Effect of hydrolytic ageing on the fracture properties of repaired carbon/polyamide 6 thermoplastic composites

Abstract

This study investigates the repair potential of carbon/polyamide 6 specimens after extensive hydrolytic degradation, using double cantilever beam (DCB) characterization. To evaluate repair effectiveness, DCB test results from repaired specimens were compared with those from aged but unrepaired specimens. The results indicate a reduction in mode I fracture toughness G_IC for most repaired specimens compared to their unrepaired counterparts (from 3.44 kJ/m² down to 0.94 kJ/m² in the non-repaired state compared to 2.44 kJ/m² down to 0.84 kJ/m² in the repaired condition). However, as ageing duration increased, the difference in G_IC values between repaired and unrepaired specimens progressively decreased, eventually converging for the longest ageing duration studied (from −29% in the unaged state down to −10% for the ultimate degradation duration). To understand this behaviour, complementary analyses, including molar mass and crystallinity ratio measurements, X-ray tomography, and SEM observations, were conducted. These investigations suggest that fibre misalignment and a weakened fibre/matrix interface after repair contributed to the observed reduction in G_IC for a given ageing duration. While prior studies have addressed the repair of thermoplastic composites, the effect of hydrolysis on their repairability remains largely unexplored. This work highlights the repair potential of thermoplastic composites even after significant irreversible degradation.

Keywords

Fracture toughness hydrolysis molar mass repair thermoplastic composites

Introduction

Composite materials have gained widespread use across various industries, including aerospace, automotive, marine, and military, due to their exceptional strength-to-weight ratio and corrosion resistance. In marine environments, where prolonged exposure to harsh conditions accelerates material degradation, composites offer significant advantages over traditional materials. However, concerns regarding their end-of-life management persist. Thermoset composites, widely used in structural applications, present significant challenges in repair and recycling due to their irreversible curing process,¹ even if recent advances with recyclable epoxies^2,3 and work performed on vitrimers^4,5 offer new possibilities for these materials. In contrast, thermoplastic composites provide a more sustainable alternative, as they can be reformed, remelted, and repaired, thereby extending their service life and reducing environmental impact.⁶

Additionally, an issue with thermoset composites is their susceptibility to delamination, which compromises structural integrity and limits repair. Thermoplastic composites, with their superior fracture toughness^7–9 and ability to be reshaped, offer a viable alternative, particularly for applications requiring long-term durability and repair. In the literature, the concept of repairing thermoplastic composites was first introduced in the 1980s by Jacquish et al.¹⁰ Following this, a significant amount of research was published on the subject in the 1990s^11–17 and there has been renewed interest since the 2010s.^18–24 A rich literature can be found on the subject, as exposed in this recent review from Barroeta Robles et al.²⁵ Similar to thermosets, repair techniques such as scarf repairs, mechanical fastening, and adhesively bonded repairs were investigated.^26–28 Then, techniques specific to thermoplastic composites (fusion bonding) were studied including induction,^29–31 resistance,^32,33 and ultrasonic welding.^34–37

To quantify the efficiency of repair, several researchers have studied the effects of repair on specimens that were previously impacted. For example, Reyes and Sharma³⁸ focused on woven thermoplastic composites (Glass/PP) subjected to low-velocity impact, which were repaired through compression moulding. They demonstrated that the repair using compression moulding significantly improved the flexural strength and stiffness compared to non-repaired specimens. Modi et al.³⁹ focused on the repair of carbon/PEEK and carbon/PEKK impacted panels using induction welding. They showed that this method improved the compression after impact (CAI) strength by 10% after repair. These two studies show that thermoplastic composites can be efficiently repaired through fusion bonding. In Double Cantilever Beam (DCB) specimens, Davies et al.⁴⁰ demonstrated that with optimized processing, C/PEEK composites could achieve post-repair (using compression moulding) fracture toughness values comparable to their pristine state. Similarly, Bolluk et al.⁴¹ reported improved fracture toughness (G_IC) values after injection repair of Elium® composites. Khan et al.⁴² also reported work on the Elium® resin on the repair of DCB specimens through healing above T_g, highlighting the effectiveness of thermoplastic repair techniques. These studies demonstrate that the repair of DCB specimens is feasible and can be utilized for research purposes.

However, in-service repair can be applied at various stages in the life of a structure, either early in its lifetime when subjected only to mechanical degradation (such as impact or delamination) or after extensive use, which combines both mechanical and chemical degradation induced by ageing. A key factor is the influence of water ingress, first investigated by the pioneering studies of Springer and colleagues.⁴³

The interaction between repair and ageing has received little attention in the literature. In the marine environment, the long-term durability of composites is crucial. Thermoplastic composites operating in such conditions are subjected to three primary degradation mechanisms: plasticization, a reversible effect caused by water absorption; oxidation, which occurs due to reactions with oxygen; and hydrolysis, an irreversible degradation process involving water-induced chain scission. While the effects of plasticization on mechanical performance have been widely studied,^44,45 fewer studies have examined the impact of irreversible degradation. In particular, little research has been conducted on the effect of ageing on fracture toughness. Ma et al.⁴⁶ investigated the effect of oxidation on the fracture toughness of C/PEEK composites and found that oxidative ageing negatively impacted fracture toughness, due to interface degradation and matrix embrittlement. Concerning hydrolysis, the authors⁴⁷ focused on the effect of hydrolytic ageing on the fracture toughness of carbon/polyamide 6 composites. Results showed that as the ageing duration increased, fracture toughness decreased. This decrease was directly associated with a loss in molecular weight, induced by chain scission and hydrolytic ageing. Additionally, a transition from ductile behaviour in the unaged state to brittle failures after ageing was observed. These two studies highlight the importance of the matrix behaviour on the overall fracture toughness response.

However, while some studies have addressed the individual effects of ageing and repair, little work has combined both aspects to assess the repair of thermoplastic composites after long-term exposure to harsh environments. Davies et al.⁴⁸ reported that flax/PLA thermoplastic composites exhibited no loss in mechanical properties after seawater ageing and recycling, suggesting that thermoplastics may retain their structural integrity even after environmental exposure (9 months). However, to the knowledge of the authors, the effect of repair on long fibre reinforced thermoplastic composites previously aged in water has not been studied. More especially, the specific interplay between chemical ageing and subsequent repair in high-performance thermoplastics remains poorly understood.

The aim of this study is to investigate the repair potential of C/PA6 thermoplastic composites, with a particular focus on their ability to undergo effective repair even after extensive chemical degradation, such as that caused by long-term service exposure in marine environments. By understanding the mechanisms governing ageing and repair through mechanical characterization, physico-chemical and imagery techniques, this research aims at evaluating repair potential of C/PA6 thermoplastic composites, ultimately contributing to extending the operational lifespan of critical structures and lower environmental footprint. This work is the first to demonstrate that press forming can restore performance in C/PA6 composites after hydrolytic ageing, even after extensive degradation.

Materials & methods

Material

The carbon/polyamide 6 material utilized in this study was provided by Celanese (Reference: CFR-TP-PA6-CF60-01) in the form of 300 mm wide unidirectional prepreg sheets with a 125 µm thickness. The prepregs have a fibre volume fraction of 48% and a fibre weight fraction of 60%. Prior to manufacture all plies were dried in desiccators at 40°C and 0% relative humidity.

Manufacturing & repair process

Unidirectional carbon/polyamide 6 panels (280 × 280 × 5 mm³) were initially fabricated in Ref. 46 through press forming using a DK Technologies press. The process involved stacking 40 plies within a mould measuring 280 × 280 mm². The panels were manufactured with a heating rate of 20°C/min up to 240°C. This temperature was then maintained for 23 min before being reduced at a cooling rate of 20°C/min down to 40°C. Pressure was applied at the end of the heating phase and maintained until the process was complete, Figure 1(a).

As stated in ASTM D5528 for Double Cantilever Beam (DCB) tests, a PTFE insert (280 × 80 × 0.025 mm³) was placed at mid-thickness on one end of each panel. DCB specimens (250 × 20 × 5 mm³) were then cut using a water jet cutting machine, such that the insert length is approximately 63 mm. The final panels exhibited a fibre volume fraction of 48%, a crystallinity ratio of 38%, and an initial molar mass of 26.2 kg/mol.

Once the specimens were tested and opened using DCB tests, they were individually repaired using a specially designed mould, as shown in Figure 1(b). It was designed to minimize mechanical clearance between the specimen and the mould, reducing the amount of material flowing out during the repair process. The repair process closely followed the original panel manufacturing procedure, with one key difference: the applied pressure. Throughout the repair process, the pressure was consistently maintained at 5 bar (red dashed curve in Figure 1(b)), as this is the minimum pressure allowed for the DK press. It should be noted that all repaired specimens in this work (unaged and aged) were tested using the DCB test prior to repair.

It is worth noting that such a repair process is not directly applicable in an industrial context. However, it is employed to enhance understanding and more especially to focus on the physico-chemical properties observed after repair.

Double cantilever beam (DCB) tests

Mode I fracture tests on the repaired specimens were conducted on an Instron 5561 testing machine, equipped with a 500 N load cell at a test speed of 1 mm/min. A Basler camera, set to a frame rate of one picture per second, was used to monitor the crack length throughout the test. Glass/polyamide 6 tabs (20 × 20 × 10 mm³) were used to ensure a good interface with the specimens. Tests were performed following ASTM 5528 using the compliance calibration (CC) method. The same test was performed both before and after repair. Three repeat specimens were tested for most conditions, with all specimens dried in desiccators at 0% humidity prior to testing. It is worth noting that for the longest ageing durations, only two specimens were successfully tested. Under these conditions, out of the three repaired specimens, the insert in one specimen slipped slightly during repair (due to lower viscosity after extensive degradation), resulting in the bonding of one side of the DCB specimen along its entire length.

Ageing

The specimens aged in this study are the same as those tested in Ref. 47. DCB specimens were immersed in deionized water at a temperature of 120°C in small pressure vessels. A pressure of 15 bar was applied to maintain liquid water during ageing. Various ageing durations were selected, ranging from 3 day to 28 days. To ensure that hydrolysis was the only degradation process occurring, the oxygen in the deionized water was removed by nitrogen saturation. The focus on hydrolysis was justified because it induces a homogeneous through-thickness degradation, unlike oxidation, and will allow a better understanding of the repair process.

After each ageing condition, all the specimens were dried in desiccators at 0% humidity until stable weight was reached. This allowed irreversible degradation to be investigated and this also ensured no additional degradation induced by residual water and hydrolysis during the repair cycle.

Differential Scanning Calorimetry – DSC

The degree of crystallinity (X_c) was determined using Differential Scanning Calorimetry (DSC) on a Q200 instrument from TA Instruments, with a heating rate of 10°C/min from ambient temperature to 300°C. In composite materials, X_c is calculated using equation (1)

X_{c} = \frac{{∆ H}_{f}}{{∆ H}_{f}^{0} \times (1 - W_{f})}

(1)

where ΔH_f represents the enthalpy of fusion of the tested polymer, ΔH_f0 is the theoretical enthalpy of fusion for a 100% crystalline material, taken as 118 J/g,⁴⁹ and W_f is the fibre weight fraction. The fibre weight fraction (W_f) was determined through Thermogravimetric Analysis (TGA) using matrix burn-off tests at 450°C for 2 h under nitrogen (N₂).

Gel permeation chromatography (GPC)

Molar mass was determined by GPC at PeakExpert Company using the method developed by Laun et al.⁵⁰ These tests enable the determination of the average molar mass by number (M_n), by weight (M_w), and the polydispersity index (PDI). In this work, only M_n values are presented. For these measurements, 10 mg samples were dissolved in 4 mL of hexafluoroisopropanol (HFiP). Detection was carried out using a Waters 2414 differential refractive index detector, and the data were analysed with PSS WinGPC Unity v7.5 SEC software. Calibration was performed using poly(methyl methacrylate) standards from PSS GmbH Mainz, Germany, with molar masses ranging from 800 to 1,600,000 g/mol, and the calibration curve was fitted using a 5th-order polynomial.

Quality control

Scanning electron microscopy – SEM

SEM microscopy was used to analyse the fracture surfaces after all DCB tests, utilizing FEI Quanta 200 equipment. Prior to observation, the samples were coated with a 60% gold and 40% palladium layer to prevent surface charging.

X-ray tomography

X-ray tomography was performed using a high-resolution GE/Phoenix tomograph at the CRT in Morlaix, France. The data were processed using VG Studio software. This technique was used to observe changes in the microstructure (such as porosities, cracks, etc.) in the unaged state, after ageing, and following repair. A 3D example of the X-ray tomography results is shown in Figure 2 (left), showing the insert region. For a more detailed analysis of the microstructure, the specimens were examined from different angles, as shown in Figure 2 (middle), for both the pristine specimen and after repair. Finally, the focus shifted to examining the cross-section in the insert zone (1’) and the crack propagation region (2’) after repair and ageing (Figure 2, right) to assess whether proper healing occurred during the repair process.

Results & discussion

Effect of repair on the unaged material

Figure 3 presents the results of the DCB tests performed on both the pristine material and the repaired specimens. Figure 3(a) illustrates the load-displacement curves, where no significant change in stiffness is observed before and after repair. However, the maximum load reached by the repaired specimens (∼160 N) is lower than that of the pristine material (∼180 N), suggesting a reduction in mechanical performance following the repair process.

Then by correlating Figure 3(a) with crack length monitoring, the critical strain energy release rate G_IC was determined, and the results are shown in Figure 3(b) After repair, a decrease in fracture toughness is observed, with G_IC values dropping from 3.44 ± 0.28 kJ/m² in the pristine state to 2.44 ± 0.47 kJ/m² post-repair, representing a 29% reduction.

Several factors could explain this reduction, including changes in the polyamide 6 matrix (such as crystallinity and molar mass), changes in the microstructure (fibre arrangement and porosity) or degradation at the fibre/matrix interface.

During the repair process, the polyamide 6 matrix undergoes a second high-temperature cycle for a certain period, which could lead to degradation due to oxidation. To monitor this, the molar mass M_n was measured. It was found to be 24.9 kg/mol after repair, compared to 26.2 kg/mol in the pristine state. This represents a 5% reduction; however, this alone cannot fully account for the observed loss in G_IC. Furthermore, as demonstrated in 47, no significant loss in G_IC was recorded within the molar mass range of 26 to 17 kg/mol. Similarly, the degree of crystallinity after repair was measured at X_c = 38%, compared to X_c = 36% in the pristine state. This slight increase also fails to explain the 29% reduction in G_IC following repair.

Therefore, further investigations were carried out with X-ray tomography to search for potential changes in the microstructure created during repair. The analysis focused on views from different angles, specifically along the three principal planes at mid-thickness. Additionally, particular attention was given to the insert region to verify its positioning after repair. Results are presented in Figure 4. In this figure, the insert appears in white, while the composite is shown in various shades of grey. Any porosities present in the material would be visible in black.

From these results, no macro-porosities are detected at this resolution (voxel size of 36 µm) in either the pristine or repaired states. This suggests that defects such as porosities are unlikely to be responsible for the observed loss in G_IC after repair. Additionally, the insert remains aligned at mid-thickness, indicating that no major fibre bundle movement occurred during the repair process.

However, the resolution of this tomography scan is not sufficient to assess finer details, such as local fibre alignment or the quality of the interface. Therefore, the fracture surfaces after the DCB tests were observed using SEM for the pristine and repaired specimens, respectively shown in Figures 5(a) and (b).

Based on these results, it is evident that the fracture surfaces before and after repair are different. In Figure 5(a), plastic deformation is observed, and the interface appears stronger than in Figure 5(b), where the fracture surface is smoother. Additionally, the fibre alignment differs noticeably between the two SEM images. To emphasize this, Figures 5(c) and (d) present the same SEM images with post-processing to highlight fibre misalignment, with each degree of misalignment represented by a different colour in the pristine and repaired states, respectively.

In Figure 5(c), all fibres are aligned, with few misalignments, whereas after repair, several fibre bundle misalignments are observed. This indicates a change in the microstructure after repair, which is associated with fibre bridging occurring during the DCB tests in the pristine state, as shown in Figure 6. When the tested DCB specimens are repositioned in the mould for repair, the bridged fibres are no longer aligned with the original 0° fibres. This explains the different fibre orientations observed in Figure 5(d).

To sum up, based on observations from X-ray tomography and SEM, the repaired composite does not exhibit macro-porosities, which could have indicated a non-controlled repair process. This absence of defects suggests that the repair process is of a good overall quality. However, significant differences were observed in the SEM fracture surface analysis. A degradation of the interface quality was noted, along with noticeable fibre misalignments induced by the repair. When combined with the physicochemical changes identified through GPC and DSC analyses, these structural changes may explain the reduction in G_IC after repair in the unaged state. The next section focuses on the effect of repair on composites that have already undergone ageing in water. The objective is to determine whether these composites can be effectively repaired after being exposed to simulated service conditions for a certain period.

Effect of repair on previously aged composites

The results from mode I fracture tests (DCB) performed on specimens aged at 120°C and then repaired are shown in Figure 7. Figure 7(a) shows the load versus displacements plots while Figure 7 presents the G_IC values as a function of crack length.

First, the plots in Figure 7(a) demonstrate that ageing has a significant impact. As the ageing duration increases, there is a corresponding decrease in load. For instance, in the unaged and repaired specimens, the peak load reaches approximately 160 N, whereas after 28 days of ageing, the peak load drops to around 80 N. Additionally, the G_IC values shown in Figure 7(b) further confirm that ageing notably affects the fracture toughness. As an example, the G_IC decreases from 2.44 kJ/m² in the unaged repaired state to 0.84 kJ/m² after 28 days of ageing and repair. Nevertheless, the G_IC value after 28 days of ageing and repair remains relatively high compared to values observed in certain thermosetting materials.⁵¹

To fully assess the effectiveness of these repairs, the results are now compared to the non-repaired data presented in Ref. 47 as a function of ageing time, Figure 8.

Results show that for the early stages of ageing (between 0 and 7 days of ageing), a significant drop in G_IC is observed after repair while results after 14 days of ageing remain relatively constant. To understand this, the changes in physico-chemical properties were analysed. Figure 9 illustrates the variations in molar masses as a function of ageing time, both before and after repair.

For both repaired and non-repaired specimens, a decrease in molar mass is observed with increasing ageing duration, dropping from 26 kg/mol to 11 kg/mol. This reduction is attributed to chain scission occurring during the hydrolysis of the amorphous phase in the polyamide 6 matrix. The results also show that, except for the 28-day ageing condition, all aged and repaired specimens exhibit an increase in molar mass of approximately 4 to 5 kg/mol after repair. This increase is significant and is associated with a repolymerization process⁵² occurring during the cooling phase of the repair. Indeed, during cooling, repolymerization takes place as the macromolecular chains cut during ageing can recombine and partially restore the polymer network. This is attributed to a polycondensation reaction between the hydrolysed end groups generated during hydrolysis. This indicates that repair not only restores structural integrity through healing but also partially reconstructs the polymer network. Additionally, results from literature suggest that molar mass can be correlated with mechanical properties, with higher molar mass generally leading to improved mechanical performance in thermoplastic polymers^53–55 and composites.^47,56 However, the decrease in fracture toughness observed in Figure 8 does not support this correlation. Therefore, further investigation is required. Figure 10 illustrates the variation in crystallinity ratio as a function of ageing time before and after repair.

Crystallinity ratio measurements show a similar trend before and after repair, with an increase in crystallinity ratio witnessed in both cases. In non-repaired specimens, this increase is attributed to the chemi-crystallization process.⁵⁷ During ageing, hydrolysis induces chain scission, allowing the resulting macromolecular chains to regain mobility. This enhanced mobility facilitates their rearrangement into new crystalline structures, explaining the increase in crystallinity ratio shown in Figure 10. For the repaired specimens, crystallinity is also influenced by the macromolecular chain length. During the repair process and more especially during cooling from the melt state, the shorter macromolecular chains regain mobility, allowing them to reorganize and form additional crystalline structures through regular crystallization. Thus, additional studies on crystallinity, such as X-ray diffraction (XRD), would be particularly valuable for analysing the amorphous and crystalline structures to help explain the loss in G_IC after ageing and repair.

However, X-ray tomography has provided valuable insights into understanding the effects of repair in the unaged state and also after ageing in Ref. 47. Therefore, additional tomography analyses were conducted after ageing and repair to investigate potential microstructural changes. Figure 11 presents cross-sections of the specimens for different ageing durations after repair. The analysis focuses both on the insert (1’) and the crack propagation zone (2’), see Figure 2, providing insights into the effectiveness of healing during the repair process.

An interesting observation from Figure 11 concerns the healing performance during the repair process following extensive ageing. In specimens that were aged and subsequently repaired after 3 and 4 weeks (see red arrows on Figure 11), healing appears to be incomplete, as revealed by the presence of a residual crack at mid-thickness, away from the insert zone. This observation is counterintuitive because a reduction in molecular weight, which occurs due to ageing, typically decreases the melt viscosity, thereby facilitating the flow of polymer chains during the repair process. This discrepancy suggests that extensive ageing may introduce additional factors that counteract the repair healing.

Also, after extensive ageing (4 weeks), transverse cracks were observed before repair, see Figure 12. These cracks are attributed to hydrolytic ageing, which causes embrittlement of the polyamide 6 matrix.⁴⁷ However, after repair (Figure 11), these transverse cracks are no longer visible. One hypothesis is that healing occurs not only in the crack propagation region but also locally at the fibre/matrix interface during the repair process.

Additionally, porosities were observed in some specimens repaired after 4 weeks of ageing, and to a lesser extent after 1 week of ageing. Two hypotheses could explain this phenomenon. First, the transverse cracks present in the non-repaired state after ageing may increase the exposed surface area during the repair process, potentially accelerating oxidation during the repair manufacturing process. Second, the reactive end groups generated by hydrolysis could further promote oxidative degradation, which is known to significantly impact the quality of composites manufactured through press forming.⁵⁸ Additionally, a previous study investigated the coupling effects between hydrolysis and oxidation in polyamide 6.⁵⁹ While evidence of accelerated degradation due to this interaction was observed, no conclusive proof was obtained to fully explain the underlying mechanisms.

Finally, Figure 12 presents the fracture surfaces of specimens following DCB tests. Figure 13(a) shows the fracture surface of a non-repaired specimen after 2 weeks of ageing. Figure 13(b) shows the fracture surface of a specimen aged in the same conditions but subsequently repaired. Lastly, Figure 13(c) displays the fracture surface of a specimen aged for 4 weeks and then repaired.

When comparing the results from Figures 13(a) and (b), three main observations can be made. First, as with the repaired specimen in the unaged shown earlier in Figure 5(d), the fibres appear less aligned after repair. This misalignment was associated with fibre bridging during the DCB test. Second, no plastic deformation is witnessed. This can be associated with hydrolysis, which changes the polyamide 6 matrix from a ductile to a brittle state as ageing progresses, as discussed in Ref. 47. Third, fewer fibres are visible and more matrix is observed in Figure 12(b). This may be associated with the healing process during repair, which may promote matrix flowing at the crack interface. Finally, when comparing Figures 13(b) and (c), no significant differences are observed, as the fracture surfaces appear similar. Although interlaminar cracks may seem more visible in Figure 13(c), this cannot be attributed to a systematic difference.

All these results demonstrate the feasibility of achieving relatively high fracture toughness values (approximately 1 kJ/m²) in repaired specimens that initially exhibited significant degradation. However, explaining the underlying mechanisms remains challenging, despite the insights provided by mechanical tests, physicochemical studies, and imaging techniques. Further research is necessary to fully understand how fracture toughness evolves with repair after ageing. Nonetheless, from a design perspective, these findings suggest that it may be possible to repair structures that have undergone extensive ageing under service conditions, provided the repair process is carefully controlled. Additionally, the repair process has not been optimized, and potential improvements could be achieved by refining the manufacturing parameters based on the degradation state of the material. Also, as stated earlier, this repair process cannot yet be directly applicable in an industrial context but allowed a better understanding of the impact of repair on the physico-chemical properties of the PA6 matrix after extensive degradation.

Conclusion

This study investigated the repair of carbon/polyamide 6 DCB specimens after extensive hydrolytic degradation. To assess this, results from DCB tests on repaired specimens were compared with those from aged but unrepaired specimens. DCB test results indicate a reduction in G_IC for all repaired specimens compared to their unrepaired counterparts. However, as the ageing duration increased, the gap between the G_IC values of repaired and unrepaired specimens progressively decreased, eventually reaching similar values for the longest ageing duration studied. To understand this, further analyses, including molar mass and crystallinity ratio measurements, X-ray tomography, and SEM observations were performed and showed that:

- Molar masses are 4 to 5 kg/mol higher compared to post-ageing values, due to repolymerization.

- Crystallinity levels remain similar to those obtained in the non-repaired state

- Fracture toughness values are lower after repair, and these are explained by changes in microstructure (such as fibre alignment), degradation of the interface, and compromised healing during the repair process.

These investigations suggested that fibre misalignment and a weakened fibre/matrix interface after repair were responsible for the observed reduction in G_IC. Despite this reduction, this study highlights the repair potential of thermoplastic composites even after significant degradation, an area that remains relatively unexplored. These findings pave the way for further research into improving composite end-of-life strategies through repair rather than replacement.

From a design perspective, the results suggest that it may be feasible to repair structures that have experienced extensive ageing under service conditions. Future work will focus on optimizing the repair process by refining manufacturing parameters to enhance mechanical performance and interface integrity.

Footnotes

Acknowledgements

This paper is dedicated to the memory of Professor George S. Springer. His pioneering studies on moisture diffusion published 50 years ago remain the reference works in our laboratory today. The authors would like to thank the scientific management department of Ifremer for supporting the REPACOMP project.

ORCID iD

Peter Davies

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

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

References

Post

Susa

Blaauw

, et al. A review on the potential and limitations of recyclable thermosets for structural applications. Polym Rev 2020; 60(2): 359–388.

Dubey

Mahanth

Dixit

, et al. Recyclable epoxy systems for rotor blades. IOP Conf Ser Mater Sci Eng 2020; 942(1): 012014, IOP Publishing.

La Rosa

Blanco

Banatao

, et al. Innovative chemical process for recycling thermosets cured with Recyclamines® by converting bio-epoxy composites in reusable thermoplastic—An LCA study. Materials 2018; 11(3): 353.

Schenk

De Calbiac

D’Elia

, et al. Epoxy vitrimer formulation for resin transfer moulding: reactivity, process, and material characterization. ACS Appl Polym Mater 2024; 6(10): 6087–6095.

Zhou

, et al. Rapidly reprocessable, degradable epoxy vitrimer and recyclable carbon fiber reinforced thermoset composites relied on high contents of exchangeable aromatic disulfide crosslinks. Compos B Eng 2020; 199: 108278.

Vodicka

. Thermoplastics for airframe applications: a review of the properties and repair methods for thermoplastic composites. Report from the Department of Defence 1996, DSTO-TR-0424.

O’Brien

. Fatigue delamination behavior of PEEK thermoplastic composite laminates. J Reinforc Plast Compos 1988; 7(4): 341–359.

Loos

. The effects of processing on interply bond strength of thermoplastic composites. J Reinforc Plast Compos 1992; 11(10): 1142–1162.

Pereira

de Moura

Moreira

, et al. Pure (I and II) and mixed-mode (I+ II) fracture characterisation of unidirectional AS4/low-melt-polyaryletherketoneTM polymer composites. J Reinforc Plast Compos 2024; 43(23–24): 1393–1404.

10.

Jaquish

Sheppard

Hill

, et al. Graphite reinforced thermoplastic composites. Contract 1980; 19(79-C): 0203.

11.

Ong

Sheu

Liou

. The repair of thermoplastic composites after impact. Tomorrow’s Materials: Today 1989; 34: 458–469.

12.

Niu

. Innovative design concepts for thermoplastic composite materials. Adv Mater: The Challenge for the Next Decade 1990; 35: 25–36.

13.

Meakin

Cogswell

Halbritter

, et al. Thermoplastic interlayer bonding of aromatic polymer composites—methods for using semi-crystallized polymers. Compos Manuf 1991; 2(2): 86–91.

14.

Davies

Cantwell

Jar

, et al. Joining and repair of a carbon fibre-reinforced thermoplastic. Composites 1991; 22(6): 425–431.

15.

Wilson

Graves

. Repair of graphite/PEEK APC-2 using thermoforming. SAMPE Q 1991; 23(1): 51–57.

16.

Rodgers

Mallon

. Repair of thermoplastic composites using induction heating. Key Eng Mater 1992; 72-74: 313–322.

17.

Kausch

Legras

. Advanced thermoplastic composites: characterization and processing. Norwich: Hanser, 1993.

18.

Kaden

. Developing a repair concept, using the advantages of carbon fibre reinforced thermoplastics, DLR-Interner Bericht. DLR-IB 435–2011/46. 2011.

19.

Nijhuis

. Repair of thin walled thermoplastic structures by melting, an experimental research, report from the national aerospace laboratory NLR. 2013.

20.

Miyake

Takenaka

. Monitoring and repair technique for interfacial de-bonding in carbon fibre reinforced thermoplastics by means of induction heating. In: ICCM: 20th international conference on composite materials, Copenhagen, 2015. ICCM.

21.

Brassard

Dubé

Tavares

. Resistance welding of thermoplastic composites with a nanocomposite heating element. Compos B Eng 2019; 165: 779–784.

22.

Villegas

. In situ monitoring of ultrasonic welding of thermoplastic composites through power and displacement data. J Thermoplast Compos Mater 2015; 28(1): 66–85.

23.

Tutunjian

Dannemann

Fischer

, et al. A control method for the ultrasonic spot welding of fiber-reinforced thermoplastic laminates through the weld-power time derivative. Journal of Manufacturing and Materials Processing 2018; 3(1): 1.

24.

Fernandez Villegas

Valle Grande

Bersee

HEN

, et al. A comparative evaluation between flat and traditional energy directors for ultrasonic welding of CF/PPS thermoplastic composites. Compos Interfaces 2015; 22(8): 717–729.

25.

Barroeta Robles

Dubé

Hubert

, et al. Repair of thermoplastic composites: An overview. Adv Manuf Polym Compos Sci 2022; 8(2): 68–96.

26.

Xiao

Hoal

Street

. Repair of thermoplastic resin composites by fusion bonding. Composites bonding 1994; 1227: 30.

27.

Davies

Courty

Xanthopoulos

, et al. Surface treatment for adhesive bonding of carbon fibre-poly (etherether ketone) composites. J Mater Sci Lett 1991; 10: 335–338.

28.

Kinloch

Kodokian

GKA

Watts

. Relationships between the surface free energies and surface chemical compositions of thermoplastic fibre composites and adhesive joint strengths. J Mater Sci Lett 1991; 10: 815–818.

29.

Ahmed

Stavrov

Bersee

HEN

, et al. Induction welding of thermoplastic composites—an overview. Compos Appl Sci Manuf 2006; 37(10): 1638–1651.

30.

Rudolf

Mitschang

Neitzel

. Induction heating of continuous carbon-fibre-reinforced thermoplastics. Compos Appl Sci Manuf 2000; 31(11): 1191–1202.

31.

Villegas

Moser

Yousefpour

, et al. Process and performance evaluation of ultrasonic, induction and resistance welding of advanced thermoplastic composites. J Thermoplast Compos Mater 2013; 26(8): 1007–1024.

32.

Ageorges

Hou

. Experimental investigation of the resistance welding of thermoplastic-matrix composites. Part II: optimum processing window and mechanical performance. Compos Sci Technol 2000; 60(8): 1191–1202.

33.

Ageorges

Hou

. Experimental investigation of the resistance welding for thermoplastic-matrix composites. Part I: heating element and heat transfer. Compos Sci Technol 2000; 60(7): 1027–1039.

34.

Villegas

. Ultrasonic welding of thermoplastic composites. Front Mater 2019; 6: 291.

35.

Jongbloed

Teuwen

Benedictus

, et al. On differences and similarities between static and continuous ultrasonic welding of thermoplastic composites. Compos B Eng 2020; 203: 108466.

36.

Tateishi

North

Woodhams

. Ultrasonic welding using tie‐layer materials. Part I: Analysis of process operation. Polym Eng Sci 1992; 32(9): 600–611.

37.

Ramarathnam

North

Woodhams

. Ultrasonic welding using tie‐layer materials. part II: Factors affecting the lap‐shear strength of ultrasonic welds. Polym Eng Sci 1992; 32(9): 612–619.

38.

Reyes

Sharma

. Modeling and damage repair of woven thermoplastic composites subjected to low velocity impact. Compos Struct 2010; 92(2): 523–531.

39.

Modi

Bandaru

Ramaswamy

, et al. Repair of impacted thermoplastic composite laminates using induction welding. Polymers 2023; 15(15): 3238.

40.

Davies

Cantwell

Kausch

. Healing of cracks in carbon fibre-PEEK composites. J Mater Sci Lett 1989; 8: 1247–1248.

41.

Bolluk

Devine

Quinn

, et al. Repair of acrylic/glass composites by liquid resin injection and press moulding. Compos B Eng 2024; 281: 111513.

42.

Khan

Irfan

Cantwell

, et al. Crack healing in infusible thermoplastic composite laminates. Compos Appl Sci Manuf 2022; 156: 106896.

43.

Shen

Springer

. Moisture absorption and desorption of composite materials. J Compos Mater 1976; 10: 2–20.

44.

Hoa

Lin

Chen

. Hygrothermal effect on mode II interlaminar fracture toughness of a carbon/polyphenylene sulfide laminate. J Reinforc Plast Compos 1992; 11(1): 3–31.

45.

Arhant

Le Gac

Le Gall

, et al. Effect of sea water and humidity on the tensile and compressive properties of carbon-polyamide 6 laminates. Compos Appl Sci Manuf 2016; 91: 250–261.

46.

Wen

Wang

, et al. Effect and mechanism of thermal ageing on mechanical properties of continuous carbon fiber‐reinforced PEEK composite laminates. Polym Compos 2024; 45(14): 13256–13271.

47.

Arhant

Lolive

Bonnemains

, et al. A study of pure hydrolysis of carbon fibre reinforced polyamide 6 composites tested under mode I loading. Compos Appl Sci Manuf 2022; 152: 106719.

48.

Davies

Le Gall

Niu

, et al. Recycling and ecotoxicity of flax/PLA composites: influence of seawater ageing. Composites Part C: Open Access 2023; 12: 100379.

49.

Mark

. Polymer data handbook. New York: Oxford Academic publishers, 2009.

50.

Laun

Pasch

Longiéras

, et al. Molar mass analysis of polyamides-11 and-12 by size exclusion chromatography in HFiP. Polymer 2008; 49(21): 4502–4509.

51.

Sela

Ishai

. Interlaminar fracture toughness and toughening of laminated composite materials: a review. Composites 1989; 20(5): 423–435.

52.

Meng

Wang

, et al. Efficient hydrolytic recycling of PA6 supported by kinetic modeling. Polym Degrad Stabil 2025; 233: 111178.

53.

Deshoulles

Le Gall

Dreanno

, et al. Origin of embrittlement in Polyamide 6 induced by chemical degradations: mechanisms and governing factors. Polym Degrad Stabil 2021; 191: 109657.

54.

Arhant

Le Gall

Le Gac

, et al. Impact of hydrolytic degradation on mechanical properties of PET-towards an understanding of microplastics formation. Polym Degrad Stabil 2019; 161: 175–182.

55.

Jacques

Werth

Merdas

, et al. Hydrolytic ageing of polyamide 11. 1. Hydrolysis kinetics in water. Polymer 2002; 43(24): 6439–6447.

56.

Le Gac

Fayolle

. Impact of fillers (short glass fibers and rubber) on the hydrolysis-induced embrittlement of polyamide 6.6. Compos B Eng 2018; 153: 256–263.

57.

Rabello

White

. Crystallization and melting behaviour of photodegraded polypropylene—I. Chemi-crystallization. Polymer 1997; 38(26): 6379–6387.

58.

De Almeida

Feuillerat

Fontanier

, et al. Determination of a degradation-induced limit for the consolidation of CF/PEEK composites using a thermo-kinetic viscosity model. Compos Appl Sci Manuf 2022; 158: 106943.

59.

Deshoulles

Le Gall

Dreanno

, et al. Chemical coupling between oxidation and hydrolysis in polyamide 6-A key aspect in the understanding of microplastic formation. Polym Degrad Stabil 2022; 197: 109851.