BAINBRIDGE ISLAND, Wash. — The toxic goo that seeps beneath wood-treatment plants is hopelessly sticky for good reason: It has to keep water from rotting railroad ties, pilings and telephone poles.
There's about 1 million gallons of it lurking in the soil and groundwater along an eastern stretch of this otherwise unpolluted island west of Seattle, making it one of the state's most heavily contaminated Superfund sites.
The Environmental Protection Agency tried a high-tech, steam-injection project, hoping to clean up a century's worth of carcinogenic creosote, pentachlorophenol and other toxins.
It was supposed to chart a promising new course for dealing with one of the toughest types of pollution to clean up. But now the EPA says it looks like the $10-million pilot project has failed.
Officials blame a series of equipment failures and a pesky chemical called naphthalene, which choked the system's pipes with crystals as it cooled down after being treated with the superheated steam.
Before that glitch, the project worked well -- so well that it flooded an old wastewater treatment plant with more sludge than it could handle.
"One piece of the question is: Can you move a lot of stuff out of the ground? The answer to that is undoubtedly yes," said Wallace Reid, an EPA project manager at the site. "The other question is, if you move a lot of stuff out of the ground, can you meet cleanup standards so that we're basically done with the project? The answer to that is no."
Wood preservatives are the main problem at more than 80 Superfund sites across the country, which raised hopes that a success at the site here, last owned by Wyckoff Inc., could lead to easier cleanups elsewhere. Although the experiment on Bainbridge Island didn't work as planned, EPA officials insist it hasn't been a total failure.
"Every lesson is valuable," said Jim Cummings, an engineer in EPA's technology innovation office. He noted that steam injection has worked at a utility pole yard in Visalia, Calif., well enough to be picked as the cleanup method for a site in Marion, La., while other types of thermal treatment are being tested at other Superfund sites.
EPA dreamed up the Wyckoff project in the late 1990s, after state and local officials balked at a plan to cap and contain the pollution rather than trying to get it all out of the ground.
Steam cleaning began in November 2002 and ended about six months later because of technical problems. Cleanup crews kept treating the groundwater, but without steam injection, they couldn't treat as much of the underground goop as originally hoped, said Cliff Leeper, operations manager for the contractor handling the project.
Since 1990, the EPA has removed more than 100,000 gallons of sludge and treated some 416 million gallons of contaminated groundwater.
On the western side of the site, a stretch of shoreline that once reeked of wood preservatives has been restored into a natural-looking beach scattered with bits of driftwood.
"This area was just a chemical pit," Reid said. "When you came in on the ferry, you could smell it."
Officials say a 100-foot-deep containment wall surrounding the site also has done its job. The rusty top edge of the wall's interlocked steel beams doesn't look pretty jutting up from the shoreline, but it's stopped the pollution from seeping into Eagle Harbor, where contamination once was so bad it gave liver lesions to fish.
To date, the cleanup has cost about $100 million. Most of the money came from the Superfund trust, which dried up recently, about eight years after then-House Speaker Newt Gingrich led a push in Congress to let a tax on polluting industries expire.
The rest came from a trust set up when Wyckoff Inc., later renamed Pacific Sound Resources, went out of business in the mid-1990s. Future funding will come from taxpayers and the local trust -- which has also paid for a cap-and-contain Superfund cleanup in West Seattle.
Most Superfund sites are surrounded by heavily industrialized areas. The Wyckoff site is nestled between wooded slopes and a stunning view of the Seattle skyline.
The City of Bainbridge Island and the local park district are working to buy the land and turn it into a waterfront park. EPA launched the steam-injection project under pressure from community leaders who wanted the site completely rid of pollution.
"People said, 'You'll do steaming and we'll be done and have a pristine area,' " Reid said.
So much for those hopes. Until the EPA decides whether steam injection can be re-engineered to work right, the agency is focusing on keeping the wastewater treatment plant running, designing a new one, and installing one more containment wall at the base of a hill on one edge of the site to reduce the amount of groundwater it has to treat.
"We're disappointed," said Libby Hudson, senior planner for the City of Bainbridge Island. "Hopefully something will work. It's a real shame to leave 1 million gallons of contaminants underground."
Meanwhile, in southwest Washington, the state Department of Ecology is raving about a steam-injection project that's cleaning part of the Port of Ridgefield that's sullied by wood preservatives.
Learning from the pitfalls at Wyckoff, Ecology engineers designed their system to keep certain pipes extremely hot, which has kept them from getting clogged with crystals.
"With a treatment system that complex, you'd expect lots of problems, but we haven't had any," said Dan Alexanian, Ecology's site manager for the $8-million project.
Mary Jane Nearman, an EPA project manager, said Wyckoff's steam-injection system could be reworked the same way, but said it would not likely be feasible. "Our cleanup standards at Wyckoff are very, very stringent because of our proximity to the sound," she said.
Ultimately, EPA officials say there's no telling how long the soil and groundwater will have to be treated at Wyckoff.