Enhanced Weathering: Grinding Rock to Cool the Planet

The reaction: carbonic acid eats rock, and the carbon goes to the sea

Every raindrop is a weak acid. As it falls, CO₂ dissolves into it and hydrates: CO₂ + H₂O ⇌ H₂CO₃ ⇌ H⁺ + HCO₃⁻, giving rain a natural pH near 5.6 even with no pollution. Those protons are what attack the silicate framework. For the archetypal enhanced-weathering mineral forsterite (the magnesium olivine, Mg₂SiO₄), the balanced dissolution is:

• Mg₂SiO₄ + 4 CO₂ + 4 H₂O → 2 Mg²⁺ + 4 HCO₃⁻ + H₄SiO₄

Read that stoichiometry carefully. Four molecules of CO₂ are consumed per formula unit of olivine, and the carbon is not destroyed — it is converted into four bicarbonate ions (HCO₃⁻) dissolved in the drainage water. All four of those carbons come from the CO₂ (the silicate rock carries none of its own), though only about two per formula unit stay locked once the ocean re-equilibrates and precipitates carbonate; the framework silicon leaves as harmless silicic acid (H₄SiO₄). For calcium silicates the algebra is the same shape — wollastonite gives CaSiO₃ + 2 CO₂ + H₂O → Ca²⁺ + 2 HCO₃⁻ + SiO₂, and calcium plagioclase (anorthite) weathers as CaAl₂Si₂O₈ + 2 CO₂ + 3 H₂O → Ca²⁺ + 2 HCO₃⁻ + Al₂Si₂O₅(OH)₄ (kaolinite clay).

The dissolved Mg²⁺ and Ca²⁺ plus bicarbonate wash through soil, into rivers, and out to the ocean. There the carbon is stored on two timescales: as bicarbonate in the vast marine dissolved-inorganic-carbon reservoir (~38,000 Gt C, residence time >10⁵ yr), and eventually as solid CaCO₃ when marine organisms build shells, sequestering it in seafloor limestone for millions of years.

Why silicate weathering, and not limestone, actually removes carbon

A common trap: if limestone dissolves in carbonic acid too, why not just spread crushed limestone? The answer is in the carbon bookkeeping. Carbonate weathering runs as CaCO₃ + CO₂ + H₂O → Ca²⁺ + 2 HCO₃⁻. Only one CO₂ is consumed, and half the carbon in the product bicarbonate was already inside the rock. When that Ca²⁺ and HCO₃⁻ reach the ocean and reprecipitate as shell CaCO₃ — Ca²⁺ + 2 HCO₃⁻ → CaCO₃ + CO₂ + H₂O — exactly one CO₂ is handed back to the atmosphere. Net over a full cycle: ≈ zero drawdown. Limestone is excellent for liming acidic soils, but it is not carbon removal.

Silicate weathering is different because the cation (Mg²⁺, Ca²⁺) arrives with no carbon of its own. When Ca²⁺ later reprecipitates as CaCO₃ and releases one CO₂, the ledger still shows a permanent loss of one atmospheric CO₂ per Ca²⁺ (two were consumed during weathering, one comes back). This asymmetry is precisely the silicate weathering thermostat that has counteracted volcanic CO₂ across geologic time. Enhanced weathering doesn't invent new chemistry — it dials the natural thermostat's rate up by orders of magnitude.

Surface area is everything: the kinetics of grinding

Olivine dissolution is proton-promoted and rate-limited by the mineral surface. The dissolution rate scales roughly as R ∝ a(H⁺)ⁿ with a reaction order n ≈ 0.5 near neutral pH, and it rises steeply with temperature — the Arrhenius activation energy for olivine is around Eₐ ≈ 60–80 kJ/mol, so a warm tropical soil weathers several times faster than a cold temperate one. But the dominant lever a geoengineer controls is specific surface area, which is set by particle size.

• A 1 mm gravel grain has a specific surface area near 0.002 m²/g.
• Milled to 10 μm, the same mass exposes roughly 0.2–1 m²/g — a 100–500× jump.
• Because dissolution is surface-controlled, cutting grain diameter by 10× cuts the weathering time by roughly 10×.

That is why enhanced weathering means fine powder, typically p80 below 100 μm and often <20 μm. The catch is energy: comminution to that fineness costs on the order of 10–30 kWh per tonne, and the CO₂ from that grinding (plus mining and transport) must be subtracted from the ~1.1 t CO₂/t captured. Life-cycle analyses put the net efficiency around 0.5–1.0 t CO₂ removed per t rock once those debits are paid — still strongly net-negative if powered cleanly.

The controlling variables in a real field

In the lab olivine dissolves briskly; in a soil, reality intrudes. Several coupled variables decide whether spread rock actually captures carbon at the projected rate:

• Water and drainage. Weathering is a reaction with water; a well-watered tropical cropland cycles far more carbonic acid past the grains than an arid one. Drainage also matters — the products (Mg²⁺, HCO₃⁻) must be flushed away, or the porewater saturates and the reaction stalls by the common-ion effect.
• pH. As protons are consumed, soil porewater pH climbs from ~5.5 toward 8–9. This is a genuine co-benefit (it counteracts soil acidification and can raise crop yields, replacing agricultural lime), but very high pH slows further dissolution and can precipitate secondary Mg-carbonates or clays that armor grain surfaces.
• Secondary minerals. Aluminous silicates like basalt shed clays (kaolinite, smectite) and iron oxides that can coat grains and throttle the reaction — a real-world reason field rates often fall 2–10× below idealized lab kinetics.
• Biology. Roots, mycorrhizal fungi, and microbes pump CO₂ and organic acids (oxalic, citric) into the rhizosphere, locally driving pH down and accelerating dissolution — a reason croplands and forest soils are favored targets over bare land.

Coastal olivine and the ocean-chemistry angle

A second deployment route skips the soil and puts olivine straight onto beaches and shelf seas. Here the dissolved products feed directly into seawater and touch the ocean carbonate buffer. Adding Mg²⁺/Ca²⁺ and consuming H⁺ raises seawater alkalinity, shifting the carbonate equilibrium CO₂ + CO₃²⁻ + H₂O ⇌ 2 HCO₃⁻ to the right and letting the surface ocean absorb more atmospheric CO₂ to re-equilibrate under Henry's law. This overlaps with ocean alkalinity enhancement, and it carries a bonus: it locally reverses ocean acidification, nudging surface pH back up from its stressed ~8.05 toward preindustrial ~8.17.

The chemistry is favorable but the caveats are chemical too. Olivine carries nickel and chromium impurities (often 0.2–0.4% Ni), which the dissolution releases into water — a real ecotoxicology constraint that caps how much can be spread. Wave energy helps by mechanically abrading grains and exposing fresh surface, but it also disperses fine dust before it fully reacts. And if seawater becomes locally supersaturated, dissolved Ca²⁺/Mg²⁺ can reprecipitate as carbonate, partially reversing the CO₂ uptake — the same runaway that the saturation horizon governs in the deep sea.

The planetary scale: how much rock, and how much cooling

Natural silicate weathering already removes on the order of 0.3 Gt CO₂ per year — the slow leak that balances volcanic outgassing over geologic time. Humanity emits about 37 Gt CO₂ per year from fossil fuels, roughly 100× faster. Enhanced weathering's ambition is to bridge part of that gap. Modeling studies suggest that spreading crushed basalt on the world's croplands (~12 million km²) at a few tonnes per hectare could remove on the order of 0.5–4 Gt CO₂ per year by mid-century — meaningful, but a supplement to emissions cuts, not a substitute.

The material demand is staggering: capturing 1 Gt CO₂ requires roughly 1–2 Gt of rock mined, milled, and spread — comparable to the entire global coal industry run in reverse. That is why basalt (abundant, low in nickel, already a mining by-product as quarry fines) is often preferred over higher-capacity but nickel-rich olivine, despite basalt's lower theoretical yield (~0.3 t CO₂/t). The permanence, though, is the real prize: unlike forests that can burn, weathering products stored as ocean bicarbonate and seafloor carbonate stay locked for 10⁴ to 10⁶ years — geological, not decadal, storage.

Silicate vs carbonate weathering as carbon-removal pathways

Property	Silicate (olivine/basalt)	Carbonate (limestone)
Example reaction	Mg₂SiO₄ + 4 CO₂ + 4 H₂O → 2 Mg²⁺ + 4 HCO₃⁻ + H₄SiO₄	CaCO₃ + CO₂ + H₂O → Ca²⁺ + 2 HCO₃⁻
CO₂ per mole rock	4 mol consumed	1 mol consumed, 1 mol was already in the rock
Net long-term drawdown	Yes — 2 CO₂ permanently removed per Mg₂SiO₄	≈ zero once carbonate reprecipitates
Reaction rate	Slow; needs fine grinding	Fast; dissolves readily
Permanence	10⁴–10⁶ yr (bicarbonate/ocean CaCO₃)	Re-releases CO₂ on reprecipitation
Best use	Genuine CDR	Liming/pH control, not net CDR