Granite – Earth-logs

How the earliest continental crust may have formed

Published on May 28, 2025 by zooks7771 Comment

Detrital zircon grains extracted from sandstones deposited ~3 billion year (Ga) ago in Western Australia yield the ages at which these grains crystallised. The oldest formed at about 4.4 Ga; only 150 Ma after the origin of the Earth (4.55 Ga). Various lines of evidence suggest that they originally crystallized from magmas with roughly andesitic compositions, which some geochemists suggest to have formed the first continental crust (see: Zircons and early continents no longer to be sneezed at; February 2006). So far, no actual rocks of that age and composition have come to light. The oldest of these zircon grains also contain anomalously high levels of ¹⁸O, a sign that water played a role in the formation of these silicic magmas. Modern andesitic magmas – ultimately the source of most continental crust – typically form above steeply-dipping subduction zones where fluids expelled from descending oceanic crust encourage partial melting of the overriding lithospheric mantle. Higher radiogenic heat production in the Hadean and the early Archaean would probably have ensured that the increased density of later oceanic lithosphere needed for steep subduction could not have been achieved. If subduction occurred at all, it would have been at a shallow angle and unable to exert the slab-pull force that perpetuated plate tectonics in later times (see: Formation of continents without subduction, March, 2017).

Landsat image mosaic of the Palaeoarchaean granite-greenstone terrain of the Pilbara Craton, Western Australia. Granite bodies show as pale blobs, the volcanic and sedimentary greenstone belts in shades of grey.

Geoscientists have been trying to resolve this paradox for quite a while. Now a group from Australia, Germany and Austria have made what seems to be an important advance (Hartnady, M. I. H and 8 others 2025. Incipient continent formation by shallow melting of an altered mafic protocrust. Nature Communications, v. 16, article 4557; DOI: 10.1038/s41467-025-59075-9). It emerged from their geochemical studies of rocks in the Pilbara Craton of Western Australia that are about a billion years younger than the aforementioned ancient zircon grains. These are high-grade Palaeoarchaean metamorphic rocks known as migmatites that lie beneath lower-grade ‘granite-greenstone’ terrains that dominate the Craton, which Proterozoic deformation has forced to the surface. Their bulk composition is that of basalt which has been converted to amphibolite by high temperature, low pressure metamorphism (680 to 730°C at a depth of about 30 km). These metabasic rocks are laced with irregular streaks and patches of pale coloured rock made up mainly of sodium-rich feldspar and quartz, some of which cut across the foliation of the amphibolites. The authors interpret these as products of partial melting during metamorphism, and they show signs of having crystallised from a water-rich magma; i.e. their parental basaltic crust had been hydrothermally altered, probably by seawater soon after it formed. The composition of the melt rocks is that of trondhjemite, one of the most common types of granite found in Archaean continental crust. Interestingly, small amounts of trondhjemite are found in modern oceanic crust and ophiolites.

A typical migmatite from Antarctica showing dark amphibolites laced with quartzofeldspathic products of partial melting. Credit: Lunar and Planetary Laboratory, University of Arizona

The authors radiometrically dated zircon and titanite (CaTiSiO₅) – otherwise known as sphene – in the trondhjemites, to give an age of 3565 Ma. The metamorphism and partial melting took place around 30 Ma before the overlying granite-greenstone assemblages formed. They regard the amphibolites as the Palaeoarchaean equivalent of basaltic oceanic crust. Under the higher heat production of the time such primary crust would probably have approached the thickness of that at modern oceanic plateaux, such as Iceland and Ontong-Java, that formed above large mantle plumes. Michael Hartnady and colleagues surmise that this intracrustal partial melting formed a nucleus on which the Pilbara granite-greenstone terrain formed as the oldest substantial component of the Australian continent. The same nucleation may have occurred during the formation of similar early Archaean terrains that form the cores of most cratons that occur in all modern continents.

News about when subduction began

Published on July 6, 2023July 5, 2023 by zooks777Leave a comment

Tangible signs of past subduction take the form of rocks whose mineralogy shows that they have been metamorphosed under conditions of high pressure and low temperature, and then returned to the surface somehow. Ocean-crust basaltic rocks become blueschist and eclogite. The latter is denser than mantle peridotite so that oceanic lithosphere can sink and be recycled. That provides the slab-pull force, which is the major driver of plate tectonics. Unfortunately, neither blueschists nor eclogites are found in metamorphic complexes older than about 800 Ma. This absence of direct proof of subduction and thus modern style plate tectonics has resulted in lively discussion and research seeking indirect evidence for when it did begin, the progress of which since 2000 you can follow through the index for annual logs about tectonics. An interesting new approach emerged in 2017 that sought a general theory for the evolution of silicate planets, which involves the concept of ‘lid tectonics’. A planet in a stagnant-lid phase has a lithosphere that is weak as a result of high temperatures: indeed so weak and warm that subduction was impossible. Stagnant-lid tectonics does not recycle crustal material back to its source in the mantle and it simply builds up the lithosphere. Once planetary heat production wanes below a threshold level that permits a rigid lithosphere, parts of the lid can be driven into the mantle. The beginnings of this mobile-lid phase and thus plate tectonics of some kind involves surface materials in mantle convection: the may be recycled.

Cartoon of possible Hadean stagnant lid tectonics, dominated by mantle plumes. (Credit: Bédard, J.H. 2018, Fig 3B, DOI: 10.1016/j.gsf.2017.01.005)

A group of geochemists from China, Canada and Australia have sought evidence for recycled crustal rocks from silicon and oxygen isotopes in the oldest large Archaean terrane, the 4.0 Ga old Acasta Gneiss Complex in northern Canada (Zhang, Q. and 10 others 2023. No evidence of supracrustal recycling in Si-O isotopes of Earth’s oldest rocks 4 Ga ago. Science Advances, v.9, article eadf0693; DOI: 10.1126/sciadv.adf0693). Silicon has three stable isotopes ²⁸Si, ²⁹Si, and ³⁰Si. As happens with a number of elements, various geochemical processes are able to selectively change the relative proportions of such isotopes: a process known as isotope fractionation. As regards silicon isotopes used to chart lithosphere recycling, the basic steps are as follows: Organisms that now remove silicon from solution in seawater to form their hard parts and accumulate in death as fine sediments like flint had not evolved in the Archaean. Because of that reasonable supposition it has been suggested that seawater during the Archaean contained far more dissolved silicon than it does now. Such a rich source of Si would have entered Archaean oceanic crust and ocean-floor sediments to precipitate silica ‘cement’. The heaviest isotope ³⁰Si would have left solution more easily than the lighter two. Should such silicified lithosphere have descended to depths in the mantle where it could partially melt the anomalously high ³⁰Si would be transferred to the resulting magmas.

Proportions of 30Si in zircons, quartz and whole rock for Acasta gneisses (coloured), other Archaean areas (grey) and Jack Hills zircons (open circles. Vertical lines are error bars. (Credit: simplified from Zhang et al. Fig 1)

Stable-isotope analyses by Zhang et al. revealed that zircon and quartz grains and bulk rock samples from the Acasta gneisses, with undisturbed U-Pb ages, contain ³⁰Si in about the same proportions relative to silicon’s other stable isotopes as do samples of the mantle. So it seems that the dominant trondhjemite-tonalite-granodiorite (TTG) rocks that make up the oldest Acasta gneisses were formed by partial melting of a source that did not contain rocks from the ocean crust. Yet the Acasta Gneiss Complex also contains younger granitic rocks (3.75 to 3.50 Ga) and they are significantly more enriched in ³⁰Si, as expected from a deep source that contained formerly oceanic rocks. A similar ‘heavy’ silicon-isotope signature is also found in samples from other Archaean terranes that are less than 3.8 Ga old. Thus a major shift from stagnant-lid tectonics to the mobile-lid form may have occurred at the end of the Hadean. But apart from the Acasta Gneiss Complex only one other, much smaller Hadean terrane has been discovered, the 4.2 Ga Nuvvuagittuq Greenstone Belt. It occupies a mere 20 km² on the eastern shore of Hudson Bay in Canada, and appears to be a sample of Hadean oceanic crust. It does include TTG gneisses, but they are about 3.8 Ga old and contain isotopically heavy silicon. So it seems unlikely that testing this hypothesis with silicon-isotope data from other Hadean gneissic terranes will be possible for quite a while, if at all.

Not-so-light, but essential reading

Published on July 25, 2013 by zooks777Leave a comment

In its 125th year the Geological Society of America is publishing invited reviews of central geoscience topics in its Bulletin. They seem potentially useful for both undergraduate students and researchers as accounts of the ‘state-of-the-art’ and compendia of references. The latest focuses on major controls on past sea-level changes by processes that operate in the solid Earth (Conrad, C.P. 2013. The solid Earth’s influence on sea level. Geological Society of America Bulletin, v. 125, p. 1027-1052), a retrospective look at how geoscientists have understood large igneous provinces (Bryan, S. E. & Ferrari, L. 2013. Large igneous provinces and silicic large igneous provinces: Progress in our understanding over the last 25 years. Geological Society of America Bulletin, v. 125, p. 1053-1078) and the perennial topic of how granites form and end up in intrusions (Brown, M. 2013. Granite: From genesis to emplacement Geological Society of America Bulletin, v. 125, p. 1079-1113).

Sea level change

Conrad covers sea-level changes on the short- (1 to 100 years), medium- (1 to 100 ka) and long term (1 to 100 Ma). The first two mainly result from local deformation of different kinds associated with glacial loading and unloading. These result in changes in the land surface, the sea surface nearby and on thousand year to 100 ka timescales to ups and downs of the sea-bed. Global sea-level changes due to melting of continental glaciers at the present day amount to about half the estimated 2 to 3 mm of rise each year. But increasingly sensitive measures show it is more complex as the rapid shifts of mass involved in melting ice also result in effects on the solid Earth. At present solid mass is being transferred polewards, but at rates that differ in Northern and Southern hemispheres and which are changing with anthropogenic influences on glacial melting. Viscous movement of the solid Earth is so slow that effects from previous glacial-interglacial episodes continue today. As a result rapid elastic movements are tending to produce relative sea-level falls in polar regions of up to 20 mm per year with rising sea level focusing on areas between 30°N and 30°S. The influence of the slower viscous mass transfer has an opposite sense: sea-level rise at high latitudes. Understanding the short- and medium-term controls is vital in predicting issues arising in the near future from natural and anthropogenic change.

Most geologists are concerned in practice with explanations for major sea-level changes in the distant past, which have a great deal to do with changes in the volumes of the ocean basins. If the global sea-floor rises on average water is displaced onto former land to produce transgressions, and subsidence of the sea floor draws water down from the land. Conrad gives a detailed account of what has been going on since the start of the Cretaceous Period, based on the rate of sea-floor spreading, marine volcanism and sedimentation, changes in the area of the ocean basins and the effects of thermally-induced uplift and subsidence of the continents, showing how each contribution acted cumulatively to give the vast transgressions and regressions that affected the late Phanerozoic. On the even longer timescale of opening and closing of oceans and the building and disintegration of supercontinents the entire mantle becomes involved in controls on sea level and a significant amount of water is chemically exchanged with the mantle.

Large igneous provinces

The Web of Science database marks the first appearance in print of “large igneous province” in 1993, so here is a topic that is indeed new, although the single-most important attribute of LIPs, ‘flood basalt’ pops up three decades earlier and the term ‘trap’ that describes their stepped topography is more than a century old. Bryan and Ferrari are therefore charting progress in an exciting new field, yet one that no human – or hominin for that matter – has ever witnessed in action. One develops, on average, every 20 Ma and since they are of geologically short duration long periods pass with little sign of one of the worst things that our planet can do to the biosphere. In the last quarter century it has emerged that they blurt out the products of energy and matter transported as rising plumes from the depths of the mantle; they, but not all, have played roles in mass extinctions; unsuspected reserves of precious metals occur in them; they play some role in the formation of sedimentary basins and maturation of petroleum and it seems other planets have them – a recipe for attention in the early 21^st century. Whatever, Bryan and Ferrari provide a mine of geological entertainment.

Granites

In comparison, granites have always been part of the geologist’s canon, a perennial source of controversy and celebrated by major works every decade, or so it seems, with twenty thousand ‘hits’ on Web of Science since 1900 (WoS only goes back that far). Since the resolution of the plutonist-neptunist wrangling over granite’s origin one topic that has been returned to again and again is how and where did the melting to form granitic magma take place? If indeed granites did form by melting and not as a result of ‘granitisation. Lions of the science worried at these issues up to the mid 20^th century: Bowen, Tuttle, Read, Buddington, Barth and many others are largely forgotten actors, except for the credit in such works as that of Michael Brown. Experimental melting under changing pressure and temperature, partial pressures of water, CO₂and oxygen still go on, using different parent rocks. One long-considered possibility has more or less disappeared: fractional crystallisation from more mafic magma might apply to other silicic plutonic rocks helpfully described as ‘granitic’ or called ‘granitoids’, but granite (sensu stricto) has a specific geochemical and mineralogical niche to which Brown largely adheres. For a while in the last 40 years classification got somewhat out of hand, moving from a mineralogical base to one oriented geochemically: what Brown refers to as the period of ‘Alphabet Granites’ with I-, S- A- and other-type granites. Evidence for the dominance of partial melting of pre-existing continental crust has won-out, and branched into the style, conditions and heat-source of melting.

English: Kit-Mikayi, a rock formation near Kis... — Typical granite tor near Kisumu, Kenya (credit: Wikipedia)

All agree that magmas of granitic composition are extremely sticky. The chemical underpinnings for that and basalt magma’s relatively high fluidity were established by physical chemist Bernhardt Patrick John O’Mara Bockris (1923-2013) but barely referred to, even by Michael Brown. Yet that high viscosity has always posed a problem for the coalescence of small percentages of melt into the vast blobs of low density liquid able to rise from the deep crust to the upper crust. Here are four revealing pages and ten more on how substantial granite bodies are able to ascend, signs that the puzzle is steadily being resolved. Partial melting implies changes in the ability of the continental crust to deform when stressed, and this is one of the topics on which Brown closes his discussion, ending, of course, on a ‘work in progress’ note that has been there since the days of Hutton and Playfair.

Share this:

Share this:

Share this: