Image Reuse and Suspicious Data Patterns in 'Tetraphenylsilane-Engineered MR-TADF Emitters for Pure-Green OLEDs with High-Efficiency, Long Lifetime, and Gamut Close to BT.2020 Standard' (Advanced Materials, DOI: 10.1002/adma.72684)

Academic fraud report · Geng Detector

Summary

Verdict: Strong indicators of image reuse and possible data fabrication. Two principal issues are flagged. First, Figure 4(e) shows inset device photographs claimed to represent both m-DBFSi and m-DBFSi-d devices; visual analysis indicates these insets are pixel-level duplicates and also match the m-DBFSi-d photo in Figure 1, including background noise and watermark placement, which is inconsistent with genuinely separate devices. Second, the transient photoluminescence decay curves in Figure 3(c) for p-DBFSi, m-DBFSi, and m-DBFSi-d share near-identical high-frequency noise patterns in the delayed fluorescence window (~10–100 µs), despite nominally distinct lifetimes (42.1, 40.5, 39.4 µs), suggesting possible copying or artificial smoothing. Reported performance (EQE 41.9%, LT80 2491 h) is exceptionally high and warrants scrutiny. Confidence is limited to visual and statistical pattern analysis; no raw data was examined. Definitive determination requires investigation by the journal or institution.

Verdict

Strong indicators of image reuse in Figure 4(e) and suspicious visual duplication of noise patterns in transient decay data in Figure 3(c). Reported at 🔴 (实锤) severity by the original Chinese report. Independent verification of raw data is required for formal misconduct findings.

Key findings

Figure 4(e) inset device photographs allegedly showing m-DBFSi and m-DBFSi-d devices appear pixel-level identical and also match the m-DBFSi-d image in Figure 1 (same background noise, watermark position).
The caption explicitly labels the insets as 'photograph of m-DBFSi-based and m-DBFSi-d-based devices,' contradicting the visual evidence.
Figure 3(c) transient PL decay curves for three distinct molecules (p-DBFSi, m-DBFSi, m-DBFSi-d) show implausibly identical noise signatures in the 10–100 µs delayed fluorescence region.
Table 1 lifetime values (42.1, 40.5, 39.4 µs) are very close, compounding concern that curves were copied or synthetically smoothed.
Reported performance metrics (EQE 41.9%; LT80 2491 h; gamut near BT.2020) are unusually high for MR-TADF emitters and heighten scrutiny.

Evidence highlights

DOI: 10.1002/adma.72684
Figure 4(e) inset image duplication; identical watermark/background noise across Figure 1 and Figure 4(e) insets.
Figure 3(c): coincident micro-noise structure across three molecules with distinct chemical structures.
Table 1 reported lifetimes: 42.1 µs (p-DBFSi), 40.5 µs (m-DBFSi), 39.4 µs (m-DBFSi-d).
Reported device performance: EQE 41.9%; LT80 2491 hours; CIE close to BT.2020 green standard.

Notes

Confidence is moderate to high based on visual patterns; raw data files (e.g., original spectrometer traces, raw device photographs, instrument log files) were not available for reanalysis.
Genuinely identical noise patterns between independent samples are highly improbable under typical experimental conditions, but shared instrument, batch, or post-processing pipelines could partially explain overlap; this does not fully account for the Figure 4(e) duplicate device photos.
Recommended actions include requesting original photographs and raw PL decay data from the authors, posting concerns on PubPeer, notifying the Advanced Materials editorial office, and informing the authors' institutional research integrity committees.
Original Chinese report includes a disclaimer that findings are AI-assisted and not a definitive misconduct determination; this English translation preserves that caveat.
No evidence of authorship fraud, figure fabrication beyond the noted panels, or statistical manipulation of Table 1 numerics was assessed.