A debate over claims of room temperature superconductivity has now boiled over into the realm of scientific publishing. Administrators of arXiv, the widely used physics preprint server, recently removed or refused to post several papers from the opposing sides, saying their manuscripts include inflammatory content and unprofessional language. ArXiv has also banned one of the authors, Jorge Hirsch, a theoretical physicist at the University of California, San Diego (UCSD), from posting papers for 6 months.
The ban is “very unfair,” Hirsch says. “I can’t work if I can’t publish papers.”
To some other scientists, arXiv’s ban and removal of papers amount to stifling scientific debate. “The scientists that care about the issue and have the expertise to evaluate the arguments on both sides should be allowed to do so by accessing the preprints in question,” Nigel Goldenfeld, a physicist at UCSD, wrote in an email to a wide range of physicists last week. “The alternative is that for cases such as this, we’ll return to the pre-arXiv days when the science of the day is discussed in privately circulated preprints that are not accessible to the wider community.” Daniel Arovas, another UCSD physicist, agreed: “Squelching what is essentially a purely scientific exchange—even one where the respective parties engage in some distasteful accusations—is highly problematic.”
But arXiv administrators argue the decision wasn’t about science. “There are no papers in this whole chain that were rejected because we did not like the scientific content,” says Ralph Wijers, a physicist at the University of Amsterdam who is the preprint server’s board chair. “People’s emotions became too affected. They got acrimonious.”
Launched in 1989, arXiv has ballooned to host more than 2 million preprints and has become a vital communication hub for physicists, astronomers, cosmologists, mathematicians, and computer scientists. Preprints typically appear within a day or two of submission, allowing near real-time discussion of complex science. Physicists embraced arXiv decades before other fields became comfortable with posting articles without peer review. “It’s a tremendous value to the physics community,” says Frances Hellman, a physicist at UC Berkeley, and president of the American Physical Society.
Not all of arXiv’s estimated 15,000 monthly submissions are accepted. Some 200 volunteer moderators scan submissions to ensure they cover legitimate scientific research that is of interest to the community. Papers that don’t appear to be scientifically sound or use “unprofessional” language can be rejected. Review boards then manage appeals.
Rejections are “rare,” perhaps 1% of submissions, says Steinn Sigurdsson, arXiv’s scientific director. But moderation helps ensure papers don’t include invective against other scientists, says Paul Fendley, a theoretical physicist at the University of Oxford and an arXiv advisory committee member. “If we allow this stuff, what is the difference between arXiv and Twitter?”
Moderators believed Hirsch crossed that line in papers critiquing a 14 October 2020 Nature publication by a team led by Ranga Dias, a physicist at the University of Rochester. The paper, reporting the discovery of a hydrogen-containing material that under intense pressure superconducts at near–room temperature, was hailed as the culmination of a centurylong quest to create a superconductor that wouldn’t have to be chilled to ultralow temperatures. But the bold claim also sparked controversy.
About a month later, Hirsch asked Dias for the raw data from some of the experiments. He says Dias rebuffed him repeatedly. Eventually, Hirsch did receive some data from one of Dias’s co-authors, and in August 2021, Hirsch submitted his own analysis to both the arXiv and Physica C. The paper was titled “On the ac magnetic susceptibility of a room temperature superconductor: anatomy of a probable scientific fraud.” After publishing it online in September, Physica C removed the article in November because it contained data published without the original team’s permission, and arXiv took it down in December.
On 29 November 2021, Dias and one of his collaborators, Ashkan Salamat, a physicist at the University of Nevada, Las Vegas, posted on arXiv a response to Hirsch’s criticisms, and included some of their raw data. In early December, Hirsch submitted two papers analyzing those raw data, and then followed up with three more papers, all of them responses to work by Dias and his colleagues. ArXiv administrators blocked all five. (Hirsch also says posting of multiple submissions has been delayed for weeks or more and papers were taken down even after they were posted.) Last week, the site also removed a paper from Dias and Salamat “due to inflammatory content and unprofessional language.” Dias and Salamat did not respond to repeated requests for comment.
Hirsch himself defended Dias and Salamat’s paper in an email to arXiv administrators. “It presents scientific arguments that the scientific community should be allowed to judge on their merits, rather than be prevented from doing so by your arbitrary self-righteous decorum standards.” Dias and Salamat have been invited to modify the offending language in their paper and resubmit it, Wijers says. And a modified version of one of Hirsch’s papers—on which he was the second author—was also reposted last week, minus a previous assertion of data manipulation in the 2020 Nature paper.
Modification is not likely to happen with Hirsch’s other offending papers, on which he was the first or only author. On 7 February, arXiv banned him from posting for 6 months, including revisions of previous papers.
Sigurdsson says he is unable to discuss the case but notes that bans can occur for reasons other than repeatedly publishing inflammatory content. Authors can also be banned for making repeated submissions in response to specific papers and contacting arXiv administrators and board members to complain—all infractions Hirsch appears to have committed. “We do not want to be flooded with separate comments” on single papers, Sigurdsson says. “Our moderators are a noise suppression machine.”
Other physicists worry moderators are making arbitrary decisions. Moderators “seem to be too keen” about removing controversial papers, says Brian Josephson, a physicist at the University of Cambridge. “And we don’t know what their prejudices might be.”
Hellman says the superconductor controversy may stem in part from the ethos of physics, which has historically encouraged combativeness. “The culture of physics is one that is more aggressive and not very welcoming,” Hellman says, which can lead to accusatory language ending up in papers. She would like to see that change. “I flinch at some of the language being used.”