R-Process Cosmo-Chronometers
Thorium and Uranium in Ultra-Metal-Poor Halo Stars
|
|
R-Process Cosmo-ChronometersThorium and Uranium in Ultra-Metal-Poor Halo Stars |
|
|
|
|
||||||
|
BPS CS29497-004
in
SCULPTOR RA 00:28:06.7 DE -26:03:03 V=14.03, [Fe/H]=-2.7, dist.= |
|||||||
|
|
as two new stars from the HERES Collaboration (The Hamburg/ESO R-process Enhanced Survey): BPS CS29491-069 in PISCIS AUSTRINUS (RA: 22:31:02.1 DE: -12:33:42) HE 1523-0901 in LIBRA (RA 15:26 DE -09:11) |
||||||
| These historic illustrations of the constellations are from Catherine Tennants The Box of Stars: A Practical Guide to the Night Sky and to Its Myths and Legends /Book, Cards and Maps. They are reproductions from Urania's Mirror, published 1825 in London. | ||||||||
The oldest objects presently known in the universe are low-mass,
very metal-poor stars in the halo of the Milky Way Galaxy. These stars
are of course younger than the universe. Thus, if it would be possible
to put more stringent constraints on the ages of such stars, a more
precise lower limit to the age of the Universe can be derived.
Radioactive elements which have very long decay-times, such as
Thorium (Th) and Uranium (U), can be used to accurately
date the age of stars, in a way that is quite similar to that used in
archaeology by means of Carbon-14 dating.
The radioactive element dating method does not make use of stellar
models such as, e.g,. the globular cluster dating method to estimate
the minimum age of the universe. Therefore, it provides an independent
estimate and the comparison of these estimates allows a cross check of
both methods.
Thorium and Uranium are heavy elements that produce faint
absorption lines in the extreme blue and UV part of the spectrum which are
extremely difficult to observe. Already around 1970, Thorium and Uranium
abundances have been reported in the class of chemical peculiar Ap stars
(For a recent
review, see C. and M. Jaschek: The Behavior of Chemical Elements in Stars,
Cambridge University Press, 1995). In 1975, the Th/U-ratio in the star HR465
was applied for an age estimate (C.R. Cowley et al., Nature 258 (1975) 311).
Only recently, faint Thorium and Uranium lines were detected in a small
sample of n-capture-element enriched, ultra-metal-poor
Halo stars as the ones presented above.
In order to use radioactive isotopes for dating, one must either determine the amount of the decaying substance and its decay product simultaneously (as in the potassium/argon method applied to volcanic ashes) and/or one must know the original amount of the decaying substance by measuring stable isotopes of the same element and assuming a constant abundance ratio. This is the case for the well-known Carbon-14 dating method. At least in popular publications, a time-invariant production of radiocarbon by cosmic rays is presumed (In reality, the cosmic ray flux is fluctuating, influenced by the variable strength of solar magnetic fields. Fortunately, this effect can be corrected by measuring wood samples absolutely dated by dentrochronology.).
 |
In the case of the Thorium and Uranium clocks (when applied to rock samples of
terrestrial or meteoritic origin), the final stable isotopes of lead can be
measured simultaneously. If applied to stars, several problems arise: The most trivial case is that one observes the Uranium and/or Thorium lines, but that the lead ones are too faint. Even if one can measure lead, there are still several problems. In the laboratory, one can determine isotopic abundances (of Pb-206, Pb-207 and Pb-208 as endproducts of the U-238, U-235 and Th-232 a-decay chains, respectively), but optical spectroscopy of stars yields chemical abundances, all decay chains mixed up. In addition, when the rocks mentioned above were forming, Thorium or Uranium were chemically fractionized from lead setting the clocks to zero. In a star, there is no chemical separation. The figures left and right show the decay chains ending in the stable lead and bismuth isotopes. The longest-lived member of the neptunium chain, 237Np, has a half-life of only 2 million years. Therefore its members die out shortly after a nucleosynthesis event and could not be detected radiochemically as was the case with the other three natural chains. [The historic names are used in the plots. Element 86Em is now called radon Rn.] |
 |
A comprehensive study of physical dating techniques (not omitting the pitfalls)
can be found in:
M.J. Aitken: Archaeological dating using physical phenomena;
Rep. Prog. Phys. 62 (1999) 1333
Three ways of determining the age of the Universe are discribed in "Ned
Wright's Cosmological Tutorial" Age of the
Universe.
|
In order to avoid the complications related to chemical Galactic evolution models, scientists have to rely on theoretical assumptions to estimate the original Thorium/Uranium concentration. The observation in recent years of ultra-metal-poor yet neutron-capture elements enriched stars in the Halo of the Milky Way Galaxy was of upmost importance for the field [see Fig. left]. These stars were born around the birth of the Galaxy. At that point, only few (might-be even only one) nucleosynthesis events producing these heavy elements including Thorium and Uranium could have contributed to the interstellar gas and dust from which these (second generation) stars were forming. And the only nucleosynthesis process occurring shortly after the Big Bang (say in the order of several hundred million years) is the rapid neutron-capture process (r-process), which takes place in type II supernovae explosions (SnII) representing the final stages in the life of massive short-lived stars. |
|
A plot
of the nuclides contributing to the r-process and the resulting abundances
is shown, superimposed on a representation of
b-lifetimes. The small black squares are the
stable isotopes, the
black line represents the limit of the known nuclides on the
neutron-rich side, and the magenta line below and to the right is a
typical r-process contour. The small magenta squares show the
nuclides that are produced when the r-process line decays.
Courtesy of Guided Tour of the Nuclear Information Service at Los Alamos |
 |
Starting with the epochal work of Burbidge, Burbidge, Fowler, Hoyle (B²FH) Synthesis of the Elements in Stars (Rev. Mod. Phys. 29 (1957) 547) much work was dedicated to the theoretical reproduction of the solar system abundance pattern of the heavy elements. A review of the classical, site-independent approach (superposition of r-process abundances at different neutron densities nn where weight and time scale obey a power law over nn) followed by the groups of F.-K. Thielemann/Basel and K.-L. Kratz/Mainz can be found in K.-L. Kratz et al., Ap. J. 403, 216(1993). The r-process involves a great number of extremely neutron-rich, very short-lived isotopes not accessible to experiment (at least in the foreseeable future). Nuclear structure properties (as nuclear ground state masses, half-lives) entering into the calculations have to be taken from theoretical nuclear models. In order to fine-tune the model parameters, nuclear spectroscopy experiments have to be driven to the extreme limits of technologies. A recent overview of experiments performed at the on-line isotope separator ISOLDE at CERN, Geneva, can be found in K.-L. Kratz et al., Nuclear Structure Studies at ISOLDE and their Impact on the Astrophysical r-Process in Hyperfine Interactions 129 (2000) 185. Important progress in this field is to be expected from the "Radioactive Ion Beam Facilities" under construction worldwide (for an example, see The RIA Project at NSCL-MSU.
|
Measured elemental abundances from the
metal-poor halo star CS22892-052 [C. Sneden et al., Ap. J. 467 (1996)
819 and priv. comm.] are
compared to the scaled solar system s- and r-process values (green and red
lines, respectively). From Ba (Z=56) on, the observed pattern closely follows
the (red) Nr,solar abundances. For a comparison to calculated r-process abundances, see Comparison with r-process calculation. |
The nearly perfect agreement brakes down for the heaviest elements, Thorium and Uranium (where only an upper limit could be given in two independent measurements). Assuming, that these radioactive elements were produced in the same unique r-process, the theoretical elemental abundances for these two elements can now be used as the sought-for initial production abundances for the age determinations. In the case of the two stars with metallicity [Fe/H]=-3 CS22892-052 and HD115444, a mean decay age of (15.6±4) Gyr was derived from the observed Thorium abundance (see J.J. Cowan et al., R-Process Abundances and Chronometers in Metal-Poor Stars in Ap.J. 521 (1999) 194).
|
|
Caption: ESO PR Photo 09c/00 shows the
wavelength region around the U II (singly ionized Uranium) 385.95 nm
line, as observed in the old star CS22892-052 with UVES
(as the fully drawn "step" line in the large panel, and as dots in the
enlarged section in the small panel).
The very faint Uranium line is discernible as a slight "depression" in the recorded spectrum. In order to estimate the critical upper limit of the Uranium abundance in the star's atmosphere, several synthetic spectra with different logarithmic abundance of Uranium (from "none" to -2.31 on a scale where the logarithmic hydrogen abundance is 12) have been computed (continuous thin lines in both diagrams). The abundance values are indicated - it can be seen that the model with [U/H] ~ -2.5 is clearly below the observed spectrum, i.e., this value is a very conservatory upper estimate of the Uranium abundance. The strong lines on either side of the U II line are due to iron (Fe I), while the weaker features are blends of lines from various elements. ( See ESO Press Release 08/00) |
|
|
Caption: ESO PR Photo 05b/01 The observed spectrum (dots) of the old star CS 31082-001 in the region of the Uranium (U II) line at 385.96 nm. The origin of some of the other spectral lines in the region is also indicated (e.g. iron, neodymium). The synthetic spectrum (thin line) was computed for the adopted abundances of the stable elements and for four different values of the abundance (by number) of Uranium atoms in the atmosphere of the star. The uppermost line (corresponding to no Uranium at all) clearly does not fit the observed spectrum at all. The best fit is provided by the middle (red) line, representing a Uranium abundance of approximately 6% of the solar value. (See ESO Press Release 02/01) |
|
First results on Thorium and Uranium abundances combined with our initial
production values (as published in ApJ 521 (1999) 194) yield an estimated age of
(12.5±3) Gyr for the star CS31082-001 derived from the Th/U ratio (R.
Cayrel et al., Nature 409 (2001) 691). This can be compared to the
most recent age determination for CS22892-052 of
(15.4±1.6rms) Gyr presented at the conference on Astrophysical Ages and Time
Scales by
Hannawald et al. Contrary to CS22892-052, the scaled Th abundance in CS31082-001 is higher than our calculated production value. The Th/Eu radioactive dating technique yields a negative age for this star. |
||
 |
The figure at the left from V. Hill's contribution to Astrophysical Ages and
Time Scales compares the abundances in CS31082-001 with two other
ultra-metal poor stars, CS22892-052 and HD115444, respectively. And whereas
the abundance patterns for Ba up to the 3rd abundance peak elements
Os and Ir are in agreement, there exist striking differences for Pb, Th and U.
The (properly scaled) Th and U abundances in CS31082-001 exceed the values for
the other two stars, whereas the Pb abundances behave just in the opposite way.
But, for very old stars of the first generations, the main contributions to the
Pb abundance ought to be from extensions of radiochemically well studied
naturally occuring a-decay chains originating
in the long-lived 232Th and 235,238U isotopes. [In
explosive nucleosynthesis events, there is an additional a-decay chain terminating in the stable 209Bi,
an element difficult to observe in old stars. As this chain encompasses only
"short-lived" isotopes, all the instable members have long been
decayed to Bi in the solar system.] Prior to derive far reaching conclusions, one ought to wait for Hubble Space Telescope (HST) observations of two of these stars. First results on Os, Pt, Au and Pb abundances obtained in 60 orbits dedicated to CS22892-052 have been reported by Sneden et al. at the 199th AAS meeting in January 2002 [#137.10] and 40 orbits to study CS31082-001 have been attributed in HST Cycle 11 (starting summer/fall 2002). |
It was mentioned above, that one reason to use Thorium and Uranium chronometry simultaneously was the assumption that nuclear structure effects would act on 232Th and 238U in the same way, so that by taking the ratio of the calculated values most of the theoretical uncertainties would cancel out. First results on a careful re-evaluation of the influence of nuclear physics input data (as different mass models, half-life evaluations, spontaneous and b-delayed fission rates) on the predicted r-process production ratios of 232Th and 238U performed by H. Schatz/NSCL(MSU) et al. and the Mainz group (applying the "Canonical r-process model" described above) were presented at the Int. Conf. on Astrophysical Ages and Time Scales by Toenjes et al.
|
Tuning the clock - Thorium and Uranium chronometers applied to CS31082-001 Int. Conf. on Astrophysical Ages and Time Scales, February 5 - 9, 2001; Hilo, Hawaii Proceedings to appear in the Astronomical Society of the Pacific ASP Conference Series |
In the figure above, the stable isotopes are shown as dark black squares, whereas the radioactive isotopes in the r-process path are the grey squares. All the coloured squares represent the precursor isotopes of 232Th and/or 238U. After neutron freeze-out, the highly unstable isotopes in the r-process path decay by b-, b-delayed neutron-, b-delayed fission-, a-decay and/or spontaneous fission back to the valley of stability. Nuclear physics data on all of these isotopes have to be known, preferentially from experiment, but in most of the cases from theoretical nuclear structure models.
|
| ||||
|
The left plot demonstrates the stability of the [U/Th] ratios
against variations in the coefficients for the power law dependance of the
weights and time scales on neutron densities for the different
components as calculated in the "canonical r-process" as described
above. These calculations were performed applying the ETFSI-Q mass model
(Pearson et al., Phys. Lett. B387 (1996) 455), which had also been used
for the abundances presented in Cowan et al., ApJ 521 (1999) 194. In the right plot, abundances derived with this mass model are displayed together with values obtained applying the different, newly developed Harteee-Fock based mass model HFBCS-1 (labeled HFBGor) from S. Goriely et al., At. Data Nucl. Data Tables 77 (2001) 311. Of special interest are the deviations in the region A=230-240, which influence the calculated Thorium and Uranium abundance ratios. The influence on the age estimate can be seen in Table 1 of the Nature article, where two theoretical estimates for the initial U/Th abundance ratios are used: on the one hand from the paper of Cowan et. al. (Ref.4 in Nature) and on the other hand from Goriely and Clerbaux (Ref. 17 in Nature, cited from Astron. Astrophys. 346 (1999) 798). One of the reasons for the discrepant estimates are differences between the two mass models which entered as nuclear physics input into the abundance calculations by the two groups cited in Nature. In this case, one obtains age estimates for CS31082-001 of 10.6 or 14.0 Gyr, respectively. |
|||||
In the near future, spectroscopic observations of metal-poor stars will continue in two ways: on the one hand, new sky surveys will look for more examples (as the Hamburg/ESO objective-prism survey HES) to get a broader statistical and systematic basis; on the other hand, high-resolution spectroscopy of individual stars will continue. With a larger sample, systematic studies can be performed as to examine the relation between metallicity and age or to look, at what metallicity different nucleosynthesis processes (as the s-process) start to contribute to the elemental abundance pattern.
|
Sneden et al. (197th AAS Meeting (2000) #44.18)
reported on the
less metal-rich ([Fe/H]=-2.0) star BD +17° 3248
(Bonner
Durchmusterung, Argelander 1859-62) in the constellation HERCULES to have an estimated Thorium age of 11.6 Gyr,
lower than for the less metal-rich stars presented above, but taking the
uncertainties into account still consistent with the other Halo stars. Further observations [reported by Cowan et al. at the 199th AAS Meeting (2002) #137.06 and displayed at the left side together with results for CS22892-052, HD115444 and CS31082-001] allowed to employ the newly detected Th, U and 3rd r-process element abundances to make chronometric age estimates for this star. The various chronometric pairs suggest an age of 13.8 ± 4 Gyr. For more detailed information, see press release: text and figures. |
 |
First indications to a contribution of s-process
nucleosynthesis were observed for the Halo star HD126238 with [Fe/H]=-1.7
(see the increased Z=56 Ba abundance in the lower-right figure on the web-page
Heavy Elements in Metal-Poor Halo Stars).
Johnson and Bolte reported on a sample of 22 metal-poor stars confirming the "universal r-process pattern". They determined the Th ages for 4 field red giants and one M92 (NGC 6341) giant. (J.A. Johnson and M. Bolte: Ap. J. 554 (2001) 888) |
Furtheron, the accuracy of the Uranium dating technique is limited at present
by incomplete knowledge of critical atomic and nuclear physics data, in
particular oscillator strenghts and production ratios of the elements produced
in the r-process. Experimental and theoretical work is already underway
and should soon lead to substantial progress in r-process cosmo-chronometry.
A first paper from the group of S. Johansson at the university of Lund
on atomic transitions in Uranium ions has been published in
"Astronomy and Astrophysics" 372, L50 (2001): H. Lundberg et al.
New
laboratory lifetime measurements of U II for the uranium chronometer.
|
Ch. Sneden et al. (ApJ 536 (2000) 85) could detect Thorium lines not only
in field Halo stars but also in stars in the globular cluster M15 (NGC7078) in
the constellation PEGASUS.
From the Th/Eu ratios an age estimate of (14±3)Gyr is obtained. This
is in fair agreement with the most recent estimates of the mean age of the
globular clusters of
(12.9±2.9)Gyr by E. Carretta et al. Ap. J. 533 (2000) 215
(astro-ph/9902086).
The age of the oldest globular cluster NGC6366 from a sample of 28 halo clusters
is given as (12.2±1.1) Gyr by M. Salaris and A. Weiss
(A&A 335 (1998) 943). In 2007, Wako Aoki et al. report the first determination of Th in an extragalactic star, COS 82 in the local group UMi dSph (Ursa minor dwarf spheroidal galaxy). The Th/Eu ratio is similar to the field Halo star BPS CS22892-052, indicating to an old age of this dwarf galaxy. |
|
HST observations of the nearest globular cluster M4 (NGC6121) in SCORPIUS
during 7 days revealed the faintest White Dwarf stars ever observed.
From the cooling rate an age of (12.7±0.7)Gyr is deduced for the oldest
stars in M4, much older than White Dwarfs in the Galactic disk
(7.3±1.5)Gyr). Results from measurements in 1995 can be
found at
"The Messier Catalog" and in forthcoming articles by
H.B. Richer et al. and
B.M.S. Hansen et al.
in Ap.J. Lett.
From the light curves of 42 SN Ia, Perlmutter et al. (ApJ 517 (1999) 565)
derived an age of the flat, accelerating Universe of (14.9 +1.4/
-1.1) Gyr.
An analysis of 11
cosmological parameters using both the fluctuation spectrum of
the cosmic microwave background (CMB) and the large scale structure of the
distribution of galaxies in space yielded a "best" value for the age
of the Universe of 14.0 Gyr, with the value between 12.1 and 14.6 Gyr with
a probability of 95% - that would fit (see
Tegmark et al.
Phys. Rev. D63 (2001) 043007).
Assuming a flat universe and structure formation from adiabatic initial
conditions
L. Knox et al. [ApJ 563 (2001)L95] constrain the age of the Universe from
CMB observations to (14.0±0.5) Gyr (and obtain an argument for
"Dark Energy" independent of supernovae observations).
In combining the cosmological results with the globular cluster data of
Salaris and Weiss (cited above) and the U age of Cayrel et al., I. Ferreras
et al. derive an age of the Universe as (13.4+1.4/-1.0)
Gyr (Mon. Not. R. Astron.
Soc., in print). This confirms a value of (13.4±1.6) Gyr derived
earlier by C.H. Lineweaver (Science
284 (1999) 1503).
In an up-date of this study submitted to Phys. Rev. D in May 2001, the age of
the Universe is given as (12.7±1.5) Gyr (X. Wang et al. Is cosmology consistent?).
It should not be forgotten that the radioactive dating determines the beginning
of nucleosynthesis of the heavy elements by stars and therefore yields only a
lower limit for the age of the Universe. To obtain from this an estimate for
the age of the Universe one has to add the time from the "Big Bang"
to the formation of the first stars and galaxies. In the absence of stars matter
(He and H from Big Bang nucleosynthesis) was in the ground state and did not
emit light, this early time of the Universe is therefore called the "Dark
Ages". In August 2001, two teams of astronomers reported on first hints
in quasar spectra of the end of the "Dark Ages" (or the beginning
of the "Reionisation Epoch") about 1 Gyr after the "Big
Bang" (R.H. Becker et
al. and S.G.
Djorgovski et al., respectively).
From a reanalysis of the "Hubble Deep Field" exposures, K.M. Lanzetta
et al. suppose, that the star forming rate peaked already about .5 Gyr
after the Big Bang (see left figure), implying changes in models of Galactic
chemical evolution.
This time span has to be added to the
Th and U ages to obtain a lower limit to the age of the Universe.
At the beginning of the article "The Enigma of Przybylki's Star",
the autors pose the question Was a weird star in Centaurus the Rosetta
Stone for the creation of rare elements? [Guy Worthey and Charles R.
Cowley, Sky & Telescope, August 2004, p. 50]Recently, Cowley, Hubrig and Bord applied the Mainz r-process abundances for estimating the age of this star from the observed Th/U ratio: Actinides in HD101065 (Przybylski's Star).
Remark: Astronomers/astrophysicists ought to cite the less known
"Canopus Stone" which also had been discovered during the campaign
of Napoleon in Egypt. Part of the inscription in three scripts and two
languages orders a calender reform. Ptolemy III Euergetes and Berenice demand
in 238 B.C. to add an additional day every 4 years in order to synchronize the
calendar with the seasons. The egyptian people did not accept this. Later on
Caesar adopted this idea under the advice of Sosigenes of Alexandria as the
Julian Calendar.
The hair of queen Berenice was projected to the sky as the constellation
"Coma Berenices" (see middle figure at top of page).
Further readings:
 |
GSI Darmstadt (Germany), May 3 - 4, 2001 |
 |
Mainzer Forscher lieferten entscheidende Daten Gabi Henkel im Feuilleton der Allgemeinen Zeitung Ausgabe Samstag 3. März 2001 |
 |
|
Datierung anhand eines langlebigen Uranisotops Georg Wolschin in NZZ Forschung und Technik Ausgabe Mittwoch 21. Februar 2001 |
 |
 |
 |
Last updated April 2006 |  |
 |
