Investigation Report: Boosting Selectivity toward Long-Chain α-Olefins in CO2 Hydrogenation through Strengthening ZnO−Fe Interactions (JACS Au, DOI: 10.1021/jacsau.5c01605)

Academic fraud report · Geng Detector

Summary

This report evaluates a 2026 JACS Au paper on Fe–Zn Fischer–Tropsch catalysts for CO2 hydrogenation and assigns a verdict of Highly Suspicious (orange). The primary confirmed issue is internal numerical inconsistency between the Abstract and the main text: the Abstract reports C4+ linear α-olefin selectivities of 50.5% and 61.7% for Fe2Zn1-A and Fe2Zn1-U, while the body of the paper gives 50.6% and 61.1%. The Abstract also cites a CH4 selectivity reduction to 8.8%, but the text states 9.5% under standard conditions; the 8.8% value appears to correspond to an extreme GHSV of 20000 h−1. Two additional concerns are flagged as insufficiently evidenced: suspiciously smooth, near-identical time-on-stream profiles across distinct catalysts in Figure 1, and over-processed in situ XPS spectra in Figure 4 showing uniform baseline suppression. Confidence in the confirmed data-consistency finding is high; confidence in the imaging-related findings is moderate pending raw data.

Verdict

Highly Suspicious (🟠). The paper contains a clear, reproducible inconsistency between Abstract and main-text catalytic performance numbers, suggesting selective data presentation for narrative effect. Two further concerns (time-series smoothing and XPS over-processing) are plausible but require raw data for confirmation.

Key findings

Internal data inconsistency (confirmed). Abstract vs. body values disagree on C4+ LAO selectivity (50.5% vs. 50.6%; 61.7% vs. 61.1%) and on CH4 selectivity reduction (8.8% in Abstract vs. 9.5% in text under standard conditions; 8.8% only achieved at GHSV = 20000 h−1).
Possible selective data cherry-picking. The Abstract appears to combine the best values from different reaction conditions into a single headline statement, which is consistent with biased data curation.
Time-on-stream data anomalies (unconfirmed). Figure 1a–h panels for Fe2Zn1-A and Fe2Zn1-U show near-identical fluctuation patterns over 0–50 h TOS, which is chemically implausible for distinct catalysts undergoing phase evolution and coking.
XPS spectrum over-processing (unconfirmed). Figure 4 in situ Fe 2p and Zn LMM spectra exhibit fully suppressed background drift and synchronized noise frequencies across fresh/reduced/spent samples, indicative of aggressive global baseline subtraction or smoothing.
No timeline anomalies. Submitted 2025-11-29, published 2026-03-03; all cited literature and instrumentation predate submission.

Evidence highlights

Abstract claim: "C4+ LAOs selectivity … 50.5% (Fe2Zn1-A) and 61.7% (Fe2Zn1-U)"; "CH4 selectivity was reduced from 13.7% … to 8.8%."
Body text (Page 1704): C4+ LAO selectivities reported as 50.6% and 61.1%; CH4 selectivity for Fe2Zn1-U given as 9.5% ("was reduced from 13.7% and 36.0% to 9.5% and 26.6%"). The 8.8% value corresponds to a separate high-GHSV (20000) test.
Figure 1 (Page 3): Fe2Zn1-A and Fe2Zn1-U TOS curves (0–50 h) in panels a–d and e–h share unusually similar fluctuation/noise signatures.
Figure 4 (Page 6): In situ Fe 2p XPS and Zn LMM auger spectra show zero background drift and identical noise frequencies across fresh, reduced, and spent states.
DOI: 10.1021/jacsau.5c01605.

Notes

The data-inconsistency finding can be verified directly from the published PDF without raw data.
The smoothing/XPS concerns are inferences from visual analysis; confirmation requires the original GC chromatograms, raw TOS data files, and XPS source files (.vms/.dat).
Differences of 0.1–0.6 percentage points between Abstract and body, while small, involve the headline performance metrics of the paper and are not typographical self-consistent.
Recommend posting the numerical discrepancies on PubPeer and requesting author clarification or an erratum.