When Earth's core segregated from its silicate mantle roughly 4.5 billion years ago, it should have swept the mantle clean of siderophile elements. Iron-loving metals like platinum, iridium, and osmium have partition coefficients favoring metallic iron by factors of ten thousand to one million. Equilibrium partitioning predicts mantle concentrations approaching zero.

Yet the modern mantle contains highly siderophile elements at roughly 0.007 times chondritic abundances—far higher than equilibrium models permit, but suspiciously close to chondritic ratios. This mismatch between absolute depletion and relative preservation is the signature of a late veneer: a final pulse of material accreted after core formation had ceased.

The implications extend beyond geochemical curiosity. If a mass equivalent to roughly 0.5 percent of Earth arrived after core closure, this same population of impactors may have delivered volatiles, including a substantial fraction of the hydrosphere. The late veneer thus sits at the intersection of core formation dynamics, terrestrial planet accretion, and the origin of habitability—a narrow chronological window during which Earth's surface chemistry was decisively shaped.

Highly siderophile elements (HSEs)—the platinum group metals along with rhenium and gold—exhibit metal-silicate partition coefficients (D_met/sil) exceeding 10⁴ under most conditions relevant to core formation. During Earth's differentiation, these elements should have followed iron into the core with near-total efficiency, leaving mantle abundances essentially undetectable.

Instead, primitive mantle estimates derived from spinel peridotites and komatiite parental melts yield HSE concentrations at roughly 0.7 percent of CI chondrite values. This is depleted, but not nearly depleted enough. More striking is the pattern: the relative abundances of Os, Ir, Ru, Pt, Pd, and Re preserve chondritic ratios to within analytical uncertainty for most elements.

This dual observation constrains the geochemistry severely. Equilibrium partitioning at any plausible pressure-temperature-oxygen fugacity condition fractionates HSEs from one another, because their partition coefficients differ substantially. A mantle equilibrated with core-forming metal should show non-chondritic HSE ratios—yet it does not.

Alternative hypotheses have been explored and largely constrained. Inefficient core formation, delayed metal-silicate equilibration, or exotic partitioning behavior at deep magma ocean conditions can reduce absolute depletion, but struggle to reproduce chondritic ratios across all HSEs simultaneously. The convergence toward chondritic patterns strongly suggests a physical addition of undifferentiated chondritic material.

The ¹⁸⁷Os/¹⁸⁸Os systematics reinforce this picture. Mantle osmium isotopes fall between enstatite and ordinary chondrite reservoirs, indicating that whatever contributed the late HSE budget shared isotopic affinities with known meteoritic materials rather than more exotic outer solar system compositions.

Takeaway

When a system preserves the fingerprints of its ingredients despite processes that should have erased them, the ingredients likely arrived after the erasing was done.

Quantifying the late veneer requires assuming its composition. If the added material approximated CI or CM chondrite in HSE content—between roughly 480 and 660 ppb total platinum group elements—then mass balance calculations yield remarkably consistent estimates.

The primitive mantle contains approximately 3.5 ppb Os, 3.5 ppb Ir, and 7.1 ppb Pt. Reconciling these concentrations with a chondritic delivery requires adding a mass equivalent to 0.3 to 0.8 percent of Earth's total mass to the silicate reservoir after core formation ceased. Central estimates cluster near 0.5 percent, or roughly 2 × 10²² kg.

This mass is significant but not implausible for the tail end of accretion. Dynamical models of terrestrial planet formation predict a stochastic late accretion phase dominated by a few large impactors following the main growth epoch. N-body simulations by Bottke and colleagues suggest the largest post-core-formation impactor on Earth may have been 2500-3000 km in diameter—approaching lunar mass.

Comparative planetology sharpens these constraints. The Moon's mantle shows HSE abundances roughly twenty times lower than Earth's, implying a much smaller late accreted mass—consistent with gravitational cross-section scaling and the shorter interval between core formation and cessation of accretion on the smaller body. Mars presents an intermediate case, with SNC meteorite data indicating a martian late veneer of perhaps 0.7 percent by mass.

The Earth-Moon HSE disparity is diagnostic. If both bodies accreted late material from the same stochastic reservoir over the same timeframe, mass-scaled expectations predict a ratio near 1200:1. The observed ratio of roughly 1500:1 is broadly consistent, though sensitive to how much lunar HSE inventory is retained versus lost to impacts.

Takeaway

Half a percent of a planet's mass sounds trivial, but arriving at the right moment, it can rewrite surface chemistry entirely.

If the late veneer delivered chondritic material, it necessarily delivered volatiles alongside HSEs. Carbonaceous chondrites contain 5-10 percent water by mass in their hydrated silicate phases; ordinary chondrites contain roughly 0.1 percent. The compositional character of the veneer therefore determines its volatile contribution.

A late veneer of 0.5 percent Earth mass composed of CI-like material would deliver approximately 1 × 10²¹ kg of water—somewhat less than Earth's surface hydrosphere (~1.4 × 10²¹ kg) but far less than plausible mantle water inventories, which may exceed several ocean masses stored in nominally anhydrous minerals and hydrous phases at depth.

The isotopic evidence is more ambiguous than the mass balance suggests. Earth's bulk water D/H ratio (approximately 1.5 × 10^-4) matches carbonaceous chondrites closely and differs markedly from most comets, favoring an asteroidal rather than cometary source. However, this signature could reflect earlier accretion of hydrated planetesimals rather than a late-arriving veneer.

Ruthenium isotopes offer a discriminating test. The ¹⁰⁰Ru/¹⁰¹Ru signature of Earth's mantle matches enstatite chondrites—which are dry—rather than carbonaceous chondrites. If the late veneer were dominated by wet outer solar system material, mantle Ru isotopes should record it. They apparently do not, at least not dominantly.

The most defensible synthesis: the late veneer probably contributed less than 10-30 percent of Earth's water inventory, with the bulk of terrestrial hydrogen delivered during main-stage accretion by hydrated planetesimals from beyond the snow line. The veneer's chemical fingerprint on HSEs is diagnostic; its role in the water budget is real but likely secondary.

Takeaway

The same event can dominate one geochemical signal while contributing marginally to another—single-cause explanations rarely survive isotopic scrutiny.

The late veneer occupies a peculiar position in Earth's history: chronologically brief, mass-wise modest, but geochemically decisive. It resolved the highly siderophile element puzzle by delivering chondritic material after the core had ceased scavenging metals, preserving relative abundance patterns that equilibrium partitioning would have destroyed.

Its role in habitability remains more nuanced. While ruthenium and osmium isotopes point toward relatively dry inner solar system sources, the veneer's volatile contribution—even if secondary to earlier accretion—may have delivered the surface reservoirs from which early oceans condensed.

As exoplanet studies mature, understanding when and how terrestrial worlds acquire their final chemical inventories becomes essential. The late veneer suggests that planetary habitability may depend not only on bulk composition but on the stochastic timing of the last few impactors—a sobering thought for statistical predictions of habitable worlds.