Geng Report: Concerns over 'High Resistivity and Low Defect Covalent Organic Frameworks for Highly Stable and Low Dose X-Ray Detection' (Angew. Chem. Int. Ed., 2026, DOI: 10.1002/anie.4087507)

Academic fraud report · Geng Detector

Summary

This Chinese-language Geng fraud-detection report raises serious concerns about the 2026 Angewandte Chemie paper by Ji et al. on covalent organic frameworks (COF366-Cu, COF366-Co) for X-ray detection (DOI: 10.1002/anie.4087507). The headline finding is that the paper's central performance metric—the limit of detection (LoD)—appears with contradictory values across the Abstract (39 nGy·s⁻¹), Introduction (36 nGy·s⁻¹), Results section (65 and 39 nGy·s⁻¹, with swapped logic), and Conclusion (39 and 65 nGy·s⁻¹, again swapped). The report also flags an implausibly low trap density (~5 × 10⁸ cm⁻³) extracted by SCLC from powder-pressed pellets—values reportedly lower than those of state-of-the-art perovskite single crystals—and unit inconsistencies in current drift (cm⁻¹ vs cm⁻²) between Abstract and main text. Overall verdict: highly suspicious, with data-inconsistency concerns. Limits: pixel-level image analysis was not possible from text alone; no original I–V or raw data were inspected. Confidence is high on textual contradictions, lower on image and SCLC-fraud claims.

Verdict

Highly suspicious. Multiple internal contradictions in the paper's most important numerical claims—specifically the limit of detection (LoD) and current-drift units—strongly suggest the manuscript was assembled from incompatible drafts or contains fabricated data. The physically implausible SCLC-derived trap density adds to these concerns. Independent verification of raw data and curves is recommended before any further reliance on the reported performance.

Key findings

Inconsistent LoD values across the manuscript (Abstract, Introduction, Section 2.4, Conclusion), with at least three different numbers (36, 39, 65 nGy·s⁻¹) and internally contradictory ordering of Cu vs. Co devices.
Implausibly low trap density (~5.2–5.8 × 10⁸ cm⁻³) reported from SCLC measurements on powder-pressed wafers, claimed to be lower than values routinely achieved by high-quality perovskite single crystals (~10⁹ cm⁻³).
Unit inconsistency in dark current drift between Abstract (~10⁻¹⁸ A·cm⁻¹·V⁻¹·s⁻¹) and Section 2.3 (1.05 × 10⁻¹⁸ and 3.12 × 10⁻¹⁸ A·cm⁻²·V⁻¹·s⁻¹).
Logically self-contradictory performance claims: Section 2.4 states COF366-Cu has LoD = 65 nGy·s⁻¹ and outperforms COF366-Co (LoD = 39 nGy·s⁻¹), but lower LoD normally means better sensitivity, so the stated ranking is inverted.
Pixel-level image forensics not performed; reviewer recommends verifying J–E, SCLC, and detector-response curves against raw data.

Evidence highlights

LoD in Abstract: 39 nGy·s⁻¹.
LoD in Introduction: 36 nGy·s⁻¹.
LoD in Section 2.4: "65 and 39 nGy·s⁻¹, respectively" with contradictory statement that the higher-LoD Cu device "continues to outperform" the Co device.
LoD in Conclusion: "39 and 65 nGy·s⁻¹, respectively"—ordering reversed relative to Section 2.4.
Figure 3g / Section 2.3: trap density 5.8 × 10⁸ cm⁻³ (COF366-Cu) and 5.2 × 10⁸ cm⁻³ (COF366-Co), measured by SCLC on pellets pressed at "one-ton pressure."
Dark current drift units: Abstract uses A·cm⁻¹·V⁻¹·s⁻¹; Section 2.3 uses A·cm⁻²·V⁻¹·s⁻¹.
DOI: 10.1002/anie.4087507.

Notes

The report cannot determine whether Figures 1–4 or the Supplementary Information contain image reuse or splicing; only textual analysis was possible from the materials provided.
The report recommends (a) requesting the original I–V traces and SCLC fitting files from the authors, (b) raising the LoD contradiction on PubPeer, (c) notifying the journal editorial office, and (d) broader quality review of the group's recent publications.
All findings are preliminary and require confirmation through institutional investigation and inspection of original laboratory data; false positives and false negatives remain possible.