Mojzsis et al.1 measured the carbon-isotope composition of graphite inclusions in grains of apatite in banded-iron formations (BIFs) from Akilia island. Their observed δ13C values ranged from −20 per mil (‰) to −50‰, using a PDB standard, indicating a biogenic origin. The BIFs are older than 3,850 Myr (ref. 2), but the age of the apatite housing the graphite material was not determined. Here we present measurements of the U-Pb age of apatites from closely related samples.

We cast the sample chip of the Akilia BIFs (approximately 1×1 cm) into epoxy-resin disks with several grains of standard apatite and polished them until they were exposed through their mid-sections. The sample apatites are about 20-30 μm in size and show similar texture to those reported by Mojzsis et al.1. We focused a primary beam of about 2.5 nA O2 to sputter an area of apatites 20 μm in diameter, and extracted the positive secondary ions using 10 kV. We found no isobaric interferences in the mass range over 204Pb and 208Pb at a mass resolution of 5,800. We obtained the 238U/206Pb ratios from the observed 238U+/206Pb+ ratios by calibration using an empirical quadratic relationship between 206Pb+/238U+ and 238U16O+/238U+ ratios of standard. The experimental details of the apatite U-Pb analysis and calibration of data are given elsewhere7.

Eleven spots on seven individual apatite grains indicate that U concentrations vary significantly from 22.8 to 132 p.p.m. and do not show any correlation with 206Pb/204Pb and 207Pb/204Pb ratios. A correlation diagram of 238U/204Pb and 206Pb/204Pb ratios of the Akilia apatites is shown in Fig. 1. A least-squares fit using the York method gives the 238U-206Pb* isochron ages of 1,504±336 (2σ; mean square of weighted deviates (MSWD)=6.4). A correlation diagram of 204Pb/206Pb and 207Pb/206Pb ratios yields the 206Pb*-207Pb* isochron ages of 1,459±160 (2σ; MSWD=1.2). Both ages agree well with each other and are younger than the 3,860±10 Myr of ref. 2 for BIFs. This suggests either that the apatites in the BIFs grew about 1,500 Myr ago, or that they grew earlier than that but were subsequently affected by recrystallization, and/or diffusive exchange with the environment, which reset the U-Pb system of the samples. If the apatites were formed during metamorphism, biogenic carbon could have been introduced during the event.

Figure 1: Correlation diagrams of 238U/204Pb and 206Pb/204Pb ratios of Akilia apatites.
figure 1

Errors are shown at the 1σ level. The dotted line shows the best fit by the York method. In the calculation of this 238U/204Pb-206Pb/204Pb diagram, an error correlation of r =0.8554 was used, derived from the correlation coefficient between δ(238U/204Pb)/(238U/204Pb) and δ(206Pb/204Pb)/(206Pb/204Pb).

The Akilia association, including the BIFs and Isua supracrustals, are the oldest known components of the Archaean craton of Greenland and were affected by several metamorphic events after their formation 3,800 Myr ago (ref. 2). The latest event recorded in rock samples is the injection of basic dykes and crustally derived granitic sheets about 1,600 Myr ago, possibly coup-led with the anatectic reheating8. Biotite from all types of gneisses in the area gives a Rb-Sr isochron age of 1,600-1,700 Myr ago (ref. 3), which is within the error of the isochron ages of the Akilia apatites.

If these apatites were formed 3,800 Myr ago, graphite inclusions within grains of apatites have also experienced a thermal event of around 600 °C about 1,500 Myr ago. The isotope composition of the carbonaceous inclusions might have been altered by the event. It has been suggested9 that the oxidation of carbonaceous matter during metamorphism could alter the initial 13C value from −10‰ to −35‰ if temperatures are higher than 500 °C and if an oxidizing system is provided. In iron formation, an oxidizing system seems a reasonable suggestion.

The young U-Pb ages on the apatites, indicating closure for Pb in an event as late as 1,500 Myr, and the theoretical possibility that the light signature is a product of fractionation since 3,850 Myr, might in isolation throw doubt on the proposal1 that the Akilia BIF hosted life when deposited at around 3,850 Myr. However, as stated by Mojzsis et al.1, the petrographic association of the apatite and carbon-rich material is well known from unmetamorphosed BIF with microfossils, and also from modern biological observations. This was considered an equally important line of evidence for biological activity when the Akilia BIF was deposited.

In conclusion, when we seek evidence of the earliest life on Earth, we need to find an apatite with a U-Pb closure age apparently older than 3,500 Myr; with a pattern of rare-earth elements that indicates a biogenic signature; and with carbonaceous inclusions that contain isotopically light carbon.

Reply - Mojzsis et al.