Geng Report: Phase change doped sand storage for isothermal thermoelectric heat delivery (Energy Materials, DOI: 10.20517/energymater.2026.73)

Academic fraud report · Geng Detector

Summary

Verdict: ⚠️ Suspicious / Inconclusive (the reviewer downgrades from the prior framing because several prior claims could not be verified against the available text). Key issues examined: (1) Image reuse / splicing — pixel-level checks were not possible; no textual contradictions were found. The authors explicitly state Figures 1 and 2 are AI-generated (Gemini 2.5 Pro), and remaining figures are reported as simulation outputs. (2) Data fabrication — efficiency figures (≈5% end-to-end, 0.74, 0.82 delivery efficiency) show realistic non-integer values consistent with COMSOL-based CFD/thermal simulation, and the authors provide a step-by-step efficiency derivation. (3) Statistical anomalies — not applicable; no biological or clinical statistics. Grid-independence checks are reported (e.g., 0.1 K change, <0.4% variation, residuals below 10⁻⁶). (4) Production-line publication concerns — timeline (Received 16 Apr 2026; Accepted 11 Jun 2026) is reasonable for a multi-physics simulation paper. Confidence limits: the reviewer could not locate the alleged Western Blot concerns, Figure 9/10 numbering error, or GitHub open-source statement in the supplied truncated text (truncated at line 973). The prior report therefore appears to contain hallucinations. Based solely on verifiable content, the paper appears technically sound but cannot be cleared without the full PDF.

Verdict

⚠️ Suspicious / Inconclusive. Downgraded from prior framing because multiple allegations in the original report could not be corroborated against the provided text.

Key findings

🔍 Image reuse / splicing: Not assessable at the pixel level. Authors explicitly disclose that Figures 1 and 2 are AI-generated (Gemini 2.5 Pro, dated March 2025); other figures are described as numerical simulation outputs. No internal textual contradictions found.
📊 Data fabrication indicators: Efficiency values (≈5% end-to-end; 0.74, 0.82 thermal delivery efficiency) are non-integer and internally consistent. The authors provide an explicit mathematical derivation (around line 176: "delivering 82% of this to a module operating near 7% conversion… efficiency on the order of 5%").
🧪 Statistical anomalies: Not applicable — the work is a COMSOL-based multi-physics simulation, with no p-values, animal data, or clinical sampling. Grid-independence checks are reported (line 442: "changed by 0.1 K"; line 449: "less than 0.4%"; line 467: residuals "fell below 10⁻⁶").
🗓️ Production-line / timeline concerns: Received 16 Apr 2026; Accepted 11 Jun 2026 (~1.5 months), reasonable for a three-layer multi-physics simulation paper. No "forced maturation" pattern detected.
⚠️ Figure numbering discrepancy: Original report alleges two Figure 9s and a missing Figure 10 in caption numbering. This claim could not be verified because the supplied PDF text is truncated at line 973 (the alleged issues are reported near lines 1015 and 1424). Marked as unsubstantiated.
⚠️ Allegations not verifiable in supplied text: References to Western Blot analysis, open-source GitHub repository (https://github.com/QiruiDING/Composite-Sand-Based...), and detailed Figure 9/10 caption errors are not present in the truncated text and likely reflect hallucinated content in the prior report.

Evidence highlights

DOI: 10.20517/energymater.2026.73
Journal: Energy Materials
Year: 2026
Authors: Qirui Ding, Jiacheng Liu, Changhui Song, Lili Zeng, Weicheng Cui
AI/AI-assisted tools statement: Gemini 2.5 Pro (March 2025) used for Figures 1 and 2 only.
Reported key efficiency figures: end-to-end ≈5%; thermal delivery 0.74 → 0.82.
Convergence diagnostics: 0.1 K change; <0.4% variation; residuals <10⁻⁶.
Timeline: Received 16 Apr 2026; Accepted 11 Jun 2026.

Notes

The paper appears to be a purely computational/CFD-thermal study; checks for Western Blot or biological image manipulation are categorically irrelevant.
A full-text PDF is required to confirm or refute the alleged Figure 9/10 numbering error and any data/code-availability statement.
Several accusations in the original Chinese report appear to be hallucinations (Western Blot concerns, GitHub link, Figure 9/10 caption mismatch). These have been excluded from the verdict.
Authors do not need to be contacted for raw data, as the work is a numerical simulation with described parameters; however, public availability of the COMSOL model files would strengthen reproducibility.
This report is AI-assisted and intended for academic discussion only; final determination of misconduct requires an institutional investigation.