Geng Integrity Report: Tetraphenylsilane-Engineered MR-TADF Emitters for Pure-Green OLEDs (DOI: 10.1002/adma.72684)

Academic fraud report · Geng Detector

Summary

This report evaluates a 2026 Advanced Materials paper by Zhuixing Xue and co-authors (Yang group) reporting tetraphenylsilane-engineered MR-TADF emitters for pure-green OLEDs. The overall verdict is 'Questionable' (yellow/medium concern). Three anomalies were flagged based on visual analysis of figures, all concerning the unusually smooth appearance of performance curves. Specifically, Figure 5d lifetime decay traces over 1421–2491 hours appear excessively smooth and suspiciously parallel across three different emitters. Figure 4c/4d EQE-L and PE/CE-L curves for three structurally distinct emitters overlap almost perfectly at low luminance (1–100 cd m⁻²). Figure 3c transient PL decay shows an unnaturally clean log-linear profile. No fabrication-level evidence (e.g., duplicated panels, impossible numerical values, identity-reused images) was identified. The assessment is constrained to figure-level visual inspection; raw data files were not available. The findings are consistent with potential cosmetic data processing rather than outright falsification, and should be treated as a call for raw-data verification rather than a definitive misconduct finding.

Verdict

Questionable (yellow) — three medium- to low-severity concerns regarding figure smoothness; no decisive evidence of fabrication identified within the scope of visual figure inspection.

Key findings

Figure 5d (Operation lifetime, p. 9): LT80 decay curves reported at 1421 h, 2006 h, and 2491 h appear unnaturally smooth and near-perfectly parallel across three distinct emitters. Sustained OLED aging tests typically show small fluctuations from grid variations, temperature control, and batch differences.
Figure 4c/4d (EQE-L and PE/CE-L, p. 8): Three emitters (p-DBFSi, m-DBFSi, m-DBFSi-d) with different substitution positions and deuteration show near-perfect overlap in the 10⁰–10² cd m⁻² range, where subtle differences in turn-on behavior and leakage current are normally observable.
Figure 3c (Transient PL decay, p. 6): Log-linear transient PL decay is unusually clean with minimal baseline noise at the tail, raising the possibility of aggressive averaging or manual trimming.

Evidence highlights

Lifetime values referenced: 1421 h, 2006 h, 2491 h (Figure 5d).
Low-luminance overlap range: 10⁰–10² cd m⁻² (Figure 4c/4d).
Emitters compared: p-DBFSi, m-DBFSi, m-DBFSi-d (deuterated variant).
DOI: 10.1002/adma.72684.
No duplicated panels, copy-pasted background regions, or impossible numerical inconsistencies were detected.

Notes

All three findings are tagged as '依据不足' (insufficient evidence) in the source report; they reflect anomalies in presentation rather than confirmed manipulation.
The concerns are consistent with over-processed plotting (e.g., excessive smoothing, binning, or selective averaging) rather than outright data fabrication.
Recommended follow-up: request raw .txt/.csv lifetime and efficiency data from the authors; consider a PubPeer comment asking for clarification of curve processing; if irregularities persist, flag to the journal editorial office for review of instrument logs.
Confidence: low–moderate. Limits: no access to raw numerical data, no access to experimental logs, and visual judgment of figure smoothness is inherently subjective. Final determination of misconduct requires institutional investigation.