Chemical and Physical Evolution of the Crust

1. Introduction

As we know it, life was engendered when organic chemicals assembled into self-replicating biomolecules, though the natural processes that precipitated this change and their timing are still unknown. The likelihood of life’s emergence at any given time is difficult to constrain due to the complex interplay of chemical (availability of essential elements), planetary (orbital dynamics and magnetic field), and stellar (luminosity and star type) factors. While no consensus exists on when life first appeared (Pearce et al., 2018), some authors have reported purported biosignatures of Eoarchean (4.0–3.6 Ga [billion years ago]) (Hofmann, 2000; Moorbath, 2005; Allwood et al., 2006; Nutman et al., 2016; Dodd et al., 2017) and Hadean (>4.0 Ga) (Bell et al., 2015) age.

There is reasonable consensus that the origin of life was in part due to chemical interactions within Earth’s lithosphere (crust + uppermost mantle). In this context, we examine tectonic regimes that may have operated from the Hadean to the Eoarchean (4.56–3.60 Ga), along with the chemical and physical nature of the crust these regimes produced.

Around 4.5 Ga, Earth’s surface was dominated by a global magma ocean (Solomatov, 2015), inhospitable to life. Based on thermal evolution models, this phase likely lasted only a few million years or less (Lebrun et al., 2013; Korenaga, 2023). Oxygen isotope compositions of zircons from the Jack Hills (JHZ) (Mojzsis et al., 2001; Harrison et al., 2008; Gamaleldien et al., 2024) indicate magmatic interaction with low-temperature surface water as early as 4.28 Ga, which places an upper limit on the duration of the magma ocean. Following this stage, crustal formation and recycling processes emerged, evolved, and continue to this day, creating oceanic and continental crust.

The continental crust, which comprises most of the subaerial portion of the entire crust, plays a critical role in regulating the chemistry of both the atmosphere and oceans and directly influences conditions for life. Erosion of continental material contributes nutrients, such as phosphorus, to marine environments and thereby affects nutrient availability for marine life (Large et al., 2018). In addition, subaerial crust enables silicate weathering, which draws down atmospheric CO₂ and contributes to climate regulation (Berner, 2004).

We have structured this article by presenting first how the crust is created and destroyed today, followed by the chemical makeup of the > 3.6 Ga crust. Next, we outline the current schools of thought with regard to the style of tectonics that may have generated the observed chemical makeup. This is followed by a synthesis of the information presented, with consideration for origin of life models.

2. Crustal Generation, Recycling, and Preservation

Today, crust is being generated both at plate boundaries (e.g., mid-oceanic ridges and subduction zones) and within plates (e.g., intraplate hotspots). It is destroyed at convergent boundaries, where denser oceanic crust subducts into the mantle, and at the base of the continental crust, where dense lower crust delaminates and sinks into the mantle (Jagoutz and Kelemen, 2015).

Most of the modern crust covering Earth’s surface is oceanic (∼60%) and is recycled within ∼200 Myr (Seton et al., 2020). While older lithologies akin to modern oceanic crust are preserved beyond 200 Myr, their proportion relative to contemporaneous felsic rocks diminishes with age due to their susceptibility to destruction. The continental crust, in contrast, is much older (∼2 Ga; Korenaga, 2018a). It preserves the oldest known fragments of continental crust at ∼4.03 Ga (Bowring and Williams, 1999; Reimink et al., 2016; Bauer et al., 2017), which are possibly as old as 4.16 Ga (Sole et al., 2025), along with the oldest known terrestrial material: the detrital JHZs (4.37 Ga; Harrison et al., 2008; Valley et al., 2014). Therefore, much of the information we can glean from ancient rocks comes from continental crust material.

3. Chemical Makeup and Importance of the >3.6 Ga Crust

The composition of life-sustaining nutrients delivered to the oceans depends upon crustal chemistry. Although the proportions and chemistry of Earth’s bimodal crust—mafic (∼50 SiO₂ wt.% and typical oceanic) and felsic (∼61 SiO₂ wt.% and continental)—can be accurately assessed today, there is broad disagreement over their past proportions (Fig. 1 in Hawkesworth et al., 2024) and compositions. To elucidate the disagreement, some studies suggest that Earth’s crust reached its present chemical diversity and volume by ∼3.5 Ga (Harrison, 2009; Greber et al., 2017; McCoy-West et al., 2019; Guo and Korenaga, 2020), while others propose a gradual shift from mafic to felsic compositions as tectonic regimes evolved to favor more differentiated crust (Condie, 1993; Tang et al., 2016; Deng et al., 2023), with modern-like crust forming by ∼3.0 Ga (Dhuime et al., 2015; Tian et al., 2023). However, Keller and Harrison (2020) refute this age and propose that the silica content as well as the mafic/felsic proportion have remained constant since 4 Ga.

OPEN IN VIEWER

FIG. 1.

Stagnant-lid model showing modes of crustal generation (plumes), heat loss (heat pipes), and crust recycling (crustal drips and delamination).

While mafic and felsic crustal compositions and proportions affect the range of crustal diversity, the mafic crustal composition is sensitive to changes in global geodynamic factors such as mantle potential temperature and seafloor spreading rates (Langmuir et al., 1992; Donnelly et al., 2004). Mafic compositional variations are useful in assessing crustal petrology and tectonics, though their variations are subtle (trace element differences in basalts) when compared with that of the continental crust, which is typified by felsic and intermediate compositions. The greater chemical variability of continental crust is due to localized variations in P, T, and fluxing agents (Lee et al., 2016; Holder et al., 2019; Nutman et al., 2020). Thus, the felsic/intermediate composition of the Eoarchean (or earlier crust, if it existed) would be the main driver of crustal chemical diversification.

Crustal diversity likely played a key role in creating and maintaining environments conducive to prebiotic chemistry. A chemically heterogeneous crust will host regions out of geochemical equilibrium, which will create redox and compositional gradients that generate the free energy required for prebiotic chemistry (Barge et al., 2017). Crustal diversity can also manifest as physical differences: compared with mafic oceanic crust, felsic continental crust has greater thickness and is more buoyant. Buoyancy differences, along with changes in factors such as ocean volume (Dong et al., 2021), mantle water flux (Korenaga et al., 2017), radiogenic heat production (Rosas and Korenaga, 2021), and lithospheric rigidity (Mai and Korenaga, 2022), can influence the amount of continental freeboard (i.e., subaerial crust above sea level).

Subaerial crustal erosion may have contributed to the stabilization of Archean cratons that could have hosted sites where life emerged (Reimink and Smye, 2024). Craton stabilization makes the lithosphere more rigid and resistant to tectonic recycling. Reimink and Smye (2024) posit that erosive forces concentrate and entrain heat-producing elements (HPEs; U, Th, and K) from the subaerial crust, which induces partial melting in the deep crust. Melting concentrates these HPEs in the partial melt and depletes them in the lower crust/uppermost mantle that reduces the T of the lower lithosphere. A colder lower lithosphere makes the craton resistant to recycling and enables erosion of the subaerial crust for a longer period of time. The extensive erosion increases nutrient flux into the oceans and produces minerals, such as clays, which are relevant to some origin of life models (Ferris et al., 1996; Ferris, 2002).

Crustal composition and diversity can influence the redox state of the hydrosphere and atmosphere by regulating oxygen levels (Lee et al., 2016; Leong et al., 2021). These studies propose that a more mafic early crust would have hosted higher concentrations of Fe²⁺ and S²⁻, both strong oxygen sinks. As crust evolved to more felsic compositions, these sinks diminished, which potentially enhanced O₂ accumulation (Leong et al., 2021).

Efforts to quantify the chemical characteristics of the early crust have led authors to hypothesize the tectonic environments in which it formed (Lenardic, 2018; Hawkesworth et al., 2020; Brown et al., 2020; Korenaga, 2021; Cawood et al., 2022; Arndt, 2023). Proposed tectonic regimes range from a stagnant-lid Earth, where the lithosphere behaves as a single rigid plate (Solomatov, 1995; Reese et al., 1999; Bédard, 2018), to episodic or sluggish-lid tectonics, characterized by intermittent subduction (Moyen and van Hunen, 2012), to early plate tectonics with sustained lithosphere recycling (Turner et al., 2014, 2020; Korenaga, 2021).

In the next section, we review the tectonic styles proposed for the Hadean to Eoarchean and the types of crust these regimes may have produced. Each of these tectonic styles finds plausible analogs among other bodies in our solar system (Stern et al., 2023).

Nearly all known Hadean and most Eoarchean rock have undergone multiple metamorphic overprints, which have obscured their original petrological signatures. Detrital crystals such as the JHZs are orphaned from their parent melts, which also eliminates any geochemical insights into their primary sediment/rock matrix. This lack of pristine primary material hampers efforts to reconstruct their geologic context, including tectonic regimes and whether a hydrosphere was in existence at the time. As a result, direct evidence for early tectonic processes remains limited; this leads to a wide range of proposed tectonic models (Moyen and van Hunen, 2012; O’Neill and Debaille, 2014; Hawkesworth et al., 2017,2020; Lenardic, 2018; Korenaga, 2018b; O’Neill and Zhang, 2019; Harrison, 2020; Cawood et al., 2022).

4.2. Mobile-lid tectonics

A mobile-lid planet is characterized by a segmented lithosphere composed of multiple interlocking plates that move horizontally relative to one another. Where plates converge, density contrasts cause the colder, older, and denser plate to sink beneath the more buoyant one. This subduction process exerts a pulling force on the trailing portion of the subducting plate (slab pull). Slab pull induces pressure gradients within the lithosphere that, in combination with stresses from an upwelling mantle, promote plate divergence and lithospheric rupture (Coltice et al., 2019).

The processes responsible for the onset of plate mobility continue to be debated. Researchers have suggested a plume-induced start (Gerya et al., 2015) where a circular oceanic plateau created by a plume was thicker at its edges, which would have created favorable conditions for subduction. Others have proposed giant impactors (>500 km) as a possible initiator (Marchi et al., 2014; O’Neill et al., 2017). Regardless of the start mechanism(s), researchers have proposed two primary modes of plate motion: sluggish-lid and plate tectonics. Plate motion in a sluggish-lid scenario involves low-angle (<30°) subduction or episodic subduction events (O’Neill et al., 2007; Moyen and van Hunen, 2012; Drabon et al., 2022). In this mode, plate boundaries are typically diffuse and span broad zones of deformation rather than the sharp, well-defined boundaries seen today. Others have proposed that modern-style plate tectonics was already active in the Hadean to the Eoarchean, based on empirical evidence (Hopkins et al., 2008; Harrison, 2009; Turner et al., 2014, 2020) and numerical models (Foley et al., 2014; Sleep et al., 2014; Foley and Rizo, 2017; Rosas and Korenaga, 2018).

While subduction on early Earth remains a plausible hypothesis, there is no universally agreed-upon geologic indicator for its onset. Diagnostic magmatic, metamorphic, or structural features are rare due to extensive metamorphism, hydrothermal alteration, and recycling of the early crust. Consequently, researchers often rely on geochemical proxies (e.g., high Th/Nb, La/Yb, Sr/Y, Th/Yb; low Y and Yb) to infer subduction-like processes.

Alternative styles of plate motion have also been proposed. For example, rapid plate tectonics (∼50 cm/year compared with today’s 1–10 cm/year) may have been necessary to reduce Hadean atmospheric pCO₂ from 100–300 bar to <1 bar by the end of the Hadean (Miyazaki and Korenaga, 2022). Another model suggests an Iceland-like setting, where upwelling mantle melts undergo fractional crystallization and assimilate hydrothermally altered crust, generating intermediate to felsic melts (Reimink et al., 2014). This produces thick (up to 46 km; Allen et al., 2002), buoyant crust that resists recycling. The following is an elucidation of the two main mobile-lid tectonic modes.

4.2.1. Sluggish lid tectonics

The earliest evidence for felsic magmatism on Earth comes from whole-rock data on tonalite–trondhjemite–granodiorite (TTG) suites. They are high silica (>64 wt.%) and sodic (3–7 wt.% Na2O) granitoids with low HREEs and Eu anomalies which is distinct from modern granitoids that are more potassic and have higher HREE and negative Eu anomalies. TTGs first appear at ∼4.03 Ga in the Slave and Superior cratons (Hoffmann et al., 2019; Moyen and Martin, 2012) and makeup the main components of Archean terranes. However, modeling suggests that older JHZ parent melts were also intermediate to felsic (Watson and Harrison, 2005; Turner et al., 2020).

While several TTG genesis modes have been proposed, the consensus is that they formed by partially melting hydrated and metamorphosed basalts at various depths (<10–25 kbar), and then fractional crystallization of these parent melts resulted in the TTG suites (Moyen and Martin, 2012). As for geodynamic regimes of formation, among the many that have been suggested, (from melting at the base of a basaltic pile to modern subduction), most authors suggest some form of subduction was likely operating, especially for melts that were generated at 10–25 kbar (Table 7.2 in Hoffmann et al., 2019).

In some locations, the TTGs are associated with ultramafic to mafic lithologies such as greenstone belts, komatiites, and basalts (Moyen and van Hunen, 2012). This bimodal co-occurrence may happen when periods of mobile-lid tectonics are punctuated by stagnant-lid periods. The former tectonic style will melt hydrated mantle at subduction-like zones to generate the felsic rocks, while the latter will create the komatiites and basalts via spreading ridges and mantle plumes.

Notable examples of these mafic and ultramafic units include the Abitibi Greenstone Belt in the Superior craton (Moyen and van Hunen, 2012), the Barberton Greenstone Belt (Lowe and Byerly, 2007), the Western Pilbara Craton (Van Kranendonk et al., 2007), and the Zimbabwe Craton (Rollinson, 2023). The Zimbabwe Craton also contains boninites (a high-Mg andesite indicative of modern-arc magmatism; Pearce and Reagan, 2019), which may or may not (Rollinson, 2023) require plate tectonics.

Simulations by Moyen and van Hunen (2012) suggest that if subduction were initiated under Archean conditions, characterized by a hotter mantle and a weaker lithosphere, the subducting slab would likely be too weak to maintain integrity and would eventually break off and fail to sustain slab pull (Fig. 2). They show that if the T_P is raised by ∼200 K above modern values, Earth favors a sluggish-lid regime.

OPEN IN VIEWER

FIG. 2.

Sluggish-lid model showing intermittent subduction and the generation of subaerial crust, along with tectonic features such as spreading ridges where basaltic crust is generated.

A sluggish-lid tectonic regime may produce magmas with geochemical signatures that resemble modern boninites or adakites intermediate composition rocks (52–63 wt.% SiO₂) commonly associated with subduction initiation (Reagan et al., 2023; Patriat et al., 2019; Martin, 1999). These rocks are typically enriched in large ion lithophile elements (e.g., Rb, K, and Sr) and depleted in high field strength elements (e.g., Nb, Ta, and Zr). However, such signatures in modeled melts and early Earth rocks are not unique to subduction and may result from multiple petrogenetic processes. Thus, while informative, they should be interpreted with caution and supported by additional lines of evidence.

The episodic nature of subduction in these settings may result from (i) slab break-off due to thermal and mechanical weakness in the early lithosphere and (ii) collision of proto-continental fragments akin to modern continental collision (Moyen and van Hunen, 2012). In a sluggish-lid Earth, vigorous mantle overturn beneath a weak lithosphere imposes substantial tensile stress on the overlying crust. Unable to subduct effectively, the crust accommodates this stress through broad, diffuse deformation zones (Lenardic, 2018). A variant of this regime, termed “ridge-only” tectonics, has also been proposed. In this model, lithospheric spreading and crustal formation occur at mid-ocean ridge analogs, but subduction is absent (Rozel et al., 2015).

4.2.2. Plate tectonics

Qualifying the scope of plate tectonics in the Eoarchean to Hadean demands caution, as the scientific debate often polarizes into opposing views advocating for either vertical or horizontal tectonics, each often proposing mutually exclusive mechanisms. Modern plate tectonics is defined by narrow plate boundaries, active subduction zones primarily driven by slab pull, and continental drift governed by mantle convection (Fig. 3). Here, we review studies that advocate for modern-style plate tectonics in the Hadean/Eoarchean via geochemical evidence and models.

OPEN IN VIEWER

FIG. 3.

Modern plate tectonics with prolonged subduction, mantle wedges generating silicic crust along with plumes.

4.2.2.1. Arguments that imply Hadean plate tectonics

The JHZs, the oldest terrestrial materials, have been extensively studied for geochemical clues to their formation mechanism to shed light on early Earth tectonism (Harrison, 2020). For these zircons, Ti-in-zircon thermometry (Watson and Harrison, 2005) and oxygen isotopes (Valley et al., 2005) suggest crystallization from felsic melts near the water-saturated granite solidus, which interacted with low-temperature surface waters. Other studies identified similarities with modern arc magmas (Turner et al., 2020; Chowdhury et al., 2023, among others) by analyzing trace element compositions of JHZs and modeling silica contents of the JH parental melts.

Hopkins et al. (2008) used the Si content in mica inclusions to argue for zircon formation in a low T, high-P environment typical of subduction zones. However, Rasmussen et al. (2011) proposed that the inclusions may be secondary and thus may not be used to define formative environments. Bell et al. (2015, 2018) later published comprehensive studies on inclusions and proposed methods on how to distinguish primary inclusions from secondary ones, and there remains considerable debate regarding the primary or secondary nature of some inclusions in JHZs.

Boehnke et al. (2018) and Guo and Korenaga (2020) reported that portions of the Hadean crust were potassic and silicic. Their studies were based on Sr isotopes in apatite inclusions in Archean zircons and modeling of Ar isotopic composition of the mantle, crust, and atmosphere, respectively. Keller and Harrison (2020) have also proposed a uniform SiO₂ content of the crust since 4 Ga. A potassic crust implies the presence of K-bearing minerals, such as micas and clays on early Earth. This, in turn, signifies crustal weathering and recycling of subaerial felsic crust. This implication is buttressed if the mica inclusions in JHZs (Hopkins et al., 2008; Bell et al., 2015) are primary. Guo and Korenaga (2020) further proposed that Earth has experienced little to no net crustal growth since the Hadean and that the rates of rock recycling characteristic of modern plate tectonics may have been established by that time.

Following a different vein, Arndt (2023) suggested that, on a stagnant lid planet, to produce voluminous felsic crust, a mafic pile needs to be buried to certain depths to promote devolatilization that will abet partial melting. They argued that this mafic pile will devolatilize before it reaches melting depths, and the rocks higher up in the pile, interacting with the fluids just released through devolatilization, will be much colder and unable to generate large volumes of melt. Thus, according to their model, a stagnant-lid Earth will not produce voluminous felsic crust.

4.2.2.2. Arguments that imply Eoarchean plate tectonics

While a sizeable body of work suggests Hadean plate mobility, many authors have presented evidence that suggests a shift in the style of tectonics around 3.8–3.6 Ga. According to these researchers, crustal recycling through plate boundary processes probably started around this time. TTGs have drawn particular interest in this debate since they dominate the volume of felsic Archean crust. Some studies support a mobile-lid origin (Martin et al., 2014; Deng et al., 2019; Huang et al., 2022), while others suggest that subduction is not required for their genesis (Johnson et al., 2017; Bédard, 2018).

Enrichment in heavier Ti isotopes in Archean TTGs, positively and negatively correlated with SiO₂ and TiO₂, respectively, has been used to infer a proto-subduction setting for some Eoarchean TTGs from Greenland (Hoare et al., 2023). Reimink et al. (2016) studied TTGs in the Acasta Gneiss Complex and reported a progressive deepening of melt generation between 4.02 and 3.55 Ga. They proposed shallow-level fractional crystallization as the dominant process to create granitoids at 4.02 Ga and proposed that it was not until 3.6–3.55 Ga that melt was being generated at depth. This is interpreted as the gradual onset of mobile-lid tectonics by 3.6 Ga.

Researchers have also relied on JHZs to advocate for Eoarchean plate tectonics. Ackerson et al. (2021) suggested that some of the Hadean and Eoarchean JHZs indicate peraluminous (high Al₂O₃) parent melts. For these melts, they alluded to a formational mechanism where melts may have assimilated high alumina detrital sediments or may have originated at depth (0.7–1 GPa), starting around 3.8–3.6 Ga. This indicates the presence of subaerial crust or the transport of crustal material to great depths only in the Eoarchean and younger. These observations are potentially corroborated by shifts in isotopic compositions of JHZs and other global Hadean–Eoarchean zircons (Bauer et al., 2020). Drabon et al. (2022) observed no correlation between O and Hf isotopes in these zircons and zircons from the Barberton belt; they interpreted this as evidence for crustal isolation from the mantle for at least the Hadean.

Apart from JHZs and TTGs, other Archean rock units have also been analyzed to propose that Archean plate tectonics was active. Greber et al. (2017) and McCoy-West et al. (2019) analyzed Ti and Mo isotopes in shales and mantle-derived rocks, respectively. They argued that the silica content of the upper crust and the degree of mantle depletion had reached modern values by ∼3.5 Ga. This implies that some homogenization process was operating in the mantle-lithosphere system then. However, this interpretation remains contested; others have proposed a gradual increase in felsic crustal content over time (Tang et al., 2016; Tian et al., 2023; Deng et al., 2023).

The Archean crust is also characterized by amphibolites, ultramafic rocks, and banded iron formations. These rocks are found extensively in the Inukjuak domain, Superior craton, Canada. Grocolas et al. (2022) analyzed these types of rocks from this region (Ukaliq Supracrustal Belt [USB]) for their ¹⁴⁶Sm–¹⁴²Nd content, bulk and trace element chemistry, and mineral chemistry to propose subduction-like magmatism in the Eoarchean.

This study corroborates the results of Turner et al. (2014), who studied the neighboring Nuvvuagituuq Supracrustal Belt—whose age is debated to be either ∼4.28 to 4.16 Ga (O’Neil et al., 2008; Sole et al., 2025) or ∼3.8 Ga (Guitreau et al., 2013). They interpreted trace element concentrations, neodymium isotopic ratios, and stratigraphic features as being consistent with early plate tectonics.

While these are compelling arguments, as previously noted, many geochemical indicators of subduction are non-unique. For example, Reimink et al. (2020) proposed that the parental melts of Hadean JHZs may have originated from hydrated mafic crust formed at shallow depths, without necessitating a subduction environment. Caro et al. (2017) used ^146,147Sm–^142,143Nd systematics of amphibolites and ultramafics from the USB and argued for episodic crustal recycling in the larger Inukjuak domain. They proposed that the Hadean crust in this region foundered after a 600 Myr quiescent period at the surface, generating felsic melts and promoting fluxed mantle melting.

In summary, we have reviewed empirical evidence and models for the type of plate tectonics that was operating in the Hadean. However, numerous authors are quite strict about the assertion that there was a shift in the style of tectonics in the Eoarchean and it was only after 3.8–3.6 Ga that plate tectonics could have started (Fig. 4). Regardless of which school of thought is correct, if some form of plate mobility was active at > 4.0 to 3.6 Ga, early Earth was probably characterized by subaerial, voluminous felsic crust.

OPEN IN VIEWER

FIG. 4.

Different schools of thought on the progression of tectonic style over time. These are broad divisions, and more granular distinctions exist within each school. Sources: School A: Stern, 2005; Stern et al., 2016; Bédard, 2018; O’Neill and Zhang, 2019. School B: Moyen and van Hunen, 2012; Tarduno et al., 2015; Dhuime et al., 2015; Holder et al., 2019; Bauer et al., 2020; Reimink et al., 2020; Drabon et al., 2022; Cawood et al., 2022; Nebel et al., 2024; Hawkesworth et al., 2024. School C: Harrison et al., 2008; Turner et al., 2014,2020; Guo and Korenaga, 2020; Korenaga, 2021; Arndt, 2023.

4.2.3. Implications of mobile-lid tectonics for the origin of life

The presence of subaerial crust complements recent work (Gamaleldien et al., 2024; Trail and McCollom, 2023) that has proposed freshwater on subaerial land at ∼4.0 Ga as well as water–rock interactions occurring at a range of temperatures. These diverse reaction environments could have contributed to an abiotic, mineral-based production of free O₂ (He et al., 2023), which could have given rise to aerobic respiration (Jabłońska and Tawfik, 2021) and subsequently multicellular life (Ward et al., 2019). A higher proportion of subaerial felsic crust would also accelerate silicate weathering, thereby enriching the oceans with bioavailable nutrients while drawing down atmospheric CO₂.

If early Earth was characterized by freshwater circulation and subaerial crust, it may have also hosted lacustrine, lagoonal, and shallow marine environments. Evidence for these environments comes from clastic, carbonate, and chemical sediments in early Earth (Nutman, 1986; Friend et al., 2008; Komiya et al., 2015; Nutman et al., 2021; Boyd et al., 2024), although metasomatic origins of the sediments have also been proposed (Rose et al., 1996; Mloszewska et al., 2013). These shallow environments with regular wet–dry cycles, detrital clay, and the ability to concentrate phosphorus, cyanides, and so on, the last characteristic being especially indicative of carbonate-rich lakes, could have been the ones to host the first biomolecules (Hazen and Sverjensky, 2010; Toner and Catling, 2019, 2020).

Along with producing subaerial crust, mobile-lid tectonics promotes efficient crustal recycling and upwelling of mantle material, which increases the flux of life-essential elements to Earth’s surface. Miyazaki and Korenaga (2022) also proposed that plate tectonics alone may not be enough for life to originate. They argued that a more rapid form of plate tectonics must have operated to draw down the pCO₂ to <1 bar by the Hadean and enable life to evolve. Vezinet et al. (2025) proposed a similar scenario based on melt inclusions in olivine: that by 4.31 Ga, ∼80% of the continental crust had been extracted from the mantle and that such extraction required plume-induced rapid subduction to drive lithospheric recycling.

5. Summary, Synthesis, and Future Directions

Sometime between 4.56 Ga and today, Earth transitioned from a lifeless, molten body to one that hosts complex, self-replicating biomolecules. Understanding this transformation requires examining how early crust formation and rock recycling processes evolved over time. Earth’s crust likely formed from a global magma ocean, driven by accretion, core formation, and the Theia impact. This phase was followed by shallow mantle melting. As the mantle cooled, crustal recycling became more efficient, and mantle convection began coupling with the lithosphere, potentially enabling subduction. However, there is no consensus on when modern plate tectonics became the dominant geodynamic regime (Fig. 4).

Current understanding of early Earth tectonics is inferred from rocks and minerals that survived billions of years of geologic processing. However, this record is fragmentary, and the preserved materials may reflect not the dominant tectonic styles of the time but those environments most resistant to destruction. Using the extant samples, tectonic regimes have been proposed that range from stagnant-lid convection to modern-style plate tectonics.

To date, no single tectonic model has been universally accepted. While certain observations support specific tectonic modes, multiple tectonic styles may have coexisted or evolved over time, and no definitive evidence has emerged to support one dominant mode. Lenardic et al. (2016) approached this uncertainty by comparing Earth and Venus—two similarly sized rocky planets. According to Lenardic et al. (2016), the divergent evolutionary paths of Earth and Venus may stem not from differences in their starting conditions but from stochastic fluctuations such as changes in surface temperature, impactor flux, or unique episodes of volatile cycling and volcanism.

Their numerical simulations argue that, following the magma ocean stage, the probability that a planet is set on either a stagnant- or a mobile-lid evolutionary path is the same. Once a random event occurs, the planet on a certain tectonic evolutionary path deviates from that path. After that event ends, there is no guarantee that the planet will revert to its previous state because it can once again take one of two equally plausible paths. To elucidate, a planet may start on a stagnant-lid path and deviate from it due to a random event, but it may, or just as likely; may not, revert to a stagnant-lid mode after the event. Furthermore, such random perturbations may initiate positive feedback loops that perpetuate the newly established tectonic mode.

These findings highlight a key limitation of tectonic models: as deterministic constructs, they cannot fully account for random and unpredictable events (i.e., noise), which, according to Lenardic et al. (2016), are not trivial. Weller and Lenardic (2012) also previously highlighted another limitation: that for the same mechanical variables, tectonic models can equally predict different tectonic regimes if they only consider starting conditions and not evolutionary histories. The perspectives reviewed here are criticized in Korenaga et al. (this issue) because the “equally plausible tectonic modes” theory does not consider the higher mantle heat content and core-to-mantle heat flux of a plate-tectonic planet. Even if both modes are “equally plausible,” according to this article, transitioning from one mode to the other would require billions of years.

Differences in tectonic style have major implications for early Earth surface conditions. For a stagnant-lid planet, the early crust was probably predominantly subaqueous, with limited subaerial exposure. This topographic configuration would reduce silicate weathering and limit the flux of nutrients into the oceans.

Conversely, plate tectonics might spur increasing crustal elevation over time due to progressive differentiation into more buoyant felsic compositions; this would enhance weathering and nutrient delivery. Greber et al. (2025) used the Si isotopic content of Archean TTGs from 4.0 to 2.9 Ga to propose that there was subaerial exposure by 3.8–3.6 Ga that promoted weathering and erosion. Independent studies also suggest a shift in global geodynamics (Bell et al., 2015) driven by a change in tectonic styles around this time (Bauer et al., 2020; Reimink et al., 2016; Drabon et al., 2022).

Several studies (Condie, 1993; Harrison, 2009; Tang et al., 2016; Greber et al., 2017; Guo and Korenaga, 2020; McCoy-West et al., 2019; Tian et al., 2023; Deng et al., 2023) suggest a temporal trend from mafic to increasingly felsic continental crust. This shift implies increasing chemical diversity and disequilibrium within the crust, potentially expanding the range of chemical environments where life could originate.

While the models reviewed here are firmly rooted in robust data and modeling techniques, they are not perfectly balanced. Authors in the early Earth community continue to propose end-member scenarios because numerous unconstrained variables limit the ability to support a tectonically nuanced and diverse early Earth model.

In summation, multiproxy models are necessary to dispel the non-uniqueness and over-interpretation that plague data on Hadean/Eoarchean material. Therefore, hypotheses should be developed that are holistic and lead to a more unified early Earth model that accounts for all observations. For example, if stagnant-lid tectonics dominated the Hadean, how can this model be modified to explain arc-like chemical signatures in the JHZs or the Iceland-like chemical makeup of the Idiwhaa unit (Reimink et al., 2014)? Conversely, if mobile-lid tectonics was active, how can that model explain paleomagnetic observations that suggest no horizontal motions (Tarduno et al., 2023) or the long lag between crust generation and recycling inferred from PGEs, and W and Nd isotopes? The models derived from paleomagnetic data remain inconclusive (Fu et al., 2024), but some solutions have been proposed to explain the apparent time lag (Korenaga and Marchi, 2023; Rosas and Korenaga, 2018). Table 1 summarizes current early Earth tectonic models and the key parameters used to support them.

OPEN IN VIEWER

Table 1.

List of Tectonic Modes Discussed in This Review and Their Major Findings

Model (M)/Empirical observation (E)	Modeled variable/Parameters used	References	Major inference
Conventional tectonic modes
Stagnant lid
M	Mantle potential temperature (Tp)	Sizova et al., 2015; O’Neill and Debaille, 2014	If modern Tp is increased by 250 K, subduction shuts down
E	Nd and W isotopes, and PGE concentration	Willbold et al., 2011; O’Neill et al., 2013; Debaille et al., 2013; Tusch et al., 2021; Maier et al., 2009	2.9 Ga mantle rocks preserve Hadean signatures implying long crustal residence times
E	Sr isotopes	Dhuime et al., 2015	Increased erosion of subaerial felsic crust since 3 Ga implying stagnant-lid until then
M + E	Xe isotopes	Parai and Mukhopadhyay, 2018	Mantle Xe flux shifts from net degassing to regassing at ∼2.5 Ga implying stagnant-lid until then
M + E	Magnetic inclusions in early Earth zircons	Tarduno et al., 2015	Weak early Earth geodynamo implying no crust delivery to the CMB
		Tarduno et al., 2023	No change in the difference between paleolatitudes of Hadean continents
Sluggish lid
M	Mantle potential temperature (Tp)	Moyen and van Hunen, 2012	If modern Tp is increased by 200 K, episodic subduction initiates
E	Bimodal association of calc-alkaline rocks and komatiites	Moyen and van Hunen, 2012	Implies short-lived subduction events interspersed with plume-dominated magmatism
Mobile lid
E	Rare Earth/Trace elements (REE/TE)	Turner et al., 2014	Hadean/Eoarchean Nuvvuagituuq rocks have similar REE/TE makeup as modern subduction zone rocks
E	Al in Jack Hills zircons	Ackerson et al., 2021	[Al] > 4 ppm implies peraluminous melts that could have been formed in subduction zones
E	Si in Mica inclusions in Jack Hills zircons	Hopkins et al., 2008	Zircons crystallized in low T, high-P environment-like subduction zones
	Al in Hornblende inclusions in Jack Hills zircons
E	Hf and O isotopes, and TE of zircons	Drabon et al., 2022	Arc-like fluxed melting at 3.6–3.8 Ga
E	TE of zircons	Reimink et al., 2020	Gradual deepening of melt generation 4.02–3.6 Ga implying subduction at 3.6 Ga
E	Ti isotopes	Greber et al., 2017	3.5 Ga to modern shales have homogenized Ti isotope contents with a felsic component implying plate tectonics
		Hoare et al., 2023	Eoarchean TTGs are similar to calc-alkaline lavas implying subduction in the Eoarchean
E	Mo isotopes	McCoy-West et al., 2019	Archean and Phanerozoic komatiites imply the current degree of mantle depletion was achieved by 3.5 Ga
Alternative modes
Impact-driven tectonics
M	Bombardment rates	Marchi et al., 2014	Before 4.4. Gyr, 60–70% of the Earth’s crust was reworked up to a depth of 20 km
M		O’Neill et al., 2017	Localized short-lived subduction events at impact sites.
Rapid plate tectonics
M	Plate motion speed	Miyazaki and Korenaga, 2022	Plate speeds of ∼50 cm/year is necessary to sequester degassed CO2 to < 1bar by the Hadean
M + E	TE and Sr isotope of melt inclusions in Olivine	Vezinet et al., 2025	∼80% of the modern continental crust had been extracted by 4.31 Ga

As improbable as it may seem for life to arise from nonreplicating inorganic matter (Frampton, 2025), unless it is extraterrestrial, it did and understanding life’s emergence requires a deep amalgamation of biology and geology. Did life originate in hydrothermal vent-like environments where H₂, a key electron donor for methanogens (Helmbrecht et al., 2025), was produced? Or was it instead produced through non-vent serpentinization (Schlute et al., 2006)? Or perhaps both pathways coexisted. Could stagnant- or mobile-lid regimes sustain the fluid circulation necessary for H₂ production? Methanogens also require an electron acceptor such as CO₂, and there is evidence for their activity by 3.5 Ga (Ueno et al., 2006). This result proposes a lower age limit on inorganic C fixation. Evidence for organics by 3.5 Ga has also been presented; Mißbach et al. (2021) reported that 3.5 Ga fluid inclusions in barites from the Dresser formation, Australia, host primary organics and gases necessary for metabolism.

Reaction pathways thus operated on early Earth to synthesize organics from inorganic carbon. Vandenborre et al. (2021) showed experimentally that carbonates may be irradiated to abiotically produce H₂ and organics such as formate, acetate, and oxalate, in the presence of deep crustal fluids. These deep crustal fluids have long residence times (>10⁹ years at the Kid Creek mine, Ontario; Holland et al., 2013) and allow organic molecules to accumulate and provide a substrate to support microbial communities.

Key questions that arise from such studies are whether fluid circulation was active at 3.5 Ga, and, if so, did it sustain hydrothermal systems capable of hosting organics? Trail and McCollom (2023) suggested that relatively oxidized hydrothermal circulation may have been occurring at ∼3.9 Ga and younger and may have transported metals such as Fe, Cu, Mn, and Zn from Earth’s interior to its surface. Dodd et al. (2017) and Papineau et al. (2022) reported evidence for possible Fe metabolizing bacteria in 3.8 Ga hydrothermal vent deposits in the Nuvvuagituuq Supracrustal Belt, Canada. If these materials can be linked to specific tectonic regimes, they could support the idea that certain tectonic settings, over others, provided life-essential elements.

Finally, the observations presented throughout this review point to the fundamental role that early Earth tectonic style and timing of transitions between tectonic modes had in shaping Hadean–Archean environments. If tectonic regimes and low-T crustal processes can be better constrained, a cohesive model of early Earth may be formulated. Paired isotopic measurements (e.g., Fe-Zn or Zr-Mg) of ancient zircons offer a way to trace concurrently magmatic and low-T processes, respectively, and provide a more refined understanding of early Earth.

Trace elements in accessory phases may also help fingerprint tectonic environments. However, their reliability as proxies will be improved if their partitioning behavior can be better constrained experimentally in specific systems (granitic, mafic, alkaline, peraluminous, etc.) as a function of intensive parameters (T, P, fO₂, etc.). Additional isotopic systems, such as Ti (Greber et al., 2017), V (Stow et al., 2024; Bekaert et al., 2025), Zr (Guo et al., 2020; Tompkins et al., 2023), Mo (McCoy-West et al., 2019), and Tl (Nielsen et al., 2017, 2021), have also provided key insights into magma evolution and crust generation. These emergent avenues of investigation are likely to experience cycles of intense attention (Aarons et al., 2021) and plateauing (Bindeman and Melnik, 2022) until the next analytical innovation.

On a broader epistemological level, Harrison and Lenardic (2022) and Harrison (2023) have argued that the study of deep time requires a rethinking of how we pose scientific questions, interpret fragmentary data, and conceptualize geodynamic evolution over billions of years. Our understanding of early Earth remains broad, but it will improve as instrumentation and modes of investigation advance. Nutman et al. (2021) have outlined how our improving knowledge of Eoarchean rocks from Greenland has progressively refined our understanding of early Earth.

While a model of Earth transitioning from stagnant- to episodic-lid and eventually to plate tectonics is heuristically useful, it may oversimplify a complex geodynamic reality. It is increasingly likely that multiple tectonic modes operated simultaneously or sequentially with overlaps, depending on local thermal and rheological conditions. Therefore, early Earth should not be viewed as a planet with a singular, global tectonic style but rather as a dynamic system with spatial and temporal heterogeneity.