Observations of Neutron-Capture
Elements in the First Stars
{
Timothy C. Beers
University of Notre Dame
SDSS
Rapid Neutron-Capture Processes in the
Universe – A QUICK Observational History
In the Olden Times (pre-1980s)…

Only objects we had to compare with
predictions of the r-process were:

The Sun and meteorites

A handful of metal-poor and very metal-poor stars

Napkins written on during conferences like this one
by a few smart people
Finding the Targets – the Very, Extremely, Ultra
Metal-Poor ([Fe/H] < -2.0, < -3.0, < -4.0) Stars
 The Bond Survey
Late 70s
 Carney / Latham Survey
Early 80s
 Ryan / Norris Survey
Late 80s
 HK Survey / Beers et al.
Mid 80s – 90s
 Hamburg/ESO Survey /
Christlieb et al.
Mid 90s – 00s
 SDSS Survey
Early 00s
 SDSS/SEGUE-I Survey
Mid 00s
 SDSS/SEGUE-II Survey
Late 00s – Early 10s
 LAMOST Survey
Early 10s – Mid 10s
 SkyMapper Survey
Mid 10s – Late 10s
The Push Below [Fe/H] = -4.0
There are now ~30,000 stars known with heavy metal
abundances below 100 times solar, ~1000 below 1000 times,
~10 below 10,000 times, two between 100,000 and
1,000,000 times, and one below [Fe/H] = -7.3
(20,000,000 times below solar)
The mad dash to the lowest metallicity is ending, IMHO —
The “Δ Science” below [Fe/H] ~ -7.0 is not sufficient
to justify the effort – there are few (n < 5) readily detectable
species ! (in the < -7.3 star: Li, C, Mg, Ca, no Fe !)
Maximal information on first-star nucleosynthesis
will come from supplementing numbers of (ideally, bright)
stars with detectable lines in the optical and near-UV
-5.5 < [Fe/H] < -3.5
Exploration of Nature’s Laboratory
for Neutron-Capture Processes
Beers & Christlieb ARAA (2005)
Stars with Nearly Pure r-Process
Frebel & Norris (2011)
Numbers of r-II and r-I Stars Known

Only dedicated survey thus far is the HERES survey
(Christlieb et al. 2004, Barklem et al. 2005)



Based on “snapshot” high-res spectra of some 350 stars with
[Fe/H] < -2.0
The results of that effort totaled 8 new r-II stars
Other “serendipitous” follow-up efforts have yielded
another ~8 new r-II stars; Roederer et al (2014) high-res
follow-up of HK Survey stars, 9 new r-II

A total of ~25 r-II thus far (roughly 5% of VMP stars)
A total of ~75 r-I stars (roughly 15% of VMP stars)

 These numbers need to be trebled or quadrupled !

Observational Facts / Speculations /
Implications about r-II and r-I Stars







Sr has been detected in one of the lowest metallicity stars known, HE 1327-2326, with [Fe/H] =
-5.7. While only upper limits on Sr and Ba are available for the handful of other stars known
with [Fe/H] <-4.5, the presence of Sr in HE 1327-2326 indicates that at least one channel exists
for the production of elements beyond the iron peak in the most metal-poor stars.
Relative numbers of r-II and r-I stars, and the dependence on [Fe/H], may reveal path to
understanding of the astrophysical origin(s) of the r process
Understanding the origin of the “actinide boost” may be a key for a given class of r-process sites
r-II stars are found to occur for a broad range of stellar evolutionary stages –- D, SG, G, RHB,
eliminating possibility that r-process enhancements are related to peculiar atmospheric
conditions in red giants
r-II and r-I stars exhibit a (lack of) radial velocity variations. indicating that they do not require
binary membership (mass transfer) to form
There appears little or no connection between r-process-element enhancement and enhanced C
(or N). CEMP-r (and CEMP-r/s) stars are formed by independent nucleosynthesis events for C
and r-process
Light element (12 < Z < 30) abundances of r-II stars are indistinguishable from those produced by
“normal” core-collapse (non r-process enhanced) supernovae; hints that r-II and r-I stars may
exhibit different first-peak (Sr, Y, Zr) element amounts
Detection of Sr ([Sr/Fe] ~ +1.0 in HE 1327-2326
(upper limit on [Ba/Fe]), with [Fe/H] = -5.7 -Indicates a production channel for (light) n-capture
elements at the earliest times
Roederer et al. (2014a)
Distribution of [Fe/H] for r-II and r-I Stars
In some r-II stars -- about 20-30% -- the abundance of U and/or
Th relative to Eu are incompatible with expected ages of stars,
but compatible with [U/Th] ages – “actinide boost”
Mashonkina et al (2014)
r-II stars have now been identified in a variety of evolutionary
stages -- not just the red-giant branch -- including at least one
main-sequence dwarf r-II star identified by Aoki et al. (2010).
This effectively eliminates the possibility that r-process
enhancement is due to peculiarities in red-giant atmospheres
Roederer et al (2014b)
r-II and r-I stars have been followed up with radial-velocity
monitoring programs (Hansen et al. 2011, Roederer et al.
2014b). There is no evidence that binarism is a requirement for
r-process enhancement (binary rate ~ 20%)
Hansen et al (2011)
There appears to be no relationship between carbon
enhancement and r-process enhancement. The CEMP-r
(CEMP-r/s) class of stars likely had C enhanced in natal clouds,
followed by pollution with r-process nucleosynthesis events
The first r-II star found, CS 22892-052,
also exhibited enhanced carbon,
[C/Fe] ~ +1.0, raising possibility that
r-process enhancement was somehow
linked. Further studies, most recently
By Roederer et al. (2014b), have shown
that only ~ 15% of r-II stars are carbon
enhanced. There is no connection.
Roederer et al (2014b)
It has been shown recently that there is little (if any) difference
between levels of the light elements (Mg, Si, 12 < Z < 30) for r-II
stars and “normal” metal-poor stars. There is evidence for
different first-peak levels (Sr, Y, Zr) between r-II and r-I stars.
Residuals of Sr, Y, Zr
for r-I stars (blue/red)
relative to r-II stars
(magenta) and
normal stars (black).
Note 0.5-1.0 dex
offset, in the sense
r-I > r-II.
May indicate different
astrophysical sites!
Sequeira Mello at al. (2014)
The Path Forward






Need to (at least) double, preferably treble or quadruple, numbers of
recognized r-II and r-I stars. This can only be accomplished with new
dedicated high-resolution survey efforts -- new surveys underway
Need to continue RV monitoring programs for r-II and r-I stars, in order to
nail the final result that mass-transfer events not required, and origin of
CEMP-r/s stars
Need to account for production of first-peak neutron-capture elements
(such as Sr) at the lowest metallicity –- First Star progenitors
Need to understand origin of the “actinide boost” phenomenon -theory needed !
Need to explore Galactic Chemical Evolution models to see if variety of
observational constraints (e.g., [Fe/H] distribution of r-II vs. r-I, first-peak
differences between r-II and r-I) can be understood in terms of a hybrid
model involving early SNe and later (and continued) NS/NS mergers. Or
GRB/non-GRB explosions. Other testable predictions from these sites
would be useful
THIS IS NOT A 25-YEAR PROBLEM (ANYMORE!) / IT IS A 5-YEAR PROBLEM