Hadean Earth

The development of the geologic timescale arose from early nineteenth century fossil correlations and thus firmly rooted in the rock record. The thinking of that time included the possibility that our planet had forever existed in a quasi-steady state. By the later part of that century, it was broadly understood that Earth must have experienced a discrete origin but that details of that event might never be discerned. The advent of radiometric dating shortly thereafter catapulted this discussion from considering upper bounds of tens of millions of years to several billions. But it was the return of lunar highland rocks in the early 1970s that revolutionized thinking about the early inner solar system, including a hypothesized impact bombardment at ca. 3.9 billion years that was thought to have obliterated planetary surfaces. It was in this context that the term “Hadean” was proposed as the earliest division of geologic time. Since then, the meaning of Hadean evolved to describe the first roughly 500 million years of Earth history, which currently coincides with the period for which we have no macroscopic rock record. The premise of this book is that there are four avenues available to seek the nature of this eon of Earth history: (1) the presumption that physical laws are time dependent, and thus limits can be placed on early Earth’s behavior using mathematical calculations; (2) that isotopic variations resulting from long-lived and extinct radioactivities preserved in mantle rocks can constrain the timing of early, planetary-scale differentiation; (3) that as much as half the rocks on the lunar surface are likely between 4.4 and 3.9 billion years in age and thus can attest to conditions then extant in the inner solar system; and (4) that detrital zircons between 4.0 and nearly 4.4 billion years preserve a lithic record of terrestrial activity in that period. Although the initial choice of the word Hadean was meant in infer that this phase of Earth history was characterized by hellish surface conditions, in classic mythology it was a cool and watery realm. Studies of these ancient zircons over the past 20 years appear to reveal an early history more akin to the latter myth than the former.

The study of Earth as an object whose history can be understood by application of physical laws dates back 200 years. This tradition is, however, rife with missteps related to as yet undiscovered physics or fundamentally incorrect assumptions. While the former is unavoidable, the latter amounts to self-inflicted wounds that may have forestalled scientific progress. Even in the absence of knowledge of initial conditions, linear mathematical relationships such as first order loss (e.g., radioactive decay) have proved useful in predicting Hadean conditions. However, more complex physical systems cannot be uniquely extrapolated back in time. For example, mantle convection, a highly non-linear, dispersive, chaotic system is, by its very nature, uninvertible. This fact has not inhibited generations of modeler’s from making ab initio predictions regarding early Earth evolution. Their results were initially limited by technological impediments and adoption of assumptions regarding the relationship between interior temperature and planetary heat loss that narrowed possible solutions. Radically new proposals regarding both the latter issue and discontinuous transitions between modes of heat loss have tempered earlier conclusions that plate-tectonic-like behavior could not arise on early Earth. Physical calculations have an important role to play in assessing the plausibility of Hadean geodynamic models, but should best be seen as “convenient fictions”.

Evidence from short- and long-lived radioisotope systems indicates that Earth was largely built from volatile-depleted planetesimals and planetary embryos during runaway accretion that occurred within 1–10 Ma of formation of the first solar system solids at 4.567 ± 0.001 Ga. Both long- and short-lived radionuclides leave isotopic signatures in mantle rocks that bear on when and how the silicate Earth formed and differentiated. The longstanding view that the Hadean mantle was compositionally undepleted appeared to be contradicted by differences between mantle and chondrite Nd isotopes which suggested a very early enriched terrestrial reservoir. Although recent work indicates this difference reflects differing irradiation histories of Earth-forming-materials and meteorites, and thus has little bearing on the timing of silicate differentiation, both terrestrial and lunar Lu–Hf zircon data appear to require global silicate differentiation by 4.50 ± 0.02 Ga billion years. Tungsten isotopic data from mantle rocks provides evidence of core formation by about 4.53 Ga and either very early isotopic isolation of silicate reservoirs or disturbance by a late chondritic veneer. Despite evidence that Moon formed substantially from proto-Earth material between about 4.53 and 4.50 Ga, the exact mechanism by which this occurred remains controversial. The once widely accepted model of collision of a Mars-sized body has lost support in light of contradictory evidence in the form of indistinguishable isotopic compositions of volatile and refractory elements between Earth and Moon. Models that appear to transcend this problem (hit-and-run collision, synestia, successive smaller collisions, magma ocean heating, etc.) are currently being evaluated. Although geochemical evidence requiring an early terrestrial magma ocean is almost entirely lacking, the sources of thermal energy available during accretion make such an appearance appear inevitable. If solidification proceeded from the bottom up, vigorous convection would have caused the lower mantle to rapidly crystallize with the upper mantle becoming largely solidified within several million years. The high abundance of highly siderophile elements in the upper mantle is strong evidence that Earth added at least half a percent of its present mass following core formation but prior to the Mesoarchean. Preservation of mantle isotopic anomalies throughout the Hadean-Archean seem unlikely to reflect sluggish mantle convention in a stagnant lid tectonic regime during that period as the plate tectonic era is associated with a large range of isolated mantle isotopic domains.

As much as half of lunar surface rocks may have originated between 4.4 and 3.9 billion years and thus observations of, and samples from, Moon could attest to conditions then extant in the inner solar system. The concept of a lunar cataclysm at ~3.9 Ga grew from seemingly contradictory observations of elemental fractionation in lunar highland rocks. U–Pb—and some Rb–Sr—data suggested recrystallization occurred between about 4.0 and 3.8 Ga. The Late Heavy Bombardment (LHB) concept that emerged appeared supported by ~3.9 Ga ⁴⁰Ar/³⁹Ar “plateau ages” of lunar impact melt rocks, although no similar spike in ages was seen in the likely more globally distributed lunar meteorites. While the ⁴⁰Ar/³⁹Ar step-heating method can reveal intragrain isotope variations, this capability has several method-specific requirements that, if not met, preclude thermochronologic interpretations. Three such issues effectively rule out the use of virtually all lunar ⁴⁰Ar/³⁹Ar data as support for the LHB hypothesis: (1) the “plateau age” approach used is an aphysical concept for the thermally disturbed samples typical of most lunar impact melt rocks, (2) laboratory artifacts destroy preserved diffusion information, or create false apparent age gradients; and (3) obtaining meaningful thermal history information from extraterrestrial samples that have differing activation energies for Ar diffusion in their K-bearing phases requires a different laboratory protocol than was used on lunar rocks. Possibly due to these issues, no case in which multiple chronometric techniques have yielded intrasample concordancy of a lunar melt rock has yet been documented. Advancements in mass spectrometry now permit ⁴⁰Ar/³⁹Ar and U–Pb dating to be undertaken on small (10 s-of-μm diameter) in situ spots on glasses and accessory minerals in lunar rocks. This approach has the potential to transcend the analytical challenge posed by the continuous impact reworking of the lunar regolith that produces fine-scale polygenetic breccias of multiple age and origins. The longstanding assumption that lunar melt rocks originated from discrete, basin-forming events is obviated by lunar imaging that show impact melts formed in small highland craters and clusters of ‘light plains’ deposits radiating outward  >2000 km from large impact basins. The latter underscores how poorly the spatial relationships between large basins and their surrounding deposits were understood when impact chronologies were developed in the 1970s. The assumption that a specific lunar melt rock from a given landing site is representative of one of the basin-forming impacts is deeply flawed. Establishing a reliable, quantitative planetary impact chronology requires that all analyzed rocks be equally suitable for the application of specific chronometers. This may not be possible given the large contrasts in incompatible trace element distributions across the lunar surface (e.g., Procellarum KREEP terrane, South Pole Aiken basin). A conservative view of the lunar chronological record is that the large nearside basins are older than 3.82 Ga but these data are consistent with most of them being older than 3.92 Ga and possibly older than 4.35 Ga.

We don’t know with confidence the mechanisms by which primitive arc basalts are modified to produce the broadly granodioritic continental crust but there is widespread agreement that plate tectonics has been doing just that for at least the past billion years. We also don’t fully understand the structure of the continental crust; popular layered models lack mechanisms to produce such structures or to recover them following tectonic homogenization. The geochemical community long favored the view that early crust was mafic, in part owing to misconceptions regarding feldspar buoyancy on a hydrous magmatic substrate and the deep stabilization of garnet (which retards crystallization of more buoyant aluminous phases from the magma). But early felsic crusts with the potential for long term stability could have emerged via crystallization of tonalitic liquids fractionated from ultramafic magmas in equilibrium with olivine or differentiating magma sheets following large impacts into early basaltic crusts. The remarkable range of estimates of the growth history of continental crust reflects a number of influences but, generally speaking, earlier growth has been increasingly favored as new age survey methodologies became available and as the effects that crustal reworking and recycling have on apparent surface age provinces became better appreciated. Isotopic data once thought to support rapid growth at ~2.7 Ga are now recognized as equally consistent with constant volume continental crust. The longstanding misapprehension that the present-day distribution of crust formation ages is equivalent to the growth history of continents strongly influenced some estimates. Instead, today’s crust represents a running balance between new growth, internal overprinting, and crustal recycling. The difficulty in deconvolving these processes is one of the two principal challenges in establishing the growth history of continental crust. The other is that crust recycled back into the mantle and thoroughly mixed leaves no trace of its past incarnations. Although the rock record has yet to yield clear, direct evidence from which to constrain the magnitude of Hadean continental crust, optimal solutions to modelling mantle isotopic data are at least as consistent with constant volume continental crust since ca. 4.4 Ga as with slow monotonic growth. Radiogenic isotopic data used to argue for an early mafic crust are contradicted by stable isotopic results that appear to support a continuously felsic continental crust of unknown volume.

Estimates of when plate tectonics began range from the last 20% of Earth history to within the first 5%. While there is no observation that precludes plate tectonics from operating at 4.3 Ga, evidence that it was is indirect. Although subduction initiation is a robust feature of the modern plate tectonic system and we can calculate with some accuracy when oceanic lithosphere attains negative buoyancy, we don’t yet understand how strong the lithosphere weakens sufficiently for subduction to initiate. Most approaches used to estimate when Earth first entered the mobile lid regime—preservation of modern plate tectonic features, detrital zircon age spectra, trace element and radiogenic isotope geochemistry, atmosphere-crust-mantle exchange, and model-based estimates—can be interpreted in multiple ways and are all underlain by assumptions that cannot be independently tested. All share the flaw that absence of evidence is not evidence of absence. Of special concern is that the Precambrian geologic record is likely biased to rock compositions most likely to resist deformation and thus exposure to erosion at newly rifted continental margins where loss to subduction erosion could occur. Thus any look-back comparison is flawed to some degree by a preservation bias. A more recently recognized limitation is the failure to consider how a hotter, early Earth would differ petrologically from, say, Phanerozoic behavior (e.g., lower incompatible trace element concentrations in mantle magmas, higher geothermal gradients). Historically, computational limitations in early geophysical modelling methods led to skepticism regarding the possibility of plate tectonics on early Earth. Influenced by this view, the geologic community was reluctant to take a dynamic view of the preserved crustal record, instead inferring the apparent absence of a Hadean rock record as evidence that there never was one. The unknown extent to which ancient continental crust was recycled into the mantle and thoroughly mixed, the abovementioned selection biases in the rock record, and the assumption of uniformitarian conditions throughout Earth history limit virtually all continental growth estimates to providing only lower age bounds and thus minimum estimate on the initiation age of subduction.

Geochemical analysis of zircons older than 4 billion years, found in Early Archean metasediments at Jack Hills, Western Australia, provide insights into the nature of Hadean Earth. Oxygen isotopes have been interpreted as indicating that protoliths of magmas from which Hadean zircons crystallized were formed in the presence of water at or near Earth’s surface. Apparent crystallization temperatures of Hadean zircons cluster at 680 °C. Given the low porosity expected in rocks under anatectic conditions, dehydration melting of micas as the principal source of the melts from which these zircons crystallized can be ruled out. Instead, a regulated mechanism producing near minimum-melting conditions during the Hadean is inferred. Combined, these results have been interpreted to reflect chemical weathering and sediment cycling in the presence of liquid water shortly after Earth accretion. ¹⁷⁶Hf/¹⁷⁷Hf ratios of Hadean Jack Hills zircons show large heterogeneities indicating a major differentiation of the silicate Earth by 4.50 Ga. A possible consequence of this differentiation is the formation of continental crust of similar order to the present. Studies of mineral inclusions within Hadean zircons indicate their crystallization from hydrous, granitoid magmas at pressures greater than 6 kbars, implying low near-surface geothermal gradients which in turn suggests their origin in underthrust environments. Given general agreement that life could not have emerged until liquid water appeared at or near Earth’s surface, a significant implication is that our planet may have been habitable as much as 500 Ma earlier than previously thought. Indeed, carbon isotopic evidence obtained from inclusions in a Hadean zircon is consistent with life having emerged by 4.1 Ga, or several 100 million years earlier that the hypothesized lunar cataclysm. Trace element analyses of aluminum, halogens, sulfur, phosphorus, rare earth elements in Hadean zircons are consistent with their origin in a range of granitoid magma types and redox conditions. Although some of the above interpretations remain subject to debate, there is now a widespread consensus that molecular water was present at or near Earth’s surface since at least 4.3 Ga. Perhaps the most remarkable feature of inferences drawn from investigations of these ancient zircons is that none were predicted from theory, underscoring the importance of observations in testing models of early Earth.

Hadean zircons have been documented from fifteen terrestrial localities in Australia, Asia, Africa, and North and South America, in stony and martian meteorites, and in lunar rocks. Extraterrestrial zircons are characterized by the absence of the positive Ce anomaly, seen in virtually all terrestrial zircons, much higher formation temperatures, and a unique suite of mineral inclusions. Remarkably little effort has been directed toward characterizing the geochemical nature of Hadean zircons from terrestrial localities beyond the Jack Hills region and thus it remains unclear how representative it is of the Hadean world. A massive analysis campaign is indicated to better understand Earth’s last true ‘dark age’.

Any successful geodynamic or environmental model for early Earth must be consistent with ten robust lines of evidence derived from geochemical and petrologic observations of Hadean Jack Hills zircons. These are: (1) a zircon sub-population enriched in ¹⁸O and depleted in ³⁰Si relative to mantle values; (2) low crystallization temperatures; (3) the presence of primary hydrous mineral inclusions; (4) the predominance of magmatic muscovite, quartz, and biotite inclusions; (5) zircon formation in relatively low heat flow environments; (6) sub-chondritic initial ¹⁷⁶Hf/¹⁷⁷Hf ratios consistent with source isolation as early as 4.50 Ga; (7) fission Xe isotope compositions indicating variable fractionation of Pu from U; (8) the absence of ultra-high pressure mineral inclusions; (9) zircon formation under a wide range of redox conditions; and (10) geochemical signatures diagnostic of felsic continental crust. Numerous models have been proposed to explain these characteristics, including an origin similar to Icelandic rhyolites or lunar KREEP terranes, crystallization from mafic igneous rocks, formation in impact melts or sagduction, plate boundary and heat pipe tectonic environments, and multi-stage scenarios involving several of these mechanisms. While an origin of Jack Hills Hadean zircons in felsic and intermediate granitoids in a plate-boundary-type setting is consistent with all ten geochemically-derived constraints, competitor models are either only partially consistent or inconsistent with the evidence.

Given the absence of a macroscopic Hadean rock record, evaluating terrestrial habitability is largely a thought experiment, but data from Hadean zircons can provide some constraints. We are certain that life as we know it would not be possible without four requirements; soluble bioactive elements (carbon, hydrogen, oxygen, nitrogen, sulfur and phosphorous), free energy, liquid water, and time. Beyond these essential ingredients, there is broad agreement that there are ten secondary factors that separate us from the other, uninhabited terrestrial planets and maintain our planet’s homeostasis. They are: (1) a galactic and planetary sanctuary for life; (2) liquid water at the planetary surface to mediate biochemistry and efficiently cool the planet; (3) dissolved water in the deep planetary interior to enhance mantle circulation and catalyze the eclogite transition; (4) a broadly solar chemical composition to provide sufficient metallicity for a stable surface platform; (5) sufficient planetary mass to retain an atmosphere and heat; (6) planetary satellite(s) to stabilize climate zones; (7) extra-planetary impactors to introduce organic building blocks and water and to create satellites; (8) long-term interior heat generation to maintain mantle circulation and the geodynamo; (9) a self-sustaining dynamo to protect the atmosphere is erosion; and (10) a mechanism to recycle surface carbon into the interior and back. Evaluating how these various factors interact is complicated but our speculations can be guided by inferences from Hadean zircon geochemistry which potentially bear on six of the ten ingredients for life—the presence of surface and interior water, the role of impacts on early Earth, internal heat generation, surface recycling, and the existence of a Hadean geodynamo. Knowledge of the geochemistry and inclusion population of Hadean zircons also permits constraints to be placed on whether mineral phases and trace elements key to biopoiesis were present during the Hadean eon.

This chapter summarizes what is known about the timing of the emergence of life on Earth from the morpho- and chemo-fossil (chemical and isotopic signals remaining from the decomposition of living organisms) records. The geologic record back to ca. 3.5 billion years includes low grade sedimentary rocks in which organic residues of microbiota present during deposition have remained substantially intact. As different metabolic mechanisms variably fractionate carbon isotopes toward isotopically light values, a longstanding strategy has been to measure δ¹³C in these organic residues, or kerogens, for biologic signatures. When compared to carbon isotopes in inorganic carbonate rocks, a consistent offset is seen throughout the past 3.5 billion years with inorganic carbon averaging δ¹³C close to 0‰ and kerogens yielding δ¹³C of approximately −25‰. As the latter value is broadly characteristic of oxygenating photosynthesis, this relationship has been seen as evidence of past biologic activity. However, as metamorphic grade increases, kerogens are reacted to simpler hydrocarbons, ultimately yielding graphitic residues. The discovery of isotopically light carbon isotopes in microscopic graphite inclusions in rocks as old as ca. 3.83 billion years and in a 4.1 Ga zircon extends the possible emergence of life on this planet back into the Hadean eon. Although inorganic mechanisms exist that could potentially produce light δ¹³C signatures, these isotopic data are consistent with molecular clock calibrations of genomic mutations which suggest a lower bound for the time of life’s origin between 4.1 and 4.4 billion years.

How the scientific community, in the absence of any observational evidence, came to a consensus that the first many hundreds of millions of years of Earth history saw a desiccated, lifeless, molten wasteland is worthy of analysis. This chapter addresses problematic aspects of our epistemology that led to this paradigm and concludes that the historical sciences can sometimes be influenced by the same existential urges for control that fueled the past four thousand years of ubiquitous creation mythology. On a more tangible level, emerging scientific views in the late 1960s provided heretofore unavailable sources of thermal energy to models of the early solar system. This occurred just prior to the return of lunar highland samples which showed that Moon had almost been globally melted almost immediately upon formation, in stark contrast to the then prevailing paradigm of cold accretion. Arguably overreacting in overthrow of that model, the community quickly adopted the view that the first many hundreds of millions of years of Earth history had been too turbulent to leave any record. Shortly thereafter, the emergence of the large radius ion microprobe provided the tool needed to first discover and then explore Hadean zircons from Western Australia. The seemingly contradictory story these results presented would take a generation to seriously challenge the orthodoxy of a protracted, hellish world. This intellectual inertia partially reflects the inevitability of the Earth and planetary sciences being more tolerant of what amounts to untestable hypotheses relative to other physical sciences. While that is understandable given the challenge that historical sciences face in attempting to understand processes operating many billions of years ago, overly elaborate models invoking speculative mechanisms that lack the basis for falsification tend to crowd out other, possibly better, models from the scientific arena.