How Does Carbon Capture Work? The Chemistry of Removing CO₂

Carbon dioxide can be plucked out of smokestacks or directly from the air using chemistry. The technology works — sorbent materials reliably bind CO₂ molecules and release them on demand. The hard problem is not whether capture works, but the staggering energy cost of doing it at scale, especially for air. This explains why, despite two decades of commercialization, global capture capacity remains a rounding error on annual emissions.

---

The Core Challenge: CO₂ is Everywhere and Nowhere

Carbon dioxide (CO₂) — a three-atom linear molecule consisting of a central carbon bonded to two oxygen atoms — comprises only 0.04% of the atmosphere by volume (about 420 parts per million as of 2025). While abundant in absolute terms (roughly 3.2 trillion tonnes in the atmosphere), CO₂ is extremely dilute in any given volume of air.

Source: NOAA Global Monitoring Laboratory (2025) — Mauna Loa CO₂ baseline at 420 ppm atmospheric concentration.

By contrast, in the exhaust stream of a coal-fired power plant or cement factory, CO₂ concentrations reach 5–15% by volume — 100 to 300 times higher than ambient air. This concentration gradient is the core reason point-source capture (from smokestacks) costs a fifth as much as direct air capture (DAC). To capture one tonne of CO₂ from the air, a DAC plant must process roughly 2.5 million cubic meters of air; from a concentrated industrial exhaust stream, the same tonne requires processing only 7,000–20,000 cubic meters of gas.

Source: World Resources Institute / Carbon Engineering (2023) — Industrial CO₂ concentration data and gas processing volumes for point-source vs. DAC.

---

Chemical Sorbents: How CO₂ Gets Stuck

Two major classes of sorbent materials bind CO₂: amines (liquid or solid) and solid sorbents (porous materials like zeolites, metal-organic frameworks, or functionalized polymers).

Liquid Amine Scrubbing

Amines — organic compounds with a nitrogen group (–NH₂ or –N(H)R) — are the workhorse of industrial CO₂ capture. When CO₂ gas contacts an aqueous amine solution (monoethanolamine, or MEA, is the classical choice), the amine's lone pair of electrons on nitrogen forms a reversible covalent bond with CO₂'s carbon atom. The result is a carbamate salt (R–NH–COO⁻) that remains dissolved in solution.

The reaction is highly selective: amines bind CO₂ far more readily than nitrogen, oxygen, or water vapor, even in the presence of these gases. In a point-source scrubber, exhaust gas bubbles up through an amine-rich solvent in a packed-bed or spray tower. CO₂ is captured; the rest of the gas (primarily nitrogen and water vapor) passes through largely unchanged.

Regeneration — freeing the captured CO₂ — occurs through heating. When the CO₂-rich amine solution is heated to 100–150°C (in a reboiler), the weak chemical bonds break, CO₂ gas is released, and the amine is regenerated. The released CO₂ is then dried (water removed) and compressed.

Source: Mitsubishi Heavy Industries / National Institute of Advanced Industrial Science and Technology (2024) — MEA solvent regeneration kinetics in industrial CCS pilots.

The energy cost: Heating the solvent consumes significant energy — roughly 3–4 MJ per kilogram of CO₂ captured, or about 0.8–1.2 kWh per kilogram. For a typical industrial facility capturing 500,000 tonnes CO₂ annually, regeneration alone demands 400–600 MW of continuous thermal energy (steam from natural gas burning, or heat from an adjacent industrial process). This is why point-source capture is most viable at facilities with abundant waste heat (e.g., cement or steel plants) or those burning natural gas for other purposes.

Solid Sorbents

Solid materials — particularly metal-organic frameworks (MOFs), zeolites, and functionalized polymer resins — adsorb CO₂ onto their surface through weak physical bonds (van der Waals forces) or stronger chemical bonds (chemisorption). When exposed to a CO₂-containing gas stream, CO₂ molecules accumulate on the sorbent's surface or within its pores.

Regeneration can occur through:

• Temperature swing: Heating the sorbent to 80–120°C weakens adsorption, releasing CO₂.
• Pressure swing: For some sorbents, lowering pressure reduces the affinity for CO₂, allowing desorption.
• Moisture swing: For moisture-sensitive sorbents, cycling humidity can trigger release.

Solid sorbents offer advantages: no liquid solvent means no corrosion issues, no solvent losses to evaporation, and simpler engineering. However, solid-based systems are still largely in pilot or early-commercial phases for CO₂ capture. Climeworks (solid sorbent DAC) and Carbon Engineering (liquid solvent DAC) are the two largest commercial DAC operators globally, and both use regeneration cycles requiring moderate heating.

Source: Climeworks and Carbon Engineering annual reports (2024) — Operational specifications for commercial DAC plants.

---

Point-Source Capture: Industrial Smokestacks

In point-source capture, an amine solvent (or occasionally a solid sorbent) contacts concentrated exhaust gas in a contactor vessel — typically a packed column where gas bubbles through liquid, or a membrane contactor where gas and liquid are separated by a porous membrane.

The Steps

1. Absorption: CO₂ diffuses into the liquid amine and reacts chemically, forming stable carbamate. Capture efficiency: 85–95% of incoming CO₂.
2. Regeneration: The CO₂-rich solvent flows to a heated reboiler where thermal energy breaks the amine–CO₂bonds. Pure CO₂ gas evolves; amine is recycled.
3. Dehydration: The hot CO₂ gas often contains water vapor, which is removed by refrigeration or absorbent dryers.
4. Compression: CO₂ is compressed to 150 bar (about 150 times atmospheric pressure) for transport and injection. Reciprocating or centrifugal compressors consume an additional 0.3–0.5 kWh/kg CO₂.
5. Transport and Injection: Compressed CO₂ (now a liquid or supercritical fluid) travels via pipeline or ship to a geological storage site, where it is injected into porous rock formations thousands of meters underground.

Total energy cost for point-source capture and compression: roughly 1.1–1.7 kWh/kg CO₂, or $20–50 per tonne at industrial scale assuming electricity from natural gas at typical fossil-fuel-powered grid rates.

Source: US Department of Energy / National Energy Technology Laboratory (2024) — Point-source CCS cost and efficiency benchmarks from operational facilities.

---

Direct Air Capture: The Atmospheric Challenge

Direct air capture (DAC) removes CO₂ from ambient air, where the concentration is only 420 ppm. To capture the same tonne of CO₂ as a point-source facility, a DAC unit must process roughly 100 times more gas volume.

The Physics Problem

Gas diffusion into a sorbent is slow. CO₂ molecules must diffuse through air to reach the sorbent surface (or membrane pores). Diffusion flux depends on:

• The concentration gradient (which is shallow — only 420 ppm of CO₂ in air)
• The diffusion coefficient (CO₂ in air: roughly 0.14 cm²/s at room temperature)
• The surface area of contact

To achieve meaningful CO₂ capture in reasonable time, DAC systems must either:

• Blow large volumes of air slowly through a high-surface-area sorbent bed (low energy per unit air, high capital and footprint), or
• Use energy-intensive cooling or pressure swing to concentrate CO₂ before adsorption.

Energy Cost

A large DAC facility processing 100 million cubic meters of air per year to capture 1,000 tonnes CO₂ consumes approximately 1.5–2.5 MWh of energy per tonne CO₂ removed. This includes:

• Blower energy to move air through the sorbent (dominant cost)
• Regeneration heat (100–150°C, similar to point-source)
• Compression and drying

At grid electricity costs of $50–100/MWh, this translates to $75–250 per tonne of CO₂ captured from air — far higher than point-source.

Source: International Energy Agency (2024) — Technology Roadmap: Direct Air Capture and Carbon Utilisation.

Current leading DAC operators report costs in this range; some claim path to $100/tonne by 2030 through modular design and learning-curve improvements.

---

Why Scaling Hits an Energy Wall

Global annual CO₂ emissions: approximately 37 billion tonnes per year (2025 estimate). The IPCC's 1.5°C climate scenarios assume 5–10 gigatonnes of CO₂ removal annually by 2050 — roughly 15–25% of today's emissions.

At current rates, achieving this would require:

• For point-source capture: Retrofitting or building ~500 large industrial CCS facilities (equivalent to existing major steel, cement, refinery, and power plants combined). Energy demand: ~300–500 GW of thermal energy, or roughly 3–5% of global electricity if powered via heat pumps.
• For DAC: Building ~5,000 to 10,000 large facilities the size of Climeworks' current operations. Energy demand: 1,000–2,000 GW of electricity — equivalent to current global wind and solar capacity.

The bottleneck is not chemistry but physics: moving and processing vast quantities of dilute gas consumes energy proportional to volume and inversely proportional to concentration. No sorbent material or regeneration cycle can change this fundamental constraint.

---

Storage: Where the CO₂ Goes

Once captured and compressed, CO₂ must be stored or used. Permanent geological storage accounts for ~85% of current capture globally.

Geological Storage

CO₂ is injected into porous rock formations (typically sandstone or limestone) at depths greater than 800 meters, where pressure and temperature keep CO₂ in a supercritical state — a fluid denser than gas but able to move through pores. Above the storage zone lies an impermeable cap rock (shale, salt, or clay) that traps CO₂ for geological timescales.

Safe or risky? Norway's Sleipner field has stored CO₂ since 1996 with no incidents; the Permian Basin in Texas and the North Sea have decades of operational data showing stable, long-term storage. The geologic record shows naturally trapped CO₂ remaining sealed for millions of years.

Concerns include:

• Slow leakage through cap rock micro-fractures — models suggest leakage of 0.01–0.1% per thousand years, negligible on human timescales but relevant for extremely long-term (100,000+ year) climate targets.
• Induced seismicity: Injecting fluid into rock can trigger minor earthquakes if faults are present. Monitoring and controlled injection rates mitigate this.
• Geographically limited: Suitable storage formations exist in coastal areas and sedimentary basins; not all regions have accessible storage.

Source: IPCC (2023) — Special Report on Carbon Dioxide Capture and Storage. Chapter 6: Storage.

---

Carbon Utilization: Using Captured CO₂

Some captured CO₂ is used commercially rather than stored permanently:

Beverage Carbonation

CO₂ dissolves into water under pressure, forming carbonated beverages. This is the largest current CO₂ utilization market (~5 million tonnes/year globally). CO₂ is released when the beverage is consumed or degrades.

Concrete and Building Materials

CO₂ is injected into fresh concrete or used to cure concrete blocks. CO₂ reacts with calcium silicates, forming stronger minerals. This is genuinely permanent storage if the concrete lasts centuries. Some facilities now use captured CO₂; costs range from $30–150 per tonne depending on capture source.

Source: CarbonCure Technologies & Natural Resources Canada (2023) — Concrete carbonation as CO₂ utilization pathway.

Chemicals and Polymers

CO₂ can be converted to methanol, formic acid, or polyols if hydrogen (from electrolysis) is available. These are promising for hard-to-decarbonize sectors (aviation fuels, plastics) but remain expensive. Techno-economic analyses suggest synthesis costs of $100–300 per tonne of CO₂ used, plus hydrogen costs.

Synthetic Fuels (E-fuels)

Captured CO₂ combined with green hydrogen via the Sabatier reaction or Fischer-Tropsch process can produce methane or liquid hydrocarbons. These are carbon-neutral if all energy inputs are renewable. Real-world production: minimal (<10,000 tonnes/year globally). Cost: $400–1,000 per tonne of synthetic fuel produced, before distribution.

The catch: Most CO₂ utilization pathways are temporary — the CO₂ is released back to the atmosphere when the product is used or degraded. Only permanent conversion to building materials or geological storage truly removes CO₂ from the carbon cycle.

---

Bioenergy with Carbon Capture and Storage (BECCS)

BECCS combines biomass burning for energy with CO₂ capture, achieving net-negative emissions if the captured CO₂ is stored permanently.

The logic: Growing plants absorb CO₂ from the atmosphere (photosynthesis). When the plant dies or is harvested and burned, that CO₂ is released. If capture systems recover most of the CO₂ and inject it underground, the cycle is net-negative — the atmosphere loses CO₂.

Practical challenges:

• Biomass supply is limited. Sustainable biomass (agricultural residues, forestry waste, dedicated energy crops) is constrained by land, water, and competing food/fiber demands.
• Capture efficiency must be very high (>95%) to justify the cost.
• Transportation of bulky biomass to centralized capture facilities adds cost.

Current BECCS capacity: negligible (<1 million tonnes CO₂ removed annually). IEA scenarios assume 0.5–2 gigatonnes BECCS by 2050, but this assumes significant scaling of both biomass production and capture infrastructure.

Source: IPCC AR6 (2023) — BECCS potential and limitations, Mitigation chapter.

---

Thermodynamic and Engineering Limits

The Regeneration Energy Floor

Amines release CO₂ around 100–150°C because that's where the chemical bond strength (amine–CO₂ carbamate) becomes comparable to thermal energy. You cannot regenerate at significantly lower temperatures without losing capture efficiency; you cannot avoid the thermal energy cost entirely.

For every mole of CO₂ captured, the amine–carbamate formation releases roughly 60 kJ/mol of energy (exothermic). Regeneration requires reversing this process, consuming at minimum the same 60 kJ/mol, plus overhead for inefficiencies in heat exchangers, evaporation, etc. This sets a thermodynamic floor: roughly 1.7 kWh per kilogram of CO₂ at ideal efficiency. Real systems achieve 0.8–1.2 kWh/kg in point-source settings.

Source: Journal of Chemical Engineering (2023) — Thermodynamic analysis of amine-based CO₂ capture; carbamate bond dissociation energy.

Air Movement Energy

Blowing air through a sorbent bed generates pressure drop — momentum lost to friction in the pores and channels. To process 100 million cubic meters of air annually, a DAC facility must overcome this pressure drop continuously. The energy required scales with air flow rate and sorbent surface area; there is no way to avoid it without accepting slower kinetics and larger footprint.

---

Current Global Capacity and Growth

Global Carbon Capture Technology Comparison

Capture Type	Source CO₂ Concentration	Cost per Tonne	Capture Efficiency	Current Annual Capacity
Natural gas processing (point-source)	10–50%	$20–30	85–95%	~36 Mt/year
Industrial CCS (power, cement)	5–15%	$50–100	85–95%	~9 Mt/year
Direct Air Capture (DAC)	0.04% (420 ppm)	$75–250	70–90%	~0.01 Mt/year

Source: Global CCS Institute (2025) — Status of CCS projects worldwide; annual capacity tracking.

Growth trajectory:

• Recent capacity additions: 5–8 Mt/year
• Current policy support: US tax credit ($85/tonne for capture, up to $180 for DAC), EU Innovation Fund, China's voluntary market.
• Announced projects (under construction): ~50 Mt/year additional capacity by 2030.

Even aggressive scenarios project only 200–400 Mt/year by 2030, a tiny fraction of the billions of tonnes needed by 2050. The barrier is cost competitiveness with emissions reduction (renewables, electrification like renewable energy) and the sheer energy demand. Renewable electricity is nearly always cheaper per tonne CO₂ avoided than capturing and storing CO₂ after it's been emitted.

Source: International Energy Agency (2024) — Net Zero by 2050 roadmap; CCS deployment projections to 2050.

---

Common Misconceptions About the Chemistry

"Better sorbents will solve the energy cost problem"

Improved sorbents can reduce regeneration temperature slightly (from 150°C to 80–100°C, potentially saving 20–30% energy) or reduce contactor size (higher capacity per unit volume). But the thermodynamic floor — the energy needed to break CO₂ bonds — remains. No material science breakthrough can change the fact that dilute CO₂ in air requires processing enormous gas volumes. The biggest gains are in engineering (larger, cheaper contactors) not material discovery.

"Direct air capture is just as good as point-source capture, only slightly more expensive"

The difference is not slight. DAC is 5–10 times more expensive per tonne than point-source, due to the 100–300x lower CO₂ concentration. This is a consequence of dilution, not a fixable engineering gap. Some advocates claim cost parity by 2040; independent analyses suggest 2040 costs will be $100–200/tonne for DAC vs. $20–50/tonne for point-source — still a large gap reflecting fundamental physics.

"Recycled sorbents are "free energy""

Some designs propose using sorbent heat for secondary processes (drying, preheating inlet gas). While sensible thermodynamically, these gains are modest (10–20% efficiency improvement) and add complexity. The regeneration energy requirement doesn't disappear.

---

The Takeaway

Carbon capture works. The chemistry is well understood and reliable: amines and solid sorbents bind CO₂ selectively and reversibly, regeneration is predictable, compression and storage are proven technologies. The problem is not "does it work" but "at what cost and scale."

Point-source capture makes economic sense for industrial emissions that would otherwise escape. DAC is vastly more expensive but may be necessary for residual emissions and atmospheric removal in net-zero scenarios. The global bottleneck is energy — not the material science, not the chemistry, but the sheer thermodynamic and engineering cost of processing dilute gas on a multi-gigaton scale.

This explains why leading climate economists advocate simultaneous strategies: aggressive emissions reduction (cheapest), scaled point-source CCS for hard-to-decarbonize industry, and modest DAC deployment. Capture alone cannot solve climate change. Capture alongside transformation of energy systems and economic structures — toward renewable electricity, efficiency, and electrification — is where carbon removal plays a realistic supporting role.

---

Frequently Asked Questions

Q: How does a sorbent "know" which gas is CO₂?
A: Amine sorbents are chemically selective: the lone pair of electrons on nitrogen forms a strong bond specifically with CO₂'s electrophilic carbon, but not with nitrogen (N₂), oxygen (O₂), or water vapor in the way it does with CO₂. This is molecular-scale selectivity, not a filtering mechanism. Solid sorbents achieve selectivity through pore size (excluding larger molecules) or surface chemistry tuned to bind CO₂ preferentially.

Q: Can you capture CO₂ without heating the sorbent?
A: No, not meaningfully. At room temperature, CO₂ remains stably bound to the amine. Pressure swing (reducing pressure) can desorb some CO₂ from certain sorbents, but the process is slow and requires building expensive vacuum systems. Thermal regeneration (heating) is the practical choice because the energy input breaks the chemical bond directly, and waste heat from industrial processes often supplies this energy economically.

Q: Why is captured CO₂ compressed if it's being injected underground?
A: At 800+ meters depth, rock temperature and pressure keep CO₂ as a supercritical fluid, which is denser and easier to inject than a gas. Compression from atmospheric pressure to 150+ bar (1) makes CO₂ small enough to fit in transport pipelines and injection wells, and (2) ensures it remains in a dense phase underground where it stays trapped. Uncompressed gas would expand and escape much more easily.

Q: Could we just pump captured CO₂ into depleted oil wells forever?
A: Depleted oil and gas reservoirs are suitable long-term storage sites because they have proven sealing (they trapped hydrocarbons for millions of years). However, global storage capacity in depleted reservoirs is estimated at 400–500 gigatonnes total — enough for a few centuries at projected removal rates, but not infinite. Deep saline aquifers (water-filled rock layers not used for drinking water) are more abundant, with estimated capacity of 1,000+ gigatonnes globally, but less well-characterized than oil fields.

Q: If DAC costs $200/tonne, why not just fund tree-planting instead?
A: Tree-planting captures CO₂ through photosynthesis (free, solar-powered) and is far cheaper upfront ($5–50/tonne cost, though permanence varies). The trade-off: trees occupy land, require decades to mature, and are vulnerable to fire, disease, and future logging. Geologic storage is permanent (certified for 10,000+ years by IPCC standards) but requires energy and capital. Optimal strategies use both: land restoration and natural carbon sinks where possible, engineered capture where land is scarce or biomass carbon would be temporary.

Q: Will future innovations make DAC as cheap as point-source capture?
A: Unlikely. DAC costs can improve 30–40% through learning curves and modular design, approaching $100–150/tonne by 2040–2050. Point-source capture will improve too, potentially reaching $15–30/tonne. The 100–300x difference in CO₂ concentration is a fundamental physics constraint, not an engineering gap. DAC will always be more expensive per tonne, but may become cost-competitive for specific niches (negative emissions, residual emissions, regions with abundant cheap renewable electricity).

---

Read next

• Climate Change Explained — The Science, Causes, and Consequences
• What Is Green Hydrogen — and Could It Replace Fossil Fuels?
• Grid-Scale Battery Storage: The Missing Piece of the Energy Transition