Sticking to the quality rules - in bonded joints

TWI Bulletin, September/October 1993

Gareth McGrath joined TWI's Engineering Department in 1989. He has been involved with adhesives, composites and polymers for the last six years, after an initial career in metallurgy. His move to composite technology began during a research assistantship at Sheffield City Polytechnic where his thesis was devoted to reclamation Prospects for advanced thermoplastic composites. Since joining TWI, Gareth has been developing techniques necessary to explain composite failure. These have included fracture toughness measurement techniques, impact damage analysis, and study of the time dependent response of joints in thermoplastics and composites. He now has a technical co-ordination function for the expanding adhesives activities at TWI.

The importance of quality assurance in manufacturing is well known and accepted, although the level of quality assurance varies from industry to industry. Gareth McGrath looks at the factors which affect quality assurance of bonded structures, while examining possible tools for achieving effective quality assurance.

Defects in bonded joints

There are several possible sources of weakness in adhesively bonded joints: the materials that are being bonded; the adhesive itself; the adhesive-adherend interface and in-service degradation.

Figure 1 shows the layers in a bonded lap joint. The joint strength depends on how the loads are transmitted by each part of the structure, and a weak link leads to premature failure. In a well made joint, the oxide and primer layers rarely fail, but if they do, it is usually referred to as adhesion failure as it occurs between the adhesive and the substrate. Failure through the adhesive (or adherend) is referred to as cohesive failure.

Fig. 1. An adhesively bonded joint: Typical thicknesses:

a) Adherend 0.5mm;

b) Surface oxide 4-300 x 10 ^-6mm;

c) Primer 0.02-0.1mm;

d) Adhesive layer 0.1-1.5mm

Fig. 2. Typical defects found in adhesively bonded joints

Adhesive

The common defects found in adhesively bonded joints are shown in Fig.2:

• Porosity is caused by gases and volatiles in the adhesive;
• Voids may be formed by coalescence of pores, but are more often caused by air entrapment during application of the adhesive, by volatiles or by insufficient adhesive being applied;
• Incorrect cure may occur locally because of contaminants or poor mixing, but it is more likely to occur throughout the whole bondline, because of incorrect formulation or mixing, or thermal exposure;
• Cracks in the adhesive are associated with curing and thermal shrinkage during manufacture, especially with some brittle, high-temperature adhesives.
  

Adhesive-adherend interface

If the adherend surface is contaminated before bonding, or if the adhesive is left too long after mixing, the subsequent bond may be weak or non-existent. A total lack of bond or a disbond of zero-volume at the interface can be verydifficult to detect afterwards and therefore, great care has to be taken when preparing surfaces for bonding. The surfaces should be cleaned, and may be treated to form more stable deposits which reduce susceptibility to environmentalattack.

Some manufacturers apply a primer to the freshly prepared surfaces while others insist that the joints are made within a fixed time. Specific bonding problems occur with the release film or peel ply used when manufacturingcomposites, which is not removed until just before bonding. In practice, it is not uncommon for the release agents used in making the peel-ply to be left on the surface of the composite, thus reducing the strength of the subsequentadhesive bond. Hot-curing epoxy adhesives are superior to cold-curing systems in this instance as they can dissolve some of the surface contamination.

Fig. 3. Possible defects in an adhesively bonded honeycomb structure

Bonded honeycomb structures

A major use of structural adhesives is in bonding sheets to a honeycomb core to form a light but very stiff structure, and particular defects occur here, as shown schematically in Fig.3. The first is when the fillet of adhesive is not properly formed and the second is when there is a complete skin-core disbond which can be caused by lack of adhesive or a gap between the skin and core caused either by adefective skin or a locally crushed honeycomb core.

In-service defects

Adhesively bonded joints are particularly susceptible to attack by aggressive chemicals and moisture. ^[1] The adhesive may be degraded so that its strength and stiffness are reduced although its ductility and toughness may be improved. However, the worst problem is degradation of the adhesive-adherend interface, shown in Fig.4. Not only is the interfacial strength reduced, but the surface of the adherend can be weakened by corrosion. It is for this reason that specific treatments have been developed, such as phosphoric acid anodising for aluminium.

Fig. 4. Degradation of adhesive bond: moisture penetrates adhesive and weakens the bond between the adhesive and surface oxide

Fig. 5. Contingent factors for adhesives uses

Quality assurance

The descriptions above show that there are several aspects of preparing and using bonded structures which must be borne in mind when specifying adhesives. These criteria are summarised in Fig.5.

Successful quality assurance procedures should be based on two principles. Firstly, quality assurance aspects must be introduced at the design stage and control maintained throughout the development phases to enable quality assurance of the finished product. Secondly, the adhesive must not be considered in isolation, rather the adhesive system must be evaluated to include adherends, dispensing, stress analysis, quality and process specification.

The quality assurance programme should be conducted concurrently with the bonding procedures to assure satisfactory product strength and durability. Materials and materials handling should meet the requirements in the applicable specifications including proper storage and shelf life stipulations. Processing steps should be carefully followed with the emphasis on maintenance of properly prepared surfaces and adequate control of the bondline thickness. Processing procedures should be checked at critical steps to assure conformance.

Joint design and adhesive selection

Joints should be designed to ensure a favourable stress distribution, ^[2] with the mechanical properties of the adhesive specified in conjunction. ^[1] It is important to use test methods which are relevant to in-service and production conditions. There is a summary of these test methods in a recent review by McGrath and Girardi. ^[3] Testing is important and a variety of tests should be considered to:

• Compare the mechanical properties of a group of adhesives;
• Act as a quality check for a batch of adhesives;
• Check the effectiveness of surface pretreatment(s);
• Measure the environmental effects;
• Determine quantitative mechanical and physical data on adhesives for structural design and analysis;
• Analyse fracture mechanisms.

Data from adhesively bonded joint tests are generally qualitative, reflecting the mechanical properties of the adhesive, the stiffness of the joint, and the degree of adhesion and the effectiveness of surface treatments. Many standard test procedures are listed by ASTM, BSI, ( e.g. BS 5350) DIN and other official bodies.

Production line design

In designing a production line, the emphasis must be on ensuring quality of the product, and previous attention to material selection will be of benefit here in ensuring that the joint is optimised. All material suppliers should be subject to quality assurance procedures, and specifications must be appropriate for the bonding process. For example, passivation procedures used for coated steels are not compatible with good bonds.

Surface preparation requirements must be laid down in light of the durability requirements and must be simple and reproducible. Monitoring of surface preparation can be through one of the following techniques: optical; surface wettability; surface analysis.

As stated earlier the adhesive system must be selected, not just the adhesive. Attention to this detail at the appropriate point in the design cycle will give confidence when dispensing the adhesive. For example, there may be more than one grade of an adhesive available and attention at the selection stage will assist production and so aid manufacture. Correct selection of the grade will enable the optimum quality of adhesive to be applied in the right place, at the right time.

When a two part adhesive is used, strict control must be applied to ensure that the required ratio is obtained, and it is worth considering adhesives that are packaged in cartridges to overcome operator error.

Having placed the correctly mixed adhesive in the right place at the right time, adequate clamping of joints must be used during cure, and variations in the jigging must be monitored if adverse effects are to be avoided during cure. The cure cycle is of course critical and any variation from the process specification should be avoided. Temperature variations, excessive humidity or condensation are all parameters which may cause a loss in bond performance, especially with moisture sensitive materials like polyurethanes. However, exceptionally low humidity may adversely affect the cure speed of cyanoacrylate adhesives.

Finally, conditions of work must be appropriate to the task, with ventilation, temperature and humidity taken into account when developing the process specification.

Manufacturing aspects

Training is a basic issue and while adhesives can be tolerant of operator error, use of untrained people will eventually cause problems, so control of fabrication procedures must start with education and training of the personnel involved. It is vital that the workforce possesses at least a qualitatively correct overall picture of the importance of the different stages of bonding, and of the Health and Safety considerations.

Control of the working environment in terms of temperature and relative humidity is important, together with general standards of cleanliness. The time elapsed between surface pretreatment and primer or adhesive application should generally be kept as short as possible to minimise contamination. Adhesive mixing, dispensing and application must be controlled carefully.

Automatic dispensing is often preferred, or alternatively use of products in special forms or else packed in suitable containers, so that weighing or measurement is not required. This also goes back to knowledge of the workforce.

Insufficient adhesive results in poor joint performance and durability problems, whilst an excess hampers assembly and necessitates time consuming cleaning operations. Joints should then be closed fully to minimise inclusion of air, whilst with close fitting parts over large bonded areas, a means of air escape must be provided.

Inspection carried out during manufacturing should be based around: visual inspection; coupon testing; sample component testing to destruction.

Non-destructive testing

NDT is obviously attractive, and the commonly used methods are summarised in the Table , which indicates which technique can be used to detect each type of defect.

Ultrasonic methods

Ultrasonic testing is the most widely used technique for NDT of adhesively bonded joints. The test may be carried out either with a single transducer in pulse-echo mode or with two transducers in through-transmission mode.

Ultrasonic bondtesters

A number of commercial bondtesters operating in the frequency range between 100kHz and 1MHz are available. The Bondascope®, manufactured by NDT Instruments, measures the magnitude and phase of the ultrasonic impedance of a bonded structure and displays the result on an oscilloscope screen. Changes in the magnitude and phase of the impedance can be related to disbonds. The Fokker Bond Tester MK II uses a spectroscopic approach where frequency and amplitude changes are monitored in the first two modes of through-thickness vibration of a system comprising a transducer coupled to the structure. Changes in resonant frequency reveal the existence and depth of disbonds.

NDT for adhesive bonds

Radiography

An X-ray source is used to produce a radiograph of the bonded structure and this technique is generally only applicable to substrates which are readily penetrated by X-rays. Disbonds or discontinuities may be revealed by the presence of previously applied penetrating liquids which are opaque to X-rays.

Conventional
Disbonds in adhesive joints made with metal adherends cannot be detected by X-radiography since absorption by the adherends is high.

Penetrant-enhanced
Many penetrants are highly toxic, and for the penetrant to be introduced the defect must be surface-breaking. It is therefore likely that penetrant-enhanced radiography will remain a laboratory tool.

Low-frequency vibration

A variety of low-frequency vibration techniques is used in NDT. The methods can be split into global techniques, in which the integrity of the whole structure is assessed using excitation and measurement at a single point of the structure, and local techniques, in which measurements are taken at several points. The local methods may be further divided into those requiring only a single excitation point and those needing excitation at each measurement point. Global methods offer the possibility of very rapid testing, but unfortunately they are not very sensitive.

The coin-tap test, in which the region of structure to be tested is tapped with a coin, the inspector listening to the resulting sound, is probably the best known local method. Defective areas sound duller than defect-free areas, and Cawley and Adams have carried out a major study of the technique. ^[5] It was found that good and defective structures can be differentiated by tapping the structure using a small hammer with a force transducer incorporated into the head, and monitoring the force input to the structure during the tap. This has led to development of an instrument which will perform the test automatically and give an objective assessment of the joint.

In conclusion

To ensure the integrity of adhesive joints, tests must be carried out before and during bonding to ensure that surface preparation is satisfactory, and there must be very careful attention to the process quality control. Both the adhesion strength of a joint and its susceptibility to environmental attack are governed by the quality of the adhesive-adherend interface layer.

^References