Evidence for an early Archaean transition to subduction

Modern plate tectonics is largely driven by slab-pull: a consequence of high-pressure, low-temperature metamorphism of the oceanic crust far from its origin at an oceanic ridge. As it ages, basaltic crust cools, become increasingly hydrated by hydrothermal circulation of seawater through it and its density increases. That is why the abyssal plains of the ocean floor are so deep relative to the shallower oceanic ridges where it formed. Due to the decrease in the Earth’s internal heat production by decay of radioactive isotopes, once oceanic lithosphere breaks and begins to descend high-P low-T metamorphism transforms the basaltic crust to a denser form: eclogite, in which the dense, anhydrous minerals garnet and sodium-rich pyroxene (omphacite) form. Depending on local heat flow, the entire oceanic slab may then exceed the density of the upper mantle to drag the plate downwards under gravity. Metamorphic reactions of any P-T regime creates minerals less capable of holding water and drive H2O-rich fluids upwards into the overriding lithosphere, thus inducing it to partially melt. Magmas produced by this create volcanism at the surface, either at oceanic island arcs or near to continental margins, depending on the initial position of the plate subduction.

A direct proof of active subduction in the geological record is the presence of eclogite and related blueschists. Such rocks are unknown before 2100 Ma ago (mid-Palaeoproterozoic of the Democratic Republic of Congo) but there are geochemical means of ‘sensing’ plate tectonic control over arc magmatism (See: So, when did plate tectonics start up? February 2016).  The relative proportions of rare-earth elements in ancient magmatic rocks that make up the bulk of continental crust once seemed to suggest that plate tectonics started at the end of the Archaean Eon (~2500 Ma). That method, however, was quite crude and has been superseded by looking in great detail at the geochemistry of the Earth’s most durable mineral: zircon (ZrSiO4), which began more than two decades ago. Minute grains of that mineral most famously have pushed back the geological record into what was long believed to be half a billion years with no suggestion of a history: the Hadean. Zircon grains extracted from a variety of ancient sediments have yielded U-Pb ages of their crystallisation from igneous magma that extend back 4.4 billion years (Ga) (see: Pushing back the “vestige of a beginning”;January 2001).  

Though simple in their basic chemical formula, zircons sponge-up a large range of other trace elements from their parent magma. So, in a sense, each tiny grain is a capsule of their geochemical environment at the time they crystallised. In 2020 Australian geochemists presented the trace-element geochemistry of 32 zircons extracted from a 3.3 Ga old sedimentary conglomerate in the Jack Hills of Western Australia, which lie within an ancient continental nucleus or craton. They concluded that those zircons mainly reveal that they formed in andesitic magmas, little different from the volcanic rocks that are erupted today above subduction zones. From those data it might seem that some form of plate tectonics has been present since shortly after the Earth’s formation. Oxygen-isotope data from zircons are useful in checking whether zircons had formed in magmas derived directly from partial melting of mantle rocks or by recycling of crustal magmatic rocks through subduction. Such a study in 2012 (see: Charting the growth of continental crust; March 2012) that used a very much larger number of detrital zircon grains from Australia, Eurasia, North America, and South America seemed, in retrospect, to contradict a subduction-since-the-start view of Earth dynamics and crust formation. Instead it suggested that recycling of crust, and thus plate-tectonic subduction, first showed itself in zircon geochemistry at about 3 Ga ago.

Detailed chemical and isotopic analysis of zircons using a variety of instruments has steadily become faster and cheaper. Actually finding the grains is much easier than doing interesting things with them. It is a matter of crushing the host rock to ‘liberate’ the grains. Sedimentary hosts that have not been strongly metamorphosed are much more tractable than igneous rocks. Being denser than quartz, the dominant sedimentary mineral, zircon can be separated from it along with other dense, trace minerals, and from them in turn by various methods based on magnetic and electrical properties. Zircons can then be picked out manually because of their distinctive colours and shapes. A tedious process, but there are now several thousand fully analysed zircons aged between 3.0 to 4.4 Ga, from eleven cratons that underpin Australia, North America, India, Greenland and southern Africa. The latest come from a sandstone bed laid down about 3.31 Ga ago in the Barberton area of South Africa (Drabon, N. et al. 2022. Destabilization of Long‐Lived Hadean Protocrust and the Onset of Pervasive Hydrous Melting at 3.8 GaAGU Advances, v. 3, article e2021AV000520; DOI: 10.1029/2021AV000520). The authors measured lutetium (Lu), hafnium (Hf) and oxygen isotopes, and concentrations of a suite of trace element in 329 zircons from Barberton dated between 3.3 to 4.15 Ga.

A schematic model of transition from Hadean-Eoarchaean lid tectonics to a type of plate tectonics that subsequently evolved to its current form, based on hafnium isotope data in ancient zircons (credit: Bauer et al. 2020; Fig 3)

The Hf isotopes show two main groups relative to the values for chondritic meteorites (assumed to reflect the composition of the bulk Earth). Zircons dated between 3.8 and 4.15 Ga all show values below that expected for the whole Earth. Those between 3.3 and 3.8 Ga show a broader range of values that extend above chondritic levels. The transition in data at around 3.8 Ga is also present in age plots of uranium relative to niobium and scandium relative to ytterbium, and to a lesser extent in the oxygen isotope data. On the basis of these data, something fundamentally changed in the way the Earth worked at around 3.8 Ga. Nadja Drabon and colleagues ascribe the chemical features of Hadean and Eoarchaean zircons to an early protocrust formed by melting of chemically undepleted mantle. This gradually built up and remained more or less stable for more than 600 Ma, without being substantially remelted through recycling back to mantle depths. After 3.8 billion years ago, geochemical signatures of the zircons start showing similarities to those of zircons derived from modern subduction zones. Hf isotopes and trace-element geochemistry in 3.6 to 3.8 Ga-old  detrital zircons from other cratons are consistent with a 200 Ma transition from ‘lid’ tectonics (see: Lid tectonics on Earth; December 2017) to the familiar tectonics of rigid plates whose basalt-capped lithosphere ultimately returns to the mantle to be involved in formation of new magmas from which continental crust stems. Parts of plates bolstered by this new, low density crust largely remain at the surface.

While Drabon et al. do provide new data from South Africa’s Kaapvaal craton, their conclusions are similar to earlier work by other geochemists based on data from other area (e.g. Bauer, A.M. et al. 2020. Hafnium isotopes in zircons document the gradual onset of mobile-lid tectonicsGeochemical Perspectives Letters, v. 14; DOI: 10.7185/geochemlet.2015), which the accompanying figure illustrates.

See also: Earliest geochemical evidence of plate tectonics found in 3.8-billion-year-old crystal. Science Daily, 21 April 2022. 3.8-Billion-Year-Old Zircons Offer Clues to When Earth’s Plate Tectonics Began. SciNews, 26 April 2022

Lid tectonics on Earth

Geoscientists have become used to thinking of the Earth as being dominated by plate tectonics in which large, rigid plates of lithosphere move across the surface. They are driven mainly by the sinking of cold, densified lithosphere in slabs at subduction zones. The volume of recycled slabs is replaced by continual supply of mafic magma to form oceanic crust at constructive margins. Such a process has long been considered to have reached far back into the Precambrian past and there are lively debates concerning when this modus operandi first arose and what preceded it. Now that we know more about other rocky planets and moons it appears that Earth is the only one on which plate tectonics has occurred. The other, more common, behaviour is dominated by stagnancy, although some worlds evidence volcanism and resurfacing as a result of giant impacts. Their subdued activity has come to be known as ‘lid tectonics’, in which their highly viscous innards slowly convect beneath a rigid, stagnant lid through which thermal energy is lost by convection: they are ‘one-plate’ systems. Although Earth loses internal heat by conduction through plate interiors, a large amount dissipates by convection associated with constructive margins: the oceanic parts of its plates lose heat laterally, as they grow older. Six papers in an advance, online issue of the free-access journal Geoscience Frontiers are concerned with the issue of terrestrial lid-tectonics and whether or not it dominated the Earth repeatedly in its Precambrian history.

A model is emerging for a hot, early Earth that was dominated by a form of lid tectonics (Bédard, J.H. 2018 Stagnant lids and mantle overturns: implications for Archaean tectonics, magmagenesis, crustal growth, mantle evolution, and the start of plate tectonics. Geoscience Frontiers, v. 9, 19-49; https://doi.org/10.1016/j.gsf.2017.01.005). Bedard’s model centres on lithosphere that was so weak because of its temperature that its subduction was impossible. Density of the lithosphere rarely increased above that of the mantle because the necessary mineralogical changes were not achieved – those involved in plate tectonics require low-temperature, high-pressure metamorphism as oceanic lithosphere is driven down at modern subduction zones. Even if such reactions did happen, the lithosphere would have been too weak to sustain slab-pull force and dense lithosphere would have simply ‘dripped’ back to the mantle. Mantle convection in a hotter Earth would have been in the form of large, long-lived upwelling zones rather than the relatively ephemeral and narrow plumes known today. Low density materials resulting from magma fractionation, the precursors of continental crust, would have been shifted willy-nilly across the face of the planet to collide. accrete and undergo repeated partial melting. In Bedard’s view, plate tectonics arose as Earth’s heat production waned below a threshold that permitted rigid lithosphere, probably in the late Archaean, to dominate after 2.5 Ga.

Bédard’s impression of an early Archaean lid-tectonic scenario. (credit: Jean H Bédard 2018, Figure 3B)

A radically different view is that stagnant-lid episodes alternated with periods of limited subduction and plate tectonics in the Archaean. Some Archaean cratons – the so-called ‘granite-greenstone terrains – seems to provide geological evidence for lid tectonics (Wyman, D. 2018. Do cratons preserve evidence of stagnant lid tectonics? Geoscience Frontiers, v. 9, 19-49; https://dx.doi.org/10.1016/j.gsf.2017.02.001). Others, such as the famous Isua supracrustal belt in West Greenland hint at plate tectonics. John Piper, of Liverpool University in Britain, argues from a series of Archaean palaeomagnetic polar wander curves that in three periods – ~2650 to 2200 Ma, 1550 to 1250 Ma, and 800 to 600 Ma – the poles shifted comparatively slowly with respect to the cratons providing the magnetic data; a feature that Piper ascribes to dominant lid tectonics (Piper, J.D.A., 2018. Dominant Lid Tectonics behaviour of continental lithosphere in Precambrian Times: palaeomagnetism confirms prolonged quasi-integrity and absence of Supercontinent Cycles. Geoscience Frontiers, v. 9, p. 61-89; https://doi.org/10.1016/j.gsf.2017.07.009). Similarly, there is some evidence based on the geochemical variation of basaltic rocks derived from the mantle. Through the Archaean, geochemical changes roughly follow cycles in the abundance of zircon radiometric ages and other geological changes that may reflect plate- and lid-tectonic episodes (Condie, K.C. 2018. A planet in transition: the onset of plate tectonics on Earth between 3 and 2 Ga? Geoscience Frontiers, v. 9, p. 51-60; https://doi.org/10.1016/j.gsf.2016.09.001). Interestingly, the age-frequency plot of almost three thousand Archaean and Hadean zircons recovered from the famous 1.6 Ga old sandstones of the Jack Hills Formation in Western Australia reveals similar cycles that may reflect such tectonic fluctuations in the Hadean (Wang, Q. & Wilde, S.A. 2017. New constraints on the Hadean to Proterozoic history of the Jack Hills belt,Western Australia. Gondwana Research, v. 55, p. 74-91; https://doi.org/10.1016/j.gr.2017.11.008). Since zircons are most likely to crystallize from intermediate and felsic magmas – i.e. precursors of continental material – their abundance in the Jack Hills rocks suggests that their source must have been in the 3.7 to 3.3 Ga gneisses on which the younger sediments rest. That is, part of those Archaean gneisses may well be made up of Hadean continental material that was repeatedly reworked and maybe remelted since such crust first appeared (in the form of surviving zircons) around 4.4 to 4.5 Ga, perhaps during vigorous lid-tectonic regimes.

Possible evolution of magmatic and tectonic styles for large silicate planets. (Credit: Stern et al. 2018, Figure 3)

Based on their reassessment of tectonic activity revealed by 8 rocky planets and moons Robert Stern of the University of Texas (Dallas) and colleagues from ETH-Zurich suggest a possible evolutionary sequence of tectonics and magmatism that Earth-like bodies might go through (Stern, R.J. et al. 2018. Stagnant lid tectonics: Perspectives from silicate planets, dwarf planets, large moons, and large asteroids. Geoscience Frontiers, v. 9, p. 103-119 ; https://doi.org/10.1016/j.gsf.2017.06.004). In their scheme plate tectonics requires certain conditions of lithospheric density and strength to evolve and suggest that, depending on planetary characteristics, slab-pull driven tectonics is likely to be preceded and followed by stagnant lid tectonics, to give perhaps a cyclical geotectonic history.

A fully revised edition of Steve Drury’s book Stepping Stones: The Making of Our Home World can now be downloaded as a free eBook