Chapter 1: The Elements in the Solar System

Modern evidence leads to the insight that stars and their planetary systems form by gravitational collapse of interstellar molecular clouds. Much of the chemistry in our solar system is governed by the original element inventory that the solar system inherited from its presolar molecular cloud about 4.6 billion years ago.

Molecular clouds are cosmic recycling bins for the elements produced in many stars from different generations. The big-bang endowed the universe with H, D, He (both isotopes ³He and ⁴He) and a dash of Li (mainly ⁷Li). Hydrogen and He serve as major nuclear fuel in stars. The light elements Li, Be and B have particular histories of their own (see below) but all other elements came into being through stellar nucleosynthesis over time.

Stars like our Sun and ones that are more massive do not exist forever. In its final evolutionary stage, a star returns most of its mass – including mass in the form of freshly synthesized elements – back to the interstellar medium. The efficiency and yields for element production depends on the initial mass of the star, which also determines whether the newly produced elements are released through a stellar explosion as a supernova, or by less violent stellar winds. The result is that the stellar “ashes” can become part of the molecular clouds in the interstellar medium from which new generations of stars can rise.

Element production in generations of stars has been ongoing since the time the universe formed some 14 billion years ago. It is still ongoing in the stars that we can see in our galaxy and in stars of other galaxies dispersed in the universe. There is no need to worry that stars will disappear soon because there is no H and He left as initial stellar fuel. On a universal scale, H and He remain the most abundant elements. At the Sun's birthplace in our galaxy, the mass fraction of all elements heavier than helium had increased only to about 1.5% between the time when element production in stars of our Milky Way Galaxy started and the solar system formed. The rather low abundance of all heavy elements compared to H and He is one reason why astronomers collectively call all elements heavier than helium “metals.” This definition of “metals” is not the traditional characterization that one normally associates with metals. On the other hand, there is the natural bias from daily life on a planet that essentially only consists of these “metals.” Typically, there are no daily concerns about the fact that most matter in the Universe was, is and will be so for quite some time H and He. We refrain here from using the astronomer's definition of metals too much to avoid confusion.

The investigations of what elements exist and what their abundances are went hand in hand. Already by 1847, the French geologist Élie de Beaumont (1798–1874) assembled a list (Figure 1.1) for the occurrences of the elements known at the time.¹  He listed the elements in order of increasing electro-negativity as suggested by Berzelius. Among the elements in his list, three entries seem unusual from the current point of view: glucinium is the old designation for Be, didymium turned to be a mixture of Pr and Nd, and pelopium, thought to be a new element found in the mineral columbite, was later shown to be impure columbium (=Nb). In his first entry column, De Beaumont used stars to mark the 16 elements that Henry de la Bèche found to be the most widely distributed over the surface of the Earth. In subsequent columns, he indicated which elements have been found in modern and ancient volcanoes, basic rock, granites, stanniferous veins, normal ore veins and geodes, mineral waters, volcanic emanations, native metals, meteorites, and organic matter. De Beaumont emphasizes that out of the 16 most abundant elements on the Earth's surface, 15 are also those that Angelot, whose data de Beaumont included in his table, had found in meteorites. In the last column of his table, de Beaumont indicates the general occurrence of the elements in “organized bodies,” or biotic matter. He observes:¹ 

De Beaumont put his table together some time before papers on the periodic table began to appear. Between 1863 and 1866, John Alexander Reina Newlands published a series of papers on the periodic relations of the elements with atomic weights. In 1869, Lothar Meyer and Dmitri Mendeleev published their notes on the periodic table. Only Mendeleev²  boldly predicted the existence of several “missing” elements, which were subsequently discovered and bestowed with patriotic names (e.g., Ga, Ge, Sc). The subject about the element occurrences and their relative abundances continued. Knowing the abundances and distribution of the elements would shed clues on the basic make-up and origins of matter. Important issues were to find the representative elemental compositions for the Earth, the Sun and the cosmos. The searches and discoveries of the missing elements in Mendeleev's periodic table involved mainly analyses of terrestrial materials. The development of optical spectral analysis by R. Bunsen and G. Kirchhoff in the early 1860s made it possible to access elements of low abundance and to quantitatively determine their concentrations in different minerals and rocks. After 1913, when H. G. J. Mosley provided the theoretical understanding for X-ray spectra, X-ray spectroscopy became another valuable analytical tool. Starting in 1919, Aston's developments of the first mass-spectrometers began to reveal the isotopic nature of the elements. The advances in micro-analytical instrumentation since the 1950s, such as the electron microprobe, neutron activation analysis and gamma-ray spectroscopy, led to a wealth of elemental and isotopic data from terrestrial, lunar and meteoritic rocks. More recently, ion-probe mass-spectrometric methods permit the analyses of the elemental and isotopic compositions on samples with micro- to nano-scale resolution that even include genuine mineral particles that formed around other stars and are hidden in meteorites.

Spectroscopy also advanced the discovery and quantitative assessment of the elements beyond the Earth. Starting in the 1860s, the analyses of the Fraunhofer absorption lines in solar photospheric spectrum, and spectral analysis of other stars, interstellar nebulae, and of comets soon revealed that the occurrence of the chemical elements is not restricted to the Earth. Analyses of meteorites, already recognized as extra-terrestrial materials by E. F. Chaldni in 1794, showed similar results. The same chemical elements as found on the Earth constitute the normal matter in other objects in the solar system and beyond in the stars. There were no stable elements in the Sun, other stars, in comets or meteorites that could not be found naturally occurring on Earth as well. Even the well-known case of He, discovered by Janssen and Lockyer in the solar spectrum in 1868, was no exception since its detection on Earth followed by Cleve and Langlet in 1895 and, independently, by Ramsey at the same time.

The story is a little different for the “missing” elements Tc (first called masurium) and Pm. Unlike the elements beyond atomic number 83 (Bi), Tc and Pm have their place between stable elements in the periodic table. All Tc and Pm isotopes are unstable and undergo radioactive decay; the half-lives of the longest-lived isotope are 4.2 million years for ⁹⁸Tc and 17.7 years for ¹⁴⁵Pm. On Earth, these elements were first known as products of artificial nuclear reactions, which succeeded in producing Tc in 1937 and Pm in 1947. These elements only occur naturally in stars that make them. In 1952, Merrill³  found Tc in several red giant stars, and Aller and Cowley likely detected Pm in a chemically peculiar star in 1970, as seems to be confirmed recently.⁴ 

Merrill's discovery of Tc in stars showed that element synthesis does happen in stars, and it came at a time when models of element synthesis in stars had gotten their first foundations. The discovery of the neutron by J. Chadwick in 1932 facilitated the understanding of element synthesis, and nucleosynthesis models started to evolve in the late 1930s, when H. A. Bethe⁵  and C. F. von Weizäcker⁶  independently proposed models how thermonuclear fusion reactions convert H into He using C, N and O as catalysts in stars. The Bethe–Weizsäcker cycle, or CNO cycle, is one of the processes for H fusion in the Sun, the other is fusion through the proton–proton reaction chains where D (just discovered by H. C. Urey in 1932) occurs as an interim product.

Henry N. Russell carried out the first quantitative analysis of the elements in the Sun in 1929.⁷  V. M. Goldschmidt's classical papers on elemental and isotopic distributions appeared in 1937⁸  where he devised basic geochemical principles that govern the observed abundances in terrestrial and meteoritic rocks. The knowledge about the abundances as well as the isotopic composition of the elements had become fairly detailed by 1956, when H. E. Suess and H. C. Urey did their classical work on cosmic abundances.⁹  In concert with abundance determinations, Alpher, Bethe and Gamov as well as ter Haar and Salpeter as well as many other researchers had provided more important groundwork on the theory of element synthesis through the 1940s and early 1950.¹⁰  In 1957, comprehensive nucleosynthesis models by E. M. Burbidge, G. R. Burbidge, W. A. Fowler, and F. Hoyle¹¹ (“B²FH”) and, independently, A. G. W. Cameron¹²  appeared to explain the synthesis of the heavy elements. Today, the refined but still growing knowledge of the elemental and isotopic abundances in our solar system, in stars of our Galaxy and beyond continues to inspire nucleosynthesis and galactic chemical evolution models, and the abundances of the elements of the solar system remain a critical test for these models.

In 1885, when most known elements populated the periodic table, the Russian scientist I. A. Kleiber published a paper on the chemical composition of celestial bodies.¹³ Figure 1.2 shows his qualitative synthesis for the general composition of celestial objects in the form of a plane periodic system following atomic numbers. This may be the first published “cosmochemical” periodic table of the elements. Note that the group for the noble gases was not yet included. Kleiber searched for periodic trends in elemental abundances in meteorites, the Sun, comets, fixed stars and meteors. He concluded that the composition of cosmic bodies is not the result of some random contribution of the elements.

Like Mendeleyev and others at the time, Kleiber noticed that elements with low atomic weights up to the iron group are more abundant than the heavier ones, with the notable exception of B. In particular, he noticed that the iron group elements are abundant in cosmic objects whereas the Pt-group elements are not, which suggested that the elements with similar chemistry are not necessarily similar in abundance.

The first comprehensive report on a quantitative determination of the relative abundances of the elements dates back to 1889 and was performed for the Earth's crust.¹⁴  Frank Wigglesworth Clarke investigated the trend of elemental abundances in the Earth's crust as a function of their atomic weights. With abundant accessible samples for chemical analyses, the rocks of the Earth's crust were a useful starting point to determine the relative elemental abundances, which might reveal something about the origins of the elements, aside of course from the economic interest of knowing the crust's elemental inventory. However, Clarke was somewhat frustrated not to find discernable elemental abundance trends. He was able to refine the conclusion that the lighter elements up to Fe are much more abundant than the heavier elements, but not much more that could shed some light on the causes for the observed element abundances.

In any case, the composition of the ∼5–50 km thick crust of the Earth cannot be representative of the composition of the entire Earth, let alone the composition of the solar system. Most of the Earth’ mass is in its silicate mantle and the core, and the crust is only about 0.4% of Earth's total mass. As a whole, the Earth mainly consists of the elements O, Mg, Si and Fe, and the other, less abundant elements in the Earth are distributed between the Fe-rich metallic core, the Mg- and Fe-silicate mantle rocks, the crust, oceans and atmosphere according to their geochemical affinities.

By 1917, several good meteorite analyses were available, and the relatively primitive nature of meteorites in contrast to the complex history of planetary rocks had already been established. Following Farrington's 1915¹⁵  suggestion that the mean composition of meteorites may resemble that of the Earth as a whole, William D. Harkins (1873–1951) averaged the composition of different stony and iron meteorites to obtain such a representative composition for the Earth.¹⁶  Then he plotted the abundance as a function of atomic number to make a fundamental discovery. The meteoritic abundances showed systematic trends with atomic number, quite different from those in the Earth's crust. Figure 1.3 is a redrawn version of his plot, which was the first illustration of the odd-even elemental abundance trend with atomic number. This trend is now known as “Harkins’ rule.”

Harkins¹⁶  describes his observations with regard to the elemental abundances in meteorites: … the elements with even atomic numbers “are in every case more abundant than their adjacent odd-numbered elements.” Harkins found that the abundant elements Mg, Si, Ca, Fe and Ni “do not only have even atomic numbers, but in addition they make up 98.6% of the material in meteorites.” He observed that odd-numbered elements with higher abundances such as Al and Co are in “between two extremely abundant even-numbered elements.”

His diagram is limited to the abundant elements from C to the Fe group for which abundances were easier to determine than for the heavier, less abundant elements. In 1917, not all the rare earth elements had been discovered and their abundances in terrestrial rocks and meteorites were sketchy at best. However, from the data available, Harkins concluded that the abundance trend in the relative odd-even abundances should apply to the lanthanides as well, which was nicely confirmed by later analyses (Figure 1.4).

Looking at meteorites and terrestrial rocks Harkins found that the abundant elements are restricted to atomic numbers between 6(C) and 30(Zn), similar to Clarke's finding that these are the most abundant elements in the Earth's crust. The absence of volatile H and He in rocks is not too surprising; however, Harkins independently repeats Kleiber's observation of the relative paucity of the light elements Li, Be and B, as well as that of the heavy elements beyond the Fe group in meteorites. Explanations for these observations and Harkin's conclusion that “in the evolution of the elements, more material has gone into the even-numbered elements than those which are odd”¹⁶  only became possible when more analytical data for solar system abundances became available and theories on nucleosynthesis evolved in the 1930–1950s.

Most of the mass, 99.86%, in our solar system resides in the Sun. Thus, the composition of the Sun should provide a good average of the element inventory that the solar system inherited from its parental molecular cloud. The gas-giant planets Jupiter, Saturn, Uranus and Neptune are the next most massive objects and make up ∼0.13% of the total mass in the solar system (Figure 1.5). The terrestrial planets Mercury, Venus, Earth and Mars contribute ∼0.0009% to the solar system's mass; the Earth alone makes up ∼0.0003%. The dwarf planets Ceres, Pluto and Eris make negligible fractions of the total solar system mass.

Ultimately, the cloud material also provided the elemental starting composition for the planets and the differences in the composition of the Sun and the planets as we see them today must reflect the chemical and physical element fractionation processes during planet formation and subsequent planetary evolution. These fractionations make it challenging to derive the overall chemical composition of the solar system from terrestrial crust and mantle rocks. Instead, it is more practical to compare planetary compositions to the solar system's composition to decipher how the element inventories of the planets were established and what differentiation processes occurred. In theory, it would be ideal to have good quantitative analyses of the elemental abundances in the Sun. In practice, this is currently only possible for a subset of the elements in the Sun's photosphere (see below).

The composition of meteorites provides a second source of information for solar system abundances. Most meteorites do not show signs of large-scale differentiation of silicates and metal as planetary objects do and it was recognized early that meteorites are the closest samples of the original, “world making” material (e.g., Merrill, 1909¹⁷ ). It is not possible to use the meteorite compositions to constrain the abundances of the noble gases or the abundances of elements like H, C, N and O, which form volatile compounds. However, for the “non-volatile” or “rocky” elements, meteorite analysis are usually more precise than photospheric measurements, and for several elements meteorites are the sole source of information on solar system elemental abundances.

As described in a separate chapter, there are compositionally different groups of meteorites. Consequently, one problem had been how the analyses of the different meteorites should be treated to derive representative abundances for the non-volatile elements in the solar system. A related issue is if there is a single, particular group of meteorites that can serve as an abundance standard.

Neither the solar photosphere nor meteorites can provide reasonable abundance estimates for some elements like the noble gases. In such cases, one can make use of the fact that most of the other normal dwarf stars in the H-burning stage have very similar relative abundances as the Sun. The younger, more massive and hotter B dwarf stars are particularly useful to derive “missing” solar elemental abundances and to check the solar abundance determinations for elements up to Fe. Yet another source for supplementing and checking the solar abundances is the analysis of interstellar diffuse nebulae such as the so-called HII regions (the notation HII indicates that ionized H dominates in contrast to HI regions with mainly neutral H). These regions are often associated with young stars, and the composition of these regions should be similar to the young B stars nearby. The elemental abundances from absorption spectroscopy of evolved giant stars or from emission spectroscopy of their descendants, the planetary nebulae, are useful for complementing the solar abundance data. (The name planetary nebulae has nothing to do with planetary systems. Through early telescopes, these dust and gas enshrouded stellar remnants appeared as “fuzzy” as the planets.) However, their usefulness to constrain solar abundances is limited because certain elements are among the giant stars’ interior nucleosynthesis products that are eventually mixed to the stellar surfaces.

In the following, “meteoritic” or “CI-chondrite” abundances refer to elemental abundances from type CI-carbonaceous chondrites, “photospheric” abundances refer to abundance determinations of the present solar photosphere and “proto-solar” or “solar-system abundances” refer to elemental abundances of the proto-sun at the time of its formation.

For completeness, it needs to be mentioned that the solar abundances occasionally were called “cosmic” abundances in the older literature. While the solar system abundances derived from the solar photosphere and the meteorites only strictly apply to the solar system, these abundances are quite similar to those found in other stars like the Sun. However, there are systematic variations in the abundances of the heavy elements (the astronomers’“metals”) in normal stars as a function of radial distance from the galactic center. These variations do not warrant making the solar system abundances representative for “cosmic” abundances.

There are two widely used relative atomic abundance scales for the elements. The astronomical abundances scale is fixed to the abundance of the most abundant element in the photosphere, hydrogen, at ε(H)=10¹² atoms. Since abundances vary over twelve orders of magnitude, this relative scale avoids dealing with negative exponents in the abundances. It is often more convenient to use the decadic logarithm for the abundances so that A(H)=log εH=12, or, for any element “El,” we have A(El)=log(εEl/εH)+12, which is the notation frequently encountered in the astronomical literature.

Geochemists and cosmochemists prefer to use a linear atomic abundance scale fixed to a silicon abundance of N(Si)=10⁶ atoms. The cosmochemical abundance scale is more convenient when dealing with planetary and meteoritic compositions because Si is a major element in rocks. The linear scale is still practical because the most abundant elements in the Sun H, He, Ne, Ar, C, N and, to some extent, O and their compounds are volatile and therefore typically less abundant in rocky planets and meteorites.

Uncertainties of photospheric and meteoritic abundance determinations are compared using the relationship U(%)=±100(10^±a– 1) where “a” is the uncertainty in dex-units quoted for abundances on the logarithmic scale and “U” is the uncertainty on the linear scale in percent. The uncertainty in logarithmic units (“dex”) is an uncertainty factor, hence the uncertainty in percent is smaller for −a than for +a, or vice versa, a given percent uncertainty yields two different uncertainty factors.

In principle, the best average for the solar system's composition would come from the analysis of the solar photosphere because the Sun contains most of the mass of the solar system. In practice, there were, and still are, various technical difficulties that prevent well-defined quantitative analyses for all elements.

The photospheric abundances are derived from the Fraunhofer lines of the solar absorption spectrum, which samples the wavelength range from the near-UV to the IR. The Fraunhofer lines mainly form in the photosphere and in the overlying photosphere-chromosphere transition region. Sixty-eight elements have been detected in the photosphere, and they are mainly present as monatomic or singly ionized ions. The quantitative determination of the photospheric abundances by spectroscopy is more elaborate than spectrochemical analysis in the laboratory, where samples and well-known standards can be measured under the same conditions. To derive quantitative abundances from the photospheric spectrum, it is necessary to know the atomic properties of the elements and the physical conditions in the solar photosphere. The information necessary and the challenges to derive quantitative abundances from stellar spectroscopy include, but are not limited to: identifications of element line positions in the spectra; presence or absence of suitable lines for certain elements in given spectral range or line blending in the measured spectra; knowledge of atomic transition probabilities and oscillator strengths (the f values); the sufficiency of the line broadening theories; and the adequacy of atmospheric structure models to describe the temperature and total pressure with depth in the atmosphere where the absorption lines originate. Finally, a still much-discussed issue is whether deviations of the excitation and ionization conditions as given by the Boltzmann and Saha equations for the local kinetic temperature (typically referred to as “local thermodynamic equilibrium” or LTE) must be considered in the abundance determinations.

The first comprehensive determination of the solar photospheric abundances by Russell in 1929⁷ for 56 elements has seen many important revisions and updates over the years. Table 1.1 summarizes current analytical data. To date, out of the 83 elements that naturally occur in the solar system (all stable elements plus Th and U) the abundances for 68 elements have been measured in the Sun with varying degrees of accuracy. Most data are from photospheric analysis, and the abundances of F, Cl and Tl are actually derived from sun spot spectra.

The noble gas abundances cannot be directly measured in the solar photosphere because suitable absorption lines are lacking. Their “photospheric” abundances are determined indirectly, which is the reason why their elemental abundances in Table 1.1 are placed in square brackets. Although He is detected in the Sun, the He abundance is mainly based on helioseismic measurements, solar interior and evolution models, or on measurements of He/H abundances in other stars. The Ne and Ar abundances are estimated from the composition of the solar corpuscular radiation (solar wind and solar energetic particles), and from Ne and Ar measurements in hotter stars, where Ne and Ar lines are accessible. The Kr and Xe abundances are quite low and their abundances are usually derived from interpolations of neighboring element abundances and nucleosynthetic arguments.

The present-day elemental abundances in the outer solar atmosphere are not affected by the ongoing H-burning nucleosynthesis in the Sun's core because there is a non-convective layer between the core and the convective layer starting right below the photosphere. However, relative to H, the heavy elements can diffusively settle from the photosphere, which needs to be considered when current photospheric abundances are used to derive the proto-solar abundances (see below).

The so-called CI-chondrites are a rare group of chondritic meteorites, which have become an important reference standard for solar system elemental abundances. More about meteorites is given in a subsequent chapter; however, because of their use as abundance standards, the data for CI chondrites are given in Table 1.1 for comparison.

Table 1.1 gives the photospheric and CI-chondrite abundances on the astronomical (normalized to log εH=12) and cosmochemical (normalized to Si=10⁶ atoms) scales.¹⁸  One important issue is how these different relative scales are linked. The problem is that the relative H abundance in meteorites is much less than in the photosphere, but the astronomical scale is tied to element/H ratios in the photosphere. However, since H abundances in meteorites are low, their element/H ratios are much higher than those in the photosphere. On the other hand, it is not a problem to compare the photospheric abundances in their usual logarithmic notation in the astronomical scale to the meteorite data on the linear cosmochemical Si-based scale. For several elements, there is a more or less constant difference when the logarithmic meteoritic values are subtracted from the photospheric values.¹⁸  The average difference (which corresponds to a factor if the scales were taken linear in the comparison) from about 35 well-determined elements is 1.533, so that the abundances on the astronomical scale (log εEl) are related to the abundances on the cosmochemical scale as: log N(El)=log εEl−1.533.

With this relation, the meteoritic abundances are converted into the astronomical scale that is normalized to the photospheric H abundance of log εH=12. Obviously, the meteoritic H abundances from the relation are less than this, because the element/H ratios are higher in the meteorites.

One could also simply re-normalize the photospheric data to Si, but this introduces uncertainty in linking the scales, because it becomes solely tied to the correctness and the quality of the Si abundances in the photosphere and in meteorites. A small change in the Si abundances from new measurements in either the photosphere or meteorites then would require re-computing the linked abundance scales. Except for historical sentiment to use Si for normalizing the cosmochemical abundance scale, there is no reason why honor is given solely to Si to link the scales. One may pick any other rocky, non-volatile element that should be fully retained in the meteorites and that is measured in the Sun. In the past, when photospheric abundance determinations were less certain, it was more reliable to calculate an average scaling factor from several elements that are well determined. The relative abundances of 35 elements agree within 10% in the photosphere and CI-chondrites, and their individual scaling factors give an average of 1.533 for linking the astronomical and cosmochemical abundance scales.

Figure 1.6 shows a plot of the photospheric versus the CI-chondritic abundances on the cosmochemical abundance scale. In most cases, the uncertainties in the abundances are smaller than the element symbols shown, but error bars are considerable for several element determinations in the Sun (Table 1.1). If the relative abundances were identical in the solar photosphere and CI-chondrites, the abundances would plot on the dotted 1 : 1 reference line. The agreement is within 10%(15%) for 31 (41) elements. This correspondence justifies the use of the CI-chondrites as an abundance standard, and to give preference to the meteoritic values of, for example, F, Cl, Ga, Rb, Ag, Cd, Au and Tl over the photospheric values because of their often-higher analytical precision. As a corollary, the CI-chondrite abundances are a useful abundance standard of elements that cannot be measured well at all in the solar photosphere (e.g., As, Se, Br, Te, I, Cs, Ta, Re, Bi, Th, U).

When compared to the photospheric abundances, the relative abundances of H, C, N, O and the noble gases are lower in CI-chondrites by several orders of magnitude, which indicates incomplete retention of volatile compounds in meteorites. Figure 1.6 shows that the Ne, Ar, Kr and Xe abundances in CI-chondrites follow a depletion trend that broadly anti-correlates with their atomic masses. Only one element, Li, has a clearly higher relative abundance in CI-chondrites than in the photosphere. The abundance of Li is well determined in both sources, and the only possible explanation is that Li was destroyed in the Sun.

Representative solar system abundances for D, Li and, possibly, Be and B cannot be derived from the Sun's photosphere. Although this section focuses on the elemental abundances, we include D here together with Li, Be and B because of their importance in cosmological models.

Deuterium and the nuclides of Li, Be and B are not produced in any significant amounts through stellar nucleosynthesis; instead, they are more likely to be destroyed in stars because of their low nuclear binding energies. Deuterium and ⁷Li were mainly produced through big-bang nucleosynthesis, and their abundances are decreased through astration (destruction in stars) over time. The ⁷Li abundance is continuously supplemented by cosmic ray spallation in the interstellar medium, which happens when heavier elements in the interstellar medium encounter bombardments from so-called cosmic rays. These energetic ions [mainly protons and α(=He²⁺) particles, but also ions of all other elements and secondary neutrons] break-up abundant C, N and O to yield the lighter element as fragments. Cosmic ray spallation is the major production mechanism of ⁶Li, Be and B. Lithium (⁷Li) is also produced in certain giant stars, so that Li is an element that has at least three different nucleosynthetic production origins. Another suggested source of the light elements is neutrino-induced nucleosynthesis in core-collapse supernovae.

Destruction of deuterium in stars is unavoidable, and happened in the Sun even before H-burning through the proton–proton chain was ignited. Within its about first million years of age, the temperatures and densities in the contracting proto-sun became favorable for D burning, which requires >0.6×10⁶ K. The young Sun was fully convective during that time, and the Sun's atmospheric element inventory could be cycled through interior regions where temperatures were sufficiently high to destroy essentially all D within about 1–2×10⁵ yr. Deuterium burning produces ³He, which increased the ³He/⁴He in the Sun, and thus the current solar ³He/⁴He as derived from the solar wind is also not representative of the proto-solar ratio.

Deuterium is essentially gone from the photosphere, but D is present in the solar wind, as seen from the analysis of the solar wind that was captured in the Al and Pt metal foils posted on the lunar surface during the Apollo missions. The D in the solar wind is a product of local spallation reactions in solar flares and therefore it is not representative of the original D inventory of the solar system. However, the overall production of D in flares of other stars is relatively small and cannot balance the loss of primordial D in stellar interiors.¹⁹  Thus, the original amount of D from the big-bang can only decrease over time and the question is how to find the D abundance that was available when the solar system formed.

Figure 1.7 shows the D/H ratios that have been determined in various settings. The primordial D/H (∼2.6×10⁻⁵) is well constrained from the cosmic microwave background measurements by the Wilkinson microwave anisotropy probe (WMAP)²⁰  and determinations of the baryon-to-photon ratios and cosmological models.^21–23  The gas in the current local interstellar medium has a low value of ∼1.5×10⁻⁵, which either reflects the depletion of D by astration since the formation of the Universe or is simply due to the preferred incorporation of D into organic and icy interstellar dust.^23,24  If the estimate for the D/H in the nearby interstellar gas and dust of ≥2.3×10⁻⁵ is correct, the decrease over time was rather modest (<12%). Within these boundaries, the initial solar system D/H should have been between 2.3 and 2.6×10⁻⁵. The D/H ratios measured in the atmospheres of Jupiter and Saturn²⁵  span this range in D/H (Figure 1.7). Jupiter, the second most massive and, like the Sun, H and He-rich object in the solar system, is probably the least altered source for the original solar system D/H ratio. Since a large portion of Jupiter is gravitationally captured gas from the solar nebula, its atmosphere should have preserved the record of the D/H ratio in the early solar system. In that case, the initial solar system D/H ratio is 2.5×10⁻⁵, as given by the average values measured in molecular H₂ by the Galileo spacecraft entry probe (2.6×10⁻⁵) and the D/H=2.4×10⁻⁵ for Jovian H₂ as determined from spectra with the ISO satellite. Within the larger uncertainties, the same applies to the D/H obtained for Saturn's atmosphere.²⁵ 

The D/H ratios for the outer giant planets, Uranus and Neptune,²⁵  are larger than the primordial and solar system D/H ratios. These giant planets contain less H and He than Jupiter and Saturn, which shows that they accumulated more solid organics and ices than solar nebula gas. The solids in the outer solar nebula were enriched in D/H, as seen in the D/H measurements from comets,²⁶  which would explain the larger D/H in Uranus and Neptune. Different meteorite groups and their components²⁷  are also enriched in D/H. The quite variable D/H ratios in the organic and hydrous phases in meteorites may in part reflect the enrichments in D and/or fractionations in D/H by ion–molecule processes that operated in the solar nebula. Back-reactions of D rich phases with relatively D-poor gas also must be considered to explain the range in meteoritic D/H ratios. The ion–molecule processes that can have occurred in the outer solar nebula are similar to those still operating in the interstellar medium, and especially to those in cold molecular clouds where quite large D/H ratios are observed. The meteorite data give weight to the notion that a lot of D in the interstellar medium today is hidden in dust grains because the ion–molecule reactions favor incorporation of D into solids. In that case, the D/H from gas phase measurements in the ISM are only lower limits to the “true” D/H ratio in the ISM.

The relatively high D/H ratios in water in the atmospheres of Venus (∼2.2 ×10⁻²) and Mars (8.1×10⁻⁴), and in the terrestrial oceans (1.56 ×10⁻⁴), cannot be representative for the original D/H ratio of the solar system. Whatever the original D/H in the terrestrial planets may have been, their currently observed D/H ratios are very likely higher than the original values because hydrodynamic escape of H is favored over D escape from all terrestrial planets.

The present-day abundance of lithium in the photosphere is ∼150-times less than that in chondrites. Within the relatively large uncertainties associated with their abundances, the relative abundances of beryllium and boron are similar in the photosphere and CI-chondrites. The nuclides of the light elements Li, Be and B are relatively fragile like D, but they require 3–8-times higher fusion temperatures than D. The required temperatures for fusion with H are lowest for Li, and highest for B; about 2×10⁶ K for ⁶Li, 2.5×10⁶ K for ⁷Li, 3.5×10⁶ K for ⁹Be and about 5×10⁶ K for ¹⁰B and ¹¹B. Complete loss of D, Li, Be and B could only occur if the temperatures increased 3–8 times above the temperatures needed for D-burning and if the Sun remained fully convective to allow processing of Li, Be and B in the entire atmosphere at the necessary temperatures. However, the relative solar Li abundance is about 150 times less than in CI-chondrites, but the Be and B abundances are about the same within uncertainties. The reason why these elements are preserved in the photosphere is that the development of a radiative zone over the core prevented full convection of the outer envelope that starts right below the photosphere. The radiative zone formed as temperatures and densities increased in the contracting Sun. As the radiative zone over the core widened, the bottom of the overlying convective zone moved up towards to lower temperatures. As the convective envelope no longer reached down to temperatures necessary for fusion, the Li, Be and B remained in the convective zone. Thus, the photospheric abundances Li, Be and B are useful diagnostics for the depth of the Sun's convective layer. However, to use them as such, the original solar Li, Be and B abundances must be known, which brings us back to the meteoritic abundances from CI-chondrites as standards for the solar system abundances.

The average solar system elemental abundances are derived from the analytical data for the solar photospheric and CI-chondrites given in Table 1.2. For the elements that have similar abundances in both sources, one can take either the average or the datum with higher analytical precision. As described above, the abundances of the volatile elements H, C, N, O and noble gases are from solar analysis or other sources than meteorites whereas abundances of Li, Be, B and elements not accessible from the Sun are based on the meteoritic data.

In deriving the proto-solar abundances from the selected meteoritic and photospheric value, one other important aspect has to be considered. Models of the Sun's evolution and interior show that currently observed photospheric abundances relative to H must be lower than those of the proto-sun because He and other heavy elements have settled from the photosphere towards the Sun's interior since the Sun formed ∼4.6 Ga ago. Therefore, the current photospheric abundances relative to H are not representative of the solar system and only the proto-solar (i.e., un-fractionated with respect to hydrogen) abundances represent the “solar system elemental abundances.”

The abundances of elements heavier than He apparently did not fractionate relative to each other because the relative abundances of many rock-forming elements in the photosphere are the same as in CI-chondrites (i.e., the abundances relative to Si are the same). However, the heavy elements are fractionated relative to H and all element/H ratios decreased. Over the Sun's age, the photosphere “lost”∼16% of the heavy elements and the He abundance decreased by ∼18% relative to H. This heavy element settling from the photosphere is taken into account when the abundances from Table 1.1 are used to derive the solar system abundances listed in Table 1.2. On the astronomical scale, the proto-solar abundances for elements heavier than He are 0.074 log-units higher than the photospheric values.

However, on the cosmochemical scale by number the solar system abundances of all elements, except for H and He, are the same as the photospheric abundances. This is the obvious outcome for element normalization to Si=10⁶ atoms. The proto-solar H abundance is only ∼0.84 times that of the photosphere, while the respective He abundance is ∼1.02 times photospheric. The proto-solar He abundance must appear slightly higher than unity on this scale because He settling from the outer layers of the sun was slightly more efficient than that of the heavy elements, including Si, which is used for normalization of the cosmochemical abundance scale. The difference in proto-solar and photospheric abundances thus expresses itself either by a higher heavy element (the astronomers’“metals”) content of the proto-sun when the H-normalized astronomical abundance scale is used or by a relative depletion in hydrogen on the Si-normalized cosmochemical scale.

The solar system elemental abundances from Table 1.2 as a function of atomic number are shown in Figure 1.8. This diagram with abundances of all stable elements plus Th and U makes an interesting comparison to Figure 1.3 with the limited data set of the element abundances available to Harkins.

The abundance trends already noticed by Harkins are clearly visible in Figure 1.8: hydrogen and He dominate and overall, abundances broadly decrease with atomic number. Major exceptions from this trend are the low abundances of Li, Be and B, and the higher abundances around Fe. Elements with even atomic numbers are more abundant than their odd-numbered neighbors. More detailed trends in the elemental abundance distributions are revealed when the abundances of the even and odd-numbered nuclides of the elements are distinguished and plotted by mass number instead of atomic number (Figure 1.9).

Figures 1.8 and 1.9 show the slight increases in abundances in regions around Ge-Sr; Xe-Ba, the rare earth elements, Os-Pt, and Pb. These abundance peaks are consequences of the larger stabilities of the isotopes of these elements as they often contain so-called magic numbers of protons and/or neutrons. The magic numbers (2, 8, 20, 28, 50, 82 and 126) correspond to closed nuclear “shells,” comparable to the closed electron shell configuration of the noble gases.

The isotopic composition of the elements and a breakdown of the solar system elemental abundances (from Table 1.2) into the isotopic abundances are given in Appendix A. Here we cannot discuss the nuclide abundance distributions as a function with mass number (instead of the elemental abundances as a function of atomic number) in any detail and only point out a few issues about the isotopes that may help to explain the elemental and nuclide abundance distributions. A survey of the 266 stable nuclides of the elements (out of 280 naturally occurring stable and long-lived nuclides in the solar system) shows the following frequency of nuclides with proton number (=atomic number Z) and neutron number N:

• Z even, N even: 159 nuclides

• Z even, N odd: 53 nuclides

• Z odd, N even: 50 nuclides

• Z odd, N odd: 4 nuclides: ²H, ⁶Li, ¹⁰B, ¹⁴N

The lower number of stable nuclides available among the odd-numbered elements contributes to the fact that these elements are less abundant than the even numbered elements. Among the odd-Z elements, 22 have one stable isotope. Two of them, V and Ta, have an additional very long-lived radioactive isotope. Several of these 22 elements have particularly low abundances, as can be seen from Figures 1.8 and 1.9. The 20 solely mono-isotopic elements are Be, F, Na, Al, P, Sc, Mn, Co, As, Y, Nb, Rh, I, Cs, Pr, Tb, Ho, Tm, Au and Bi. The elements V and Ta each have abundance contributions from a long-lived second isotope (⁵⁰V with a half-life of 1.4×10¹⁷ years, and ¹⁸⁰Ta with >1.2×10¹⁵ years), which for practical purposes can be regarded as stable.

Some of the longer-lived naturally occurring radioactive nuclides (see Chapter 3), for example, ⁴⁰K (Table 1.3), have odd proton and neutron numbers, most of them have odd Z and even N, but none of these long-lived radionuclides are combinations of even numbers of protons and even numbers of neutrons.

Conversely, stable nuclides with even proton and neutron numbers are favored in frequency and abundance. Elements with even atomic numbers typically have more than one isotope. The extreme case is Sn with ten stable isotopes (Sn also has a magic number of Z=50 protons). The occurrence of several isotopes for even-Z elements contributes to the larger abundances of the even numbered elements seen in Figures 1.8 and 1.9.

These simple comparisons show that nuclear properties influence the abundances of the elements. Detailed studies of the abundances of the nuclides as a function of mass number have led to the establishment of detailed rules and explanations for the regularities in the abundance distributions of the stable elements.^9,28,29 

The issue here is that the nuclear make-up and the stabilities of the nuclides, as well as timing of solar system formation, controlled the overall abundances of the elements in the solar system. The overall abundances are not controlled by the elements’ electron shell structures, which determine their chemical behavior and position in the periodic table.

The reason for the nuclear control of the elemental abundances is that nuclear properties ultimately determine the yields during stellar element synthesis. The principal nucleosynthesis origins of the elements are indicated in Figure 1.8 by the dashed lines. Briefly, for the lightest elements, “BB” indicates primordial production through the big-bang and “X” indicates production through spallation reactions (see discussion for D, Li, Be and B above). Hydrogen fusion to He through the proton–proton chain and CNO cycle is the main occupation of stars for most of their lives. (About 10 billion years for the Sun, which has converted about 0.3% of its H into He in the past ∼4.6 billion years.) In low and intermediate mass stars (up to ∼8 solar masses), the fusion of three ⁴He yields ¹²C, and reaction of ¹²C with He can produce elements up to Ne in later stages of stellar evolution. The fusion of 3 He with atomic number 2 to ¹²C with atomic number 6 circumvents the production of the elements with atomic numbers 3–5 (Li, Be and B), which is the reason why these elements are so low in abundance. In contrast, elements with isotopes that have proton and neutron numbers equal to multiples of the ⁴He nucleus (¹²C, ¹⁶O, ²⁰Ne, ²⁴Mg, ²⁸Si, ³²S, ³⁶Ar, ⁴⁰Ca) are particularly abundant (Figure 1.9), a fact that was already noticed in 1914 by the Italian chemist Guiseppe Oddo (1865–1954). Several of these nuclides have “doubly magic” compositions (¹⁶O with Z=N=8 and ⁴⁰Ca with Z=N=20). The production of elements beyond C up to Fe through successive C, O and Si burning requires stars with more than 8 solar masses. The reactions leading to the Fe-peak elements involve disintegration of Si and other lighter nuclei to alpha particles and reassembly of the Fe peak nuclei in a steady state setting of nuclear reactions (nuclear statistical equilibrium).

The net fusion reactions of lighter elements to heavy elements up to Fe are exothermic. This is because the average binding energies of the protons and neutrons in nuclei up to mass numbers 56 (isotopes of Fe, Ni) increases. The most abundant nucleus in the iron peak region produced by supernova nucleosynthesis is actually ⁵⁶Ni with equal “magic” numbers of 28 neutrons and 28 protons. However, ⁵⁶Ni is unstable with a half-life is 6.1 days and decays to ⁵⁶Fe, which is why Fe is much more abundant than Ni. The reason why the doubly magic ⁵⁶Ni is not a stable nucleus can be understood from the nuclear shell model. Strong spin–orbit coupling occurs and destabilizes nuclei with doubly magic numbers of 28 and above, which is not the case for nuclei with the lower doubly magic numbers.

Beyond the region of the Fe group elements, nuclear binding energies decrease with increasing mass number and fusion reactions of the “lighter” elements such as Fe or Ni to form heavier ones up to Th and U require energy. The stability of the Fe-group nuclei puts a natural block on fusion yields from lighter elements, and this can explain why the Fe and Ni abundances spike above the overall abundance trends given by other elements or nuclides (Figures 1.8 and 1.9).

The elements beyond the iron group are mainly built through neutron capture by pre-existing, lighter elements (mainly Fe itself). Figure 1.10 shows the relative contributions from different processes to each element beyond Fe. There are two main neutron capture processes, called S and R processes.^11,12  One occurs on a slow timescale (hence S process) so that unstable nuclides produced through neutron capture may undergo beta-decay to a stable nuclide before another neutron can be captured. The S process operates in low and intermediate mass stars that have reached the giant star stage. It was in these types of stars where Merrill found Tc, which is naturally made by the S process. The S process also operates in massive stars that eventually blow up as supernovae, but the products and relative yields are different (Figure 1.10). The S process in massive stars is called the “weak S” process³⁰  to distinguish it from the “main S” process in giant stars that produces a much wider mass range of the heavy nuclides that contribute to the elements.³¹ 

During the R process, nuclides can capture neutrons rapidly (hence R process) before they undergo beta decay. This builds up unstable, neutron rich nuclides. After neutron exposure stops, these nuclides go through a chain of beta decays until a heavy stable or long-lived nucleus is produced. Thorium and U can only be produced this way. Larger contributions of isotopes by the R process occur for elements such as Te–Xe, Ir–Pt and the heavy rare earth elements (Figure 1.10). The R process requires high neutron densities, which may be realized in supernovae, but the exact site where the R process takes place is still under debate.

In addition to the S and R processes that produce neutron rich isotopes, a so-called- P-process is needed to explain the production of relatively proton-rich isotopes that contribute smaller fractions to a few elements (Figure 1.10). These proton-rich nuclides cannot be obtained through neutron capture processes and subsequent radioactive decay. One possibility is photodisintegration and “boil-off” of neutrons from more neutron-rich isotopes. The P process is believed to operate in supernovae, but appears to be the least understood process for heavy element production.

Since the different nucleosynthesis processes operate in different types of stars, it follows that the element inventory of the solar system is the result of nucleosynthesis in many different types of stars from different stellar generations. On a Galactic scale, the composition of the solar system is a snapshot of the composition at our location in the Galaxy 4.6 billion years ago. More importantly, it is the composition that, once concentrated to the solar system, through chemical and physical processing gave rise to the diversity of planetary objects that we have in our solar system today.

D. Arnett, Supernovae and Nucleosynthesis, Princeton University Press, Princeton, 1996, 598 pp.

B. E. J. Pagel, Nucleosynthesis and Chemical Evolution of Galaxies, Cambridge University Press, Cambridge, 1997, 378 pp.

N. Prantzos, Origin and evolution of the light nuclides, Space Sci. Rev., 2007, 130, 27.

G. Wallerstein, Synthesis of the elements in stars: forty years of progress, Rev. Mod. Phys., 1997, 69, 995.

S. E. Woosley, A. Heger, and T. A. Weaver, The evolution and explosion of massive stars, Rev. Mod. Phys., 2002, 74, 1015.