Why Study Stellar Chemical Compositions?
The elements of the periodic table do not just magically exist! They have been created, mostly in nuclear fusion reactions that occur in the extremely hot and dense interiors of stars. Each star generates heavy elements and returns them to the interstellar medium as it dies. Thus the overall metal content of our Galaxy has grown with time. But our knowledge of elemental nucleosynthesis is far from complete, so much so that a panel of the National Research Council's board on physics and astronomy named "How were the heavy elements from iron to uranium made?" as number three in their list of The 11 Greatest Unanswered Questions of Physics.

A general overview of stellar chemical compositions and nucleosynthesis of the elements: Here is the cover page from the November/December 2005 issue of McDonald Observatory's StarDate magazine. StarDate writer Rebecca Johnson wrote this copyrighted article, which you are welcome to read on-line by clicking on its title: Chemical Universe.
Chris Sneden Chris Sneden

The record of our Milky Way Galaxy's initial burst of element generation is contained in the chemical compositions of metal-poor stars of the Galactic halo. I am involved in several related efforts to apply high resolution, high S/N spectroscopy to the determination of abundances in stars with metallicities ranging over three orders of magnitude. Here I briefly discuss a few aspects of these projects, These results are products of the collective efforts of my friends who are co-authors on the various papers.

Neutron-Capture Elements: the r-Process

Neutron-capture elements are those whose major isotopes are created in nuclear fusion regions via the ingestion of free neutrons by lighter seed nuclei (such as Fe, for example). The neutron-capture elements defined in this way comprise essentially all elements of the periodic table with atomic numbers greater than 30, thus comprising nearly two-thirds of the stable elements in the periodic table. Neutron-capture and beta-decay rates compete with each other in building the neutron-capture elements. Low neutron fluxes (the so-called s-process) create unstable nuclei that undergo beta-decays between successive neutron captures, and element synthesis proceeds along the valley of beta-stability. High neutron fluxes (the r-process) create extremely heavy isotopes of nuclei (up to the neutron-drip line) that then rapidly decay back toward the valley of β-stability after the cessation of the neutron flux. The s-process occurs during quiet stellar evolution (such as during helium burning), while the very violent r-process is associated in some way with the explosive deaths of high-mass stars in Type II supernovae.

The most metal-poor of our Galaxy's stars exhibit enormous variations in their bulk contents of neutron-capture elements, as a glance at these spectra show.
Chris Sneden

A small portion of the near-UV spectrum in the Sun and two very metal-poor stars: The top panel contains the solar spectrum (Delbouille et al. 1973). The bottom panel shows spectra of two very metal-poor giant stars: HD 122563, deficient in neutron-capture elements; and CS 22892-052, strongly enhanced in such elements. Click on picture to see larger image
Studies by our group and others have shown that the dominant mode of neutron-capture element production in extreme cases such as CS 22892-052 (e.g Sneden et al. 2003) and CS 31082-001 (e.g Siqueira Mello et al. 2013) and many other halo stars has been the r-process. This is illustrated in the elemental abundance plot for HD 108317 shown here. There is a very nice match to the expected abundance distribution from the r-process, but no match at all to an s-process pattern. This indicates the dominance of explosive synthesis of the neutron-capture elements in the early Galactic history. The availability of HST/STIS high-resolution spectra for a few r-process-rich stars has been crucial in filling in the abundance curves with detections of elements such as Ge, Cd, Te, Lu, Os, Pt, and Au.

Observed and predicted abundances for neutron-capture elements in HD 108317: This plot is taken from Roederer et al. 2012). The black symbols are the observed abundances, the open triangles are upper limits, the blue curve is the solar-system (meteoritic) s-process abundance distribution normalized to the observed Ba abundance, and the red curve is the solar-system r-process distribution normalized to the observed Eu abundance. The abundance units are logarithmic number densities on a standard scale in which log ε(H) = 12. The lower panel shows the differences between the observed abundances and the scaled solar-system curve. Click on picture to see larger image
Chris Sneden
The very heavy neutron-capture element thorium (Th) is created only via the r-process, but it is radioactively unstable, decaying with a half-life of 14 billion years. This element has been detected in a lot of very metal-poor, r-process-rich stars. In the above figure, the relative abundance of thorium to the other neutron-capture elements less than predicted, because as time has gone on thorium has decayed away. This yields an "age" for Th and other neutron-capture elements of about 15 billion years, possibly providing an independent age estimate for the halo of our Galaxy. For a general discussion of this work, please see my interview by StarDate writer Rebecca Johnson in 2002 (copyright University of Texas at Austin). For a technical summary of our r-process abundance studies in halo metal-poor stars, please see Cowan & Sneden (2003, Nature, 440, 1151). A selected list of other papers emphasizing the nucleosynthetic interpretation of these abundance results may be found John Cowan's website.

Neutron-Capture Elements: the s-Process

In the Galactic halo live many "blue metal-poor" (BMP) stars. These are stars that are on the main sequence but are bluer and more luminous than the normal stars of the main-sequence turnoff. If they are as old as the rest of the Galactic halo, BMP stars should have long-since evolved away from the main sequence. George Preston discovered (2000, AJ, 120, 1014) that about 2/3 of BMP stars are spectroscopic binaries, ones with long orbital periods but very low eccentricities (that is, nearly circular orbits). The simplest conclusion is that BMP stars that we see today are past victims of substantial mass transfer from companions that now are compact objects. The mass transfer happened during the expansion of the companion envelopes during their red-giant branch (RGB) or asymptotic giant branch (AGB) phases.

Further proof of this scenario has come from chemical composition studies of BMP stars. We discovered (Sneden, Preston, & Cowan 2003, ApJ, 592, 504) that several of the most metal-poor BMP binaries have greatly enhanced neutron-capture abundances with a distinct s-process distribution. The heaviest elements in the s-process chain, lead (PB) and bismuth (Bi) were discovered in one BMP star, as the following illustration shows.

Chris Sneden

Selected spectra of neutron-capture elements in the BMP star CS 29497-030: These plots, taken from Ivans et al. 2005), illustrate observed and synthetic spectra of several strong transitions. Without very large overabundances of neutron-capture elements, these spectral lines would be undetectably weak. The detection of bismuth (top panel) is the first detection of this element in metal-poor stars. Click on picture to see larger image

Detailed comparisions with predicted s-process abundance distributions confirm its dominance in BMP binaries, as illustrated in the nex figure.

Comparison of observed and predicted neutron-capture abundances in CS 29497-030: This plot is taken from Ivans et al. 2006). In the top panel the observed abundances are open circles (upper limits are depicted with arrows), a predicted pure s-process abundance set is the red dotted line, and an s-process distribution with the addition of a prior enrichment by the r-process is the blue solid line. In the bottom panel, the red and blue symbols are the values of the observed abundances minus the pure s-process and the r+s-process predictions, respectively. The black bars represent the observational uncertainties. The abundances are in the standard bracket notation, defined as the logarithmic abundance ratios by number X/Fe minus the same ratios in the Sun. Click on picture to see larger image
Chris Sneden

Globular Cluster Abundances

Chris Sneden

Images of two bright globular clusters: On the left is a ground-based picture of M13 On the right is a color picture of M15 from the Hubble Heritage program. Click on pictures to see larger images
Chris Sneden

The elemental abundance patterns of globular cluster stars exhibit a rich (and often bewildering) variety. Very large star-to-star and cluster-to-cluster abundance differences are the rule, not the exception. The light elements of the periodic table (those elements with atomic numbers 13 or less) have the most obvious variations. The abundances of elements C, N, O, Na, Mg, and Al can be different in otherwise very similar giant stars of a single globular cluster. Fortunately, these abundances do not vary at random, but instead are interlinked in various correlations and anticorrelations. But some stars will have relatively large abundances of O and Mg and small abundances of Na and Al, while other stars will show the opposite abundance pattern. Good examples are the plots shown in the diagrams below.

Chris Sneden
Chris Sneden

Light element abundance correlations in M13 (left) and in several clusters (right): Please see the papers for detailed explanation: Kraft et al. 1997) and Carretta et al. 2009) and. Click on pictures to see original images

A simple nucleosynthetic scenario covers many of the observed light element variations: very high temperature (up to 80 million degrees Kelvin) hydrogen fusion cycles. At the same high interior temperatures in which O can be converted to N in the ON-cycle hydrogen burning, spectroscopically undetectable) Ne can and will be turned into Na The anticorrelation between O and Na therefore occurs quite naturally, and the difference between high O, low Na stars and low O, high Na stars is that the latter stars have had a combination of higher interior temperatures and better convective envelope mixing than have the former stars. But the high central temperatures needed to make this idea work are daunting. Recent studies focus instead on the idea that earlier generations of cluster stars (now dead) seeded the intra-cluster medium with their elemental output, and the presently observed stars gathered in this material at birth. Please see, for example, Carretta et al. (2009) and the review by Carretta et al. (2009) for an extended discussion of these issues.

Support for the multiple-population idea has come from the identification in recent years of several globular clusters with multiple color-magnitude diagram sequences. These clearly indicate age/metallicity/abundance spreads in at least these clusters. Particularly intriguing is the case of M22. Not only is a split red giant branch (RGB) obvious in the figure below, but the discriminants in the color-magnitude diagram appear to be related to two chemically-distinct populations: a relatively metal-poor group (mean [Fe/H]=-1.82) that has neutron-capture element abundances characteristic of r-process dominance, and a less metal-poor group (mean [Fe/H]=-1.67) that exhibits of significant enhancements of carbon and s-process abundances. These results are described in detail by Marino et al. (2012) and references therein.

Chris Sneden

Red giant branch (RGB) sequences in M22 Left panel: I versus the Str¨omgren index m1. Stars belonging to the two RGBs have been represented in red and blue colours. Right panel: Stars selected in the double RGB of the I-m1 diagram, have been represented in the U-(U-B) CMD. (from Marino et al. 2012). Click on picture to see larger image


All of the research described here has resulted from the collective efforts of large numbers of people, including my current and former students, and investigators at many astronomical institutions around the world. Their contributions are gratefully acknowledged. For many years I have benefited from grants by the U.S. National Science Foundation and the National Aeronautics and Space Administration. Their support has been crucial to my research.