Integrity Review: High-mobility p-type semiconducting two-dimensional β-TeO2 (Nature Electronics, 2021)

Academic fraud report · Geng Detector

Summary

This review applies a multi-pass analytical framework to assess the paper 'High-mobility p-type semiconducting two-dimensional β-TeO2' (DOI: 10.1038/s41928-021-00561-5, Nature Electronics, 2021) by Zavabeti et al. Verdict: no systematic indicators of data fabrication or academic misconduct were detected. Key supporting observations include (1) physically plausible device-to-device variance (field-effect mobility 146 ± 42 cm² V⁻¹ s⁻¹, ~29% coefficient of variation across 10 devices; subthreshold swing 103 ± 3 mV per decade), (2) curve-fit R² values of 0.91 (EELS bandgap) and 0.93 (temperature-dependent mobility) consistent with realistic noise rather than artificially perfect fits, and (3) internally consistent instrument/software timeline (Thermo Scientific K-alpha XPS; JEOL 2100F TEM acquired 2011; FEI Nova NanoSEM 2007/2014; MATLAB R2019b; Digital Micrograph 1.8.4). Limitations: pixel-level image-duplication analysis was not possible from the provided text. Overall, the textual evidence is consistent with genuine experimental data.

Verdict

No systematic indicators of data fabrication or academic misconduct were identified in the textual content of the paper. The reported statistics, fit quality, and methodological timeline are internally consistent and consistent with authentic experimental data.

Key findings

Physically realistic device-to-device variance: field-effect mobility of 146 ± 42 cm² V⁻¹ s⁻¹ across 10 devices (coefficient of variation ≈ 29%), and subthreshold swing of 103 ± 3 mV per decade. Such spread is typical of nanomaterial device fabrication and contact-resistance variation, not of synthetic data.
Honest curve-fit quality: reported R² values of 0.91 (EELS bandgap fit, Figure 3a) and 0.93 (temperature-dependent mobility fit, Figure 4e). Imperfect fits are inconsistent with overly curated or fabricated data.
Internally consistent instrument and software timeline: XPS (Thermo Scientific K-alpha), TEM (JEOL 2100F, acquired 2011), SEM/EDX (FEI Nova NanoSEM, 2007/2014), MATLAB R2019b, and Digital Micrograph 1.8.4 — all consistent with a 2021 publication.
Descriptive consistency: AFM-measured thickness (~1.5 nm) attributed to a bilayer structure is consistent with the crystallographic model; no internal contradictions in the text.

Evidence highlights

Reported mobility error bar: $146 \pm 42\ \text{cm}^2\ \text{V}^{-1}\ \text{s}^{-1}$ (n = 10 devices).
Reported subthreshold swing: $103 \pm 3\ \text{mV per decade}$.
Reported fit coefficients of determination: $R^2 = 0.91$ (EELS bandgap) and $R^2 = 0.93$ (variable-temperature mobility).
DOI: 10.1038/s41928-021-00561-5.
Instrument acquisition/usage timeline: TEM 2011; SEM 2007/2014; software versions (MATLAB R2019b, Digital Micrograph 1.8.4) precede the 2021 publication date.

Notes

Scope: this assessment is limited to textual and statistical indicators in the provided content; pixel-level image duplication, splicing, and Western blot forensics were not feasible without access to the original figure files.
The coefficient-of-variation magnitude (~29%) for mobility is taken as an authenticity-supportive signal rather than a quantitative fraud-detection criterion.
A clean result from this review does not constitute formal clearance; authoritative determination of misconduct requires institutional investigation and access to raw data and original images.