It seemed like game-changing technology: Take carbon dioxide out of the ocean, and turn it into a carbon-neutral fuel that could be used in today's current gas tanks.

When scientists at X—formerly Google X, the "moonshot factory" at Alphabet known for driverless cars and Wi-Fi on balloons—learned about a new process for turning seawater into fuel, they partnered with the researchers behind it to try to make it real. Two years later, despite the fact that the technology worked, they killed the project.

Like some other projects at X, it started when someone happened to read a study on the new technology. Someone else invited the researcher behind it to come to the X lab to give a tech talk. After he went into more detail, they were even more interested.

"We knew if you could make a carbon-neutral fuel at a price point that would allow for commercialization, you would have such potential for impact," says Kathy Cooper, team lead for the project, which was eventually named Foghorn.

Transportation makes up around 14% of global greenhouse gas emissions, and while electric cars are very slowly becoming more common, other transportation sectors—like airplanes or cargo ships—don't yet have a simple way to stop polluting. What is key is that the new fuel could be used in existing vehicles.

The fuel makes use of the rising carbon dioxide levels in the ocean; as CO2 increases in the atmosphere, concentrations of dissolved CO2 also rise in the sea, ending up in a form called bicarbonate that makes the ocean more acidic.

"The process that we're using, in short, essentially shifts the pH of the ocean," says Matt Eisaman, one of a team of PARC scientists who originally developed the technology. By sucking ocean water into a tank and making it more acidic, it's possible to collect CO2 as a gas. Using another process, it's possible to also pull hydrogen from the water. If the CO2 and hydrogen are reacted together, they become a liquid fuel.

After the X team heard Eisaman's initial presentation, they asked him to do some quick calculations to see if making fuel like this might be commercially viable. The analysis had a huge degree of uncertainty, but it seemed like it might actually be something they could sell some day. X formed a partnership with the PARC researchers.

"We decided it was worth digging a little deeper, doing more experiments and prototyping to really understand if we could make it work," says Cooper. "And thus Foghorn was born."

Most ideas at X don't last long; some are killed within hours. To be viable for the innovation lab, projects have to help solve a problem that affects millions or billions of people, the technology has to be "audacious," and there has to be a chance that it can make it to market within roughly 5 to 10 years.

If something could be commercialized sooner, the reasoning is that another company is probably working on it already; if it takes longer, it might not provide a good return, and the technology also might become outdated by the time it's ready.

Often, back-of-the-envelope calculations make it clear that an idea couldn't become financially viable quickly enough. When something passes the earliest tests—like Project Foghorn—it moves on to a next stage, with a "rapid evaluation" team that tries to understand the potential project's biggest risks.

When Foghorn started, the researchers from PARC already had a working proof-of-concept; the biggest uncertainties were around cost. The team got to work building a bigger prototype in the lab, planning to take more accurate measurements that it could plug into an economic model.

It didn't take long before a problem appeared: They realized that if the machine ran for a longer period of time, minerals would build up on the machine's membrane, basically destroying the system. "So we had to invent our way out of that problem, which we did," says Eisaman.

The team also found another important solution. If they partnered with desalination plants, they realized, they could avoid the expense of building pipes in the ocean, helping bring the cost down. It brought the process much closer to their target—$5 for a gallon of seawater fuel, within five years.

Still, it wasn't quite that cheap, and the scientists realized that there was another problem; there just aren't that many desalination plants in the ocean today. Even if they partnered with all of them, it would only be possible to produce a relatively small amount of fuel, only enough to offset about four coal plants' worth of emissions.

Eventually—about two years after the project began—the team pulled the plug. Renewable hydrogen, a key ingredient used with their carbon dioxide to make the fuel, was too expensive to produce; the carbon dioxide itself was a little too expensive to pull out of the seawater.

"In the end, it was just a question of opportunity costs," says Cooper. "For Matt and I, we're always looking for what can X put its resources into in order to make the largest impact. We especially care about the issue of climate change, so we tend to look into that sector, though we've also evaluated others. We just thought the resources that we would put into this project we'd probably have a greater impact if it was put toward something else."

While it was hard to pull away, the culture of the lab supports the idea of failure—and even hands out bonuses when teams agree that it's time to give up on something.

"Because X is premised on the idea of pursuing highly risky projects, there's just an understanding that a lot of them aren't going to work," says Cooper. "So it's not seen as surprising or the fault of anyone if something doesn't work. It's just sort of seen as the nature of the work. And that depersonalizes it in a way that's very helpful."

It's possible that the viability of the technology could change sooner than expected. A price on carbon, an R&D breakthrough, and increasing fossil fuel prices could help. But without those things, it's hit a dead end for now.

The researchers are working on a peer-reviewed paper about the technology—and costs—that they can share with other scientists. There's little data now about the costs of "negative emissions" technologies like this, which suck carbon out of the atmosphere rather than just slowing down its buildup.

At some point, most climate scientists think that we'll need to use those technologies to meet the goals of the Paris climate agreement. It's easier to pull CO2 from the ocean than the air, so there's a good chance that Foghorn, or something like it, may eventually be used.

"In some sense [we'll] kick it back to the research community," says Eisaman. "Then if we wait maybe 5 to 10 years and let that research and development take place, I think eventually something like this will be commercialized. It was just a bit ahead of its time."

Have something to say about this article? You can email us and let us know. If it's interesting and thoughtful, we may publish your response.