A closer look - adhesive bond inspection using NDT

TWI Bulletin, September/October 1994

Ian Munns graduated from the University of Hertfordshire in 1992 and joined TWI later that year. Since then he has been widely involved with the NDT of plastics, composites, ceramics and adhesively bonded joints. Ian is currently working in the Numerical Modelling Section in the Structural Integrity Department.

George Georgiou graduated from Imperial College in Mathematics in 1972, and stayed on a further year as a research student studying theoretical fluid dynamics. He was a full-time mathematics lecturer at Tottenham College of Technology until 1983, gaining his PhD in 1984. Since 1990 he has worked at TWI on a variety of NDT related problems. In particular he is working on NDT of plastics, sprayed coatings and adhesives and is currently involved with BSI and CEN committees which are drafting standards for ultrasonic inspection of welds. He is currently Section Manager of the Numerical Modelling Section in the Structural Integrity Department.

There is great need for a non-destructive means of assessing the integrity of adhesively bonded joints. Non-destructive testing (NDT) methods are available, but the information they are able to provide is far from comprehensive, as Ian Munns and George Georgiou report.

This review examines the nature of defects which can arise in adhesively bonded joints, and then goes on to look at NDT methods which may be suitable for their detection and measurement. It also introduces some of the more recent NDT techniques which are being investigated with respect to adhesively bonded, joint inspection. The merits and limitations of each technique are assessed and discussed in the final section.

Industry's wider exploitation of modern structural adhesives remains hampered by inadequate quality inspection procedures. Stringent checks may be performed as part of the manufacturing process, but the degree of confidence with which bonded joint strength can be stated is limited. Consequently, at the design stage, highly conservative safety factors are incorporated into many load-bearing adhesive joints. Thus, the full potential of adhesively bonded joints is not being realised. This highlights the need for more effort to be directed to developing effective NDT methods.

Defect types and their significance

The most common types of defect in adhesively bonded joints are shown in Fig.l. Voids, porosity, cracking and poor cure are present within the bulk of the adhesive layer and serve to reduce the cohesive strength of the bonded joint. Defects, such as the zero-volume disbond, located at the adhesive-adherend interface tend to reduce the adhesive strength of the bonded joint.

Several NDT methods are able to detect voids, areas of porosity and cracking within the adhesive layer, and also areas of disbonding between adherend and adhesive. These methods are widely recognised and are applied industrially. It is the inability of existing NDT methods to detect poor adhesive strength with any reliability which is inhibiting more widespread use of adhesives.

Reduced adhesive strength can result from poor preparation of the adherend surfaces during manufacture. This type of defect may not immediately affect the strength of the joint and, as such, a poorly prepared joint may satisfy any destructive tests devised to ensure the quality of a batch of joints. However, a joint of this type is especially susceptible to environmental degradation. In a relatively short time, the interaction of moisture, temperature and cyclic loading can cause a drastic reduction in joint durability. Figure 2 illustrates the importance of correct surface preparation on the ability of an aluminium alloy epoxy lap joint to withstand accelerated ageing in water at 50°C. An NDT method which is capable of inspecting the thin interfacial layers at the adherend-adhesive interface is required if the potential of adhesively bonded joints is to be fulfilled.

Methods of non-destructive testing

Several techniques are commercially available for detecting major defects such as disbonds and voids in adhesively bonded joints. Perhaps the most widely used of these is the Fokker Bond Tester. The ability of this device to detect disbonds or voids which are at least as large as the diameter of the probe used is undisputed. However, the MkII Fokker Bond Tester also claims to measure the cohesive strength of adhesively bonded joints. This has been contested ever since the Fokker Bond Tester was introduced in the 1950s and care should be taken when using this device to measure cohesive strength.

Normal incidence ultrasonic techniques offer a popular means of detecting disbonds and voids. Computerised normal incidence ultrasonic systems, with the ability to create C-scan images of the bonded structure, have the added attraction of producing a hard-copy image of the bondline and an indication of the extent of any disbanding.

Other NDT methods commonly applied to bonded structures in an industrial environment are radiography and low frequency vibration techniques. Radiography is limited by the type of adherend and orientation of defect but, in some cases, can provide a reliable means of detecting disbonds and voids. Low frequency vibration techniques are much less sensitive than the ultrasonic techniques commercially available. However, they do not require a coupling medium and their speed and ease of operation make them particularly attractive for detecting gross disbonds in large adhesively bonded structures.

The techniques discussed above, together with the more novel techniques such as thermography, holography and Nuclear Magnetic Resonance Imaging (NMRI), can reliably detect major defects, such as disbonds and voids, within adhesively bonded joints. However, no reliable NDT methods exist capable of detecting poor cohesion or adhesion. Consequently, this area has been the focus of much research.

The Fokker Bond Tester, in certain instances, can provide a measure of cohesive strength. A number of other NDT methods such as NMRI, ultrasonic and dielectric techniques have been used to assess the bulk properties of the adhesive layer and to provide an indication of the cohesive strength of an adhesively bonded joint. However, most researchers recognise that the value they obtain is only a check on the cohesive properties of the adhesive layer. joint strength depends on a number of other factors, such as the way in which the joint is loaded.

It is common practice for operational stress levels in the bondline of critical structures to be low in comparison with the strength that is demanded by mechanical quality control tests. Therefore, in the rare instances where a bond with poor cohesive strength is manufactured, it will be evident from mechanical tests on the associated test coupons. Also, it is likely that a joint with poor cohesive strength will contain readily detectable defects, such as voids and porosity.

It is widely recognised that the main factor inhibiting industry's more widespread use of adhesively bonded joints is the inability of NDT methods to detect poor adhesion. Potential solutions to this problem fall into one of two categories; a dielectric approach, or an ultrasonic approach.

Use of dielectric measurements to inspect adhesively bonded joints is only a relatively recent development. It has shown potential in its ability to detect disbonds and voids but, more significantly, in the way it is able to detect the adverse effects of moisture in the oxide/primer layers. The technique has also demonstrated an ability to detect areas of poor wetting at an adhesive-adherend interface.

However, these results are preliminary, in as much as the technique has only been applied to a small number of substrate and adhesive types. Also, the dielectric technique can only be applied to adhesively bonded joints where both adherends are conductive and where electrical contact can be made to both parts of the joint. Ultrasonic techniques are applicable to both conductive and non-conductive materials, and those techniques that only require access to one adherend surface have successfully been investigated.

Research has shown that those ultrasonic techniques which introduce a nondestructive shear stress at the adherend-adhesive interface are potentially more suited to detecting areas of poor adhesion. Of the techniques based on this principle, it is the interfacial wave technique which is thought to offer significant benefits over other methods. An interfacial wave, by its very nature is more sensitive to variations in quality at the adhesive-adherend interface. Also, by increasing the wavelength of the interfacial wave, so that it is large in comparison with the thickness of the adhesive layer, it is possible to assess the bulk properties of the adhesive itself.

A technique which propagates ultrasound along the length of the adhesively bonded joint in this manner has the potential to inspect a greater proportion of the bondline per test than those localised methods which rely on ultrasound transmitted across the thickness of the joint. Another valuable consideration is the accessibility offered by an interfacial technique.

The stress distribution in a single lap joint is shown in Fig.3. Clearly, the most critical areas are the two ends of the bonded overlap. A defect in either of these areas has the ability to reduce the overall strength of the lap joint considerably. Many of the conventional, through-thickness, techniques cannot reliably detect defects in these areas and as these are the most likely areas for water to enter the bondline or for cracks to initiate, this is a serious short-fall in inspection capability. It is believed that an interface wave method would not be restricted in this manner and could provide a means of assessing the integrity of these critical areas.

Finally, it must be realised that interaction of ultrasound with the various layers of an adhesively bonded joint results in a very complicated waveform. It is here that theoretical mathematical modelling techniques, in conjunction with modern computing facilities, can be extremely useful in gaining qualitative and quantitative information about the nature of this interaction.

This article is a summary of a more in-depth article soon to be published as a TWI members report. ^[1]

TWI currently have a Group Sponsored Project (GSP 5658) which is addressing some of the critical issues discussed above and which is relevant to the aerospace, automotive, power and offshore industries.

Conclusions

• NDT methods exist for detecting major defects such as disbonds, voids and porosity, within the bondline.
• There is no widely accepted technique for detecting poor cohesion and, most importantly, poor adhesion.
• Dielectric techniques for adhesively bonded joint inspection are in their infancy, but show a sensitivity to imperfections at the adhesive-adherend interface.
• Research has shown that those ultrasonic techniques which introduce a non-destructive shear stress at the bondline are potentially able to detect poor adhesion.
• Ultrasonic interface wave techniques are thought to offer significant advantages over other techniques for adhesively bonded joint inspection.

References