3. Defect Types

Metallic materials or their alloys are a class of elementary materials, such as aluminium, steel, titanium and nickel alloys, all of which are crystalline when solid. Given pure metals, some of the important defect types can be point defect, line defect and plane defect (Gilbert 2020). A point defect involves only a single particle (called a lattice point). A line defect is limited to a row of lattice points. A plane defect involves an entire plane of lattice points in a crystal. A vacancy occurs where an atom is missing from the crystalline array, constituting a tiny void in the middle of a solid.

There are four fundamental mechanisms for introducing a point defect into the structure of a solid (Hiroshi 2014; Fang 2018), such as (a) when a particle is missing at one or more lattice sites, a vacancy is attained; (b) when a particle forces its way into a hole between lattice sites, interstitial impurity is attained; (c) substitutional impurities result from replacing the particle that should occupy a lattice site with a different particle and (d) dislocations are unidirectional defects caused by holes that are not large enough to be a vacancy.

When a fraction of the original materials are replaced by impurities, a solid solution can be attained. Alloys are examples of solid solutions. Lattice distortions of the crystalline materials often occur when impurities are added to a solid. Thus, point defects often determine the properties of a material. Point defects can change the mechanical properties, such as strength, malleability or ductility. Dissolving a small percentage of carbon in pure iron (i.e. making it a steel) makes it stronger than iron; however, higher percentages of carbon can make the steel harder and more brittle.

A dislocation mechanism (screw or edge dislocation types) can weaken a metal, as it allows planes of atoms in a solid to move one row at a time. Interestingly, they can also strengthen a metal when work hardened during heating, hammering, cooling, reheating and reworking. In the course of the work hardening process, intersecting dislocations (i.e. when planes of atoms move one row at a time) that impede the movement of planes of atoms are created.

Most metallic materials are polycrystalline in nature (i.e. structure with many crystallites of varying size and orientation), whereas a group of crystals is called grains. Crystal grains in polycrystalline metallic materials deform by slips on specific slip systems. The place where two grains meet is called a grain boundary. The movement of a deformation through a solid polycrystalline tends to stop at a grain boundary. Therefore, managing the grain size in solids is necessary to obtain a desirable mechanical property, and fine-grained polycrystalline materials are usually stronger than coarse-grained ones.

From early aircraft to the most advanced ones, different types of materials are used in the aerospace industry. Metals have been the most preferred materials and served as the primary choice of materials for many years because of its versatile features and properties. Although the use of advanced composites is continuously increasing in aircraft, metallic materials still constitute 45% of the total weight (20% aluminium, 15% titanium and 10% steel) of the Boeing 787 aircraft. Aluminium is exploited in wings and tail leading edges; titanium is primarily exploited on engine parts and fasteners while steel is used in several places including landing gears, leading edge of the wings, engine pylons, hinges, cables, fasteners, etc.. Airbus A350 has a similar material distribution, 20% Al, Al–Li alloys, 14% titanium and 7% steel by weight (Criou 2007).

Defects and its prevention in aerospace materials are uttermost concerns since undetectable flaws can cause catastrophic consequences for aircraft and passengers. The defects can be categorized, from the origin point of view, under four headings: (a) due to manufacturing, (b) during assembly, (c) during transport and (d) during service. This subsection is intended to shed light on the in-service related defects of aerospace materials. Defects during in-service mainly occur because of either inadequate material specification; in other words, inappropriate material choice and operation beyond the intended design parameters (Archer and McIlhagger 2015).

The common characteristic of in-service damage is that they occur unexpectedly, and it might be difficult to predict and diagnose it. Table 3.1 shows the most common causes of failure. The following subsections describe the major defects or failure types encountered in metallic structures.

Disbonds are both manufacturing and in-service defects. They are defined as unplanned non-adhered or unbonded areas within a bonded interface. They can be caused by adhesive or cohesive failure and may occur at manufacturing or at any time during the life of the structure. Disbond is generally a result of misprocessing during manufacturing such as incorrect initial surface cleaning, pressure/vacuum fail or an excessive load of the structure during service. Actually, this type of damage is dependent on the integrity of the adhesive layer and is affected by the presence of manufacturing defects as well as in-service loads. Manufacturing defects include a wide range of misprocesses such as poor surface preparation, contamination, improper curing, inaccurately applied pressure, the geometrical mismatch between the adherends and trapped air/moisture in the adhesive mixture.

Poor surface preparation is one of the leading causes of adhesively bonded joint failures. The creation of porosity and voids can also degrade the adhesive properties, as well as, create stress concentration during loading. During service, disbond can occur due to excessive stress applied at the bonding interface, impact damage, environmental degradation such as moisture ingression and aging of the adhesive layer (US Department of Transportation, FAA-AR-09-4 2009 (Tomblin, Seneviratne and Pillai, 2009)). Disbond can be found in any bonding surface, some of the examples include laminate/metal or laminate/laminate bonding surface, laminate skin to core (honeycomb or foam) bonding surface in sandwich structure and core to the core in multiple sandwich structure. Typical disbonds are shown in Figs. 3.24, 3.25, 3.26, and 3.27.

Delaminations are both manufacturing and in-service defects. In composites, they can be defined as a local failure in the adhesion between two successive layers causing the two layers to separate from one another at that location. Delamination can be caused by misprocessing during manufacturing (such as pressure/vacuum failure) or by impact damage or excessive load during in-service. Delamination is caused due to matrix properties having lower fracture toughness, strength and resistance against inter-laminar shear and transverse tension as compared to the reinforcement constituent. During the manufacturing process, thermal stresses and resin cure shrinkage can give rise to residual/inter-laminar stresses. These stresses can be high enough to cause delamination due to the mismatch between the properties of two adjacent layers (Bolotin 1996).

In-service delamination may be due to temperature cycling, local loads, impacts and fatigue. Failure due to delamination often initiates from free edges, surface defects and stress concentration points, which leads to the loss of overall stiffness/strength and under compressive loading leads to reduced buckling load limits (Vorontsov et al. 1990). Delaminations can also occur at free edges, in which a local three dimensional stress state phenomena plays an important role. Mainly explained by the mismatch of the elastic material properties between two adjacent dissimilar laminate layers, the free-edge effect is characterized by the concentrated occurrence of three-dimensional and singular stress fields at the free edges in the interfaces between two layers of composite laminates (Mittelstedt and Becker 2007). Several delamination cases are shown in Figs. 3.28, 3.29, and 3.30.

Defects in the composites may also occur during other processes, e.g. during drilling (Figs. 3.31 and 3.32), where delamination and extracted fibres are visible (Tyczynski, Śliwa, Ostrowski 2015).

Foreign inclusion is a typical manufacturing defect. It stands for the presence of a foreign body within the composite structure. Generally, foreign inclusions are pieces of material that have inadvertently not been removed from the surface before bonding. Most foreign inclusions in composites are the remains of unremoved pieces of: (a) protective sheets of preimpregnated plies, either paper, nylon or teflon sheets, (b) masking tape, (c) flash breaker tap and (d) peel ply that is an extra layer of synthetic material (glass, nylon or other) which is laid upon the outer surface of the composite during fabrication. This layer is intended to be peeled off in a subsequent bonding step before bonding. Several foreign inclusion examples are shown in Figs. 3.33 and 3.34.

Matrix cracking is both a manufacturing and in-service defect (Fig. 3.35). It is characterized by cracks within the resin cementing the fibres of a laminate structure. It is often caused by mechanical stress lateral to the direction of the fibres, either during manufacturing or in-service. It is usually the first damage to occur when a composite laminate is subjected to mechanical stress such as quasi-static/cyclic tensile loading (Gayathri et al. 2010). Matrix cracking is characterized by damage in one or more layers without reaching the opposite surfaces. It turns into ‘fracture’ when the damage extends through the entire layers of the laminate.

During manufacturing, matrix cracking is often a consequence of thermal and/or residual stresses. It occurs in the through-thickness and transverse directions and can run perpendicular and parallel to the fibres, respectively. The through-thickness cracks occur because of the significantly lower matrix strength/stiffness as compared with the reinforcement and the strain magnification within the matrix pockets. Similarly, transverse cracks appear due to a mismatch between the in-plane Poisson’s ratio of the plies in the loaded and off-axis directions (Abrate 1991). This type of damage usually occurs during service as a result of tensile, fatigue and impact loading and is affected by the polymer matrix (Hu 2007).

Improper curing is a matrix-dominated defect because during the cure cycle the resin changes its physical state from liquid to rubbery to glassy. Incomplete/non-uniform curing creates parts with thickness variation causing residual stress to build up, thus affecting the stiffness and toughness of the composites, whereas over-curing can lead to matrix cracking or burning. Improper cure can also entrap volatiles, which can lead to void formation (Antonucci et al. 2006).

Mechanical processing steps such as drilling, trimming and machining can introduce inter-laminar matrix cracking. Matrix cracking can affect moisture absorption by providing more space for the moisture to penetrate the laminate. It can also act as a stress concentration point that causes the part to fail.

Porosity in composites is a manufacturing defect caused by incomplete gas microbubbles outflow from the part during the process. A part of the microbubbles that should leave the part remains stuck to the layers and/or bonding interfaces and often combine to form discrete bubbles (Fig. 3.36). The discrete bubbles themselves can also combine and lead to macroscopic delamination or disbond. Porosity is generally defined per volume percentage. High porosity level (several percent) critically affects mechanical properties and thus aviation companies tend to severely limit porosity level to low values on critical parts (1.5% or 2% maximum). When the porosity level is high, the individual gas microbubbles often merge into voids. Actually, porosity and voids affect compressive, shear and bending strengths, which are dominated by the matrix properties. They also act as stress concentration points and possible damage initiation sites (Rubin and Jerina 1995).

Fibre breakage occurs when the applied stress is greater than the fracture strength of the fiber. Typically, failure due to fibre breakage occurs in steps. First, the fibre with an existing damage fails, which in turn, increases the load concentration in the surrounding intact fibres. Since the load distribution is now higher for the remaining fibres, the next weakest fibres tend to fail. This process will repeat until the entire structure fractures. Fibre breakage is generally caused in-service due to foreign-object impact, lightning strike, applied load, erosion, scratches and abrasion.

Apart from the defects mentioned above, composite laminate may have other defects such as non-uniform laminate thickness and wrinkles or waviness. Non-uniform laminate thickness is a typical manufacturing defect generally related to non-uniform pressure or moulding tool misplacement during fabrication. An example of non-uniform laminate thickness in radius is shown in Fig. 3.37. Wrinkles (waviness) is also a typical manufacturing defect generally related to uncontrolled layer displacement during processing. Wrinkles can significantly affect the mechanical properties of composite laminates. A typical laminate wrinkle in a radius is shown in Fig. 3.38. Waviness can be formed during the cool-down phase of composite manufacturing due to a mismatch in the thermal expansion between fibres and matrix and/or composite part and the tool plate (Kugler and Moon 2002). Also, during the fabrication stage, the lamina oriented in different directions attempts to slip with respect to one another. If this slippage is restricted by the neighbouring plies, waviness can form (Vanclooster et al. 2009). Because of waviness, shear stresses can be induced because of the offset angle between the fibres and the applied load. This shear stress produces shear strains, which leads to greater misalignment of fibres. Therefore, waviness reduces the compressive strength of a composite (Wisnom 1993).

Apart from honeycomb to laminate disbond mentioned above, honeycomb cores may also suffer additional defects (see Figs. 3.39, 3.40, 3.41, and 3.42) that may occur both at manufacturing or in-service steps as defined: (a) honeycomb to honeycomb disbond in multicore structures, (b) condensed core that is a lateral distortion of the size of the cells relative to their axis, (c) buckled core that is a general lateral shift of the cells relative to their axis, (d) crushed core that is a collapse of the core parallel to the cell axis and (e) torn core that is characterized by cell breaking in a perpendicular direction to the cell axis, and this defect is often related to excessive core crushing.

Apart from foam to adhesive disbond defects mentioned above, foam cores may also experience other types of defects that may occur (see Fig. 3.43) both at manufacturing and in-service steps as defined: (a) core cracking that is generally related to excessive stress applied to the core and (b) core subsurface cracking that can be considered as a typical foam disbond defect. Since at bonding surface involving foam cores, an adhesive is always showing some slight penetration within the foam (less than 1 mm in thickness), the disbond appears within the foam just below the adhesive penetration line rather than at the foam limit line.

The material properties of fibre composites are also influenced by moisture and temperature. The effect is notably large at higher temperatures and moisture. Figure 3.44 shows the effect on the material properties obtained on small specimens (T300, fiber volume fraction = 0.6, 914C). The matrix 914C is one of the first generation used in aerospace. However, the general behaviour for different resin systems is similar.

Fatigue is the weakening of a material caused by cyclic loading that results in progressive and localized structural damage and growth. In metallics, the growth of cracks is considered. Fibre reinforced composites shows three phases of damage growth: (a) onset of first matrix cracks. With the growing number of cycles, the crack density increases and gets more compact. Crack density reaches its saturation level when the characteristic crack pattern is reached, (b) formation and growing of delamination and (c) buckling of delaminated areas and degradation until collapse. Different examples of fatigue loaded structures are as follows.

Figure 3.45 illustrates the characteristic crack pattern of a [(0, +45, −45, 90) 2]_S laminate under R = −1 fatigue loading. Although the panel looks quite damaged the residual strength is still 85%. Figure 3.46 shows the influence of small impact damage on damage tolerance. For the static case, the influence of strength is reduced significantly to about 30-40%. However, under fatigue loading, these values are almost constant and only very slightly reduced to 25% strength, which is very damage tolerant. This behaviour is different from metallics where the influence of cracks is smaller for the static case and much larger for the fatigue loading. Figure 3.47 shows the fatigue behaviour of a [0₂,+45,0₂,−45,0,90]_S laminate with an artificial delamination of ø10 mm in the middle of the plate. The stress amplitude is ±400 MPa and the stress ratio R = −1. The number of cycles until collapse is reached at 14.2 × 10⁴. It can be seen that the artificial delamination is almost no growing until collapse. However, there are more delaminations at the edges due to the free edge effects. Figure 3.48 shows the results of a stiffened CFRP panel cyclically loaded by axial compression. The panel was partly artificial pre-damaged by separation of the second stringer by a Teflon layer. In each cycle, the panel was loaded almost by collapse load. One cycle took 7 s. The figure shows thermography images at different cycles until collapse. In the first figure, the artificial separation of the second stringer is visible. In the next steps, other stringers start to separate from the skin. But the artificial separation does not grow until collapse (Degenhardt et al. 2008).

One of the main aspects of applying surface coatings to the substrate is to improve the overall performance of the coated components. This can be achieved through the proper application of coatings technology, their characterization, testing and evaluation. As an example (Fig. 3.49), for corrosion-resistant applications, most aircraft surfaces are painted with a shielding layer consisting of chromatic conversion on the substrate, an epoxy primer and a polyurethane topcoat (typical total thickness of about 50–125μm). The role of epoxy primers is to inhibit corrosion of underlying substrates, whereas the topcoat (polyurethane) is a layer that is resistant to a range of chemicals and weather, largely flexible layer and provides the desired surface appearance. A layer of sealant coat is useful at faying surfaces (surfaces that are in contact at a joint) to help provide the flexibility of the layer and prevent cracking (Tiong and Clark 2010).

Some of the manufacturing defects in paint layer during the application, drying and curing stages may include bleeding (leeching out of the painted layer), sagging (dripping of paint from top to bottom forming wavy appearance), wrinkling (formation of undulating wrinkling film), crawling (inability to form a film), crating (formation of small bowl-shaped depressions), lifting (by layers of coating on existing paint), prolonged drying time (inability to dry) and loss of gloss (caused by severe absorption of undercoat).

Beyond corrosion-resistant applications where paints are most suitable, over the last few decades, the number of coating of film deposition methods for wear protection in high-temperature applications such as chemical vapour deposition, physical vapour deposition, electrodepositing, radio frequency ion source implantation, electron beam implantation, thermally sprayed coating (air plasma, flame spray, high-velocity oxy-fuel), chemical catalytic reduction deposition, vacuum-diffused deposition and plasma arc deposition has evolved (Awang et al. 2020). There are many but some of the materials as a thin protective coating for aircraft applications are polymer, ceramic, diamond-like carbon (DLC), TiN, NiTi, (TiB+TiC)/Ti64, TiAlN, AlCrN, Al₂O₃, WS₂ and stellite.

Considering thermal spray coating technology in the aerospace and defence sector, it is used in a range of applications which includes landing gear parts (bearings, axles, pins), jet engine parts (blades, vanes, combustion liners, compressor casings, nozzles), actuation parts (pistons, pumps, flaps), engine cowling and wing span structures. Thermal spray coatings (e.g. using thermal barrier materials) protect engine turbine blades from high temperatures and ensures high reliability for a long period of time. In landing gear parts, thermally sprayed coatings are replacing hard chromium plating as the preferred coating method to provide improved mechanical performance. Some of the coating materials manufactured using thermal spray coating techniques are metals (e.g. MCrAlY), ceramics (e.g. zirconia, alumina), cermets (e.g. WC-Co) and composites (Gansert 2013).

Taking an example of alumina (Al₂O₃) coatings sprayed onto steel substrates using thermally sprayed techniques, have been used in aircraft parts because of their wear resistance properties (Thirumalaikumarasamy et al. 2014). Figures 3.50 and 3.51 shows some of the manufacturing features and defect in thermally sprayed coatings (air plasma spray-APS and high-velocity oxy-fuel-HVOF). The sprayed surfaces (Fig. 3.50a, b) show that the Al₂O₃ particles have spread and non-melted or semi-molten particles not distinguishable. Polished cross-sections of the sample (Fig. 3.50c, d) show a qualitatively higher porosity for the APS alumina (conventional powder) than HVOF alumina (fine powder). Average coating porosity was measured as 2.8% and 8.8% for the HVOF and APS coatings, respectively (Ahmed et al. 2012). The microscopic investigation revealed pores, voids, un-molten particle, non-bonded inter-splat areas and vertical cracks in splats, including columnar grain structure in lamellae (Fig. 3.51).

All such coatings with underlying substrates require an optimized process parameter for coatings to adhere for a long period of time under service conditions. However, each of the coating deposition methods could induce some manufacturing defects if the process parameters are not appropriately managed. This can include poor adherence of coating to the substrate, high porosity, micro-cracks, non-uniform mechanical properties, through-thickness tensile residual stresses, poor wear resistance (sliding/erosive/abrasive), poor thermo-mechanical behaviour, etc. However, with the advancement in the coating deposition technologies and process optimization and control, it has been anticipated that the coating materials deposition quality measured during manufacturing is critical for enhancing our understanding of future generation applications (e.g. high-speed aircraft). Post-manufacturing coating characterization can include various microstructural characterizations and understanding the role of structure-property relationship through advanced characterisation tools (e.g. nanomechanical testing (Faisal et al. 2014; Faisal et al. 2017), acoustic emission sensor based in-situ mechanical testing (Faisal et al. 2011a), neutron diffraction residual strain testing (Ahmed et al. 2012), tribomechanical testing (Ali et al. 2017)). Overall, the coating tests can include mechanical, physical, chemical and weathering, whereas, the coating evaluation can include evaluation under various environmental conditions (e.g. under normal operation and/or in corrosive and wear conditions). All these testing including testing with sensors have the potential to improve our understanding of the structure-property relationships and failure characteristics of current and future generation coatings or thin films.

In-service defects in coatings can be classified according to the nature of environmental damage (corrosion, wear) and physical damage (thermo-mechanical loading) in aircraft structure. Overall, the testing of the integrity of coatings onto the substrate is extremely important for the evaluation of the coating-substrate system. Considering the extreme loading conditions in which the aircraft operates (more for the military than civilian aircraft), the obligation of very high mechanical strength of coating-substrate system deviates other gradual concerns (e.g. corrosion, wear) (Korb and Olson 1997). However, in all cases, it is important that coating is resistant to corrosion, has high adhesion strength with the substrate or between various layers and minimize creep in high-temperature conditions (Pokluda 2010).

Several coatings have been applied to aircraft engine parts (e.g. compressor blades) to develop multifunctional erosion and corrosion mitigations in harsh in-service conditions (Sun et al. 2020). In this work, the multi-arc ion plating technique was applied for TiN/Ti coating fabrication on the TC4 (Ti-5.60Al-3.07V, wt%) substrate. Note that such nitride coating depends not only on high mechanical strength but also on the complex mix of environmental conditions such as erosion and corrosion. Figure 3.52 shows the morphology and mechanism of pitting corrosion of TiN/Ti coating onto TC4 after 576 h salt spray corrosion). From the Tafel polarization plots in Fig. 3.52b, it was observed that the corrosion potential difference promoted the corrosion cells between TiN and Ti, which were the cathode and anode for electrochemical corrosion (Sun et al. 2020), respectively, with corrosion process schematic shown in Fig. 3.52c. It was concluded that the structural design of multilayer dense coatings can be an improved way to prevent the medium of corrosion solutions.

Common defects in paint coatings during service life could be peeling or flaking, chalking, blistering, rust staining, algae or fungi growth, etc. For each of these defects, it requires a range of preventive and remedial methods that are necessary to rectify any unexpected defect.

Wear or damage of the coated surface due to the impact of small sand particles or large objects (such as bird strike), including indentation of the structure during handling process is what is thought to be the most common reason of inducing surface defects. If the coating layer is brittle, then potentially failure such as peeling or flaking can occur (e.g. Fig. 3.53). Physical damage to the coated surface can include thermal shock and fatigue while operating at various altitudes.

In most cases, the coating and substrate materials have different thermal expansion properties, therefore, a combination of thermal shock and fatigue loading can induce coating failure. During in-service conditions, as an example, the engine does not run constantly; therefore, during every flight, a change in thermo-mechanical loading will cause significant strain mismatch at the interface of the coating-substrate system, leading to thermal fatigue damage to the coatings mainly.

The ductile-to-brittle transition (DBT) is important for the understanding of fracture or failure processes (Milne et al. 2003). In aircraft engine parts with coatings, if the ductile-to-brittle transition temperature of coatings is within the temperature range of the engine start phase, then the thermo-mechanical loading stress can cause failure in coatings. As the aircraft flies in airspace with various moisture contents and potentially a corrosive environment with excessive temperature would have an add on effect (oxidation and reduction of materials) to those effects mentioned above, leading to coating failure.

Coated components in aircraft structures when properly tested and in-situ monitored using sensors during in-service operations, can ensure a satisfactory lifespan before being further used. It is understood that regular maintenance may reduce the occurrence of in-service failures in coatings and associated parts by replacing suspect parts. However with the nature of some of the aircraft parts (e.g. turbine engine, landing gear), it is difficult all the time to inspect thoroughly each element of a given assembly. As such the structural health monitoring (SHM) using advanced sensing tools may be a way forward.

Adhesive bonding of primary aircraft structures has been in use for more than 60 years. The qualification process of new adhesive technology is a protracted process involving fluid immersion testing, high, low and cyclic thermal performance testing followed by static buckling and fatigue testing. Adhesive bonding is for example used in bonding stringers to the fuselage and to the skin to stiffen the structures against buckling. It is also used to manufacture the honeycomb structures used in the flight control structures such as elevators, ailerons, spoilers etc. Defects can be introduced both in manufacturing and during service. Thus, the characterization of defects both at the manufacturing and during service conditions is important. The qualification of an adhesive system is carried out not to assess the static performance but to evaluate the long-term durability. The evaluation is usually in comparison with an existing proven technology.

The aim of defect identification and characterization in adhesive bonds is to understand the mechanisms that lead to the creation of the defect, the effect of the defect on the performance of the adhesive joint and ultimately the damage tolerance of the joint. As with fastener-based joints, where the repair can be performed and the joint performance reinstated to 100% post-repair through a simple replacement of the fastener, a 100% regain of the adhesive joint strength is impossible unless a complete reapplication of the adhesive is performed. Hence, the defect type (e.g. Fig. 3.54) and severity must be identified accurately to assess the residual strength of the adhesive bond.

The adhesive bonding of two substrates is a complex phenomenon involving various fields of study. Several theories have been proposed to explain the bond formation, some of them being, mechanical interlocking, interface layer formation, weak boundary layer formation etc. The defects observed in adhesive bonds can be broadly classified into two categories namely those created during the manufacture and in-service defects. Manufacturing defects are created due to improper fabrication techniques related to substrate preparation and adhesive application. Conversely, service defects can be created due to environmental and operational loading factors.

Manufacturing defects in adhesive bonds are classified into three main types: (a) complete voids, disbonds, porosity. This type of defect occurs because of air trapped within the adhesive during the fabrication stage (e.g. Fig. 3.55), the presence of volatiles, defects in the application of the adhesive or insufficient quantity of adhesive. The mixing of the adhesive and hardener during fabrication can sometimes introduce air trapped within the mixture (b) poor adhesion, the improper bond between the adherends and the adhesives. This kind of defects occurs due to improper surface preparation and impurities being present on the surfaces of the adherends before fabrication and (c) kissing defects caused by the local disbonding of the adhesives and adherends. These are also called zero volume disbonds due to the very small dimensions along with the thickness. The wider category of surface disbonds into which kissing bonds fall into that created by the application of adhesive only on one of the adherends in the fabrication stage. Kissing bonds are the worst case of this.

The definition of kissing bonds varies, and several conflicting statements have been made in the literature. For example, Nagy (1991) defined kissing bonds as contact between two compressed but otherwise unbonded surfaces. Jiao and Rose (1991) defined them as two surfaces in perfect contact with each other, however without any ability to transmit shear stresses. However, this definition is not valid in the case of real adhesive bonds as no two surfaces can be in perfect contact and so can be applied in modelling studies only. They also proposed another method of modelling the joint with a thin layer of liquid present between the two disbonded surfaces. Due to the intimate contact of the surfaces in a kissing bond (e.g. Fig. 3.56), where some or all of the surface asperities on both the surfaces are in contact, despite the complete lack of adhesive strength, they are very difficult to detect with many of the conventional non-destructive evaluation techniques. With this combination of undetectability and severity, the presence and detection of kissing bonds present a significant practical problem in the application of adhesive bonds in primary load-carrying structures.

Problems with curing in terms of temperature, cure time or the improper mixing of the different parts of the adhesive mixture also cause porosity and adhesive cracks within the adhesive layer. Poor cure sometimes adjusts itself as the adhesive cures, though slowly. Voids are sometimes also caused because of the relative motion of the adherends with respect to each other during the cure (Fig. 3.55). The presence of voids has been shown to not have a significant effect on the failure initiation in lap joints (Karachalios et al. 2013). This was because, in most of the joints, the failure strain of the adhesive is quite low compared to the yield strength of the adherend materials used. However, if a rubber toughened epoxy were to be used along with high strength alloys for the adherends, the joint strength is almost proportional to the bonded area (Karachalios et al. 2013).

A brief discussion on the effect of the external contaminants introduced in the fabrication stage is given in the following paragraphs. Surface contaminants can be broadly categorized into two groups, namely organic and inorganic. The surface preparation required before bonding is different in each case, organic contaminants being the easier to remove through a simple degreasing process. Inorganic contaminants, conversely, need to be removed by degreasing, de-smutting and in some cases, deoxidization. Surface contamination tends to reduce the surface free energy. The effect of surface contaminants is studied by Smith and Crane (1980) using different levels of the controlled introduction of contaminants on the adherend surfaces. Contaminants like silicone grease, oils, fingerprints, cigarette smoke residues, mucous, drink residues, lubricating oil were introduced on aluminium adherends which were chromium acid and phosphoric acid anodized. In joints with lubricating oil introduced, no significant reduction in the lap shear strength has been observed at contaminant thicknesses of 10 nm. However, in the case of silicone grease, a significant reduction in strength was observed in the contaminant thickness range of 3.5–20 nm. This is accompanied by a tendency towards a dominant interfacial failure. In the automotive industry, tolerance of adhesive bond strength to surface contamination of adherends has been tested. For example, Minford (1981) evaluated the effect of adherend immersion in lubricant solution on the joint strength. He reported that the presence of the oil did not make any difference till the surface concentration reached 0.95mg/cm². The strength of the bond dropped by around 50% and the failure transitioned from 90% cohesive to more than 40% interfacial adhesive failure. Similarly, the bonds retained their original strength till an oil concentration of 0.62 gm/cm² when exposed to a humid or salt spray environment for 180 days. The effect of the presence of oil has been found to reduce the glass transition temperature (T_G) during the cross-linking process while curing. This has been noticed with an adhesive containing 6% dicyandiamide and with the same adhesive containing CaCO₃ filler. The reduction in TG was not noticed in the case of fully cross-linked adhesive. Similarly, Anderson (1993) studied the effects of HD-2 grease contamination on one of the adherends on the double tapered cantilever beam and butt joint tensile test performance. Hysol EA913 and EA946 were tested and the former exhibited a 50% reduction in the tensile strength at an oil contamination of 400 mg/m², whereas the latter did not exhibit any reduction in strength.

Interfacial degradation appears in adhesive bonds when exposed to environmental factors such as high temperature and humidity. The bond formation across the interface is due to secondary and dispersive forces across. The interface work of adhesion is found to be a function of the interfacial free energies of the adherend and adhesive and the surface free energy. Metals tend to form oxide layers on the surface which are polar. This polar nature attracts water molecules which themselves are polar because of the hydrogen bonds. This leads to a disruption in the interfacial bonding between the adhesives and adherends. The work of adhesion in an inert environment tends to be positive leading to a strong bond, whereas in the presence of moisture, the work of adhesion could be negative leading to a disbond along the interface. In addition, certain metal oxides, such as aluminium oxide, react with water to form hydroxides which exhibit loose adhesion to the metal surface leading to a weak interfacial layer.

The general effect of hot and humid environments on the adhesive bond strength is summarised in Fig. 3.57. It shows the aluminium alloy lap joint strength degradation with respect to time of exposure to humidity. As shown, lower levels of humidity have significantly less effect on the bond strength compared to higher levels. This leads to the argument that there exists a critical concentration of moisture in the adhesive below, in which there will be no joint strength degradation. The dotted line in the plot indicated the strength recovery in a joint exposed to high humidity levels after exposure to lower humidity levels. This was attributed to the failure in the primer close to the oxide layer rather than directly at the interface. The effect of moisture diffusion into the adhesive cannot be reversed and the joint strength degradation is irreversible. Moisture ingress in adhesive bonds can either be through direct absorption and diffusion through the adherends and the adhesive. In some cases, where the adherend diffusion is not possible, the cracks and voids present within the bond line promote the moisture uptake. Moisture usually leads to plasticization of the adhesive and/or changes in the glass transition (T_G) temperature in some cases. The rate at which the moisture diffuses into the bond line depends on the diffusion coefficient which in turn depends on the environmental temperature. Hence, hot and wet environments pose the most aggressive form of accelerated ageing to adhesive bonds.

Moisture diffusion in adhesives is dictated by Fick’s law where the rate of diffusion of water is directly proportional to the square root of the exposure time. Where case II diffusion dynamics are in play, the diffusion rate is directly proportional to the exposure time. This is characterized by a saturated and swollen diffusion front travelling through the pristine polymer. Though Fick diffusion is seen in most adhesives, environmental conditions such as high temperature and high humidity tend to promote case II diffusion. Evidence has also been recorded of bulk adhesives that exhibit Fickian diffusion behaviour exhibit case II diffusion when bonded in a joint (Liljedahl et al. 2009). Once diffusion occurs, the moisture either occupies the free spaces or voids within the bond line as free water or exists as bound water. Bound water is responsible for the volumetric changes within the bond line, leading to residual stresses and undesirable stress concentrations which further promotes case II diffusion (Adamson 1980). The bound water can further be classified into Type I and Type II, in which Type I is responsible for single hydrogen bonds that result in plasticization of the adhesive and reduction in the TG value of the adhesive (Zhou and Lucas 1995; Zhou and Lucas 1999a). Type II is where the water molecules form multiple hydrogen bonds lead to a promotion in secondary cross-linking and thus do not have a significant effect on T_G as that of Type I (Zhou and Lucas 1995). Typically, Type I water can be removed by low-temperature heating, Type II might need heating to a higher temperature (Zhou and Lucas 1999b). Kinloch et al. (2000) have reported that relatively viscous adhesives have difficulty penetrating the asperities and pores within the adherend and so promote the diffusion of moisture along the interface leading to premature rupture of the joint. However, when a low viscosity primer was applied to the adherend before adhesive application, the strength reduction has decreased drastically as the primer fills in the gaps which could be occupied by water. Similarly, moisture absorption in composite pre-pregs before curing is known to cause significant problems in bonding. The absorbed moisture diffuses to the surface of the substrate upon curing thus completely preventing any adhesion along the interface.

Metallic materials can be joined by a variety of methods including welding, brazing, adhesive bonding and mechanical fastening. For aerospace applications, friction stir welding (FSW) is very attractive because of ability to weld butt, lap- and T-joints, the ability to join difficult to weld classical alloys, ability to join dissimilar alloys, possible elimination of cracking in the fusion and heat-affected zone (HAZ), lack of weld porosity, etc. (Campbell 2006; Myśliwiec and Śliwa 2018; Śliwa et al. 2019; Myśliwiec et al. 2019).

Friction stir welding (FSW) is a solid-state joining process. It involves rotating a cylindrical tool with a short protrusion or ‘pin’, which is plunged between two metal plates (Fig. 3.58).

High pressure and shear strain plastically deform and consolidate the workpieces using material extrusion from the front to the back of the tool. The plates are clamped with a sturdy fixture to the backing plate with an anvil piece of hardened steel underneath the path of the FSW tool, counteracting the vertical and horizontal forces arising during welding. The combination of frictional and deformation heating around the immersed rotating pin and at the interface between the shoulder of the tool and the plates leads to the consolidation of the two metal plates as the tool traverses along the joint line. FSW process was invented in the Welding Institute (TWI) in the UK in 1991 and has been researched extensively since then and applied in various fields such as the automotive, marine and aerospace industries, where aluminium alloys are heavily used (Thomas et al. 1991; Huang et al. 2018).

Global trends in CO₂ emission and gas price have attracted extensive attention from the automotive manufacturing industry to produce lighter, safer and environmentally friendly vehicles (Guo et al. 2019). In conventional FSW, a tool consisting of a probe and a shoulder was commonly used. Generally, the diameter of the shoulder is about three times bigger than that of the probe. However, this type of tool is associated with several issues. One is the significant through-thickness temperature gradient because the heat generated by the shoulder is much higher than that by the probe, with the peak temperature developing at the top surface, which affects joint microstructure and properties. Another issue is the generation of the flash and arc corrugation because some plastic material moves out of the weld (Zhang et al. 2012).

To receive good welds of metals using friction stir welding (FSW) and avoid defects of such joints it is necessary to know characteristics of the material flow of alloys in FSW (e.g. Colligan 1999 ) by using a tracer material. Xu et al. (2001) developed two finite element models to predict the material flow during the FSW of Al alloys. Mishra and Ma (2005) described the current state of understanding and development of FSW. Fujii et al. (2006) demonstrated the influence of tool geometry, welding parameters and joint configurations on material flow and temperature distribution in FSW. Vijayan and Raju (2008) outlined the detailed parameters governing the joining process, including the rotation rate, welding speed, axial force and tool geometry. Kulkarni et al. (2018) investigated the influence of the type of the backing plate material on weld quality. Hook defects on AS (advancing side) and on RS (retreating side) of SSFSLW (stationary shoulder friction stir lap welding) joints at different welding speeds have been identified by Wen et al. (2019).

The FSW weld does not demonstrate many of the defect encountered in normal fusion welding and distortion is significantly less. But to consider the influence of all conditions and the process parameters on the final result of joining, it could be shown that the main possible defects (e.g. Fig. 3.59) occur when there is an inappropriate choice of parameters for a given case, especially for welding very thin sheets, and when plastic flow and stir materials in the welding zone are difficult to realize. The welds are created by the combined action of frictional heating and plastic deformation of the joined materials due to the rotating tool. To avoid defects, all important parameters, mainly the speeds and feeds influencing the heat input during welding, must be carefully chosen.