On the response to hygrothermal aging of pultruded FRPs used in the civil engineering sector

This paper presents the effects of hygrothermal aging on the durability of a pultruded flat sheet, immersed in distilled water at 25 o C, 40 o C, 60 o C or 80 o C for a period of 224 days. Elevated temperatures noticeably increase the moisture diffusion coefficient and moisture uptake behaviour. Measured changes in the tensile and in-plane shear mechanical properties were examined after 28, 56, 112 or 224 days. Tensile properties remained practically unaffected by aging whereas matrix dominated shear properties revealed an initial drop which was recovered to a substantial degree after further hygrothermal aging. Visco-elastic property changes due to the superimposing mechanisms of plasticization, additional cross-linking etc. were recorded. Scanning Electron Microscopy micrographs indicate that the fibre/matrix interface remained practically intact, even after the most aggressive hot/wet aging. X-Ray Energy Dispersive Spectroscopy analysis showed no chemical degradation incidents on the fibre reinforcement surfaces and infrared spectroscopy revealed superficial chemical alteration in the aging matrix. Optical microscopy revealed matrix cracking in samples aged at 80 o C for 112 days. Lastly, Computed Tomography scans of un-aged material showed internal imperfections that undoubtedly enhanced moisture transport. After aging at 60 o C for 112 days, Computed Tomography detected preferentially situated water pockets.


Introduction
Fibre Reinforced Polymer (FRP) materials are used in engineering structures as load carrying elements as they can provide relatively high resistance to aggressive environmental conditions [1,2]. Owing to their relative high strength and stiffness to weight ratios, and perceived enhanced durability benefits over traditional structural materials, there is growth in using FRP shapes and systems. The application of FRPs to repair and strengthen existing infrastructure is routinely practiced [3].
Exploitation of FRPs as primary and secondary structural elements [1,2] is less developed within the civil engineering sector, and is a topic for materials and structural engineering research and development. The specific properties of FRPs [4][5][6] compared with structural grades of steel has activated applications in bridge deck constructions for short span highway bridges, as well as lightweight pedestrian footbridges. In many civil engineering applications, the composite processing method is pultrusion [4,7,8], since it is cost efficient in producing continuous lengths of constant cross-section shapes, having fibre volume fractions of 40-60%. Both mechanical fasteners and onsite adhesive bonding are used as methods of connection to fabricate primary structures from individual pultruded shapes [9]. Owing to the method of production, pultruded shapes (herein referred to as PFRP shapes) are highly orthotropic materials and their thin-walled cross-section is determined by the die's shape [10]. The direction of pultrusion is also the longitudinal direction for the PFRP shape. Often, off-the-shelf pultrudates normally have reinforcement of glass, and a matrix based on either a polyester, vinylester or polyurethane resin.
Despite service lives of 50 years or higher, civil engineering structures often lack the routine inspection and maintenance found in the aerospace and marine sectors. As a consequence, knowledge of their long-term service performance is a prerequisite for economic, safe and reliable design. Research in the field of assessment and standardization for the longterm mechanical properties of FRPs, and their structures, is therefore in high demand. A growing number of research studies is continuously being published, which can be seen as aiming to offer designers a comprehensive database for the durability performance of FRPs. Karbhari et al. [2] have documented the reasons that created the 'gaps' in the transfer of knowledge and technology from the aerospace to the civil engineering sector.
Moisture ingress is one of the phenomena that can adversely affect the long term durability of FRPs. The penetration of moisture penetrates from exposed surfaces induce both reversible and irreversible changes to the composite constituents [11]. Reversible changes are physical in nature, involving both property (durability) [12] and dimensional changes (swelling) [13]. Depending on the conditions and extent of hygrothermal exposure, some aging effects i.e. plasticization or softening, can be recoverable at the initial stages of aging when the absorbed water is removed and no chemical reaction occurs [14,15]. Prolonged environmental exposure leads to irreversible changes that induce permanent property alterations within the matrix, the fibre surfaces and the fibre/matrix interface. Matrix micro-cracking [16], chain scission, residual cross-linking, hydrolysis, oxidation and plasticization are the major effects of the presence of moisture and moisture gradients [17]. The severity of the exposing conditions increases with the energy from increasing temperature.
The fibre/matrix interface and the interphase region around each individual fibre, of a few tens of nanometer thickness, are susceptible to deterioration in the presence of moisture. With adhesion between the fibres and matrix reduced due to deterioration [18], there is enhanced interfacial capillary action along the fibres that can further promote their degradation.
It has been also reported that moisture attacks the fibre reinforcement leading to mechanical property loss [19]. Capillary action or 'wicking' along the fibre reinforcement has been shown to be more pronounced than moisture diffusion in the matrix of a composite, thus degrading more prominently the fibre/matrix interface rather than the matrix or/and fibres individually [20]. The latter observation is verified by the higher amount of absorbed moisture uptake in reinforced composites vs. unreinforced plain resin polymers [14,21]. Based on that fact, moisture-induced local damage in a FRP such as hydrolysis, chain scission and permanent plasticization, is expected to be more pronounced than that in the respective plain polymer composite, due to the fundamental difference in the relative amount of total absorbed moisture [22]. All the above affect the durability of the materials and therefore modify their short and long-term performance.
Investigations towards an understanding of the long-term durability of FRPs in practical time periods can be simulated by 'short-term' accelerated aging [12,[23][24][25][26][27][28][29][30][31][32][33]. Accelerated aging corresponds to aging in severe conditions that provokes degradation and subsequently reduces aging time to a realistic short-term period. Additional energy applied to the system in the form of elevated temperature, increases the material's molecular mobility accelerating the chemical reactions which occur during the degradation process. However, the applicability of the accelerated aging approach depends on the material's chemical nature (chemical stability, degree of cure etc.) which may alter at high temperatures [34][35][36].
In a pure polymer matrix system, moisture absorption usually follows a Fickian trend [34]. Extreme environmental conditions, such as elevated temperatures and prolonged exposure times induce intrinsic chemical changes leading to deviations from the classical Fickian model [21,35,37]. Observed deviations can be attributed to polymer relaxation [38] and incomplete polymer curing, which ultimately allows for residual cross-linking and the leaching out of low molecular weight segments [21]. These major structural changes lead to an anomalous behaviour that impedes data interpretation and confuses the long-term prognosis. For engineering quantification the extrapolation of mechanical property variations, from aging at elevated temperatures, is typically employed to make long-term property predictions through an Arrhenius-type approach [39].
The coupling of moisture uptake with increased temperature is found to significantly influence the mechanical properties of the matrix, especially when the aging temperature approaches the Glass Transition temperature (Tg). In the early stages of aging, the presence of moisture depresses Tg by acting as a plasticizer of the matrix. Beyond a certain aging point, matrix and/or fibre/matrix interface/interphase degradation may result in leaching of low molecular weight segments. This loss of species, seen as a mass loss in the gravimetric measurements, increases the rigidity of the polymer and effects positively the Tg [12]. Leaching of non-bound substances causes an increase in Tg since non-bound segments usually act as plasticizers and subsequently their absence causes less mobility in the polymer in the longer-term [35]. Zhou and Lucas reported a similar phenomenon of Tg alterations based on un-bound and bound water molecules [15]. They reported the presence of Type-I bound water at the initial stages of aging that causes polymer plasticization and depression of Tg. After prolonged aging, the Type-II bound water leads to additional side cross-links that contribute to a Tg recovery, or a lowering in the reduction. Zhou and Lucas reported that Tg is minimum when the material moisture peaks and that Tg values are expected to elevate post-saturation [15]. Moreover, they witnessed that the higher the aging temperature the greater Tg was recovered.
Interestingly, should the composite be partially cured, additional, or 'residual', cross-linking may take place, complicating further any characterization evaluation of hygrothermal aging. Additional cross-linking due to re-activation of curing, causes Tg to rise [40,41].
The presence of moisture in polymer composites induces swelling which develops internal stresses that may instigate micro-cracking [42]. The generation of micro-cracks may increase the rate of moisture ingress and hence increase the severity of deterioration.
It is undeniable that the interference of the different mechanisms which may change when transitioning through Tg [43], may trigger fluctuations in the durability of the material which eventually may lead to negligible deterioration or even improvement of the mechanical properties [34]. The complexity of the hygrothermal aging process increases when the different mechanisms superimpose with different rates and initiation times [44].
Fundamental works towards the understanding of the durability of FRPs have been conducted by Surathi et al. [34], Berketis et al. [46], Schutte [47], Sethi et al. [48] and Maxwell et al. [30]. Keller et al. studied the effects of natural weathering on the durability of FRPs in real service conditions [49]. In the latter, it was reported that the system and material stiffness remained unaffected after 17 years of exposure to severe Alpine environment. On the contrary Keller et al. found that the material's strength was markedly reduced, without yet influencing the structural safety of the construction [49]. The use of the Arrhenius-type protocols, such as in the proposed model from Bank et al., [39] has found a growing interest amongst those wishing to make predictions for mechanical property values in the long-term, throughout short-term hot/wet aging tests [50]. Using such protocols the durability of polymeric composites exposed to various environments has been studied in terms of the: FRP itself [26]; fibre reinforcement [51,52]; matrix [51,53]; fibre/matrix interface and interphase [54][55][56].
Despite the growing body of research in this field, the unquantified durability performance of FRPs is perceived as a major drawback hindering their wider exploitation in civil engineering structures [34]. Poor understanding of how materials change internally with in-service conditions can, undoubtedly, lead to poor designs and executions, requiring maintenance, or replacement much sooner than is desirable to achieve an acceptable level of sustainable construction [34]. This paper reports experimental findings of an extensive study concerning the effects of moisture on an 'off-the-shelf' Microscopy-SEM and Optical microscopy); chemical (X-Ray Energy Dispersive Spectroscopy-EDS); tomographic (Computing Tomography-CT). The comparative study based on the evaluation of a portfolio of analysis results is used to extend the understanding on the behaviour of the changing properties of a PFRP flat sheet exposed to aggressive aqueous environments.

Material
The material is a five-layered glass PFRP flat sheet (FS040.101.096A) from Creative Pultrusions Inc., USA. The nominal thickness is approximately 6.4mm. The outer surfaces of the laminate are covered by a thin protecting and non-structural polyester veil, which has the dual functions of retarding moisture ingress and protecting the PFRP from UV radiation. E-CR glass fibres serve as the reinforcement and the matrix is of an isophthalic polyester resin with unknown additives. Fig.   1 displays the structure of material comprising three Continuous Strand Mat (CSM) layers (for 2.75mm thick) and two layers (for 3.65mm thick) of unidirectional (UD) reinforcement. The fibre volume fractions in the two reinforcement types are 33.3% and 54.5%, respectively, giving an overall volume fraction of 45%.

Testing
Sub-sections 2.2.1-2.2.3 provide summary details regarding the tests for moisture uptake, measurement of tensile and shear mechanical properties, and of viscoelastic properties. Sub-sections 2.2.4-2.2.6 introduce the SEM, EDS, FTIR and X-ray Computed Tomography (CT-scan) characterization methods.

Moisture absorption
For the moisture ingress measurements specimens were oven-dried at 30 o C for 72h in order to remove absorbed moisture from the environment and to ensure standard initial conditioning. Square plates of 200mmx200mm and 6.4mm thickness (h) were immersed in distilled water at 25 o C, 40 o C, 60 o C or 80 o C in thermostatically controlled Grant SUB36 water tanks.
The specimen size was chosen so that it resembled real conditions with moisture absorption occurring mainly in the thickness direction. Prior to immersion specimens were weighed using a digital scale with 0.001g sensitivity. Following immersion they were removed at pre-determined times in order to weigh the water 'uptake' in accordance with ASTM D5229. Mass (M(%)) change was determined using: where Mt is the measured mass at time t and Mo the initial dry mass. Moisture uptake measurements were conducted over a period of 224 days [39]. To calculate moisture diffusion coefficient the second Fick's law is commonly employed assuming uniform moisture and temperature conditions within the sample: where x the is the distance in the thickness (h) direction, C is the concentration of the water and D a constant diffusion coefficient [57]. Moisture absorption and diffusion processes are functions of temperature, the type of FRP, fibre orientation and the fibre volume fraction [58,59]. By assuming that classical Fickian diffusion takes place, D is calculated from the initial linear relationship of the moisture uptake with time. From the gravimetric curve shown in Fig. 2 the slope is: In Equ. (3) M1 and M2 are the masses of the specimen at times t1 and t2 obtained from the linear part of the absorption process. The maximum absorbed moisture content is M∞. In Fickian behaviour, D is a constant with respect to moisture concentration and temperature. According to [57] the solution of Equ. (2) for the one-dimensional approximation is: Equ. (4) can be simplified for a short-term and a long-term approximation, as: for Dt/h 2 < 0.04, and (5) for Dt/h 2 > 0.04.
As long as M∞ can be established from the gravimetric curve, D for a statistical homogeneous material can be determined from: where l, w and h are for length, width and thickness of the rectilinear shaped specimen. In formulating Equ. (7) it is assumed that the moisture diffusion uptake occurs predominantly in the through-thickness direction. The final dimensional parameter in Equ. (7) accounts for the contribution in moisture absorption through the (cross-section) edges of the material [60]. Prolonged exposure can induce a deviation from this classical model and non-Fickian diffusion is observed [ For the tensile coupon testing a DARTEC servo-hydraulic testing machine equipped with a 100kN load cell calibrated to BS EN ISO 7500-1 was used. To determine the strength and modulus of elasticity in the pultrusion direction, testing conducted on straight sided specimens ( Fig. 3(a)) having dimensions of 250mmx25mm in accordance with ISO 527-4:1997. Aluminium end tabs were bonded using Araldite 2015 epoxy adhesive in order to prevent from undesirable gripping failure during testing. Load was applied at a constant stroke rate of 2mm/min. FLA 10-11 direct strain gauges were bonded at mid-positions on one of the 25mm wide surfaces in order to record longitudinal strain.
To determine in-plane shear properties, both the (anticlastic) Plate-twist and Iosipescu test methods were employed. Figs.

3(b) and 3(c)
show the test set-ups with the two different sized and shaped coupons. Loading was applied using a 50kN capacity Instron (3369 series) testing machine that conforms to BS EN ISO 7500-1. Plate-twist tests were conducted following ISO 15310:1999, whereas Iosipescu testing according with ASTM D5379/D5379M-98. Plate-twist specimens were 230mm×230mm×6.4mm and as shown in Fig. 3(b) the specimens were twisted using a specific rig having four loading points spaced at ~305 mm. A second specific rig was required for the Iosipescu test method, and as shown in Fig.   3(c) the specimen of 20mm×76mm×6.4mm has two opposite-facing V-notches. In both shear tests a constant stroke rate of 1mm/min was utilized.
For each temperature and aging time in the text matrix, batches of five or three nominally identical specimens were tested to establish the changing tensile or in-plane shear properties.

Dynamic Mechanical Thermal Analysis
DMTA was used to examine and compare the visco-elastic properties between the un-aged and aged materials. Samples of the PFRP flat sheet were analysed using a TTDMA Dynamic Mechanical Analyser (Triton technology, UK) on dual cantilever beam mode from 20 o C to 200 o C. Tg was defined by the peak of the tanδ curve.

Scanning Electron Microscopy
An JSM-6480LV (JEOL, Japan) SEM was employed to scrutinize the aging effects at the micro-structural level under variable pressure mode at 60 Pa to alleviate charging that would impair the imaging quality. Both un-aged and aged PFRP samples were inspected. All examined samples were metal coated with a 10nm chromium layer to enhance the detection resolution. Captured SEM images were analysed on a chemical basis using X-Ray Energy Dispersive Spectroscopy (EDS).

Fourier Transform Infrared Spectroscopy
Un-aged and aged materials were analyzed in the infrared spectra with the aim of identifying any chemical changes in the matrix due to aging. Particles of material from the bulk PFRP were collected, and for scanning, the matrix was separated from the fibres. Scans were carried out in the range of 500cm -1 to 4000cm -1 , using a Perkin Elmer Frontier spectrometer, equipped with a diamond MIRacle Attenuated Total Reflectance (ATR) accessory (PIKE Technologies, USA) and a Deuterated Triglycine Sulfate (DTGS) detector with KBr optics (PerkinElmer, USA).

Computed tomography
The internal structure of the PFRP flat sheet was examined using a Nikon XT H 225 CT-scanner. In the case of un-aged material the tomography was employed to intrinsically identify cracks and voids. Interrogating aged samples (in their wet condition) had the objective of finding internal wet regions where the absorbed moisture had concentrated. The 'wet' samples were sealed in a plastic bag in order to prevent moisture loss during the scanning process. As illustrated in Fig. 4 five flat sheets of 40mmx40mm were stacked to have a sample of ~32mm thickness.

Results and Discussion
In this section of the paper the test results are presented, evaluated and discussed in terms of developing an understanding of the competing mechanisms that change the mechanical properties with aging temperature and aging duration up to 224 days. days, prior to M(t) dropping to a significantly lower mass. In this context, secondary effects refer to chemical decomposition mechanisms that lead to mass loss; that was previously reported and analyzed in a recent work by

Moisture absorption
Grammatikos et al. [45]. Weight changes are known to be due to: i) moisture absorption; ii) leaching out of un-bound segments (decomposition); iii) leaching out of hydrolysis products (decomposition) [12,15,45]. The specific behaviour at 80 o C can be used to propose that the overall chemical decomposition might have been activated or, alternatively, have been more prominent after a relative short soaking period. This highlights the fact that moisture absorption and chemical decomposition mechanisms are superimposed. Moreover, it is observed that the moisture absorption process is more prominent in the initial immersion period and that chemical decomposition is most prominent in later stages, after giving a clearly identifiable peaked response in the M(%) values.
On the assumption that the limit of absorption is time, not temperature, dependent, it is possible that M∞ at the three lowest temperatures can be equal or similar to that indicated and measured at 80 o C. Based on the work of Chin et al. [61] it could take several years for a composite laminate at 40 o C to attain moisture saturation.  Table 1 and Fig. 5(b) indicate that the diffusion coefficient exponentially increases with T, thereby verifying the responses plotted in Fig. 5(a).    As seen by inspecting the results in Fig. 6(b) the retention of EL seems to fluctuate around the un-aged value (100%) at each temperature, following an initial enhancement at 28 days. EL is seen to remain fairly constant to 224 days at 25 o C and  Due to being matrix dominated, the equivalent shear properties (Fig. 7) have been affected more than the longitudinal tensile properties (Fig. 6), which expectedly remain practically constant. The monotonic reduction in shear properties for 112 days is mainly attributed to plasticization of the polymer matrix of the composite. Nevertheless, other degrading mechanisms such as micro-cracking might have contributed to the depression of shear properties. It is worth to mention that beneficial mechanisms such as additional cross-linking might also be present, however, their effect is not likely to be distinguished unless it overcomes the potential of plasticization. In return, the final improvement in shear properties at the end of the aging regime could be attributed to additional cross-linking that occurs in the partially-cured matrix and significant leaching of low molecular weight segments. Secondary cross-links between the polymer chains and the water molecules have been also reported to occur [15]. Decomposition at this stage is present it is however being overshadowed by extensive residual curing.    Table 2 lists the Tg values of the un-aged and aged materials. The important finding is that Tg has a general decreasing trend with exposure time, which is known to correspond to plasticization and micro-cracking due to moisture-induced swelling of the bulk of polymer composite [27]. These material changes were clearly mirrored by way of the initial decreasing trend in the matrix dominated shear properties presented in  Increases in moisture content have been shown to lead to increases in the magnitude and shifts to a lower temperature in loss factor curves (tan) [12,62]. The drop in Tg shown in Fig. 8 may be due to the increasing presence of water within the polymer. Distribution of intermolecular chemical bonds will induce a further decrease in Tg that will affect negatively the mechanical properties. In this characterization work, it is found that storage modulus (E'), tan and Tg all vary according to the different structural mechanisms that control, at different times, the physicochemical and mechanical properties [12]. incidents owing to the lack of the necessary thermal energy to re-activate curing. Samples aged at 40 o C ( Fig. 9(b)), highlighted a more complicated behaviour. There is an increase in E' (glassy region) after 28 days, followed by a decrease for the 56 and 112 days and a final recovery after 224 days. In this case the beneficial effects of cross-linking and leaching of low-molecular segments were more dominant than plasticization over the 28 days period and therefore the material was found to increase in stiffness. This behaviour was not continued for the following 112 days in which E' returned to values similar to the un-aged material. The effect of plasticization is verified by the broadening of the tanδ (Figs. 9(a) and 9(b)) curves with aging. This behaviour is more obvious for samples aged at 25 o C than 40 o C. Broadening of tan response implies higher molecular mobility which paves the way for plasticization [12,40,46]. At 60 o C (Fig. 9(c) The visco-elastic behaviour of aged PFRP at 80 o C ( Fig. 9(d)) was similar to that at 60 o C. It is worth mentioning that samples aged at 60 o C and 80 o C reached higher E' values as opposed to samples aged at 25 o C and 40 o C revealing the larger effect of additional cross-linking and leaching of low-molecular weight segments. (-··-).

Scanning electron microscopy
Using the approach introduced in Sub-section 2.2.4 SEM micrographs of fractured surfaces from failed tensile specimens are shown in Figs. 10 and 11. Images for the PFRP aged to 224 days can be compared with the un-aged material. Fig. 10 presents fractured surfaces with both matrix and fibres, whereas in Fig. 11 the higher resolution images are for the matrix only. Initial inspection finds that the fibres are firmly connected/embedded into the matrix indicating the presence of an acceptable fibre/matrix adhesion at the four aging temperatures. Closer examination of the fibre surfaces themselves in  The surface of the fibre reinforcement was analyzed using Energy Dispersive Spectroscopy (EDS).  Table 3 is attributed to the presence of residual matrix particles on the surface of the fibres.

Fourier Transform Infrared Spectroscopy
Samples of matrix (separated from the fibres) from both un-aged and aged material were scanned in the infrared spectra.
Infrared signatures of the aged PFRP (red curves) have been plotted against the un-aged spectra (blue curve) in Fig. 12, with the aim of identifying any significant chemical reactions that might have occurred during the hygrothermal aging. It is known that ester groups in polyester resins are susceptible to hydrolysis and the higher the aging temperature the more severe hydrolytic attack is expected to occur. Figs Owing to the presence of O-H there is a broad peak with weak intensity in the four parts of Fig. 12, in the wavelength region of ~3700 cm -1 . The next weak peak which sets at around 2960 cm -1 is attributed to C-H stretching. A third peak at ~1730 cm -1 indicates the presence of ester groups in the polyester (the carboxyl acid groups). The severity of decomposition can be determined by the maximum intensity of the O-H peak at 3700 cm -1 and the C-H peak at 2960 cm -1 . Overall, evaluation of the IR spectrums of the aged materials suggests that chemical decomposition by hydrolysis was insignificant, since, with increasing aging temperature, the change in the peaks of the O-H band is ~1% and < 2% [64]. This finding from the FTIR analysis enhances the aforementioned understanding reported in Section 3.2 for the causes of the observed changes in the mechanical properties and PFRP material's microstructure.

Computed tomography scan
Un-aged (dry) and aged (wet) samples were scanned with the aim of identifying internal imperfections and wet regions, respectively. In each case, a stack of 5 replicates was inspected for statistical purposes.

Optical microscopy
The exposed faces (cross-sections) of samples aged for 112 days at 60 o C and 80 o C were analyzed with an optical microscope with a view to identifying the presence of surface corroding deterioration caused by hot/wet aging. The three images in Fig. 14 allow for comparison between the un-aged and aged PFRP materials. As can be seen by the 'black' areas appearing in Fig. 14(b), aging induces characteristic 'pitting' corrosion around the fibres. The severity of this phenomenon leads to cracks through the matrix, as seen in Fig. 14(c). Such matrix cracking will accelerate moisture absorption and could trigger the development of other degradation mechanisms such as delaminations, hydrolysis, fibre corrosion etc.

Concluding Remarks
This paper presents an examination of the effects of hygrothermal aging on the durability of a pultruded E-CR glass fibre reinforced polyester matrix. Samples cut from a 6.4mm thick flat sheet were immersed in distilled water at 25 o C, 40 o C, days. Physico-chemical effects were characterized by analysis using SEM, EDS, FTIR, CT-scanning and Optical microscopy. It was found that:  Elevated temperatures accelerate the moisture absorption rate and moisture diffusion coefficient (0.42 to 3.2610 -6 mm 2 /s). Aging at 80 o C (which is more than 40 o C below the un-aged material's Tg) revealed substantial mass loss that can be attributed to decomposition and leaching out of low-molecular weight segments into the water medium.
 Tensile strength and modulus of elasticity in the direction of pultrusion remained virtually unaffected, except for aging at 80 o C.
 Determination of in-plane shear modulus and in-plane shear strength showed that matrix dominated properties were significantly affected by hygrothermal aging. After an initial lowering, the 224 day test results showed that these mechanical properties had significantly increased compared to test results for 112 days. Changes witnessed from the mechanical testing data are believed to be due to the superimposing mechanisms of swelling, plasticization, leaching out of low-molecular weight segments and additional cross-linking.  Energy Dispersive Spectroscopy (EDS) analysis showed that no significant chemical degradation occurred on the E-CR glass fibre surfaces after hygrothermal aging for 224 days.
 Infrared spectroscopy (FTIR) analysis revealed minimal chemical changes in the aged matrix. Matrix cracking, that enhances moisture ingress and subsequently accelerate aging, was detected in exposed faces (cross-section edges) of samples after aging for 112 days at 80 o C via optical microscopy.
 Computed tomography scanning provided imaging information to locate, in un-aged material, internal imperfections in the form of cracks along the unidirectional roving bundle lengths that undoubtedly, promotes moisture transport into the body of the laminate. CT-scanning also detected that there were preferentially pockets of water inside samples aged at 60 o C for 112 days.
Using a standard accelerated testing protocol and combining mechanical testing with a programme of material characterization analysis methods, a qualitative understanding has been developed on the effects of competing degradation mechanisms on the durability of a commercially available pultruded flat sheet. The reported test results revealed the matrix and interface/interphase dominance on the material's response during aging.