Investigative 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 flags several anomalies in the Nature Materials paper (DOI: 10.1038/s41563-025-02141-w) describing a vapour–liquid–solid–solid growth route for 2D non-layered β-Bi₂O₃ crystals. The verdict is 'highly suspicious' (not confirmed fraud). The most notable issue is a numerical coincidence: the abstract's claimed maximum hole mobility of 136.6 cm² V⁻¹ s⁻¹ equals the reported mean plus one standard deviation (130.4 + 6.2 cm² V⁻¹ s⁻¹) for 2.4 nm devices, suggesting the headline value was derived from summary statistics rather than a genuine measurement peak. A second arithmetic inconsistency appears in the droplet composition (40% O, 35% Bi, 15% Cl, 5% Na; total 95%) reported by XPS/EDS, whereas the crystal-edge composition sums correctly to 100%. Minor concerns include OCR-grade garble in the abstract ('136.6θcm2 V−1 s−1', 'on/oϕ'), an unusually long review timeline of ~30 months (Received 2 July 2022; Accepted 15 January 2025), and a unit-mismatched sentence in the Methods. No pixel-level image duplication has yet been verified. Confidence is moderate; resolution requires the authors' raw mobility traces and EDS/XPS spectra.

Verdict

Highly suspicious. Multiple internal numerical inconsistencies and contextual anomalies are documented, but none constitutes definitive proof of misconduct. Independent verification of raw data is required.

Key findings

Statistical coincidence in mobility values: maximum reported (136.6 cm² V⁻¹ s⁻¹) equals the 2.4 nm-device mean plus one standard deviation (130.4 + 6.2 cm² V⁻¹ s⁻¹). This is the classic fingerprint of a fabricated or back-calculated headline number.
Compositional sum discrepancy: droplet composition given as 40% O, 35% Bi, 15% Cl, 5% Na (total = 95%), while the crystal-edge composition (60% O, 39.5% Bi, 0.25% Cl, 0.25% Na) correctly sums to 100%.
Abstract text corruption: garbled tokens such as 136.6θcm2 V−1 s−1 and on/oϕ current ratio appear in the abstract; the underlying PDF/HTML source should be inspected.
Abnormal review timeline: 2 July 2022 submission to 15 January 2025 acceptance (~30 months), atypical for Nature Materials.
Methods-section wording error: 'The heating and deposition rates were 20 °C min⁻¹ and 30 min, respectively' conflates a rate (°C min⁻¹) with a duration (min).
Image-integrity check pending: HAADF-STEM panels (Fig. 1d/e, Fig. 2g) flagged for potential splicing; pixel-level analysis not yet performed.

Evidence highlights

Maximum hole mobility (Abstract): 136.6 cm² V⁻¹ s⁻¹.
Mean ± SD for 2.4 nm device (p. 694): 130.4 ± 6.2 cm² V⁻¹ s⁻¹; 130.4 + 6.2 = 136.6 (exact).
Droplet composition (p. 689 / Fig. 2 caption): O 40%, Bi 35%, Cl 15%, Na 5% → 95%.
Crystal-edge composition: O 60%, Bi 39.5%, Cl 0.25%, Na 0.25% → 100%.
Dates: Received 2 July 2022; Accepted 15 January 2025 (≈30 months).
DOI: 10.1038/s41563-025-02141-w.

Notes

Findings 1 and 2 are the strongest signals; both are verifiable by arithmetic alone and do not depend on image analysis.
Finding 3 may be a PDF/OCR artefact rather than author error, but the abstract should be inspected in the published HTML version.
Finding 5 (long review period) is circumstantial, not evidence of misconduct.
The report recommends requesting the authors' raw field-effect transistor transfer curves, the underlying mobility-extraction files, and unprocessed EDS/XPS spectra before escalating to PubPeer or the editorial office.
Affiliations named in the source material include Nanjing University of Science and Technology and Nanjing University; institutional referral is listed only as a contingent step.