Washington

Physicists are coming to terms this week with one of the most audacious scientific frauds ever uncovered. The research in question was carried out at one of the world's best-known industrial laboratories and published in top journals — including this one (see editorial statement on page 425).

Jan Hendrik Schön: fired from Bell Labs after an inquiry found falsification of data. Credit: BELL LABS

An independent committee charged with reviewing work done by Jan Hendrik Schön at Bell Laboratories in Murray Hill, New Jersey, released its findings on 25 September. It found a “preponderance of evidence” that he falsified or fabricated data in 16 of the 24 alleged cases of misconduct that it looked at, involving 25 published research papers.

“Hendrik Schön showed reckless disregard for the sanctity of data in the value system of science,” said the report of the committee, which was chaired by Malcolm Beasley, an electrical engineer at Stanford University in California. Bell Labs — part of telecommunications corporation Lucent — fired Schön the evening before the findings were released. But the affair is far from over, with physicists asking not only what could have been done to prevent the fraud, but also what will become of more than 100 papers that Schön authored since arriving at Bell from Germany in 1998.

In a statement accompanying the report, Schön admitted to making “mistakes”, but said that he “never wanted to mislead anybody or to misuse anybody's trust”. He added: “I truly believe that the reported scientific effects are real, exciting, and worth working for.” He was unavailable for further comment.

Schön's work to create transistors out of single molecules and induce superconductivity in carbon 'buckyballs' won him acclaim, and his findings adorned the covers of Nature and Science. But researchers struggled to replicate his results, and as Schön's publication list grew, so did the community's scrutiny. In May, a group of researchers informed Bell Labs that they had discovered a series of three graphs that appeared identical in different publications, down to what should be random noise (see Nature 417, 367; 2002). This revelation, together with additional suspect graphs, led Bell to convene Beasley's panel.

Battlogg: regrets in retrospect.

“The report shows that scientific misconduct occurred,” says Beasley. “But there are other questions to ask.” The panel also took a close look at Schön's most frequent co-authors: Christian Kloc, Zhenan Bao and his supervisor Bertram Batlogg. “We found it very difficult to deal with the issue of the responsibility of co-authors,” says Beasley. The committee cleared Kloc and Bao of any wrongdoing, but it raised questions about Batlogg, who left Bell in autumn 2000 for the Swiss Federal Institute of Technology in Zurich, but whose name continued to appear on Schön's papers throughout 2001. The report asks if Batlogg should have “insisted on an exceptional degree of validation of the data in anticipation of the scrutiny that a senior scientist knows such extraordinary results would surely receive” or even “crossed the line of trust and questioned the integrity of the data”. In the end, it says: “The Committee does not consider itself qualified to make a specific judgement in this case.”

Some researchers close to the case claim that some responsibility must rest with the co-authors, particularly Batlogg. “You can't just take the kudos of authorship and then say, 'It's not my fault,'” says Charles Lieber, a chemist at Harvard University in Cambridge, Massachusetts, who was one of the first to note problems with Schön's data. “I think the buck stopped at Batlogg,” asserts Arthur Ramirez of Los Alamos National Laboratory in New Mexico, who has devoted his own research to replicating one of Schön's results.

“As co-author I acknowledge a responsibility to ensure the validity of data in publications,” Batlogg said in an e-mail. “I have learned, with the deepest of regrets, that the verification measures I have followed in this extraordinary case were not adequate to prevent or uncover scientific misconduct. I have placed, in retrospect, too much trust in my highly talented collaborator.”

Bao and Kloc prepared materials for Schön's experiments and are not regarded as sharing responsibility for his findings.

Researchers are now asking if Bell Labs management should have had a tighter grip on what its staff were doing. “This has to be a management failure on some level,” says Nobel laureate Philip Anderson of Princeton University in New Jersey, a former Bell researcher. Cherry Murray, vice-president of physical sciences at Bell, says that the laboratory responded promptly to two complaints it received about Schön's work, in October 2001 and February 2002. “In both cases Schön was very apologetic and the explanations did seem to reflect honest mistakes,” Murray says, “but with 20:20 hindsight, these cases should have alerted us to fraud.” She adds that steps have been taken to reinforce peer review inside the lab — including the creation of an internal server where all papers must be posted before submission for external publication.

Beasley: headed investigation.

Others ask why editors at journals, such as Nature and Science, weren't more suspicious of Schön's astonishingly high publication rate. Some critics charge that journals ignored negative feedback from the community because they wanted to publish high-profile work. “Nature's editorial and refereeing policy seems to be influenced by the newsworthiness of the work, not necessarily its quality,” says Anderson. “And Science seems to be caught up in a similar syndrome.”

But Karl Ziemelis, physical sciences editor at Nature, says that there was no evidence of wrongdoing when the papers were accepted for publication. “On several papers, referees had questions over the interpretation of the data,” Ziemelis says. “But they were nearly unanimous on the importance of the findings” (see Opinion, page 417).

“You can't design a peer-review system that's an absolute guarantee against clever fraud, but obviously when something like this happens, you're going to look at your practices,” adds Donald Kennedy, Science's editor-in-chief.

Journal publishers now have to decide what to do about the status not just of the 25 papers reviewed by the Beasley panel, but also of the rest of Schön's published work. Kennedy and Ziemelis both say that their journals will request retractions of the papers named as fraudulent by the committee. Science had not decided what to do about the other work, but the co-authors on all Schön's Nature papers will be asked if they want to retract.

Ramirez says that he now questions everything Schön has published. Colleagues have passed him eight more suspicious figures, including one that appeared in a Nature article not previously implicated (J. H. Schön, Ch. Kloc and B. Batlogg Nature 406, 702; 2000). Lieber says that “any paper Schön was first author on since arriving at Bell Labs should be withdrawn”.

Along with many other physicists, Myriam Sarachik of the City University of New York, president-elect of the American Physical Society (APS), believes that science is self-correcting, and that Schön's fraudulent work will eventually be either ignored or discarded. “The scientific method does ultimately weed out this sort of thing,” she says.

Nevertheless, Sarachik says, the APS will review its misconduct policy in light of the Schön case and another high-profile fraud case involving the claimed discovery of element 118 (see Nature 418, 261; 200210.1038/418261b). The society may also develop programmes to teach graduate students good laboratory practice, and designate an ombudsman to receive complaints of misconduct. “These latest episodes have made us realize that we'd better look at how we conduct our business,” Sarachik says.