June 15, 2018
Can we really save the climate by turning air into gasoline?
Last week The Atlantic picked up a story about a Canadian company that is reporting significant progress in reducing the costs of direct air capture (DAC) of CO2 from the atmosphere. Under the rather provocative headline “Climate Change Can Be Stopped by Turning Air Into Gasoline,” The Atlantic‘s coverage was rather credulous and missed some key issues.
At a basic level, converting CO2 (a combustion product) into a liquid hydrocarbon (a combustion fuel) is going to require an input of energy. That’s the first law of thermodynamics. Also, any process to separate and purify a mixture of gases (air and CO2) is going to require an additional input of energy . That’s the second law of thermodynamics. The Atlantic pretty much entirely missed these important points:
Finally, the carbon dioxide is combined with hydrogen and converted into liquid fuels, including gasoline, diesel, and jet fuel. This is in some ways the most conventional aspect of the process: Oil companies convert hydrocarbon gases into liquid fuels every day, using a set of chemical reactions called the Fischer-Tropsch process. But it’s key to Carbon Engineering’s business: It means the company can produce carbon-neutral hydrocarbons.
The process as described by Carbon Engineering requires inputs of energy that generally come from fossil sources, meaning the resulting fuels will not be carbon-neutral.
First of all, the DAC process requires energy input in the form of electricity. The authors present configurations in which all electricity is generated on site, or delivered by the grid. In areas with clean grid electricity, the latter would make sense. In areas with a carbon-intensive grid, the electricity is generated on-site from additional natural gas combustion, with the resulting CO2 captured and added to the output CO2 stream.
Second, the DAC process requires energy input in the form of natural gas. Even when clean grid electricity is available, the calciner unit needs natural gas since calcining calcium carbonate requires high temperatures (their design temperature is 900 C. “Calcining” is the reaction of CaCO3 –> CaO + CO2). In their “baseline” configuration their natural gas requirement is 8.81 GJ for every tonne of CO2 captured from the air, which translates into an additional 480 kg (0.48 tonnes) of CO2 from natural gas for every tonne of CO2 captured from the air. Even in their “minimum gas” configurations, the 5.25 GJ/tonne CO2 translates to an additional 290 kg (0.29 tonnes) per tonne captured from the air.
In a carbon capture & sequestration application, you could argue that this is no big deal, since the additional CO2 from gas combustion is in the same stream as the CO2 from calcining reaction, so can be readily compressed for geological sequestration (although having more gas to compress means a larger compression energy requirement, reducing the efficiency of the overall process).
In a synfuels context the use of natural gas would complicate things, since a portion of the carbon in the resulting fuel would have been derived from fossil sources rather than being “recycled” from the air. Using Carbon Engineering’s numbers, this fraction would be at least 22%, under the “minimum gas” configuration and assuming that zero-carbon grid electricity was used. But the carbon intensity of the final synfuel would also depend on the source of the hydrogen. If electrolyzed using zero-carbon electricity, then this would not increase the effective carbon intensity of the product. While the authors note that this is a possibility, the reality is that most hydrogen is actually produced by reforming natural gas. Thus, this represents yet another fossil carbon input into this process.
You might argue that a 78% reduction in carbon intensity is not perfect, but is much better than a 0% reduction. Would we be better off powering our cars with zero-carbon grid electricity or electrolyzed hydrogen? From a purely environmental standpoint, sure. From an economic/environmental standpoint… I would need to run more numbers, and it would depend on the total cost of producing the synfuels. Josiah Neeley over at R Street Institute has already pointed out that the range of costs for Carbon Engineering’s DAC process ($94 – $232 / ton of CO2 removed) are well above both current estimates of the social cost of carbon, and the costs of implementing many mitigation technologies.
To their credit, the CBC (who always love a local-boys-do-good story), got it right, noting
Because the plant currently uses some natural gas, by the time the fuel it produces has been burned it has released a half-tonne of carbon dioxide for every tonne removed from the air. That gives it a carbon footprint 70 per cent lower than a fossil fuel, he said.
So is DAC a waste of time and resources?
There is (or may soon be) a compelling use case for a fuel like this is in aviation, where hydrocarbon fuels offer incomparable technical performance. Why? Liquid hydrocarbons have outstanding energy density (high energy per liter and per kg) which is crucial when you need to carry all your fuel with you on a long journey. Liquid hydrocarbons are also easy to store and distribute, and can fit readily into any shape imaginable (e.g. wing tanks). With rapidly growing demand for air travel and no plausible pathway to electrification in that sector (at least not yet), a synthetic fuel that reduces carbon emissions by even 70% would be a welcome improvement.
Moreover, the theoretical minimum energy for separating CO2 from air is only about 0.5 GJ per tonne of CO2, or less than 10% of the gas requirement for this process. No one seems to have worked out a practical process that comes close to that theoretical minimum, but it doesn’t mean we should stop trying.
We need to be realistic about what technology can deliver and what it can’t. And realistically, while a vast improvement over competing DAC processes, this technology is not ready to deliver “carbon neutral hydrocarbons” or to “stop climate change by turning air into gasoline,” as claimed by The Atlantic.