Integrity Review Report: "Real-valued gain enabled higher-order exceptional point under anti-parity-time symmetry" (Zhu et al., 2025)

Academic fraud report · Geng Detector

Summary

This report assesses potential integrity concerns in Zhu et al. (2025), published in Applied Physics Letters (DOI: 10.1063/5.0283263). The overall verdict is "Questionable" (suspect). The analysis is limited because only plain text was provided; image-based reuse and Photoshop splicing checks (Formulas 1 and 3) could not be performed. Key findings include: (1) a description of resistor arrays with "61% tolerance," which is implausible because no commercial resistor series uses a 61% tolerance class and the figure is likely a typo or OCR artifact (e.g., "1%" misread as "61%"); (2) an apparent misnaming of the impedance analyzer as "Keysight 4990A" instead of the correct "E4990A." Timeline verification passed (submitted 2025-05-30, accepted 2025-07-03, published 2025-07-18). No raw data tables or scatter plots with error bars were available for last-digit or outlier analysis. Confidence in the flagged text anomalies is moderate; pixel-level conclusions cannot be drawn.

Verdict

🟡 Questionable — Based purely on textual content. No image-reuse or splicing checks could be performed because figures were not available for analysis. Findings are suggestive but not conclusive of misconduct.

Key findings

Implausible component tolerance: The methods section states the use of "Yageo, 61% tolerance" resistor arrays to set loss rates γ2 = γ3 = γ0. No standard resistor series carries a 61% tolerance; common classes are 1% (F), 5% (J), or 10% (K). Most likely a typo or OCR misread of "1%," but it raises questions about editorial scrutiny.
Instrument model misnamed: The impedance analyzer is listed as "Keysight, 4990A." The standard model designation is E4990A, suggesting an omitted prefix.
No anomalous timeline: Submitted 2025-05-30; accepted 2025-07-03; published 2025-07-18. Approximately one-month review is plausible for Applied Physics Letters.
Limited raw-data visibility: The paper relies on analytical expressions, equivalent circuits, and characteristic curves; absence of numerical tables and detailed scatter plots with error bars precludes last-digit and statistical anomaly checks.

Evidence highlights

Location: Page 4, right column, first paragraph — direct quote: *"The loss rates γ2=γ3=γ0 are ensured by using identical resistance arrays R2 and R3 (Yageo, 61% tolerance)."*
Location: Page 4, right column, first paragraph — instrument referenced as "Keysight, 4990A."
DOI: 10.1063/5.0283263
Timeline: Submission 2025-05-30 → Acceptance 2025-07-03 → Publication 2025-07-18 (review ≈ 34 days).

Notes

The "61% tolerance" and "4990A" issues are most plausibly typographical / OCR artifacts, not deliberate fabrication, but they underscore insufficient copy-editing.
Without figures, no assertion can be made regarding image duplication, splicing, or fabricated spectra.
Recommended next step: verify against the original PDF to confirm whether "61%" and "4990A" are rendering errors; perform image reuse (Formulas 1 and 3) checks if high-resolution figures become available.
Confidence: moderate on textual anomalies; n/a on image integrity.