Gasoline additive's going, but far from gone.

Last month, the Environmental Protection Agency announced plans to phase down, and perhaps out, use of a suspected carcinogen that refiners have added to gasoline since the 1970s. An octane booster, methyl tert-butyl ether--better known as MTBE--also reduces a vehicle's emissions of carbon monoxide and ozone. Owing to myriad gasoline leaks, however, MTBE now taints drinking-water supplies throughout the nation.

That contamination, and the cleanup challenges it poses, was explored in some 50 papers presented during a 4-day symposium held last week as part of the American Chemical Society's spring national meeting in San Francisco.

What's clear "is that we know how to treat MTBE and remove it from drinking water," notes Andrew Stocking of the consulting firm Malcolm Pirnie in Oakland, Calif.

Unfortunately, no one has yet cleaned up an MTBE-tainted aquifer--and there are many, observes Lisa S. Dernbach, an engineering geologist with the California Regional Water Quality Control Board in South Lake Tahoe.

Lakes and streams have been tainted with MTBE-treated gasoline spewed by two-stroke engines such as those that power personal watercraft. Much of the U.S. contamination, however, threatens groundwater. Underground fuel tanks, many beneath gas stations, are a primary source. Some 250,000 of them leak MTBE-laced gasoline, according to an analysis that will appear in the May 1 ENVIRONMENTAL SCIENCE & TECHNOLOGY.

How much MTBE has escaped from these tanks "remains the big unknown," says coauthor John S. Zogorski of the U.S. Geological Survey in Rapid City, S.D. No one knows either how long most of these tanks have leaked or the typical MTBE concentrations in tainted aquifers, he explains. Among the water samples that have been analyzed, MTBE concentrations range up to a few parts per million.

Zogorski argues that to predict how MTBE contamination will move through the environment, hydrologists need better data on drinking-water extraction rates than are now available. As he explains, "The more water that is pumped from an aquifer, the more you pull [nearby] contaminants into that aquifer."

It's an issue that Dernbach knows well. Currently, 15 of the 36 drinking-water wells serving her area have been taken out of service to avoid pumping up noxious-tasting MTBE-laced water. Yet she found unanticipated impacts of these shutdowns on two of the MTBE plumes that she's studying. Closing the wells greatly altered the plumes' path and accelerated their spread.

Dernbach now suspects that in some cases, it will prove easier and less costly to clean drinking water after it's pumped from the ground than it is to chase an expanding underground plume with cleanup systems.

With so much MTBE already flowing through aquifers, however, some underground treatment may be necessary. The half-life for this pollutant in untreated water ranges from 2 to 3 years, according to new laboratory studies by Clinton D. Church of the Oregon Graduate Institute in Portland. In other words, he explains, it would take 15 to 20 years for a concentration of 100 parts per billion to fall to around 1 ppb.

Numerous ways to speed the process are under investigation. For instance, Douglas MacKay of the University of Waterloo in Ontario is boosting MTBE's degradation by naturally occurring bacteria. Most tainted aquifers have microbes that would chow down on the pollutants if they had lots of oxygen, he notes. Yet in the California aquifer that he's studying, he says, "we can detect no oxygen anywhere in, near, or around the MTBE plume."

To remedy the problem, he's inserted a system of plastic pipes into an aquifer to diffuse a bounty of oxygen to the microbes there. As water sweeps by the porous pipes, its oxygen concentration becomes saturated, he says. This spurs the microbes' appetite for the pollutant. MacKay reports, "We can knock down MTBE concentrations from several hundred parts per billion to below detection limits--which means below 5 ppb." The 5 ppb concentration, which gives a turpentine flavor, is California's MTBE limit for water. The additive is considered toxic only at higher concentrations.

Marc A. Deshusses of the University of California, Riverside is taking another tack. He's trickling tainted water through filters packed with MTBE-eating microbes. Once he "acclimatizes" the bacteria to their pollutant-enhanced diet, each gram (dry weight) of the microbes can degrade 5 or 10 milligrams of MTBE to carbon dioxide per hour. At high starting concentrations of MTBE, breakdown takes hours to days.

Ironically, MacKay notes, "despite all the concern over MTBE in water, there have been very few incentives--and almost no funding--to study its remediation [in groundwater]." He says that it's as if, once the additive is banned, the problem will go away. "But that is absurd," he maintains, because the current water contamination will persist for decades.

Until it's gone, Dernbach argues, MTBE will threaten ecosystems and drinking water like "a time bomb," requiring costly cleanups. Water in one plume that she's following has for 2 years been pumped up, treated, and returned to the aquifer or drained into sewers. Finishing the cleanup may take another decade, she says, and cost $5 million.