Isotope Ratio Measurements Firm Up Knowledge of Earth’s Formation

A complicated interplay of processes occurred 4.6 billion years ago in the early stages of the Solar System as material from the initial solar nebula condensed and collided to form aggregates, planetesimals, and eventually planets. And on Earth as well as the other terrestrial planets and the larger asteroids, heat from accretion and from the decay of short-lived isotopes such as aluminum-26 was sufficient to melt—at least partially—the assembled material, which led to the segregation of the iron-rich core and the silicate mantle.

Determining when Earth’s core formed can provide constraints on models of the planet’s formation. Studying Earth’s isotopic composition can provide such constraints, but this approach has led to conflicting answers.

Recently, three reports—by Ronny Schoenberg, Ken Collerson, and coworkers at the University of Queensland and the University of Bern; ¹ by Qingzhu Yin, Stein Jacobsen, and colleagues at Harvard University and the Ecole Normale Supérieure in Lyon, France; ² and by Thorsten Kleine, Klaus Mezger, and coworkers at the University of Münster and the University of Cologne ³ —have independently established a chronometry, based on careful measurements of tungsten isotopic compositions, that appears to have settled the timescales.

The core of the matter

As the Solar System formed, according to prevalent models, Earth accreted rapidly—over a period of a few million to a few tens of millions of years. ⁴ But confirming such models requires resolving what happened in the first few tens of millions of years in the 4.6-billion-year-old Solar System.

The hafnium–tungsten (Hf–W) system is well suited for dating the formation of Earth’s core. One Hf isotope, ¹⁸²Hf, decays to stable ¹⁸²W with a halflife of 9 million years. Produced like other heavy elements (including ¹⁸²W) in supernova explosions, ¹⁸²Hf was present in the original solar nebula, but the initial supply of ¹⁸²Hf has long since decayed. Thus, its concentration can’t be extrapolated back, unlike the concentrations of uranium-235 and uranium-238—whose decays to lead-207 and lead-206 form another chronometer of the early Solar System. The erstwhile presence of ¹⁸²Hf is revealed, however, in the isotopic composition of W. Today’s deviations from the average Solar System ratio of ¹⁸²W to other stable, nonradiogenic W isotopes reflect variations in the Hf/W ratio in the earliest period of the Solar System.

The chemistry of the Hf–W system also is vital to dating core formation. Lithophile (“stone loving”) Hf stayed in the silicate mantle when the metallic core formed. Tungsten, in contrast, is moderately siderophile (“iron loving”), preferentially dissolving in the metallic core. Thus, following the core formation, the mantle had a higher Hf/W ratio than the core. And so, if Earth’s mantle and core have different ¹⁸²W abundances, mantle–core separation must have occurred before the ¹⁸²Hf vanished.

Tungsten samples are easily obtained from the silicate Earth—just break open an incandescent light bulb—but not from the ferrous core. Fortunately, there is another source of W samples that can be tapped: meteorites. Carbonaceous chondrites are a class of primitive meteorites that are widely thought to have the same chemical composition (except for volatile elements such as hydrogen, helium, and the alkalis) as the solar nebula from which Earth and the other planets formed. Such meteorites can thus serve as proxies for establishing a baseline from which to look for deviations in W isotope ratios. (For more on what can be learned from meteorites, see the article by Thomas Bernatowicz and Robert Walker in Physics Today, December 1997, page 26 .)

Measuring W isotope ratios in chondrites is challenging, however. The assays require a sensitive mass spectrometer to detect differences of less than one part in 10⁴ in W isotope abudances, and the chondrites themselves contain W only at the partsper-billion level. Furthermore, corrections must be made for instrument effects and for background signals from other elements.

Following reports ⁵ in the mid-1990s by Jacobsen and Charles Harper (Harvard) that iron meteorites—thought to be the cores of early planetesimals—showed a deficiency in ¹⁸²W, Der-Chuen Lee and Alex Halliday (then at the University of Michigan) reported the first measurements of the ¹⁸²W abundance in carbonaceous chondrites; they found essentially no difference from the composition of the silicate Earth. ⁶ The apparent absence of any significant deviation indicated that the fractionation of Hf and W due to core formation must have occurred after the ¹⁸²Hf was extinct. Lee and Halliday thus concluded that, unless planetary accretion was a slow process occurring over tens of millions of years, Earth’s core must have formed at least 50 million years after the birth of the Solar System.

The three recent experimental efforts, which reexamine the Hf–W chronometry based on new measuremerits of chondrites, have challenged the late formation time. The figure shows the new results for the abundance of ¹⁸²W in various chondrite and terrestrial samples. These results each reveal a clear difference in the ¹⁸²W content, with the chondrites having about 2 parts in 10⁴ less than Earth’s mantle and crust.

At the 2002 Gold-schmidt Conference, held last August in Davos, Switzerland, Halliday, now at ETH Zürich, reported finding chondritic W concentrations that agree with the new sets of results, rather than with his earlier measurements, and said he has reproduced the Münster results. Thus the question of the ¹⁸²W abundance in chondrites appears to be settled.

A new consistency

The new findings point to an earlier formation of Earth’s core, when there was still ¹⁸²Hf left in the hafnium-enriched mantle to decay and thereby increase the ¹⁸²W fraction. Determining how much earlier requires additional information: the relative abundances of Hf and W, and the initial ratio of ¹⁸²Hf to other Hf isotopes at the birth of the Solar System.

Each team drew on existing estimates of 12–18 for the Hf/W ratio in bulk silicate Earth. The Münster group inferred an initial ¹⁸²Hf/¹⁸⁰Hf ratio of about 1.0 × 10⁻⁴ from W isotope measurements of various phases in a well-dated chondrite that formed very soon after the beginning of the Solar System. The Harvard group found the same ratio from its study of chondritic meteorites. The Queensland researchers obtained a value 50% higher based on other reported measurements on iron meteorites, but found a value comparable to that of the other groups when considering their own measurements on iron meteorites.

Conclusions about the core formation date also rest on assumptions regarding how the core formed. Models commonly assume that small planetesimals formed early and had differentiated cores. As those objects collided to form bigger ones, the accreting material was at least partly molten, allowing some degree of equilibration between the silicate and metal melts before the metal sank to Earth’s core. This equilibration is what the Hf–W studies date, but the extent of equilibrium varies from model to model.

Putting all the isotope ratios together, and assuming fully equilibrated element distribution between the mantle and core, the three teams reported a consistent value of about 30 million years as the latest time after the birth of the Solar System for Earth’s core to have formed in a single event. Such an earlier formation time is in better agreement with other chronometries ⁷ and with models. Arguing that Earth’s core is more likely to have segregated continuously as the planet grew, the Harvard group calculated 11 million years as the mean core formation time.

The new chronometry also has implications for other parts of the Solar System. For example, the Moon, which has a chemical composition similar to that of Earth’s mantle, is commonly thought to have been formed by a giant impact with Earth. Based on their data, the Harvard and Münster groups estimate that the Moon formed 25–33 million years after the beginning of the Solar System, comparable—and perhaps related—to the date they obtained for Earth’s core formation.