Barium content of Archaean continental crust reveals the onset of subduction was not global

The Archaean Eon (4.0–2.5 billion years ago [Ga]) represents a critical period in Earth's history, marked by the formation and evolution of its earliest continental crust. This crust is predominantly composed of tonalite-trondhjemite-granodiorite (TTG) suites, making their petrogenesis crucial for understanding the prevailing geodynamic regime. The tectonic setting of TTG magmatism remains controversial, with ongoing debate surrounding the role of subduction. Some hypotheses propose that TTG formation occurred in the absence of subduction, perhaps through processes like delamination or plume activity. Conversely, other models strongly emphasize the involvement of subduction, suggesting that the characteristic geochemical signatures of TTGs are indicative of magmatic processes associated with subduction zones. Understanding the timing and nature of the onset of subduction is fundamental to resolving this debate. Determining whether subduction was a global process from the outset or a more gradual, regionally diachronous phenomenon has significant implications for our understanding of early Earth dynamics and the evolution of plate tectonics. This study leverages detailed petrological modeling and a comprehensive analysis of global TTG geochemistry to address these questions, focusing on the use of barium (Ba) as a proxy for subduction-related processes. The importance lies in providing a better constraint on the timing and global extent of early subduction, influencing models of early Earth evolution and the development of plate tectonics.

Previous research on TTG petrogenesis has yielded diverse interpretations. Studies have explored partial melting of hydrated basaltic sources, highlighting the role of pressure and temperature conditions in generating TTG magmas. The Sr/Y and La/Yb ratios in TTGs have been used as indicators of the depth of melting, classifying TTGs into high, medium and low-pressure types. However, the reliability of these ratios as proxies for melting depth has been questioned, due to the potential for post-magmatic processes like fractional crystallization and assimilation to alter their values. Experimental petrology and thermodynamic modeling have attempted to simulate TTG formation through partial melting of basaltic protoliths, but these have faced challenges in accurately reproducing the observed high-MgO compositions of some TTGs. The potential role of fluid-present melting in generating TTGs has also been investigated. Fluid-rich environments, like those found in hot subduction zones, are likely to significantly affect the melt composition and mineral assemblages. While enriched Archaean tholeiitic basalts have been proposed as a suitable source rock for TTGs, the variability in their trace element compositions needs to be considered. Some studies suggest that the early Archaean was characterized by a stagnant lid regime, dominated by vertical tectonic motions, rather than the horizontal plate movements of modern plate tectonics. This transition to plate tectonics and its timing remain critical questions in the field. Existing proxies, such as the appearance of specific rock types or metamorphic facies, often provide upper age limits rather than precise constraints on the onset of global subduction.

This study employs petrological modeling to investigate the relationship between geothermal gradients and the Ba content of TTG melts. The researchers used THERMOCALC v. 3.45 with the internally consistent dataset (ds62) of Holland and Powell for phase equilibria modeling. This involved calculating phase diagrams for an average enriched Archaean tholeiitic basaltic composition under a range of pressure-temperature (P-T) conditions, representing both hot and cold subduction settings, as well as a higher geothermal gradient representative of a stagnant lid regime. Two typical geothermal gradients (450°C GPa⁻¹ and 900°C GPa⁻¹) were modeled. Water-saturated conditions (7.0 wt.% H₂O) were included to assess the effects of fluid-present melting. The modeling considered the evolution of melt composition under different geothermal gradients and water contents to explore the influence of these factors on Ba concentrations. Trace element modeling was conducted using a batch melting equation, incorporating mineral/melt partition coefficients from the literature. The researchers accounted for accessory mineral solubility (zircon and apatite), calculating their retention in the residual based on solubility equations and degrees of melting. The study analyzed a global dataset of TTG Ba contents, utilizing a Bayesian change-point algorithm (conjugate partitioned recursion) to identify statistically significant shifts in Ba concentrations through time. This helped determine the timing of the onset of subduction in different cratons and whether the transition was globally synchronous or diachronous. The statistical analysis involved assessing the lognormal distribution of Ba data to confirm the validity of high-Ba TTGs as an endmember instead of outliers. Data used included >1000 analyses of enriched Archaean tholeiitic basalts.

The petrological modeling demonstrated a strong dependence of Ba concentration in TTG melts on the geothermal gradient. Specifically, only low geothermal gradients, characteristic of hot subduction zones, produced Ba-rich TTGs (>1000 ppm). Higher geothermal gradients (representing stagnant lid conditions) resulted in significantly lower Ba concentrations. The analysis of the global TTG Ba dataset revealed three statistically significant positive shifts in Ba content at approximately 3.7, 3.1, and 2.8 Ga. Crucially, these shifts were not synchronous globally. Instead, the onset of Ba enrichment varied diachronously across different cratons. For instance, the Slave and Kaapvaal cratons exhibited high Ba values as early as 4.0 and 3.5 Ga, respectively, while others showed later increases, extending to as late as 2.7 Ga (Superior craton). This diachronous pattern supports a model where subduction initiated regionally and gradually propagated globally. The study's findings indicate that Ba content in TTGs serves as a robust proxy for subduction, largely unaffected by fractional crystallization and magma assimilation. The correlation between the ages of Ba positive shifts in different cratons and the ages of their oldest rocks suggests a relationship between rock preservation and subsequent subduction initiation. The upper bound for the onset of global subduction is estimated to be sometime after 2.7 Ga. This is consistent with Earth's thermal history and mantle potential temperature estimates.

The findings of this study challenge the notion of a globally synchronous onset of subduction in the Archaean. The diachronous pattern of Ba enrichment in TTGs across different cratons strongly supports a model of subduction propagation, where subduction began regionally and progressively expanded globally. This regional to global transition from a stagnant lid to a plate tectonic regime aligns well with estimates of Earth's thermal evolution. The use of Ba as a proxy for subduction provides a significant advancement in understanding early Earth geodynamics. The Ba proxy offers a novel constraint on the timing of the transition to plate tectonics, addressing limitations associated with other proxies that mainly provide upper age bounds. The correlation between Ba shift ages and the ages of oldest rocks suggests that subduction initiation might be linked to the preservation of existing continental crust. Future research should focus on applying this approach to other regions and integrating it with other geochemical and geophysical data to further refine our understanding of the evolution of early Earth tectonic regimes.

This research demonstrates that the Ba content of Archaean TTG suites is a valuable proxy for the onset of subduction. The diachronous nature of Ba enrichment across different cratons reveals that the transition from a stagnant lid to a plate tectonic regime was not a globally synchronous event. Instead, subduction began regionally and progressively expanded, with an upper bound of 2.7 Ga for the onset of global subduction. This study improves our understanding of early Earth geodynamics, providing a robust constraint on the timing of a major transition in Earth's tectonic history.

The study relies on a global dataset of TTG Ba contents, and the completeness and representativeness of this dataset could influence the interpretation of the results. The accuracy of the petrological modeling depends on the chosen input parameters and models, such as the average Archaean tholeiitic basalt composition and the mineral/melt partition coefficients used. Uncertainties inherent in these parameters could affect the precise predictions of Ba concentrations under different P-T conditions. The interpretations presented assume a direct relationship between Ba enrichment in TTGs and subduction, and further investigation might be needed to completely eliminate alternative explanations for observed Ba variations. The precise timing of individual subduction initiation events within each craton might also depend on the spatial and temporal resolution of the analyzed TTG samples.