Integrity Detection Report: Phenylboronic acid modified Ag₂S NPs for bacterial targeting photothermal antibacterial therapy and promotion of wound healing (RSC Advances, DOI 10.1039/d5ra09845a)

Academic fraud report · Geng Detector

Summary

This report assesses the paper 'Phenylboronic acid modified Ag₂S NPs for bacterial targeting photothermal antibacterial therapy and promotion of wound healing' (RSC Advances, DOI 10.1039/d5ra09845a). Overall verdict: highly suspicious (report states 99.3% posterior probability, BF = 2.53e+03; this is an algorithmic estimate, not a formal finding of misconduct). Confirmed internal contradictions include: in vivo PTT laser irradiation duration stated as 10 minutes in Methods (Section 2.8) but 5 minutes in Results (Section 3.5); two mutually exclusive wound-area values (17.03±1.30% and 7.32±1.17%) attributed to the same treatment group; implausible reagent units and stoichiometry (200 mg/mL nanoparticle concentration, 3 mL vs 3 μL FITC, AgNO₃/Na₂S quantities differing by ~8 orders of magnitude). Weaker, non-conclusive signals include statistically 'too regular' data presentation, anomalous last-digit/SD distributions (χ²=373.6, p<0.0001; odd/even ratio 33/134), and image-forensics copy-move/noise flags across 7 embedded images, all of which have partial benign explanations (multi-panel figure conventions, formatting conventions, smoothing) and require original data verification. Key limitation: analysis is AI-assisted, based on the PDF alone; original records were not available.

Verdict

Highly suspicious (🟠/🔴 per report; auto-synthesized posterior 99.3%, 95% credible interval [93.4%, 99.9%], Bayes Factor 2.53e+03). This is an AI-assisted screening assessment, not a determination of academic misconduct. Original data have not been examined.

Key findings

Method–Results contradiction (confirmed): Section 2.8 states 808 nm laser at 2.5 W cm⁻² for 10 minutes on days 1, 3, and 5; Section 3.5 states 5 minutes on the same days. Thermal imaging description (Fig. 5A) again references 10 minutes. All three descriptions point to the same treatment procedure.
Group mislabeling in results (confirmed): Section 3.5 attributes two different terminal wound-area values — 17.03±1.30% and 7.32±1.17% — to the same "PBA-Ag₂S@PC NPs" group; contextually the latter likely belongs to the "PBA-Ag₂S@PC NPs + PTT" group.
Reagent units and stoichiometry (confirmed): Nanoparticle concentration consistently written as 200 mg/mL throughout (likely intended μg/mL); FITC labeled amount "3 mL" likely 3 μL; synthesis uses 1 mL of 0.2 mM AgNO₃ (0.2 μmol Ag) with 18.2 mg (~0.076 mol) Na₂S — a ~8-order-of-magnitude molar discrepancy that is chemically implausible.
Overly regular data presentation (suspected): Nearly identical error bars (Fig. 1E, 5C), perfectly parallel curves (Fig. 2C), near-linear uniform decline in viability (Fig. 3A); SDs appear proportionally scaled to means (21.70±2.17%, 17.03±1.30%, 12.72±1.61%, 7.32±1.17%). Possible benign explanations (smoothing, material homogeneity); the paper does not declare smoothing/normalization.
Last-digit and SD anomalies (suspected): Last-digit distribution highly non-uniform (χ²=373.6, p<0.0001); odd/even ratio 33/134; adjacent-digit correlation deviation 10.8σ; all SDs reported to exactly two decimals. Tests mixed quantities and have limited sample size — downgraded to a flag.
Image forensics (suspected): Copy-move matches at consistent offsets (e.g., (184,0), (0,32), (32,0); 44–124 block pairs) detected in all 7 embedded images (img-000 to img-006); noise variance inconsistent in 6 images (CV 0.49–1.32). Multi-panel figure conventions offer partial benign explanation (caption-supported); noise findings lack positive support in the text. Low-confidence visual lead: Fig. 4I/4J Control background textures possibly related.

Evidence highlights

Laser duration: 10 min (Methods, §2.8) vs 5 min (Results, §3.5) vs 10 min (Fig. 5A thermal imaging narrative).
Wound area: 17.03±1.30% and 7.32±1.17% both assigned to "PBA-Ag₂S@PC NPs treatment group".
Stoichiometry: 0.2 μmol AgNO₃ vs ~0.076 mol Na₂S (~10⁸-fold excess); 200 mg/mL concentration implausible for nanoparticle dispersion.
Statistics: χ²=373.6, p<0.0001 (last-digit); odd/even 33/134; adjacent-digit deviation 10.8σ.
Forensics: 7/7 images with copy-move matches; 6/7 with noise variance inconsistency; report cites logLR contributions of 1.59, 1.59, 1.19, 1.31, 1.08 for top evidence items; 18 candidate leads excluded on review.

Notes

DOI: 10.1039/d5ra09845a; journal: RSC Advances; publication year listed as 2026; source file: d5ra09845a.pdf.
Findings 1, 2, and 6 were marked as confirmed on review; findings 3, 4, and 5 retained as flags due to plausible benign explanations (multi-panel layout, reporting conventions, smoothing).
Recommended actions: request original data from authors (photothermal raw traces, CFU plates, in vivo irradiation protocol records, AgNO₃/Na₂S batching records, raw figure data); PubPeer posting, journal notification, and institutional reporting not yet done.
Analysis is AI-assisted; the Bayesian synthesis treats evidence as conditionally independent, which may overstate joint strength. Final determination of misconduct requires official investigation.