View all news

Building the Earth

23 February 2006

Assessing the rates and processes of crustal growth requires linking the apparently contradictory information from the igneous and sedimentary rock records.

In the Earth’s early history there appears to be peaks of volcanic activity at 2.7, 1.9 and 1.2 billion (Ga) years. These peaks imply that crust was rapidly generated in response to the emplacement of mantle 'super-plumes', rather than by the process of subduction which results in more continuous crustal generation.

However, these peaks have to be reconciled with the smooth curves for the evolution of the continental crust inferred from the variations in neodymium isotopes in sedimentary rocks. Sediments provide average compositions of the source rocks from which they were formed, but these averages cannot be interpreted unless we can individually determine the evolution of both the igneous and sedimentary rocks in the continental crust.

Zircons are tiny minerals that form when magmas cool. They are difficult to destroy and so have a tendency to be retained as ‘detrital’ minerals in sediments. They can also be ‘inherited’ as grains in granites when granites form by melting crustal rocks. Zircons therefore encapsulate a more representative record of igneous events than the general geology exposed at the surface. The hafnium isotope ratios measured in zircons reflect the time since the crustal source rocks of the magma in which they crystallized separated from the mantle. The difficulty is that sediments yield hybrid source ages that reflect mixing processes during erosion and sedimentation rather than the age of crust forming events. Thus we have to distinguish magmas derived from sedimentary and igneous source rocks, and this has can be done using oxygen isotopes, which are high in sedimentary rocks. Kemp and the team reported in Nature [ 2 Feb 2006], the first study that integrates hafnium and oxygen isotopes, all measured in situ on the same, precisely dated detrital zircon grains.

Previous crustal evolution models require revision

The zircons sampled in this study are from south east Australia in an area that was part of the supercontinent called Gondwana. Our findings reveal that crustal growth was fundamentally episodic and confined to two major pulses at 1.9 Ga and 3.3 Ga. However, the crystallization ages of the zircons define prominent age peaks at the much younger ages of 500-650 Ma and 0.9-1.2 Ga. We conclude that most of the zircons crystallized in magmatic episodes that involved remelting of the pre-existing crust, and while they preserve the record of crust generation at 1.9 Ga and 3.3 Ga, the younger events involved relatively little generation of new crust.

In light of these findings, previous crustal evolution models that were based on neodymium in sedimentary rocks require revision. Such models invoked the generation of new continental crust on average every 500 million years, but we now know that that is unlikely in the late Archaean and much of the Proterozoic. Such models also required that new crust would be eroded and contribute to continental sediments soon after it was formed. The new data indicate that this too is unlikely and instead it may take about one billion years before new crust dominates the sediment record. It is this rate of response that results in the smooth curves for the isotope ratios in sediments with deposition age in the continental crust, in sharp contrast to the pulsed peaks of new crustal growth recorded in igneous rocks.

Kemp, Hawkesworth and Paterson/Earth Sciences Department

Edit this page