Engineering Failure Analysis 84 (2018) 34–45

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal

Effect of surface treatment on adhesively bonded aluminium-

T
aluminium joints regarding aeronautical structures
Sergio Correiaa, Vitor Anesb,c, Luis Reisa,b,⁎
a
Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1, 1049-001 Lisbon, Portugal
b
IDMEC, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1, 1049-001 Lisbon, Portugal
c
GI-MOSM, Instituto Superior de Engenharia de Lisboa, Instituto Politécnico de Lisboa, Rua Conselheiro Emídio Navarro, 1, 1959-007 Lisbon,
Portugal

AR TI CLE I NF O AB S T R A CT

Keywords: The structural integrity of several structures could be determined by their joints strength. Over
Adhesive the years, adhesively bonded joints have been often chosen to achieve a compromise between
Aeronautical structures mass reduction and higher mechanical strength. Among others, the reduction in stress con-
Adhesively bonded joints centrations, the ability of producing smooth surfaces with no discontinuities and the reduced
Single lap joint
weight penalties are some of the factors that make this type of joints so attractive. Normally, to
Surface treatment
increase the bond strength, the materials to be bonded must be subjected to a kind of surface
treatment. For metals, and more specifically, for aluminium alloys, phosphoric acid anodizing
and chromic acid anodizing have been the most used treatments worldwide. However, recent
investigations show that these kinds of anodizing are detrimental to health due to the release of
carcinogenic substances. With this in mind, it is of the utmost importance to find alternative
surface treatments that can ensure an effective bond. In this paper, a vast experimental study was
performed based in the single lap joint ASTM D 1002 standard method, with the objective of
determining the best alternative surface treatment (Sulfuric Acid Anodizing and Boric-Sulfuric
Acid Anodizing), for aluminium-to-aluminium joints, using two types of adhesives, namely the
AF 163 and the EA 9658 AERO. Results show that the optimum surface treatment is different for
each type of adhesive and this fact has a huge influence on mechanical behavior of this type of
aeronautical adhesive joints.

1. Introduction

With the advances in manufacturing techniques, several improvements have been verified in structural adhesives. As a direct
result of these improvements, the use of bonded joints in several aeronautical structures became very common nowadays.
Between other advantages, adhesively bonded joints allow a better load distribution, increases the service life, reduces machining
cost and reduces the production complexity [1,2].
Despite the many advantages, bonded joints also suffer from a significant number of limitations. Among others, one can identify
the bad adhesion to some substrates, weak resistance to cleavage stresses, degradation due to exposure to hostile environments and
the difficult compromise between the quality of bonded joints and their costs [3].
However, in contrast to more conventional joining techniques, the major concern related with adhesively bonded joints is the lack
of accepted guidelines to predict the joint strength. The aerospace industry, which has been developing the technology, is still

⁎
Corresponding author at: IDMEC, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1, 1049-001 Lisbon, Portugal.
E-mail address: luis.g.reis@ist.utl.pt (L. Reis).

http://dx.doi.org/10.1016/j.engfailanal.2017.10.010
Received 7 June 2017; Received in revised form 6 October 2017; Accepted 22 October 2017
Available online 06 November 2017
1350-6307/ © 2017 Elsevier Ltd. All rights reserved.
S. Correia et al. Engineering Failure Analysis 84 (2018) 34–45

Fig. 1. Single lap joint, taken from A. Çalik [8].

designing joints mainly based in previous experience and rules of thumb.

Several studies show that the bond strength of an adhesively bonded joint depends on many factors, such as the bonded area, the
environment in which the structure will be operating, the service temperature, the type of surface treatment applied to the adherents
and, of course, the type of adhesive used [4–6].
Actually, some of the adhesives usually used to repair engine nacelle structures are being discontinued, so in some cases, it is of
the upmost importance to find new adhesives that present a mechanical behavior equal or better than the old ones specified in the
original design.
Moreover, there is a need of using surface treatments not detrimental to health, meaning that surface treatments such as phos-
phoric acid anodizing (PAA) and chromic acid anodizing (CAA) cannot be used due to regulatory directives, although they were used
in the original design [7].
In this context, the objective of this paper is to find new alternatives to the Hysol EA 9689 adhesive, an adhesive produced by
Henkel Loctite commonly used in aircraft repairs, which was recently discontinued. To achieve that, the mechanical behavior of
aluminium-aluminium adhesively bonded joints was studied, considering two different adhesives, namely the AF163 and the EA9658
AERO. To improve adherence between adhesives and subtract, two different surface treatments were also considered, the Sulfuric
Acid Anodizing (SAA) and the Boric-Sulfuric Acid Anodizing (BSAA). These surface treatments are not detrimental to health and are
good alternatives to the PAA and CAA surface treatments.
The experiments were made using the ASTM D 1002 standard, which is based on the Single Lap Joint (SLJ). The SLJ present in
Fig. 1, is one of the bonded joints most used in an aircraft, especially due to the ease of produce, inspect and repair [8].

2. Theoretical background

When the SLJ is analyzed theoretically or computationally, it is normally assumed a perfect adherence between the adhesive and
adherents, with the failure occurring in the adhesive layer.
However, in real joints, this is not always the case and other types of failure can occur, especially failures related to the adherence
between adhesive and adherent, please see Fig. 2. Due to this reason, the adherents' surface is usually treated with electrolytic
passivation processes such as PAA or CAA. Furthermore, there are products that must be applied after the surface treatment and prior
to the adhesive application, the so-called adhesion promoters and primers that, as well as the surface treatments, helps to improve the
adhesion and, consequently, the strength of the joint.

2.1. Types of failures

When destructive tests are being performed in bonded joints, good indicators of the joint strength behavior are given by analyzing
the adhesive layer after failure. In an adhesive joint, there is the need to assure a perfect adhesion between the adherents and the

Fig. 2. Failure types in a bonded joint: cohesive failure on the top left; adhesive failure on the top right; adhesive-cohesive failure on the bottom left; adherent failure
on the bottom right, taken from J. Tomblin et al. [10].

35
S. Correia et al. Engineering Failure Analysis 84 (2018) 34–45

adhesive. In this case, it is expected that the failure occurs in the adhesive layer and never in the interface between adherent and
adhesive. Fig. 2 shows four types of failure modes that can be identified in an adhesive joint, namely:

1. the cohesive failure, represented in the upper left corner of Fig. 2, is characterized by failure of the adhesive itself;
2. the adhesive failure, represented in the upper right corner, is characterized by a failure of the joint at the adhesive/adherent
interface and it is typically caused by inadequate surface preparation, chemically and/or mechanically. Specimens that fail ad-
hesively show a lack of adhesion resulting from unsuitable surface qualities of the adherent [9];
3. the adhesive-cohesive failure, represented in the lower left corner, the failure presents the previous two types simultaneously;
4. and the adherent failure, represented in the lower right corner, it is characterized by the failure of the adherent instead of the
adhesive. In metals, this occurs when the adherent yields.

As explained, the adherents surface treatment and the use of adhesion promoters or primers play a major role in the type of failure
that is obtained, as it can be seen in Subsection 2.2 and 2.3.

2.2. Surface treatments

The purpose of applying a surface treatment in a material is to modify the material surface to provide the best possible bonding
conditions.
One of the main problems related with the surface treatment is the impossibility of predicting which treatment is more suitable for
a certain kind of adhesive or application without the realization of experimental tests, especially when no qualitative or quantitative
data is available [11].
There are several types of surface treatments for metals which can be classified as micro or macro-mechanical, chemical or a
combination of these [12]. In this work, only chemical surface treatments will be introduced and tested, namely the Sulfuric Acid
Anodizing and the Boric-Sulfuric Acid Anodizing. Every surface treatment has the main goal of enhancing the bond strength, either in
dry or humid environments [13]. This is extremely important in the aerospace industry since an aircraft, and consequently its bonded
joints, is subjected to great temperature and humidity variations during all phases of its flight [6].
Several works have been carried out along the years emphasizing the importance of surface treatments in the bond strength.
Vaillancourt and Abele [13] performed a very extensive work with SLJ, mainly based on experimental tests and observations,
obtaining the failure loads, for various types of adhesives and several types of surface treatments, including different types of etches
and anodizing. The main conclusion from the work developed it was that treated specimens presented a much better mechanical
behavior, when compared with the ones with no kind of treatment. Also, the anodized specimens were more efficient than the etched
ones, assuring better bonding conditions. Teixeira de Freitas and Sinke [14,15] developed a new type of peel tests to study the
adhesion properties of bonded composite-to-aluminum joints. They concluded that in adhesion tests the failure mode is more im-
portant than the failure load, in addition they concluded that the adherents surface treatment play an important role in the interface
adhesion.
In this study, four types of chemical surface treatment will be presented, being them the Phosphoric Acid Anodizing (PAA), the
Chromic Acid Anodizing (CAA), the Sulfuric Acid Anodizing (SAA) and the Boric-Sulfuric Acid Anodizing (BSAA). The first two types
must be presented since they are the treatments that the manufacturers of the adhesives tested, in this work, advise to use. The last
two are the ones that will be under investigation in this work. Vaillancourt and Abele [13] made a review of some of this treatments,
being their work quite useful to understand the effects that they cause in the metal surface.

2.2.1. Phosphoric acid anodizing

Anodizing is one of the most worldwide used surface treatments of metals, especially in the aerospace industry [16]. During the
process, stable coatings are formed on the adherent surface in a wide variety of electrolytes, producing an oxide layer in the metal
surface. Microscopically, this causes the appearance of “whiskers” in the surface, increasing the bond surface of the metal, despite not
being seen at naked eye, please see Fig. 3.
In Phosphoric Acid Anodizing (PAA) surface treatment, the specimens must be submerged in a bath of 9 to 12%, in weight, of
phosphoric acid, at temperatures ranging from 19 to 25 °C. The anodizing voltage ranges from 9 to 16 V under a direct current. The
immersion time is around 20 to 25 min [13].
This type of treatment is advised, normally, by the adhesive manufacturer Henkel Loctite, but has been progressively discontinued
because this kind of anodizing is detrimental to health due to the release of carcinogenic substances.

2.2.2. Chromic acid anodizing

The Chromic Acid Anodizing (CAA) is very like to the phosphoric acid anodizing method. The chromic acid creates a very strong
layer of oxide aluminium in the metal surface, developing an environment quite propitious for strong bonding.
As in the PAA case, the CAA is falling into disuse, due to the presence of hexavalent chromium on its constitution. Hexavalent
chromium is carcinogen and has been phased out from metal finishing processes [17].

2.2.3. Sulfuric acid anodizing

Sulfuric Acid Anodizing (SAA) is basically used for structural adhesive bonding in high-humidity environments. It is a surface
preparation process very common in the automotive and consumer cookware industries. It involves submerging the sample in a

36
S. Correia et al. Engineering Failure Analysis 84 (2018) 34–45

Fig. 3. Draw of the microscopic oxide structures formed after Phosphoric Acid anodizing, taken from A. Vaillancourt and T. Abele [13].

sulfuric acid bath and then an electric current is applied through it, this creates a very thin oxide layer on the sample surface which
improves the bond strength [18,19].

2.2.4. Boric-sulfuric acid anodizing

Finally, the last anodizing process mentioned is the Boric-Sulfuric Acid Anodizing (BSAA). This treatment has been introduced as
an alternative to the CAA treatment. The oxide film from the boric-sulfuric electrolyte has a paint like adhesion that is equal, or
superior, to the one formed on chromic acid [20]. The process is voltage controlled and is ramped to 15 V.

2.3. Adhesion promoters and primers

Some adhesives may provide only marginal adhesion to some substrates. This effect could be related with low surface energy of
the adhesive or to a boundary layer that is cohesively weak. The substrate may also be porous, allowing moisture and other en-
vironmental chemicals to easily pass through to the substrate to the adhesive inter-phase, causing disbonding between adhesive and
adherent [21].
Generally, attempts to overcome these problems include applying a surface treatment to the adherents or even change the
adhesive composition. However, when none of these approaches work, adhesion and durability can be increased by primers or
adhesion promoters [20].
These types of materials work in a similar way to improve adhesion, adding a new, usually organic, layer at the interface between

Fig. 4. Primer in a bonded join [17].

37
S. Correia et al. Engineering Failure Analysis 84 (2018) 34–45

Fig. 5. SLJ specimen dimensions.

adhesive and substrate, please see Fig. 4. Both are strongly adsorbed onto the surface of the substrate and this adsorption can be so
strong that, instead of being merely physically adsorbed, it has the nature of a chemical bond.
The main difference between primers and adhesion promoters is that primers are liquids that are applied directly to the adherents
as a relatively heavy surface coating prior to the application of the adhesive. On the other hand, adhesion promoters are liquids that
form a very thin layer between the substrate and the adhesive. Plus, adhesion promoters can be applied by either incorporating them
into the adhesive formulation (if the adhesive is liquid) or by applying directly on the substrate, identically as a primer [21]. For the
adhesives studied in this work, a special primer for each one is advised.

3. Materials and methods

3.1. Methods

As stated, the evaluation of the mechanical behavior of the single lap joint (SLJ) will be the focus of this work. The usual test for
this type of joints is the ASTM D 1002 [22], consisting essentially in the steps described below:

• The specimen must be placed in the grips/claws of the testing machine;

• The load shall be applied to the specimen. In this case the crosshead speed applied was 1.3 mm/min;
• The load was applied until failure.
The overlap used and the geometry of the specimen can be seen in Fig. 5. The thickness of the metal, t, is 1.6 mm, being the
adhesive thickness dependent on the type of adhesive used.
Carrying on this test, the failure Load of the joint, P is obtained. To obtain the apparent failure shear stress from P, Eq. (1) can be
used.
P
τ =
bL (1)

where P represents the applied load, b = 25.0 mm the SLJ width and L = 12.5 mm the bonded overlap.

3.2. Materials

Since the main interest of the work is to study the adhesive behavior, one must guarantee that the adherents are in the elastic
region. This was verified using a simple analytical method proposed by Adams and Wake [1]. In this method, the maximum overlap
length must be lower than 13.88 mm. Since the overlap used was 12.5 mm, the one suggested by the ASTM D 1002 standard, thus it is
verified that the adherents will be in the elastic regime during testing.
The only adherent material tested was the aluminium 2024-T3 alloy. Regarding the adhesives, two types were tested in this work:
the Henkel Loctite EA 9658 Aero [23] and the 3 M Scotch-Weld AF 163-2 [24]. Both are considered structural adhesives, manu-
factured in film form and widely used in the aerospace industry. These two adhesives are tested to replace an old one, the Hysol EA
9689 [25], also produced by Henkel Loctite.
Each adhesive is advised to be used with a primer to improve their adherence to the substrates. The Scotch-Weld EC-3924B primer
must be used with the adhesive AF 163-2, while Henkel Loctite EA 9258.1 AERO primer is advised to be used along with the EA 9658
AERO adhesive. Furthermore, each manufacturer advises the use of a single type of anodizing to achieve the optimum bonding
strength. Properties of each adhesive are presented in Table 1.

38
S. Correia et al. Engineering Failure Analysis 84 (2018) 34–45

Table 1
Properties of each adhesive and surface treatment suggestion given by the manufacturer, τf - apparent shear stress at failure, t - adhesive
thickness [22–24].

Properties EA 9689 EA 9658 AF 163-2

phased out

τf [MPa] 23.4 36.5 39.3

t [mm] – 0.60 0.24
Anodizing PAA PAA CAA
Primer EA 9205R EA 9258.1 EC-3924B

3.3. Specimens manufacturing

The manufacturing process had two main phases: surface treatment followed by bonding. To apply the surface treatment, firstly it
is necessary to clean the adherents' surface. This is essential to remove any type of grease or impurity from the aluminium adherent.
Immediately after, the aluminium SJL plate is submerged in a bath that promotes the creation of micro-pores in the surface.
Next, the SLJ plates are submerged in another bath solution during 45 min for anodization. Two bath solutions were separately
used in different coupons, namely, the Sulfuric Acid Anodizing and the Boric-Sulfuric Acid Anodizing.
The bath temperature ranges from 24 to 29 °C, and a direct current is applied on the submerged SLJ plates at 15 V. The electric
current passing through the anodize solution converts the aluminium surface in a form of oxide, increasing the contact area. Finally, a
surface finish is applied.
This process is typically used to increase corrosion resistance and promote, even more, adherence of paints and adhesives.
For both film adhesives (EA 9658 and AF 163-2) the process of bonding is quite similar. All the bonding process was manually
performed in a clean room, under controlled environment and in a shortest possible time.
It starts with the primer application and its cure. After the primer curing, the bonding region is degreased and the film adhesive is
applied in the desired area. To prevent stress concentrations, and to obtain more reliable results, an adhesive fillet is added manually
to each side of the bond layer.
Finally, the adhesive is cured, being the specimens closed together inside of a peel-ply bag to create a vacuum environment, which
promotes the contact between the adhesive and adherents.
The cure method was different for each adhesive. Due to the high cure temperature, the coupons bonded with the EA 9658 AERO
adhesive were cured with thermal blankets, while the coupons bonded with the AF 163-2 adhesive were cured in a drying oven.

3.4. Number and type of specimens

Two types of SLJ were tested, one representing a new joint (NJ), while the other was representing a repaired joint (RJ). In the NJ
type both adherents surface is anodized and painted with primer. In the RJ case, only one substrate was prepared with a type of
anodizing, while the other subtract was only cleansed with solvent wipe to remove any oil or grease contamination.
In many bonded joints repairs it is not possible to perform anodizing due to constraints in the anodizing tank. Also, the original
anodizing treatment is lost due to maintenance procedures, thus is makes sense analyze the RJ case. Both types of joints were
combined with the two adhesives, as well as with the two types of anodizing.
The ASTM D 1002 advices to perform at least 5 tests for each type of specimen studied. To increase the reliability of the results, 7
tests were conducted for each one. Altogether, 56 tests were performed. In Table 2, all the combinations used are presented.

4. Results and discussion

The results presented will include the failure load, the apparent failure shear stress and the load-displacement curves. Also, all the
results will be presented from a statistical point of view.

Table 2
Tests performed for each specimen.

Adhesive Treatment Joint type

NJ RJ

AF 163-2 BSAA 7 7
AF 163-2 SAA 7 7
EA 9658 AERO BSAA 7 7
EA 9658 AERO SAA 7 7

39
S. Correia et al. Engineering Failure Analysis 84 (2018) 34–45

Fig. 6. Statistical representation of the failure loads obtained in the performed tests: The upper black line indicates the maximum observation, the upper blue indicates
the upper quartile, the red line indicates the median, the lower blue line indicates the lower quartile, and the lower black line indicates the minimum observation. (For
interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

4.1. EA 9658 AERO adhesive

4.1.1. Failure load

The results obtained for the apparent failure shear stress of EA 9658 AERO adhesive are presented graphically in Fig. 6. Although
that the failure load, Pf, is the value obtained at rupture in the ASTM D 1002 test, one choose to present the apparent shear strength,
τf , which is obtained by dividing Pf by the bonding area, please see Eq. (1).
Using the apparent shear strength for correlation allows the comparison between the experimental results and the values given by
the adhesives manufacturers. All the tests were performed successfully, with valid results.
Analyzing Fig. 6, one can see that the boxes dimensions are quite reduced, indicating that the results, for each type of specimen
studied, do not present a wide dispersion. Also, the small values of the standard deviations support this statement. The BSAA NJ
specimens are the ones with the best apparent failure shear stress, i.e. τ = 23 ± 1.75 MPa . Overall, the BSAA surface treatment
provides a better mechanical behavior for joints bonded with the EA 9658 AERO, either for NJ and RJ. The values of the standard
deviations are quite small, indicating that the average values are quite significant, since there is not a great variation between tests for
each type of joint.
Based on these results, one can conclude that the surface treatments improve the bond strength – the simulated repaired joints
presents a much lower bond strength, due to the lack of adherence in the face that is not treated. This effect will be analyzed in detail
further on.
Despite the improved results achieved in the treated coupons, they are a bit far from the ones indicated by the EA 9658 AERO
technical datasheet [23] and shown in Table 1. Recalling that the manufacturer performed this same standard test, being the alu-
minium plates treated with the PAA surface treatment. With this treatment, the supplier indicates that the joint can handle a value of
τ = 36.5 MPa which means that the best results obtained (for the BSAA treatment) are still 36% worse when compared with the
reference data retrieved from technical datasheet.

4.1.2. Load-displacement curves

For each test, a load-displacement curve was obtained. In this way, it is interesting to present a typical curve for each type of joint
studied, please see Fig. 7. The curve chosen was the one with the highest failure load, for each type.
Fig. 7 shows that, initially, the specimens present a very similar behavior, with the curves having the same slope for displacements
under 0.1 mm, approximately. With the increasing displacement, the specimens start to behave differently.
The BSAA NJ yields for an approximate displacement of 0.1 mm. Due to this effect, the slope of the curve decreases a little and
remains constant until it fails. The failure is completely brittle, and occurs for a displacement around 1.3 mm, with a corresponding
load of 7827 N.
The SAA NJ presents a linear behavior until displacement reaches 0.3 mm. Here, the joint integrity changes, with the curve going
to an almost horizontal step. After this phenomenon, the joint keeps resisting the increasing displacement, having the curve a similar
slope as the BSAA NJ curve. The failure is completely brittle once again and occurs for a load of 6739 N.
The BSAA RJ and the SAA RJ presents a similar behavior, failing at very low loads. Both failures are still brittle, being the failure
load for the BSAA RJ a bit higher when compared with the SAA RJ specimen.

4.1.3. Type of failure

In Section 2 the types of failure in an adhesively bonded joint were introduced. Analyzing the failure surfaces of the joints, two

40
S. Correia et al. Engineering Failure Analysis 84 (2018) 34–45

Fig. 7. Typical load-displacement curve for each type of joint studied.

types of failure were observed: adhesive failure (A) and adhesive-cohesive failure (AC). Furthermore, the type of failure is strictly
related with the joint strength. Specimens with an AC type of failure have higher failure loads, while the specimens presenting an A
failure, have much lower failure loads.
Adhesive-cohesive failure is observed in all the NJ specimens, either treated with BSAA or SAA. In Fig. 8, an example of a
specimen bonded with the EA 9658 AERO is presented. The AC failure is easily identified, since, in both surfaces, clean areas and
areas with adhesive are observable. An example of one adhesive and one cohesive failure areas is presented with letters A and C
respectively.
Also, adhesive failure occurs in all the RJ specimens, for both treatments. The joint always fails due to the lack of adherence in the
non-treated plate (NTP). In Fig. 9, this type of failure is presented, for a BSAA RJ specimen, where the problem of adherence is clearly
visible in the right plate (TP). In fact, neither the primer adheres to the surface, as it can be seen in the left plate (NTP) in the region
pointed by the arrow A.

4.2. AF 163-2 adhesive

4.2.1. Failure load

The results obtained for the apparent failure shear stress are presented graphically in Fig. 10. Again, all the tests were performed
successfully, with valid results.
Analyzing the results for the AF 163-2 adhesive, one can conclude the following: Again, as for the first adhesive, the results are
quite reliable. This statement is based in the small dimensions of the boxes of Fig. 10, allied to the low values of the standard
deviations, which indicate that the results do not have a wide dispersion.
The SAA NJ specimens are the ones with the best apparent failure shear stress: τ = 25 ± 1.3 MPa . Overall, the SAA surface
treatment provides a better mechanical behavior for joints bonded with the AF 163-2, either for NJ and RJ.
As seen in the EA 9658 AERO adhesive, the standard deviations found in the AF 163-2 results are quite small, indicating a small
variation between tests for each type of joint.
Again, the surface treatment is essential to improve the bond strength – the simulated repaired joints presents a much lower bond
strength, due to the lack of adherence in the face that is not treated.
Despite the improved results achieved in the treated coupons, as seen in the EA 9658 AERO adhesive, they are still a bit far from

Fig. 8. Adhesive-cohesive failure surface on a BSAA NJ.

41
S. Correia et al. Engineering Failure Analysis 84 (2018) 34–45

Fig. 9. Adhesive failure surface on a BSAA RJ.

Fig. 10. Statistical representation of the failure loads obtained in the tests performed: The upper black line indicates the maximum observation, the upper blue
indicates the upper quartile, the red line indicates the median, the lower blue line indicates the lower quartile, and the lower black line indicates the minimum
observation. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

the ones indicated by the AF 163-2 technical datasheet [24] and shown in Table 1. The manufacturer performed the same standard
test, being the aluminium plates treated with CAA. With this treatment, the supplier indicates that the joint can handle a value of
τ = 39.3 MPa which means that the best results obtained (for the BSAA treatment) are again 36% worse when compared with the
reference data retrieved from technical datasheet.

4.2.2. Load-displacement curves

Same as for the previous adhesive, for each test, a load-displacement curve was obtained. A typical curve for each type of joint is
presented, please see Fig. 11. The curve chosen was, once again, the one with the highest failure load, for each type.

Fig. 11. Typical load-displacement curve for each type of joint studied.

42
S. Correia et al. Engineering Failure Analysis 84 (2018) 34–45

Fig. 12. Adhesive-cohesive failure surface on a BSAA NJ.

From Fig. 11, it is observable that, differently from the previous adhesive, the specimens present a different behavior, even for low
loadings. The BSAA NJ behavior changes pattern for an approximate load of 1000 N. Due to this change, the slope of the curve
decreases a little and remains constant until it fails. The failure is completely brittle. The SAA NJ presents a linear behavior until
values around 6000 N. Here, the adhesive changes behavior with some inconstancy, implying that a part of the joint had already
failed. After this phenomenon, the joint can still resist the load increasing, having the curve the same slope as the BSAA NJ. The
failure is completely brittle once again.
The BSAA RJ and the SAA RJ presents a similar behavior, failing at very low loads. The main difference is related with the failure
behavior, which is more ductile for the repaired joint treated with BSAA.

4.2.3. Type of failure

Again, two types of failure were observed: adhesive failure and adhesive-cohesive failure. As it was already discussed, specimens
with AC failure present higher bond strength. Similarly, to the EA 9658 AERO adhesive, adhesive-cohesive failure is observed in all
the NJ specimens, either treated with BSAA or SAA.
Fig. 12 shows an example of a specimen bonded with the AF 163-2. The adhesive-cohesive failure is easily identified, since, in
both surfaces, clean areas (arrow A) and areas with adhesive (arrow C) are observable.
Also, as expected, the adhesive failure occurs in all the RJ specimens, for both treatments. As discussed earlier, the lack of
adherence occurs in the non-treated plate, which is the main reason for the low bond strength.
The exact same problem was observed in the EA 9658 AERO adhesive. Fig. 13 shows this type of failure for a BSAA RJ specimen,
where the problem of adherence is clearly visible in the non-treated plate on the left. Also, examples of adhesive and cohesive areas
are presented for both types of failures.

4.3. Comparison between adhesives

A comparison between both adhesives is presented in Fig. 14, where the average values of τ are presented, for each adhesive,
surface treatment, and type of specimen.
The dashed line represents the reference value that one wants to achieve to selected the right combination of adhesive/surface
treatment in order to replace the Hysol EA 9689. This reference line is given by the apparent failure shear stress of the old adhesive,
the Hysol EA 9689, with the value of τ = 23.4 MPa .
From Fig. 14, it is clearly observable that the adhesives behave differently for each type of joint.
The specimens SAA NJ, bonded with the adhesive AF 163-2, are the ones which presents a better mechanical behavior. The
apparent failure shear stress for this specimen is τ = 25 MPa representing an improvement of 8% relatively to the old adhesive.
Furthermore, the BSAA NJ specimens also present an improvement of 3% relatively to the old adhesive. This means that, for spe-
cimens NJ treated with AF 163-2, the main goal was achieved, i.e. it was found a replacement for the Hysol EA 9689 compatible with

Fig. 13. Adhesive failure surface on a BSAA RJ.

43
S. Correia et al. Engineering Failure Analysis 84 (2018) 34–45

Fig. 14. Comparison between adhesives.

BSAA, an environmentally accepted anodizing surface treatment.

The EA 9658 AERO adhesive, applied to aluminium treated with BSAA, also presents good results – only 2% worse than the old
adhesive.
When analyzing the BSAA RJ and SAA RJ, it is clearly observable that the EA 9658 AERO presents a much better mechanical
behavior when compared with the AF 163-2. The main issue for these results is that there is not any type of reference line for the old
adhesive, since this kind of test was not performed by the manufacturer. Consequently, it is not possible to infer if there is any kind of
improvement relatively to the Hysol EA 9689. Furthermore, due to the low values obtained, the security of the joint cannot be
guarantee, and further tests shall be performed.
One very interesting effect is the fact that the optimum surface treatment is different for each adhesive. For the AF 163-2, the best
surface treatment is the SAA, either for NJ or RJ, and for the EA 9658 AERO the BSAA is the one which presents a better mechanical
behavior.

5. Conclusions

This work presents an experimental study regarding the selection of surface treatments and adhesives to be applied on aluminium-
aluminium adhesively bonded joints of aeronautical structures. The main objective was to select a new combination of adhesive and
surface treatment that could assure safety in repaired and new bonded joints. Recently, some adhesives usually used in aerospace
structures have been discontinued, as well as some surface treatments have also been banned due to health issues. These two reasons
are the motivation that lead to the pursuit of an improved combination of surface treatment and adhesive. To achieve that, and based
on single lap joint ASTM D 1002 standard method, several tests were performed. Results show that the combination of the AF 163-2
adhesive with the SAA surface treatment managed to achieve a result 8% better than the old combination, i.e. better than the
combination of the discontinued adhesive and the banned surface treatment suggested in the original design of some commercial
aircraft. However, the proposed combination presents a very poor mechanical behavior when applied to RJ cases (cases in which the
application of surface treatments is not possible), whereby the appliance of this adhesive in this type of joints (repaired joints) can
jeopardize the safety of the structure. Furthermore, despite being one of the most used experimental test in the world regarding
adhesives characterization, the results of the ASTM D 1002 standard cannot be used only by itself to fully characterize the mechanical
behavior of the adhesive and surface treatment selected in this work. One must note that, despite being used in aluminium-aluminium
joints [26], the chosen adhesive will also be applied in honeycomb sandwich panels. In this sense, some more standard tests shall be
performed to ensure the quality of the bonding, both in aluminium-aluminium joints and honeycomb panels. The proposed tests are
the ASTM C 273 - Shear Properties of Sandwich Core Materials, ASTM C 297 - Flatwise Tensile Strength of Sandwich Constructions,
ASTM C 364 - Edgewise Compressive Strength of Sandwich Constructions and the ASTM C 393 - Flexural Properties of Sandwich
Constructions, and peel tests such as floating roller or t-peel. Moreover, some more SLJ should be tested, with different overlap
lengths and adhesive thicknesses, in order to infer about the effects of these two properties in the joint strength.

Acknowledgements

This work was supported by FCT, through IDMEC, under LAETA project UID/EMS/50022/2013. The second author gratefully
acknowledge financial support from FCT - Fundação para Ciência e Tecnologia (Portuguese Foundation for Science and Technology),
for the Ph.D. Grant PD/BD/52344/2013.

44
S. Correia et al. Engineering Failure Analysis 84 (2018) 34–45

References

[1] R.D. Adams, J. Comyn, W.C. Wake, Structural Adhesive Joints in Engineering, Springer Science & Business Media, 1997.
[2] M. Samaei, M. Zehsaz, T.N. Chakherlou, Experimental and numerical study of fatigue crack growth of aluminum alloy 2024-T3 single lap simple bolted and
hybrid (adhesive/bolted) joints, Eng. Fail. Anal. 59 (2016) 253–268.
[3] J. Cui, Finite Element Modeling of Adhesive Failure with Adherend Yielding, National Library of Canada, 2001.
[4] M.D. Banea, L.F. da Silva, Adhesively bonded joints in composite materials: an overview, Proceedings of the Institution of Mechanical Engineers, Part L: Journal
of Materials: Design and Applications, 223 2009, pp. 1–18.
[5] K.L. Mittal, The role of the interface in adhesion phenomena, Polym. Eng. Sci. 17 (1977) 467–473.
[6] V. Anes, R. Pedro, E. Henriques, M. Freitas, L. Reis, Bonded joints of dissimilar adherends at very low temperatures-an adhesive selection approach, Theor. Appl.
Fract. Mech. 85 (2016) 99–112.
[7] G.W. Critchlow, K.A. Yendall, D. Bahrani, A. Quinn, F. Andrews, Strategies for the replacement of chromic acid anodising for the structural bonding of alu-
minium alloys, Int. J. Adhes. Adhes. 26 (2006) 419–453.
[8] A. Calik, Effect of adherend shape on stress concentration reduction of adhesively single lap joint, Eng. Rev. 36 (2016) 29–34.
[9] R.D. Campilho, M.D. Banea, J. Neto, L.F. da Silva, Modelling adhesive joints with cohesive zone models: effect of the cohesive law shape of the adhesive layer,
Int. J. Adhes. Adhes. 44 (2013) 48–56.
[10] J. Tomblin, W. Seneviratne, P. Escobar, Y. Yoon-Khian, Shear Stress-Strain Data for Structural Adhesives, Wichita State University Department of Aerospace
Engineering, 2002.
[11] R.L. Patrick, Treatise on Adhesion and Adhesives, CRC Press, 1981.
[12] D.S. Cobb, M.A. Vargas, T.A. Fridrich, M.R. Bouschlicher, Metal surface treatment: characterization and effect on composite-to-metal bond strength, Oper. Dent.
25 (2000) 427–433.
[13] A. Vaillancourt, T. Abele, Adhesive Technology: Surface Preparation Techniques on Aluminum, Worcester Polytechnical Institute, Worcester, 2010.
[14] S.T. de Freitas, J. Sinke, Adhesion properties of bonded composite-to-aluminium joints using peel tests, J. Adhes. 90 (2014) 511–525.
[15] S.T. de Freitas, J. Sinke, Test method to assess interface adhesion in composite bonding, Appl. Adhes. Sci. 3 (2015) 9.
[16] L.E. Fratila-Apachitei, I. Apachitei, J. Duszczyk, Thermal effects associated with hard anodizing of cast aluminum alloys, J. Appl. Electrochem. 36 (2006)
481–486.
[17] P. Goyal, D.T. Masram, Nano-biomaterial for decontamination of carcinogenic metal from waste water, Nanotechnol. Res. J. 8 (2015) 447.
[18] V.R. Capelossi, M. Poelman, I. Recloux, R.P.B. Hernandez, H.G. De Melo, M.G. Olivier, Corrosion protection of clad 2024 aluminum alloy anodized in tartaric-
sulfuric acid bath and protected with hybrid sol–gel coating, Electrochim. Acta 124 (2014) 69–79.
[19] T. Mertens, B. Rico-Oller, M. Kolb, J. Wehr, Influence of etching processes on the oxide layer formation during anodizing of Ti6Al4V, J. Surf. Interfaces Mater. 3
(2015) 75–82.
[20] L. Zhang, G.E. Thompson, M. Curioni, P. Skeldon, Anodizing of aluminum in sulfuric acid/boric acid mixed electrolyte, J. Electrochem. Soc. 160 (2013)
C179–C184.
[21] E.M. Petrie, Handbook of Adhesives and Sealants, McGraw-Hill, 2000.
[22] A. International, Standard Test Method for Apparent Shear Strength of Single-Lap-Joint Adhesively Bonded Metal Specimens by Tension Loading (Metal-to-
Metal), ASTM International, 2010.
[23] Henkel Loctite, Loctite EA 9658 AERO Epoxy Film Adhesive Technical Process Bulletin, (2013).
[24] 3M Scotch-Weld, Structural Adhesive Film AF 163-2 Technical Datasheet, (2009).
[25] Henkel Loctite, Hysol EA 9689 Epoxy Film Adhesive Technical Process Bulletin, (2001).
[26] S.T. de Freitas, J. Sinke, Failure analysis of adhesively-bonded skin-to-stiffener joints: metal–metal vs. composite–metal, Eng. Fail. Anal. 56 (2015) 2–13.

45