Carbon Capture: The Chemistry of Scrubbing CO₂ From the Air

Why 420 ppm makes capture so hard

CO₂ is a stable, closed-shell, nonpolar linear molecule (O=C=O), and at atmospheric abundance it is astonishingly dilute. At 420 ppm its partial pressure is only ≈0.042 kPa out of 101 kPa. Thermodynamics sets a hard floor on separating a dilute component: the minimum work to concentrate CO₂ from 420 ppm to pure is

• W_min = RT · ln(1/x) ≈ (8.314 J·mol⁻¹·K⁻¹)(298 K) · ln(1/0.00042) ≈ 19 kJ/mol, or about 0.43 GJ per tonne CO₂.
• From a 12% flue gas the same term drops to ≈5 kJ/mol — an order of magnitude cheaper, which is exactly why capturing at a smokestack is far easier than from open air.

Real plants run at 5–20× the thermodynamic minimum because reversible chemistry, heat losses, and fan/pump work all pile on. The engineering fight is to find a sorbent whose binding is strong enough to grab CO₂ at 0.042 kPa yet weak enough to release it without boiling the whole system. That is a Goldilocks problem in ΔG = ΔH − TΔS: you want ΔH negative enough for spontaneous uptake but small enough that a modest temperature swing flips the equilibrium.

Amine scrubbing: the carbamate handshake

The dominant post-combustion technology washes flue gas up a column against an aqueous amine — classically 30 wt% monoethanolamine (MEA), HOCH₂CH₂NH₂. The amine's lone-pair nitrogen (pKaH ≈ 9.5) is a Brønsted-Lowry base that grabs CO₂. In the absorber (~40 °C) the fast reaction is carbamate formation, which consumes two amine molecules per CO₂:

• Zwitterion step: R-NH₂ + CO₂ ⇌ R-NH₂⁺COO⁻
• Deprotonation: R-NH₂⁺COO⁻ + R-NH₂ ⇌ R-NHCOO⁻ + R-NH₃⁺
• Net: 2 R-NH₂ + CO₂ ⇌ R-NHCOO⁻ + R-NH₃⁺  (ΔH ≈ −85 kJ/mol)

At higher CO₂ loading, and with hindered or tertiary amines, the slower bicarbonate route takes over — CO₂ hydration catalyzed by the base:

• CO₂ + H₂O ⇌ H₂CO₃ ⇌ H⁺ + HCO₃⁻
• R-NH₂ + CO₂ + H₂O ⇌ R-NH₃⁺ + HCO₃⁻  (only 1 amine per CO₂ — a higher theoretical capacity)

The rich solution is pumped to a stripper and boiled with steam at 100–120 °C. Heat is Le Chatelier's lever: the exothermic uptake reaction reverses, releasing a nearly pure CO₂ stream (>99%) and regenerating the amine to cycle back. The catch is the regeneration duty, ≈3.5–4 GJ per tonne CO₂, which is why capturing a coal plant's exhaust can consume 20–30% of its electricity output — the notorious energy penalty.

Direct air capture: hydroxides and solid amines

Pulling CO₂ from ambient air needs a stronger base because the partial pressure is 100–300× lower than flue gas. Two chemistries dominate:

• Aqueous hydroxide (Carbon Engineering–style). Air blows through a solution of KOH: CO₂ + 2 OH⁻ → CO₃²⁻ + H₂O. The carbonate is then swapped onto calcium: Ca(OH)₂ + CO₃²⁻ → CaCO₃↓ + 2 OH⁻. The limestone pellets are heated in a calciner at ~900 °C to drive off CO₂: CaCO₃ → CaO + CO₂ (ΔH ≈ +178 kJ/mol), and the CaO is slaked back to Ca(OH)₂ with water. That 900 °C step is the reason this loop is energy-hungry.
• Solid supported amines (Climeworks–style). Amine-functionalized porous solids adsorb CO₂ as carbamate/bicarbonate at ambient temperature, then release it with temperature-vacuum swing at only ~80–100 °C — low-grade heat, which is why Climeworks' Orca and Mammoth plants run on geothermal heat in Iceland.

Either way DAC costs 5–10 GJ per tonne and roughly $250–600/t today, because you cannot cheat the 19 kJ/mol thermodynamic floor and you must move enormous volumes of air. Global DAC capacity in the mid-2020s was still only tens of thousands of tonnes per year — a rounding error against ~37 billion tonnes of annual fossil CO₂.

Turning CO₂ into stone

The most permanent fate for captured carbon is to mimic — but vastly accelerate — the silicate weathering thermostat that regulates Earth's climate over geologic time. Basalt and peridotite are rich in Ca²⁺, Mg²⁺, and Fe²⁺ silicates that react with carbonic acid to precipitate stable carbonate minerals:

• Forsterite: Mg₂SiO₄ + 2 CO₂ → 2 MgCO₃ + SiO₂ (magnesite)
• Wollastonite: CaSiO₃ + CO₂ → CaCO₃ + SiO₂ (calcite)
• In-situ (CarbFix, Iceland): CO₂ is dissolved in water and injected into basalt, where it weathers the rock and mineralizes as calcite and ankerite.

The CarbFix pilots showed something startling: >95% of injected CO₂ mineralized within about two years, not the centuries once assumed. Once it is carbonate rock it is locked for >10,000 years and cannot leak. The catch is stoichiometric mass: fixing one tonne of CO₂ needs roughly 2–3 tonnes of reactive rock ground fine enough to react, plus large volumes of water. Mineralization is thermodynamically downhill (ΔG < 0, exothermic) — nature does it for free — but kinetically slow, so the whole enterprise is a race to speed a reaction that would otherwise take millennia.

The controlling variables: temperature, pKa, and loading

Every capture cycle is an exercise in tuning chemical equilibrium. The knobs that matter:

• Temperature swing. Because uptake is exothermic (ΔH ≈ −60 to −90 kJ/mol), heating shifts equilibrium toward release. The van 't Hoff relation, d(ln K)/d(1/T) = −ΔH/R, means a larger |ΔH| gives a sharper temperature response but a bigger regeneration bill. Designers hunt for the sweet spot near −50 to −70 kJ/mol.
• Sorbent basicity (pKaH). A stronger base binds dilute CO₂ better but resists release. The amine window sits around pKaH 9–10; hydroxides (effective pKa ~15.7 for OH⁻/H₂O) bind hard enough for air but need 900 °C to regenerate.
• Working capacity (loading). Carbamate chemistry caps at ~0.5 mol CO₂ per mol amine (two amines per CO₂); the bicarbonate route can reach ~1.0 but is slower. Blended amines (e.g. MEA + methyldiethanolamine + a piperazine promoter) split the difference — fast kinetics plus high capacity.
• Degradation and corrosion. MEA oxidizes and reacts irreversibly with SO₂, NO₂, and O₂ in flue gas, forming heat-stable salts and corrosive products; make-up amine and reclaiming are real operating costs.

Why it matters for climate — and where it fits

Atmospheric CO₂ has climbed from ~280 ppm preindustrial to ~420 ppm, and roughly a quarter of every year's emissions dissolves into the sea, pushing surface-ocean pH down about 0.1 units (from ~8.2 to ~8.1) — a 30% rise in H⁺ that stresses the carbonate-shell builders of the marine biological pump. Capture technology addresses the source rather than the symptom, in two distinct roles:

• Point-source CCS catches CO₂ from cement kilns, steel mills, and gas plants — sectors with no easy electric substitute. Cement alone emits ~8% of global CO₂, half of it chemically unavoidable from CaCO₃ → CaO + CO₂ in the kiln, so capture is essentially the only decarbonization route.
• Negative emissions (DAC + mineral storage) remove CO₂ already in the air, the only way to draw down legacy carbon or offset aviation and agriculture.

The honest scale check: to matter for climate, capture must reach billions of tonnes per year. In the mid-2020s all operating CCS captured ~50 million tonnes annually — about 0.1% of emissions. Capture is necessary but not a license to keep emitting: at ~$300/t, offsetting a single trans-Atlantic flight's CO₂ via DAC costs more than the ticket, a stark reminder that not emitting is almost always cheaper than un-emitting.

Three routes to catching CO₂, compared by the chemistry that binds it

Property	Amine scrubbing (post-combustion)	Direct air capture (DAC)	Mineral carbonation / storage
CO₂ concentration handled	~4–14% (flue gas)	~0.042% (ambient air)	Injected supercritical CO₂
Binding chemistry	Carbamate/bicarbonate with amine	KOH/Ca(OH)₂ or solid amine on air	Ca²⁺/Mg²⁺ silicates → carbonate
Energy to regenerate	≈3.5–4 GJ/t CO₂ (100–120 °C steam)	≈5–10 GJ/t CO₂ (up to 900 °C calciner)	Exothermic; ΔH ≈ −60 to −90 kJ/mol
Permanence	Depends on where CO₂ is stored	Depends on storage	>10,000 years (rock)
Cost per tonne CO₂ (2020s)	≈$40–90	≈$250–600	≈$10–30 (in-situ basalt)