On April 20, 2010, an explosion tore through Deepwater Horizon, an oil rig in the Gulf of Mexico. Over the next three months, the exposed wellhead released some 750 million liters of oil into the Gulf.

To cope with the record-breaking volumes of oil, authorities decided to dump 7 million liters of a dispersant called Corexit into the Gulf. This substance would break the oil slick into smaller clumps, preventing it from washing onto beaches, or clogging the fur and feathers of coastal wildlife. The smaller particles would also be easier fodder for oil-digesting microbes, which have evolved to break down hydrocarbons that naturally seep from oceanic vents.

But a new study by Samantha Joye at the University of Georgia shows that, at least in terms of the latter goal, the dispersants failed miserably.

By simulating the Deepwater spill in their laboratory, Joye's team found that the dispersants actually suppressed oil-busting bacteria and slowed their ability to degrade oil. Instead, they favored microbes that, well, excel at digesting dispersants.

“You could argue that it would have been better if we had left the organisms alone,” says Joye.

“This work is a wonderful proof of the inefficacy of [Corexit],” says Jack Gilbert from the University of Chicago. “It's also another example of why it is always better to test the microbial system you wish to exploit prior to widespread application of a compound.”

Dispersants have been used in oil spills before and had often been considered as the main line of defense against such catastrophes. But when Joye started looking into the evidence behind this position, she became concerned. Some papers claimed that they help bacteria to break down oil, others said that they slow the process, and yet others found no effect. The problem, Joye realized, was that they were all using different measurements, most of which were indirect.

To get better answers, her team, including ex-student Sara Kleindienst, collected water from a site in the Gulf of Mexico where hydrocarbons naturally seep into the oceans. Back in the lab, they flooded the water with either oil, dispersants, or both, at concentrations and temperatures that mimicked the Deepwater spill and its aftermath. “The goal was to recreate the Deepwater plume as best we could,” says Joye.

The team analyzed these microcosms with a battery of techniques. They counted microbial cells, and sequenced them to identify the species that were present. They quantified the enzymes made by those microbes to gauge how active they were. They checked the production of microbial molecules used to break down the oil. And they used slightly radioactive oil compounds to directly measure the rate at which those substances were being degraded. “No one has ever done all of that in a single experiment,” says Joye.

These techniques showed that dispersants fueled the growth of Colwellia bacteria, which went from 1 percent of the total microbes to up to 43 percent. Colwellia has sometimes been billed as an oil-degrader, but its presence was minimal in the oil-only water. If anything, it was more of a Corexit-degrader.

Marinobacter, however, is a bona fide oil-busting specialist. In oil-only water, its went from 2 percent of the total microbes to 42 percent. But Corexit curtailed this ascendance, either by affecting Marinobacter directly, or by boosting competitors like Colwellia. Either way, the dispersants suppressed all strains of Marinobacter. And, according to the team’s varied techniques, the dispersants reduced the breakdown of hydrocarbons from the oil.

“So what happened to the hydrocarbons?” asks Joye. In another recent paper, written on the fifth anniversary of the disaster, she estimated that microbes broke down 43 to 61 percent of the oil, while 2 to 15 percent sank to the sea floor. “We we really can't account for 24 to 55 percent of the oil. Where is it? I don't know and it really bothers me. I suspect that some is in marshes, some is on beaches, and probably more than we think is on the seafloor. But we did a very poor job—and I include myself here—of measuring hydrocarbon degradation.”

Should authorities avoid dispersants in the future? “That's an extraordinarily complicated question,” says Joye. Corexit has its problems, but it does seem to keep oil away from coasts. “Nobody wants to see oiled birds, turtles, and dolphins, but the bottom line is that if you disperse that oil, it's still in the water. You feel better, but is it improving the situation? My gut instinct is that I would put my faith in the microbial communities to do their job.”

But Terry Hazen at the University of Tennessee, who previously suggested that oil-eating microbes were indeed blooming in the Deepwater plume, is not convinced by the team’s analysis. He argues that these microcosm experiments create a “bottle effect,” which can't adequately simulate the complicated dispersal of oil and dispersants through deep seawater.

Joye counters that her team also repeated their experiments using water collected from an actual ongoing oil slick, albeit one happening in surface waters, and saw the same suppression of hydrocarbon breakdown. “That says this isn’t a fluke of the experiment,” she says. “It’s a pattern.”

Leila Hamdan from George Mason University also agrees with the team’s conclusions. She previously showed that Corexit is toxic to Marinobacter, and has upcoming data showing that dispersants also affect microbes growing on underwater surfaces, like the hulls of ships. “The deep ocean is home to the largest population of microorganisms on earth—what I call ‘the invisible majority,’” she says. “It’s their world, and this paper provides evidence that their ability to do their job is greater when we do not introduce other chemicals into their environment.”

That said, it may be worth finding stimulants that actually boost the oil-degrading microbes. For example, Joye found that “a little bit of nitrogen makes Marinobacter very happy.” The concept of introducing nitrogen, so commonly associated with algal blooms, into the oceans would be heresy to many scientists, but Joye wonders if adding nitrogen-making bacteria would work better.

“There's a lot of stuff we need to know answers to before tomorrow arrives,” she says. “When you look at the wells being drilled, it's not a matter of if, but when. Do you think you could seal a well faster now than we could six years ago? I think we're going to be looking at many of the same problems.”