The case for direct air capture

Direct air capture (DAC) sounds too good to be true. Humanity is threatened by an excess of carbon dioxide in the atmosphere, so build a giant DAC machine that can remove it.

And it can be done. Right now in Iceland, a plant built by the DAC start-up Climeworks is pulling CO₂ out of the air at a rate of about 4,000 metric tons (t) per year and pumping it deep underground. The CO₂ reacts with basalt rock there to form stable carbonate minerals. The plant, named Orca, runs on low-carbon geothermal electricity and waste heat, so it is by all accounts a carbon-negative process.

But DAC is expensive and energy intensive. Critics of DAC say that other forms of carbon removal are cheaper and that reducing emissions in the first place will always be a better investment in the fight against climate change. So the question isn’t if DAC works; it’s if DAC makes economic sense. There’s a lot riding on the answer.

In an influential report in 2018, the United Nations’ Intergovernmental Panel on Climate Change (IPCC) concluded that humanity can keep global warming below 1.5 °C and dodge the worst consequences of climate change if we can reach net-zero greenhouse gas emissions by 2050.

The strategies for getting there fall into two buckets: emission reduction and carbon removal. Emission reduction is anything that results in less CO₂ being released, including point-source carbon capture, improved energy efficiency, electrification, renewable energy, regenerative agriculture and forest management, and reengineering of industrial processes. Carbon removal is anything that pulls CO₂ out of the air, including nature-based methods such as planting trees and technologies such as DAC.

Though the two approaches sometimes compete for investment and for buyers of the emissions-offset credits they create, many experts say both will be needed. Marcius Extavour, vice president for energy and climate at the XPrize Foundation, says the IPCC report and its follow-ups lay out the need clearly. “If we want to avoid warming beyond 1.5 °C, every scenario is going to require some kind of carbon removal, in addition to reducing emissions immediately and as fast as possible.”

Even within the area of carbon removal, DAC has competition from other methods, most of which are lower cost. “In the nature-based solutions community, they say, ‘It’s a no-brainer. We’re the cheapest or the most abundant,’ ” Extavour says. At the same time, he adds, many people in DAC “think direct air capture is the only thing that’s really scalable.”

Extavour is leading the Carbon Removal XPrize, a competition that will award $100 million to groups developing DAC and other carbon removal methods. Eighty million of that will be split between up to four teams that demonstrate systems that remove 1,000 t of CO₂ per year and present a clear techno-economic path to expanding to a megaton scale. On April 22, Earth Day, XPrize will announce 15 contestants that will receive $1 million each in milestone funding to develop their concepts.

DAC has a lot in common with point-source carbon capture, in which CO₂ is scrubbed from the flue gas of power plants or other industrial facilities. But the two address very different markets and needs, according to Thomas McDonald, the former CEO of Mosaic Materials. Mosaic has developed CO₂ sorbent materials based on metal-organic frameworks for both DAC and point-source carbon capture.

Applied to combustion power plants, McDonald says, point-source carbon capture can result in clean electricity that competes with renewable and nuclear power. It can also handle process emissions such as the CO₂ produced by yeast during ethanol fermentation or by refineries and chemical plants.

DAC comes into play to deal with CO₂ emissions that are unavoidable and can’t be captured with point-source methods, such as those from aviation. It’s also necessary to remediate historical greenhouse gas emissions—the CO₂ the world has been spewing for the past 100-plus years.

Exactly what emissions can and can’t be avoided is the subject of some debate. “There’s going to be a fight across the economy for every sector to be considered hard to abate so they don’t have to change,” says John Noël, a senior climate campaigner at the environmental nonprofit Greenpeace.

The CDR Primer, a digital textbook about carbon dioxide removal published in 2021 by a team from the University of Pennsylvania, estimates truly hard-to-abate emissions at 1.5–3.1 gigatons (Gt) of CO₂ per year. That number assumes that the electricity and industrial sectors can be completely decarbonized. It thus includes only some shipping, aviation, and heating CO₂, along with emissions of nitrous oxide—a greenhouse gas more potent than CO₂—from agriculture, wastewater, and landfills.

Other estimates classify some emissions from power and manufacturing as hard to abate, including those from the chemical industry. They range from 5 to 29 Gt per year.

According to the International Energy Agency, an intergovernmental organization, 19 DAC plants, including Orca, currently operate, for a total capacity of roughly 10,000 t per year. Most are pilot-scale facilities built to develop the technology rather than to capture a meaningful amount of CO₂; several are not even net negative in carbon.

DAC companies are developing two basic business models: sequester the CO₂ underground and sell credits to companies looking to offset their emissions, or sell the CO₂ to companies that can use it.

Some industries, such as beverage production, already buy CO₂—about 180 million t per year worldwide. And CO₂-to-chemical projects, such as Covestro’s program to make polyurethanes from CO2, are underway. Aether, a start-up that makes lab-grown diamonds from CO₂ supplied by Climeworks, raised $18 million in series A investment earlier this month.

The aviation industry is one hoped-for customer for CO₂ from DAC. Carbon dioxide can be combined with hydrogen to produce hydrocarbons such as kerosene-type jet fuel, though that technology is also not yet at commercial scale.

“Direct air capture is an essential step to decarbonization in the sense that it should be, in the longer term, the only way to produce synthetic kerosene,” says Matteo Mirolo, sustainable aviation policy officer at the nonprofit Transport and Environment. Flying all European Union–originating flights on such “e-kerosene” would consume 365 million t of captured CO2 per year, Mirolo says.

Climeworks is bullish on fuels derived from captured CO₂. It’s part of a consortium called Norsk e-Fuel that recently announced plans for a plant in Norway that it expects to produce 25 million L of jet fuel per year from DAC CO₂ and renewable electricity.

But critics say better options will be available to decarbonize jet fuel too. Biomass-derived fuels can in principle be carbon negative if plants use waste biomass and are fitted with carbon capture. The biofuel company Gevo, for example, has aviation fuel partnerships with Archer Daniels Midland, Chevron, and Delta Air Lines. And Strategic Biofuels is working with regulators on an injection site in Louisiana for a jet fuel project consuming forestry waste.

Other DAC firms are racing to join Climeworks in having steel in the ground. Carbon Engineering is working with Oxy Low Carbon Ventures, a subsidiary of Occidental Petroleum, on megaton-scale DAC plants in Norway and Texas. Carbon Engineering’s system uses a set of calcium-chemistry “loops” to isolate the CO₂ it catches using a potassium hydroxide solution.

The firm’s Norwegian plant will inject its CO₂ underneath the ocean floor. Oxy will use the Texas plant’s CO₂ to push more oil and gas out of depleted wells, a practice called enhanced oil recovery (EOR). Global Thermostat, with backing from ExxonMobil and the US Department of Energy, is working on both EOR and ways to use captured CO₂ as a feedstock for fuels and plastics.

Greenpeace’s Noël calls using captured CO₂ to extract more oil a “joke” and warns that partnering with the oil and gas industry is a dangerous path. “It is a Trojan horse for these companies to maintain their access and influence over climate policy and design and implementation going forward at the state, federal, and global level,” he says.

At the same time, EOR operators are willing to pay some of the highest prices for CO₂, and DAC needs customers in order to grow.

Former Carbon Engineering CEO Steve Oldham, who left the firm in January, says EOR from DAC can be carbon neutral. But the firm is also looking at e-fuels and pure sequestration. “DAC is fundamentally a service,” Oldham says. “You and I pay someone to take garbage away from our house. DAC is a model that makes sense for hard-to-decarbonize sectors.”

The market for removing carbon the way we remove the trash is gaining steam, McDonald says. “There’s an incredible amount of demand that’s coming to purchase these, and there’s actually not enough supply of high-quality, additional, verifiable credits.”

Despite that need, DAC has struggled to catch on because of high prices and limited capacity.

Thermo Fisher Scientific’s experience is instructive. The company launched a line of carbon-neutral mass spectrometers in 2021, using carbon removal credits to offset the emissions created during the instruments’ life cycle. The firm’s inorganic mass spectrometry group calculated that 15 kg of CO₂ is emitted for every 1 kg of spectrometer delivered.

“Emission reduction is obviously the most important long-term goal for decarbonizing the economy,” Mario Tuthorn, marketing manager for the group, said in a presentation on the program. But that’s hard to do in precision manufacturing, where performance of the final device is very sensitive to material and design choices, he added. “If we want to make a change today, the most practical approach is to offset our emissions.”

Avoided emissions such as those gained from planting or protecting trees were not good enough in the team’s analysis because they are hard to calculate and track. “Will the forest you paid to protect still be protected in 50 to 100 years’ time?” Tuthorn asked. “A much more rigorous offsetting approach is carbon dioxide removal.”

In a white paper, the group detailed how it shopped for carbon removal credits. It emphasized permanent removal of CO₂ that would not have been removed other ways. Priority also went to technologies that have the potential to scale but need investment now to drive down costs.

Though Thermo Fisher sounds like the exact customer DAC firms are courting, DAC won only 3%—which went to Carbon Engineering—of the program’s 2021 carbon removal spending. Instead, 78% went to construction with sustainable timber and insulation with recycled paper. Biochar and bio-oil projects, which heat plant matter to convert it to charcoal or pyrolysis oil that is then buried or injected into old wells, shared 19%.

Though Tuthorn described DAC as the gold standard for carbon removal alongside biomass-based energy plants that sequester their CO₂ emissions, he said it wasn’t ready and affordable yet. “We want to remove our emissions today,” he said, “and this means we must limit our investment in direct air capture, where plants are still under construction.”

Price was also a concern. Thermo Fisher was willing to spend an average of $80 per metric ton of carbon removed, more than most other initiatives want to spend. Market watchers estimate the current price of DAC offsets at about $400 per metric ton. At that price, DAC could be only a small part of the pie. Over time, Tuthorn said, DAC will gain a bigger share of the credits the program buys.

DAC similarly won only a small chunk of Microsoft’s fiscal 2021 carbon removal credit purchases. The software giant bought 1,400 t of removal credits from Climeworks, 4,000 t from biochar and bio-oil projects, and 193,000 t from soil carbon management. Most of its purchases—1.1 million t—were from forestry projects.

In 2022, Microsoft replaced DAC in its offset portfolio with a project to mineralize CO₂ in old concrete. The company paid an average of $19.40 per metric ton for its 2022 offsets.

Microsoft detailed the problem in its 2022 carbon offset white paper. “High-durability solutions are critical, but supply is limited and expensive, and many companies cannot yet afford them at scale,” it says, referring to biochar, bio-oil, mineralization, and DAC. “We need expensive, higher durability solutions to become more affordable.” Microsoft says it is both an offset credit customer of Climeworks and an investor in the firm through its Climate Innovation Fund.

“It’s great to have these customers,” says Christoph Beuttler, head of climate policy at Climeworks. But it’s a chicken-and-egg moment: for prices to come down, DAC needs to scale up, and to do that, DAC needs customers buying carbon offsets. “We are a technology, right? So we’re expensive now; we get cheap as—if—we scale up. We need the money now,” Beuttler says.

He estimates that DAC needs $100 billion in R&D investment to reduce the price to a point where the technology can really compete.

In a 2021 paper in Industrial and Engineering Chemistry Research, Klaus Lackner and Habib Azarabadi of Arizona State University were more optimistic. They argue that DAC costs could fall with increased use, as solar photovoltaic costs have; it might take only $200 million of carbon credit purchases to bring the cost of DAC down to $100 per metric ton (DOI: 10.1021/acs.iecr.0c04839).

“In just the chemistry, there’s still a lot of potential that we can tap into,” Beuttler says. “But obviously, there’s also a lot to gain from economies of scale. The ones behind me are handmade in Switzerland,” he says, pointing to a photo of a Climeworks DAC module. That is “probably not the most cost-efficient way to produce them,” he adds. “If we continue making them by hand, it will remain a luxury product.”

Climeworks is planning to build a plant in 2024 that will be 10 times as large as Orca, moving the firm in the right direction on both chemistry advances and economies of scale, Beuttler says.

Nearly everyone working in DAC agrees that emission reduction is the low-hanging fruit in fighting climate change and should be harvested as quickly as possible. The importance of carbon removal will grow over time, though, and DAC may be the most scalable, trackable, and tradable way to cover the “last mile” of hard-to-abate and historical emissions.

“Big-picture, long-term thinking is what we need,” Beuttler says. “And we mustn’t think about the current status but about the net-zero world or net-negative world that we need to design.”