Adhesively bonded composite structures cannot be certified for primary load-bearing use because non-destructive testing (NDT) cannot reliably detect weak adhesive bonds, above all kissing bonds (full physical contact without chemical adhesion), cannot inspect many real structures without water immersion, and cannot translate ultrasonic indications into residual-strength acceptance criteria. The source study (Yılmaz and Jasiūnienė, 2020) demonstrated high-frequency immersion acoustic microscopy with post-processing for contamination-induced weak bonds in CFRP/epoxy single-lap joints and explicitly flagged kissing bonds, structure-dimension/immersion limits, and generalization as unresolved. Existing literature (2020–2026) reports method-level progress on nonlinear ultrasonics, guided waves, air-coupled testing, machine learning and model-assisted probability-of-detection, but no validated, certification-capable resolution of the source-raised question: reliable kissing-bond detection in realistic (non-immersion) inspection configurations. The quantified mapping between ultrasonic observables and mechanically measured residual bond strength, which the source paper does not itself pose, remains a related field-level barrier.

Background

Adhesive bonding is increasingly used to join carbon-fibre-reinforced polymer (CFRP, a polymer-matrix composite reinforced with woven carbon fibres) laminates in aerospace, automotive and marine structures, typically in single-lap joint geometry (two overlapping adherend plates bonded by a thin adhesive film). Its main safety limitation is inspectability: non-destructive testing (NDT) must confirm bond quality after manufacture. Gross defects such as disbonds, voids and porosity are readily found by conventional ultrasonic testing, but weak bonds are far harder. A weak adhesive bond has near-normal geometry yet reduced adhesion, and its limiting case is the kissing bond: an interface in intimate physical contact but with absent or inadequate chemical bonding, so that the ultrasonic impedance mismatch that normally reveals a disbond essentially vanishes while the joint remains mechanically weak (Nagy, 1991; Adams and Cawley, 1991). Kissing bonds were shown early on to be nearly invisible to linear pulse-echo ultrasound, and compressive loading or nonlinear ultrasonic indicators were proposed to make them detectable (Brotherhood et al., 2002; Yan et al., 2009). The detection problem is aggravated by two further constraints. First, the highest-resolution ultrasonic method for bondlines, scanning acoustic microscopy (a C-scan technique using a focused transducer of order tens of MHz, which requires the part to be immersed in water), cannot be applied to large or water-sensitive structures, whereas contactless air-coupled ultrasound operates only at much lower frequencies (hundreds of kHz) with correspondingly poorer resolution. Second, even when a bond anomaly is detected, certification requires acceptance criteria tied to mechanical strength, and it has long been unclear whether any ultrasonic observable can predict residual joint strength at all — the non-destructive evaluation of bondline quality is repeatedly described as elusive. Within this context, Yılmaz and Jasiūnienė (2020) inspected CFRP/epoxy single-lap joints (HexPly M21 satin prepreg adherends, 3M AF163 film adhesive) containing four bond qualities — perfect bond, disbond (release-film inclusion), and weak bonds with two levels of release-agent contamination — using a 50 MHz focused immersion transducer in pulse-echo, wavelet denoising, time-of-flight alignment, and amplitude/phase interface visualizations in both time and frequency domains. They quantified each post-processed image with a bonding characteristic value (BCV), their own image metric proportional to the mean-square deviation of pixel values relative to the maximum observed value, which indicates defect likelihood. Detection of the contamination-induced weak bonds succeeded, but the authors explicitly noted the limitations regarding structure dimensions and water immersion, and stated that different bonding qualities such as kissing bonds should be studied separately. Subsequent work by the same group and others compared inspection modalities, built model-assisted probability-of-detection (POD, the statistical relationship between defect size and detection reliability) frameworks, and developed nonlinear ultrasonic methods — contact acoustic nonlinearity (weak-contact vibration that generates ultrasonic harmonics), wave mixing (interaction of two ultrasonic waves whose combined response reveals interface nonlinearity), and nonlinear guided waves (ultrasonic modes travelling along the bonded plates) — yet a 2026 review of composite bonding repairs still lists the limited capability of NDT to detect all bondline defects, quantify bond strength and assess bond quality as a central certification barrier.

Problem Statement

Can weak adhesive bonds of the kissing-bond type — interfaces in full physical contact but without adequate chemical adhesion — be detected and characterized ultrasonically in composite-adhesive joints with validated reliability under inspection conditions compatible with real structures, and over a stated envelope of adherend/adhesive systems? The question, raised by Yılmaz and Jasiūnienė (2020) as the explicit boundary of their contamination-weak-bond study, comprises two jointly linked sub-objectives: (i) ultrasonic detection and characterization of kissing bonds as distinct from contamination-induced weak bonds (the source study's partially bonded, release-agent-contaminated specimens, as against the essentially zero interfacial adhesion of a kissing bond), whose detectability and signatures were conjectured to differ; and (ii) applicability beyond the immersion acoustic-microscopy configuration of that study, i.e. to larger or non-immersible structures via dry, contact or air-coupled implementations, recognizing the resolution loss documented for low-frequency air-coupled testing. A quantified relationship between the ultrasonic indication (in that study, the bonding characteristic value — an image-level defect-likelihood metric — and, in later work, linear or nonlinear ultrasonic observables) and the mechanically measured residual strength of the joint is a well-documented field-level open problem that the source paper does not itself pose; it is relevant context for the certification barrier but is not part of the completion criterion for this question. The problem is not restricted to any single transducer, frequency, waveform, or analysis method; an answer demonstrating, with adequate statistical power over a stated envelope of tested systems, observables and configurations, that ultrasonic observables cannot carry the required kissing-state information would equally resolve the question in the negative within that stated envelope.

References

  1. Yılmaz B, Jasiūnienė E (2020). Advanced ultrasonic NDT for weak bond detection in composite-adhesive bonded structures. International Journal of Adhesion and Adhesives 102:102675. DOI 10.1016/j.ijadhadh.2020.102675. https://doi.org/10.1016/j.ijadhadh.2020.102675
  2. Nagy PB (1991). Ultrasonic detection of kissing bonds at adhesive interfaces. Journal of Adhesion Science and Technology 5:619–630. DOI 10.1163/156856191X00521. https://doi.org/10.1163/156856191X00521
  3. Adams RD, Cawley P (1991). A review of defect types and nondestructive testing techniques for composites and bonded joints. NDT & E International 24:105–114. DOI 10.1016/0963-8695(91)90924-R. https://doi.org/10.1016/0963-8695(91)90924-R
  4. Brotherhood CJ, Drinkwater BW, Guild FJ (2002). The effect of compressive loading on the ultrasonic detectability of kissing bonds in adhesive joints. Journal of Nondestructive Evaluation 21:95–104. DOI 10.1023/A:1022584822730. https://doi.org/10.1023/A:1022584822730
  5. Yan D, Drinkwater BW, Neild SA (2009). Measurement of the ultrasonic nonlinearity of kissing bonds in adhesive joints. NDT & E International 42:459–466. DOI 10.1016/j.ndteint.2009.02.002. https://doi.org/10.1016/j.ndteint.2009.02.002
  6. Yılmaz B, Asokkumar A, Jasiūnienė E, Kažys RJ (2020). Air-coupled, contact, and immersion ultrasonic non-destructive testing: comparison for bonding quality evaluation. Applied Sciences 10:6757. DOI 10.3390/app10196757. https://doi.org/10.3390/app10196757
  7. Yılmaz B, Smagulova D, Jasiūnienė E (2022). Model-assisted reliability assessment for adhesive bonding quality evaluation with ultrasonic NDT. NDT & E International 126:102596. DOI 10.1016/j.ndteint.2021.102596. https://doi.org/10.1016/j.ndteint.2021.102596
  8. Boll B, Willmann E, Fiedler B, Meißner RH (2021). Weak adhesion detection – enhancing the analysis of vibroacoustic modulation by machine learning. Composite Structures 273:114233. DOI 10.1016/j.compstruct.2021.114233. https://doi.org/10.1016/j.compstruct.2021.114233
  9. Samaitis V, Yılmaz B, Jasiūnienė E (2023). Adhesive bond quality classification using machine learning algorithms based on ultrasonic pulse-echo immersion data. Journal of Sound and Vibration 546:117457. DOI 10.1016/j.jsv.2022.117457. https://doi.org/10.1016/j.jsv.2022.117457
  10. Solodov I, Kornely M, Philipp J, Stammen E, Dilger K, Kreutzbruck M (2023). Linear vs nonlinear ultrasonic testing of kissing bonds in adhesive joints. Ultrasonics 132:106967. DOI 10.1016/j.ultras.2023.106967. https://doi.org/10.1016/j.ultras.2023.106967
  11. Attar L, Ech Cherif El Kettani M, Leduc D, Predoi MV, Galy J (2023). Detection of kissing bond type defects and evaluation of the bonding quality in metal/adhesive/composite structures by a wavenumber-frequency insensitive SH mode. NDT & E International 137:102841. DOI 10.1016/j.ndteint.2023.102841. https://doi.org/10.1016/j.ndteint.2023.102841
  12. Brunetti J, Khor W, D'Ambrogio W, Fregolent A, Ciampa F (2024). Contact acoustic nonlinearity and local damage resonance for the detection of kissing bonds in structural adhesive joints. Journal of Sound and Vibration 573:118202. DOI 10.1016/j.jsv.2023.118202. https://doi.org/10.1016/j.jsv.2023.118202
  13. Momm GG, Fleming D (2022). Estimate residual strength of degraded bonded joints by combining analytical models with non-destructive evaluations. The Journal of Adhesion 99(5):783–802 (online 2022; print issue 2023). DOI 10.1080/00218464.2022.2056027. https://doi.org/10.1080/00218464.2022.2056027
  14. Jiao J, Li Z, Li L, Li G, Lu X (2024). Nondestructive evaluation of adhesive joints using nonlinear non-collinear wave mixing technique. Journal of Nondestructive Evaluation 44(1):4 (online December 2024; print issue March 2025). DOI 10.1007/s10921-024-01142-0. https://doi.org/10.1007/s10921-024-01142-0
  15. Guan R, Li H, Ni W, Huang T, Wang K, Han X (2026). Kissing bond damage identification and evaluation in CFRP-reinforced steel plates using mixed-frequency ultrasonic guided waves. Sensors 26:1531. DOI 10.3390/s26051531. https://doi.org/10.3390/s26051531
  16. Guan R, Zhang H, Deng J, Kang Q, Wang K (2026). Kissing bond damage detection based on the non-reciprocity of nonlinear guided waves in CFRP-reinforced steel plates. Materials 19:1859. DOI 10.3390/ma19091859. https://doi.org/10.3390/ma19091859
  17. Chen W, Liu X, Zhu W, Weng S (2026). Integrated triple-frequency transducer for one-way wave mixing: theory, verification and weak bond detection. Measurement 287:122569. DOI 10.1016/j.measurement.2026.122569. https://doi.org/10.1016/j.measurement.2026.122569
  18. Sawalmeh A, Zitoune R, Crouzeix L, Navarro P, Marguet S, Chandra Shekar A, Hof LA, Villa L, Giasin K (2026). Review on composite bonding repairs: scientific challenges on machining, surface preparation, and structural health monitoring. Composites Part B: Engineering 323:113781. DOI 10.1016/j.compositesb.2026.113781. https://doi.org/10.1016/j.compositesb.2026.113781
  19. Žukauskas E, Smagulova D, Jasiūnienė E (2026). Numerical determination of weak adhesive bonds using ultrasonic guided waves. Computer Modeling in Engineering & Sciences 146(3):1–10. DOI 10.32604/cmes.2026.077492. https://doi.org/10.32604/cmes.2026.077492