Mechanical Engineering Professor Ranga P. Dias will see his second official paper retraction, following the results of an investigation by premier physics journal Physical Review Letters (PRL).
Previously lauded as an “editor’s suggestion” piece by PRL, the 2021 paper on manganese disulfide (MnS2) marked a continuing step on the journey to achieve room-temperature superconductivity — an effort that has proven unsuccessful for decades.
Dias’ 2021 paper focuses on the electrical properties of MnS2, including a dramatic, reversible reductio
A recognized name in the physics community, Dias’ superconductor research rocketed him to the world stage. Recognized as a TIME100 Next innovator in 2021, his work with superconductors is widely considered to be a promising step towards the technology of the future.
Superconductivity is a phenomenon where certain materials can conduct electricity with no resistance and no energy loss. The caveat is that superconductivity usually only occurs at incredibly cold temperatures that hover just above absolute zero, a condition that is highly cost-intensive to maintain.
Today, superconductors play critical roles in high-current wires, scientific accelerators like the one at CERN, and even magnetic resonance imaging (MRI) machines. The advent of a room-temperature superconductor could be the key to cost-effective superconductivity — and the key to engineering fiction’s wildest dreams.
“Let’s be clear: hoverboards, magnetic levitation trains and resistance-free power lines are not coming this year or next. But thanks to Ranga Dias, they’re closer than they ever were,” said Time.
Dias’ work is even slated to “find a place in physics textbooks,” according to the University News Center.
Last year, however, following concerns expressed regarding the validity of the MnS2 data, PRL conducted an internal investigation with a number of independent experts. Now concluded, the results of the inquiry support the accusations against Dias and his team.
Reports also indicate noticeable differences between the published data and the alleged “raw data” that Ashkan Salamat, Associate Professor of Physics at the University of Nevada, Las Vegas, and a fellow author on the MnS2 paper, was asked to submit for the investigation. Salamat is alleged to have deliberately obstructed the investigation by submitting manipulated data.
Additionally, the data from the PRL publication bears striking resemblances to Dias’ own PhD thesis, completed in 2013.
What led to the retraction
Upon discovering a match between a resistivity plot for germanium selenide in Dias’ 2013 thesis and the plot for MnS2 in the 2021 paper, Professor James Hamlin, a physicist currently with the University of Florida, reached out to Dias and the rest of the paper’s authors with his findings.
Simon Kimber, a coauthor of the 2021 PRL paper and a materials scientist previously with the University Burgundy Franche-Comté, reached back.
According to Kimber, “it was absolutely clear that something very strange was going on.”
Dias’ thesis and the PRL paper were on two completely different materials, with three different datasets, three different conditions, and three different pressures. Yet, after scaling the data – they were “basically identical,” he said. “To me, as a physicist, it was completely obvious that this was impossible.”
Concerned by the apparent similarities, Kimber emailed PRL to request a retraction of the 2021 paper, kicking off the investigation that has since concluded.
PRL has formally requested the paper’s retraction — and, failing the authors’ cooperation, will unanimously retract the paper on their own authority.
This is not the first of Dias’ papers to be brought into question — nor the first to face retraction.
Three of his papers now hold contention with the science community: his 2013 PhD thesis from Washington State University; his retracted paper on carbonaceous sulfur hydride, published by Nature in 2020; and his now-retracted 2021 paper on manganese disulfide, published by PRL.
A recently published database of retractions shows that only four in every 10,000 papers are retracted — and, as pointed out by Science, “relatively few authors are responsible for a disproportionate number of retractions.”
PRL operates under the standards set forth by the Committee on Publication Ethics (COPE), a non-profit group advising journal editors and publishers worldwide. COPE’s retraction guidelines provide overarching recommendations for editors, and recommend retractions for findings that are evidently “unreliable, either as a result of major error […] fabrication or falsification.”
The first retraction
Three years ago, Dias and his team were initially praised for their claims of a breakthrough in physics: the discovery of the first room-temperature superconductor in the form of carbonaceous sulfur hydride (CSH).
The results of their research were published in the 14 October 2020 issue of Nature, to global acclaim.
However, the results were controversial — not only due to the stunning claims, but also a lack of clarity regarding the paper’s methodologies.
The American Physical Society (APS) dove into the detailed allegations of misconduct against Dias and his team for their 2020 paper. The allegations, which included accusations of plagiarism, data manipulation, and poorly constructed data, resulted in the CSH paper’s retraction the following year.
Among Dias’ skeptics was Jorge E. Hirsch, a condensed-matter theorist at the University of California, San Diego. A long-time critic of room-temperature superconductivity — especially in hydrogen-based materials, as Dias used — Hirsch decided to check over the paper for flaws after its initial publication.
He focused on the measurements of magnetic susceptibility, a property that describes how well a material will magnetize, and what he found concerned him.
Magnetic susceptibility should drop sharply when the material enters the superconducting state. Then, as the temperature continues to fall, the curve should flatten or rise very slowly, according to the APS.
To Hirsch, the shape of one of the susceptibility plots in the CSH paper (“Extended Data Figure 7d”) seemed strange, as the low-temperature slope showed a sharp jump upwards.
Curiously, the data bore some resemblance to the magnetic susceptibility of europium, as published in a PRL paper in 2009.
Hirsch first reached out to one of the authors, Matthew Debessai, but was met with no response. So, he moved onto another author — James Hamlin, the physicist who found the similarities between Dias’ PhD thesis and the 2021 paper.
According to Hamlin, he and Hirsch found “one issue after another” with the europium paper, including a section of magnetic susceptibility data that appeared to have been copy-and-pasted across various temperature ranges.
Debessai’s europium results were retracted in December of 2021, with further investigation by another of the paper’s authors finding no evidence of the element’s superconductivity.
Hamlin took his concerns to Dias and Salamat, but they allegedly seemed unconcerned about the seemingly manipulated susceptibility data. They were, instead, concerned that the allegations could go viral.
“I had also raised concerns about the resistivity data, the electrical resistivity data, which were kind of never responded to by the authors,” Hamlin said.
In an analysis published after the CSH paper’s retraction, Hamlin found that Dias’ electrical resistance data displayed discrete steps in some places, and smooth slopes in others.
Digitization creates smooth slopes and discrete “steps” when data is more or less precise, respectively. When data is especially precise, there should be no visible steps — but the inclusion of both in Dias’ data implies at least two signals of differing digital precision.
This, according to Hamlin’s research, raised “questions concerning the methods used to obtain the published data.”
While Hirsch and Hamlin dissected the meat of the data, others hoped to replicate the CSH results, as described by the APS. Among them were Alexander Goncharov, a materials scientist at D.C.’s Carnegie Institute, and Mikhail Eremets, of the Max Planck Institute for Chemistry in Germany.
The description of how to synthesize CSH was “scarce but still sufficient,” according to Goncharov, but after months of work, both parties’ independent experiments grounded nothing.
They turned to Dias for guidance, but, according to the APS, were given no help. No one could reproduce Dias’ data, and for just over a year, Dias refused to provide the CSH data files, despite considerable urging on Hirsch’s behalf.
“Finally, in December of 2021, which is a year and two months after the paper was published, [Dias and Salamat] did release the data,” said Hirsch. “Meaning the raw data and the background signal data, from which the published data were obtained.”
Hirsch would then review the data with Dirk van der Marel, a condensed-matter physicist at the University of Geneva. According to Hirsch, “something was very wrong” with the numbers.
When a signal is measured, it contains an element of background noise, thanks to ambient influences like electrical interference. Independently measured signals will have independent noise — and subtracting one signal from another should result in a cleaner overall signal with at least as much noise as the original sources.
Dias’ data, however, seemed to show the opposite.
In the CSH paper, Dias and his colleagues claimed to have made two independent voltage measurements — the measured, or raw signal, from the superconducting CSH sample, and a background signal from a non-superconducting CSH sample.
With this in mind, the math is fairly simple: subtracting the background from the raw signal should produce a somewhat noisy end result. But Dias and Salamat’s results yielded the opposite, and showed comparatively lower noise instead.
According to Hirsch, after confronting Dias and Salamat with this concern, the two then claimed to have constructed the data via means that they refused to disclose, even when Hirsch asked for further clarification.
“It was a nonstandard method, and we hadn’t disclosed it. So that was the reason to retract,” Dias told the APS. “[Nature] hasn’t questioned the validity of our data […] the data is valid.”
Concerning the alleged plagiarism
Beyond the accusations of data fabrication and manipulation, Dias’ data is also haunted by allegations of plagiarism.
While Hamlin was looking into the CSH data, he was struck by some lines in Dias’ write-ups that read as familiar — lines that he himself had written in his 2007 PhD thesis.
Hamlin and Kimber would go through Dias’ thesis by hand, after their initial introduction. According to an analysis shared with Science, they found Dias’ thesis to contain “6300 words that are identical to passages from 17 sources” — or about 21% of the thesis overall.
An independent comparison by Physics Magazine found “dozens of paragraphs that match word for word” and two figures with remarkable similarities. Reporting by Undark, Science, and The New York Times previously referenced the degree of similarities between the two theses, though Dias claims to have appropriate citations.
Dias’ University website also contains text copied without attribution, according to Hamlin and Kimber.
“The scientific community is going to need answers on this,” said Hamlin — answers, some feel, that lie in a precedent set decades ago.
Peter Armitage, a condensed-matter physicist at Johns Hopkins University, made reference to the infamous Schön scandal of 2002.
German physicist Jan Hendrik Schön, who made a series of apparent superconductor breakthroughs, was met with worldwide acclaim. His work, however, was found to be vastly fraudulent after physicists noted that the noise patterns within one of his graphs was eerily reminiscent of another.
Schön was found to have manipulated or outright fabricated data in at least 17 papers during his time. Despite a number of attempts, his work could not be reproduced.
In regard to Dias’ research, Armitage found a clear similarity — especially with Dias’ latest 2023 Nature paper.
“I’ve talked to other people who came to the same conclusion — you cannot produce the data in the new paper by the analysis procedure they’ve claimed they perform,” he said. “It just doesn’t come out. There’s no way to get it.”
Similarly concerning to him is the lack of action over the allegations against Dias.
The Schön case was met with swift action by Schön’s employer, Bell Labs — including a committee investigation and a public release of the inquiry report. On top of several retractions, Schön was later stripped of his doctoral degree.
Washington State University has declined to comment on whether they have carried out an investigation into the validity of Dias’ thesis, according to the APS.
“I would point out that there’s a precedent of people losing PhDs for plagiarizing far less of their thesis,” Armitage said. “If someone was plagiarizing 21% of their [reporting], they wouldn’t have a job, and there’s no difference in science.”
Dias’ retraction also comes on the heels of Stanford President Marc Tessier-Lavigne’s resignation, as first publicized by The Stanford Daily. Following a review of allegations first proposed by The Daily last November, Tessier-Lavigne will see the retraction of five widely cited papers due to “unusual frequency of manipulation of research data and/or substandard scientific practices,” according to the Stanford-sponsored investigation.
University of Rochester’s response
The University, comparatively, has conducted two internal inquiries into the CSH data-manipulation allegations, according to Dean of the Hajim School of Engineering and Applied Sciences Wendi Heinzelman.
“In response to concerns related to Professor Dias’ 2020 Nature paper, the University conducted two inquiries into the issues raised, which were completed in May 2022,” she said. “In both inquiries it was determined that there was no evidence that supported the concerns.”
A third inquiry would be launched into Dias’ 2020 Nature paper following its retraction, and similarly found no evidence of misconduct.
Although the details of the investigations are not publicly available, Heinzelman said that both were conducted according to University policy. The lack of formal University response, however, has its own fair share of critics.
“The normal practice in academia is that you tend not to share referee communications, things with editors and so on,” said Kimber. “In the circumstances, I don’t give a shit. Every time I’ve heard following due process, privacy, confidentiality over the past six months — it’s just been used as an excuse to stop things from progressing.”
“As my mother would always say, sunlight is the best disinfectant,” he added. “I think, in such extraordinary circumstances, that it’s completely reasonable for information to become public.”
Controversy also surrounds Dias’ latest 2023 Nature paper, which claims to have discovered another room-temperature superconductor in nitrogen-doped lutetium hydride.
Heinzelman said in correspondence with the Campus Times that Dias is “actively participating in those ongoing [investigation] efforts within the scientific community,” and efforts to reproduce his results are ongoing.
She made reference to a New York Times article by Kenneth Chang, who had previously covered Dias’ at-the-time breakthrough research. The article, she said, “reported on a University of Chicago research team verifying a key measurement from [Dias’] study.”
According to Chang’s article, the University of Chicago’s team did verify the “apparent vanishing of electrical resistance,” but “this result does not prove that the material is a room-temperature superconductor.”
Other behavior from Dias has also piqued public interest — such as his startup company, Unearthly Materials, which bills itself as “powering the century of superconductivity,” with “a record of revolutionary breakthroughs.”
Digital science magazine Quanta discusses Unearthly Materials’ history of fundraising, including Dias’ claim of raising over $20 million from a number of high-profile investors, such as Sam Altman, CEO of OpenAI, and Daniel Ek, co-founder and CEO of Spotify.
Following the Quanta story, Dias and his team reportedly rescinded the claims — saying that the statements were only “aspirational,” and that the names listed were purely prospective investors.
Dias’ claim, initially made during a 2021 talk organized by the Sri Lanka Association for the Advancement of Science, was later posted to YouTube. The video, however, is no longer available.
However, these claims, aspirational or not, are currently supported with private and governmental backing.
Two years ago, he was also the recipient of a $1.6 million grant from the Gordon and Betty Moore Foundation, to support his “groundbreaking efforts,” according to the University.
In addition, Dias’s lab has funding support from a National Science Foundation CAREER award and a grant from the US Department of Energy, specifically in pursuit of room-temperature superconductors. His current CAREER grant intends to award his lab a total of over $794,000, which is slated to pay out until 2026.
Amidst these controversies, the University has featured Dias’ lab in promotional material concerning campus research.
His most recent superconductive effort, the nitrogen-doped lutetium hydride of his 2023 paper, was featured in a University video alongside guest speakers such as Heinzelman. The video paints Dias as someone set to open the floodgates of industry and education in Rochester, thanks to his work at the University.
“We envision the clear possibility of making Rochester a center for everything that is related to superconductivity,” said Renato Perucchio, Chair of the University’s Mechanical Engineering department.
In the video, Dias humorously calls the alleged superconductor “reddmatter,” after a Star Trek material of the same name — but to him, the discovery seems to be anything but a joking matter.
“These kinds of huge claims need a careful, thorough examination,” he said. “So, because of that, [we] went through a rigorous review process. Multiple referees, multiple rounds, and we made all the data and everything that we discussed in the paper, our methods and everything, very open, so everybody can take a look at it and go through it and really make sure that they come to the same conclusion.”
The Campus Times has repeatedly reached out to Dias for comment regarding the allegations, but has yet to receive a response.
Editor’s Note (8/1/23): Wording has changed slightly in the second and third paragraphs to more clearly explain the details of Dias’ 2021 paper.
