Forensic Review: Effects of joint size and isothermal aging on interfacial IMC growth in Sn-3.0Ag-0.5Cu-0.1TiO2 solder joints (DOI: 10.1016/j.jallcom.2016.12.131)

Academic fraud report · Geng Detector

Summary

This report assesses a 2017 Journal of Alloys and Compounds paper by Z.L. Li, L.X. Cheng, G.Y. Li, J.H. Huang, and Y. Tang regarding interfacial IMC growth in Sn-3.0Ag-0.5Cu-0.1TiO2 solder joints. The overall verdict is 'Questionable.' The principal red flag concerns stated growth-rate (diffusion coefficient) values of 7.32, 27.04, 79.78, 3.77, 12.98, and 67.83 cm²/s in Section 3.2. Solid-state Cu diffusion physically cannot occur at cm²/s; back-of-envelope calculation from the reported ~1–3 µm thickness change over 480 h yields D on the order of 10^-15 cm²/s, strongly suggesting a missing scientific-notation exponent in the typeset PDF (likely an OCR/superscript loss). A secondary, lower-severity concern is the suspiciously smooth monotonic Cu-concentration decline in Table 3 across 0/120/240/480 h with SDs confined to 0.01–0.05. No image-based analysis was possible. Confidence is moderate; verdict depends on whether the unit is a transcription artefact or an author error.

Verdict

Questionable (🟡). The report flags one likely serious typesetting/extraction issue and one statistical-quality concern; neither alone proves misconduct, but together they warrant author clarification.

Key findings

Implausible diffusion-coefficient units: Section 3.2 reports growth rates of 7.32, 27.04, and 79.78 decreasing to 3.77, 12.98, and 67.83 cm²/s. Solid-state metal diffusion coefficients are physically on the order of 10^-12 to 10^-15 cm²/s. The values are likely missing an exponent such as ×10^-15 cm²/s.
Quantitative cross-check supports a missing exponent: Using y = √(Dt) with y ≈ 3 µm and t = 480 h gives D on the order of 10^-15 cm²/s, consistent with the missing-exponent hypothesis.
Overly regular Cu-concentration decline in Table 3: Average Cu concentration drops monotonically with aging time (0 → 480 h) for 90 µm (1.01→0.53→0.42→0.38), 150 µm (1.13→0.62→0.51→0.45), and 200 µm (1.35→0.84→0.63→0.56) pads, with SDs consistently in the 0.01–0.05 range and no outlier scatter typical of grain-boundary heterogeneity.
No image-based checks possible: Figures 7–9 (SEM micrographs) were not available for pixel-level duplication or splicing analysis in text-only mode.

Evidence highlights

DOI: 10.1016/j.jallcom.2016.12.131
Reported growth-rate values: 7.32, 27.04, 79.78 → 3.77, 12.98, 67.83 cm²/s (units as printed).
Thickness change used for estimate: ~1 µm → ~3 µm over t = 480 h.
Table 3 Cu-concentration trajectories for pad sizes 90, 150, 200 µm; SDs 0.01–0.05 throughout.

Notes

The diffusion-coefficient issue is most plausibly a superscript loss during PDF text extraction rather than an authorial physics error, but this must be verified against the original PDF (48.pdf).
The smoothness of the EDX concentration table is statistically unusual but not by itself evidence of fabrication; it may reflect averaging over many measurements.
Follow-up actions: verify the original PDF rendering of Section 3.2; contact authors to clarify physical units; review co-author group publications (refs [5–9]) for similar patterns.
Limitations: text-only review; no Western blot or SEM noise/splicing analysis performed; final misconduct determination requires institutional investigation.