Surface Treatment and Repair Bonding

CHAPTER SEVEN

Surface Treatment and Repair Bonding Andrew N. Rider*, David R. Arnott*,a, James J. Mazza† *Defence Science & Technology Group, Fishermans Bend, VIC, Australia † U.S. Air Force Research Laboratory (AFRL/RXSA), Dayton, OH, United States

1 INTRODUCTION Adhesion can be seen as the force or energy of attraction between two materials or phases in contact with each other [1]. In order to achieve intimate contact with or “wet” the adherend surface, the adhesive must flow somewhat like a liquid. Heat or pressure may be applied to facilitate liquid-like behaviour of the adhesive. Once formed, the adhesive bond is expected to carry loads throughout the life of the joint. Although many substances can act as an adhesive, the current chapter is restricted to toughened epoxy adhesives used to bond to metallic or composite materials. The following chapter also focuses on surface treatments and bonding procedures typically employed in application of bonded repairs applied to aircraft structure. Both fundamental and practical aspects of bonded repair application will be described, complementing the work in Chapter 6. The reproducible development of durable bonds is a fundamental requirement for bonded repair technology [2].

1.1 Surface Energy and Wetting The complex interface between an adhesive and a metal adherend is best described as an interphase in which critical dimensions are measured in nanometres. There is controversy over the exact nature of the interactions between epoxy polymers and metal oxides on the adherend [3]. However, some of the earliest studies in practical bonding applications suggested that hydrogen bonds produced by the interaction of polar groups present in the epoxy and metal oxide played a major role in bond formation [4]. Atomic scale modelling has recently been used to examine the complex interactions a

Retired, formerly employed at Defence Science and Technology Group.

Aircraft Sustainment and Repair https://doi.org/10.1016/B978-0-08-100540-8.00007-8

© 2018 Elsevier Ltd. All rights reserved.

253

254

Andrew N. Rider et al.

g lv Vapour

Liquid g sl

q

g sv

Solid

Fig. 1 Balance of surface tensions for a liquid drop on a solid surface.

present at the epoxy and aluminium oxide interface [5] in the presence of water [6] and hydrogen bonding was identified. With the development of new spectroscopic techniques there is also evidence that covalent bond formation at the epoxy-aluminium interface, in the presence of organosilane coupling agents, affects the hydrolytic stability of the adhesive bond [7]. The interactions between an adhesive and an adherend are often described in thermodynamic terms with expressions derived for the case of a liquid drop adsorbed on a flat, homogeneous substrate in equilibrium with the liquid vapour [8]. The balance of forces between the liquid drop, solid substrate and the liquid vapour (Fig. 1) can be expressed in terms of the Young equation [9]: γ sv ¼ γ sl + γ lv cos θ

(1)

where γ represents the relevant surface energy at the three-phase contact point (i.e. solid–vapour (sv), solid–liquid (sl) and liquid vapour (lv)) and θ is the equilibrium contact angle. The value of the contact angle depends on the minimisation of the surface energy. For clean, high-energy surfaces such as aluminium oxide, the surface energy is minimised by the spreading of the relatively lower surface energy liquid, such as water, leading to low or zero contact angles on uncontaminated surfaces. The issues of wetting are complex, particularly in response to chemical inhomogeneity [9], rough surfaces [1,10], capillary forces [11] and the dynamic spreading of viscous liquids [11]. Theoretical considerations indicate that the external pressure to assist the capillary driving pressure and heat (or solvent) to lower the viscosity of the adhesive will aid wetting and penetration [12,13].

1.2 Bondline Pressurisation and Adhesive Cure The structural film adhesives are cured thermally using controlled heating rates. During heating, the adherends are pressurised either mechanically

255

Surface Treatment and Repair Bonding

109

1

Viscosity (Nsm-2)

107 106

Extrapolated

105

0.1

4

10 1000 100 10

0

20

40 60 80 100 Temperature (°C)

120

Plate separation (mm)

108

0.01 140

Fig. 2 Pressurised sandwich panel showing viscosity changes with temperature and consequent calculated plate separation for a typical thermoset epoxy film adhesive.

or hydrostatically. As the temperature is increased, the viscosity of the adhesive approaches a minimum, before increasing as the polymer crosslinking reactions begin [14]. In a pressurised sandwich of two metal plates separated by a film adhesive, the adhesive will flow during the low-viscosity phase and the plate separation will decrease (Fig. 2). A quadratic pressure profile is developed within the adhesive [14]. The local pressure in the adhesive at the centre of the sandwich is higher than the applied load on the plate and can lead to deformation of thin adherend plates loaded by hydrostatic pressure. The thickness of the bondline at the plate edges can be less than at the centre for cures conducted using vacuum bag procedures. The pressure profile also applies a hydrostatic constraint to bubble development in the adhesive and it is not uncommon for voids to develop at the periphery of a repair when using a vacuum bag for bond pressurisation. Structural film adhesives are typically designed for a positive pressure constraint of volatile gases to minimise bubble development.

1.3 Adhesive Bond Performance A strong adhesive bond does not imply a long-lasting or durable bond. Water is the environment most commonly assessed in the literature, although other fluids such as fuel and hydraulic fluid may degrade a bond [15,16]. The current chapter will focus on the critical role of adherend surface treatment on the environmental resistance of an adhesive bond exposed to a humid atmosphere [17]. While much has been written on the subject of adhesive bonding, knowledge is still largely empirically based, and the engineering tools available for the through-life management of adhesively bonded structure rely on

256

Andrew N. Rider et al.

quality assurance processes in place during the initial bond fabrication [18]. While reliable bonding practices can typically guarantee the initial strength of an adhesive bond, the difficulty facing any bonded structure or repair is guaranteeing the long-term environmental resistance. In the case of epoxy bonds to metallic substrates, the displacement of the adhesive bonds by moisture is energetically favoured [19], consequently there is a need, through fundamental experimentation and practical validation, to prove an adhesive bond will maintain the mechanical performance over the intended life in the specific environment. Therefore, research will be presented that will focus on surface treatments for repair bonding, giving consideration to the atomic nature of the bond interface and the relationship between microscopic behaviour and macroscopic mechanical properties. Interfacial chemistry plays a major role in determining the environmental resistance of an adhesive bond.

1.4 Standards and Environments for Adhesive Bonding The facilities, environment, conditions, skills and techniques available for adhesive bonding vary widely. However, it must be emphasised that the quality and long-term performance of an adhesive bond relies on attention to standards and the skill of the technician, together with controls over processes and procedures for all bonding situations. 1.4.1 Bond Integrity and Standards Adhesively bonded components are manufactured, and bonded repairs are conducted, without the benefit of a comprehensive set of effective nondestructive process control tests or techniques to fully assess the through-life integrity of the bonded product. Standard nondestructive inspection (NDI) techniques may be able to detect physical defects leading to voids or airgaps in bondlines, but they cannot detect weak bonds or bonds that may potentially weaken in service. Recently, however, there has been a proof test developed based on shock waves generated by a high-peak power, short-pulse laser, which provides some hope that a localised measurement technique with the ability to consistently verify bond strength will be available in the future [20,21]. In the meantime, the quality and integrity of the bonded component will rely on a fully qualified bonding procedure, together with the assurance that the process was carried out correctly. The Aloha Airlines Boeing 737 incident in April 1988, where the aircraft lost part of the cabin roof in an explosive decompression [22,23], illustrates

Surface Treatment and Repair Bonding

257

the importance of bond durability and more importantly, the ease with which this issue can be overlooked. In the repair environment, experience has shown that some bonded repair designs and application procedures have little chance of success and can, in some cases, decrease the service lives of components [24]. A survey of defect reports conducted at one Royal Australian Air Force (RAAF) Unit [24–26] indicated that 53% of defects outside structural repair manual limits were related to adhesive bond failure. In addressing the standards applied to adhesively bonded repairs, the RAAF [27] have established a substantial improvement in the credibility of bonded repair technology. 1.4.2 Adhesive Bonding Environments The performance of an adhesive bond is sensitive to the adherend surface treatment and the environmental conditions under which the bond is prepared. Facilities located adjacent to operational airbases or in industrial environments need to have concern for the effect of hydrocarbon contamination. Facilities in tropical locations need special consideration for the effect of heat and high humidity. Factory manufacture uses specialised facilities and staff. The facilities will include vapour degreasing or alkaline cleaning, etching tanks, anodising tanks, jigs, autoclaves and appropriate environmental controls. Adhesives will be stored in freezers, and monitoring procedures will be in place. There is a well-trained workforce with skills maintained through production volumes, and highly developed inspection procedures are available. At the other extreme, field repairs are generally conducted with relatively unsophisticated facilities, minimal surface treatments, vacuum bag or reacted force pressurisation and little or no environmental control. Staff multiskilling and rotation influence the currency of experience and hence the quality and performance of adhesive bonds [28]. The requirement for environmental controls, the attention to bonding procedure detail and the need for staff training and supervision are of particular concern. If the use of training measures can be combined with regular monitoring, then any deviation in quality of repairs or bonding operations being undertaken can be identified. At one RAAF repair depot, the ongoing review of wedge test data enabled deviation in standard practices or degradation in application equipment to be identified and remedied [29]. The use of quality control tools can also aid in continued monitoring of processes to improve reliability [30,31]. Depot-level repairs are conducted with facilities and staff skills that vary considerably. Some depots have almost factory-level facilities and high level

258

Andrew N. Rider et al.

of staff skill. Other depots are capable of only low-level bonded repairs and are little removed from a field repair capability. Laboratory experiments are designed to establish knowledge and principles. It is easy to overlook important detail from factory or field experience since most laboratories are held to close environmental tolerances and do not resemble the workshop environment. 1.4.3 Constraints for On-Aircraft Repairs On-aircraft repairs impose additional constraints on processes and procedures. The considerations include: accessibility of the area, limitations in the use of corrosive chemicals, adequacy of environmental controls and constraints on the tools for pressurisation and heating of the bond during cure. Safety, health and environmental issues are more demanding for on-aircraft bonding since it is harder to control, contain and clean-up hazardous chemicals. Constraints on the use of electrical power on fuelled aircraft, or those with inadequately purged fuel tanks, can restrict the range of treatment and bonding methods available. The surrounding aircraft structure imposes constraints on the choice of surface preparation, heating arrangements and pressurisation tools.

2 MECHANICAL TESTS 2.1 Loading and Failure Modes The most common method used to assess the relative performance of an adherend surface treatment involves loading an adhesive joint asymmetrically in tension, as shown in Fig. 3, described as mode I opening. The stresses leading to failure are localised in a region adjacent to the crack tip. The extent of this region depends on the stiffness of the adherends, the toughness of the adhesive and, importantly, the effectiveness of the adherend surface treatment. The mechanical performance of a bond should be accompanied by an inspection of the fracture surface. Visual inspection assisted with optical

Fig. 3 Asymmetric tension or mode I opening of an adhesive joint.

Surface Treatment and Repair Bonding

259

microscopy will provide macroscopic information concerning the locus of fracture and the presence of voids or defects. The term cohesive failure describes fracture totally within the adhesive, leaving adhesive on both separated adherends. The term adhesion failure describes a fracture at one interface with the adherend, resulting in one face having the visual appearance of the adherend material and the mating face with the appearance of the adhesive. Visual inspection alone does not convey the complete picture. Because an adhesive bond is formed as a result of atomic interactions, closer inspection of adhesion failures with surface composition analysis techniques can provide detailed insight into the material causing the weakness at the fracture site.

2.2 Qualification of Bonding Procedures and Performance An adhesive bond represents a complex system of materials, treatments and processing steps. The issue of qualification of the adhesive system is complex since specific requirements depend on the application. The focus must be on mechanical performance and durability because the bonded joint is expected to transfer load for the service life. For structural joints, strength is typically evaluated using shear tests (for static properties and fatigue) and toughness with cleavage tests. For honeycomb structure, properties are typically evaluated with flatwise tension and peel tests. Tests are conducted at representative temperatures experienced throughout the service environment, including the operating extremes. Tests are also conducted using moisture-conditioned specimens to evaluate durability performance. Other conditioning may include exposure to salt fog, SO2, hydraulic fluids, fuels, de-icers, fuel and more [15,16]. Subcomponent or component testing normally follows coupon testing. The failure modes of test specimens are as important as the strength or toughness values obtained. Failure modes at interfaces between the treated metal surface and the adhesive or primer are generally not acceptable. The primary objective is for the mechanical properties of joint to be limited by the properties of the cured adhesive, not the surface treatment. Qualification of the adherend surface treatment procedure is of particular importance. Many surface preparations can provide adequate initial bond strength, however, maintaining this strength for the life of a system in its operating environment is a more difficult challenge. Moisture durability is of primary concern. However, for certain titanium applications, longterm durability at elevated temperature is important.

260

Andrew N. Rider et al.

3 STANDARD TESTS 3.1 Wedge Durability Test The ASTM D 3762 wedge test is usually employed to assess the environmental resistance of adhesive bonds. A crack is initiated in a bonded joint through insertion of a wedge into the bondline (Fig. 3). The test specimen is then typically exposed to hot/wet conditioning and crack growth is monitored. The initial preexposure fracture is expected be cohesive within the adhesive layer, and the equilibrium crack length is therefore expected to reflect the toughness of the adhesive system under dry conditions. An excessive initial crack length accompanied by interfacial failure, even before environmental exposure, reflects a poor surface treatment. Cracks that remain within the adhesive indicate that the surface preparation is not the weak link in the bonded joint. Poor surface pretreatments lead to interfacial failures and long crack lengths. The wedge test is a quality control test employed to compare a surface preparation against a control under the same testing conditions. The wedge test can be misused because ASTM D 3762 is not fully prescriptive. Difficulties in the comparison of published data may occur if test conditions and systems are not fully specified. It should also be noted that tough adhesives place higher demands on the performance of the surface treatment than do brittle adhesives. The testing of tough adhesives introduces the essential requirement to conduct a simple calculation to ensure that the adherends will not plastically deform in cases where fracture energy measurements are made [32]. However, bonded joints that strain significantly when exposed to hot/wet environment may provide a less rigorous test and, generally, the wedge test is not used to provide quantitative data. The wedge test is a severe test, since the adhesive is at its breaking stress at the crack tip while directly exposed to the conditioning environment. For this reason, surface preparations that allow limited interfacial failures may be satisfactory. The RAAF Engineering Standard DEF(AUST)9005 [27] uses crack growth criteria and allows some interfacial failure in relation to one particular tough adhesive based on service histories of RAAF aircraft. However, as a general rule cohesion failure of the adhesive bond should be specified. There is ongoing pressure to establish a relationship between service life and the performance of an accelerated durability test. Although the wedge test has been correlated to adhesive bond service life for limited applications,

261

Surface Treatment and Repair Bonding

similar durability performance for new treatments does not imply similar service lives [33]. With current understanding and the complexity of the bonded joint, there is no straight forward way to accelerate nature to obtain a quantitative correlation [34]. Despite the complexity of relating wedge test performance with field service performance, some recent efforts have examined this relationship. The loading of composite patches bonded to aluminium honeycomb beams under fatigue loading in tropical field trials showed that poor treatments identified by wedge testing failed extremely rapidly, whereas marginal and good surface treatments showed good long-term performance [35]. A long-term trial has also reported on how wedge tests exposed to harsh marine conditions perform compared to tests conducted in humid laboratory conditions [36]. Further examples of teardown investigation of bonded repairs exposed to aircraft service environments will be described at the end of the Chapter (Section 12).

3.2 Fracture Mechanics and the Cleavage Specimen Fracture mechanics has been applied to the cleavage specimen in an attempt to quantify the test. The elastic energy release rate, GI, is the energy delivered from the stressed cantilevers to create a unit area of fresh fracture surface. For the double-cantilever beam specimen [32,37–40]: GI ¼


3E  h3 w 2 1 2  4 Jm 4 16a ð1 + ðλ0 =λÞ0:64ðh=aÞÞ

(2)

where 

3k λ¼ Ebh3

1=4



3k0 and λ0 ¼ Ebh3

1=4 and k ¼

1 ð1=k0 Þ + ð1=ka Þ

(3)

with k0 ¼

2Eb 2Ea b and ka ¼ h ha

(4)

In Eq. (2), h and E are the thickness and modulus of the adherends, respectively, w is the load point displacement and a is the effective crack length. Corrections can be made to the measured crack length to allow for adherend rotation and adhesive thickness, ha, as shown by the λ and λ0 terms, where adherend width, b, is required [40–42]. A small fraction of GI is dissipated in breaking atomic bonds, while the remainder is dissipated as thermal energy as a result of deformation processes in the stressed polymer. Eq. (2) shows

262

Andrew N. Rider et al.

that for a plane double-cantilever beam bonded specimen, G1 decreases as the crack grows since the stress intensity at the crack tip decreases. It is common practice to assess a critical elastic energy release rate, G1c, at some arbitrarily long time where the crack velocity is small. It is now becoming more common practice to use longer and thicker adherends for durability tests [43], primarily to avoid plastic bending and to provide a longer initial crack length, providing a more accurate measure of GI and GIc. To avoid adherend bending, the adherend thickness must exceed a critical value, hcrit given by [32]: hcrit ¼

3:GI  E σ 2y

(5)

where σ y is the yield stress of the adherends.

4 FUNDAMENTALS OF DURABLE BONDING The employment of complex surface treatments to prepare highenergy surfaces (such as metals) prior to bonding is primarily conducted to ensure adequate service life of the joint when it is exposed to aqueous environments. Moisture can gain access to a bonded joint through the adhesive bondline. The surface preparation should ensure that: (1) the adhesive bond is hydrolytically stable and (2) the moisture levels that may exist at the adhesive and adherend interface will not lead to traditional corrosion processes that can form weak hydrated oxide layers that would fail under load. The choice of the surface treatment should consider the nature of the substrate, its initial condition, the type of adhesive to be used and the intended service environment [44]. Surface treatments modify both the physical and the chemical properties of the adherend. In general, the relative contribution of the physical roughness and the chemical character of the adherend to bond strength and durability are not known as it is quite difficult to design experiments to separate these effects. A review of experiments to highlight the relative effect of physical and chemical properties of the adherend on bond durability is described later.

4.1 Surface Roughness and Bond Durability The surface roughness profile can affect the fracture toughness of an adhesive bond when the bond is degraded by exposure to a humid environment

263

Surface Treatment and Repair Bonding

a = 60°

GIc (J/m2)

1000

Grit-blast 100

a = 30° a = 0°

10

0

0.5

1.5

1.0 tan a (°)

2.0

Fig. 4 GIc determined from wedge tests conducted in 95% R. H. at 50°C show the effect of surface roughness profile angle (α.). The Al2024 clad adherends were ultramilled or grit-blasted. The ultramilling profiles produced angles of 0, 30 or 60 degrees.

[45,46]. Wedge tests conducted with aluminium adherends, surface prepared using an ultramilling method, showed that the elastic energy release rate at the slow-crack velocities in a humid environment, GIc, depended strongly on the adherend surface profile angle (α) (Fig. 4). The ultramilling method created either a flat adherend surface with a 0 degree profile angle, or sawtooth profile angles of either 30 or 60 degrees and a peak to valley depth of 10 μm. The surface relief on the 0 degree ultramilled terraces was less than 5 nm (Fig. 5), indicating that mechanical interaction between the adhesive and the adherend was not significantly contributing to the durability of the joint. The flat 0 degree ultramill surface

Ultramilled (a = 0°)

Gritblast

µm

nm 140

1.5 40

40 µm

(A)

20

20 0 0

µm

40

0 40

(B)

20 20 µm

0

µm

0

Fig. 5 Atomic force micrographs of (A) ultramilled (α ¼ 0 degree) and (B) grit-blasted (α  45 degrees) surfaces. The relief on the terraces of the α ¼ 0 degree surface is less than 5 nm.

264

Andrew N. Rider et al.

a = 60°

Shear

a = 0°

d

d a Peel

Fig. 6 The potential peel and shear components of interfacial loading for 0 and 60 degrees sawtooth surfaces used in Fig. 4

provided a baseline for bond durability in the absence of mechanical effects. In this case, load transfer depends primarily on basic chemical interactions occurring at the interface between the adhesive and the flat aluminium adherend. It was hypothesised that surface roughness may introduce an interfacial shear component (Fig. 6) at the adhesive to adherend interface that influences fracture toughness in humid environments. More recent finite element analysis modelling (FEA) of nanoscale roughness suggests that the influence on fracture toughness is a complex combination of factors including the real to nominal surface area ratio and the levels of tooth-tip damage occurring ahead of the crack [47]. The complexity of the surface roughness influence on fracture toughness has also been shown for simple bulk epoxy systems with different levels of roughness [48]. However, FEA of macroscale roughness has predicted that an increase in the aspect ratio and height of the surface features should increase fracture toughness due in part to a crackkinking process [49]. Development of hierarchical roughness at the micron scale dimension using a combined grit-blasting and micromesh embossing technique on aluminium adherends has also been shown to have a significant effect on shear strength [50].

4.2 Surface Hydration and Bond Durability Epoxy resins have a high polarity which provide strong hydrogen bonding attraction between epoxy molecules and metal oxides [51]. DeBruyn [4] showed that the nominal breaking stress of an aluminium-epoxy single lap joint depended strongly on the hydroxyl content of the epoxy. Recent theoretical modelling has also shown that hydrogen bonds formed between the epoxy hydroxyl groups and the aluminium oxide were responsible for the main adhesion force [5]. This observation leads to the expectation that the interfacial strength of an adhesive bond would similarly depend on the

Surface Treatment and Repair Bonding

265

hydroxyl content of the surface oxides on the adherend. Plasma oxidation experiments were conducted on ultramilled-aluminium adherends [46,52] to systematically change the hydroxyl concentration of planar γ-alumina films from very low levels to concentrations expected for pseudoboehmite-type hydrated oxides. These experiments demonstrated that the environmental resistance of the adhesive bonds was independent of the oxide hydroxyl concentration. Derivatisation experiments on the epoxy adhesive suggested that the density of metal hydroxyl group was always in large excess of the active epoxy groups. It was also concluded that changes in a metal adherend surface roughness played an important role in the environmental resistance of the bond determined from mode I tests. Recent atomistic-scale modelling of epoxy and aluminium oxide interfaces in the presence of water has suggested that, whilst the presence of water alters the hydrogen bonding interactions, the adhesion force is not changed [6]. These results indicate the influence of water at the interface of epoxy-aluminium bonds is complex and many of the macroscopic imperfections in practical joints need to be considered when developing predictive degradation models. It is important that the surface oxide is cohesively strong, that is the oxide should not fail or separate from the metal surface. It is well known that hydration and growth of the oxide in water can lead to weakness of the oxide structure and that these conditions should be avoided. In the plasma experiments, planar cohesive γ-alumina films were formed and the contributions of a weak oxide were avoided [46,52].

4.3 Surface Contamination and Bond Durability It is universally acknowledged that an unprepared surface covered with thick layers of hydrophobic contamination leads to a weak adhesive bond with very poor long-term durability. Reduction of the contaminant concentration will ultimately lead to adequate initial bond strength, limited by fracture in the adhesive, but the long-term durability may still be poor due to the overriding influence of environment-induced failure at the adhesive to adherend interface. The adhesive bond durability is very sensitive to the presence of hydrophobic contaminant on the adherend, but the dependence involves a complex combination of the nature, the concentration and the distribution of the contaminant. Studies of bond durability with one epoxy film adhesive following deliberate contamination of prepared aluminium adherends showed sensitivity to the nature of the hydrocarbon contaminant [53]. The durability was remarkably tolerant to contamination with aviation kerosene and a homologous

266

Andrew N. Rider et al.

series of alkanes of lower chain length than C16. This suggested that the adhesive was capable of displacing sufficient area of some surface contaminants for the adhesive to make good bonding attachment with the adherend. It is expected that the durability response to the nature of the contaminant will be adhesive specific as it is well known that some adhesives are formulated for application to grossly soiled surfaces [54]. In some surface treatments, an aqueous organosilane coupling agent is applied to improve bond durability performance [53]. The ability of this organosilane coupling agent solution to wet an adherend surface is very sensitive to the presence of contaminant. Aviation fuel contamination before organosilane application leads to a marked reduction in adhesive bond durability, whereas contamination after the organosilane is dried has minimal effect on durability (Fig. 7) [55]. Hydrocarbon contaminants are not uniformly distributed over the surface. Angle resolved X-ray photoelectron spectroscopy (XPS) studies show that hydrocarbon is distributed as islands on the surface [55]. Some surface treatments will accentuate the island distribution of residual contaminant, whereas others will lead to a more uniform distribution. Abrasion and grit-blasting processes roughen the surface, but an estimated 5–10 atomic layers of residual contaminant remain on the surface, Time (h) 100

0

25

100

GB + Avtur dip + SCA

90 Crack length (mm)

400

GB (band)

80 70 60

SB + wipe(MEK) + SCA

50

SB + wipe(water) + SCA

40 GB + SCA (band)

30

GB + SCA + Avtur dip 20 0

5

10

15

Root time

(h0.5)

20

25

Fig. 7 Wedge test results conducted at 95% R. H. and 50°C for 2024 clad aluminium bonded with FM 73 adhesive showing the effect of abrasion with Scotchbrite (SB) or grit-blast (GB), cleaning with Methyl Ethyl Ketone (MEK) or water and fuel (Avtur) vapour exposure, prior to organosilane coupling agent (SCA) application.

267

Surface Treatment and Repair Bonding

76⬚

30⬚

Fig. 8 Contact angle between a 5 μL water droplet and a polished aluminium surface ultrasonically degreased in high purity MEK is 30 degrees and for MEK containing tissue plasticisers is 76 degrees.

distributed nonuniformly [17,55]. By contrast, wiping the surface with a solvent-soaked cloth or tissue will spread solvent containing dissolved organic material uniformly across the surface. The solvent evaporates leaving the surface uniformly covered with hydrophobic contaminant. A contact angle experiment conducted on polished aluminium, as shown in Fig. 8, illustrates the importance of considering the solvent, and the tool used to apply it, as a source of unwanted contamination [56]. The contamination deposited by the solvent can have a dramatic effect on bond durability, as illustrated in Fig. 9. Here, clad 2024 aluminium alloy Time (h) 25

Crack length (mm)

120

100

400

SB + wipe(MEK) SB + wipe(water)

100

GB

80 60

GB + SCA 40 PAA 20

0

5

10

15

20

25

Root time (h0.5)

Fig. 9 Wedge test results conducted at 95% R. H. and 50°C for 2024 clad aluminium bonded with FM 73 adhesive, showing the effect of abrasion with Scotchbrite pads (SB) or grit-blast (GB), cleaning with methyl ethyl ketone (MEK) or water and organosilane coupling agent (SCA) application compared to phosphoric acid anodisation (PAA), representing a factory benchmark treatment [57].

268

Andrew N. Rider et al.

was surface prepared by abrasion with Scotch-Brite abrasion pads, followed by debris removal, or grit-blast, then bonded with FM 73 adhesive. The debris removal with aerospace wipes soaked in methyl ethyl ketone (MEK) leads to very poor durability, whereas the debris removal with wipes soaked in water lead to durability approaching that of the grit-blast. Composition analysis of the failure surface (Fig. 10) shows that the very poor durability of the solvent-wiped adherend is associated with weakness at the adhesive to metal oxide interface due to the contaminant. The waterwiped and grit-blasted surfaces show the presence of aluminium oxide on both fracture surfaces.

SB + wipe (MEK)

SB + wipe (water)

Surface concentration (atom%)

80 60

Adhesive face

C

40

C

Adhesive face

Al

20 Al 0 20 40 Metal face

60

Metal face

O

O

Adhesive

Adhesive

Al - oxide

Al - oxide

Al

Al

Fig. 10 Surface composition of both fracture faces measured with XPS for the treatments (i) Scotch-Brite abrade plus wipe with MEK soaked tissues (SB + wipe (MEK)) and (ii) Scotch-Brite abrade plus wipe with water soaked tissues (SB + wipe (water).

269

Surface Treatment and Repair Bonding

4.4 Bond Durability Model A stress-based diffusion model (Fig. 11) was used to help describe the bond degradation processes and fracture for a series of aluminium/epoxy wedge test specimens exposed to humid environments [46,58,59]. Disbonded regions were observed along the interface of wedge test specimens exposed to humid environments [60]. Observations by other workers [61] and scanning electron microscopy (SEM) micrographs (Fig. 12) indicate that microcavities may develop at the interface in the crack-tip region for stressed wedge test samples. Under stress, the size and distribution of these microcavities could be determined by variations in bond strength or hydrolytic Polymer hydrolysis

Moisture ingress

Polymer desorption

Oxide degradation

Adhesive Moisture ingress Oxide Metal

Weak bond region

Strong bond region

Fig. 11 Stress-based diffusion model that describes moisture ingress to the adhesive metal interface and three possible degradation reactions. The magnitude of microcavities ahead of the crack tip could control the rate of moisture diffusion and the dominant degradation reaction path could determine the position of bond weakening and the dominant locus of failure.

270

Andrew N. Rider et al.

Unstressed bond

Humid interfacial fracture

Initial dry cohesive fracture

3 µm Voids at the stressed interface ahead of crack

4 µm

400 nm

Fig. 12 SEM micrographs showing disbonded areas in the stressed regions just ahead of the crack tip of a wedge test specimen exposed to 95% R. H. and 50°C.

stability along the interface. The variations in bond strength may be caused by the distribution of contaminant and when combined with inhomogeneous stress distributions induced by surface roughness may influence moisture diffusion rates and concentrations along the interface. Firstly, all practical adherends will have several atomic layers of residual contaminant distributed nonuniformly on the surface [55]. During cure, adhesive bonds will form more efficiently in the uncontaminated surface regions. Increased concentrations of contaminant may lead to less effective bonding. This has the capacity to explain changes in the interfacial moisture diffusion and concentration levels with changes in contaminant levels.

Surface Treatment and Repair Bonding

271

Inhomogeneous stress distributions induced by surface roughness have also been used by several workers to describe adhesive joint fracture [62,63]. At a microscopic level, surface roughness may introduce a surface shear component to the existing surface peel component in a normally stressed bond and thereby would change the effectiveness of the interfacial transfer of load. This may potentially influence the size of microcavity formation. Microcavities represent paths for moisture ingress. The size and distribution of microcavities will determine the ease of moisture access and the fractions of the adherend surface subjected to rapid bond degradation. At these localised sites of moisture ingress, the adhesive bond may degrade by at least three reaction paths, as shown in Fig. 11, polymer degradation, polymer desorption or oxide degradation. With time, the sites of adhesive bond degradation expand and coalesce, leading to a reduction in the fracture strength and crack propagation [45,46]. Moisture in contact with the metal oxide at the bond interface will lead to oxide growth and hydration. The presence of localised sites of oxide growth some hundreds of microns ahead of the crack tip is illustrated by the Scanning Auger Micrographs of the fracture surface of a forcibly opened wedge test specimen with flat, ultramilled adherends (Fig. 13). This supports the concept that bond degradation occurs in advance of the crack tip. Mechanical strength may be sustained by stress sharing networks in the polymer bridging load across the degraded regions of the interface. The stress-based diffusion model may aid in developing a more comprehensive view of adhesive bond durability and its dependence upon interfacial stress, contaminant concentration, surface roughness, coupling agent performance and other relevant influences. An opportunity exists for modelling the micromechanics of the bondline and moisture diffusion behaviour. Finite element models have begun to be applied to the influence of surface roughness on mode I fracture toughness [47,49] and more comprehensive models incorporating moisture effects have also been developed to aid in bonded joint design [64–66]. Comprehensive studies on the factors affecting organosilane treatments have also been undertaken and indicated that there may be a diffusion zone near the crack region that comprises an aluminium oxide and organosilane phase where failure may begin [67].

272

Andrew N. Rider et al.

Direction of crack Crack tip Metal failure surface

Region behind tip Region behind tip

2 mm

A.

Max.

Min. 30 µm

30 µm

30 µm

Oxide

1370 1390 1410 Kinetic energy (eV)

Counts (Arb.)

Counts (Arb.)

B.

Oxide

Metal

1370 1390 1410 Kinetic energy (eV)

Fig. 13 Scanning Auger Microprobe (SAM) analysis of the crack-tip region of a failed wedge sample pretreated with a 180° ultramill. The SAM maps indicate the distribution of metallic aluminium at positions ahead of the crack tip and the spectra show the relative concentrations of aluminium oxide and aluminium metal at the locations indicated.

5 REQUIREMENTS OF SURFACE PREPARATION Adherend preparation typically involves a series of steps, each with an important purpose. While the surface preparation affects the initial bond strength, it is the influence of surface preparation on the long-term environmental durability of bonded structures that is of particular concern.

Surface Treatment and Repair Bonding

273

5.1 Degreasing Degreasing is conducted to remove gross organic contamination. Greases and oils are usually present on alloys received from the manufacturer. After prolonged service, components are expected to have a wide variety of organic contaminants, with some so aged and firmly adsorbed that they are resistant to solvents. Degreasing is used to decrease the concentration of these organic contaminants on the adherend surface prior to subsequent preparation steps. Factory facilities often use vapour degreasing to reduce the concentration of organic contaminant on components to be bonded. A solvent such as trichloroethylene is evaporated in a closed space then allowed to condense and drip from the soiled components. Organic contaminants are slowly dissolved in the liquid phase, whilst it is in contact with the component and are transported to the solvent reservoir under gravity. In many cases, tank-based cleaning using alkaline aqueous solutions has replaced vapour degreasing due to environmental and safety concerns [68,69]. Degreasing is typically the only surface preparation used for aluminium honeycomb. Due to the benefits of solvent vapour degreasers, they may still be used in some specialist applications [70]. In the repair environment, degreasing is frequently conducted with solvent-soaked tissues or cloths. Common solvents used are MEK and acetone. Again, the solvent dissolves the organic contaminant on the soiled component, but it is important to ensure that sufficient solvent volume is swept over the component to ensure a solvation gradient. A unidirectional sweep to the edge of the region is important to ensure that the dissolved organic contaminant is adequately flushed from the zone undergoing the degreasing process. None-the-less, a volatile solvent will evaporate and can leave a thin film of organic contaminant uniformly spread over the adherend surface. Further, the solvent can dissolve polymers and residual greases from the tissue or cloth and deposit these on the adherend surface as the solvent evaporates (Fig. 9) [56]. Some efforts to replace solvents from petrochemical production have examined materials such a limonene, which has shown promise in specific degreasing applications [71,72]. The waterbreak test is widely used to assess the presence of contaminant on an adherend surface. However, the waterbreak test must be treated with caution because contaminants such as water-displacing fluids yield low contact angles [56]. Recent research has developed a portable method to apply the contact angle test easily and locally [31].

274

Andrew N. Rider et al.

5.2 Abrasion, Grit-Blasting or Etching Abrasion, grit-blasting or etching is conducted to remove loosely adherent oxides, to prepare a contaminant-free active surface and to generate a rough surface topography. Abrasion physically removes metal by the action of hard particles bonded to a carrier cloth or pad. This creates a surface with furrows and leaves residual metal debris on the surface. The practice of removing debris from an abraded aluminium alloy adherend using clean tissues must be approached with caution. Debris removal with tissues soaked in MEK leads to poorer bond durability than removal using tissues soaked in water (Fig. 9) [56]. The reason is that the MEK solvent can dissolve organic material from the tissue and leave the contaminant distributed across the surface as the solvent evaporates. This contaminant then interferes with the bond between the adhesive and the metal surface, allowing moisture to diffuse into the stressed bondline at a faster rate than for a bondline where there is better attachment between the adhesive and the metal surface [53]. The grit-blast process uses fine abrasive particles carried in a high velocity stream of clean, dry air or nitrogen to impact the adherend surface. Plastic deformation of the metal surface is more dominant than metal removal and crater formation in the surface is evident [73] (Fig. 14). Preexisting debris is consolidated into the surface during impact deformation of the metal surface.

(i)

450 kPa

(ii) (iv)

2 x 450 kPa

(iii)

50 µm

10 µm

Fig. 14 Micrograph cross sections showing (i) no grit-blast, (ii) single impact grit-blast and (iii) double impact grit-blast. The micrograph of the double impact grit-blast shows the degree of overfolding and cavities formed on the surface (iv).

275

Surface Treatment and Repair Bonding

20 10 0

4h Lab exposure

5h

30

Exposure to kerosene combustion

1 min

40

Abrade + water wipe

50

MEK wipe

Contact angle (°)

60

As received

70

20 min

80

Gritblast

Oven dry @110°C

90

Fig. 15 The contact angle for clad 2024 aluminium following degreasing (MEK wipe), abrasion with Scotch-Brite plus debris removal with water soaked tissues (SB + wipe (water)), dry at 110°C, grit-blast (GB), recontamination with either exposure to burnt aviation kerosene or a laboratory environment.

The grit-blast treatment improves the hydrophilic wetting of the surface (Fig. 15) [53] and the durability of the bond (Fig. 9) over that of the abraded surface. Some of the improvement in wetting can be attributed to a decrease in the concentration of hydrophobic contaminant and some to the roughening of the surface. Measurements with XPS show that hydrocarbon concentrations are decreased by solvent degreasing and further reduced by abrasion [53]. Curve fitting of the C1s peak in XPS spectra of solvent degreased and grit-blasted surfaces shows the reduction of a species containing a C]O bond, which could be attributed to the removal of residual ketone-based materials. The severity of grit-blast must be controlled. Insufficient grit-blasting leads to ineffective preparation of a contaminant-free active surface. The threshold grit-blast density for 50 μm alumina grit to achieve full surface coverage is 0.5 g cm2 [74]. It is recommended that 1 g cm2 should be used to ensure complete surface impact and contamination removal. Excessive grit-blasting does not improve the durability performance of the adhesive bond, although additional deformation of the surface will result. The additional surface deformation gives rise to subsurface cavities in the adherend which become sites for the entrapment of air, moisture and other volatile materials. During cure of the adhesive at elevated temperatures, these

276

Andrew N. Rider et al.

volatile materials form bubbles in the adhesive leading to voids in the cured bond [14,75,76]. Void formation will be discussed in Section 6.2. Chemical etching dissolves metal from the surface in a complex process involving dissolution and regrowth of the oxide film. Aluminium alloys immersed in chromic acid etching solutions develop a relatively flaw-free oxide due to the absence of contaminants [77]. The chemical dynamics of the etching process leads to a microporous structure [78] on the metal surface which leads to similar bond durability performance to that of the grit-blast treatment [73].

5.3 Creation of a High-Energy Surface Oxide Creation of a high-energy surface oxide implies the optimisation of surface wetting. It may be an automatic result of abrasion, grit-blasting or etching processes. Alternatively, processes such as anodising may be employed to create hydrophilic surface oxides. The creation of a high-energy surface oxide implies that steps to minimise the readsorption of hydrophobic contamination have been taken. Surface oxides form rapidly on almost all metal and alloy surfaces. Abrasion and chemical surface treatments are conducted to reduce contaminant, to remove preexisting loose oxides, to generate compact mechanically robust surface oxides and to produce hydroxylated oxide surfaces on the metals which will bond to the polar functional groups in the adhesive. Most surface treatments are optimised by trial and error with very little fundamental understanding of properties at the nanometre scale. The oxides formed on structural metals such as aluminium, steel and titanium will be sufficiently hydroxylated to form strong bonds [46]. However, the possibility of forming a weak oxide on the adherend must be considered. Extended etching of a nickel surface in a nitric acid solution produced an appropriately rough surface, but the smut and thickened oxide produced a weak bond interface [79]. Anodising is a process involving electrolytic treatment of metals in which a stable, porous oxide is intentionally grown on the surface of the metal [80,81]. This oxide is mechanically cohesive and tenaciously adheres to the metal surface. In a typical anodising bath, the metal alloy is connected as the anode in an electric circuit, is immersed in an oxidising electrolyte of usually low pH and a positive dc (direct current) potential is applied [82]. Whilst most anodising baths use acidic electrolytes and dc potentials, alkaline electrolytes and alternating current may also be used. The oxide is formed on the metal surface as a result of controlled chemical dissolution

Surface Treatment and Repair Bonding

277

of the metal and electrochemical oxidation of the surface [83]. The microtopography of the outer oxide is very sensitive to bath composition, anodising parameters, alloy composition and the surface finish on the metal. Since the best durability performance is generally obtained with a high degree of micro- or nanoscale roughness of the metal oxide surface, it is not surprising that attention to details of the bath and alloy is essential to obtain optimum surface-oxide film properties. There is extensive literature on factory anodising [83]. In field repair applications, tank anodisation is often not practical. The electrolyte is either formed into a self-supporting gel or special approaches for the containment of circulating electrolyte over the treatment region are used. The use of anodisation for on-aircraft repairs needs to consider issues including the potential for unwanted corrosion, the potential for hydrogen embrittlement of fasteners, difficulties with the removal of the electrolyte, difficulties with the use of electrical equipment in the vicinity of flammable vapours and the potential for damaging the oxide film with the postanodising processes.

5.4 Coupling Agent The purpose of the coupling agent is to provide a chemical link between the metal oxide and the adhesive which may improve the hydrolytic stability of the bond. Organosilane coupling agents are normally amphiphilic with an organic-head group which is chemically compatible with the adhesive and the silanol groups, formed during hydrolysis (Fig. 16), that react with

Fig. 16 Organosilane coupling agent γ-glycidoxypropyltrimethoxy silane after hydrolysis leading to silanol formation.

278

Andrew N. Rider et al.

the metal-oxide hydroxyl groups either via hydrogen [84] or, has been shown with surface spectroscopy, covalent bonding [7]. The organosilane forms strong polysiloxane networks, which play a significant role in interfacial durability enhancement [7,84]. Characteristics of a good coupling agent are the crosslinking with the adhesive and hydrolytic stability to water, which can hydrate the oxide film. Both are essential, but the crosslink connection into the adhesive and increased load sharing appears to also be important. Nitrilotris methylene phosphonic acid (NTMP) is an outstanding hydration inhibitor and was reported to improve bond durability [85]. However, other research [59,86] suggests that the inability of NTMP molecules to crosslink through primary chemical interactions prevents a cohesively strong film forming and leads to reduced performance as a durability improver. Studies have been conducted on the process parameters used in the ‘the Australian Silane Surface Treatment’, which identified the optimum organosilane concentration, hydrolysis time, application time, drying time and drying temperature [87,88]. Optimal parameters identified were similar to processes and conditions prescribed in the Royal Australian Air Force Engineering Standard [27]. While initial results suggested that the optimum temperature for drying organosilane coupling agent was 93°C for 90 min, later work suggested that higher temperatures might be required for void minimisation, when bonding to grit-blasted aluminium adherends [76,89]. An examination of the significant factors affecting organosilane performance on aluminium also showed that the influence of the preexisting surface roughness played a critical role in the environmental resistance of the adhesive joint as determined using the wedge test [86]. A number of other papers have also examined the mechanisms and factors that affect the organosilane coupling agent treatment on aluminium surfaces [67,90] and have suggested that drying temperatures and delays in applying the organosilane do not affect joint performance critically. The presence of contaminant on the metal surface plays a significant role in the ability of the organosilane coupling agent to attach to the hydroxyl groups on the metal oxide. This is illustrated in Fig. 7 where it was shown that the durability enhancement afforded by the organosilane coupling depends strongly on the surface treatment prior to the application of the coupling agent. Analysis of the failure surface with XPS indicates that as the durability improves, fracture shifts from the oxide film towards the interface between the coupling agent and the metal oxide [53]. The effectiveness

Surface Treatment and Repair Bonding

279

of the coupling between the organosilane and the metal oxide has an influence on the ability of water to diffuse to the interface between the adhesive and the metal adherend together with the hydrolytic stability of the metal oxide. In practical terms, it is essential to ensure that the surface treatment has produced a hydrophilic surface and to ensure that contamination is avoided before the application of the organosilane coupling agent [56]. Brushes or tissues used to apply organosilane can transfer organic contaminant to the prepared metal surface leading to deterioration in bond durability performance [56].

5.5 Adhesive Primer Most adhesive suppliers recommend a primer for metal bonding. Primers protect the adherend surface from contamination or chemical changes between surface preparation and the bonding process. Since they are low-viscosity fluids, primers can readily penetrate surface roughness and microporosity developed by the surface preparation and better wet the adherend surface. Primers also help protect the bonded adherend from moisture attack in order to improve long-term durability. For this reason, adhesive primers often contain corrosion inhibitors [91]. Popular adhesive primers are epoxy-phenolic based and contain chromate inhibitors. Whilst hexavalent chromium ions provide the best protection to the metal oxide surface, these materials are toxic to the environment, are known carcinogens and as a result there is pressure to find alternative inhibiting agents. Application of a corrosion-inhibiting adhesive primer to enhance bond durability for on-aircraft repair is as desirable as it is for factory and depot processes. However, priming on aircraft is more difficult to control both from the application and the environmental hazard point of view. Wipe-on or brush application of the primer is frequently substituted for spray application used in off-aircraft bonding. It is important to recognise that improperly applied primer, especially if it is too thick, can lead to inadequate initial bond strengths. Primer thickness may be measured by eddy current thickness gauges. Primers also may have pigments added as an aid to visual thickness control, particularly in the case of sprayed application. However, this optical method limits thickness control to fractions of a micron and the choice of the pigment can influence the visual sensitivity. The performance on thickness

280

Andrew N. Rider et al.

control is very dependent on operator skill and ability. The RAAF Engineering Standard DEF(AUST)9005 [27] does not include a primer specification. This follows concern over the ability of technicians to adequately control primer thickness and cure properties in a repair environment. The elimination of primer usually results in some reduction in wedge test performance, but RAAF service experience has indicated good performance using only the grit-blast/silane treatment. Some studies have shown that the primer performance may be affected by cure temperature of the adhesive, with lower cure temperatures leading to reduced benefit offered by the primer [92]. A comprehensive study on primers used in repair bonding of aluminium has been conducted more recently that has examined a series of primers which are either water based or have had the toxic chromiumbased pigments removed or both [93]. Generally, the use of water-based primers like Cytec BR6747–1 can offer equivalent performance to the traditional BR127 solvent-based primer [94], but removal of the strontium chromate usually reduces overall performance. However, recent studies show good results for a chromate-free, water-based primer when used with sol–gel AC-130 and AC-130-2 surface treatments (refer to Section 8.3.1) [95].

5.6 Drying Drying the surface thoroughly following any treatment involving solvents or water is absolutely essential to minimise the evolution of volatile materials responsible for void formation during the cure phase of the adhesive. Measurements of water evolution from a grit-blasted aluminium alloy surface indicate that sufficient steam can be generated by a poorly dried surface to eject most of the adhesive from the bondline [14]. Some of the water on the adherend is physisorbed and some is bound in the hydrated surface-oxide film (Fig. 17). Curing an adhesive bond at elevated temperature will release even the chemically bound water as steam [76]. Experiments have shown that drying at 110°C for at least an hour is essential to minimise void formation in some epoxy adhesives where cure is conducted using vacuum bag pressurisation. Conducting the bonding process in low humidity to avoid readsorption of moisture is also essential (Fig. 17).

281

Surface Treatment and Repair Bonding

Temperature 20°C

Ramp temperature to 177°C Temperature ramp to 177°C

0.25

Dried 80°C

0.3 Dried 40°C

Absorbance

0.25

Absorbance

0.2

0.2

0.15

Dried 110°C Re-exposed 30°C 70% RH

0.1

Dried 110°C

0.05 Dried 110°C

0.15

0

0

500

1000

1500

Time (s)

0.1 0.05 0 0

10

20

30 40 Time (min)

50

60

70

Fig. 17 FTIR water evolution profile for grit-blasted clad 2024 aluminium alloy illustrating the chemical affinity of the adsorbed water with the aluminium oxide surface and the effect of drying for 1 h at a range of temperatures.

6 ADHESIVE APPLICATION Adhesives (and primers) have a limited shelf life, and refrigerated storage in sealed containers is essential for most. Repacking of application lots of film adhesives in sealed polyethylene bags must be conducted in controlled atmospheric conditions to minimise moisture absorption. Quality assurance requires careful documentation of histories and conditions of storage. The quality of the bonding process relies on the flow of adhesive and the curing conditions. Process control of the temperature ramp rate and final temperature is essential. Aircraft structure has differential heat sinks leading to a requirement for zoned heating and looped control to avoid local overtemperature and under-temperature regions. Experience has shown that there is a general lack of care with ensuring that temperature-sensing devices are properly located and properly calibrated to ensure that the temperature sensed reflects the temperature in the curing adhesive. Trivial mistakes such as leaving the peel plies or separator sheets on the adhesive are committed even by the most experienced technicians.

282

Andrew N. Rider et al.

Fortunately, the consequences are quite apparent. However, the use of peel plies and other materials containing a silicone release agent is a more insidious error of judgement. The RAAF has taken steps to ban these materials from all bonding operations.

6.1 Factors Controlling Bondline Thickness The mechanical properties of an adhesive joint are influenced by the thickness of the bondline [96–98]. Many film adhesives contain a supporting scrim, which exercises some control over bondline thickness. The scrim may also act as a strength-limiting defect, can toughen the bond, or can provide a path for moisture ingress into the bond.

6.2 Void Formation and Minimisation Periodically, high levels of void formation in the bond have been observed for repairs conducted at tropical airbases [99]. Calculations show that the water content of the bondline in which the adhesive and the adherend are both exposed to humid, tropical conditions can be sufficient to produce enough steam during an elevated temperature cure to eject most of the adhesive from the bondline (Fig. 18) [14]. The resultant degradation in mechanical properties has airworthiness implications and, therefore, must be addressed. Recent studies show that fracture toughness is also reduced by composite adherend moisture absorption, which cannot be fully recovered with drying [100]. The volatile gases present in the bond have three potential sources [75]: Firstly, the adhesive supplied from the manufacturer may contain solvents

(A)

1 mm

(B)

Fig. 18 Void formation in FM 300 epoxy film adhesive where both the adhesive and adherend were exposed to (A) temperate (50% R.H. at 20°C) and (B) tropical (70% R.H. at 30°C), conditions.

Surface Treatment and Repair Bonding

283

with their concentration limits defined by a specification designed for positive pressure cure. There will be batch-to-batch variations in the concentration of these volatile materials [75]. Secondly, the adhesive can quickly absorb water [75] if exposed to hot, humid tropical conditions. The equilibrium concentration of moisture depends strongly on the relative humidity of the atmosphere. Thirdly, moisture is readily adsorbed by the rough overfolded surface (Fig. 14) of the abraded and grit-blasted aluminium alloy adherend [75,101]. This moisture is both physisorbed and chemisorbed by the aluminium alloy surface (Fig. 17). While each source of volatile gases on its own may be relatively innocuous, the combined effect of all three sources can lead to a level of voiding which will affect mechanical properties (Fig. 18) [76]. Void minimisation strategies are generally centred on minimising the moisture content of the bondline or by constraining the volatile gases using hydrostatic pressure. The most effective strategies are to conduct bonding in a temperature controlled, dehumidified atmosphere and to use positive pressurisation of the bondline with an autoclave, press or pressurised bladder. Aside from moisture content, entrapped air can cause problems and a variety of methods are available to assist in trapped air removal, such as embossing adhesives to provide an air path [102]. In the case of bonding to composites, the availability of partially impregnated prepreg materials can also facilitate air path development and enable vacuum cure to be applied [102].

7 SURFACE TREATMENT QUALITY CONTROL Tools to assist with quality control during the production of an adhesive bond and nondestructive evaluation of the bonded joint need to be applied carefully to ensure reliable outcomes. The production of a strong and durable adhesive joint depends critically on the skill and the integrity of qualified personnel manufacturing the bonded joint. Strict adherence to a qualified procedure is essential. A benchmarking activity conducted for the RAAF [28] indicates that adherence to well-defined qualified procedures, underpinning standards, staff training, regular reviews of staff qualification and continuity of experience were of utmost importance to achieving acceptable bond durability performance in the repair environment. The use of suitable quality control monitoring can help identify deviations in process standards and enable quick adjustment to ensure

284

Andrew N. Rider et al.

maintenance of standards [29,30]. New portable quality control monitoring devices are now being developed to assist in field and depot level adhesive bonding activity [103].

7.1 Waterbreak Test The aerospace industry routinely uses the tendency of clean water to ‘bead’ or ‘break’ as an indication of the presence of hydrophobic contaminant on an adherend during surface preparation. In practice, the waterbreak test relies on the skill and experience of the technician and is not necessarily reliable. Surface roughness has a significant influence on the outcome. Some contaminants have hydrophilic characteristics and, therefore, lead to a waterbreak free indication. Examples of hydrophilic contaminants commonly encountered are water-displacing fluids used in aircraft maintenance. As mentioned, new portable waterbreak test methods are now being applied in quality control applications [31,103].

7.2 Surface Work Function Methods The electron work function of a metal surface is very sensitive to its chemical state. A number of methods based on work function are in use to assess contaminated surfaces [104–106]. However, the work function methods have difficulty in discriminating between oxide growth, contaminant concentration and surface roughness. The Fokker surface contamination tester is based on the Kelvin vibrating capacitor surface potential difference method [104]. The physical size of the measurement area, the sensitivity to interelectrode spacing and potential contamination of the gold reference electrode are limitations. The Optically Stimulated Electron Emission (OSEE) method is also commercially available [107]. The method is based on the emission of photoelectrons with energies of less than 6.7 eV, stimulated by ultraviolet radiation from a mercury vapour source. The photocurrent is measured in air using an electrically biased plate located 1–6 mm from the sample surface [106]. Aluminium metal with a work function of 4.08 eV [108] will emit photoelectrons, whereas aluminium oxide with a work function greater than 6.7 eV will not [106]. It was found [109] that the principal impediments to quality control in the base repair environment are the convoluted response to contaminant concentration, oxide growth and surface roughness. More recent work, however, has used the OSEE method combined with glossmeter readings to show that the technique can be used reliably to assess

Surface Treatment and Repair Bonding

285

contaminant levels on grit-blasted aluminium plates in controlled laboratory environments [110]. Foreseeably, the OSEE method could be used in production or depot maintenance with controlled environments to ensure surface cleanliness levels are maintained in defined treatment processes.

7.3 Fourier Transform Infrared Spectroscopy Advances in Fourier transform infrared (FTIR) spectroscopy have led to some promise as a technique to evaluate contaminants on a rough metal surface. However, the complexity of infrared spectroscopy leads to the requirement for a skilled analyst to interpret the data [103]. Commercial instrument manufacturers [111,112] have developed infrared spectroscopic systems for noncontact evaluation of surfaces. It should, however, be clearly understood that the depth discrimination of infrared spectroscopy is typically in the hundreds of nanometres range and that special grazing incidence techniques are required to bring the sensitivity down to monolayer coverage levels. With the correct accessories, contaminant on metal surfaces can be detected using the new handheld instruments [112].

7.4 Optical Reflectivity Inspection of the quality of a grit-blasted surface is performed visually and, for aluminium surfaces prepared with fine grit, optical reflectance characteristics are a good indicator of the severity of grit-blast impact [74,109]. Commercial reflectance colorimeters have been used to quantitatively assess the reflectivity of aluminium alloys and show an exponential dependence on the density of impact of fine 50 μm alumina particles [74,109]. A very simple hand-held instrument was shown to adequately monitor grit-blast severity on aluminium alloys and has previously been used as a training and qualification tool by the RAAF. More recent developments have used a commercially available gloss-meter to reliably assess the surface roughness of grit-blasted aluminium surfaces in training applications for RAAF [110]. The polarisation response of reflected polarised light is used as a quality control tool for phosphoric acid anodised surfaces. The method indicates defects in the thickness and maturity of the porous anodised film [113].

7.5 Process Control Coupons (Traveller or Witness Specimens) In the absence of a comprehensive array of quantitative quality control tools, it is common practice for process control coupons, also known as traveller or

286

Andrew N. Rider et al.

witness specimens, to be prepared in parallel to the production or repair task. This procedure is useful in guaranteeing that the materials, processes and application environments being used are satisfactory, but does not guarantee the quality of the bonding in the part that would be provided by a proof test. The RAAF Engineering Standard DEF(AUST)9005 [27] places emphasis on quality control through a strategy of qualification of processes, procedures and personnel. Nevertheless, some recent efforts have looked at the use of satellite coupons which are bonded at the perimeter of the repair to provide a very close representation of the bonded repair environment during application [114]. More details are provided in Chapter 6.

7.6 Practitioner Education, Skill and Standards Quality control in the manufacture and repair of bonded components relies on a strategy of qualification of processes, procedures and personnel [24]. It is essential for the procedures to include a quality assurance trail to ensure that the task was performed in strict accordance with the qualified specification. It is to be noted that the regulatory framework used to manage the structural integrity of bonded components is currently not capable of identifying bonded components which are susceptible to time-dependent bond degradation [23]. There is much work to be done towards developing regulations addressing preproduction validation tests to eliminate practices which lead to bond degradation. The lack of reliable nondestructive evaluation tools has led to a reluctance by many engineers to accept the engineering risks involved in an adhesively bonded joint, particularly for primary load-bearing structure [2].

8 SURFACE PREPARATIONS FOR ALUMINIUM ADHERENDS Historically, the majority of failures in aluminium-adhesive joints in aircraft have been initiated by moisture [115]. Thus, the employment of complex chemical treatments ensures adequate service life of the joint when it is exposed to aqueous environments. In general, the surface preparation of aluminium is designed to remove weak boundary layers (oxide scale and organic contaminants) and to form stable layers that adhere well to the base metal and are chemically and physically compatible with the adhesive or primer [80,116]. This is required of both factory and field surface preparation processes, including on-aircraft treatments. The extent of the surface treatment depends upon the demands of the application.

Surface Treatment and Repair Bonding

287

8.1 Factory Processes Considerable research has been conducted regarding preparation of aluminium for adhesive bonding, particularly for 2000 and 7000 series alloys since they are most used in aerospace applications. Most factory processes for aluminium surface preparation involve etching in acid solutions or electrochemical anodisation or both. These treatments are intended to produce a stable oxide layer that tenaciously adheres to the metal surface. The best durability performance is generally obtained with a high degree of microroughness on the surface. Durability is further increased when the adherend surface is also chemically compatible with the adhesive or primer [80]. 8.1.1 Phosphoric Acid Anodise (PAA) Phosphoric acid anodise (PAA) is currently the most widely used anodisation for aluminium prebond treatment [81]. This is largely due to the superior performance the treatment demonstrated during the United States Air Force’s Primary Adhesively Bonded Structure Technology (PABST) programme [117,118]. PAA is characterised by its simple chemistry, room temperature requirements, low electrical needs, relatively good environmental acceptability and fairly wide tolerances for process parameters [119]. The PAA process produces an approximately 400 nm thick amorphous aluminium oxide characterised by a thin inner barrier layer, an outer porous layer [115]. In relation to alternative anodisation processes, the PAA oxide will not hydrate or ‘seal’, is much thinner and has the largest pore sizes [82]. Anodisation is carried out in 10% aqueous H3PO4 (by weight) at 10 V. The exact details of the procedure vary slightly between using organisations [44,120,121]. The best durability performance is obtained by employing an acid etch prior to anodising [122]. Both FPL and P2 etch [122] have been employed in this way. In order to obtain repeatable durability results, key process variables, such as tank make-up and rinse water, must be controlled. The microporous PAA oxide is fragile and must be primed or bonded as soon as practical. A low-viscosity primer can penetrate the pores, stabilise the oxide layer and protect the adherend surface from corrosion, given that the thin PAA oxide provides little inherent corrosion protection. Although the use of a few rubber-containing 177°C curing adhesive primers led to suboptimal bond strengths when used with PAA, most primers are compatible with PAA oxides [123–125]. PAA-primer systems yield good bond strengths and are generally considered to provide the best long-term durability performance for joints made with typical aerospace aluminium alloys and epoxy adhesives [44,81,126].

288

Andrew N. Rider et al.

8.1.2 Chromic Acid Anodise (CAA) Chromic acid anodise (CAA) is the other common, high-performance anodisation used for aluminium alloys. The overall oxide thickness is nearly four times that of PAA (approximately 1500 nm). This thicker, denser oxide is less fragile than the PAA oxide and provides greater corrosion protection to the base alloy. Although the CAA oxide provides less-developed porosity than does the PAA oxide, primers have been shown to penetrate the pores. Anodisation is conducted in 5% aqueous CrO3 (by weight) at a temperature in the range of 32–42°C with the potential gradually raised from 40 to 50 V. The oxide morphology can be altered by varying the process parameters [120]. Cleaning and deoxidising steps are similar to those for PAA. The pretreatment acid etch step for the CAA process influences the porosity of the outer surface of the oxide [115]. The CAA treatment is optimal when preceded by a chromic-sulphuric acid etch [127]. Corrosion protection provided by CAA can be enhanced by sealing the oxide with hot water or dichromate solution [128], but sealing fills the pores and decreases adhesion. Although the attributes of sealed CAA have performed well in some applications [129], the process is generally not recommended for adhesive bonding applications [44,127]. The unsealed oxide may be primed with a corrosion-inhibiting adhesive primer or an organosilane coupling agent. The CAA-primer systems yield good initial bond strengths and durability. Performance in moist environments is slightly inferior to PAA [44,81,130] when tested with toughened epoxy adhesives. In actual service with vinyl-phenolic adhesives, CAAtreated aluminium joints have an outstanding durability record. The CAA process has recently become less popular as environmental and safety concerns regarding hexavalent chromium are making it more expensive and difficult to use. Recently, boric sulphuric acid anodise (BSAA) has been developed to produce an oxide similar to the CAA oxide without the use of chromium. Although the treatment was intended originally for paint adhesion applications, a variant called sulphuric boric acid anodise (SBAA) was optimised for aluminium adhesive bonding [131]. 8.1.3 Optimised Forest Products Laboratory (FPL) Etch The optimised Forest Products Laboratory (FPL) etch produces a 40 nm thick amorphous Al2O3 film [132] with an outer network of shallow pores on top of a thin barrier layer. This microroughness is less pronounced than is present in the PAA oxide.

289

Surface Treatment and Repair Bonding

The optimised FPL etch is 5% by weight Na2Cr2O72H2O, 26.7% H2SO4 and 68.3% water with a small quantity of 2024 aluminium to seed the bath [120]. Etching is conducted for 10 min at a temperature of about 65°C. Pretreatment steps include solvent degreasing and/or alkaline cleaning. As with the other chemical processes, rinse steps are important. The FPL etch variables must be carefully controlled. The etched aluminium surface is usually primed with a corrosion-inhibiting adhesive primer prior to adhesive bonding. The FPL etch process is less expensive and time-consuming than the anodise procedures. It yields good initial bond strengths, but inferior moisture durability compared to the CAA or PAA anodises [44,81,130,132,133]. 8.1.4 P2 Etch The P2 etch uses ferric sulphate in place of the toxic sodium dichromate as the oxidiser (15% by weight FeSO4, 37% H2SO4 and 48% water [120]) and produces an oxide with a similar morphology to those obtained using the various chromic-sulphuric etches [80]. The P2 etch produces similar initial bond strengths and durability to that of FPL etch. The performance as an anodisation pretreatment is comparable to the FPL etch [80,122,134]. Three factory aluminium surface preparations are compared on the basis of wedge test data, indicated in Fig. 19, for Cytec BR 127-primed Al7075T6 alloy bonded with Loctite Hysol EA 9628 adhesive [122]. The failure modes are indicated on the plot, where ‘Adhesive’ failure represents interfacial failure between the adhesive and the metal surfaces and ‘Cohesive’ failure represents failure within the adhesive layer.

Cracklength (mm)

40 Adhesive failure

FPL PAA P2

36

Adhesive/ cohesive

32 100% Cohesive

28 0

2

4

6

8

10

12

14

Root time (h0.5)

Fig. 19 Wedge test data comparing the relative durability performance of the PAA, FPL and P2 factory-based surface pretreatment processes for BR 127 primed Al-7075-T6 alloy bonded to EA 9628 adhesive and tested at 50°C/95% RH [122].

290

Andrew N. Rider et al.

8.2 On-Aircraft Acid Anodisation and Acid Etch Processes The limitations imposed by the on-aircraft environment lead to a demand for simple surface preparations. Solvent degreasing and manual abrasion alone lead to extremely poor long-term durability [44,126]. Adaptations of factory acid etch and anodisation processes have been made to enable their use on aircraft to obtain adequate bond performance. Care must be exercised with the use of acids since these can cause either corrosion in joints and around fasteners or embrittlement of high-strength steel fasteners [135]. Some aircraft structural repair manuals specify primer application with cheesecloth [126]. Inadequate control over primer thickness, often encountered during aircraft repairs, such as with wipe-on or brush applications, can lead to poor initial bond strength. The elimination of the primer can decrease the long-term bond durability. Decisions regarding priming and the entire surface preparation process must often be made on a case-by-case basis for on-aircraft repair. 8.2.1 On-Aircraft Phosphoric Acid Anodise PAA has been adapted to on-aircraft use. Treatment steps involve the removal of organic coatings, solvent degreasing, manual abrasion and drywipe removal of abrasives and debris. Two anodising approaches have been developed: The Phosphoric Acid Non-Tank Anodise (PANTA) process uses phosphoric acid, thickened with fumed silica, in a gauze sandwich with a stainless steel mesh cathode [136]. Precautions to keep the gel moist and to avoid trapping hydrogen gas are essential during anodisation. After anodisation, the acid must be removed quickly and without damaging the fragile oxide surface. The surface is then dried prior to priming [126]. The Phosphoric Acid Containment System (PACS) contains the phosphoric acid under a double vacuum bag [137]. For the PACS process, the steel cathode screen is placed on top of a nylon breather material. Vacuum is used to pull phosphoric acid through the bag, and anodising is conducted for 25 min with a continual flow of acid over the repair area. Rinsing is accomplished by drawing clean water through the vacuum bag. Final rinsing is conducted after removing the vacuum bag, breather and cathode screen. The surface is then dried and primed prior to bonding. The advantages of the PACS process over PANTA include acid containment, the ability to conduct overhead applications, the avoidance of

Surface Treatment and Repair Bonding

291

electrolyte drying and the minimisation of trapped gas against the repair area. However, limited access to the repair area or leaking fasteners under the containment bag could prevent the use of PACS in some applications. Morphology studies of PANTA and PACS oxides indicate they are similar to those obtained with the factory PAA process. Mechanical testing shows initial bond strength and durability comparable with the tank PAA process [126,136,138]. Both the PANTA and PACS preparations should be followed by application of a corrosion-inhibiting adhesive primer to protect the anodised surface. 8.2.2 Acid Paste Etching Processes The popular chromic-sulphuric acid etches have also been adapted for onaircraft application using fumed silica, barium sulphate or other suitable material to thicken the acid. Ambient temperature etching increases the application time. Typical treatment steps include a solvent degrease and manual deoxidisation by abrasion or grit-blasting. The paste etches can be difficult to apply or rinse. They generally provide good initial bond strengths but poorer durability than the factory etching processes. Pasa-Jell 105 (Semco Products Division of PRC-DeSoto International) is an inorganically thickened blend of acids, activators and inhibitors that is specified in many aircraft repair manuals. Durability results obtained using Pasa-Jell 105 are much better than the simple hand cleaning approaches but are inferior to those obtained using the factory optimised FPL etch and PAA processes as well as PANTA [126]. The P2 paste etch uses a thickened version of the factory P2 solution and delivers performance similar to that obtained with the chromic-sulphuric etches. Ideally, a corrosion-inhibiting adhesive primer is applied after rinsing and drying the repair surface. 8.2.3 Chromate Conversion Coatings Certain conversion coating processes that are primarily intended to promote the adhesion of paint to aluminium have been used as prebond treatments. In general, conversion coatings should not be used where good adhesive bond durability [139] and cohesive failure modes [140] are desired. The best results employ a 2% hydrofluoric acid etch prior to conversion coating within 15 min of the etch [140]. Durability results show the process can be superior to Pasa-Jell 105 but inferior to factory processes and PANTA. Although the 2% hydrofluoric acid method has been included in aircraft

292

Andrew N. Rider et al.

repair manuals, it is really no longer a viable on-aircraft surface preparation due to the hazardous chemicals involved.

8.3 On-Aircraft Processes Based on SoldGel and Organosilane Chemistry Two main treatment processes have been used over the past 20 years that have utilised organosilane coupling agent chemistry [87–89] and sol–gel chemistry [141,142]. In 1993, the Materials and Manufacturing Directorate of the United States Air Force Research Laboratory (AFRL/ML) became active in metal surface preparation research and funded multiple efforts through 1997. The technologies investigated included excimer laser, plasma polymerisation, ion beam enhanced deposition (IBED), plasma spray and sol–gel [141]. The sol–gel work clearly showed the greatest potential for application due to its performance as well as scale-up potential. For this reason, AFRL focused its efforts on optimising sol–gel technologies. 8.3.1 SoldGel Chemistry The area of sol–gel chemistry represents a branch of polymer science that, beyond silicones, has only been exploited for aerospace applications in the past 15 years [142]. The term ‘sol–gel’ is a contraction for solutiongelation and refers to a series of reactions where soluble precursors (typically metal alkoxides, substituted metal alkoxides or metal salts) undergo hydrolysis and condensation to form a sol and then crosslink to become a gel. This crosslinked structure can be deposited as a coating and consolidated by dehydration. Many different metal atoms can be used to produce films with a wide variety of properties. Organic functionalities can be attached to the hybrid metal’oxygen framework to create organometallic polymer systems with even more diversity. Sol–gel technology has revolutionised metal adhesive bonding by providing an environmentally compliant, high-performance, simple and inexpensive approach for surface preparation [142]. One of the big advantages to sol–gel chemistry is its versatility and the ability to be tailored for specific applications. Solution chemistry can be controlled to vary deposited film density, porosity and microstructure. Ideally, adhesion is via direct chemical bonding at the coating/substrate interface as well as the coating/adhesive interface [143]. Several sol–gel chemistries have been developed by a number of organisations [144,145] and is commercially available and applied widely in the aerospace industry. AFRL originally focused its efforts on technology that

Surface Treatment and Repair Bonding

293

led to a surface preparation similar to the existing organosilane processes, while using a more reactive water-based silicon’zirconium chemistry [142,146,147]. This approach led to improved performance, which is also compatible with paste adhesive technology [148]. As with the organosilane preparations, sol–gel chemistry investigated by AFRL eliminates the dependency on strong acids and bases. Also, there are no power requirements as required for anodisation, and wastewater is greatly reduced over conventional wet chemistry approaches since rinsing is not required. There is no need for hexavalent chromium in the sol–gel process, and it may be possible to eliminate chromates from the priming step. Research has reported encouraging results for chromate-free primers [95]. Simple on-aircraft repair versions intended for use in repair bonding of noncritical structure are now available, without adhesive primer, and can be applied with both film and paste epoxy adhesives [149]. Sol–gel methodologies have been developed for adhesive bonding applications for factory, rework and on-aircraft applications. In the past 15 years there has been extensive development of the sol–gel treatment for a range of metals. A summary of the optimisation work on the metals aluminium, titanium and stainless steel and the effect of abrasion treatments and adhesive types as well as sol–gel ageing effects is detailed in Refs [150,151]. Research examining the effect of primers [93], operating windows [152], abrasion methods, [153] application and cure methods [154] and application with room temperature curing adhesives has also been reported [155]. The sol–gel process is underpinned by extensive testing and research and is now providing a reliable surface treatment alternative to some of the more traditional methods that rely on harsh chemicals. 8.3.2 Organosilane Coupling Agent Surface Preparation The application of γ-glycidoxypropyltrimethoxysilane coupling agent following grit-blast produces very good durability results [87–89]. The organosilane coupling agent is applied from a dilute aqueous solution which deposits an ultrathin layer that dramatically improves the durability of adhesive bonds formed with a grit-blasted aluminium surface (Fig. 7). The fundamental chemistry that explains the mechanism of the organosilane layer is explained in Section 5.4. Versions of the grit-blast/silane surface preparation, known as the ‘Australian Silane Surface Treatment’, are popular for aluminium treatment for repair bonding applications. The primary reason for the use of the organosilane treatment is the ability to achieve highperformance bond strength and environmental durability without the use

294

Andrew N. Rider et al.

Cracklength (mm)

100 80

100% Adhesive PANTA Pasa-Jell 105 Silane

60 40 100% Cohesive

20 0

5

10

15

20

Root time (h0.5)

Fig. 20 Wedge test data comparing the relative durability performance of the PANTA, Pasa-Jell 105 [156] and Silane on-aircraft [89] surface treatment processes for BR 127primed Al7075-T6 alloy bonded with FM 73 adhesive and tested at 50°C/95% R.H.

of acids that can be problematic for on-aircraft processing. Application of BR 127 chromate-containing primer after the silane surface treatment can offer improved durability in applications where harsher environmental conditions may be experienced. The relative performance of PANTA, Pasa-Jell and organosilane coupling agent on-aircraft aluminium surface preparations is indicated by the wedge test data shown in Fig. 20 for BR 127-primed Al-7075-T6 alloy bonded to FM 73 adhesive [89,156]. The failure modes for each treatment are also indicated on the plot. 8.3.3 Hot Solution Treatment for Adhesive Bonding A process termed the ‘Hot Solution’ treatment has been developed at DST Group in response to the need for environmentally friendly surface treatments. The process involves immersion of aluminium alloys in boiling water followed by immersion in a 1% solution of epoxy silane. Wedge test experiments indicate that the durability of this treatment may perform as well as phosphoric acid anodisation for some aluminium alloy and epoxy adhesive combinations [157]. Fundamental research has identified that optimum durability is achieved for immersion of the aluminium between 4 min and 1 h in the distilled water heated to between 80°C and 100°C. These conditions enable a platelet structure to grow in the outer film region, which, combined with the formation of hydrolytically stable adhesive bonds made to the epoxy silane, appears to be critical in the development of the excellent bond durability [157,158]. The process can also be successfully applied at lower temperatures, but longer times are required [159].

Surface Treatment and Repair Bonding

295

9 SURFACE PREPARATIONS FOR TITANIUM ADHERENDS Titanium alloys are found in specialised aerospace applications due to their high strength-to-weight ratios, retention of mechanical properties at elevated temperatures (exceeding 400°C), excellent fracture toughness and corrosion resistance [160]. The most widely used aerospace titanium alloy, Ti-6Al-4 V, will be the focus of this discussion. As with aluminium, moisture is the environment of primary concern, limiting the long-term durability of the bonded joint. Long-term exposure to high temperatures is also a durability concern for titanium bonded joints.

9.1 Factory Processes Most factory processes for titanium surface preparation involve etching or electrochemical anodisation in acidic or alkaline baths. The best durability performance is obtained when the treatment creates microroughness, moderate to good durability is obtained with treatments generating macroroughness and poor durability performance is generally obtained with smooth titanium surfaces [160]. Plasma spray is one potential factory treatment, driven by the need to find a surface preparation for titanium that can withstand long-term exposure to elevated temperatures. Good durability results have been obtained when Ti-6Al-4 V powder was sprayed on the same alloy. A microscopically rough metallic film is deposited. Its morphology is more random than those obtained with the chemical treatments, having deep pores and many knob-like protrusions [161]. 9.1.1 Chromic Acid Anodise (CAA) and Alkaline Anodise (NaTESi) The CAA process is a widely utilised and accepted process for prebond treatment of titanium, particularly Ti-6Al-4V. CAA leaves a durable, porous layer of amorphous TiO2 that is microrough to increase surface area for physical and chemical bonding [160]. Both 5-V and 10-V CAA processes exist. These differ from the CAA procedures used for aluminium preparation since they contain hydrofluoric acid. A study conducted by the U.S. Navy in 1982 found the 5-V CAA process to provide the best overall moisture durability as determined by the wedge test (ASTM D 3762) with conditioning at 60°C and 100% (condensing) relative humidity. Four epoxy film adhesives, 121°C-curing and 177°C-curing, were used. For the study, various organisations provided treated titanium panels to the U.S. Navy for

296

Andrew N. Rider et al.

testing. The 10-V process also provided very good durability [162]. There is regulatory pressure in many areas to eliminate the use of hexavalent chromium used in the CAA process. A number of studies have used an alkaline-based anodisation process using sodium hydroxide solutions called the NaTESi process [163,164]. The results indicate similar performance to CAA but do not require the toxic chromium. 9.1.2 Phosphate Fluoride Several phosphate fluoride procedures have been developed. Although phosphate fluoride treatments are still used for titanium prebond surface preparation, long-term durability is generally not good. These processes ranked at the bottom of the 1982 U.S. Navy study, with markedly poorer wedge test results than the other processes evaluated [162]. 9.1.3 Pasa-Jell 107-M Pasa-Jell 107-M is a proprietary (Semco Products Division of PRC-DeSoto International) blend of mineral acids (nitric and chromic), activators and inhibitors that is specifically formulated for treatment of titanium. It is intended to clean and chemically activate the titanium surface to improve chemical bonding of the adhesive or primer. The process, including rinsing, is conducted at ambient temperature (less than 38°C) [165]. Pretreatment steps include degreasing and mechanically abrading the titanium surface. Solvents or alkaline cleaners can be used for the former. Dry or wet abrasive blasting with aluminium oxide grit is recommended for the latter. In the 1982 U.S. Navy evaluation, the Pasa-Jell 107 process, preceded by a liquid hone step, provided very good wedge test durability, refer Fig. 21. This treatment ranked just behind the 10-V CAA process [162]. The dry hone step prior to Pasa-Jell 107 treatment produced poorer durability. 9.1.4 TURCO 5578 TURCO 5578 (TURCO Products Division of Henkel Surface Technologies) is an alkaline etchant containing sodium hydroxide. The process, including rinsing, is conducted at elevated temperature (80–95°C). TURCO 5578 can be used to remove contaminants and evenly etch titanium surfaces. It does not contain chromates, phenol or hydrofluoric acid [166]. The TURCO 5578 process does not etch titanium as quickly as the common acid etchants and it is slightly more difficult to maintain in the process tank, however, it does not cause the hydrogen embrittlement

297

Surface Treatment and Repair Bonding

Cracklength (mm)

100 100% Adhesive

90 CAA(10V) CAA(5V) Phosphate Fluoride Pasa-Jell 107-M Turco 5578

80 70

100% Cohesive

60 50

0

5

10

15

20

25

30

35

40

Root time (h0.5)

Fig. 21 Wedge test data comparing the relative durability performance of the CAA, phosphate fluoride, Pasa-Jell 107M and TURCO 5578 factory-based surface treatment processes for Ti-6Al-4V alloy primed with BR 127, bonded with FM 300 K adhesive and tested at 60°C/100% RH [162].

that can be a concern with acid etchants. Although the TURCO 5578 process was not found to produce a great deal of microroughness on the treated titanium surface [160], the U.S. Navy showed that it provided very good wedge test durability, similar to the two CAA processes [162]. Five factory titanium surface preparations are compared on the basis of wedge test data indicated in Fig. 21 for BR 127-primed Ti-6Al-4V alloy bonded with FM 300 K adhesive [162]. The failure modes for each surface treatment are indicated in the plot. Recent work has examined how the TURCO 5578 can be modified using a plasma process to deposit a silica-like layer on the etched titanium substrate [163]. The plasma treatment can improve wedge test performance of the TURCO 5578 treated titanium almost to the level of the NaTESi anodisation treatment. 9.1.5 SoldGel and Silane Coupling Agent Treatments The basic chemistry of the organosilane coupling agent makes it suited to use across a range of metals. Studies have shown that the titanium surface roughness affects the environmental resistance of adhesive bonds with increased surface roughness provided by grit-blasting improving durability [167]. The use of grit-blasting and organosilane combined with new water-based epoxy primers can also lead to some durability improvements, but generally the use of the sol–gel treatment (Section 8.3.1) combined with water-based primers, provides superior performance [94]. Much of the initial

298

Andrew N. Rider et al.

development work for the sol–gel treatment of titanium adherends is detailed by AFRL reports [150,151] and was described in Section 8.3.1. 9.1.6 Laser Treatment Some effort has examined the use of lasers to pretreat titanium for bonding applications to remove the problems associated with chemical-based processes. Lap shear results indicated a relationship in strength and durability associated with laser ablation conditions [168,169]. Work has shown that the laser treatment can offer a replacement for common chemical treatments and does not impact on fatigue performance of the titanium [170].

9.2 On-Aircraft Processes Due to restrictions imposed by the on-aircraft environment, there are fewer titanium treatment options. The key on-aircraft challenges are the inability to use an elevated-temperature process, the difficulty in containing and rinsing the highly acidic or alkaline etchants and controlling hazardous materials. Although on-aircraft CAA could be conducted in a manner similar to on-aircraft PAA for aluminium, this does not appear to be common. The viable options for on-aircraft repair are grit-blasting, Pasa-Jell 107 (a thickened version of the tank etchant), grit-blast/silane identical to the process used for aluminium (Section 8.3.2) and the sol–gel technology (Section 8.3.1). 9.2.1 Grit-Blasting Grit-blasting is often used as a stand-alone titanium prebond treatment [171]. In contrast to the case for aluminium, grit-blasting treatment of titanium is one of the best procedures for obtaining good initial joint strength. For this reason, and the fact it is relatively easy and nonhazardous, gritblasting is often used for on-aircraft titanium treatment. However, although adequate joint durability can often be obtained, long service life in moisture or other aggressive environments requires alternate approaches [44]. It is best to use grit-blasting as a pretreatment step for Pasa-Jell 107 or the silane and sol–gel processes combined with primers. 9.2.2 Pasa-Jell 107 The paste version of Pasa-Jell 107 (inorganically thickened Pasa-Jell 107-M) can be used to treat titanium on aircraft since it is an ambient-temperature process. Pretreatment via grit-blasting with aluminium oxide is required for best durability performance. Care must be taken to contain the acid and properly rinse the aircraft components after etching.

Surface Treatment and Repair Bonding

299

10 SURFACE PREPARATIONS FOR STEEL ADHERENDS Surface preparations for steels, particularly chemical treatments, are greatly influenced by the nature of the substrate and its initial condition. The large number of steels makes the objective of achieving a universal treatment difficult [171]. The approach of intentionally forming a coherent, adherent oxide with fine microroughness on the surface of steel is not an effective strategy for good adhesion [171]. In general, the important factors for steel surface preparation are cleanliness and descaling or rust and oxide removal, and passivation for stainless steels [172]. Care must be taken during the preparation process, since many steel alloys rapidly form surface oxides. For instance, drying cycles after treatment can be critical. Also, alcohol is often found in treatment solutions, and alcohol rinses may be used after water rinses. Primers are also desirable to help protect the bonded joint from moisture attack [173].

10.1 Factory Processes For surface preparation, three general approaches exist: mechanical abrasion, chemical etching and conversion coating. 10.1.1 Grit-Blasting The formation of a macrorough surface using grit-blasting is a very common surface treatment for steels. Angular alumina grit is often used for this process. This approach can readily yield good, reproducible initial bond strengths [44] and adequate durability may be realised for many applications. However, to obtain the longest service life, additional treatment is usually required [44,171]. 10.1.2 Acid Etches The morphologies produced on acid-etched steel surfaces are a function of the steel microstructures. Acid etchants can create surface roughness by attacking the grain boundaries of the metal. Some of these processes include the following: nitric-phosphoric acid, phosphoric acid-alcohol, chromic acid, nitric-hydrofluoric acid, sulphuric acid-sodium dichromate, sulphuric acid-sodium bisulphate, oxalic-sulphuric acid and hydrochloric acid-ferric chloride [171,173,174]. Many of the acid etches leave a deposit of carbon, known as smut, on the steel surface. Therefore, a desmutting step, typically using another acid, must also be conducted.

300

Andrew N. Rider et al.

It has been suggested that chemical etches for steels, other than stainless, are not desirable. The many different etches investigated do not tend to outperform grit-blasting, even in durability testing. The ultimate performance for stainless steel bonded joints is obtained when the steel is chemically treated, although little is known about the mechanisms that lead to this improved performance [44]. Furthermore, there is no consensus regarding which treatment is best for a particular alloy. 10.1.3 Conversion Coatings Corrosion-resistant conversion coatings are available for steel and are used as treatments prior to painting. Several of these have also been evaluated for adhesive bonding applications. The conventional phosphate coatings for steel did not provide adequate bond durability and were adversely affected by elevated-temperature adhesive cure cycles [174]. Improved results are obtained when the phosphating process is closely controlled [171,174]. 10.1.4 SoldGel Treatments The sol–gel treatment (Section 8.3.1) has also been successfully applied to stainless steel and when combined using grit-blasting and water-based primer exhibits very durable performance after 48 h in a 60°C/98% R.H. environment [94]. Detailed reports on the application of the sol–gel treatment to stainless steel are provided in AFRL reports [150,151].

10.2 On-Aircraft Processes Preparation of steel is even more difficult for on-aircraft adhesive bonding. Most of the factory processes are impractical since they rely on strong acids, typically used at elevated temperatures. Grit-blasting is a viable option for on-aircraft repair of steel as it was for titanium. In order to improve environmental durability, a corrosion-inhibiting adhesive primer should be applied. The sol–gel treatment combined with BR6747-1 primer can be used successfully for field and deport level repairs as described in Section 10.1.4 and offers the best performance for simple surface treatments. The silane surface preparation has also been applied successfully to prepare D6AC steel wing skin surfaces for bonded repair using boron/epoxy patches bonded with epoxy adhesive [175].

Surface Treatment and Repair Bonding

301

11 SURFACE PREPARATIONS FOR THERMOSETTING-MATRIX COMPOSITES Many types of composite materials exist and, as with metals, the nature of the adherend to be bonded determines the best surface preparation. The types of composites used in structural applications in the aerospace industry are typically fibre-reinforced thermosetting resins. The fibres contribute strength and stiffness to the composite while the resin matrix transfers load. This discussion concentrates on epoxy matrices reinforced with graphite or boron fibres and bonded with epoxy adhesives. Whilst most people realise adhesive bonding of metal structures requires strict adherence to proper processes, many pay little attention to the need for proper processing for adhesive bonding of composite adherends [176]. Obviously, surface preparation of composites is critical since, impact damage aside, the only in-service failures of bonded composite structures have been interfacial and most relate to durability problems. The durability issue primarily concerns marginal surface preparations that typically result in some surface contamination [44]. Weak initial bonds cannot be nondestructively tested, and in-service loading may lead to bond failure [176]. Two main techniques are used to prepare thermosetting matrix composites for bonding: (1) the peel-ply method and (2) solvent cleaning and abrasion, often conducted after a peel-ply surface has been exposed [177]. Solvent cleaning followed by mechanical abrasion is the primary means to remove contamination from a composite surface. For badly contaminated surfaces, a solvent-soak process using reagent-grade acetone has been recommended [178]. If the condition of the surface is poor, Scotch-Brite abrasion may be employed. Pumice has also been used as an abrasive. Deionised water should be used, especially for the final rinsing, to prevent surface contamination [178]. The standard waterbreak test can be used to verify cleanliness of the composite surface [176]. Grit-blasting should be conducted after abrasive scrubbing and/or solvent degreasing procedures. A light grit-blast with aluminium oxide results in optimal bond strengths with a minimum of scatter [178]. Practice is essential, and limiting the blast pressure is critical to preventing surface damage. Very little material should be removed, and the blasted surface should have a dull or matt finish [178]. A pressure of 140 kPa has been recommended using No. 280 dry alumina grit [178].

302

Andrew N. Rider et al.

Manual abrasion with 80–120 grit aluminium oxide paper can be employed as an inferior alternative to grit-blasting. If the composite surface ply is unidirectional, sanding should be in parallel with the fibres to minimise damage [177]. Removal of the grit and other debris can be achieved by pressurised jet of clean, dry nitrogen or air, or wiping. Dry-wiping is preferred given the potential for solvent wiping to recontaminate the surface (refer to Section 4.3). Hart-smith et al. [178] suggest a final cleaning operation with isopropyl alcohol since it is more miscible in water than the stronger solvents and can be removed by rinsing with deionised water, and then waterbreak tested. The present authors recommend dry removal of blasting residue followed by a waterbreak test using deionised water. Composite surfaces must be dry as well as cleaned prior to bonding. After the waterbreak test, the surface should be dried at 120°C [176]. However, the real moisture problems result from environmental moisture absorbed by the matrix resin. Drying times depend on the laminate thickness. For example, laminates 6.3 mm thick require 24 h of drying at 135°C to enable the moisture to migrate out. It is most important to dry laminates slowly and thoroughly prior to bonding at elevated temperature. Special care should be taken if honeycomb or foam core is present since any moisture in the cells could convert to steam and destroy the component. Drying temperatures should be limited to 70°C in these cases [178]. Following surface preparation and drying, composite substrates are sometimes primed prior to bonding to take advantage of flow and wetting properties of the primer [178]. However, priming does not necessarily improve bond performance. Laser surface treatments have also been applied to composite materials for repair bonding [179] and fabrication applications [180]. Lasers offer the ability to texture and remove the matrix resin to provide a reliable surface pretreatment method without use of peel plies, chemicals or abrasion methods that can be dependent on operator skills. Some research has also examined the use of atmospheric plasma units to apply surface treatments to composite surfaces prior to bonding [179,181]. The plasma can provide a reliable method to remove contaminant from the surface of composite following chemical preteatments or release agents from peel plies and could potentially increase the functional group concentration on the matrix surface.

Surface Treatment and Repair Bonding

303

11.1 Precured Patches Precured patch preparation typically involves peel-ply removal prior to bonding. Debate continues as to the type of peel ply that should be used and whether additional treatment is needed after peel-ply removal [44,176,182]. The peel ply can prevent gross contamination of the patch surface up to the time it is removed. In principle, the peel ply should tear a very thin layer of resin from the composite to create a fresh clean surface. However, peel plies, particularly nylon [182] and those using release agents, have the potential to transfer material to the composite surface. Peel plies containing silicone release agents must be avoided. Although polyester peel plies seem to be better than nylon, the best approach for patch preparation is a light grit-blasting step [182].

12 ENVIRONMENTAL DURABILITY OF FIELDED REPAIRS The successful application of bonded repairs to a range of aircraft structures over the past 25 years has provided significant opportunities for engineers to assess the relative environmental resistance and overall durability for a range of surface treatment, adhesive, metal and composite combinations. Both residual strength measurement and teardown inspection of bonded structure and repairs that have seen extended service lives are extremely valuable. The results help in developing a database which can directly correlate laboratory accelerated testing with real service performance. A number of recent examples will be briefly described below.

12.1 C-141 Boron Doubler Reinforcements for Lower Wing Panels From 2005 through 2007, the United States Air Force (USAF) conducted a test programme to evaluate the effectiveness and long-term durability of bonded repairs applied to USAF C-141 aircraft during the 1990s. The overarching goal of this programme was to provide information that would enable USAF engineering authorities to have the confidence in adhesively bonded repair technology necessary to use this repair approach, when and where appropriate. The programme included efforts to validate the design, analysis, materials, processes, NDI techniques and related items associated with bonded composite repairs applied to the C-141 aircraft, primarily to address residual strength [183].

304

Andrew N. Rider et al.

In the early 1990s, C-141 aircraft were experiencing structural cracking initiating at cross-drilled holes in the stiffening risers of the lower wing panels, referred to as ‘weep holes’ since they allowed fuel to transfer between the compartments created by the integral risers. A repair was devised consisting of three adhesively bonded boron patches: one patch bonded to each side of the riser on the inside of the wing over the weep hole and crack, and a third patch bonded to the outer mould line (OML) of the lower wing skin beneath the cracked riser (Fig. 22). This repair configuration was utilised for cracks at weep holes that could not be removed by oversizing the hole followed by cold working. About 850 bonded boron patch repairs (over 2500 individual patches) were installed across the C-141 fleet, mostly from 1993 to 1995 [184]. Several organisations installed C-141 weep hole repairs using precured boron/epoxy patches fabricated from Textron/Specialty Materials 5521 prepreg. Patch shapes and sizes varied slightly due to both the organisation executing the repair and repair location on the aircraft. Precured patches were prepared for bonding by removing a peel ply cured with the patch as its outer ply and/or by abrasion. Prebond surface preparation used for the 7075-T6 aluminium wing structure depended on the organisation installing the repair. The vast majority of repairs were installed by the USAF and Composite Technology, Inc. (CTI), and these two organisations used similar grit-blast/silane processes for metal surface preparation [87–89]. Boron patches were bonded using one of three similar 121°C-curing modified epoxy film adhesives.

OML patch

Riser patches (installed inside wing over weep hole)

Test specimen

Fig. 22 C-141 wing panel with overlay showing location of bonded patches and test specimen.

Surface Treatment and Repair Bonding

305

For the test programme, wing panels were ‘harvested’ from C-141 aircraft stored at the USAF Aerospace Maintenance and Regeneration Center (AMARC) at Davis-Monthan Air Force Base near Tucson, Arizona. Repairs were identified for residual strength testing and machined from the wing panels. Care was taken to select repairs made to similar structure and those using similar installation procedures, however small differences did exist between test specimens. NDI of repair areas indicated no crack growth under repair patches and no growth of any acceptable disbonds identified at the time of patch installation. A total of 52 valid residual strength tests were conducted on service-aged repairs having between 2100 h to 7600 flight hours and ranging from just under 9 years to almost 13 years between installation and testing dates [183]. Specimens were tested to failure in tension at ambient laboratory conditions while strains on the patches and surrounding surfaces were monitored using full-field stereo-optic (3-D) strain measurement as shown in Fig. 23 [184]. For the test programme, what is commonly referred to as Design Limit Load (DLL) was defined as Design Limit Stress (DLS) to account for variations in cross-sectional area present among the specimens. DLS is the largest stress a component must withstand during its operational life as a result of mission-related loads, and Design Ultimate Stress (DUS) is defined as 150% of DLS. Successful test specimens were those that failed at or above

Fig. 23 Load frame with test specimen in grips showing speckle pattern for strain measurement. Octagonal-shaped OML patch can be seen in the close-up (inset).

Percent design ultimate stress (%DUS)

306

Andrew N. Rider et al.

200

Failure locations & modes Panel @ discontinuity away from repaired weep hole (fracture or net section yield) Repair @ Bondline or within patch (prior to overall specimen failure at panel discontinuity) Panel @ Saw cuts (atypical) – INVALID TESTS

150

100

Requirement (Design Ultimate Stress)

Design Limit Stress

50

53

51

49

47

45

43

41

39

37

35

33

31

29

27

25

23

21

19

17

15

13

9

11

7

5

3

1

0

Specimen number

Fig. 24 C-141 weep hole bonded repair residual strength.

DUS either at or away from the repair area. To be conservative, the highest level of DLS for the C-141 wing structure was selected. Results for the residual strength tests are shown in Fig. 24. Most residual strength specimens failed well above DUS, although stereo-optic strain measurement indicated some damage to the OML repair patch or adhesive bondline for seven of these prior to specimen failure away from the repair area. An additional two specimens showed indications of damage in the repair area at stresses below the DUS requirement, though overall specimen failure was above DUS and away from the repair area. The cause for one was a separator sheet from the boron/epoxy prepreg left in the laminate between the outermost plies. The other failure was at the interface between the adhesive and bond primer, but its cause could not be determined. Due to the nature of these specimens harvested from actual aircraft structure, failure was at features outside the patch area so ultimate patch strengths were not obtained. The test programme also included unrepaired specimens for baselines and tests of newly repaired specimens following the original procedures, as well as failure predictions and statistical analyses [183]. In addition, a limited amount of fatigue testing was conducted with service-aged and new repairs of similar composition [185]. In all cases, the service-aged repairs behaved similarly to the new repairs. Evaluation of service-aged C-141 weep hole bonded repairs showed no evidence of crack growth under repairs or disbond growth during service.

Surface Treatment and Repair Bonding

307

Residual strength testing of 52 specimens showed the repairs held up remarkably well for the roughly 10 years the repairs were in service. Destructive evaluation of bonded patches revealed no evidence of corrosion in adhesive bondlines or failure at metal-primer interfaces, indicating the grit-blast/silane surface preparation performed well in service for riser patches exposed to fuel and OML patches exposed to the outside environment during service. Despite the results of the test programme, the effort’s overarching goal was not met since USAF engineering authorities did not gain enough confidence in bonded repairs to change their approach to rely on these repairs to carry DLL for safety-of-flight-critical structures. Unfortunately, residual strength tests conducted for the purpose of meeting the programme’s goal were not useful for determining the remaining environmental durability of the repair bondlines. Future testing of C-141 weep hole repairs left intact when residual strength test specimens failed away from the patches may be conducted by the USAF in the future to gain better insight into service-aged adhesive bondlines.

12.2 F-111C Aluminium Doubler Reinforcements to Honeycomb Structure Following the retirement of the RAAF F-111C fleet in December 2010, DST Group undertook a programme to assess the condition of adhesively bonded repairs that had been applied to the aircraft’s external honeycomb panels over the service life [186]. Aluminium doublers were bonded over honeycomb structure to repair different types of service damage. The metal honeycomb panels were used to stiffen the external fuselage or for control surfaces. The programme involved identifying repairs which had been applied using the methods detailed in the RAAF Engineering Standard [27] and, as such, would have been conducted by specially trained technicians under controlled conditions. Repairs were inspected using basic taptesting and then the bondline condition was measured using a Pneumatic Adhesion Tensile Testing Instrument (PATTI) shown in Fig. 25. Following PATTI testing, the repairs were removed from the panel skin and inspected for anomalies. Whilst the PATTI test represents an out-of-plane loading condition not intended in the repair design, recent studies have shown that it is possible to predict the pull-off tensile strength of bonded repairs using cohesive zone modelling [187]. Consequently, it is possible to relate a PATTI test result to an adhesive shear strength if consideration is given to the stress concentration effects caused during loading of thin, flexible substrates during the pull-off test [188].

308

Andrew N. Rider et al.

Pulling -force

Reaction plate Gasket

Pressure hose

Aluminium skin Adhesive

Gasket

Pull stub Adhesive

Routed skin

Honeycomb core

Fig. 25 Configuration of the PATTI tester used to measure the adhesive bond strength of the bonded repair doublers applied to the external skin of damaged honeycomb panels on RAAF F-111C aircraft during the service life.

Laboratory tests were used to establish the variability in the PATTI test results that were inherent in the test and the geometric variations associated with testing on aircraft panels with curvature. It was established that pull-off tensile strengths of any repairs that were below 10 MPa, or that exhibited high levels of adhesion failure, were indicative of reduced adhesion. The reduction in adhesion could either be the result of deficiencies in the initial application and/or service degradation. A total of 236 repairs were tested of which 98 repairs had full service histories. All of the repairs inspected used FM300 adhesive from Cytec and the clad Al-2024T3 skin was treated using the grit-blast and silane treatment (Section 8.3.2). Fig. 26 shows that the average repair strength as a function of the accumulated flight hours did not vary in a consistent trend. A similar result was observed for the repairs with known service lives, but unknown flight hours. It was determined that for repairs that had experienced more than 1500 flight hours or had service lives up to 15 years, there was no indication of strength deterioration caused by exposure time. The average strength was around 15 MPa with a 95% confidence limit of 3 MPa. In the very few cases where individual PATTI tests showed low results, teardown inspection of the bonded repairs revealed that the causes were corrosion, heavy voiding of the bondline, poor adhesive wetting or inadequate grit-blasting. Apart from corrosion, these factors were a result of deficiencies during initial application.

309

Surface Treatment and Repair Bonding

Average repair strength (MPa)

25

20

15

10 y = 0.0006x + 15.198 R2 = 0.0095

5

0 0

500

1000

1500

2000

Aircraft flight hours (days)

Fig. 26 Correlation between flight hours experienced by each F-111C bonded repair and the average repair strength, where the error bars represent the 95% confidence interval for all repairs measured with the same accumulated flight hours. The broken red line represents the lower limit expected for undegraded bonds accounting for possible geometrical variations in loading.

Corrosion was believed to have occurred on aircraft stored after retirement, where paint film degradation and water pooling had led to conditions that would be nonrepresentative of a service aircraft. It should also be noted that the low number of defects identified only occurred in small and localised areas of individual repairs and represented a very low percentage of the surface area of the bondline. The overall conclusion from the repair analysis programme was that bonded repairs provide a reliable technology when applied by trained technicians in fit-for-purpose facilities. There was no clear evidence to indicate that the bonded repairs degraded as a result of conditions experienced during flight or storage of the airframe.

12.3 F-111C Boron Doubler Reinforcements to Lower Wing Skin As a part of the adhesively bonded doubler assessment programme detailed in Section 12.2, three F-111C wings were also recovered that had a bonded composite repair installed to the lower wing skin to restore the residual strength, which had been compromised due to cracking at the Forward

310

Andrew N. Rider et al.

Fig. 27 The FAS 281 boron doubler repair specimen positioned into the 2 MN test machine showing the outer skin with the repair and the inner skin with strain gauges.

Auxiliary Spar Station (FASS) 281 [189]. The 5521/4 boron epoxy doublers were bonded using FM73 adhesive and the aluminium skin was prepared using the grit-blast and silane treatment detailed in Section 8.3.2. Further details of the repair are detailed in Chapter 6. The wing skins with the doublers were excised and cut to dog-bone shape to enable uniaxial testing to determine if the bonded repairs would survive 1.2 times design limit loading (1.2  DLL), as shown in Fig. 27. The testing required careful design of the specimen, grips and constraint fixture to ensure out-of-plane bending was minimised. Strain gauge monitoring and NDI before and after testing confirmed that all three repairs, which had experienced more than 600 aircraft flight hours, survived the 1.2  DLL loading without developing any damage. The testing confirmed the repairs could restore the strength of primary aircraft structures and meet the durability requirements of the airworthiness standards.

12.4 Boron Doubler Reinforcements to Mirage IIIA More than 150 boron doubler reinforcements were applied to the RAAF Mirage IIIA in the late 1970s to reduce cracking from the lower wing skin fuel-decant hole into the main spar [190]. A 7-ply unidirectional boron doubler patch was bonded using 3M AF-126 epoxy adhesive, with the PANTA process (Section 8.2.1) used to prepare the aluminium wing skin [191]. DST group recovered 17 wings in 2001 and undertook inspection using the same

311

Surface Treatment and Repair Bonding

7

14

Mean: 777 Std. dev. 482 Min: 0 Max: 1880

12

6 5

8 1520 psi calibration strength

6

Count

Count

10

3 2

2

1 0

500

1000 POTS (psi)

1500

2000

1520 psi calibration strength

4

4

0

(A)

Mean: 1497 Std. dev. 242 Min: 994 Max: 1908

0

(B)

0

500

1000 1500 POTS (psi)

2000

Fig. 28 Pull-off tensile strength (POTS) measured for PATTI tests conducted on boron doublers adhesively bonded to Mirage IIIA lower wing skins. Strength values for (A) defect regions identified using NDI and (B) in regions where no indications were observed.

methods described in Section 12.2 for the F-111C bonded aluminium doublers. The bonded boron doublers saw flight times up to 2000 h and service lives up to 9 years [192], followed by 12 years storage at the Woomera Defence site in South Australia. Fig. 28 shows the strength distributions measured using the PATTI test across the boron doublers. In regions where no indications were observed (Fig. 28B) from NDI inspection, the average strength is close to that measured for calibration specimens made in the laboratory. In regions with indications (Fig. 28A), there is a reduction in strength which is associated with locations at the edge of the patches where high levels of porosity were observed. While the reduction in strength at some locations on the patch perimeter was marked, these regions accounted for less than 4% of the total bonded area examined. It should be noted these doublers were bonded under field conditions where humidity control was difficult. Despite the porosity observed across the patches, there was very limited evidence of adhesion failure, with only one patch exhibiting an area covering about 5% of the total patch area. Even though the boron patches were bonded in the early stages of the DST Group bonded repair research programme and the RAAF Engineering Standard [27] had yet to be developed, the results show that durable adhesive bonds can be prepared when reliable surface treatment methods are employed. Fig. 29 shows that the PANTA process provides similar performance to the grit-blast and silane treatment in accelerated mode I testing in

312

Andrew N. Rider et al.

2500

GI (J/m2)

2000

1500

GB + PANTA (2024-T6) GB + GPS (2024-T6) GB (2024-T6)

1000

500

0 0

5

10

15

20

25

30

35

Time (h0.5)

Fig. 29 Long crack extension tests for Al2024-T6 aluminium alloy pretreated using the Phosphoric Acid Non-Tank (PANTA), grit-blast and silane (GPS) or Grit-blast (GB) processes and bonded with AF-126. The fracture toughness (GI) was determined in a 50°C/95% R.H. environment.

humid conditions. The evidence from both the F-111C doubler programme and the Mirage IIIA repair inspections suggest that the surface treatments used provide durable adhesive bonds and this can be correlated with accelerated laboratory test results.

12.5 Boron Doubler Reinforcements for C-130E Wing-Risers Boron composite doublers were applied to wing risers on C-130E aircraft between 1978 and 1992 [193]. The doublers were applied ahead of rivet holes to prevent stress corrosion cracking of the Al-7075T6 alloy. The doublers were 25 mm by 45 mm, 4 ply laminates of 5521/4 boron epoxy prepreg, bonded with AF-126 or FM73 epoxy adhesives. The metal pretreatment used methyl ethyl ketone solvent degreasing before and after grit-blasting with 50 μm alumina particles directed with nitrogen propellant of 450 kPa pressure. Pressure and heating during cure of the adhesive was achieved with toggle clamps and silicone rubber heating mats, respectively. DST Group undertook teardown inspection of boron doublers applied to a retired wing between 2003 and 2006 after the C-130E retirement in 2000, using flatwise tension tests to examine the residual strength of the adhesive bond as shown in Fig. 30. The wing-riser containing the patch

313

Surface Treatment and Repair Bonding

Load

FWT block

Al-7075 T6 Wing riser

Paste adhesive Boron patch AF-126/FM73

Fig. 30 Configuration of the flatwise tension test used to measure the adhesive bond strength of boron doublers bonded to C-130E wing-risers, which had been applied between 1978 and 1992.

was cut from the wing and loading blocks bonded to the external doublers were used to test the adhesive bond strength. Flatwise tension testing of AF-126 and FM73 bonded to Al-7075 T6 flatwise tension blocks under laboratory conditions indicated that the bond strength was in excess of 40 MPa. Testing of bonded boron doublers showed they failed in interlaminar mode at 20 MPa or higher. Subsequent testing of C-130E riser doublers from the service wing then assessed any bond degradation for cases where adhesive or adhesion failure occurred at strength values below 20 MPa. The results in Fig. 31 show a considerable number of cases where the strength is below 20 MPa and these represented both adhesion failure at the adhesive and metal interface and cohesive failure within the adhesive. The results show that both the inferior surface treatment and the absence of guidance that would have been provided by the RAAF Engineering Standard have led to the poorer field performance compared to the cases for the Mirage IIIA and F-111C bonded doubler performance discussed in Sections 12.2 and 12.4. The results also are consistent with the mode I accelerated test results in Fig. 29 and provide further support for the correlation in the accelerated testing and field performance of bonded repairs and the importance of the surface treatment.

314

Andrew N. Rider et al.

14 >20 MPa Calibration strength

12

Count

10 8 6

AF126 FM73 Average: 13.6 MPa Std dev: 6.4 MPa

4 2 0 0

5 10 15 20 Flatwise tension strength (MPa)

25

30

Fig. 31 Flatwise tension strengths measured for boron doublers adhesively bonded to C-130E wing risers using either AF-126 or FM73 adhesive and a grit-blast surface treatment. Doublers were bonded to the wing-risers between 1978 and 1992.

REFERENCES [1] J.R. Huntsberger, Interfacial energies, contact angles and adhesion, in: R.L. Patrick (Ed.), Volume 5 – Treatise on Adhesion and Adhesives, Marcel Dekker Inc., New York, 1981. [2] A.A. Baker, Bonded composite repair of primary aircraft structure, Compos. Struct. 47 (1999) 431–443. [3] A.J. Kinloch, Adhesion and Adhesives, Science and Technology, Chapman and Hall, London/New York, 1986, pp. 56–57. [4] N.A. DeBruyne, The adhesive properties of epoxy resins, J. Appl. Chem. 6 (1956) 303. [5] T. Semoto, Y. Tsuji, K. Yoshizawa, Molecular understanding of the adhesive force between a metal oxide surface and an epoxy resin, J. Phys. Chem. C 115 (2011) 11701–11708. [6] T. Semoto, Y. Tsuji, K. Yoshizawa, Molecular understanding of the adhesive force between a metal oxide surface and an epoxy resin: effects of surface water, Bull. Chem. Soc. Jpn. 85 (2012) 672–678. [7] M.-L. Abel, R.P. Digby, I.W. Fletcher, J.F. Watts, Evidence of specific interaction between γ-glycidoxypropyltrimethoxysilane and oxidized aluminium using high-mass resolution ToF-SIMS, Surf. Interface Anal. 29 (2000) 115–125. [8] J.R. Huntsberger, Surface energy, wetting and adhesion, J. Adhes. 12 (1981) 3–12. [9] P.C. Hiemenz, Principles of Colloid and Surface Chemistry revised and expanded second ed., Marcel Dekker, New York and Basel, 1986, (chapter 6). [10] R.N. Wenzel, Resistance of solid surfaces to wetting by water, Ind. Eng. Chem. 28 (1936) 988. [11] W.D. Bascom, R.L. Partick, The surface chemistry of bonding metals with polymer adhesives, Adhes. Age 17 (10) (1974) 25. [12] N.A. DeBruyne, Aero Research Tech Notes No 168 (1956) p. 1.

Surface Treatment and Repair Bonding

315

[13] D.E. Packham, in: K. Mittal (Ed.), Adhesion Aspects of Polymeric Coatings, Plenum, New York, 1983, p. 19. [14] D.R. Arnott, A.R. Wilson, P.J. Pearce, G. Mathys, M.R. Kindermann, A. Camilleri, M.J. Davis, G. Swan, Void development in aerospace film adhesives during vacuum bag cure, International Aerospace Congress 97, Sydney Australia 24–27 February 1997 p. 15. [15] A.N. Rider, E. Yeo, The Chemical Resistance of Epoxy Adhesive Joints Exposed to Aviation Fuel and Its Additives, DSTO, Melbourne, Australia, 2005. [16] Y. Sugita, C. Winkelmann, V. La Saponara, Environmental and chemical degradation of carbon/epoxy lap joints for aerospace applications, and effects on their mechanical performance, Compos. Sci. Technol. 70 (2010) 829–839. [17] D.R. Arnott, A.R. Wilson, A.N. Rider, L.T. Lambrianidis, N.G. Farr, Studies of the degradation of metal-adhesive interfaces with surface analysis techniques, Appl. Surf. Sci. 70 (71) (1993) 109. [18] J.D. Minford, Handbook of Aluminium Bonding Technology and Data, Marcel Dekker, New York, 1993. [19] A.J. Kinloch, Adhesion and Adhesives-Science and Technology, Chapman and Hall, London, 1987. [20] R. Bossi, K. Housen, C.T. Walters, D. Sokol, Laser bond testing, Mater. Eval. 67 (7) (2009) 819–827. [21] M.J. Piehl, et al., Validation of laser bond inspection technology (FA8650-15-C5701), 2016 Defense Manufacturing Conference, Denver, CO, 29 November 2016. [22] M. Job, Air Disaster Volume 2, Aerospace Publications, Weston Creek ACT, 1996. Chapter 11. [23] M.J. Davis, Deficiencies in regulations for certification and continuing airworthiness of bonded structures, International Aerospace Congress 97, Sydney Australia, IEAust, 24–27 February 1997 p. 215. [24] M.J. Davis, Proceedings 41st International SAMPE Symposium 1996, p. 936. [25] M.J. Davis, Proceedings International Symposium on Composite Repair of Aircraft Structure, Vancouver, 9–11 August 1995. [26] L.J. Hart-Smith, M.J. Davis, Proceedings of 41st International SAMPE Symposium and Exhibition, Anaheim 25–28 March 1996. [27] Royal Australian Air Force, Composite Materials and Adhesive Bonded Repairs Standard, Report No. DEF(AUST)9005, 2006. [28] P.J. Pearce, A. Camilleri, C.L. Olsson-Jacques, D.R. Arnott, A Benchmarking Review of RAAF Structural Adhesive Bond Procedures, DSTO-Aeronautical and Maritime Research Laboratory, Report DSTO-TR-0267, 1999. [29] A.N. Rider, R. Vodicka, G.I. Mathys, I. Stoyanovski, Review of RAAF Procedures for Qualifying Bonded Repair Technicians, Report DSTO-TR-1876, Defence Science and Technology Organisation, 2006. [30] P.R. Underhill, A.N. Rider, P. Livingstone, Fatigue behaviour of aluminum bonded joints as a function of wedge test performance, J. Adhes. Sci. Technol. 23 (4), 555–566. [31] R.G. Dillingham, E.S. Oseas, A.D. Gilpin, F.C. Ganance, Device and Method to Measure Wetting Characteristics. US Patent 8,272,254, 2012. [32] G. Gurney, R. Amling, Adhesion Fundamentals and Practice, McLaren, New York, 1969, pp. 211–217. [33] A.A. Baker, in: A.A. Baker, R. Jones (Eds.), Bonded Repair of Aircraft Structures, Dortrecht, Martinius Nijhoff, 1988, p. 118. [34] I.A. Ashcroft, R. Digby, S.J. Shaw, Accelerated ageing and life prediction of adhesively-bonded joints, Abstracts Euradhesion 98/WCARP 1, Garmisch Partenkirchen Germany, 6–11 September 1998 p. 285.

316

Andrew N. Rider et al.

[35] A. Rider, I. Williams, E. Shum, L. Mirabella, Environmental Durability Trial of Bonded Composite Repairs to Metallic Aircraft Structure, Report DSTO-TR1685, 2005. [36] J.J. Mazza, K.M. Storage, J.A. Smith, P.K. Childers, Evaluation of adhesive bond durability in a marine environment. CAMX 2014-Composites and Advanced Materials Expo: Combined Strength. Unsurpassed Innovation, 2014. [37] D.A. Hardwick, J.S. Ahearn, J.D. Venables, Environmental durability of aluminium adhesive joints protected with hydration inhibitors, J. Mater. Sci. 19 (1984) 223. [38] J. Cognard, The mechanics of the wedge test, J. Adhes. 20 (1986) 1. [39] D. Broek, Elementary Engineering Fracture Mechanics, Martinius Nijhoff, Hingham, Mass, USA, 1986, p. 128. [40] S. Bistac, J. Guillemenet, J. Schultz, Fracture of metal/polymer/metal assemblies: viscoelastic effects, J. Adhes. 78 (11) (2002) 987–996. [41] M.H. Stone, T. Peet, Evaluation of the Wedge Test For Assessment of Durability of Adhesive Bonded Joints Royal Aircraft Establishment Tech Memo, Mat 349, 1980. [42] S. Mostovoy, P.B. Crosley, E.J. Ripling, Use of crack-line-loaded specimens for measuring plane-strain fracture toughness, J. Mater. 2 (3) (1967) 661. [43] J. Grosko, Georgia Division (Communication). Lockheed Aeronautical Systems Co., Georgia, USA. [44] A.J. Kinloch, Adhesion and Adhesives Science and Technology, Chapman and Hall, London, 1987, pp. 123–151. [45] A.N. Rider, D.R. Arnott, The influence of adherend topography on the fracture toughness of aluminium-epoxy adhesive joints in humid environments, J. Adhes. 75 (2001) 203–228. [46] A.N. Rider, Surface Properties Influencing the Fracture Toughness of Aluminium Epoxy Joints (Ph.D. thesis), University of New South Wales, Australia, 1998. [47] E.D. Reedy, Effect of patterned nanoscale interfacial roughness on interfacial toughness: a finite element analysis, J. Mater. Res. 23 (2008) 3056–3065. [48] R. Atif, F. Inam, Influence of macro-topography on damage tolerance and fracture toughness of monolithic epoxy for tribological applications, J. Eng. Technol. 4 (2016) 335–360. [49] T. Suzuki, R. Matsuzaki, A. Todoroki, Y. Mizutani, Prediction of the macroscopic fracture toughness of a composite/adhesive interface with periodic surface microstructures, Int. J. Adhes. Adhes. 60 (2015) 16–22. [50] J. Salstela, M. Suvanto, T.T. Pakkanen, Influence of hierarchical micro-micro patterning and chemical modifications on adhesion between aluminum and epoxy, Int. J. Adhes. Adhes. 66 (2016) 128–137, https://doi.org/10.1016/j.ijadhadh.2015.12.036. [51] R.G. Schmidt, J.P. Bell, Epoxy adhesion to metals, in: K. Dusˇek (Ed.), Epoxy Resins and Composites II. Advances in Polymer Science, vol. 75, Springer, Berlin, Heidelberg, 1986, p. 41. [52] A. Rider, N. Brack, S. Andres, P. Pigram, The influence of hydroxyl group concentration on epoxy–aluminium bond durability, J. Adhes. Sci. Technol. 18 (2004) 1123–1152. [53] D.R. Arnott, A.N. Rider, C.L. Olsson-Jacques, L.T. Lambrianidis, A.R. Wilson, P.J. Pearce, R.J. Chester, A.A. Baker, C.E.M. Morris, M.J. Davis, G.R. Swan, Bond Durability Performance – The Australian Silane Surface Treatment, 21st Congress of ICAS, 13–18 September, Melbourne Australia, 1998. [54] S.G. Hong, F.J. Boerio, Adhesive bonding of oil-contaminated steel substrates, J. Adhes. 32 (2–3) (1990) 67. [55] C.L. Olsson-Jacques, A.R. Wilson, A.N. Rider, D.R. Arnott, The effect of contaminant on the durability of bonds formed with epoxy adhesive bonds with alclad aluminium alloy, Surf. Interface Anal. 24 (9) (1996) 569–577.

Surface Treatment and Repair Bonding

317

[56] D.R. Arnott, A.R. Wilson, A.N. Rider, C.L. Olsson-Jacques, L.T. Lambrianidis, P.J. Pearce, M.J. Davis, G.R. Swan, Research Underpinning the Adherend Surface Preparation Aspects of the RAAF Engineering Standard C5033, International Aerospace Congress 97, Sydney Australia 24–27 1997 P. 41. [57] J.D. Venables, Adhesion and durability of metal-polymer bonds, J. Mater. Sci. 19 (8) (1984) 2431–2453. [58] A.N. Rider, D.R. Arnott, A.R. Wilson, O. Vargas, Influence of simple surface treatments on the durability of bonded aluminium alloy plates, Mater. Sci. Forum 189–190 (1995) 235–240. [59] A.N. Rider, D.R. Arnott, Durability of bonds formed between epoxy adhesive and aluminium alloy treated with phosphonate inhibitors, Surf. Interface Anal. 24 (9) (1996) 583. [60] D.R. Arnott, M.R. Kindermann, Durability testing of epoxy adhesive bonds, J. Adhes. 48 (1–4) (1995) 101–119. [61] J. Cognard, Cleavage and static toughness, J. Adhes. 57 (1–4) (1996) 31. [62] J.R. Evans, D.E. Packham, Adhesion of polyethylene to metals: the role of surface topography, J. Adhes. 10 (3) (1979) 177. [63] D.E. Packham, Preparation of metallic surfaces with microfibrous topography and their effect on adhesion of hot-melt and thermoset adhesives, Int. J. Adhes. Adhes. 2 (4) (1986) 225–228. [64] Y. Hua, A.D. Crocombe, M.A. Wahab, I.A. Ashcroft, Continuum damage modelling of environmental degradation in joints bonded with EA9321 epoxy adhesive, Int. J. Adhes. Adhes. 28 (2008) 302–313. [65] X. Han, A.D. Crocombe, S.N.R. Anwar, P. Hu, W.D. Li, The effect of a hot-wet environment on adhesively bonded joints under a sustained load, J. Adhes. 90 (2014) 420–436. [66] K.B. Katnam, J.P. Sargent, A.D. Crocombe, H. Khoramishad, I. A. Ashcroft, Characterisation of moisture-dependent cohesive zone properties for adhesively bonded joints, Eng. Fract. Mech. 77 (2010) 3105–3119. [67] M.-L. Abel, R.D. Allington, R.P. Digby, N. Porritt, S.J. Shaw, J. F. Watts, Understanding the relationship between silane application conditions, bond durability and locus of failure, Int. J. Adhes. Adhes. 26 (2006) 2–15. [68] J. Von Grote, C. H€ urlimann, M. Scheringer, K. Hungerb€ uhler, Reduction of occupational exposure to perchloroethylene and trichloroethylene in metal degreasing over the last 30 years: influences of technology innovation and legislation, J. Expo. Anal. Environ. Epidemiol. 13 (2003) 325–340. [69] J. Lavoue, D. Begin, M. Gerin, Technical, occupational health and environmental aspects of metal degreasing with aqueous cleaners, Ann. Occup. Hyg. 47 (2003) 441–459. [70] J. Tourigny, Vapour degreasers make a comeback, Cleanroom Technol. 22 (2014) 22–24. [71] G. Paggiola, S. Van Stempvoort, J. Bustamante, J.M.V. Barbero, A.J. Hunt, J. H. Clark, Can bio-based chemicals meet demand? Global and regional case-study around citrus waste-derived limonene as a solvent for cleaning applications, Biofuels Bioprod. Biorefin. 10 (2016) 686–698. [72] J.H. Clark, T.J. Farmer, A.J. Hunt, J. Sherwood, Opportunities for bio-based solvents created as petrochemical and fuel products transition towards renewable resources, Int. J. Mol. Sci. 16 (2015) 17101–17159. [73] A.R. Wilson, N.G. Farr, D.R. Arnott, L.T. Lambrianidis, A.N. Rider, J. Shenton, Relationship of surface preparation of clad 2024 aluminium alloy to morphology and adhesive bond strength, Australian Aeronautical Conference Melbourne 9–11 October 1989. pp 221–225

318

Andrew N. Rider et al.

[74] L.T. Lambrianidis, D. R. Arnott, A.R. Wilson, J. van den Berg, O. Vargas, The effect and evaluation of grit-blast severity on adhesive bond durability for aircraft repairs, Second Pacific and International Conference on Aerospace Science and Technology and Sixth Australian Aeronautical Conference, 20–23 March, Melbourne Australia 1995, pp. 355–360. [75] D.R. Arnott, P.J. Pearce, A.R. Wilson, R.J. Chester, R.J. Camilleri, J. van den Berg, The effect on mechanical properties of void formation during vacuum bag processing of epoxy film adhesives, Proceedings Second Pacific and International Conference on Aerospace Science and Technology, The Australian Institute of Engineers, Melbourne, 21–23 March 1995, pp. 811–816. [76] P.J. Pearce, D.R. Arnott, A. Camilleri, M.R. Kindermann, G.I. Mathys, A. R. Wilson, Cause and effect of void formation during vacuum bag curing of epoxy film adhesives, J. Adhes. Sci. Technol. 12 (6) (1998) 567. [77] D.R. Arnott, W.J. Baxter, S.R. Rouze, The role of defects in the fracture of oxide films on metals, J. Electrochem. Soc. 128 (4) (1981) 843–847. [78] J.D. Venables, Adhesion and durability of metal-polymer bonds, J. Mater. Sci. 19 (1984) 2431–2453. [79] A. Baker, R. Chester, R. Solly, Bonded Repair of a Gun Carriage Using Electroformed Nickel Reinforcements, Air Vehicles Division, Defence Science and Technology Organisation, DSTO-TR-1930, 2006. [80] H.M. Clearfield, D.K. McNamara, G.D. Davis, in: H.F. Brinson (Ed.), Engineered Materials Handbook, vol. 3 Adhesives and Sealants, ASM International, Materials Park, OH, 1990, p. 259. [81] A.H. Landrock, Adhesives Technology Handbook, Noyes Publications, Park Ridge, NJ, 1985, p. 66. [82] J.A. Marceau, in: E.W. Thrall, R.W. Shannon (Eds.), Adhesive Bonding of Aluminum Alloys, 51 Marcel Dekker, Inc., New York, 1985, p. 51 [83] L. Young, Anodic Oxide Films, Academic Press, New York, 1961, pp. 1–3. [84] R.G. Schmidt, J.P. Bell, Epoxy adhesion to metals, in: K. Dusˇek (Ed.), Epoxy Resins and Composites II. Advances in Polymer Science, vol. 75, Springer, Berlin, Heidelberg, 1986, pp. 33. [85] G.D. Davis, J.S. Ahearn, L.J. Matienzo, J.D. Venables, Use of hydration inhibitors to improve bond durability of aluminium adhesive joints, J. Mater. Sci. 20 (1985) 975. [86] A.N. Rider, Factors influencing the durability of epoxy adhesion to silane pretreated aluminium, Int. J. Adhes. Adhes. 26 (1) (2006) 67–78. [87] R.J. Kuhbander, J.J. Mazza, Proceedings 38th International SAMPE Symposium, May 10–13 (1993) p. 1225. [88] A.A. Baker, R.J. Chester, Int. J. Adhes. Adhes. 12 (1992) 73. [89] J.J. Mazza, J.B. Avram, R.J. Kuhbander, Grit-blast/Silane (GBS) Aluminium Surface Preparation for Structural Adhesive Bonding, WL-TR-94-4111 (Interim Report under U.S. Air Force Contracts F33615-89-C-5643 and F33615-95-D-5617). [90] R.P. Digby, S.J. Shaw, The international collaborative programme on organosilane coupling agents: an introduction, Int. J. Adhes. Adhes. 18 (4) (1998) 261–264. [91] H.M. Clearfield, D.K. McNamara, G.D. Davis, Engineered Materials Handbook, vol. 3 Adhesives and Sealants, H.F. Brinson (technical chairman), ASM International, Materials Park, OH (1990), p. 254. [92] A. Rider, P. Chalkley, Durability of an off-optimum cured aluminium joint, Int. J. Adhes. Adhes. 24 (2004) 95–106. [93] J.J. Mazza, K.M. Storage, D.B. McCray, J.A. Smith, 2010. The evaluation of adhesive bond primers for repair bonding of aluminum alloys, in: International SAMPE Technical Conference. Presented at the 2010 SAMPE Fall Technical Conference and Exhibition.

Surface Treatment and Repair Bonding

319

[94] J. Aakkula, O. Saarela, Silane based field level surface treatment methods for aluminium, titanium and steel bonding, Int. J. Adhes. Adhes. 48 (2014) 268–279. [95] L. Chen, D. Salnikov, K. Hansen, M. Newman, P. Rajtar, C.A. Praggastis, J. Orrock, J. Stone, S. Kazle, Development of a water based, non-chromated, sol–gel compatible adhesive bonding primer, in: International SAMPE Technical Conference. Presented at the SAMPE Baltimore 2015 Conference and Exhibition, Soc. for the Advancement of Material and Process Engineering, 2015. [96] W.A. Dukes, R.W. Brient, The effect of adhesive thickness on joint strength, J. Adhes. 1 (1969) 48. [97] H.F. Wolfe, C.L. Rupert, H.S. Schwartz, AFWAL-TR-81-3096, 1981. [98] L.F.M. Da Silva, G.W. Critchlow, M.A.V. Figueiredo, Parametric study of adhesively bonded single lap joints by the taguchi method, J. Adhes. Sci. Technol. 22 (2008) 1477–1494. [99] Royal Australian Air Force, Amberley Air Force Base, Internal Communication. [100] S. Budhe, A. Rodrı´guez-Bellido, J. Renart, J.A. Mayugo, J. Costa, Influence of prebond moisture in the adherents on the fracture toughness of bonded joints for composite repairs, Int. J. Adhes. Adhes. 49 (2014) 80–89. [101] A.R. Wilson, M.R. Kindermann, D.R. Arnott, Void development in an epoxy film adhesive during vaccum bag cure, Proceedings Second Pacific and International Conference on a Melbourneerospace Science and Technology, The Institution of Engineers Australia, 20–23 March 1995, pp. 625–630. [102] M. Preau, P. Hubert, Processing of co-bonded scarf repairs: void reduction strategies and influence on strength recovery, Compos. A: Appl. Sci. Manuf. 84 (2016) 336–345. [103] R.J. Crossley, S. Ratchev, A. Smith, Emerging technologies for use in aerospace bonded assemblies, SAE Int. J. Aerosp. 6 (2013) 383–391, https://doi.org/ 10.4271/2013-01-2134. [104] P.F.A. Bijlmer, in: K.L. Mittal (Ed.), Characterisation of the Surface Quality by Means of Surface potential Difference in Surface Contamination, Genesis Detection and Control, vol. 2, Plenum Press, New York, 1979, p. 723. [105] T. Smith, Photoelectron emission from aluminum and nickel measured in air, J. Appl. Phys. 46 (1975) 1553. [106] R. Gause, A Non Contacting Scanning Photoelectron Emission Technique for Bonding Surface Cleanliness Inspection Fifth Annual NASA NDE Workshop, Cocoa Beach, Florida, 1–3 December (1987). [107] Photo Emission Technology, 766 Lakefield Rd. Suite h, Westlake CA 91361. [108] R.C. Weast (Ed.), CRC Handbook of Chemistry and Physics, 54th ed., Chemical Rubber Co., 1973/74, p. E80 [109] C.L. Olsson-Jacques, D.R. Arnott, L.T. Lambrianidis, A.R. Wilson, M.R. Kindermann, G. Theodossiou, Toward quality monitoring of adherend surfaces prior to adhesive bonding in aircraft repairs, The International Aerospace Congress 1997 – 7th Australian Aeronautical Conference, Sydney Australia, 24–27 February 1997, pp. 511–520. [110] A.N. Rider, Prebond Inspection Techniques to Improve the Quality of Adhesive Bonding Surface Treatments. Air Vehicles Division, Defence Science and Technology Organisation, Report DSTO-TR-1919, 2006. [111] Foster Miller Inc. 195 Bearhill Rd. Waltham MA 02451-1003 and [email protected] [112] J. Seelenbinder, Detection of Trace Contamination on Metal Surfaces Using the Handheld Agilent 4100 ExoScan FTIR, Agilent Technologies, Connecticut, 2011. [113] J.D. Minford, Handbook of Aluminium Bonding Technology and Data, Marcel Dekker, New York, 1993, p. 58.

320

Andrew N. Rider et al.

[114] A. Baker, A.J. Gunnion, J. Wang, P. Chang, Advances in the proof test for certification of bonded repairs–increasing the technology readiness level, Int. J. Adhes. Adhes. 64 (2016) 128–141. [115] H.M. Clearfield, D.K. Mcnamara, G.D. Davis, Engineered Materials Handbook, vol. 3 Adhesives and Sealants, H. F. Brinson (technical chairman), ASM International, Metals Park, OH, 261, (1990). [116] A.J. Kinloch, Adhesion and Adhesives Science and Technology, Chapman and Hall, London, 1987, pp. 101–103. [117] E.W. Thrall, Failures in Adhesively Bonded Structures (Lecture No. 5), Douglas Paper 6703, Presented to AGARD-NATO Lecture Series 102: Bonded Joints and Preparation for Bonding, Oslo Norway and The Hague, Netherlands, 1979, p. 2–3. [118] R.W. Shannon, et al., Primary Adhesively Bonded Structure Technology (PABST) General Material Property Data, AFFDL-TR-77-107 (Report for U.S. Air Force Contract F33615-75-C-3016), 1978. [119] T.J. Reinhart, A.A. Baker, R. Jones (Eds.), Bonded Repair of Aircraft Strucures, Martinus Nijhoff Publishers, Dordrecht, The Netherlands, 1988, p. 23. [120] H.M. Clearfield, D.K. Mcnamara, G.D. Davis, Engineered Materials Handbook, vol. 3 Adhesives and Sealants, H. F. Brinson (technical chairman), ASM International, Materials Park, OH (1990), p. 260. [121] ASTM D 3933-93, 1997 Standard Guide for Preparation of Aluminum Surfaces for Structural Adhesives Bonding (Phosphoric Acid Anodising), Annual Book of ASTM Standards, vol. 15.06, American Society for Testing and Materials, West Condshohocken, PA, 1997, pp. 287–290. [122] C. Griffen, D.R. Askins, Non-Chromate Surface Preparation of Aluminum, AFWAL-TR-88-4135 (Interim Report for U.S. Air Force Contract No. F3361584-C-5130), August 1988. [123] J.A. Marceau, in: E.W. Thrall, R.W. Shannon (Eds.), Adhesive Bonding of Aluminum Alloys, Marcel Dekker, Inc., New York, 1985, p. 55 [124] D.R. Askins, D.R. Byrge, Evaluation of 350°F Curing Adhesive Systems on Phosphoric Acid Anodised Aluminum Substrates, AFWAL-TR-86-4039 (Interim Report for U.S. Air Force Contract Nos. F33615-82-C-5039 and F33615-84-C-5130), 1986. [125] E.E. Peterson, D.B. Arnold, M.C. Locke, Compatibility of 350°F curing honeycomb adhesives with phosphoric acid anodising, Proceedings of 13th National SAMPE Technical Conference, (1981), pp. 177–188. [126] M.H. Kuperman, R.E. Horton, in: E.W. Thrall, R.W. Shannon (Eds.), Adhesive Bonding of Aluminum Alloys, Marcel Dekker, Inc., New York, 1985, pp. 430–446 [127] P.F. Bijlmer, in: E.W. Thrall, R.W. Shannon (Eds.), Adhesive Bonding of Aluminum Alloys, Marcel Dekker, Inc., New York, 1985, pp. 28–32 [128] N.L. Rogers, in: E.W. Thrall, R.W. Shannon (Eds.), Adhesive Bonding of Aluminum Alloys, Marcel Dekker, Inc., New York, 1985, pp. 41–49 [129] N.L. Rogers, J. Appl. Polym. Sci.: Appl. Polym. Symp. 32, Wiley, New York, 1977, 37–50. [130] E.W. Thrall, Jr., Failures in Adhesively-Bonded Structures, Douglas Aircraft Company Paper 6703, (1979), pp. 2–3. [131] G.B. Gaskin, et al., Investigation of Sulfuric-Boric Acid Anodizing as a Replacement for Chromic Acid Anodization: PhaseI, Proceedings 26th International SAMPE Technical Conference, Atlanta GA, 1994, p. 258–264. [132] H.M. Clearfield, D.K. McNamara, G.D. Davis, Engineered Materials Handbook, vol. 3 Adhesives and Sealants, H. F. Brinson (technical chairman), ASM International, Materials Park, OH (1990), pp. 260–263. [133] D.E. Packham, in: D.E. Packham (Ed.), Handbook of Adhesion, Longman Scientific & Technical, UK, 1992, p. 201.

Surface Treatment and Repair Bonding

321

[134] K.M. Adelson, E.A. Garnis, R.F. Wegman, Evaluation of the Sulfuric Acid-Ferric Sulfate and Phosphoric Acid Anodise Treatments prior to Adhesive Bonding of Aluminum, ARSCD-TR-82013 (U.S. Army Final Report), September 1982. [135] W.B. Pinnell, Hydrogen Embrittlement of Metal Fasteners Due to PACS Exposure, AFRL-ML-WP-TR-2000-4153, vol. 5, (1999) (Report for Delivery Order 0004, Task 2 of U.S. Air Force Contract F33615-95-D-5616). [136] M.C. Locke, W.M. Scardino, Anodize Optimization and Adhesive Evaluations for Repair Applications, AFML-TR-78-104, July 1978. [137] I. Pergan, On-aircraft phosphoric acid anodizing, Int. J. Adhes. Adhes. 19 (2–3) (1999) 199–204. [138] S.S. Saliba, Phosphoric acid containment system (PACS) evaluation for on-aircraft anodisation of aluminum surfaces, Proceedings of 38th International SAMPE Symposium, vol. 38, (1993), pp 1211–1224. [139] E.A. Podoba, D. McNamara, et al., Adhesive bond durability with conversion coatings, Appl. Surf. Sci. 9 (1–4) (1981) 359–376. [140] M.H. Kuperman, R.E. Horton, in: E.W. Thrall, R.W. Shannon (Eds.), Adhesive Bonding of Aluminum Alloys, Marcel Dekker, Inc., New York, 1985, pp. 430–446 [141] J.J. Mazza, Advanced Surface Preparation for Metal Alloys, AGARD Report 816, 1997, pp. 10-1–10-12. [142] K.Y. Blohowiak, Sol–gel technology for space applications, Proceedings National Space and Missile Materials Symposium, San Diego CA, 27 February–2 March 2000. [143] C.F. Baes, R.E. Mesmar, in: C.J. Brinker, G.W. Scherer (Eds.), Sol–Gel Science-The Physics and Chemistry of Sol–Gel Processing, Academic Press, San Diego, 1990. [144] T. Tiano, M. Pan, W. Dorogy, K. Jayaraj, Functionally Gradient Sol–Gel Coatings for Aircraft Aluminum Alloys, WL-TR-96-4108 (Final Report for U.S. Air Force Contract F33615-95-C-5621), 1996. [145] C. Chu, H. Zheng, Sol–Gel Deposition of Active Alumina Coatings on Aluminum Alloys, WL-TR-97-4114 (Final Report for U.S. Air Force Contract F33615-94C-5605), 1997. [146] K.Y. Blohowiak, J.H. Osborne, K.A. Krienke, US Patents 6,037,060 (2000), 5,958,578 (1999), 5,939197 (1999), 5,869,141 (1999), 5,869,140 (1999), 5,849,110 (1998), 5,814,137 (1998). [147] J. Mazza, G. Gaskin, W. Depiero, K. Blohowiak, Faster durable bonded repairs using sol–gel surface treatment, Proceedings of the Fourth Joint DoD/FAA/NASA Conference on Aging Aircraft, St. Louis MO 2000. [148] D.B. McCray, J.J. Mazza, Optimization of sol–gel surface preparations for repair bonding of aluminum alloys, Proceedings 45th International SAMPE Symposium and Exhibition, Long Beach CA 2000, p. 53–54. [149] D.B. McCray, et al., An ambient-temperature adhesive bonded repair process for aluminum alloys, Proceedings 46th International SAMPE Symposium and Exhibition, Long Beach, CA, 2001, p. 1135–1147. [150] K.Y. Blohowiak, R.A. Anderson, R.R. Stephenson, Sol–Gel Technology for Surface Preparation of Metal Alloys for Adhesive Bonding and Sealing Operations, AFRLML-WP-TR-2003-4169, 2001. [151] J.J. Mazza, G.B. Gaskin, W.S. De Piero, K.Y. Blohowiak, Sol–Gel Technology for Low VOC, Nonchromated Adhesive Bonding Applications, AFRL-ML-WP-TR2004-4063, 2004. [152] B. Daniel, McCray, The Development of On-Aircraft Surface Preparations Utilizing Sol–Gel Coatings for Adhesive Bonding Aluminum Alloys, AFRL-ML-WP-TR2002-4131, 2002.

322

Andrew N. Rider et al.

[153] K.Y. Blohowiak, R.A. Anderson, R.R. Stephenson, K.S. Preedy, Advanced Sol–Gel Adhesion Processes AFRL-ML-WP-TR-2003-4173, 2003. [154] D.B. McCray, J.A. Smith, P.K. Childers, Support for Aging Aircraft Technology Transition, AFRL-ML-WP-TR-2006-4057, 2005. [155] B. Daniel, McCray, The Evaluation of Ambient Temperature Processes for Repair Bonding of Aluminium Alloy, AFRL-ML-WP-TR-2002-4043, 2002. [156] M.C. Locke, R.E. Horton, J.E. McCarty, Anodize Optimization and Adhesive Evaluations for Repair Applications, AFML-TR-78-104 (Final Report for U.S. Air Force Contract F33615-73-C-5171), 1978. [157] A.N. Rider, D.R. Arnott, Int. J. Adhes. Adhes. 20 (2000) 209. [158] A.N. Rider, The influence of porosity and morphology of hydrated oxide films on epoxy-aluminium bond durability, J. Adhes. Sci. Technol. 15 (2001) 395–422, https://doi.org/10.1163/156856101300157524. [159] P.R. Underhill, A.N. Rider, D.L. DuQuesnay, Warm water treatment of aluminum for adhesive bonding, Int. J. Adhes. Adhes. 23 (2003) 307–313, https://doi.org/ 10.1016/S0143-7496(03)00048-4. [160] D.K. Shaffer, H.M. Clearfield, J.S. Ahearn, in: J.D. Minford (Ed.), Treatise on Adhesion and Adhesives, vol. 7, Marcel Dekker, Inc., New York, 1991, pp. 437–444 [161] H.M. Clearfield, et al., J. Adhes. 29 (1989) 81–102. [162] S.R. Brown, G.J. Pilla, Titanium Surface Treatments for Adhesive Bonding, NADC82032-60 (Phase Report for U.S. Navy Airtask No. WF61-542-001), 1982. [163] T. Mertens, F.J. Gammel, M. Kolb, O. Rohr, L. Kotte, S. Tsch€ ocke, S. Kaskel, U. Krupp, Investigation of surface pre-treatments for the structural bonding of titanium, Int. J. Adhes. Adhes. 34 (2012) 46–54, https://doi.org/10.1016/j.ijadhadh.2011.12.007. [164] M. Marı´n-Sa´nchez, A. Conde, M. Garcı´a-Rubio, A. Lavia, I. Garcı´a, Durability of titanium adhesive bonds with surface pretreatments based on alkaline anodisation, Int. J. Adhes. Adhes. 70 (2016) 225–233, https://doi.org/10.1016/j.ijadhadh.2016.07.001. [165] Semco®, Pasa-Jell 107 Technical Data Sheet, 1996. [166] TURCO®, 5578 Technical Data Bulletin, 1999. [167] N. Brack, A.N. Rider, The influence of mechanical and chemical treatments on the environmental resistance of epoxy adhesive bonds to titanium. Int. J. Adhes. Adhes. 48 (2014) 20–27, https://doi.org/10.1016/j.ijadhadh.2013.09.012. [168] C.T. Walters, J.A. Smith, 2007. Laser pretreatment of surfaces for adhesive bonding, in: International SAMPE Technical Conference. Presented at the 39th International SAMPE Technical Conference-From Art to Science: Advancing Materials and Process Engineering. [169] F.L. Palmieri, K.A. Watson, G. Morales, T. Williams, R. Hicks, C.J. Wohl, J. W. Hopkins, J.W. Connell, Laser ablative surface treatment for enhanced bonding of Ti-6Al-4V alloy, ACS Appl. Mater. Interfaces 5 (2013) 1254–1261, https://doi. org/10.1021/am302293m. [170] F.L. Palmieri, A. Crow, A. Zetterberg, J. Hopkins, C.J. Wohl, J.W. Connell, M.A. Blecher, K.Y. Blohowiak, 2014. Further investigation into the use of laser surface preparation of TI-6AL-4V alloy for adhesive bonding, in: International SAMPE Technical Conference. Presented at the SAMPE Tech Seattle 2014 Conference, Soc. for the Advancement of Material and Process Engineering. [171] H.M. Clearfield, D.K. McNamara, G.D. Davis, Engineered Materials Handbook, vol. 3 Adhesives and Sealants, H. F. Brinson (technical chairman), ASM International, Materials Park, OH (1990) p 264-273. [172] R.C. Snogren, Handbook of Surface Preparation, Palmerton Publishing Co., Inc., New York, 1974, p. 265 [173] A.H. Landrock, Adhesives Technology Handbook, Noyes Publications, Park Ridge, NJ, 1985, pp. 72–75. [174] R.F. Wegman, Surface Preparation Techniques for Adhesive Bonding, Noyes Publications, Park Ridge, NJ, 1989, pp. 66–70.

Surface Treatment and Repair Bonding

323

[175] A.A. Baker, R.J. Chester, M.J. Davis, J.D. Roberts, J.A. Retchford, Reinforcement of the F-111 wing pivot fitting with a boron/epoxy doubler system—materials engineering aspects, Composites 24 (6) (1993), 511–521. [176] L.J. Hart-Smith, D. Brown, S. Wong, in: S.T. Peters (Ed.), Handbook of Composites, Chapman and Hall, London, 1998, pp. 667–685. [177] A.H. Landrock, Adhesives Technology Handbook, Noyes Publications, Park Ridge, NJ, 1985, pp. 105–106. [178] L.J. Hart-Smith, R.W. Ochsner, R.L. Radecky, Engineered Materials Handbook, vol. 3 Adhesives and Sealants. H. F. Brinson (technical chairman), ASM International, Materials Park, OH (1990), p. 840–844. [179] A.N. Rider, N. Brack, J. Gopalakrishna, Surface treatments and adhesives for bonded repairs to high temperature carbon-bismaleimide composite structure, J. Adhes. Sci. Technol. 26 (2012) 911–937, https://doi.org/10.1163/156856111X593630. [180] F.L. Palmieri, M.A. Belcher, C.J. Wohl, K.Y. Blohowiak, J.W. Connell, Laser ablation surface preparation for adhesive bonding of carbon fiber reinforced epoxy composites, Int. J. Adhes. Adhes. 68 (2016) 95–101, https://doi.org/10.1016/j. ijadhadh.2016.02.007. [181] M.A. Belcher, K.L. Krieg, P.J. Van Voast, K.Y. Blohowiak, 2013. Nonchemical surface treatments using atmospheric plasma systems for structural adhesive bonding, in: International SAMPE Technical Conference. Presented at the SAMPE 2013 Conference and Exhibition: Education and Green Sky-Materials Technology for a Better World, pp. 567–571. [182] L.J. Hart-Smith, G. Redmond, M.J. Davis, The curse of the nylon peel ply, Proceedings of 41st International SAMPE Symposium, vol. 41, (1996) 303–317. [183] L.M. Butkus, et al., Bonded Boron Patch Repair Evaluation: Final Report, USAFATR-2007-06, 2007. [184] N.J. Jacobs, Quick Reaction Evaluation of Materials and Processes, Delivery Order 0010: Bonded Boron Patch Repair Evaluation, AFRL-ML-WP-TR-2007-4145, 2006. [185] C.B. Guijt, The Effect of In-Service Conditions and Aging on the Fatigue Performance of Bonded Composite Repairs, USAFA-TR-2007-07, 2007. [186] E.S.Y. Yeo, A.N. Rider, F-111 Adhesive Bonded Repairs Assessment ProgramProgress Report 1: Analysis of FM300 Repairs, DSTO-TR-2849, June 2013. [187] M. Lee, E. Yeo, M. Blacklock, M. Janardhana, S. Feih, C.H. Wang, Predicting the strength of adhesively bonded joints of variable thickness using a cohesive element approach, Int. J. Adhes. Adhes. 58 (2015) 44–52. [188] M. Lee, E. Yeo, C.H. Wang, Effects of adherend thickness and taper on adhesive bond strength measured by portable pull-off tests, Int. J. Adhes. Adhes. 44 (2013) 259–268. [189] A.B. Harman, A.N. Rider, Verification of Adhesive Bond Strength for Composite Repairs Applied to Primary Aircraft Structure with more than 600 Flight Hours of Service, DST-Group-TR-3310, 2017. [190] A.A. Baker, R.J. Callinan, M.J. Davis, R. Jones, J.G. Williams, Repair of mirage III aircraft using the BFRP crack-patching technique, Theor. Appl. Fract. Mech. 2 (1) (1984) 1–15. ISSN 0167-8442, https://doi.org/10.1016/0167-8442(84)90035-1. [191] M.J. Davis, J.D. Roberts, “Procedure for Application of Boron-Fibre Reinforced Plastic Patch to Mirage Lower Wing Skin Decant Region”, Materials Technical Memorandum 373, DSTO, commonwealth of Australia, 1981. [192] DSTO File: BM240/Part1 from the Period 1977 to 1982 Entitled, “Adhesive Bonded Repairs for Mirage”, Memorandum Addressed to Chief Superintendent, ARL from HQSCENG, 29th June 1982, page 62. [193] A. A. Baker, M. M. Hutchison, Fibre composite reinforcement of cracked aircraft structures”, 33rd Annual Technical Conference, 1978 Reinforced Plastics/Composites Institute, The Society of the Plastics Industry, Section 17-E, pp. 1–8, 1978.