Integrity Review Report: 'Vapour–liquid–solid–solid growth of two-dimensional non-layered β-Bi2O3 crystals with high hole mobility' (Nature Materials, 2025)

Academic fraud report · Geng Detector

Summary

This report reviews a 2025 Nature Materials paper reporting vapour–liquid–solid–solid growth of 2D β-Bi2O3 crystals (DOI: 10.1038/s41563-025-02141-w). The verdict is "highly suspicious" (not proven misconduct). Three textual/numerical inconsistencies were flagged. (1) Raman peak positions for the same vibrational modes disagree between two sections: 102/129/466 cm⁻¹ vs. 97/125/464 cm⁻¹, with no thickness-related explanation given. (2) EDS weight fractions for initial droplets (40% O, 35% Bi, 15% Cl, 5% Na) sum to 95% rather than 100%, whereas crystal-edge values sum to 100%. (3) The 2.5-year review period (submitted 2 July 2022, accepted 15 January 2025) and the listing of a 2024 National Key R&D grant (2024YFB3612400) raise timeline questions, though late-stage grant attachment is plausible. No image-level analysis was possible from text. Confidence is moderate; findings are confined to verifiable text/numbers and warrant author response and editorial attention.

Verdict

🟠 Highly suspicious — multiple internal inconsistencies in reported numerical data, but no image-based or fabricated-data finding has been independently verified.

Key findings

Inconsistent Raman peak assignments for the same material between two sections of the manuscript.
EDS weight fractions do not sum to 100% for the droplet composition, while a later composition (crystal edge) does sum correctly.
Abnormally long review timeline (≈2.5 years from submission to acceptance) and temporal mismatch between the paper's submission date and a 2024 funding grant acknowledged in the paper.
Image-level analysis not performed; original micrographs/spectra were not available for forensic review.

Evidence highlights

Raman peak contradiction — Page 689: "three Raman peaks at 102 cm⁻¹, 129 cm⁻¹ and 466 cm⁻¹." Page 693: "four active Raman modes: A₁(1) mode at 97 cm⁻¹, B₁(1) mode at 125 cm⁻¹, B₂(2) mode at 208 cm⁻¹ and A₁(3) mode at 464 cm⁻¹." The A₁(1) mode shifts from 102 → 97 cm⁻¹; B₁(1) from 129 → 125 cm⁻¹; A₁(3) from 466 → 464 cm⁻¹. No explanation for thickness-dependent shift is provided.
EDS arithmetic anomaly — Reported droplet composition: "40% O, 35% Bi, 15% Cl and 5% Na" → 40 + 35 + 15 + 5 = 95%, leaving 5% unaccounted. Crystal-edge composition reported as "60% O, 39.5% Bi, 0.25% Cl and 0.25% Na" sums to 100%.
Submission/acceptance dates — Submitted 2 July 2022, accepted 15 January 2025 (review period ≈ 2.5 years).
Funding — Acknowledged grant: National Key R&D Program of China, no. 2024YFB3612400 (project year 2024), attached to a paper submitted in 2022.

Notes

All quantitative figures above are reproduced verbatim from the report; no values were inferred or invented.
Image-level checks (Western blot duplication, HAADF-STEM manipulation, cloning in Figure 1 and Figure 3) could not be performed because only extracted text was supplied; reviewers with access to the original figures should conduct an independent forensic pass.
Possible innocent explanations have not been ruled out: e.g., the Raman discrepancy could reflect genuinely different sample thicknesses (though no such discussion is provided); the 5% mass shortfall in EDS could reflect undeclared light elements (e.g., H) or rounding — but this is not stated in the text, which makes the omission itself a concern; the 2024 grant may have been added during a late revision, but this typically requires editorial disclosure.
Recommended next steps: (i) request raw Raman spectra and EDS spectra/dat from the authors; (ii) post structured queries on PubPeer regarding the Raman shift and the missing 5%; (iii) verify the official start date of grant 2024YFB3612400 against the experimental timeline; (iv) consider editorial notification only after the authors have had an opportunity to respond.