Ask ten people what carbon capture does and you'll get ten versions of the same vague answer: "it sucks CO₂ out of the air." That's true in the way that "a car moves you around" is true. It skips every part that matters — what actually happens inside the plant, how much energy the process eats, and why, after decades of pilot projects, we still aren't drowning in working capture plants.
I've followed this technology closely for years, argued about it at industry meetups, and spent a genuinely embarrassing amount of time reading engineering reports on amine solvents. What I've learned is that carbon capture is neither the savior the oil majors like to advertise nor the greenwashing scam its loudest critics claim. It's a real industrial process with real physics behind it — and real limits that nobody selling it wants to put on the brochure.
Let me walk you through how it actually works, where the numbers get uncomfortable, and why the honest answer to "can this save us?" is: partly, expensively, and not on its own.
Key Takeaways
- Carbon capture separates CO₂ from other gases in a flue stream, then compresses it for transport and permanent storage underground.
- The three main approaches — post-combustion, pre-combustion, and oxy-fuel — differ mostly in when you pull the carbon out of the process.
- Capture rates of 85–95% are achievable, but the "net" number is always lower because the capture process itself burns energy.
- Storage capacity in deep geological formations is large; the bottleneck is cost, pipeline infrastructure, and injection rate per site.
- Safety risks exist but are manageable and well understood: leaks, induced seismicity, and groundwater contamination are monitored, not mystical.
- Capture is a tool for hard-to-abate sectors, not a blanket license to keep emitting.
Carbon capture technology explained without the marketing gloss
Strip away the press releases and carbon capture is a separation problem. You have a gas stream — exhaust from a cement kiln, a steel furnace, a gas turbine — and roughly 4% to 15% of it is CO₂. Your job is to pull that fraction out, concentrate it to near-pure CO₂, and put it somewhere it won't reach the atmosphere for thousands of years.
The concentration matters enormously. A natural gas power plant's flue gas is dilute, maybe 4% CO₂. A cement kiln runs higher, sometimes 14–20% because the limestone decomposition itself releases CO₂. The more dilute your stream, the more work — and energy — you need to extract the carbon. That single fact explains why capture costs vary by a factor of three or more depending on what you're attaching it to.
How does carbon capture work, chemically?
Most commercial capture uses a chemical reaction, not a filter. The dominant method is amine scrubbing: flue gas is bubbled through a solvent containing amine compounds, and the CO₂ molecules react with the amines to form a weakly bonded chemical compound. The cleaned gas exits the top. Then you heat the solvent to around 120°C, the bond breaks, and you get a nearly pure CO₂ stream while the solvent gets recycled.
That heating step is the catch. Roughly 20–30% of a plant's energy output can go into running the capture and regeneration loop. It's why early capture projects were economically painful, and why newer designs chase lower-energy solvents, solid sorbents, and membranes that separate CO₂ without a thermal regeneration cycle at all.
Other configurations exist too. In pre-combustion capture, you gasify the fuel first, shift it to a hydrogen and CO₂ mixture, and pull the CO₂ out before burning the hydrogen. In oxy-fuel systems, you burn fuel in nearly pure oxygen, producing a flue gas that's mostly water and CO₂ — easy to separate, but you need an air separation unit, which is its own energy hog.
What is carbon capture and storage, and where does the CO₂ go?
Here's the part most explainers rush. You don't capture CO₂ and magically delete it. You have a physical substance that has to be moved and buried.
Once separated, the CO₂ is compressed to a supercritical state — above about 31°C and 74 bar — where it behaves like a dense liquid. In that form it's pumped through pipelines to a storage site. And the storage site isn't a cavern or a tank. It's a porous rock formation deep underground, typically 1 to 3 kilometers below the surface, capped by an impermeable layer of shale or salt that seals it in.
Why deep geological storage actually works
The mechanism is the same one that kept oil and gas trapped for millions of years. Porous sandstone holds the CO₂ in its pore spaces; a dense caprock above stops it from migrating upward. Over time, a fraction of the CO₂ dissolves into the formation's brine, and on longer timescales it can mineralize into solid carbonate rock. That mineralized form is essentially permanent.
Geologists have mapped suitable formations across sedimentary basins worldwide, and the theoretical capacity dwarfs what we'd need for centuries of industrial emissions. Capacity isn't the real constraint. Cost, injection rate per well, and — critically — having a pipeline network connecting capture sites to storage sites are.
- Saline aquifers: the largest potential, but you're injecting into rock nobody's ever drilled.
- Depleted oil and gas fields: well-characterized geology, existing wells, but limited volume.
- Basalt formations: rapid mineralization, promising for permanent trapping.
One honest caveat: I've seen storage capacity estimates quoted with wild confidence, and the truth is that only a small fraction of mapped capacity has been proven injectable at real-world flow rates. Theoretical capacity is not the same as accessible capacity. Anyone who tells you the number casually is skipping that distinction.
Carbon capture and storage examples that actually run
Enough theory. What exists in the field?
The longest-running industrial example is a gas processing project in Norway, capturing CO₂ from natural gas since the 1990s and injecting it into a saline formation beneath the North Sea. It's been operating for over two decades and is one of the few projects where the storage has been independently monitored for that long.
In Canada, a coal power unit was retrofitted with post-combustion capture and has captured millions of tonnes, though not without well-documented operational hiccups and cost overruns in its early years. In the US, a large-scale project on a coal gasification plant sells captured CO₂ for enhanced oil recovery — injecting it into oil fields to push out more crude, which is profitable but leaves the climate arithmetic muddier.
And that's the pattern that irritates me. Most deployed carbon capture capacity is attached to oil and gas, and much of the captured CO₂ gets used to extract more fossil fuel. Whether that counts as climate progress depends entirely on how you account for the barrel of oil it helped produce.
The limits of carbon capture nobody puts on the slide
Now the uncomfortable part. You asked how it works and what its limits are, so here they are, in order of how much they actually matter.
Is carbon capture safe?
Broadly yes, at the operational level, with caveats that deserve honesty. The CO₂ itself isn't toxic in the way people imagine, but it's an asphyxiant: a large fast release in an enclosed space is dangerous because it displaces oxygen. Pipeline transport runs at high pressure, so leak detection matters. Underground, the risks are induced seismicity from injection pressure changes, and the possibility of CO₂ migrating into drinking water aquifers if a caprock fails.
Here's the thing, though: those risks are monitored continuously. Injection wells have pressure sensors, seismic arrays, and groundwater sampling programs. The technology for monitoring is mature. The failure mode that worries me isn't a sudden catastrophic leak — it's slow, poorly funded long-term monitoring after a project's operator moves on. Storage needs to be permanent; corporate attention spans are not.
The energy penalty, and the net capture problem
This is the limit people underestimate most. If a capture system removes 90% of the CO₂ from a plant's flue gas but consumes 25% of the plant's output to do it, the effective emissions reduction is lower than the headline number. You've also got emissions embedded in the solvent manufacturing, the pipeline construction, and the energy to compress and inject.
Net capture rates in real deployments land meaningfully below the gross figures. Anyone quoting "90% capture" as though it's the whole story is leaving out the energy overhead and the upstream footprint. A fair number is usually somewhere in the 60–80% range of true net reduction for a typical retrofit, depending on the energy source powering the capture loop.
Cost and infrastructure: the wall you hit
Capture costs today run from roughly $50 to over $120 per tonne of CO₂, depending on the concentration of the stream and the technology used. Dilute streams cost more. That's before transport and storage, which can add another $10 to $30 per tonne if you need a new pipeline.
| Capture approach | Typical strength | Main drawback |
|---|---|---|
| Post-combustion amine | Retrofits onto existing plants | High energy penalty for solvent regeneration |
| Pre-combustion | Higher CO₂ concentration, cheaper separation | Requires gasification, big capital cost |
| Oxy-fuel | Nearly pure CO₂ stream, simple separation | Needs an air separation unit — huge energy draw |
| Direct air capture | Works anywhere, no point source needed | Extremely dilute input, very high cost per tonne |
Notice direct air capture at the bottom. It's the darling of the offset market, and it's also the most expensive route by far, because air is only about 0.04% CO₂. Pulling carbon from the atmosphere is thousands of times harder than pulling it from a smokestack. I find the enthusiasm for it baffling when we still aren't capturing most of the easy point-source emissions.
Why carbon capture is important — and where it stops being the answer
Cement. Steel. Chemicals. These three sectors together account for a huge share of global industrial emissions, and — this is the key point — their CO₂ isn't mostly coming from burning fuel. It's coming from the chemistry of the process itself. You cannot electrify your way out of limestone decomposing into lime and CO₂. There's no fuel to switch.
That's where capture earns its place. For industries with process emissions, it's one of very few tools that exist. For a gas power plant, it's competing with wind, solar, and batteries, which is a much harder fight.
The mistake is treating it as universal. My position on this is firm: capture should be reserved for emissions that can't be avoided any other way, and any policy that lets a company keep a coal plant running indefinitely on the promise of capture is a policy that's been captured itself. The technology is real. The exemption from actually cutting emissions is the problem.
Does carbon capture remove carbon permanently?
For mineralized CO₂ in basalt formations, yes, effectively permanently. For CO₂ stored as supercritical fluid in saline aquifers, it's trapped by the caprock and gradually dissolves or mineralizes over centuries to millennia. "Permanent" in this context means geologically permanent, not forever in the literal sense — but far longer than any timescale relevant to climate policy.
Can carbon capture be scaled to make a real dent in global emissions?
Technically the geology and engineering allow for far more than current deployment. Practically, scaling requires trillions in capital, a continent-scale pipeline network, and sustained policy support over decades. Based on how slowly infrastructure of that kind gets built, I'd be skeptical of any timeline promising gigatonne-scale capture this decade.
The honest verdict on carbon capture
Carbon capture is a genuine engineering achievement with a narrow but critical role. It works. It separates CO₂, compresses it, and buries it in rock that has held fluids for geological time. The physics is sound, the monitoring technology is mature, and for cement and steel it may be the only lever we have.
What it is not is a substitute for cutting emissions at the source. Every tonne you don't produce is cheaper, cleaner, and less risky than a tonne you produce and then try to catch. The energy penalty means capture always runs at a deficit relative to not emitting in the first place. And the economics only pencil out where a policy or a buyer forces them to.
So the next time someone tells you carbon capture will save the climate, ask them which sector. If they say cement, they might be right. If they say "all of them," they're selling you something.
