Integrity Analysis Report: 'Vapour–liquid–solid–solid growth of two-dimensional non-layered β-Bi2O3 crystals with high hole mobility' (DOI: 10.1038/s41563-025-02141-w)

Academic fraud report · Geng Detector

Summary

This report analyzes the Supplementary Information of a 2025 Nature Materials paper on two-dimensional β-Bi2O3 crystals. The overall verdict is Highly Suspicious, based on multiple red flags for citation malpractice and potentially fabricated quantitative data. Key findings include: (1) Reference 33, cited to support a 562 nm PL peak attributed to layered BiOCl, actually pertains to nickel-oxide thin-film transistors, indicating citation misattribution; (2) Reference 66 cites 'Nat. Commun. 21, 4309 (2020)', but Nature Communications Volume 21 does not correspond to 2020 (the correct volume is 11), suggesting a fabricated or hallucinated citation; (3) EDS elemental composition reported as exactly 60% O, 39.5% Bi, 0.25% Cl, 0.25% Na (summing to 100.0%) is implausibly precise for semi-quantitative EDX of trace elements; (4) Statistical wording around n=10 temperature-dependent measurements is ambiguous regarding device-vs-technical replication. Image and blot inspection was not possible from text-only input. Confidence is moderate-to-high for citation errors, low-to-moderate for data fabrication claims.

Verdict

Highly Suspicious (🟠). Multiple strong indicators of citation malpractice were identified from text alone. Quantitative data anomalies raise additional concerns but cannot be confirmed without raw spectra. Pixel-level image analysis was not performed due to input limitations.

Key findings

Citation misattribution (Finding 1, 🔴): In Supplementary Note 2, the claim that a 562 nm PL peak confirms layered BiOCl cites Reference 33, which is actually "Xu, W. et al. p-type transparent amorphous oxide thin-film transistors using low-temperature solution-processed nickel oxide. J. Alloy. Compd. 806, 40–51 (2019)." The cited paper concerns NiO TFTs and is wholly unrelated to BiOCl photoluminescence or VLS–VSS growth mechanisms.
Implausible/incorrect citation (Finding 2, 🔴): Reference 66 lists "Nat. Commun. 21, 4309 (2020)." Nature Communications Volume 21 was not the 2020 volume (article 4309 in 2020 belongs to Volume 11). The mismatched volume is consistent with a fabricated or AI-hallucinated reference.
Implausibly perfect EDS composition (Finding 3, 🟠): Reported crystal-edge composition of 60% O, 39.5% Bi, 0.25% Cl, 0.25% Na sums to exactly 100.0%. Semi-quantitative EDX rarely yields such exact integer totals, particularly for trace species (Cl, Na at 0.25%). The stoichiometric alignment with Bi2O3 suggests possible manual calculation rather than instrumental measurement.
Ambiguous statistical reporting (Finding 4, 🟡): Supplementary Fig. 42 description states "n=10 for both 2.6-nm-thick and 4.8-nm-thick β-Bi2O3 FET devices…ten independent measurements…for each FET device." The phrasing conflates device-level and technical replicates, a known pattern for inflating sample sizes when device yield is low.
Incomplete checks: Image duplication, Western blot reuse, and last-digit distribution tests could not be performed because only extracted text from the Supplementary Information PDF was supplied.

Evidence highlights

DOI under review: 10.1038/s41563-025-02141-w
Ref 33 (line ~146 of SI): cited for BiOCl PL peak → actually NiO TFT paper, J. Alloys Compd. 806, 40–51 (2019).
Ref 66 (line ~906 of SI): "Nat. Commun. 21, 4309 (2020)" — Volume 21 is inconsistent with the 2020 publication year for Nature Communications.
SI Note 2 (line ~147–148): EDS composition 60% O / 39.5% Bi / 0.25% Cl / 0.25% Na, summing to exactly 100.0%.
SI Fig. 42 caption (line ~689–696): n=10 designation ambiguous between per-device repeats and per-device independent measurements.

Notes

Limitations: analysis is restricted to the Supplementary Information text; the main manuscript, original EDS spectra, raw transport data, and figure images were not available for inspection.
Recommended follow-up: obtain the main manuscript and high-resolution figures; verify References 33 and 66 directly in the publisher's reference list; request raw EDS spectra and original temperature-dependent transport data from the authors; consider posting the citation concerns on PubPeer and notifying the Nature Materials editorial office.
The findings above are indicators, not a final determination of misconduct; authoritative conclusions require institutional investigation.