Integrity Concerns in "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 a Nature Materials paper (DOI: 10.1038/s41563-025-02141-w) by Y. Xiong et al. on the VLSS growth of 2D β-Bi2O3 as highly suspicious. The central mechanism rests on two chemical equations that fail basic mass balance: Eq. (1) loses oxygen atoms and Eq. (2) is unbalanced (likely meant 3BiOCl → Bi2O3 + BiCl3). EDS droplet composition is reported as 40% O, 35% Bi, 15% Cl, 5% Na, summing to only 95%, with suspiciously round integers atypical of real micro-EDS data. FET mobilities (42.2 ± 4.0, 60.4 ± 4.8, 130.4 ± 6.2 cm²/V·s) show anomalously uniform fractional standard deviations for n=5. The TESCAN VEGA 3 SEM is a basic tungsten-filament instrument not normally used for EBL fabrication of nanodevices. The review cycle was ~2.5 years (July 2022 – January 2025). No image-level analysis was performed. Verdict: highly suspicious, not definitive proof of fraud; the unbalanced equations and 95% sum are reproducible textual anomalies requiring author response.

Verdict

Highly suspicious. Multiple reproducible textual and arithmetic anomalies undermine the credibility of the central VLSS growth mechanism and supporting quantitative data. Marked as concerning, not as confirmed misconduct, pending author response and image-level verification.

Key findings

Unbalanced core chemical equations (Eqs. 1 and 2, p. 692). Eq. (1), Bi2O3(s) + NaCl(s) → NaxBiOy(l) + BiOxCly(l), cannot satisfy oxygen mass balance: the LHS has 3 O, while plausible products (NaBiO + BiOCl) provide only 2 O. Eq. (2), BiOCl(s) → Bi2O3(s) + BiCl3(g), is not stoichiometrically balanced; the intended form is most likely 3BiOCl(s) → Bi2O3(s) + BiCl3(g). The proposed VLSS mechanism is chemically incoherent as written.
EDS microanalysis of droplet sums to 95% (p. 690). Reported composition is 40% O, 35% Bi, 15% Cl, 5% Na, totaling 95%. The residual 5% is unaccounted for, and the values are unusually round integers (multiples of 5) for spot-mode EDS on a microscopic droplet.
Suspiciously uniform FET statistics (p. 694, Fig. 4d). Mobility values 42.2 ± 4.0, 60.4 ± 4.8, 130.4 ± 6.2 cm²/V·s show fractional SDs of ~9.5%, ~7.9%, and ~4.8% respectively, with n=5. Inter-device variability is implausibly narrow for nanodevices sensitive to contact resistance and interface traps.
Equipment-method mismatch. Methods state that FETs were fabricated via electron-beam lithography on a TESCAN VEGA 3 SEM, a basic tungsten-filament instrument typically used for imaging, not for nanometer-precision EBL; no external beam-blanking/pattern-generator system is mentioned.
Extended review timeline. Received 2 July 2022; accepted 15 January 2025 — a ~2.5-year review cycle, atypical for Nature Materials and suggestive of significant post-review revision or dispute.

Evidence highlights

DOI: 10.1038/s41563-025-02141-w
Eq. (1) text: Bi2O3(s) + NaCl(s) → NaxBiOy(l) + BiOxCly(l) (oxygen not conserved)
Eq. (2) text: BiOCl(s) → Bi2O3(s) + BiCl3(g) (Bi, O, Cl not conserved; correct form would be 3BiOCl(s) → Bi2O3(s) + BiCl3(g))
EDS droplet composition: 40 + 35 + 15 + 5 = 95%
Mobility statistics: 42.2 ± 4.0 (n=5), 60.4 ± 4.8 (n=5), 130.4 ± 6.2 (n=5)
Instrumentation: TESCAN VEGA 3 SEM cited for EBL
Timeline: Received 2 July 2022; Accepted 15 January 2025

Notes

No image-level analysis (pixel splicing, noise-pattern, or filter-overprocessing checks) was performed on Figures 1–4; the report is based solely on text, equations, and numerical data in the PDF.
The unbalanced equations may reflect editorial/typesetting errors rather than fabrication, but they invalidate the mechanistic claim as written and require author clarification or correction.
The 95% EDS sum and integer-like values are anomalies, not proof of fabrication; rounding conventions and unreported elements (e.g., adventitious C) should be clarified by the authors.
The 2.5-year review window is not itself evidence of misconduct but is consistent with the other flagged issues.
A formal PubPeer post or direct correspondence asking the authors to (i) provide balanced equations, (ii) explain the EDS sum, and (iii) clarify the EBL toolchain is the proportionate next step.