For more than a century, humanity has been adding carbon dioxide to the atmosphere at an extraordinary rate.
Burning coal, oil and natural gas releases carbon that had been locked underground for millions of years. Deforestation and changes in land use add further emissions.
The result is an atmosphere containing significantly more carbon dioxide than it did before the industrial era.
Reducing future emissions is essential.
But scientists increasingly face another question:
What can we do about the carbon dioxide that is already in the atmosphere?
That question has created a rapidly developing field known as carbon dioxide removal, or CDR.
Researchers are exploring ways to capture carbon from the atmosphere and store it for long periods—or accelerate natural processes that move carbon from the atmosphere into ecosystems, oceans and geological formations.
The technologies range from enormous machines that filter carbon dioxide directly from the air to approaches involving forests, minerals, soils and the ocean.
None is a magic solution.
Removing carbon can be expensive, energy-intensive and technically difficult.
Yet scientists increasingly view carbon removal as a potential complement to rapid emissions reductions, particularly for addressing residual emissions from sectors that are difficult to fully decarbonize.
The race is now underway to determine which methods can actually work at meaningful scale.
Imagine filling a bathtub.
Turning off the tap slows the flow of water.
But if the bathtub is already overflowing, slowing the tap isn't enough.
You also need to remove water.
Climate change isn't exactly like a bathtub, but the analogy illustrates the basic difference between reducing emissions and removing carbon.
Solar power, electric vehicles and energy efficiency can reduce the amount of new carbon dioxide entering the atmosphere.
Carbon removal attempts to take some carbon dioxide back out.
That distinction matters because some emissions may be extremely difficult to eliminate completely.
Industries such as aviation, shipping, cement and certain manufacturing processes may continue producing residual emissions even as cleaner technologies expand.
Carbon removal could potentially help address part of that remaining carbon burden.
But scientists emphasize an important principle:
Carbon removal is not a substitute for cutting emissions.
Removing carbon is generally more difficult than preventing an equivalent amount from being emitted in the first place.
One of the most futuristic approaches is direct air capture.
The basic concept sounds simple.
Build machines that pull ordinary air through specialized materials capable of capturing carbon dioxide.
The air contains only a relatively small concentration of CO₂ compared with nitrogen and oxygen.
That makes the task difficult.
A direct-air-capture system therefore needs to process enormous quantities of air.
Once the carbon dioxide has been captured, the system must separate it from the material and produce a concentrated stream that can be transported and stored.
The captured carbon could potentially be injected into suitable geological formations for long-term storage.
Some systems use solid materials.
Others use liquid chemical solutions.
Researchers are working to reduce energy consumption, improve capture materials and lower costs.
The technology is still at an early stage compared with conventional energy infrastructure, but pilot and commercial-scale projects are being developed.
One of the least visible parts of the carbon-removal race is materials science.
Scientists are searching for materials that can selectively capture carbon dioxide while using as little energy as possible.
An ideal material would have several characteristics.
It would capture CO₂ efficiently.
It would work at relatively low concentrations.
It would release the captured carbon without requiring enormous amounts of heat or electricity.
It would remain stable through thousands of cycles.
And ideally, it would be inexpensive and environmentally responsible to manufacture.
This is a difficult combination.
A material that captures carbon extremely strongly may require too much energy to release it.
A material that releases carbon easily may not capture enough.
Researchers are therefore investigating new porous materials, membranes, sorbents and chemical systems.
Artificial intelligence could also accelerate this search by helping scientists identify promising molecular structures before testing them experimentally.
Another approach doesn't rely on giant machines.
It relies on rocks.
Certain minerals naturally react with carbon dioxide.
Over geological timescales, these reactions help move carbon from the atmosphere into stable forms.
Scientists are investigating whether this natural process can be accelerated.
The idea is known as enhanced weathering.
Rock materials can be crushed into smaller particles, increasing their surface area and potentially speeding up chemical reactions.
These materials could then be applied to suitable environments, such as agricultural land, where weathering processes may help remove atmospheric CO₂.
The challenge is scale.
Mining, crushing and transporting enormous quantities of rock requires energy and infrastructure.
Researchers therefore need to determine whether the carbon removed can significantly exceed the emissions associated with deploying the technology.
If the balance is favorable, enhanced weathering could become one part of a larger carbon-removal portfolio.
Plants naturally pull carbon dioxide from the atmosphere through photosynthesis.
But when plants die and decompose, much of that carbon eventually returns to the atmosphere.
What if some plant material could be transformed into a more stable form?
That is the idea behind biochar.
Biomass can be heated under controlled conditions with limited oxygen, producing a carbon-rich material.
When appropriately produced and applied to soils, some of that carbon can remain stored for extended periods.
Biochar may also influence soil properties, although its effects depend heavily on the material and local conditions.
The concept is attractive because it connects biological carbon capture with waste management and agriculture.
But again, the details matter.
Where does the biomass come from?
How is it processed?
How long does the carbon remain stored?
What happens to soil and ecosystems?
A carbon-removal method isn't automatically beneficial simply because it starts with plants.
Its entire lifecycle has to be considered.
Perhaps the oldest carbon-removal technology is also one of the simplest.
Trees.
Forests absorb carbon dioxide as they grow and store carbon in wood, roots and soils.
Protecting existing forests is therefore important for climate mitigation.
Restoring degraded forests can also increase carbon storage under appropriate ecological conditions.
But biological carbon storage has limitations.
Forests can burn.
They can be damaged by drought, insects and disease.
Carbon stored in vegetation can eventually return to the atmosphere.
This makes biological storage fundamentally different from storing carbon deep underground in suitable geological formations.
Nature-based solutions can play an important role, but researchers must account for permanence and potential reversals.
A forest isn't a permanent vault.
It is a living ecosystem.
The world's oceans already absorb substantial amounts of carbon dioxide from the atmosphere.
Scientists are investigating whether ocean-based approaches could increase that carbon uptake.
One area of research involves ocean alkalinity enhancement.
Adding certain alkaline materials to seawater can alter ocean chemistry and potentially increase its capacity to absorb atmospheric CO₂.
Other proposals involve stimulating biological processes in the ocean.
But marine carbon removal is particularly complicated.
The ocean is a massive interconnected ecosystem.
Changing its chemistry could have consequences beyond carbon storage.
Researchers therefore need to understand not only whether a method removes carbon, but also what it does to marine ecosystems, nutrient cycles and ocean chemistry.
This is an area where scientific caution is especially important.
There is another challenge that sounds technical but could determine whether the entire industry succeeds:
How do you prove how much carbon was actually removed?
Capturing carbon is not enough.
Scientists need to measure how much was removed, where it went and how long it will remain stored.
This is sometimes described through measurement, reporting and verification, or MRV.
For a tree plantation, researchers may need to estimate carbon stored in biomass and soil.
For direct air capture, they need to measure captured CO₂ and confirm where it is ultimately stored.
For mineralization, they need to verify that carbon has become part of stable mineral compounds.
Without reliable measurement, carbon credits and climate claims can become difficult to trust.
As carbon removal grows, accurate monitoring could become as important as the removal technology itself.
Cost remains one of the biggest barriers.
Removing a ton of carbon dioxide from the atmosphere can be significantly more expensive than avoiding a ton of emissions in many situations.
That means researchers need to reduce costs dramatically if carbon removal is ever going to operate at the scale required for meaningful climate impact.
Several factors influence the price:
Economies of scale could eventually lower costs.
Technology improvements could also make systems more efficient.
But the industry has another problem.
The carbon dioxide being removed has little direct economic value.
A company may spend money to capture it and then permanently store it.
The financial incentive therefore often depends on carbon markets, government policy or other mechanisms.
This makes carbon removal partly a technology challenge and partly an economic and political one.
Artificial intelligence could influence almost every stage of carbon removal.
Scientists can use AI to search for new capture materials.
Engineers can optimize equipment.
Researchers can model geological storage sites.
Satellite systems can monitor forests and land-use changes.
Machine-learning systems can analyze enormous environmental datasets.
AI could even help determine where a carbon-removal facility should be built based on energy availability, geological conditions, transportation infrastructure and environmental constraints.
But AI cannot eliminate the physical challenges.
A computer can identify a promising material.
Scientists still have to manufacture it.
An algorithm can optimize a carbon-capture process.
Engineers still have to build the machine.
The biggest breakthroughs will likely come from combining computational discovery with real-world experimentation.
Perhaps the hardest part of carbon removal is understanding the sheer scale involved.
Global fossil-fuel emissions are measured in tens of billions of tonnes of CO₂ each year.
Removing a meaningful fraction would require an enormous industrial system.
That could mean millions of machines.
Huge amounts of renewable or low-carbon energy.
Large quantities of materials.
Pipelines and storage infrastructure.
Monitoring networks.
And potentially massive amounts of land or geological storage capacity, depending on the technologies used.
This is why scientists generally emphasize a portfolio approach.
There probably won't be one technology that solves carbon removal.
Instead, multiple methods may contribute different amounts in different locations.
The most interesting thing about carbon removal is how many scientific disciplines it brings together.
Climate science determines how much removal might be needed.
Chemistry develops capture materials.
Geology studies permanent underground storage.
Biology explores forests, soils and microorganisms.
Oceanography examines marine approaches.
Engineering builds the equipment.
Computer science helps optimize systems.
Economics determines whether they can scale.
It is therefore not simply an environmental technology.
It is becoming a multidisciplinary scientific field.
And the stakes are unusually high.
The climate challenge ultimately isn't a choice between reducing emissions and removing carbon.
Humanity likely needs both.
Every tonne of emissions avoided today reduces the amount of carbon that needs to be dealt with later.
At the same time, some carbon removal may be necessary to address residual emissions and potentially reduce atmospheric CO₂ over longer periods.
The important question is whether researchers can develop methods that are effective, affordable, measurable and genuinely durable.
That answer is still being worked out.
Direct-air-capture machines may become cheaper.
New minerals may accelerate natural carbon storage.
Biochar could expand in appropriate agricultural systems.
Restored ecosystems could store more carbon.
Ocean-based approaches may or may not prove viable at scale.
Some technologies will succeed.
Others will probably fail.
That is normal for scientific innovation.
The carbon-removal race is still young.
But it represents a remarkable shift in humanity's relationship with the atmosphere.
For most of industrial history, we treated the atmosphere as a place where emissions could simply disappear.
Now scientists are trying to reverse part of that process.
The ultimate goal isn't to give society permission to keep emitting.
It is to develop another tool for dealing with a problem that has already accumulated.
The atmosphere is a vast carbon reservoir—but scientists are beginning to ask whether humanity can build a system capable of taking some of that carbon back.
The answer could become one of the defining scientific and engineering challenges of the century.