As a way to resolve a scientific dispute, it was always likely to be fraught. In March 2005, nuclear engineer Rusi Taleyarkhan of Purdue University in West Lafayette, Indiana — known for his controversial claims to have achieved 'bubble fusion' — formally joined forces with one of his most prominent critics, physicist Seth Putterman of the University of California, Los Angeles.

But few could have predicted that this collaboration would end in such disarray. After concerns about Taleyarkhan's work were reported in Nature earlier this year1, Purdue carried out an inquiry, but has shrouded the results in confidentiality, a decision that has frustrated other researchers in the field. The findings could resolve the long-standing controversy surrounding bubble fusion, but whether they will ever be made public now seems to rest on a technicality: did $250,000 of US taxpayers' money help fund the disputed work?

The story began in 2002, when Taleyarkhan reported fusion in collapsing bubbles within a liquid, an effect also called sonofusion2. His work made headlines worldwide: if the effect could be harnessed, it would promise almost unlimited energy. But others in the field were not convinced.

In the hope of settling the resulting argument, in March 2005 DARPA, the Pentagon's research agency, paid Taleyarkhan and his critics to work together to replicate the bubble-fusion experiment. Putterman was principal investigator on the $812,000 grant and so has to account for all the expenditure. Taleyarkhan was allocated $318,000 of the grant, and Ken Suslick of the University of Illinois at Urbana-Champaign was given $145,000.

It was clear things weren't going well when concerns about the validity of Taleyarkhan's bubble-fusion work were reported on 8 March1. These included an analysis by Putterman's postdoc, Brian Naranjo3, showing that the neutrons described in Taleyarkhan's latest paper, published in Physical Review Letters (PRL) in January4, came not from fusion as claimed but from the radioactive decay of standard lab material.

Rusi Taleyarkhan denies that federal dollars funded certain aspects of his research. Credit: P. EFIRD/KNOXVILLE NEWS-SENTINEL/NEWSCOM/SHNS

Purdue's provost Sally Mason called the reports “very serious” and announced a three-month review, the outcome of which she promised to publish. But once the review was complete, Purdue said the results and any future steps would be kept internal and confidential. On completion of a review, Purdue's stated policy is either to close the matter or to proceed to a fully fledged misconduct investigation. The university won't say which has occurred.

As a principal investigator, do I have any right to know what's happening on my project?

Many in the field are disappointed by the lack of information, including Putterman. He is sure that money from the shared DARPA grant was used for the work Purdue reviewed, so he is particularly keen to know whether an investigation is under way. “As a principal investigator, do I have any right to know what's happening on my project?”

In cases where federal money is involved, there is a responsibility to tell the government and the taxpayer how it was spent — and any misuse of federal dollars can have serious consequences for the researchers involved. If DARPA money was used for any of the disputed work, Purdue would be required to notify the agency of any investigation, and share information relating to it. Such communications could eventually be made public under the US Freedom of Information Act.

Taleyarkhan did not acknowledge the DARPA grant in the January PRL paper. But when Nature asked Putterman to confirm no DARPA money was used, he requested the relevant accounts from Purdue. Putterman now says: “I've reviewed the books, and I am confident that the paper relied on federal money that was not acknowledged.” For example, Taleyarkhan and his colleagues claimed DARPA salaries in the run up to submission of the PRL paper (see 'Where did the money go?').

Taleyarkhan has declined to communicate with Nature directly. But he said through a third party — Brian Josephson at the University of Cambridge, UK — that Putterman's interpretation of how the work was funded is “off-base and wrong”. Josephson also provided part of an e-mail in which Taleyarkhan strongly denies using the DARPA grant on the disputed work. Taleyarkhan says the experiments were completed by May 2005, several months before the paper was submitted, and that start-up funding from Purdue paid for them. (The university told Nature that this funding totalled $58,607.) He adds that he and the others involved worked on the project outside the normal eight-hour day.

In the e-mail, Taleyarkhan also says that the bubble fusion described in the PRL paper is different from that reported in his previous papers, on which he has warmly acknowledged DARPA funding. He does not give details of what the $251,044 of DARPA money he spent was used for, if not the disputed work.

Taleyarkhan's explanation may make little difference if the case is investigated. “If any part of salary is allocated to a grant awarded by a federal agency, then federal funding is involved,” says Mark Frankel, director of the Scientific Freedom, Responsibility and Law programme at the American Association for the Advancement of Science in Washington DC. Nature has confirmed this general interpretation with an investigator at a federal funding agency.

Purdue says it has not queried Taleyarkhan's assertion that no federal money was used. “The authors of the paper are the best source of information on the source of support for research,” says a spokesman. “We have no reason to question the source of support stated.”