A nanotechnology paper is spending the summer in a state of electronic limbo, after Science published it online but declined to follow up with the usual publication in print.

The article, which shows evidence for the possibility of combining the hot fields of spintronics and molecular transistors, appeared on Science's website Science Express on 18 April. Science Express papers normally appear in the print journal within “several weeks”, according to the website. But after the paper's lead author Jan Hendrik Schön, a researcher at Bell Labs in Murray Hill, New Jersey, was accused of fraud in May, the paper's print publication was delayed. A panel chaired by Malcolm Beasley of Stanford University in California is investigating several papers authored by Schön, who has denied the allegations (see Nature 417, 367–368; 2002).

Science would usually send a final version of the paper, edited for print publication, to the authors for approval, but has not done so in this case, says Eldon Emberly of the NEC Research Institute in Princeton, New Jersey, a co-author of the paper. Science's editor-in-chief, Donald Kennedy, says that the hold-up is due to the ongoing investigation into Schön's work. Although Kennedy does not know whether this particular paper is part of the investigation, he says that a decision to assign it a page number will be put on hold until the results of an official inquiry are made public later this summer. “If there is a question about a paper, why would we make the financial investment in printing it? That would just be flat-out silly,” says Kennedy.

But some other journal editors fear that the situation will confuse researchers, including those who are considering citing the paper, as it will not be clear whether the work was ever published in Science or not. Because the paper can be cited by others, it needs to be clear what will happen if it never appears on Science's pages, says Martin Bloom, editor of Physical Review Letters, which has considered its online version the “journal of record” since 2001. Bloom adds that his journal would have no choice but to publish an erroneous paper if it had already appeared online. “We would still print it along with a withdrawal,” he says, adding that Science “needs to set out a policy”.

Others worry that the electronic paper, were it to be retracted, might be lost entirely, potentially undermining the integrity of the scientific record. Clear policies are needed for managing online papers, given the fact that they can be altered or removed from websites, says Paul Ginsparg, a professor of physics and information science at Cornell University in Ithaca, New York, and founder of the Los Alamos preprint server — a service that allows scientists to post papers online before peer review. Although Science has a right not to publish the paper in print, he says, it ought to leave the original paper online, with appropriate updates on its status.

Nature editor Philip Campbell says that this journal's existing policy would prevent such an occurrence. Whereas Science Express papers pass through a final edit before appearing in print, a Nature paper that appears online is considered to be final. Furthermore, authors whose papers are published online are assured that their papers “will be published in print as soon afterwards as space is available”. Campbell says that, in a similar set of circumstances, Nature would not reconsider its policy.

Kennedy says that any decision about the precise fate of Schön's paper will have to wait until the investigation is published, but that the work will not be removed from the electronic archive. “The paper isn't in limbo at all — it's published,” he says. “I really don't think it's a problem.”

But Ginsparg disagrees. “The question is: will that URL on Science Express be there 10–20 years from now?” he asks.