Science Journal Vindicates Microsoft Quantum Paper, Lifts “Expression of Concern” After 4-Year Dispute
The journal Science is preparing to lift its editorial expression of concern (EEoC) issued for one of the most high-profile papers in recent years in the field of quantum computing. The article, published in 2020, described an experiment aimed at creating topological superconductors—materials capable of hosting error-resistant Majorana particle states, seen as a promising foundation for scalable quantum computing.
At the heart of the study were nanowires fully coated in superconducting material. The authors asserted that these structures could induce a topological superconducting phase—a distinct state of matter where Majorana quasiparticles emerge. These exotic entities, neither standard fermions nor bosons, blend the properties of matter and antimatter and are believed to possess extraordinary resistance to external noise. Microsoft’s quantum chip architecture, dubbed Majorana 1, was based on this concept as a potential alternative to traditional error-corrected qubit systems.
However, in 2021, Science appended an EEoC to the paper following criticism from experimental physicists Sergey Frolov (University of Pittsburgh) and Vincent Mourik (Forschungszentrum Jülich). They raised concerns that the published tunneling spectroscopy graphs—a technique for probing energy states in superconductors—had been selectively chosen and failed to represent the full experimental landscape.
The forthcoming revision will lift the EEoC and replace it with a formal correction: the editors acknowledge that while the paper contains no errors in calculation or interpretation, it did lack transparency in equipment calibration and omitted a complete dataset. From a formal standpoint, the publication remains valid.
For Charles Marcus, a professor of physics and former director of the Microsoft Quantum Lab in Copenhagen, the decision serves as a form of vindication. The University of Copenhagen’s Research Integrity Committee likewise found no misconduct. Marcus stated that both Science and Denmark’s National Science Council affirmed the propriety of his actions—though, he added, the four years spent defending his reputation cannot be reclaimed.
Frolov and his colleagues remain unconvinced. They argue that the correction does not resolve the core issue of data cherry-picking, casting doubt on the reliability of the conclusions. Frolov noted that this is the first instance of Science revoking an EEoC without retracting the paper, despite lingering questions over its transparency.
The controversy surrounding this publication is part of a broader conflict between Microsoft and independent physicists researching Majorana particles. Over the past five years, Frolov has successfully pushed for the retraction of two Microsoft-affiliated articles in Nature and the imposition of EEoCs on several others, including one in Science.
In 2019, at the request of an editor, Frolov reviewed theoretical and experimental manuscripts submitted by Microsoft, Copenhagen, UC Santa Barbara, and Yale. According to him, both were rejected, but later merged and accepted as a single paper. Microsoft even issued a press release proclaiming a “scientific breakthrough.”
When the authors refused to provide the full dataset, Frolov appealed to the editorial board and was eventually sent 50 MB of experimental data. He estimated this represented a single day’s work, while a full analysis would have required several weeks. In his view, the data failed to substantiate the persistent presence of Majorana signals. He filed a formal complaint, which became the basis for the initial EEoC.
Further support for his claims came from an Austrian research group working with the same Microsoft chips. They shared 600 MB of their own measurements with the critics and also found no evidence of Majorana particles. These findings were published in both Nature and Science, marking a rare consensus between two leading journals in discrediting a previous discovery.
Frolov accuses Microsoft of consistently highlighting the most striking segments of data while disregarding the rest. This, he argues, is why several papers were ultimately retracted. His perspective is echoed by theorist Henry Legg (University of St Andrews), and criticism has previously come from Simone Severini, head of quantum technologies at Amazon.
Marcus counters that Legg is not an experimentalist and underestimates the complexity of equipment calibration. He points to 15 adjustable parameters required to establish conditions favorable for Majorana states. Moreover, he believes Frolov and Mourik’s campaign is personal—characterized by repeated data requests, conference pressure, appeals to funding agencies, and messages to students. Following the EEoC, Marcus observes, interest in Majorana research in academia has largely waned.
His stance is supported by Michael Manfra (Purdue University, Microsoft Quantum Lab), who stated, “There was no fabrication in the paper,” and affirmed that Science’s decision to lift the warning was justified.
Despite the sharpness of the dispute, the scientific core remains paramount. It centers on an exceptionally complex and still poorly understood class of quasiparticles—states that, in theory, could underpin robust and scalable quantum computers. But experimental verification is fraught with difficulty—the signals are faint, unstable, and sensitive to numerous variables. This makes the entire domain vulnerable to error, overinterpretation, and “positive selection” of data.
Microsoft, for its part, continues its research and emphasizes that open publication, debate, and doubt are vital components of the scientific process. Whether the company will ultimately succeed in transforming these theoretical constructs into practical technology, however, remains an open question.
