Kissing bonds — zero-volume interfacial defects in which adhesive and adherend touch but transfer little load, e.g. through a thin contaminant layer — are nearly invisible to conventional ultrasound because they transmit sound almost perfectly. Yan, Drinkwater and Neild (2010) showed that finite-amplitude ultrasound generates higher harmonics at such imperfect contacts, and that a nonlinear-interfacial-stiffness model predicts the response quantitatively for compressively loaded fracture-surface bonds, while contamination-layer bonds were detected only qualitatively. The open problem is to turn this principle into a quantitative inspection capability for contamination-layer kissing bonds: the defect and its interfacial condition must be measurable against an independently known ground truth, not merely observed to produce a nonlinear response qualitatively, while accounting for the multilayer transfer characteristics that bandpass-filter the interface-generated harmonics. Related open challenges identified in the literature include pulse-echo (single-sided), scanning-capable implementations; the separation of interfacial from adhesive bulk nonlinearity; statistically quantified detection thresholds and false-positive rates across contaminant types and distributions; and the robustness of the measurement and post-processing strategy across transducers, frequencies and industrial geometries. Despite substantial partial progress, no solution of this core is known in the literature to date.

Background

Structural adhesive joints are increasingly used in aerospace and automotive structures because they distribute stress evenly and avoid fastener holes, but their certification is blocked by the difficulty of inspecting bond quality. A kissing bond is a zero-volume interfacial defect in which the adhesive and the adherend (the substrate being bonded) remain in intimate contact but transfer little or no load — either because disbonded surfaces are pressed together (a dry, compressively loaded contact) or because a very thin layer of contaminant, such as release agent, oil or moisture, couples them. Because the acoustic impedance (the product of density and sound speed) mismatch across such a contact is small, a kissing bond reflects very little ultrasound and transmits sound almost like a perfect bond, so conventional pulse-echo inspection, which images reflected energy, largely misses it; this insensitivity of linear ultrasonics was established by Nagy, Adler and Sheppard (1992) and quantified by Brotherhood, Drinkwater and Dixon (2003), who found that linear detectability collapses once the contact is closed by even modest pressure.

The nonlinear route exploits the fact that an imperfectly contacting interface is strain-dependent: under a finite-amplitude wave the contact opens and closes ('clapping') and can frictionally micro-slip, so its stiffness varies over the vibration cycle. This contact acoustic nonlinearity (CAN) generates higher harmonics — signal components at integer multiples of the excitation frequency — that a well-bonded joint would not produce at comparable amplitude. Measuring the ratio of the second-harmonic amplitude to the fundamental, A2/A1A_2/A_1, after transmitting a high-power tone burst (a packet of many cycles of a sinusoid) through the joint therefore provides a defect-specific signature. Yan, Drinkwater and Neild (2009, 2010) demonstrated this in a through-transmission set-up (a transmitting transducer on one side of the joint and a receiver on the other), using the transmitter's naturally limited bandwidth to suppress transmitter-generated harmonics and time-window/area-averaging post-processing to isolate the interfacial contribution. They modelled the interface as a nonlinear spring, i.e. a lumped interfacial stiffness that depends on interface strain, and obtained good quantitative agreement between model and experiment for kissing bonds made from compressively loaded fracture surfaces; however, for the practically important contamination-layer kissing bonds the detection remained qualitative. They also identified a systemic complication: the adhesive material itself and the layer stack of the joint generate and filter harmonics (Rothenfusser, 2001; Yan, Neild and Drinkwater, 2012). In particular, the linear dynamics of the multi-layer bonded system act as a series of bandpass filters on the harmonics generated at the kissing interface, so the measured nonlinearity depends strongly on the adhesive thickness to wavelength ratio; robust measurements require these transfer characteristics to be known and compensated.

The 2010 study demonstrated the principle in through transmission and achieved good quantitative agreement between its nonlinear-interfacial-stiffness model and experiment only for the compressively loaded, fracture-surface type of kissing bond; for the practically important contamination-layer kissing bonds the detection remained qualitative. Achieving quantitative nonlinear ultrasonic detection and characterization of contamination-layer kissing bonds — measuring the defect and its interfacial condition against an independently known ground truth, with robust nonlinear measurement that accounts for the multilayer filtering characteristics which, in the words of Yan, Neild and Drinkwater (2012), 'should be known and compensated for' — defines the problem addressed here. The subsequent literature identifies several further open challenges in converting the demonstrated principle into a practical non-destructive evaluation (NDE) technique: the sensitivity, detection thresholds (minimum contaminant mass or thickness) and false-positive rates of contamination-layer kissing-bond detection have not been quantified for a range of contaminant types and distributions; the approach has only been implemented in through-transmission, whereas field inspection of single-sided structures requires pulse-echo (transmitter and receiver on the same side, as in conventional ultrasonic scanning); the contributions of adhesive bulk material nonlinearity and interfacial nonlinearity to the measured harmonics have not been separated, which limits unambiguous defect characterization; and the robustness of the transmitter natural-filtering strategy and the time-window/area-averaging post-processing across different transducer choices, excitation frequencies and industrial specimen geometries has not been established. These challenges serve one objective: a nonlinear-ultrasonic inspection of adhesive joints that can reliably accept or reject bonds containing kissing defects.

Problem Statement

Establish quantitative nonlinear ultrasonic detection and characterization of contamination-layer kissing bonds in adhesive joints, extending the through-transmission demonstration of Yan, Drinkwater and Neild (2010), in which harmonics generated at the kissing interface were extracted and a time-domain model based on a nonlinear interfacial stiffness showed good quantitative agreement only for the compressively loaded (fracture-surface) type of kissing bond, while detection of contamination-layer kissing bonds remained qualitative. The capability sought is quantitative nonlinear ultrasonic detection and characterization of contamination-layer kissing bonds: the defect and its interfacial condition must be measurable against an independently known ground truth — that is, a reference for the defect and its interfacial condition obtained independently of the nonlinear ultrasonic measurement itself, for example from specimen fabrication or a separate characterization method — not merely observed to produce a nonlinear response qualitatively, while accounting for the multilayer transfer characteristics identified by Yan, Neild and Drinkwater (2012), which act as bandpass filters on the interface-generated harmonics and which, in the source's own words, 'should be known and compensated for' in order to make robust nonlinear measurements. Pulse-echo (single-sided), scanning-capable implementation; separation of adhesive bulk from interfacial nonlinearity; statistically formatted detection-performance reporting (for example, quantified detection thresholds and false-positive rates across contaminant types and distributions); and validated operational best practice are recognized open challenges that a solution may address, but they are not imposed here as acceptance requirements. A qualitative detectability demonstration (for example, with a new detection modality) does not by itself resolve the problem.

References

  1. Dawei Yan, Bruce W. Drinkwater, Simon A. Neild (2010), 'Experimental and theoretical characterization of kissing bonds in adhesive joint using non-linear ultrasonic measurement', AIP Conference Proceedings 1211 (Review of Progress in Quantitative Nondestructive Evaluation Vol. 29), 1190-1197, DOI 10.1063/1.3362191, https://doi.org/10.1063/1.3362191 (source paper)
  2. Dawei Yan, Bruce W. Drinkwater, Simon A. Neild (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
  3. Dawei Yan, Simon A. Neild, Bruce W. Drinkwater (2012), 'Modelling and measurement of the nonlinear behaviour of kissing bonds in adhesive joints', NDT & E International 47, 18-25, DOI 10.1016/j.ndteint.2011.12.003, https://doi.org/10.1016/j.ndteint.2011.12.003
  4. C. J. Brotherhood, B. W. Drinkwater, S. Dixon (2003), 'The detectability of kissing bonds in adhesive joints using ultrasonic techniques', Ultrasonics 41, 521-529, DOI 10.1016/S0041-624X(03)00156-2, https://doi.org/10.1016/S0041-624x(03)00156-2
  5. P. B. Nagy, L. Adler, W. Sheppard (1992), 'Ultrasonic detection of kissing bonds in adhesive joints', NDT & E International 25, 99, DOI 10.1016/0963-8695(92)90550-Z, https://doi.org/10.1016/0963-8695(92)90550-z
  6. M. Rothenfusser (2001), 'Nonlinear ultrasonic effects in adhesive joints', AIP Conference Proceedings 557, 1253-1260, DOI 10.1063/1.1373898, https://doi.org/10.1063/1.1373898
  7. Igor Solodov, Mike Kornely, Jens Philipp, Elisabeth Stammen, Klaus Dilger, Marc Kreutzbruck (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
  8. Do-Kyung Pyun, Lucas W. Koester, Daniel J. Barnard, Leonard J. Bond (2022), 'Nonlinear Ultrasonic Inspection of the Effect of Contaminants on Material Properties of Epoxy-Adhesive', Journal of Nondestructive Evaluation 41, 75, DOI 10.1007/s10921-022-00904-y, https://doi.org/10.1007/s10921-022-00904-y
  9. Bengisu Yılmaz, Elena Jasiūnienė (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
  10. Bengisu Yilmaz, Damira Smagulova, Elena Jasiuniene (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
  11. Jacopo Brunetti, Weeliam Khor, Walter D'Ambrogio, Annalisa Fregolent, Francesco Ciampa (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
  12. R. Houzibe, M. Ech-Cherif El-Kettani, P. Maréchal, E. Siryabe, G. Betchewe (2023), 'Discrimination between weak adhesion, delamination and kissing bond at the interfaces of composite bonded assemblies', Wave Motion 118, 103105, DOI 10.1016/j.wavemoti.2022.103105, https://doi.org/10.1016/j.wavemoti.2022.103105
  13. Jack N. Potter, Jingwei Chen, Anthony J. Croxford, Bruce W. Drinkwater (2016), 'Ultrasonic phased array imaging of contact-acoustic nonlinearity', Proceedings of Meetings on Acoustics 29, 045002, DOI 10.1121/2.0000409, https://doi.org/10.1121/2.0000409
  14. Y. Ohara, S. Yamamoto, T. Mihara, K. Yamanaka (2008), 'Ultrasonic Evaluation of Closed Cracks Using Subharmonic Phased Array', Japanese Journal of Applied Physics 47, 3908, DOI 10.1143/JJAP.47.3908, https://doi.org/10.1143/jjap.47.3908
  15. C. Jeenjitkaew, F. J. Guild (2017), 'The analysis of kissing bonds in adhesive joints', International Journal of Adhesion and Adhesives 75, 101-107, DOI 10.1016/j.ijadhadh.2017.02.019, https://doi.org/10.1016/j.ijadhadh.2017.02.019
  16. M. F. Najib, A. S. Nobari (2015), 'Kissing bond detection in structural adhesive joints using nonlinear dynamic characteristics', International Journal of Adhesion and Adhesives 63, 46-56, DOI 10.1016/j.ijadhadh.2015.08.004, https://doi.org/10.1016/j.ijadhadh.2015.08.004
  17. T. Singhal, E. Kim, T. Kim, J. Yang (2017), 'Weak bond detection in composites using highly nonlinear solitary waves', Smart Materials and Structures 26, 055011, DOI 10.1080/1361-665x/aa6823, https://doi.org/10.1080/1361-665x/aa6823
  18. P. J. Steinbild, R. Höhne, R. Füßel, N. Modler (2019), 'A sensor detecting kissing bonds in adhesively bonded joints using electric time domain reflectometry', NDT & E International 102, 114-119, DOI 10.1016/j.ndteint.2018.11.013, https://doi.org/10.1016/j.ndteint.2018.11.013
  19. Dacho Dachev, Mihalis Kazilas, Giulio Alfano, Sadik Omairey (2025), 'Towards Reliable Adhesive Bonding: A Comprehensive Review of Mechanisms, Defects, and Design Considerations', Materials 18, 2724, DOI 10.3390/ma18122724, https://doi.org/10.3390/ma18122724