Direct air capture (DAC) is the use of chemical or physical processes to extract carbon dioxide () directly from the ambient air. If the extracted is then sequestered in safe long-term storage, the overall process is called direct air carbon capture and sequestration (DACCS), achieving carbon dioxide removal. Systems that engage in such a process are referred to as negative emissions technologies (NET).
DAC is in contrast to carbon capture and storage (CCS), which captures from point sources, such as a cement factory or a bioenergy plant. After the capture, DAC generates a concentrated stream of for sequestration or utilization. Carbon dioxide removal is achieved when ambient air makes contact with chemical media, typically an aqueous alkaline solvent or sorbents. These chemical media are subsequently stripped of CO2 through the application of energy (namely heat), resulting in a CO2 stream that can undergo dehydration and compression, while simultaneously regenerating the chemical media for reuse.
DAC has yet to be integrated into emissions trading because, at over US$1000, the cost per ton of carbon dioxide is many times the carbon price. The current high cost of DAC is driven by the scale of deployment and energy factors. In 2025 DAC costs exceeded $1000 per tonne . Some claim that for plant scales of 1 Mtpa (million tonnes per annum) and above, DAC cost could be reduced below $200 per tonne of atmospheric removed, but others dispute this. Future innovations may reduce the energy intensity of this process.
DAC was suggested in 1999 and is still in development. Several commercial plants are planned or in operation in Europe and the US. Large-scale DAC deployment may be accelerated when connected with economical applications or policy incentives.
In contrast to carbon capture and storage (CCS) which captures emissions from a large point source such as a factory, DAC reduces the carbon dioxide concentration in the atmosphere as a whole. Flying across oceans and getting into space are expected to emit a significant amount of greenhouse gases in the mid-21st century. DAC can be used to capture carbon dioxide before or after its use in jet fuel. Although biofuels are an alternative to efuels, as of 2026 the most sustainable balance between them is unclear.
Methods of capture There are the three stages of capture in DAC: the contacting stage, the capture stage, and the separation stage. In the contacting stage, the DAC system transports atmospheric air containing to the equipment using large-scale fans. Subsequently, in the capture stage, rapidly and effectively binds with liquid solvents in chemical reactors or solid sorbents in filters, which must possess binding energies equivalent to that of . Later in the separation stage, external energy sources facilitate the separation of from the solvents or sorbents, yielding pure and regenerated solvents or sorbents. Following the completion of these three stages, the separated pure is either utilized or stored, while the recovered solvents or sorbents are recycled for reuse in the capture process.
Generally, solid sorbents DAC (S-DAC) uses low temperature process DAC, while liquid (amine or metallic hydroxides) sorbents DAC (L-DAC) uses low or high temperature process. S-DAC and L-DAC feature different properties in terms of kinetics and heat transfers. Currently, L-DAC and S-DAC represent two mature technologies for industrial deployment. Additionally, several emerging DAC technologies, including electro-swing adsorption (ESA), moisture-swing adsorption (MSA), and membrane-based DAC (m-DAC), are in different stages of development, testing, or limited practical application.
More recently, Ireland-based company Carbon Collect Limited has developed the MechanicalTree™ which simply stands in the wind to capture . The company claims this 'passive capture' of significantly reduces the energy cost of Direct Air Capture, and that its geometry lends itself to scaling for gigaton capture.
Most commercial techniques use a liquid solvent—usually amine-based or caustic—to absorb from a gas. For example, a common caustic solvent: sodium hydroxide reacts with and precipitates a stable sodium carbonate. This carbonate is heated to produce a highly pure gaseous stream. Sodium hydroxide can be recycled from sodium carbonate in a process of causticizing. Alternatively, the binds to solid sorbent in the process of chemisorption. Through heat and vacuum, the is then desorbed from the solid.
Among the specific chemical processes that are being explored, three stand out: causticization with alkali and alkali-earth hydroxides, carbonation, and organic−inorganic hybrid sorbents consisting of amines supported in porous adsorbents.
Other explored methods The idea of using many small dispersed DAC scrubbers—analogous to live plants—to create environmentally significant reduction in levels, has earned the technology a name of artificial trees in popular media.
Moisture swing sorbent In a cyclical process designed in 2012 by professor Klaus Lackner, the director of the Center for Negative Carbon Emissions (CNCE), dilute can be efficiently separated using an anionic exchange polymer resin called Marathon MSA, which absorbs air when dry, and releases it when exposed to moisture. A large part of the energy for the process is supplied by the latent heat of phase change of water. The technology requires further research to determine its cost-effectiveness.
Metal-organic frameworks
Other substances which can be used are metal–organic frameworks (MOFs).
Metal–organic frameworks are crystalline porous materials composed of metal ion clusters linked by organic molecules, resulting in highly ordered three-dimensional structures with exceptionally large internal surface areas. Compared to conventional solid sorbents such as activated carbons and zeolites, MOFs offer greater tunability: their pore size, geometry, and surface chemistry can be systematically varied by changing the metal centers or organic linkers, enabling materials to be designed with specific CO₂ binding properties. For DAC applications, amine-functionalized MOFs have attracted particular interest, as the amine groups can selectively bind CO₂ through chemisorption even at the extremely low atmospheric concentration of approximately 400 ppm. A key challenge for MOFs in DAC is moisture stability: since atmospheric air always contains water vapor at concentrations far exceeding CO₂, water molecules compete for the same binding sites and can cause structural degradation of the framework over repeated adsorption–desorption cycles. Engineering MOFs that are both selective for CO₂ and resistant to humidity under real-world operating conditions remains an active area of research.
Identifying effective MOF sorbents for DAC is challenging due to the vast number of possible framework combinations. In 2023, researchers from Meta AI's Fundamental AI Research (FAIR) team and Georgia Institute of Technology released the Open Direct Air Capture (OpenDAC) dataset to address this problem computationally. The dataset contains results from approximately 400 million CPU hours of quantum chemistry calculations screening MOF candidates for CO₂ adsorption properties, making it one of the largest open computational datasets in the field.[] Alongside the dataset, machine learning interatomic potential models were released to enable rapid screening of candidate materials at a fraction of the cost of traditional density functional theory simulations, with the goal of accelerating the discovery of viable DAC sorbents.
Membranes Membrane-based separation (m-DAC) employs semi-permeable membranes. This method requires little water and has a smaller footprint. Typically polymeric membranes, either glassy or rubbery, are used for direct air capture. Glassy membranes typically exhibit high selectivity with respect to carbon dioxide; however, they also have low permeabilities. Membrane capture of carbon dioxide is still in development and needs further research before it can be implemented on a larger scale.
Electro-Swing Adsorption Electro-swing adsorption (ESA) has also been proposed.
Rock flour Rock flour, soil ground into nanoparticles by glacier ice, has potential both as a soil conditioner and for carbon capture. Glacier melting deposits one billion tons of rock flour annually, and one ton of Greenlandic rock flour can capture of carbon.
Environmental impact DAC is a carbon negative technology, with its greenhouse gas emissions (GHG) estimated to range from 0.01 t emitted per t captured when renewable electricity is used to 0.65 t emitted per t captured when grid electricity and natural gas (NG) heating are used. The energy source emission factor of DAC is the primary driver of DAC's GHG emissions. The combination of renewable wind and grid electricity would also be carbon negative, providing high carbon removal benefits. The emissions factors for renewable wind and grid electricity are typically less than 0.1 t emitted per t captured if the wind plant supplies at least 50–80% of the plant capacity factor (capacity usage) when grid electricity emission factors do not exceed 0.3077 kg/kWh. Higher grid electricity emission factors could still be used, but this would require their use to be less than 20% to achieve very high carbon removal.
Proponents of DAC argue that it is an essential component of climate change mitigation. Researchers posit that DAC could help contribute to the goals of the Paris Agreement (namely limiting the increase in global average temperature to well below 2 °C above pre-industrial levels). The IEA estimates that capture of at least 85 million tonnes and 980 million tonnes of CO2 annually by 2030 and 2050, respectively, are needed to achieve net zero. However, others claim that relying on this technology is risky and might postpone emission reduction under the notion that it will be possible to fix the problem later, and suggest that reducing emissions may be a better solution. It is important to see DAC as a complementary solution that is necessary in helping to achieve climate targets.
Opponents of DAC argue that the resources required to operate DAC technologies, are an immense burden that may outweigh the goal of the technology itself. A 2020 analysis revealed that DAC 2 technology may be an unsuitable option to capture the projected 30 Gt- per year as it requires an enormous amount of materials (16.3–27.8 Gt of NH3 and 3.3–5.6 Gt of EO) The same study found that DAC 1 technology requires at least 8.4–13.1 TW-yr (46–71% TGES), an estimate that was calculated with the exclusion of the associated energy costs for carbon storage. However, the IEA net zero approaches require CO2 capture from DAC in the magnitude of 0.1 Gt- (980 million tonnes of CO2) annually in 2050, which is significantly lower than 30 Gt- per year that opponents of DAC were assessing.
Energy cost concerns were explored in 2021 and found that in order for DAC technology to maintain a carbon removal of 73-86% per ton of captured, DAC would demand land occupation and renewable energy equivalent to what is needed for a global switch from gasoline to electric vehicles, with approximately five times higher material consumption. However, the material demand of DAC is mostly from common materials, such as steel, concrete and earth minerals (like zeolites and metallic hydroxides). The use of electric vehicle may require substantial accessibility to critical materials, and this limited availability of critical materials may not be able to sustain the demand needed for net zero.