News

Technologies to remove carbon dioxide from the air: A reality check

MIT study finds that many climate-stabilization plans are based on questionable assumptions about the future cost and deployment of “direct air capture” technology and therefore may not bring about the promised reduction in carbon dioxide in the atmosphere.

Nancy W. Stauffer MITEI

In 2015, 195 nations plus the European Union signed the Paris Agreement and pledged to undertake plans designed to limit the global temperature increase to 1.5°C. Yet in 2023, the world exceeded that target for most if not all of the year—calling into question the long-term feasibility of achieving that target.

To do so, the world must reduce the levels of greenhouse gases in the atmosphere, and strategies for achieving levels that will “stabilize the climate” have been both proposed and adopted. Many of those strategies combine dramatic cuts in carbon dioxide (CO2) emissions with the use of direct air capture (DAC), a technology that removes CO2 from the ambient air. As a reality check, a team of researchers in the MIT Energy Initiative (MITEI) examined those strategies, and what they found was alarming: The strategies rely on overly optimistic—indeed, unrealistic—assumptions about how much CO2 could be removed by DAC. As a result, the strategies won’t perform as predicted. Nevertheless, the MITEI team recommends that work to develop the DAC technology continue so that it’s ready to help with the energy transition—even if it’s not the silver bullet that solves the world’s decarbonization challenge.

DAC: The promise and the reality

Including DAC in plans to stabilize the climate makes sense. Much work is now under way to develop DAC systems, and the technology looks promising. While companies may never run their own DAC systems, they can already buy “carbon credits” based on DAC. Today, a multibillion-dollar market exists on which entities or individuals that face high costs or excessive disruptions to reduce their own carbon emissions can pay others to take emissions-reducing actions on their behalf. Those actions can involve undertaking new renewable energy projects or “carbon-removal” initiatives such as DAC or afforestation/reforestation (planting trees in areas that have never been forested or that were forested in the past). And DAC-based credits are especially appealing for several reasons, explains Howard Herzog, a senior research engineer at MITEI. With DAC, measuring and verifying the amount of carbon removed is straightforward; the removal is immediate, unlike with planting forests, which may take decades to have an impact; and when DAC is coupled with CO2 storage in geologic formations, the CO2 is kept out of the atmosphere essentially permanently—in contrast to, for example, sequestering it in trees, which may one day burn and release the stored CO2.

Will current plans that rely on DAC be effective in stabilizing the climate in the coming years? To find out, Herzog and his colleagues Jennifer Morris and Angelo Gurgel, both MITEI principal research scientists, and Sergey Paltsev, a MITEI senior research scientist—all affiliated with the MIT Center for Sustainability Science and Strategy (CS3)—took a close look at the modeling studies on which those plans are based.

Their investigation identified three unavoidable engineering challenges that together lead to a fourth challenge—high costs for removing a single ton of CO2 from the atmosphere. The details of their findings are reported in a paper published in the journal One Earth on September 20, 2024.

Challenge 1: Scaling up 

When it comes to removing CO2 from the air, nature presents “a major, non-negotiable challenge,” notes the MITEI team: The concentration of CO2 in the air is extremely low—just 420 parts per million, or roughly 0.04%. In contrast, the CO2 concentration in flue gases emitted by power plants and industrial processes ranges from 3% to 20%. Companies now use various carbon capture and sequestration (CCS) technologies to capture CO2 from their flue gases, but capturing CO2 from the air is much more difficult. To explain, the researchers offer the following analogy: “The difference is akin to needing to find 10 red marbles in a jar of 25,000 marbles of which 24,990 are blue [the task representing DAC] versus needing to find about 10 red marbles in a jar of 100 marbles of which 90 are blue [the task for CCS].”

Given that low concentration, removing a single metric ton (tonne) of CO2 from air requires processing about 1.8 million cubic meters of air, which is roughly equivalent to the volume of 720 Olympic-sized swimming pools. And all that air must be moved across a CO2-capturing sorbent—a feat requiring large equipment. For example, one recently proposed design for capturing one million tonnes of CO2 per year would require an “air contactor” equivalent in size to a structure about 3 stories high and 3 miles long.

Recent modeling studies project DAC deployment on the scale of 5 to 40 gigatonnes of CO2 removed per year. (A gigatonne equals one billion metric tonnes.) But in their paper, the researchers conclude that the likelihood of deploying DAC at the gigatonne scale is “highly uncertain.”

Challenge 2: Energy requirement

Given the low concentration of CO2 in the air and the need to move large quantities of air to capture it, it’s no surprise that even the best DAC processes proposed today would consume large amounts of energy—energy that’s generally supplied by a combination of electricity and heat. Including the energy needed to compress the captured CO2 for transportation and storage, most proposed processes require an equivalent of at least 1.2 megawatt-hours of electricity for each tonne of CO2 removed.

The source of that electricity is critical. For example, using coal-based electricity to drive an all-electric DAC process would generate 1.2 tonnes of CO2 for each tonne of CO2 captured. The result would be a net increase in emissions, defeating the whole purpose of the DAC. So clearly the energy requirement must be satisfied using either low-carbon electricity or electricity generated using fossil fuels with CCS. All-electric DAC deployed at large scale—say, 10 gigatonnes of CO2 removed annually—would require 12,000 terawatt-hours of electricity, which is more than 40% of total global electricity generation today.

Electricity consumption is expected to grow due to increasing overall electrification of the world economy, so low-carbon electricity will be in high demand for many competing uses, for example, in power generation, transportation, industry, and building operations. Using clean electricity for DAC instead of for reducing CO2 emissions in other critical areas raises concerns about the best uses of clean electricity.

Many studies assume that a DAC unit could also get energy from “waste heat” generated by some industrial process or facility nearby. In the MITEI researchers’ opinion, “that may be more wishful thinking than reality.” The heat source would need to be within a few miles of the DAC plant for transporting the heat to be economical; given its high capital cost, the DAC plant would need to run nonstop, requiring constant heat delivery; and heat at the temperature required by the DAC plant would have competing uses, for example, for heating buildings. Finally, if DAC is deployed at the gigatonne per year scale, waste heat will likely be able to provide only a small fraction of the needed energy.

Challenge 3: Siting

Some analysts have asserted that, because air is everywhere, DAC units can be located anywhere. But in reality, siting a DAC plant involves many complex issues. As noted above, DAC plants require significant amounts of energy, so having access to enough low-carbon energy is critical. Likewise, having nearby options for storing the removed CO2 is also critical. If storage sites or pipelines to such sites don’t exist, major new infrastructure will need to be built, and building new infrastructure of any kind is expensive and complicated, involving issues related to permitting, environmental justice, and public acceptability—issues that are, in the words of the researchers, “commonly underestimated in the real world and neglected in models.”

Two more siting needs must be considered. First, meteorological conditions must be acceptable. By definition, any DAC unit will be exposed to the elements, and factors like temperature and humidity will affect process performance and process availability. And second, a DAC plant will require some dedicated land—though how much is unclear, as the optimal spacing of units is as yet unresolved. Like wind turbines, DAC units need to be properly spaced to ensure maximum performance such that one unit is not sucking in CO2-depleted air from another unit.

Challenge 4: Cost

Considering the first three challenges, the final challenge is clear: The cost per tonne of CO2 removed is inevitably high. Recent modeling studies assume DAC costs as low as $100 to $200 per ton of CO2 removed. But the researchers found evidence suggesting far higher costs.

To start, they cite typical costs for power plants and industrial sites that now use CCS to remove CO2 from their flue gases. The cost of CCS in such applications is estimated to be in the range of $50 to $150 per ton of CO2 removed. As explained above, the far lower concentration of CO2 in the air will lead to substantially higher costs.

As explained under Challenge 1, the DAC units needed to capture the required amount of air are massive. The capital cost of building them will be high, given labor, materials, permitting costs, and so on. Some estimates in the literature exceed $5,000 per tonne captured per year.

Then there are the ongoing costs of energy. As noted under Challenge 2, removing one tonne of CO2 requires the equivalent of 1.2 megawatt-hours of electricity. If that electricity costs $0.10 per kilowatt-hour, the cost of just the electricity needed to remove one tonne of CO2 is $120. The researchers point out that assuming such a low price is “questionable,” given the expected increase in electricity demand, future competition for clean energy, and higher costs on a system dominated by renewable—but intermittent—energy sources.

Then there’s the cost of storage, which is ignored in many DAC cost estimates.

Clearly, many considerations show that prices of $100 to $200 per tonne is unrealistic, and assuming such low prices will distort assessments of strategies, leading them to underperform going forward.

The bottom line

In their paper, the MITEI team calls DAC a “very seductive concept.” Using DAC to suck CO2 out of the air and generate high-quality carbon-removal credits can offset reduction requirements for industries that have hard-to-abate emissions. By doing so, DAC would minimize disruptions to key parts of the world’s economy, including air travel, certain carbon-intensive industries, and agriculture. However, the world would need to generate billions of tonnes of CO2 credits at an affordable price. That prospect doesn’t look likely. The largest DAC plant in operation today removes just 4,000 tonnes of CO2 per year, and the price to buy the company’s carbon-removal credits on the market today is $1,500 per tonne.

The researchers recognize that there is room for energy efficiency improvements in the future, but DAC units will always be subject to higher work requirements than CCS applied to power plant or industrial flue gases, and there is not a clear pathway to reducing work requirements much below the levels of current DAC technologies.

Nevertheless, the researchers recommend that work to develop DAC continue “because it may be needed for meeting net-zero emissions goals, especially given the current pace of emissions.” But their paper concludes with this warning: “Given the high stakes of climate change, it is foolhardy to rely on DAC to be the hero that comes to our rescue.”


Research Areas

Press inquiries: miteimedia@mit.edu

We're hiring! Learn more and apply