Download a PDF of this article.

When Frank Zeman made a device to mop carbon dioxide out of the air of his laboratory at Columbia University in New York, it didn't look like a machine that could save the planet. Black tape held together plastic parts eaten away by lye; baking soda encrusted the outside. If someone walked behind the air intake (which looked like a grey hair dryer), their exhalations would interfere with the results. But the contraption worked.

Such a device, if scaled up and perfected, could be used to dial back Earth's greenhouse thermostat by taking CO2 straight out of the sky. Although Zeman's fully functioning desktop device has not yet made it out of the lab, others have developed parts of bigger and more ambitious devices, some of which are heading for commercialization. All are imperfect, but they all work, and that undeniable fact is turning air capture from a 'what-if' pub discussion into a serious proposal.

"Nobody doubts it's technically feasible," says Zeman, now director of the Center for Metropolitan Sustainability at the New York Institute of Technology.

Climate special

Increasingly it looks like air capture will be needed. Efforts to limit CO2 emissions will need to be strengthened massively if they are to keep concentrations from reaching dangerous levels, so there may be little choice but to remove some of the CO2 already in the air (see page 1091) or cool the planet in other ways (see page 1097). "Without having something that is carbon negative, the possibility of avoiding high levels of CO2 is basically zero," says Peter Eisenberger, former director of the Lamont–Doherty Earth Observatory at Columbia University and co-founder of the air-capture company Global Thermostat.

In a recent analysis, Roger Pielke of the University of Colorado in Boulder put some numbers on the task ahead. Assuming a middle-range scenario projected by the Intergovernmental Panel on Climate Change (IPCC), humanity must somehow prevent itself from emitting (or must soak up) 650 gigatonnes of carbon by 2100 to keep concentrations under 450 parts per million (p.p.m.) at that point1. To put that in perspective, humans added about 9 Gt of carbon to the atmosphere last year.

Economic studies suggest that some reductions could come affordably, or even at a profit, from fairly obvious places. Deeper cuts would require serious money. A report from the international consultancy McKinsey estimates that energy-efficiency measures, conversion to low-carbon energy sources, and forestry and agriculture management could — with serious effort — cut about 10 Gt of carbon emissions annually by 2030, for under US$300 per tonne. But it will be much harder and more expensive to get at any fraction of the remaining 9 Gt of annual emissions expected that year in a business-as-usual scenario2. Pielke is one of many beginning to wonder whether mopping up CO2 with chemicals and machinery — a strategy with an ironically un-green image — might be part of the answer.

Klaus Lackner has imagined huge 'farms' featuring thousands of air-scrubbing devices that could soak up billions of tonnes of carbon from the atmosphere. Credit: STONEHAVEN PRODUCTIONS

It could be an unbeatable idea. Sponging CO2 from the air has a direct, immediate and measurable effect on the source of the problem, avoiding the possible side effects of geoengineering. Air-capture devices can be sited anywhere, although preferably on cheap land with an untapped renewable energy supply and a geological reservoir that could serve as a dump for the captured gas. In principle, there is no limit to how much CO2 you can extract: name an atmospheric concentration you'd like to end up with, and the technology can get you there.

To many in the 1990s, that cost seemed ridiculously high. In engineering circles, the dogma ran that the ease of extracting a gas was proportional to its concentration. At 0.04%, CO2 in the atmosphere seemed exceedingly difficult; the effort and money needed to extract and store CO2 from industrial flue stacks, where it can make up perhaps 10% of all gas, is already high, estimated by the IPCC to cost between $70 and $260 per tonne of carbon (see Nature 442, 620–623; 2006). The assumption was that filtering CO2 out of the atmosphere would be 250 times harder and vastly more expensive.

That assumption turns out to be wrong. The benefit of air capture is that it deals with a nearly infinite and relatively clean source, so there is no need to scrub out polluting gases before beginning and no need to take out every last bit of CO2. Thermodynamically, the task proves to be about twice as hard as flue-stream capture3. Better still, the technology to make such devices is already available.

Although air capture has been ignored by the IPCC and sidelined by scientists, that is changing. Researchers in Canada, the United States and Switzerland have come up with plans, tested prototypes, filed patents and founded companies to pursue the idea.

It is the most expensive climate-mitigation technology. And that's a good thing. Frank Zeman

Which technology will win out is yet to be seen. The Virgin Earth Challenge, launched by airline entrepreneur Richard Branson and former US vice-president Al Gore in February 2007, offers up to $25 million for the first demonstrably viable commercial design to remove significant amounts of greenhouse gases from the atmosphere (the exact criteria are unclear). As yet the prize goes unclaimed.

The bare-bones chemistry of carbon capture is simple. The simplest thing to do is to expose air to a sorbent of lye (NaOH). This reacts with CO2 to create a solution of sodium carbonate. It's so simple that Klaus Lackner, also of Columbia University, once helped his daughter to do it for a school science project. To get the carbon out of solution, a trick can be borrowed from the pulp and paper industry: when slaked lime (Ca(OH)2) is added to the mix, particles of calcium carbonate settle out. Throw this into a kiln and you are rewarded with a pure stream of captured CO2 and quicklime (CaO), from which the sorbent can be renewed.boxed-text

Crude prototypes

This is how Zeman's desktop device worked, and also how David Keith of the University of Calgary in Alberta, Canada, is pursuing the problem. Keith built a large-scale machine a few years ago to see how much CO2 could be sucked up in practice. He calls it the 'Russian tractor' technique — not especially high-tech, but proven to work. A prototype featured on the Discovery Channel in 2008 mopped up a few kilograms of carbon overnight.

Keith didn't build the second half of the scheme — the 900 °C kiln that spits out concentrated CO2 — because that's already a known industrial process. It's also the energy-intensive and costly part. Nevertheless, he is setting up a company called Carbon Engineering, convinced the idea is worth pursuing, and is working to reduce costs.

Keith has chosen the most obvious approach to the problem but admits that others have "much more clever" schemes. That includes a material being developed by Lackner for the company Global Research Technologies, based in Tucson, Arizona, and funded by a $5-million donation from the late billionaire Gary Comer. (Comer, founder of the Lands' End clothing-catalogue company, donated money to fight climate change after he sailed through the Northwest Passage in 2001 without being blocked by ice.) In April 2007, Global Research Technologies had its first demonstration of air capture with a prototype device. It was a success, widely lauded in the press, but it needed further work. For one thing, it just vented the captured carbon out the back. For another, it didn't behave as it was expected to. "When we closed the door on it, something was happening we didn't understand," says Lackner.

David Keith and his carbon-capture machine. Credit: K. BENDIKTSEN, UNIV. CALGARY

The device used a commercially available wet resin to mop up CO2. When its designers analysed the results, however, they realized the material was better than they thought. Not only did it turn CO2 into carbonate, but in a dry environment it would go a step further to bicarbonate. When they exposed the resin to water, the bicarbonate flipped back to carbonate, releasing CO2 and water vapour. They didn't need a kiln — they just needed to expose their loaded resin to water in a relative vacuum, and then pressurize the result to condense the water out. "All you pay for is making the vacuum, pumping and pressurizing," says Lackner.

Others argue that kiln-temperature heat isn't necessarily a problem. In Zurich, Aldo Steinfeld and colleagues at the Swiss Federal Institute of Technology are using the Sun-tracking mirrors used by solar-power plants to heat up their air-capture reactor to 800 °C. They have a fully functioning lab model, and hope to have a larger field prototype within a few years to hand to an industrial partner.

Eisenberger, on the other hand, needs only low temperatures — under 100 °C, achievable using waste heat from power plants or cement factories — to run his system. To test these ideas, Eisenberger founded a company called Global Thermostat in 2006 with Graciela Chichilnisky, an economist and mathematician at Columbia University.

Eisenberger imagines a future in which air-capture devices start to be deployed by 2015; by 2020, half of new power generators are matched with air capture, and by 2040, some 9 Gt of carbon is being pulled from the air per year, to a total of 650 Gt by 2100 — the amount that Pielke also estimated would be needed. (Coincidentally, that total roughly matches the IPCC's estimate of the Earth's geological capacity to act as a garbage dump for buried gas). This whole operation could be accomplished by, say, 35,000 facilities that each took a quarter of a million tonnes of carbon per year out of the air. The combined footprint of this global operation would total less than 300 square kilometres — a fraction of the size of London.

Because Eisenberger assumes the world will also make substantial cuts in emissions over the same period, his air-capture scenario would return atmospheric concentrations to 380 p.p.m. of CO2 by 2100, and they would continue to decline thereafter. The price? About $60 trillion for the air capture, or roughly $660 billion per year. That's on the same scale as the US economic stimulus package against the current recession, but every year for a century.

The price is the hardest thing to estimate, since no one has yet built a full-scale device. When Lackner first put out figures of about $100 per tonne of carbon in 2006, many saw it as massively over-optimistic — some joked that the real price was one mysterious 'Lackner' per tonne, given the apparently magical capacities of his material, the identity of which was kept under wraps for commercial reasons at the time. Today, Eisenberger's estimate is slightly cheaper still.

Cost competitive

At the other end of the scale, Keith has estimated it might cost $500 per tonne of carbon using today's technologies (see 'A way to pay for capturing carbon dioxide'). That would rack up a bill of $325 trillion to soak up 650 Gt of carbon, but Pielke notes that such a price tag would still only be 2.7% of global economic output by 2100. That compares favourably with price estimates of the IPCC (–1 to 5% of global economic output) and economist Nicholas Stern (–2 to 4%) for stabilizing air concentrations at 450 p.p.m. without air capture.

"We should be looking into it, at least," Pielke concludes. To put the cost issue in perspective, he notes, if all the emissions from US cars were sucked up by air capture using today's technology, and the cost tagged onto the price of petrol, motorists in the United States would still have one of the lowest pump prices in the world.

Many air-capture enthusiasts talk about countering something on the scale of global aircraft emissions, projected to reach about 0.25 Gt of carbon per year by 2030. (Technology can reduce carbon emissions from power plants and cars, but it is difficult to reduce such emissions from planes.) This is where Roger Aines of Lawrence Livermore National Laboratory in California sees air capture playing a potential part. He and his colleagues are making an overview assessment of the strategy, and estimate that the quarter-gigatonne target could be met by, say, a thousand 250,000-tonne air-capture facilities requiring a total of 900,000 gigawatt-hours of energy per year. This is slightly more than the total electricity generated by the 104 nuclear power plants in the United States. If wind were to supply the power, the world would need something like 135,000 additional 1.5-megawatt turbines. That would approximately double the current global wind-power capacity.

Such a scenario is within the realm of possibility, but it demands an increase in energy production just at a time when we should be trying to break our energy addiction. For some, that's a critical problem. Every dollar spent on air capture instead of shifting to renewables is "a long-term loss to society", says Mark Jacobson of Stanford University in California. His concern is that researching a 'get out of jail free' card for climate change would provide an excuse to continue unabated emissions.

That worry is voiced by many, but it is also dismissed by many. "For some people there's concern that if there's hope that air capture will work, it reduces the incentive to reduce emissions," says Pielke. "That makes as much sense as saying we shouldn't have open-heart surgery because it stops people from lowering their cholesterol. We need both."

No one argues that air capture is a cure-all. Eisenberger sees it as a necessary bridge to get us more painlessly to our goal of a renewable-energy economy. Despite the 'reasonable' price tag of air capture, it is still cheaper, and more sensible, to capture large-industry pollutants at source and to reduce energy use. "Air capture would be a back-stop technology to fill in the gap between what we can achieve and what our goals are," says Pielke.

"It is the most expensive climate-mitigation technology," agrees Zeman. "And that's a good thing. It has this role as the upper bound on solving the climate problem." No matter what we have to do to get the atmosphere settled, it won't cost more than this.