES
Baraffe I., Takahashi K, 1993, A&A, 280, 476
Burbidge E. M., Burbidge G. R., Fowler W. A, Hoyle F., 1957,
Rev. Mod. Phys., 29, 547
Cameron A G. W., 1982, Ap&SS, 82, 123
Clayton D. D., 1988, MNRAS, 234, 1
da Silva L., de la Reza R., de Magalhiies S. D., 1990, in VangioniFlam E., Casse M., Audouze J., Van J. T. T., eds, Astrophysical
Ages and Dating Methods. Editions Frontieres, Gif-surYvette, Paris, p. 419
Edvardsson B., Andersen J., Gustafsson B., Lambert D. L., Nissen
P. E., Tomkin J., 1993, A&A, 275, 101
Franc;ois P., Spite M., Spite F., 1993, A&A, 274, 821
Gilroy K K, Sneden c., Pilachowski C. A, Cowan J. J., 1988, ApJ,
327, 298
Gratton R. G., Sneden C., 1988, A&A, 204, 193
Gratton R. G., Sneden c., 1994, A&A, 287, 927
Griffin R., Gustafsson B., Vieira T., Griffin R., 1982, MNRAS, 198,
637
Hartmann K, Gehren T., 1988, A&A, 199,269
Hollowell D., Then I., Jr, 1989, ApJ, 340, 966
Kappeler F., Beer H., Wisshak K, 1989, Rep. Prog. Phys., 52,
945
Kappeler F., Gallino R., Busso M., Picchio G., Raiteri C. M., 1990,
ApJ, 354, 630
Kipper T., J!ilrgensen U. G., 1994, A&A, 290, 148
Lambert D. L., 1989, in Waddington C. J., ed., Cosmic Abundances
of Matter. Am. Inst. Phys., New York, p. 168
Lambert D. L., Smith V. V., Busso M., Gallino R., Straniero 0.,
1995,ApJ,450, 302
Luck R. E., Bond H. E., 1991, ApJS, 77, 515
McWilliam A, Preston G. W., Sneden c., Searle L., 1995, AJ, 109,
2757
Magain P., 1995, A&A, 297, 686
Malaney R. A, 1986, MNRAS, 223, 709
Malaney R., Fowler W. H., 1989, MNRAS, 237, 67
Mathews G. J., Bazan G., Cowan J. J., 1992, ApJ, 391, 719
Mathews G. J., Cowan J. J., 1990, Nat, 345, 491
Mishenina T. V., Klochkova V. G., Panchuk V. E., 1995, A&AS,
109, 471
Morell 0., Kiillander D., Butcher H. R., 1992, A&A, 259, 543
Norris J., Da Costa G. S., 1995, ApJ, 447, 680
Pagel B. E. J., 1965, R. Obs. Bull., No. 104
Pagel B. E. J., 1989, Rev. Mex. Astron. Astrofis., 18, 153
Pagel B. E. J., 1991, in Oberhummer H., ed., Nuclei in the Cosmos.
Springer-Verlag, Berlin, Heidelberg, New York, p. 89
Pagel B. E. J., Tautvai§iene G., 1995, MNRAS, 276, 505 (Paper
I)
Plez B., Smith V. V., Lambert C. L., 1993, ApJ, 418, 831
Prantzos N., Hashimoto M., Nomoto K, 1990, A&A, 234, 211
Primas F., Molaro P., Castelli F., 1994, A&A, 290, 885
Raiteri C. M., Gallino T., Busso M., 1992, ApJ, 387, 263
Ryan S. G., Norris J. E., Bessell M. S., 1991, AJ, 102, 303
Smith V. V., Cunha K, Lambert D. L., 1995, AJ, 110, 2827
Sneden c., Parthasarathy M., 1983, ApJ, 267, 757
Sneden c., Preston G. W., McWilliam A, Searle L., 1994, ApJ,
431, L27
Sneden c., McWilliam A, Preston G. W., Cowan J. J., Burris D. L.,
Armosky B. J., 1996, ApJ, 467, 819
Spite M., Spite F., 1978, A&A, 67, 23
Tautvai§iene G., 1997, MNRAS, in press
Tautvai§iene G., StraiZys V., 1989, Izv. Spets. Astrofiz. Obs., 28,
88
Truran J. W., 1981, A&A, 97, 391
Tomkin J., Lambert D. L., 1983, ApJ, 272, 722
Vanture A D., 1992, AJ, 104, 1997
Wallerstein G., Greenstein J. L., 1994, ApJ, 139, 1163
Wheeler J. c., Sneden c., Truran J. W., 1989, ARA&A, 27, 279
Woolf V. M., Tomkin J., Lambert D. L., 1995, ApJ, 453, 660
Yoshii Y., Tsujimoto T., Nomoto K, 1996, ApJ, 462, 266
Zhao G., Magain P., 1990, A&A, 238, 242
Zhao G., Magain P., 1991, A&A, 244, 425
© 1997 RAS, MNRAS 288, 108-116
© Royal Astronomical Society • Provided by the NASA Astrophysics Data System