Fraud Detection Report: Solvent-Free Chemical Recycling of Polyesters and Polycarbonates by Magnesium-Based Lewis Acid Catalyst (Angew. Chem. Int. Ed. 2024, DOI: 10.1002/anie.202420688)

Academic fraud report · Geng Detector

Summary

Verdict: Highly suspicious (高度可疑). The report identifies three principal concerns in Zhao et al.'s paper on Mg-based Lewis acid-catalyzed solvent-free recycling of polyesters and polycarbonates. (1) A likely copy-paste error in the Results and Discussion attributes a crystallization temperature (Tc) of 2060 °C to ε-caprolactone — a value far above the decomposition temperature of any organic compound, consistent with misplacement of a 'by 2060' forecast phrase. (2) Catalytic recycling data claim triplicate averages (Tables 1, 2 footnotes) but report no standard deviation or error bars; Figure 5 shows 10 reuse cycles tightly clustered at 92–96% yield with no error bars — implausibly uniform for real catalytic runs. (3) Figure 2 TGA curves appear excessively smooth, lacking baseline noise typical of genuine thermogravimetric traces, raising suspicion of post-acquisition smoothing or simulated data. Confidence is moderate-high; concerns (1) and (2) are well supported, while (3) requires raw instrument files for confirmation.

Verdict

Highly suspicious (高度可疑). Multiple methodological and data-integrity red flags were identified. The most substantive issues are a physically implausible temperature claim and the absence of error statistics alongside suspiciously uniform catalytic recycling data.

Key findings

Physically implausible crystallization temperature (2060 °C) for ε-caprolactone (ε-CL). Reported in the first paragraph of Results and Discussion. 2060 °C is far above the thermal decomposition threshold of any organic compound and is inconsistent with known PCL/ε-CL crystallization behavior. The figure strongly suggests a copy-paste artifact from a 'by 2060' projection elsewhere in the manuscript, but it nonetheless reveals serious drafting, proofreading, and peer-review lapses.
Recycling yield data over-uniform and lacking error bars. Tables 1 and 2 footnoted as averages of three independent replicates, yet no standard deviation (SD) or standard error (SEM) is reported. Figure 5 reports ten successive catalyst reuse runs at 92–96% yield with perfectly level bars and no error bars, an implausibly narrow distribution for real catalytic recycling experiments subject to mass-transfer and deactivation variability. Data selection, fitting, or fabrication cannot be ruled out from the paper alone.
Suspiciously smooth TGA curves (Figure 2). Visual inspection indicates the weight-loss traces lack the baseline noise and micro-drift characteristic of genuine TGA output. Possible explanations include over-aggressive software smoothing or computer-generated/simulated curves. Severity rated medium pending inspection of raw instrument files (.txt or proprietary TGA export formats).

Evidence highlights

DOI: 10.1002/anie.202420688
Anomalous value cited verbatim: "because the Tc value of ɛ-CL is as high as 2060 °C" (Results and Discussion, first paragraph).
Yield window in Figure 5: 92%–96% across 10 cycles, with no error bars.
Tables 1 and 2 footnote: data reported as "The average separation yield of three times", but no SD/SEM provided anywhere.
Figure 2: TGA curves for PCL under different Lewis acids visually smooth, lacking baseline noise.

Notes

Confidence is moderate-to-high for findings 1 and 2; finding 3 (TGA smoothness) is suggestive but requires raw TGA exports and laboratory notebooks for confirmation.
The 2060 °C figure may reflect an editorial/copy-paste error rather than intentional misrepresentation, but its survival through peer review at Angewandte Chemie is itself a serious quality concern.
Recommended follow-up: (a) request raw, instrument-level TGA data and replicate-level yield spreadsheets from the corresponding author; (b) raise a PubPeer comment requesting error statistics for Tables 1, 2 and Figure 5; (c) notify the journal editor to audit the original experimental records.
This AI-assisted assessment is for academic discussion only and is not a formal finding of misconduct.