A Hadean to Paleoarchean geodynamo recorded by single zircon crystals

See allHide authors and affiliations

Science  31 Jul 2015:
Vol. 349, Issue 6247, pp. 521-524
DOI: 10.1126/science.aaa9114

Unlocking Earth's ancient magnetic past

The magnetic field protects Earth's surface from deadly cosmic radiation and provides clues about the planet's interior. Tarduno et al. found that some of the oldest minerals on Earth, Jack Hills zircons, preserved a record of a magnetic field over 4 billion years ago (see the Perspective by Aubert). Earth's magnetic field appears to have been fully operational a mere few hundred million years after the planet formed. This suggests an early start for plate tectonics and an ancient cosmic radiation shield that was important for habitability

Science, this issue p. 521; see also p. ARTICLE


Knowing when the geodynamo started is important for understanding the evolution of the core, the atmosphere, and life on Earth. We report full-vector paleointensity measurements of Archean to Hadean zircons bearing magnetic inclusions from the Jack Hills conglomerate (Western Australia) to reconstruct the early geodynamo history. Data from zircons between 3.3 billion and 4.2 billion years old record magnetic fields varying between 1.0 and 0.12 times recent equatorial field strengths. A Hadean geomagnetic field requires a core-mantle heat flow exceeding the adiabatic value and is suggestive of plate tectonics and/or advective magmatic heat transport. The existence of a terrestrial magnetic field before the Late Heavy Bombardment is supported by terrestrial nitrogen isotopic evidence and implies that early atmospheric evolution on both Earth and Mars was regulated by dynamo behavior.

The oldest previously reported geomagnetic field values, from 3.2 billion– to 3.45 billion–year-old magnetite bearing single feldspar and quartz phenocrysts from igneous rocks of the Nondweni and Barberton Greenstone Belts (Kaapvaal Craton, South Africa) (13), indicate a relatively strong field, but the prior history of the geodynamo is unknown. Some thermal evolution models predict no geodynamo before ~3.5 billion years ago (Ga) (4).

For magnetic minerals to be suitable recorders, they must be small, in the single to pseudosingle domain state (5), and have remained pristine since formation. The metamorphism that has affected Paleoarchean and older rocks makes paleointensity determination especially difficult. These metamorphosed rocks typically contain large multidomain magnetic grains (MD) with short relaxation times, secondary magnetic remanence carriers, and minerals with a propensity to alter during thermal demagnetization. The single crystal paleointensity (SCP) method was developed to isolate ideal magnetic carriers and behavior (13). In this method, a host silicate grain separated from a bulk rock is the focus of paleointensity study. The silicate is itself not of intrinsic magnetic interest; instead, it acts as the host to minute magnetic inclusions that have been shown to record magnetic fields with high fidelity. Previous SCP studies focused on feldspar (1, 2), quartz (2, 3), and olivine (6) hosts. Here, we present SCP data from zircons.

The Jack Hills (JH) conglomerate bears the oldest known zircon populations, with grains up to ~4.4 billion years old (7, 8). Unlike prior SCP studies, these zircons are detrital; their source has yet to be discovered and may have been lost to erosion. We evaluate the potential of events after and before the depositional age of the conglomerate (~3 Ga) to impose secondary magnetizations through paleomagnetic conglomerate tests and sensitive high-resolution ion microprobe (SHRIMP) analyses, respectively.

The JH belt has been metamorphosed to at least ~475°C (9) and variably deformed. Metamorphism is common in Archean terrains, but key samples that have locally escaped the more extreme effects of chemical and physical alteration accompanying metamorphism can sometimes be isolated (3). For example, some of the JH conglomerates are far from rare late intrusive volcanic rocks, which are expected to impose thermal and/or chemical remagnetizations. The interior of some cobble-sized clasts preserve centimeter-sized regions with a minimum of internal deformation and secondary mineralization, and a magnetite-dominated magnetization that passes a conglomerate test after demagnetization to temperatures of 500° to 550°C (10). Thus, locally components of the JH sediments have the potential to preserve magnetizations at high unblocking temperature in single or pseudosingle domain magnetic grains.

Our samples for zircon analysis were collected near the Discovery outcrop (7), ~1 km East of the cobble-bearing sites (10). Quartz clasts near the Discovery outcrop are pebble-sized (>4 mm, <64 mm), and because of this small size, they have suffered penetrative deformation; they are too deformed for a meaningful macroconglomerate test, but the zircons have survived with little or no internal deformation. We conducted a microconglomerate test on oriented small (~500 to 800 μm) samples, each centered on a single large (200 to 300 μm) zircon (11) to test whether the characteristic magnetization held by zircons from our samples has also survived post-depositional geological events. We used an ultrahigh-resolution three-component direct-current superconducting quantum interference device (SQUID) magnetometer (William S. Goree, Inc., Sand City, CA) that affords an order of magnitude greater sensitivity than that of other high-resolution SQUID rock magnetometers (11) in order to meet the challenge of full-vector measurement of such small samples with low bulk magnetizations. Thermal demagnetization by use of a CO2 laser (2) revealed unblocking between 565° and 580°C, which is consistent with a magnetite carrier (fig. S1, A to G). The characteristic magnetization of individual samples is well defined (median angular dispersion <10°), but together, the directions from seven samples cannot be distinguished from those drawn from a random distribution, indicating a positive microconglomerate test (fig. S1H). This test also addresses the debate over the primary nature of inclusions within JH zircons (12, 13). If magnetite carrying the characteristic magnetization was formed during metamorphism after conglomerate deposition, it should record a consistent direction that is contrary to our results. As in the larger-scale conglomerate test, multiple directions that are distinct from the high unblocking temperature magnetization (on a per sample basis) are isolated at lower unblocking temperatures, excluding modern lightning effects. The low unblocking temperature magnetizations reflect post-depositional geological events.

A zircon might also have experienced a reheating event from the time spanning its formation to incorporation into the JH conglomerate. We tested for the presence of reheating in two ways. First, we searched for Pb loss in the SHRIMP data that could be associated with hypothetical geologic events between initial zircon crystallization and emplacement in the conglomerate. Inhomogeneous redistribution of Pb within zircon at the nanometer scale is common in zircon that has experienced high-temperature metamorphism (11). The secondary ion mass spectrometry (SIMS) analysis sputters zircon to a depth of 700 to 1000 nm over ~15 min so that each individual analysis represents a depth-time series. In our second approach, we searched the secondary beam-normalized Pb counts over the analyses for nonsystematic variations indicative of Pb redistribution at the submicrometer scale during amphibolite- to granulite-grade metamorphism (11).

We separated JH zircons by hand using nonmagnetic techniques for SCP studies. We focused on larger zircons from the JH population that are greater than 150 μm in one dimension and separated them from searches of several thousand zircon crystals. We used 0.5-mm fused quartz sample holders to reduce blanks and used a routine that stacks complete magnetometer measurements at each demagnetization temperature step (11). Thermal methods are best suited for retrieving any thermoremanent magnetization (TRM) recorded by the zircons (5). Demagnetization experiments revealed natural remanent magnetization (NRM) versus temperature decay mainly between ~565° and 580°C, which is consistent with our microconglomerate results and the dominance of near-end-member magnetite (11). We used the Thellier-Coe method with MD grain tail checks (Fig. 1, A, D, and G, and fig. S2) (11). After paleointensity analyses, we analyzed zircons using the Geological Survey of Canada SHRIMP (Fig. 1, B, C, E, F, H, and I) (11). Data from 19 zircons met an extensive set of selection criteria (tables S1 to S3) (11). Age data from one sample (ZTC8) show evidence for Archean Pb loss that could represent greenschist metamorphism at ~2.6 Ga or an older event. However, the data for this and the other samples do not indicate nonsystematic variations in Pb counts. For 207Pb/206Pb ages between 3.38 and 3.66 Ga, paleofield values range between ~4 and 29 μT. These values are above threshold detection levels [~0.6 μT, defined by the interaction of the solar wind and an unmagnetized planet (1, 11)], even with the consideration of cooling-rate effects (11), and suggest the presence of a geomagnetic field from Paleoarchean to Eoarchean times.

Fig. 1 Thellier-Coe paleointensity and SHRIMP age data from individual zircon crystals.

(A) NRM lost versus TRM gained [circles; red are selected data (11)], with least-squares line fit to determine paleointensity value for sample ZTC7. Open triangles are partial TRM checks; solid green triangles are MD tail checks. (Inset) Orthogonal vector plot of field-off steps of Thellier-Coe data. Labeled points are degrees Celcius. Blue, declination; red, inclination. (B) Scanning electron microscope image of grain prepared for SHRIMP geochronological analyses. Scale bar, 20 μm. (C) Concordia diagram showing SHRIMP geochronological data. Uncertainty ellipses are 2σ. (D to F) Analyses shown in (A) to (C) for sample ZTC9. (G to I) Analyses shown in (A) to (C) for sample ZTC15.

Only ~12% of zircons from the Discovery site are >3.9 billion years old (7). To explore whether our samples preserve magnetic and age evidence for Hadean fields, we next applied a simplified paleointensity method to select zircons. This determination consisted of an evaluation of paleointensity at a single temperature (565°C) at which the characteristic remanent magnetization was defined. The 565°C determinations can provide a paleointensity estimate that is within a factor of approximately two of the full Thellier-Coe value (fig. S3). Most conservatively, however, we only used these data to test for the presence or absence of a geomagnetic field by comparison with threshold values. Data from 25 of these zircons met selection criteria (tables S2 to S4) (11). SHRIMP geochronological analyses yield 207Pb/206Pb ages between ~3.26 and ~4.22 Ga for these samples. These data also lack nonsystematic variations in Pb counts, but age data for our two oldest samples show some complexity. For our oldest sample (Z565-12; 4.22 Ga), we interpret this complexity as the influence of greenschist grade metamorphism at ~2.6 Ga (Fig. 2B). For sample Z565-5 (4.11 Ga), we cannot exclude the possibility of reheating between 2.6 and 3.9 Ga. We note that this age complexity is not seen in our samples dating to ~4 Ga and younger (Fig. 2, B to C). Paleointensities recorded by the 565°C determinations (Fig. 3) also exceed external field values and therefore suggest the presence of a core dynamo in the Hadean Eon.

Fig. 2 Paleointensity determinations analyzed at 565°C and SHRIMP age data from individual zircon crystals.

(A) Concordia diagram (left) showing SHRIMP geochronological analyses (uncertainty ellipses are 2σ). Line (blue) shows fit to data assuming Pb loss at 2650 ± 50 million years ago. (Right) Scanning electron microscope image zircon analyzed for sample Z565-12. Scale bar, 20 μm. (B) Analyses shown in (A) for sample Z565-11. (C) Analyses shown in (A) for sample Z565-24.

Fig. 3 Paleointensity versus time.

Solid blue squares are Thellier-Coe results from single zircons. Solid red circles are paleointensity determinations analyzed at 565°C. Diamonds are equatorial equivalent values assuming a dipole-dominated field from prior studies of Archean single-silicate crystals as follows: green, Barberton Greenstone Belt (BGB) Dalmein pluton (2); yellow, BGB Kaapvaal pluton (2); tan, Nondweni Greenstone Belt dacite (3); and pink, BGB dacite (3). All age and Thellier-Coe paleointensity uncertainties are plotted at 1σ. Uncertainties for 565°C paleointensity determinations are a factor of two bounds (fig. S2). Pink solid line and blue shaded region are, respectively, mean equatorial field value and standard deviation for the recent field derived from a bootstrap resampling of data from the past 800,000 years (11). LHB, Late Heavy Bombardment (23). Dashed line is the detection limit imposed by the external field (1).

Before the onset of inner core growth, the geodynamo was likely driven by thermal convection, in which case, the heat flow at the core-mantle boundary (CMB) must have exceeded the adiabat. Recent estimates suggest an adiabatic value around 15 TW (14, 15), although some are lower (16). Neither stagnant lid convection, as has been proposed for the Archean (17), nor the modeled heat transfer across a basal magma ocean (4) is consistent with an operating dynamo during this interval. Advection via heat pipes (18) and/or plate tectonics were likely the main mantle heat transfer processes operating on the early Earth.

In the absence of a core dynamo field, atmospheric N2 would be susceptible to ionization and removal by solar wind pickup (1, 19). The presence of a magnetic field is therefore consistent with the lack of nitrogen isotopic fractionation deduced by the study of geologic samples of the Archean atmosphere (20). However, the weakest paleointensity values (Fig. 3) are comparable with a magnetic field only ~12% of that of today, which would imply magnetopause standoff distances less than three Earth radii (1). A coronal mass ejection added to this steady-state solar wind could have resulted in the pulsing of volatiles and water from Earth’s atmosphere and perhaps implantation of nitrogen on the Moon (21). This in turn argues that Earth’s water budget was initially much greater than today—and/or that it was replenished during the delivery of a water-rich late veneer (22) or during the tail of the Late Heavy Bombardment (23) at ~3.9 Ga—to account for today’s water inventory. In addition to magnetic shielding, total losses may have been in part mitigated by bottlenecks in transport through the atmosphere and/or an early hydrogen envelope (1, 19), the latter restricting extreme atmospheric expansion.

Our paleointensity and age data suggest the presence of a core dynamo >750 million years earlier than prior estimates. An early start for the geodynamo is similar to that of the early Martian magnetic field (24), but the subsequent collapse of the Martian dynamo probably facilitated atmospheric stripping (25). In contrast, the early start and persistence of atmospheric shielding attendant with the long-lived geodynamo was likely a key factor in the development of Earth as a habitable planet.

Supplementary Materials

Materials and Methods

Figs. S1 to S3

Tables S1 to S4

References (2647)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: We acknowledge the late William Goree for his efforts in developing the ultrahigh-sensitivity SQUID magnetometer used in this study. We also thank G. Kloc for assistance in sample preparation; T. Pestaj for assistance in the SHRIMP laboratory; and S. Wilde, J. Valley, D. Trail, and C. Manning for helpful discussions. The data presented in this manuscript are tabulated in the supplementary materials. This work was supported by the U.S. NSF (grants EAR-0619467 and EAR-1015269).

Stay Connected to Science

Navigate This Article