Impact of adverse environmental conditions on the fatigue behavior of carbon epoxy adhesive joints under mode II fracture

Abstract

This work analyses how environmental degradation affects the behaviour of adhesive joints in carbon-epoxy composites. These joints, made from a unidirectional carbon fibre-reinforced epoxy matrix and bonded with an epoxy-based adhesive, were tested under mode II fracture static and fatigue loading conditions. The experimental analysis was carried out using the End Notched Flexure (ENF) test method to evaluate both the fatigue initiation phase (G-N curves) and the fatigue crack growth phase (G-da/dN curves). Specimens were subjected to controlled exposure in a salt spray chamber and a climatic chamber (hygrothermal) with different exposure periods: no exposure, 1 week, 2 weeks, 4 weeks and 12 weeks. Fatigue initiation data were analysed using a probabilistic Weibull distribution model. Key findings reveal that environmental degradation significantly reduces the fatigue resistance of the adhesive joints. Under static loading, the load capacity of the adhesive joints decreased by approximately 20% after saline exposure and 25% after hygrothermal exposure. Additionally, fatigue crack growth rates were found to change depending on the type and duration of environmental exposure, underscoring the critical role of environmental conditions on joint behaviour.

Introduction

Carbon fibre-reinforced polymer (CFRP) composites have become increasingly important in various industries due to their exceptional specific strength, stiffness, and durability¹. These materials, typically composed of a polymeric matrix (often epoxy, but also potentially polyester, vinyl ester, or thermoplastic matrices such as polyetheretherketone (PEEK)) reinforced with carbon fibres, are known for maintaining superior performance even under extreme conditions. CFRP laminates are commonly manufactured by stacking unidirectional plies; however, other configurations such as woven fabrics, multidirectional layups, and hybrid arrangements are also frequently employed, depending on the specific mechanical requirements. The versatility of these layup configurations allows engineers to optimize performance for various structural applications in industries such as aerospace, automotive, and energy². However, the layered structure of these materials, formed by stacking unidirectional plies, makes them susceptible to the initiation and growth of cracks between layers. This phenomenon, known as delamination, is one of the main failure modes that limit their service life.

Adhesive bonding has gained traction as a joining method in many industrial applications due to its advantages over traditional techniques like welding or riveting^3,4. Adhesives not only reduce structural weight but also improve stress distribution and eliminate the stress concentration points often found in mechanical fasteners. Additionally, adhesive bonding allows for cleaner, more aesthetic finishes, enabling the design of complex and integrated structures. These benefits, along with the ability to join dissimilar materials, have made adhesive joints increasingly popular in modern structural designs.

However, the bond between surfaces, especially at the adhesive-material interface, remains a critical challenge, as this area is often the most prone to crack initiation and propagation. To address this, significant research has focused on improving key properties such as toughness. Efforts include modifying the composition of the adhesive^5,6 introducing intermediate layers⁷plying surface treatments⁸ and analysing different adhesive thicknesses to determine the optimal configuration^9,10.

The behaviour of the adhesive interface is crucial for ensuring the mechanical strength and durability of joints, particularly under environmental influences. Factors such as temperature variations^{11, 12–13} and dynamic loading, especially the fatigue phenomenon^{14, 15, 16–17}. are among the most studied. Optimizing stress distribution is equally important to avoid concentrations that could weaken the joint. As a result, single-lap joints^{18, 19–20} have become a standard configuration for studying adhesive joints, particularly in CFRP structures. This configuration enables detailed analysis of stress concentrations, crack propagation, and mechanical behaviour under varying load conditions, providing insights for enhancing the strength and durability of adhesive joints in structural applications.

While Mode I fracture has been extensively studied for its critical relevance^{21, 22, 23–24} it is not the only failure mode encountered in CFRP components. Mode I involves direct tensile stresses on the adhesive, which facilitate crack propagation and can lead to failure. Theoretical modelling and experimental characterization²⁵ have yielded reproducible tests and predictive models, establishing Mode I as a key focus in adhesive research.

However, real-world conditions often involve mixed-mode fractures^26,27 where combined loads create complex stress states. Among these, mode II fracture is particularly significant. It is associated with sliding or shear fractures, which commonly occur in applications where structures are subjected to shear stresses. Examples include components under bending or misalignment, such as in aircraft, automobiles, and vibrating structures. These parallel forces on bonded surfaces promote sliding or shear fractures. Consequently, research into mode II fracture^{28, 29, 30–31} has expanded significantly, offering a more comprehensive understanding of adhesive joint behaviour.

Combining studies of Modes I and II^{32, 33–34} provides a more realistic and thorough analysis of adhesive joint performance in CFRP structures, improving design and structural durability.

Another critical aspect of this research is analysing the fracture surface to understand how environmental conditions affect the integrity of adhesive joints^{35, 36, 37–38} Prolonged exposure to saline environments can degrade the adhesive’s mechanical properties, reducing its fracture resistance and cohesion with the composite material³⁹.

This study aims to evaluate the fatigue behaviour of adhesive joints during delamination under mode II fracture, focusing on different exposure durations in two environments: a saline environment and a high-humidity combined with high-temperature environment (hygrothermal). The chosen substrate was a unidirectional carbon fibre-reinforced epoxy composite, and the adhesive was epoxy-based. The key parameter analysed was the energy release rate under mode II fracture loading, examining how degradation processes in saline and hygrothermal environments affected adhesive joint performance in both fatigue initiation and fatigue crack growth phases.

Materials

Below are the materials used in this study, describing the fundamental characteristics of the chosen composite substrate, as well as the adhesive employed.

Composite material

The material chosen for this study, commercially known as MTC510-UD300-HS-33%RW, is a unidirectional carbon fibre-reinforced epoxy matrix composite. The laminate consists of 8 plies, with a fibre volume fraction of approximately 65.4%, determined in accordance with ASTM 2584 − 02⁴⁰ Table 1 shows the typical mechanical properties of this composite.

Table 1. Mechanical properties of the composite.

Mechanical Property	Value	Coefficient of Variation (CV %)
Longitudinal elastic modulusa	122.0 GPa	8.5
Transverse elastic modulusa	8.5 GPa	8.0
Longitudinal tensile strengtha	1156.0 MPa	12.5
Transverse tensile strengtha	28.0 MPa	11.8
Shear modulusb	5.2 GPa	9.8
Shear strengthb	37.0 MPa	2.0

(a) ASTM D 3039M⁴¹ (b) ASTM D 3518M⁴².

The composite was fabricated using vacuum molding at a suction pressure of approximately 0.9 bar, aiming to apply a pressure on the laminate as close as possible to atmospheric pressure and cured according to the schedule provided by the supplier. During the first hour of the curing cycle, the temperature was gradually increased from room temperature to 100 °C. It was maintained at 100 °C for three and a half hours before being raised to 120 °C at a rate of 40 °C per hour. The material remained at 120 °C for one hour before being allowed to cool to room temperature inside the oven.

Adhesive

A commercially available structural epoxy-based adhesive, Loctite^® EA 9461, was used to bond the composite parts. According to the manufacturer, this adhesive was cured 1 h at 80 °C. This procedure aligns with the standard recommendations provided by the supplier. The adhesive thickness in the bonded specimens was measured using an optical microscope at several locations, resulting in an average thickness of 0.28 mm. Table 2 summarizes the key properties of the adhesive.

Table 2. Properties of Loctite^® EA 9461 adhesive.

Property	Value
Density	1.35 g/cm³
Viscosity	72.0 Pa·s
Elastic modulus	2.76 GPa
Tensile strength	30.0 MPa
Strain at break	3.5%
Curing time (80 °C)	1 h

Experimental methodology

This section describes the most important aspects of the experimental program carried out to characterize the adhesive joint in the selected composite material, focusing on delamination under mode II fracture loading in both static and fatigue conditions. The study also evaluates the influence of exposure to saline and hygrothermal environments over varying periods.

Surface preparation

The composite material used as the substrate was prepared by manually sanding it with P220 aluminium oxide sandpaper. After sanding, the surfaces were cleaned using a two-step process: first, compressed air at 6 bar was applied with an air gun to remove loose particles, and then the substrates were gently wiped with isopropyl alcohol using lint-free cloths to eliminate any remaining dust. To assess how the surface changed before and after sanding, a surface roughness tester was employed. Because the unidirectional fibre orientation can affect the surface finish, the substrates were measured in both the 0° and 90° directions. In each condition (natural and sanded), three measurements were taken in different areas of each substrate to capture variability in surface roughness. The parameters evaluated include R_a (arithmetic mean roughness value), R_z (mean peak-to-valley height over a sampling length), and R_max (maximum peak-to-valley height). The resulting data are presented in Table 3.

Table 3. Surface roughness before and after sanding.

Direction	State	Ra (µm)	Rz (µm)	Rmax (µm)
0°	Natural	2.24 2.58 2.02	9.48 10.70 9.83	9.99 11.60 12.70
	Sanded	2.98 3.03 3.27	11.50 11.20 12.01	13.70 12.30 14.85
90°	Natural	1.74 1.84 2.02	8.59 8.19 10.10	9.75 9.8 11.60
	Sanded	3.21 3.17 3.57	11.40 13.30 10.70	13.30 18.50 12.00

Degradation processes

Saline environment

A Köhler salt spray chamber (model DCTC 1200 P) was used to replicate a saline environment. The chamber was maintained at an average temperature of 35 °C (± 2 °C), 89% relative humidity, an air pressure of 1.2 bar, and a saline solution prepared by dissolving analytical-grade sodium chloride (50 g/L) in distilled, demineralized water. The resulting solution had a relative density between 1.0255 and 1.04, a pH ranging from 6.5 to 7.2, and a flow rate of 1–2 mL/h. At the end of each curing process, the specimens were removed and any remaining saline solution was rinsed off. The selected exposure periods in the salt spray chamber were 1, 2, 4, and 12 weeks.

Hygrothermal environment

This environment was simulated using a Vötsch climate chamber (model VC2020), programmed to maintain constant operating conditions. Specifically, the selected settings were a temperature of 60 °C and 70% relative humidity, applied over the same exposure durations as in the saline environment.

The goal of these degradation processes was to determine the behaviour of the adhesive joint as a function of exposure time to both temperature/humidity conditions and the saline environment.

Characterization of delamination behavior

To understand how exposure to these selected environmental agents affects the material’s delamination process, the chosen comparison parameter was the mode II energy release rate, measured under static and dynamic (fatigue) loading conditions. ENF-type (End-Notched Flexure) specimens were used, with dimensions of 20 mm in width, 150 mm in length, and a thickness of 4.3 ± 0.1 mm. An initial delamination length of 60 mm was introduced from one end by means of a 12 μm PTFE (Teflon) film. The methodology outlined in ASTM D7905/D7905M⁴³ was followed for testing.

To determine the mode II energy release rate (G_IIC), the experimental compliance calibration (NPC) method was applied, as shown in Eq. (1):

1

Where m represents a parameter function of the flexibility of the specimen, P_MAX is the maximum load applied, a₀ is the initial crack length and B represents the width of the specimen. For all tests, a servo-hydraulic MTS Model 810 testing machine with a 5 kN load cell was used. Crack propagation was monitored using a high-resolution camera. Figure 1 shows a photograph illustrating the arrangement of the specimen with the testing equipment.

Fig. 1 [Images not available. See PDF.]

Specimen arrangement in the testing machine.

Figure 2 shows the representative load–displacement curves obtained from static ENF tests for different environmental exposure conditions. The force–displacement curves under Mode II clearly illustrate how environmental exposure affects the behaviour of adhesive joints.

Fig. 2 [Images not available. See PDF.]

Load–displacement curves under environmental exposure.

In the saline environment, after one week of exposure, a noticeable reduction in both the maximum load and the initial stiffness is observed compared to the unaged material. This decrease suggests an early loss of properties in the adhesive or at the interface. After twelve weeks, the degradation is even more pronounced: although the maximum load remains similar, the displacement at failure is reduced, indicating a steeper slope in the curve and thus a more brittle behaviour, with a lower deformation capacity before fracture.

Similarly, in the hygrothermal environment, the one-week curve shows a response that is similar to or slightly better than the unaged material, possibly due to post-curing thermal effects. However, after twelve weeks, a significant drop in both the load and the final displacement is observed, reflecting a substantial loss of ductility and a more abrupt failure. This suggests a more severe chemical degradation of the adhesive compared to the saline environment, consistent with the accelerated effects of combined heat and humidity.

In order to analyse how exposure time in a saline environment and in a hygrothermal environment can affect the selected adhesive joint, considering its behaviour under mode II fracture delamination and fatigue loading, both the fatigue delamination initiation phase and the subsequent crack growth phase were studied. Regarding initiation, it was considered to occur once the fatigue loading process produced a visible crack. During this phase, the ΔG–N fatigue curves were determined, reflecting the number of cycles necessary for the onset of delamination at a given energy release rate.

To define the energy levels applied to the test specimens, the average values of the critical fracture energy obtained from previous static characterizations were used as reference. Seven different percentages of these reference values were selected. All fatigue tests were conducted under displacement control in the testing equipment. For each selected level, the maximum energy release rate was kept constant, while the minimum rate was defined by using an asymmetry ratio R = 0.1.

Crack initiation was determined through direct observation of the specimen, and the number of initiation cycles was defined as the number of loading cycles from the start of the test until a crack was visually detected at the crack front.

The crack growth rate was determined for different fractions of G_IIc during the fatigue test under displacement control, by measuring the evolution of crack length. These measurements were carried out using a 70× microscope placed in front of one of the lateral faces of the specimen and mounted on a device with micrometric adjustments, which allowed visualization and measurement of the crack. Although performing the test under these conditions is complex and time-consuming, it is considered a sufficiently reliable methodology, and the data obtained are valid, even taking into account variations in crack growth across the specimen width. As in the crack initiation tests, asymmetry ratios R = 0.1 were used.

Once the onset of delamination was identified, the fatigue process was paused every 500 cycles, and the specimen was brought back to the average displacement position. At that stage, the crack front was pinpointed using a microscope, and the amount of crack growth was measured. Based on these observations, the propagation rate (da/dN) was calculated. To determine the energy release rate over the final 500 cycles, the mean crack length and the maximum and minimum load values recorded in that interval were used. All fatigue tests were run at a frequency of 3 Hz.

Analysis of the fracture surface

Analysing fracture surfaces is essential for understanding failure mechanisms in adhesive joints, particularly when they are subjected to mode II fatigue under different environmental conditions. In this study, direct observation of both lateral edges of the specimen was performed during testing to visualize the delamination propagation trend in the adhesive joint.

The crack growth was monitored during the tests using a PULNiX TM-7CN camera equipped with a 50× magnification lens. After completing the tests, the fracture surface was analysed using a JEOL-JSM5600 scanning electron microscope to identify potential differences among the various types and aging periods examined.

In order to assess the presence of voids or bubbles in the adhesive, 100% of the specimens tested under Mode I⁴⁴ were examined. No initial defects were observed on the fracture surfaces. However, this inspection method has proven to be ineffective for Mode II, as the fracture surface is erased due to the friction between the two parts of the specimen during testing.

The fracture surfaces resulting from mode II fatigue tests were also analyzed using SEM (Scanning Electron Microscopy) to gain detailed insight into the failure mechanisms and the progression of material degradation over time.

Results and discussion

The following sections present the results obtained from the experimental study.

Static test

Table 4 shows the average values and coefficients of variation of the experimental results obtained under quasi-static loading conditions for the adhesive joint under study. It lists the critical energy release rate calculated for each fracture mode after different exposure periods in both the salt spray chamber and the hygrothermal chamber. Five specimens were tested for each exposure period.

Table 4. Mode II fracture behaviour as a function of exposure time in the hygrothermal chamber and salt spray chamber.

Exposure period [Weeks]	Saline GIIc [J/m2]	Coefficient of Variation (CV)	Hygrothermal GIIc [J/m2]	Coefficient of Variation (CV)
0	2097	7.0	2097	7.0
1	2226	6.1	2375	5.6
2	2631	6.9	2722	8.5
4	3305	1.3	2166	2.9
12	1936	2.4	1677	1.4

Based on these results, the same trend is observed for both degradation processes: the fracture energy increases with exposure time, reaching a peak before twelve weeks. At twelve weeks, a reduction in delamination resistance is noted when compared with the unexposed material. The most critical exposure period in the saline environment is four weeks, where the highest energy release rates are observed relative to the baseline (unexposed) values. In contrast, at twelve weeks of exposure, the energy release rate decreases by about 8%.

Fatigue test

Initiation of the fatigue delamination process

To increase the reliability of the experimental results, five tests were performed for each exposure condition and for each of the seven selected load levels, which corresponded to 25%, 30%, 35%, 40%, 45%, 50%, and 55% of the critical ERR obtained from the static tests. A probabilistic approach was applied across the full fatigue-life domain. All fatigue tests were conducted at a frequency of 3 Hz.

Several models are available for this kind of analysis^45,46. In this study, the Weibull regression model proposed by Castillo et al.^47,48 was selected because it enables normalization over the entire fatigue-life range and has proved effective in other investigations involving composite materials.

Figure 3 (a, b, c and d) presents the fatigue-initiation curves under mode II fracture loading for the different exposure periods considered (1, 2, 4 and 12 weeks), compared with the fatigue behaviour of material without exposure. These plots show the maximum energy release rate as a function of the number of cycles at a 5% probability of fatigue failure.

Fig. 3 [Images not available. See PDF.]

Fatigue initiation curves, showing maximum energy versus number of cycles under mode II fracture loading, for the exposed material at a 5% probability of failure, taking the unexposed material as a reference.

Figure 4 (a, b, c and d) presents the fatigue initiation curves under mode II fracture loading for the different exposure periods considered (1, 2, 4 and 12 weeks), compared with the fatigue behavior of the material without exposure. These graphs show the fatigue load level, expressed as a percentage of the value reached under static loading, versus the number of cycles, for a 5% probability of fatigue failure.

Fig. 4 [Images not available. See PDF.]

Fatigue initiation curves, showing load level versus number of cycles under mode II fracture, for the exposed material at a 5% probability of failure, using the unexposed material as a reference.

For the unexposed material, the fatigue-life field under mode II fracture ranges from 1257 J/m² to 415 J/m², corresponding to low-cycle and high-cycle (fatigue limit) regions, which represent 60% and 20% of its static capacity, respectively. For the material exposed to the selected saline environment for one week, the fatigue-life field ranges from 1113 J/m² and 493 J/m², corresponding to 50% and 22% of its static capacity. For exposure to the hygrothermal environment over the same period, the fatigue-life field ranges from 1054 J/m² and 337 J/m², representing 44% and 14% (fatigue limit), respectively, of its static capacity. This indicates greater degradation under the hygrothermal environment, especially in the high-cycle fatigue region.

After two weeks of exposure in the salt spray chamber, the results are 60% and 26% of its static capacity in the low-cycle and high-cycle regions, respectively. In the same regions, for hygrothermal exposure, the results are 47% and 20%, indicating poorer performance under the hygrothermal environment, although the trend is less pronounced than in the one-week case.

After four weeks of exposure, the trend persists. In the salt spray chamber, the results are 61% and 26% of its static capacity in the low-cycle and high-cycle regions, respectively. Under hygrothermal exposure, the results are 47% and 25% in those same regions.

Finally, after twelve weeks of exposure in the salt spray chamber, the results are 59% and 18% of its static capacity in the low-cycle and high-cycle regions, respectively. Under hygrothermal exposure for the same period, the results are 35% and 15%, which, like in the previous cases, indicates poorer performance of the material in a hygrothermal environment, more pronounced at this exposure duration.

Figure 5a and b presents the fatigue curves for the studied adhesive joints under the two degradation processes, grouping the four exposure periods into a single statistical sample. Figure 5a shows the load level, expressed as a percentage of the critical energy release rate (obtained from the prior static characterization), versus the number of cycles endured during the fatigue test. Figure 5b shows the maximum delamination energy applied to the tested specimens versus the number of cycles.

Fig. 5 [Images not available. See PDF.]

Fatigue behaviour for the two studied degradation processes and for the unexposed material, based on the load level, at a 5% probability of failure.

When the dominant variable is considered to be the percentage of fatigue strength relative to delamination resistance under static loading, grouping the four exposure periods into a single simple, it becomes apparent that the worst performance occurs under relatively high moisture and temperature (hygrothermal exposure). This poorer performance is observed throughout the entire fatigue-life range. For the material exposed to the saline environment, the effect is less pronounced; at least in the intermediate region of its fatigue behaviour, it shows trends similar to those of the unexposed material.

When using the maximum delamination energy under fatigue as a comparison parameter, it is clear that hygrothermal exposure is the most aggressive condition for the adhesive joint across the entire fatigue-life range, whereas exposure to the saline environment actually improves its fatigue behaviour compared with the unexposed material.

Table 5 summarizes the extreme load levels reached under fatigue loading for each exposure condition, shown both as a percentage of the static reference values and as the maximum delamination energy achieved in each extreme region of the fatigue curves.

Table 5. Extreme load levels reached under fatigue loading.

	Mode II [%]		Mode II [Gmax (N/m2)]		Mode II [Gmax (N/m2)]
	Low-cycle region	High-cycle region	Low-cycle region	High-cycle region	CV [%] Low-cycle	CV [%] High-cycle
Unexposed	60	19	1257	425	10.2	9.4
Hygrothermal	50	18	991	416	12.1	13.2
Saline	60	20	1519	472	9.7	11.6

These results highlight the influence of the two degradation processes on the adhesive joint. While under static loading, the first weeks of exposure produced an increase in delamination resistance (more pronounced under saline exposure) once fatigue loading was applied, the values obtained for saline-exposed material matched those of the unexposed material, in contrast to what was observed under static conditions. Under hygrothermal exposure, there is a sharp reduction in load-bearing capacity in the low-cycle region (but not in the high-cycle or theoretical infinite-life region), leading to a significant decrease in the overall fatigue-life range. In both saline and hygrothermal exposures, there is a superposition of temperature and humidity effects. In polymers of this type, moderate temperature can lead to increased toughness, while moisture has the opposite effect. In saline exposure, water molecules are accompanied by large sodium chloride molecules that slow down the degradation by delaying diffusion. For these reasons, both static and fatigue behaviour initially improves, followed by deterioration. This deterioration is delayed in the saline environment compared to the hygrothermal one, due to the reason previously explained.

Growth of the fatigue delamination process

Figure 6 shows, for mode II fracture, the crack growth rate under fatigue loading versus the normalized maximum total energy release rate. This normalization is based on the critical energy release rate obtained in the prior static characterization. The figure includes representative specimens of the material’s behaviour, which were subjected to the different exposure periods (1, 2, 4 and 12 weeks) in saline and hygrothermal environments, and it compares them with the unexposed material. Panels (a), (b), (c), and (d) in Fig. 6 correspond to the four respective exposure periods. The experimental data obtained during the fatigue tests and their trend lines are presented.

The results reveal that, overall, the trend is similar but with certain nuances. The unexposed material exhibits a lower crack propagation rate, indicating better performance than the exposed materials. The graphs show that the load level required to achieve low crack growth rates is higher in the unexposed material. However, it can be observed that, in some cases and at specific exposure times, the exposed specimens tend to approach the curve of the unexposed material. This finding implies that, although environmental exposure affects the adhesive and can increase the crack propagation rate, the material still retains strength characteristics like the unexposed material within certain loading ranges. This behaviour is more evident in the shorter exposure periods (1 and 2 weeks), where degradation is less pronounced.

Comparing the two exposure types (hygrothermal and saline) shows that specimens exposed to a saline environment generally demonstrate faster crack propagation and reduced resistance as the exposure period increases. Specimens subjected to the hygrothermal environment follow a similar pattern; however, they show slightly less, and more stable degradation compared to the saline-exposed specimens.

Fig. 6 [Images not available. See PDF.]

Fatigue crack growth rate versus normalized maximum energy release rate for the different exposure periods studied, under the selected hygrothermal and saline conditions.

Figure 7 shows the fatigue crack growth rate as a function of the normalized maximum energy release rate applied to the adhesive joint during the crack growth phase, for both degradation processes. It compares the adhesive joint without exposure to those with one, two, four, and twelve weeks of hygrothermal (Fig. 7a) and saline (Fig. 7b) exposure. As a reference, the figure includes the range of crack growth rates between 10^− 2 y 10^− 3 mm/cycle.

Fig. 7 [Images not available. See PDF.]

Fatigue crack growth rate versus normalized maximum energy release rate for the different exposure periods studied.

From the experimental results shown, for the hygrothermal exposure case (Fig. 7), similar crack growth rates are observed for one, two, and four weeks of exposure. After twelve weeks of exposure, the crack growth rate is slightly slower, tending to approach the values obtained for the unexposed material. When examining the crack growth rate range between 10^− 2 y 10^− 3 mm/cycle, it can be observed that, for all exposure periods, this range occurs at about 10% of the critical energy release rate under static loading. In contrast, for the unexposed material, it is around 16%. This suggests that hygrothermal exposure generally induces some embrittlement in the adhesive.

In the case of saline exposure, the behaviour is noticeably different. For the longer exposure periods of four and twelve weeks, the behaviour is practically the same and quite like that of two weeks, differing primarily in the low crack growth rate region, which occurs at lower energy levels. The least predictable behaviour is seen after one week of exposure, where the trend indicates that, in the slow crack growth region, propagation occurs at very low energy rates, lower than those for the other exposure periods and the unexposed material. However, in the high crack growth rate region, the energy required for propagation is considerably higher than in the other exposure periods and the unexposed material.

As in the previous case, if the crack growth rate range between 10^− 2 y 10^− 3 mm/cycle is taken as a reference, it can be observed that, in this region, crack growth occurs with an amplitude of the energy release rate between 11% and 17% of the critical energy release rate under static loading for the unexposed material and for all exposure periods, except for the one-week exposure, which shows an amplitude of 40%. This finding suggests that some degree of adhesive plasticization occurs during the first week of exposure to the selected saline environment.

From the analysis of the experimental data, a distinction can be made between the behaviour observed at the onset of crack propagation and during the stable growth phase. The initial fracture toughness reflects the resistance required to trigger crack growth and is closely associated with the local strength of the adhesive and its interfacial bond with the substrate. In contrast, the energy dissipated during crack propagation is more sensitive to mechanisms such as plastic deformation within the adhesive, frictional effects at the crack surfaces or damage accumulation. It was observed that in all environmental conditions, the propagation values tended to be higher than the initiation ones, suggesting a toughening effect once the crack was initiated. This behaviour may be attributed to phenomena such as microcrack bridging, energy dissipation due to surface roughness or localized plasticity ahead of the crack tip. The difference between initiation and propagation behaviour could also reflect changes in the local stress field or progressive degradation of the interface as crack length increases. These aspects highlight the importance of evaluating both initiation and growth properties when assessing the long-term performance of adhesive joints under variable environmental conditions.

Fracture surface

As is well known, during a three-point bending test on an ENF specimen, the displacement of the upper surface (above the neutral axis) and the lower surface (below the neutral axis) occurs in opposite directions, resulting in a shear stress that causes the crack to propagate in mode II.

Shear stresses generate the fissures at the crack tip, and during crack growth, the cusps of the fissures merge, leading to its propagation. Figure 8 shows a schematic representation of the ENF test showing the relative displacement between the upper and lower surfaces, resulting in mode II fracture. The fissures at the crack front are also depicted, along with the shear (τ). The principal tensile stresses (σ₁), that cause the fissures to open, are also showed.

Fig. 8 [Images not available. See PDF.]

Schematic of tangential and principal stresses at the crack tip.

In Fig. 9a, a schema of the fracture pattern with fissures at 45-degree angles can be observed, which is caused by the relative movement between the upper and lower surfaces. Meanwhile, in Fig. 9b, these fissures can be seen in a real specimen. As the crack grows, the fissures open up, giving rise to cusps.

Fig. 9 [Images not available. See PDF.]

Cracks generated under mode II test.

For a more detailed examination of the delamination process evolution, specific areas were selected on the fracture surface of specimens subjected to fatigue in the region near the insert that initiates the delamination process. These areas were analyzed using Scanning Electron Microscopy (SEM) to determine their topography and identify the various failure modes arising from the studied degradation processes, as well as those induced by mode II fatigue to which the tested specimens were exposed. The orientation of these cusps can provide information on the general direction of crack growth and on how the fracture surface varies depending on the degradation process experienced by the material.

In this context, the behavior of the adhesive joint is strongly influenced by the environmental conditions to which it is exposed and by the type of loading applied. Factors such as prolonged exposure to moisture, salinity, or cyclic loads can lead to significant adhesive degradation, compromising its ability to transfer loads and affecting its resistance to delamination.

To more accurately evaluate the influence of the environmental conditions considered (saline and hygrothermal environments) and the chosen exposure periods (no exposure, 1, 2, 4 and 12 weeks) on the behavior of adhesive joints subjected to mode II fatigue, an exhaustive analysis of the fracture surfaces was conducted. This approach enabled the identification of how the exposure conditions affect the adhesive’s response under the shear-dominated stresses characteristic of mode II fracture.

Figure 10 provides a summary of selected images representative of the surfaces, showing both the material and adhesive behavior in relation to the fracture type (mode II) and the aging conditions to which the specimens were subjected. Saline (b and c) and hygrothermal (d and e) environmental exposed specimens were examined at 300× magnification, and three aging conditions are presented: no exposure (a), after 1 week of exposure (b and d) and after 12 weeks of exposure (c and e), which was the maximum exposure duration studied.

Overall, the resulting fracture surfaces exhibit cohesive-type failures. In specimens exposed to the saline environment, a progressive loss of definition in the longitudinal markings characteristic of mode II is observed as the exposure time increases. This phenomenon is associated with interfacial slipping due to shear loads, consistent with previous observations⁴⁹ and other recent studies⁵⁰. Meanwhile, images obtained from the climatic chamber suggest a similar degradation pattern, indicating that more extended exposure results in greater surface deterioration. This similarity in how the surfaces evolve in both aging conditions suggests that both environments induce some degree of wear in the adhesive.

Regarding the influence of exposure time, the fracture surfaces of the unexposed material do not exhibit significant changes in the fracture plane, although some areas show a certain degree of cracking, seemingly due to minor adhesive brittleness. After one week, and for both exposure processes, the observed surfaces become more irregular, with more pronounced changes in the fracture plane, which may translate into an increase in the energy consumed during the fracture process. After twelve weeks, there are no major changes in the fracture plane, but there is noticeably more detached material, particularly under saline exposure.

This deterioration can be attributed to the combined effect of moisture absorption in the saline environment and prolonged thermal conditions in the climatic chamber. Both factors accelerate the chemical and physical degradation of the adhesive. The similarities observed in the failure patterns under both aging conditions underscore the sensitivity of the epoxy adhesive to extreme environments, although the degradation appears more rapid in the presence of salinity.

Fig. 10 [Images not available. See PDF.]

Fracture surfaces of adhesive joints subjected to mode II fatigue at different exposure times in saline and hygrothermal environments. Images obtained at 300× magnification.

In summary, the fracture surfaces of specimens subjected to Mode II fatigue generally exhibit cohesive-type failure morphologies, although subtle differences between conditions allow for further interpretation of the degradation mechanisms involved. Unexposed specimens tend to show cleaner, more planar surfaces, often with brittle fracture features and well-defined longitudinal striations aligned with the direction of sliding. These markings are typical of shear-dominated delamination and indicate a relatively stable crack front and uniform energy dissipation.

In contrast, specimens exposed to saline and hygrothermal environments show rougher, more irregular topographies. At shorter exposure durations (e.g., 1 week), the fracture surfaces exhibit signs of increased plastic deformation, such as smoother cusp formation, local adhesive stretching, and partial coalescence of shear bands. These features suggest early signs of adhesive plasticization or matrix softening, likely due to moisture absorption or thermal effects that reduce the local yield strength of the adhesive.

As the exposure duration increases, particularly after 12 weeks, the surfaces display a progressive loss of longitudinal markings and the appearance of micro-voids, fractured inclusions, and particle detachment. This trend is more pronounced in the saline environment, where the combination of moisture and salt may lead to chemical degradation of the adhesive and interfacial weakening. In several SEM images, interfacial debonding zones are observed, especially near the fibre-adhesive interface, indicating that environmental exposure reduces the integrity of the bond line and promotes mixed failure modes. Evidence of microcrack branching and secondary fracture paths further supports the occurrence of complex crack propagation mechanisms under long-term degradation.

Overall, the evolution from well-defined cohesive failure in unexposed specimens to more complex and irregular fracture patterns in exposed specimens reflects a transition from brittle to more ductile-like behaviour in the early stages of exposure, followed by increasing brittleness and damage accumulation at longer durations. These findings are consistent with the mechanical test results and reinforce the conclusion that both moisture and thermal effects significantly alter the fracture behaviour and load-transfer capability of epoxy adhesive joints under Mode II fatigue.

Conclusions

An experimental analysis was conducted on epoxy-based adhesive joints on unidirectional carbon/epoxy laminates under mode II delamination fatigue (initiation and crack growth phases), considering different exposure durations in saline and hygrothermal environments.

Analysis of the load–displacement curves confirmed that both environments degrade joint performance, with a noticeable loss of stiffness and strength observed after just one week, particularly under saline conditions. After twelve weeks of exposure, a marked reduction in strength and displacement was evident in both environments, indicating a transition to a more brittle and degraded fracture behaviour. These findings align with the decrease in G_IIc values and the trends observed during fatigue tests.

Under static loading, moderate exposure durations (1, 2, and 4 weeks) led to an increase in the energy release rate that the joints could withstand, especially under saline conditions. However, after twelve weeks, a significant reduction in load-bearing capacity was observed, particularly under hygrothermal conditions.

During the fatigue initiation phase, the hygrothermal environment resulted in the most pronounced deterioration across the fatigue-life range, while the saline environment caused a less severe decrease in performance, with trends similar to the unexposed material. In the fatigue crack growth phase, hygrothermal ageing accelerated crack propagation, although after twelve weeks, the behaviour approached that of the unexposed material. Saline exposure produced variable effects depending on exposure duration, with notably lower energy thresholds for slow crack growth observed after one week of exposure.

Overall, the results demonstrate the significant influence of environmental ageing on the mechanical integrity of adhesive joints under both static and fatigue loading. The findings underscore the importance of considering moisture and temperature effects in the design and material selection of adhesive joints intended for long-term service in offshore and humid industrial environments to ensure structural durability and reliability.

Acknowledgements

The authors acknowledge financial support from the Vice-Rectorate for Research at the University of Oviedo through the Research Support and Promotion Plan, project PAPI-22-PF-16. This research did not receive external funding.

Author contributions

Conceptualization, Antonio Argüelles.; methodology, Miguel Lozano.; validation, Jaime Viña; investigation, Paula Vigón. All authors have read and agreed to the published version of the manuscript.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. Budhe, S; Banea, MD; de Barros, S. Da silva, L. F. M. An updated review of adhesively bonded joints in composite materials. Int. J. Adhes. Adhes.; 2017; 72, pp. 30-42.

2. Ibrahim, A. H., Watson, B., Jahed, H., Rezaee, S. & Cronin, D. S. An experimental-cohesive zone model approach to predict fatigue life of adhesive joints with varying modes of loading and joint configurations for automotive applications. J. Adhes. 1–27. https://doi.org/10.1080/00218464.2024.2408372 (2024).

3. Wei, Y; Jin, X; Luo, Q; Li, Q; Sun, G. Adhesively bonded joints – A review on design, manufacturing, experiments, modeling and challenges. Compos. B Eng.; 2024; 276, 111225.

4. Budzik, M. K. et al. Testing mechanical performance of adhesively bonded composite joints in engineering applications: an overview. J. Adhes.98, 2133–2209. https://doi.org/10.1080/00218464.2021.1953479 (2022).

5. Ghabezi, P. & Farahani, M. Effects of nanoparticles on nanocomposites mode I and II fracture: a critical review. Reviews of Adhesion and Adhesives5, 414–435. https://doi.org/10.7569/RAA.2017.097312 (2017).

6. Zhang, D., Huang, Y. & Wang, Y. Bonding performances of epoxy coatings reinforced by carbon nanotubes (CNTs) on mild steel substrate with different surface roughness. Compos Part. Appl. Sci. Manuf147, (2021).

7. Kaiser, I; Zhang, C; Tan, KT. Mechanical behavior and failure mechanisms of CFRP and titanium tubular adhesive lap joints at extreme temperatures. Compos. Struct.; 2022; 290, 115528.

8. Yenigun, B., Chaudhry, M. S., Gkouti, E. & Czekanski, A. Characterization of Mode I and Mode II Interlaminar Fracture Toughness in CNT-Enhanced CFRP under Various Temperature and Loading Rates. Nanomaterials13 (2023).

9. da Silva, LFM; Rodrigues, TNSS; Figueiredo, MA; de V, MFSF; Chousal, J. A. G. Effect of adhesive type and thickness on the lap shear strength. J. Adhes.; 2006; 82, pp. 1091-1115.

10. Azari, S; Papini, M; Spelt, JK. Effect of adhesive thickness on fatigue and fracture of toughened epoxy joints – Part I: experiments. Eng. Fract. Mech.; 2011; 78, pp. 153-162.

11. Cao, J; Meng, X; Gu, J; Zhang, C. Temperature-dependent interlaminar behavior of unidirectional composite laminates: property determination and mechanism analysis. Polym. Compos.; 2021; 42, pp. 3746-3757.

12. Coronado, P; Argüelles, A; Viña, J; Viña, I. Influence of low temperatures on the phenomenon of delamination of mode I fracture in carbon-fibre/epoxy composites under fatigue loading. Compos. Struct.; 2014; 112, pp. 188-193.

13. Meng, J et al. Mode I fracture toughness with fiber bridging of unidirectional composite laminates under cryogenic temperature. Compos. Sci. Technol.; 2024; 246, 110386.

14. Castro Sousa, F; Akhavan-Safar, A; Carbas, RJC; Marques, EAS. & Da silva, L. F. M. Experimental behaviour and fatigue life prediction of bonded joints subjected to variable amplitude fatigue. Int. J. Fatigue; 2025; 190, 108583.

15. Singh, A; Aizen, S; Mega, M; Rifkind, S. Banks-Sills, L. Fatigue delamination propagation: various effects on results. Fatigue Fract. Eng. Mater. Struct.; 2023; [DOI: https://dx.doi.org/10.1111/ffe.14179]

16. Floros, I; Tserpes, K. Fatigue crack growth characterization in adhesive CFRP joints. Compos. Struct.; 2019; 207, pp. 531-536.

17. Casas-Rodriguez, JP; Ashcroft, IA; Silberschmidt, V. V. Damage in adhesively bonded CFRP joints: sinusoidal and impact-fatigue. Compos. Sci. Technol.; 2008; 68, pp. 2663-2670.

18. Shang, X et al. Fracture mechanism of adhesive single-lap joints with composite adherends under quasi-static tension. Compos. Struct.; 2020; 251, 112639.

19. Gai, D et al. Mechanical and failure analysis of outer single lap adhesive joints of carbon fiber reinforced plastics under hygrothermal conditions. Int. J. Adhes. Adhes.; 2024; 134, 103793.

20. Quaresimin, M; Ricotta, M. Fatigue behaviour and damage evolution of single lap bonded joints in composite material. Compos. Sci. Technol.; 2006; 66, pp. 176-187.

21. Campilho, R. D. S. G. & da Silva, L. F. M. Mode I fatigue and fracture behaviour of adhesively-bonded carbon fibre-reinforced polymer (CFRP) composite joints. in Fatigue and Fracture of Adhesively-Bonded Composite Joints 93–120 (Elsevier, 2015). https://doi.org/10.1016/B978-0-85709-806-1.00004-5

22. Li, J; Yan, Y; Zhang, T; Liang, Z. Experimental study of adhesively bonded CFRP joints subjected to tensile loads. Int. J. Adhes. Adhes.; 2015; 57, pp. 95-104.2015IJIE..79..95L

23. Takeda, T. Mode I fracture toughness determination and environmental durability evaluation of adhesive bonds by wedge test. Int J. Adhes. Adhes127 (2023).

24. Bello, I; Alowayed, Y; Albinmousa, J; Lubineau, G; Merah, N. Fatigue crack growth in laser-treated adhesively bonded composite joints: an experimental examination. Int. J. Adhes. Adhes.; 2021; 105, 102784.

25. María, L. et al. TESIS DOCTORAL Análisis Experimental y Numérico de Reparaciones Adhesivas de Laminados Delgados.

26. Kouno, Y; Imanaka, M; Hino, R; Omiya, M; Yoshida, F. R-curve behavior of adhesively bonded composite joints with highly toughened epoxy adhesive under mixed mode conditions. Int. J. Adhes. Adhes.; 2021; 105, 102762.

27. Rubiera, S; Argüelles, A; Viña, J; Rocandio, C. Study of the phenomenon of fatigue delamination in a carbon-epoxy composite under mixed mode I/II fracture employing an asymmetric specimen. Int. J. Fatigue; 2018; 114, pp. 74-80.

28. Yan, X et al. Mode-II fracture toughness and crack propagation of pultruded carbon Fiber-Epoxy composites. Eng. Fract. Mech.; 2023; 279, 109042.

29. Shamchi, SP; de Moura, MFSF; Zhao, Z; Yi, X; Moreira, P. M. G. P. Dynamic mode II interlaminar fracture toughness of electrically modified carbon/epoxy composites. Int. J. Impact Eng.; 2022; 159, 104030.

30. Abdel-Monsef, S; Renart, J; Carreras, L; Turon, A; Maimí, P. Effect of environment conditioning on mode II fracture behaviour of adhesively bonded joints. Theoret. Appl. Fract. Mech.; 2021; 112, 102912.

31. Akhavan-Safar, A., Jalali, S., da Silva, L. F. M. & Ayatollahi, M. R. Effects of low cycle impact fatigue on the residual mode II fracture energy of adhesively bonded joints. Int J. Adhes. Adhes126 (2023).

32. Vigón, P., Argüelles, A., Lozano, M. & Viña, J. A. Fatigue behavior of adhesive joints under modes I and II fracture in carbon-epoxy composites, influence of exposure time in a saline environment. J. Adv. Join. Process.9 (2024).

33. Saikia, PJ; Muthu, N. Estimation of CZM parameters for investigating the interface fracture of adhesively bonded joints under mode I, mode II, and Mixed-Mode (I/II) loading. Fatigue Fract. Eng. Mater. Struct.; 2024; [DOI: https://dx.doi.org/10.1111/ffe.14457]

34. Kotrotsos, A; Geitona, A; Kostopoulos, V. On the mode I and mode II fatigue delamination growth of CFRPs modified by electrospun Bis-maleimide resin. Compos. Sci. Technol.; 2023; 237, 110000.

35. Mohammadi, R., Assaad, M., Imran, A. & Fotouhi, M. Fractographic analysis of damage mechanisms dominated by delamination in composite laminates: A comprehensive review. Polymer Test.. https://doi.org/10.1016/j.polymertesting.2024.108441 (2024).

36. Moraes, CE; Santos, LFP; Leal, TPFG; Costa, ML; Botelho, EC. Moisture sorption behavior of an epoxy-based structural adhesive and its impact on the mode I interlaminar fracture toughness of bonded joints in the oil and gas industry. Int. J. Adhes. Adhes.; 2024; 134, 103799.

37. Katafiasz, TJ; Greenhalgh, ES; Allegri, G; Pinho, ST; Robinson, P. The influence of temperature and moisture on the mode I fracture toughness and associated fracture morphology of a highly toughened aerospace CFRP. Compos. Part. Appl. Sci. Manuf.; 2021; 142, 106241.

38. DU, Y; MA, Y; WANG, Z. Effect of hygrothermal aging on moisture diffusion and tensile behavior of CFRP composite laminates. Chin. J. Aeronaut.; 2023; 36, pp. 382-392.

39. Vigón, P., Argüelles, A., Lozano, M. & Viña, J. Mode II delamination under static and fatigue loading of adhesive joints in composite materials exposed to saline environment. Materials16 (2023).

40. ASTM D2584-02 Standard Test Method for Ignition Loss of Cured Reinforced Resins.

41. ASTM D3039-76. Standard test method for tensile properties of Fiber-Resin composites. Am. Soc. Test. Mater. (2017).

42. ASTM D3518-76. Stardard practice for inplane shear Stress-Strain response of unidirectional reinforced plastics. Am. Soc. Test. Mater. (2019).

43. ASTM. ASTM D7905/D7905M-14: standard test method for determination of the mode II interlaminar fracture toughness of unidirectional Fiber-Reinforced polymer. Am. Soc. Test. Mater.. https://doi.org/10.1520/D7905 (2014).

44. Argüelles, A., Viña, I., Vigón, P., Lozano, M. & Viña, J. A. Study of the fatigue delamination behaviour of adhesive joints in carbon fibre reinforced epoxy composites, influence of the period of exposure to saline environment. Sci Rep12 (2022).

45. Apetre, N; Arcari, A; Dowling, N; Iyyer, N. Phan, N. Probabilistic model of mean stress effects in Strain-Life fatigue. Procedia Eng.; 2015; 114, pp. 538-545.

46. Barbosa, J. F., Correia, J. A. F. O., Freire Júnior, R. C. S., Zhu, S. P. & De Jesus, A. M. P. Probabilistic S-N fields based on statistical distributions applied to metallic and composite materials: State of the art. Adv. Mech. Eng. https://doi.org/10.1177/1687814019870395 (2019).

47. Castillo, E; Fernández-Canteli, A. A general regression model for lifetime evaluation and prediction. Int. J. Fract.; 2001; 107, pp. 117-137.

48. Castillo, E; Fernández-Canteli, A; Pinto, H. López-Aenlle, M. A general regression model for statistical analysis of strain-life fatigue data. Mater. Lett.; 2008; 62, pp. 3639-3642.2008MatL..62.3639C

49. Ashcroft, IA; Wahab, MMA; Crocombe, AD; Hughes, DJ; Shaw, SJ. The effect of environment on the fatigue of bonded composite joints. Part 1: testing and fractography. Compos. Part. Appl. Sci. Manuf.; 2001; 32, pp. 45-58.

50. Vigón, P; Argüelles, A; Lozano, M; Viña, J. Fracture analysis under modes I and II of adhesive joints on CFRP in saline environment. Npj Mater. Degrad.; 2024; 8, 117.