In a move that surprised very few people, the journal Nature retracted a paper claiming a major advance in high-temperature superconductivity. This marks the second paper the journal retracted over the objections of Ranga P. Dias, a faculty member at the University of Rochester who led the research. Or at least it’s implied that he objected to this retraction, as he apparently refused to respond to Nature about the matter.
Dias’ work on superconductivity has focused on hydrogen-rich chemicals that form under extreme pressures. Other research groups have shown that the pressure forces hydrogen into crystals within the material, where it encourages the formation of electron pairs that enable superconductivity. This allows these chemicals to superconduct at elevated temperatures. Dias’ two papers purportedly described one chemical that could superconduct at room temperatures and extreme pressures and a second that did so under somewhat lower pressures, putting it within reach of more readily available lab equipment.
But problems with the first of these papers became apparent as the research community dug into the details of the work. Dias’ team apparently used a non-standard method for calculating the background noise in a key experiment and didn’t include the details of how this was done in the paper. In other words, the data in the paper looked good, but it wasn’t clear whether it accurately reflected the experimental results. As a result, Nature retracted it, although all nine authors of the paper objected to this decision at the time.
So it was surprising that the same journal accepted a paper describing similar work from the same research group. It was perhaps less surprising that similar problems cropped up. In this case, eight of the paper’s 11 authors say they aren’t at all confident the paper displays data in a way that accurately represents what went on in the lab. As the retraction notice puts it, “They have expressed the view as researchers who contributed to the work that the published paper does not accurately reflect the provenance of the investigated materials, the experimental measurements undertaken, and the data-processing protocols applied.”
A rough translation from the academic language: “We have little idea how the images of the data in the paper were generated.”
As noted above, Dias, along with two colleagues at the University of Rochester, hasn’t responded to the retraction. His spokesman apparently told The New York Times that “Professor Dias intends to resubmit the scientific paper to a journal with a more independent editorial process.” It is not clear how “independent” would translate to “finds it acceptable that most of the people who supposedly generated the data worry it could be faked.”
If anything, Nature’s failure here is that it seemingly did handle the peer review of the second paper as if it were independent from the first. In a sense, that was idealistic, ignoring any social context and simply focusing on what was presented in the paper. But it was also naive, given that the earlier paper was retracted precisely because the paper didn’t present an accurate picture of the experiments.
As for Dias, this may end up being the least of his worries. A third paper he was on, published in Physical Review Letters, has also been retracted (again, over Dias’ objections). In this case, there are indications that a graph purportedly showing recent data was simply copied from Dias’ thesis, which was on a different topic entirely. There are also accusations that his thesis contained plagiarized material. The University of Rochester has started a review of Dias’ work, and although the findings of those reviews are generally kept confidential, any consequences resulting from them will be difficult to miss.