Washington

Efforts to replicate an experiment that detected 'bubble fusion' in a small jar are struggling to win funds, prolonging a controversy over the validity of the result.

When the original experiment was published in March (Science 295, 1868–1873; 2002), it sparked a keen debate over whether nuclear fusion had actually occurred. A team led by Rusi Taleyarkhan, a nuclear engineer at Oak Ridge National Laboratory in Tennessee, claimed success based on measurements of two major by-products of fusion, neutrons and tritium.

But critics said that an internal review at Oak Ridge six months earlier had called those measurements into question. Recalling the discredited 'cold fusion' experiments a decade earlier, researchers asked for an independent effort to replicate the results.

Lee Reidinger, deputy director of Oak Ridge, convened a multidisciplinary team, including Taleyarkhan, to refine the original set-up and use better detectors. The effort was expected to cost up to $500,000.

Reidinger says he hoped to use part of Oak Ridge's budget that is reserved for research conducted at the laboratory's own discretion. But he says that officials at the Department of Energy told him this could only be used for original experiments, not for verifying existing work. Now the group is to submit a formal proposal for funding — a process that will take several more months, with uncertain prospects of success.

Taleyarkhan, meanwhile, has won a new grant from the Defense Advanced Research Projects Agency (DARPA), believed to be worth around $250,000, to continue his own investigation. The funding from DARPA, which specializes in high-risk projects, will be used to scale up his apparatus. Taleyarkhan will be looking at the process's potential not just for fusion power, but also for “more immediate applications” as a neutron source, perhaps for radiography or detection of explosives.

He also says he is helping several groups worldwide to reproduce his original result, although he declined to identify them for fear that publicity could adversely affect their funding proposals. “I believe it will take at least four to six months before you start seeing results,” he adds.