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

Academic fraud report · Geng Detector

Summary

Highly suspicious. Five issues were raised based on textual and logical analysis; image-pixel forensics could not be performed. (1) Funding–timeline inconsistency: the manuscript was received 2 July 2022 yet acknowledges a National Key R&D Program of China grant no. 2024YFB3612400, which by convention was launched in 2024 — the funding cannot legitimately have supported work submitted ~2 years earlier. (2) The authors claim a 0.28–0.3 nm 'half-unit-cell' thickness for a non-layered β-Bi₂O₃ (c = 5.634 Å = 0.56 nm); non-layered oxides cannot be cleaved to half a unit cell without catastrophic dangling bonds, suggesting mismeasurement, surface reconstruction, or over-interpretation. (3) Reported FET mobilities show anomalously low dispersion — e.g., 130.4 ± 6.2 cm² V⁻¹ s⁻¹ (CV ≈ 4.7%) and 60.4 ± 4.8 cm² V⁻¹ s⁻¹ — far narrower than typical 15–30%+ spread for CVD-grown non-layered 2D devices, consistent with data cherry-picking. (4) EBL patterning is attributed to a TESCAN VEGA 3 (a basic W-filament SEM), an unusual choice for sub-μm high-performance FETs. (5) STEM/TEM images (Fig. 1d–f,i; Fig. 2g) could not be pixel-analyzed; flagged for follow-up. Image-level duplication is unverified.

Verdict

Highly suspicious. Multiple independent concerns — including a clear funding–submission date contradiction and physically implausible half-unit-cell 2D oxide — warrant formal editorial and institutional investigation. Image-pixel duplication has not been confirmed due to lack of access to original high-resolution files.

Key findings

Funding–timeline contradiction (red flag): Received 2 July 2022; Acknowledgements list National Key R&D Program of China grant no. 2024YFB3612400, a 2024-launched program. A 2024 grant cannot have supported a manuscript submitted in mid-2022; this is a hard inconsistency in funding attribution.
Implausible 'half-unit-cell' 2D oxide: Non-layered β-Bi₂O₃ is 3D covalently/ionically bonded (c = 5.634 Å, i.e. 0.56 nm). A stable 0.28–0.3 nm (half-unit-cell) flake violates surface-energy physics; the claim likely reflects AFM substrate artifact, surface oxidation/reconstruction, or over-interpretation of a VESTA-rendered model as experimental evidence.
Implausibly tight device statistics: 2.4 nm devices average μ = 130.4 ± 6.2 cm² V⁻¹ s⁻¹ (CV ≈ 4.7%); 1.0 nm devices average μ = 60.4 ± 4.8 cm² V⁻¹ s⁻¹. Such narrow dispersions are atypical for CVD non-layered 2D FETs and are consistent with selective data curation.
EBL equipment mismatch: EBL is described as performed on a TESCAN VEGA 3 SEM — a basic W-filament instrument generally used for imaging, not standard for sub-μm high-mobility FET fabrication; omission of a dedicated EBL system (e.g., Raith, JEOL JBX) is suspicious.
Image integrity unverified: TEM/STEM panels (Fig. 1d–f,i; Fig. 2g) were not pixel-analyzed; periodic copy–paste artifacts in lattice fringes and abrupt BiOCl→β-Bi₂O₃ phase interfaces are flagged for visual review.

Evidence highlights

DOI: 10.1038/s41563-025-02141-w
Journal/Year: Nature Materials, vol. 24, May 2025
Received: 2 July 2022; Accepted/Published: 2025
Grant listed: National Key R&D Program of China, no. 2024YFB3612400
Crystal data: c = 5.634 Å (0.56 nm); reported 2D thickness 0.28–0.3 nm
FET mobilities: 130.4 ± 6.2 cm² V⁻¹ s⁻¹ (2.4 nm); 60.4 ± 4.8 cm² V⁻¹ s⁻¹ (1.0 nm)
Reported highest mobility: 136.6 cm² V⁻¹ s⁻¹
Lithography tool stated in Methods: TESCAN VEGA 3 SEM

Notes

Confidence: medium-high for textual/logical issues; low for image duplication (not yet examined at pixel level).
Limitations: review performed from PDF text and figures only; no raw I–V curves, AFM height maps, or original STEM/TEM image stacks were inspected.
Recommended actions: (i) request raw I–V and transfer curves plus per-device statistics from the authors; (ii) flag the 2024YFB3612400 funding–date contradiction to the Nature Materials editorial office; (iii) open a PubPeer thread on the half-unit-cell thickness claim and the unusually small standard deviations; (iv) conduct pixel-level inspection of Fig. 1d–f,i and Fig. 2g once high-resolution originals are available; (v) escalate to the authors' institutional research integrity committee if raw-data review confirms cherry-picking.
The funders had no role in this review.