Here’s the crazy thing about the plan to build an almost mile long, 90-foot deep, subterranean ice wall around the Fukushima nuclear plant: It’s not really very crazy at all. Building cryogenic barriers sounds like the specialty of an obscure supervillain, but it’s a well-established technique in civil engineering, used regularly for tunnel boring and mining. Ground freezing was even tested as a way of containing radioactive waste in the 1990s at Oak Ridge National Laboratory and performed admirably.
Joe Sopko, the civil engineering firm Moretrench’s director of ground freezing, has spoken with several consultants about the details of the project, and he’s convinced it’s certainly possible. “This is not a complicated freeze job. It really isn’t,” he told me. “However, the installation, because of the radiation, is.” Ed Yarmak of Arctic Foundations, which installed the system at Oak Ridge, agreed. “It’s a large system, but I don’t think it’s out there, where people can’t do it and can’t do it efficiently.” Here’s the problem this technology could solve. The Fukushima nuclear plant, which was devastated by an earthquake and tsunami in March of 2011, is located on a slope. This fact of topology means that groundwater running down from the Abukuma plateau to the east pass right into the site.
Japanese officials estimate that 400 tons of water reach the plant every day and mix with water used to cool the reactors. That’s roughly 96,000 gallons of radioactive water. About 300 tons (or 72,000 gallons) of contaminated waterflow out to sea daily, according to Japan’s National Resources and Energy Agency.
TEPCO, the Japanese utility, has been trying to deal with these problems, but the volume of water poses a formidable logistical challenge, and some of it continues to leak out to the ocean.
So, what to do? Engineers have been pumping and treating the water, but the scale has made that too difficult and prevented other cleanup work from happening. They’ve contemplated diverting groundwater approaching the site into the sea or building clay walls, too. But it appears that officials have settled on the ice-wall containment strategy, as suggested by Japanese contractor Kajima.
In news stories, this operation was presented as difficult on account of the scale. “The technology has been used before in the construction of tunnels, but never on the massive scale that the Fukushima plant would require,” CNN wrote. Even Cabinet Secretary Yoshihide Suga told reporters, “There is no precedent in the world to create a water-shielding wall with frozen soil on such a large scale.”
But the more I dug into ground freezing, the more I realized it was one of those corners of engineering that’s been quietly helping the world’s infrastructure get built for decades. There are journal articles about it and books, too. There’s J.S. Harris’ Ground Freezing in Practice and the definitive textbook, Frozen Ground Engineering by Orlando Andersland and Branko Ladanyi. There have been hundreds of ground-freezing projects, an evaluation by the Department of Energy, and dozens of international conferences.
Here’s how it works. Freeze pipes, made from normal steel, are sunk into the ground at regular intervals. The spacing is normally about one meter. Then, some type of coolant is fed into the pipes. Sopko uses a brine—salty liquid which can be cooled far below the freezing point of fresh water without turning into a solid. On the surface, a big refrigerator chills the liquid, which is pumped into the pipes. The liquid extracts heat from the ground, and returns to the chiller, where it is recooled and sent back down. It’s not a fast process and can take many months. (Sometimes, for speed’s sake an expendable refrigerant like liquid nitrogen is used, but it requires trucking in tanks full of the stuff.)
First, ice forms in columns around the freeze pipes. Then, as time goes on, the ice spreads out, linking the columns. Finally, an impermeable wall forms. For containment, it’s important that the ice extend all the way down to the bedrock, so that the walls of ice form a box with the bedrock at the bottom. If an earthquake cracks the ice or the power goes out for a period of times, refrigerating the ground again re-seals the wall.
“You have all this cold frozen soil that water wants to leak through,” Yarmak said. “But as the water leaks its way through, it freezes, and the wall heals itself back up.”
All this to say: As crazy as it sounds, humans regularly freeze vast chunks of Earth… because we can. (I am reminded of our unofficial motto: This is your world. Look at it.)
The Fukushima plant is not even the largest ice wall ever attempted. Sopko, Andersland’s last PhD student at Michigan State, told me that in the 1990s, he’d designed and installed a 3.5 kilometer perimeter wall that required 1,950 pipes, 159,000 meters of drilling, and eight 1,500 horsepower compressors for theAquarius gold mine in Ontario, Canada. Unfortunately, while the pipes were being installed, the price of gold plummeted and the system was never switched on.
“Right now, we’re currently involved in a pilot test in the oil sands. The proposed job would be 8 kilometers,” Sopko told me. “We’re right in the middle of freezing a pilot test.”
While some technologies need to change a lot as they scale up, ground freezing isn’t one of them. As Harris notes in Ground Freezing in Practice, “The method is not limited by problems of scale.”
“The three really large jobs that I’ve looked at, the only thing that makes those different than the small [mining] shaft is the coolant distribution system, being able to pump enough coolant through the pipes,” Sopko said. “It’s pretty easy to do, though. The pipes are the same and the compressors are the same.”
What’s really surprising is that the operation does not take that massive an amount of power. Sopko walked me through how much power one might need to get the job done. Japanese authorities have said the wall’s perimeter would be roughly 1,400 meters at a depth of 30 meters. We assumed they’d place a freezing pipe every meter and want the wall to be 2 meters thick. With those numbers in mind, Sopko made the back-of-the-envelope calculation that TEPCO would need about 6,000 horsepower of compressor to do the refrigeration during the active freezing period, which would probably take a couple of months. After that, the maintenance of the ice wall would require about 3,000 horsepower. Yarmak thought 6,000 horsepower was a pretty good “guesstimate,” as well.
In electrical terms, that’s about 4.5 megawatts of power, which is substantial, but less than a percent of a large power plant’s output.
The key problem ground freezing projects can run into, Sopko said, was fast flowing groundwater. Flow rates above 1 meter per day can make it difficult for the freeze wall to form. But he said that he’d spoken with people with knowledge of the site, who said the rate was a tenth of that, or about 10 centimeters per day.
The most difficult thing, as in all cryogenic barrier construction, is the drilling.
“The holes have to go in straight. They have to be parallel to each other,” Sopko said. “If the pipes deviate too far apart from each other, then, you don’t get closure between the two.” In other words, you’d have holes in your wall.
Arctic Foundation’s Yarmak also noted that the difficulty of the drilling would vary. The installation of the pipes on the inland side of the complex would be relatively easier because the water you’d encounter would be less contaminated. It’s on the other side, after the water has passed through the plant, that the drilling could get tricky.
“If it’s contaminated material, then everything gets really expensive, and things slow down. And you have to make sure you’re keeping your people safe and not screwing up the environment more than it already is,” Yarmak said.
However, if the engineers can get the inland and wing walls to form, then the amount of water flowing through the plant could drop enough to make drilling on the ocean side a little easier.
Still, working on a contaminated site is just difficult. At Oak Ridge, Yarmak’s crew had to stay on a patch of pavement that had been plopped down over an old cooling pond. “You couldn’t walk off the pavement. The pavement was clean, but the woods were not. You couldn’t go into the woods. If the leaves came down, you had to blow them away because they were contaminated,” he said. “It was quite an interesting job, but it was a little stressful. You wanted to make sure your crew stayed safe.”
At Fukushima, those problems will be even more extreme, but the cost of doing nothing is even higher.
Alexis Madrigal is a senior editor at The Atlantic, where he oversees the Technology channel.