So Physical Review Letters formally retracted that paper about manganese sulphide, in the limelight for having been coauthored by Ranga P. Dias, yesterday. The retraction notice states: “Of the authors on the original paper, R. Dias stands by the data in Fig. 1(b) and does not agree to retract the Letter.” Figure 1(b) is reproduced below.
The problem with the second plot is that its curves reportedly resemble some in Dias’s doctoral thesis from 2013, in which he had examined the same properties of germanium tetraselenide, a different kind of material. Curves can look the same to the extent that they can have the same overall shape; it’s a problem when they also reproduce the little variations that are a result of the specific material synthesised for a particular experiment and the measurements made on that day.
That Dias is the only person objecting to the retraction is interesting because it means one of his coaouthors, Ashkan Salamat, agreed to it. Salamat heads a lab in the University of Nevada, Las Vegas, that’s been implicated in the present controversy. Earlier this year, well after Physical Review Letters said it was looking into the allegations against the manganese sulphide paper, Scientific American reported:
Salamat has since responded, suggesting that even though the two data sets may appear similar, the resemblance is not indicative of copied data. “We’ve shown that if you just overlay other people’s data qualitatively, a lot of things look the same,” he says. “This is a very unfair approach.”
Physical Review Letters also accused Salamat of attempting to obstruct its investigation after it found that the raw data he claimed to have submitted of the group’s experiments wasn’t in fact the raw data. Since then, Salamat may well have changed his mind to avoid more hassle or in deference to the majority opinion, but I’m still curious if he could have changed his mind because he no longer thought the criticisms to be unfair.
Anyway, Dias is in the news because he’s made some claims in the past about having found room-temperature superconductors. A previous paper was retracted in September 2022, two years after it was published and independent researchers found some problems in the data. He had another paper published in March this year, reporting room-temperature but high-pressure superconductivity in nitrogen-doped lutetium hydride. This paper courted controversy because Dias et al. refused to share samples of the material so independent scientists could double-check the team’s claim.
Following the retraction, The New York Times asked Dias what he had to say, and his reply seems to bring back the bus under which principal investigators (PIs) have liked to throw their junior colleagues at signs of trouble in the past:
[He] has maintained that the paper accurately portrays the research findings. However, he said on Tuesday that his collaborators, working in the laboratory of Ashkan Salamat, a professor of physics at the University of Nevada, Las Vegas, introduced errors when producing charts of the data using Adobe Illustrator, software not typically used to make scientific charts.
“Any differences in the figure resulting from the use of Adobe Illustrator software were unintentional and not part of any effort to mislead or obstruct the peer review process,” Dr. Dias said in response to questions about the retraction. He acknowledged that the resistance measurements in question were performed at his laboratory in Rochester.
He’s saying that his lab made the measurements at the University of Rochester and sent the data to Salamat’s lab at the University of Nevada, where someone else (or elses) introduced errors using Adobe Illustrator – presumably while visualising the data, but even then Illustrator is a peculiar choice – and these errors caused the resulting plot to resemble one in Dias’s doctoral thesis. Hmm.
The New York Times also reported that after refusing in the past to investigate Dias’s work following allegations of misconduct, the University of Rochester has now launched an investigation “by outside experts”. The university doesn’t plan to release their report of the findings, however.
But even if the “outside experts” conclude that Dias didn’t really err and that, honestly, Salamat’s lab in Las Vegas was able to introduce very specific kinds of errors in what became figure 1(b), Dias must be held accountable for being one of the PIs of the study – a role whose responsibilities arguably include not letting tough situations devolve into finger-pointing.