Effect of volume fiber and crack length on interlaminar fracture properties of glass fiber reinforced polyester composites ( GF / PO composites )

A delamination is one of the main damage modes of two-dimensional fabrics reinforced structural composites parts [1-2]. In fact, a delamination is the result of interlaminar crack propagation. The interlaminar fracture toughness becomes one of the important indexes of the properties for the matrix [3-4]. In general, a delamination within a composite structure would be subjected to a mixture of mode Ι (opening), mode II (forward shear), mode III (anti-plane shear) crack driving forces [5-6]. It is more appropriate that interlaminar fracture toughness of the composites is described by critical energy release rate GIc, GIIc and GIIIc [5]. Mode Ι delamination is the lowest fracture energy mode; consequently most attention has been focused on assessing mode Ι delamination propagation as a measure of the damage tolerance of structural polymer composites parts. Previous literature showed characterization of mode I and II interlaminar fracture behaviors was developed using carbon fiber reinforced high performance matrixes such as epoxy resins (PEEK) [7-9]. Consequently, these testing protocols had been predominantly applied to the study of carbon fiber based materials. There were few investigations of the interlaminar fracture behaviors of other thermosetting composites which were reinforced by glass fiber. Until now, there are no studies on the interlaminar fracture behaviors of glass fiber mat reinforced polyester composites. In this paper, mode Ι interlaminar fracture behaviors of the pure polyester resin and GF/PO composites by use of double cantilever beam (DCB) tests were studied and compared with that of a standard, unidirectional carbon fiber reinforced epoxy composite(CF/EP composites). The fracture area along crack propagation was analyzed by scanning electron microscope (SEM) to identify the fracture characteristics.


Introduction
A delamination is one of the main damage modes of two-dimensional fabrics reinforced structural composites parts [1][2].In fact, a delamination is the result of interlaminar crack propagation.The interlaminar fracture toughness becomes one of the important indexes of the properties for the matrix [3][4].In general, a delamination within a composite structure would be subjected to a mixture of mode Ι (opening), mode II (forward shear), mode III (anti-plane shear) crack driving forces [5][6].It is more appropriate that interlaminar fracture toughness of the composites is described by critical energy release rate G Ic , G IIc and G IIIc [5].Mode Ι delamination is the lowest fracture energy mode; consequently most attention has been focused on assessing mode Ι delamination propagation as a measure of the damage tolerance of structural polymer composites parts.Previous literature showed characterization of mode I and II interlaminar fracture behaviors was developed using carbon fiber reinforced high performance matrixes such as epoxy resins (PEEK) [7][8][9].
Consequently, these testing protocols had been predominantly applied to the study of carbon fiber based materials.There were few investigations of the interlaminar fracture behaviors of other thermosetting composites which were reinforced by glass fiber.Until now, there are no studies on the interlaminar fracture behaviors of glass fiber mat reinforced polyester composites.
In this paper, mode Ι interlaminar fracture behaviors of the pure polyester resin and GF/PO composites by use of double cantilever beam (DCB) tests were studied and compared with that of a standard, unidirectional carbon fiber reinforced epoxy composite(CF/EP composites).The fracture area along crack propagation was analyzed by scanning electron microscope (SEM) to identify the fracture characteristics.

Materials
The material used in this study is a plate made of unsaturated polyester resin and 30 layers of glass-fiber mat (V f =30%, density 270 g/m²); fabricated by the TEN CATE ADVANCED COMPOSITE MATERIALS company.

Mode Ι interlaminar fracture test
Mode I interlaminar fracture tests were carried out Fig. 1 Mode I DCB test at 20°C according as ISO 15024 [10].Force-displacement data were obtained using an Instron Universal Testing Machine operating at a crosshead speed of 0.3 mm/min.Dimensions of a DCB specimen for glass fiber based composites were shown in Fig. 1.A Polytetrafluoroethylene (PTFE) film of thickness 13 m was inserted in the central layer at one end of the test specimens before the process, which was employed to create a initial crack from initial load.A pair of metallic hinges was glued to the loading end of the specimen in order to enable the load to be applieled (Fig. 1).Observation of the crack propagation during tests was facilitated by coating edges of the specimens with white acrylic paint and placing indicator marks along the coated edges.Marks were drawn at 5 mm intervals along edges of the specimen, extending at least 60 mm beyond the tip of the insert.Additionally, the first 10 mm and last 5mm were marked at 1 mm intervals.The crack propagation energy values were calculated using the corrected beam theory: where P is the load, δ is the displacement of the notch lip, the crack length and b is the specimen width.The corrected beam theory requires the determination of a correction factor Δ, which takes into account crack tip rotation and shear deformation.a is the total crack length (a = a 0 +Δa) (Δa is crack propagation length and a 0 is measured along the curved coordinate scale fixed to the specimen, as described by Fig. 2), N is a correction factor for the stiffening caused by the metallic hinge (described by Eq. ( 2)) [l 1 is the distance from the center of the loading pin to the midplane of the specimen; l 2 is the distance from the loading pin center to its edge of the load metallic hinge], F is a correction factor for large displacements described by Eq. ( 3).
Fig. 2 The DCB specimen under the load

Effect of V f (the volume of fiber) on the interlaminar fracture properties
Typical mode Ι interlaminar fracture data were shown in Figs. 3 and 4. Fig. 3 showed the forcedisplacement plots for GF/PO composites and CF/EP composites [11].The displacement of the GF/PO composite was much greater than that of CF/EP composite, which showed the high flexibility of the specimen and the need to account and correct for large displacement, F (described by Eq. ( 3)).

Fig. 3 displacement curve
The force-displacement plot of CF/EP composite was linear prior to crack initiation, after which the force dropped as the crack propagates.The sudden drop in force immediately after crack propagation was typical of brittle materials.Crack propagation in this type of composite was relatively stable.In contrast, the plot of GF/PO composites exhibited distinct curvature prior to initiation, which occurred at a much greater displacement.Crack propagation was unstable and there was evident repeated crack initiation and crack arrest.The force-displacement plots reflected the difference in toughness of the pure polyester resin compared with that of the epoxy resin.Mode Ι force-displacement plots were used to form delamination-resistance curves (R-curves) in Fig. 4, which showed the delamination resistance with delamination length (crack propagation length Δa).Fig. 4 showed that G Ic varied significantly during crack propagation, with values lying between the upper and lower bounds with a variation in G Ic of 1.0 KJ/m 2 , corresponding to crack initiation and crack arrest [12].In contrast, the R-curve for CF/EP composite was relatively flat and a single propagation value can be obtained, which also showed crack propagation in this type of the composite was relatively stable.
Values of G Ic determined from Mode Ι interlaminar tests were summarized in Table .Values of G Ic for GF/PO composites were significantly higher than that for the pure polyster resin, which were explained by the additional force required to fracture fiber bridging in GF/PO composites, also described by Moura and Morais, et al [13,14].In contrast, CF/EP composite was produced using individually stacked plies of unidirectional CF preimpregnated with the epoxy resin and had little interlaminar fiber bridging.Consequently, during DCB test, fiber bridging can be negligible and crack propagation along the midplane was relatively smooth, which was in accordance with.
To demonstrate the occurrence of fiber bridging, comparative DCB tests were carried out using a hybrid composite, in which two plies of unidirectional continuous fiber were incorporated into the central ply to displace the corresponding plies of GF/PO composites, also employed by Morais, et al [14].The total number of glass fiber plies for the hybrid composite was the same as that for GF/PO composites (V f is 30%).The central two unidirectional plies can significantly reduce interlaminar fiber penetration and eliminate fiber bridging during crack propagation.R-curves for GF/PO composites (V f is 30%) and the hybrid composite were shown in Fig. 5.There was markedly less variation in G Ic with crack length for the hybrid composite, the value of G Ic for the hybrid composite was about 0.7 KJ/m 2 , which was similar to that of the pure polyster resin in Table 1.It can be seen that the value of fracture toughness G Ic was slightly improved with the increase of the volume of fiber V f .Additionally, the value of fracture toughness G Ic was affected by the types of weave, fiber orientation, the types of the matrix, the thermal treatment and so forth [15].In mode Ι interlaminar fracture tests, different initial crack lengths a 0 were used in different literatures.We also found that a 0 affected values of G Ic in the experiments.Fig. 6 showed effects of a 0 on values of G Ic .When a 0 was Fig. 6 Typical R-curves of different initial crack length a 0 25 mm, values of G Ic varied markedly and were much dispersed.As a 0 increased, the degree of dispersing was reduced, which was in accordance with Jiao and Todo [16,17].Because the inserted film would influence the resin flow and form the rich area in the resin.Values of G Ic determined in this area may be invalid.Therefore, the initial crack length may not be too long.When a 0 was 40 mm, values of G Ic showed little variation.Therefore, 40 mm was chosen as initial crack length in this study.

Fractography
After mode Ι interlaminar fracture tests, the fracture areas of GF/PO composites were analyzed by SEM to identify the fracture characteristics and to verify the deformation behaviour predicted from the modelling.Figs.7  and 8 were SEM photographs of the fracture areas when the distance of crack propagation length Δa = a -a 0 was about 45 mm.It showed the surface of the fiber was not bare and clean, which indicated good interface bonding between the fiber and the matrix.The crack propagation can be transferred from the matrix to the interface between the fiber and the matrix.So the crack propagated along the interface between the fiber and the matrix, and finally was transferred to the fiber through the good interface [18][19][20][21].In general, the fiber can endure high applied load, therefore the fiber based composites can achieve high mechanical properties.Cook-Gordon Damage Theory mentioned values of G Ic for thermosetting materials would be mainly decided by the interface between the fiber and the matrix [3,11].As a result, GF/PO composites can achieve high mechanical properties because of good interface bonding.

Conclusions
Mode Ι interlaminar fracture behaviors by use of DCB tests were studied.G Ic varied significantly during crack propagation.The occurrence of fiber bridging led to significant variation of G Ic (1.1~2.3KJ.m -2 ) of GF/PO composites and higher G Ic than that of the pure polyster resin (0.65 KJ.m -2 ).When initial crack length was 40mm, G Ic showed little variation.Crack propagation can be transferred from the matrix to the interface between the fiber and the matrix, finally to the fiber, which indicated good interface bonding in GF/PO composites.

N. Mahmoudi EFFECT OF VOLUME FIBER AND CRACK LENGTH ON INTERLAMINAR FRACTURE PROPERTIES OF GLASS FIBER REINFORCED POLYESTER COMPOSITES (GF/PO COMPOSITES)
S u m m a r y Double-cantilever beam (DCB) specimens are widely used to measure the mode I critical-energy release rate G Ic of glass-fiber reinforced polyester composites (GF/PO composites).
It showed G Ic varied significantly during crack propagation.G Ic of GF/PO composites are higher than that of the pure polyester (PP) resin because of the additional force required to fracture fiber bridging in GF/PO composites.Comparative DCB tests were carried out to demonstrate the occurrence of fiber bridging.G Ic varied markedly and were much dispersed, which would induce little variation when initial crack length was 40 mm.The fracture characteristics were analyzed by scanning electron microscope.

Fig. 4 R
Fig. 4 R-curves derived from force displacement curves of FG/PO composites and CF/EP

Fig. 5
Fig. 5 Comparative R-curves of FG/PO composite and the hybrid composite 3.2.Effect of initial crack length on the interlaminar fracture properties

Fig. 7
Fig. 7 SEM photograph of the fracture area in the region of Δa is about 45 mm (×3000)

Table Values of
G IC for the pure polyester resin, GF/PO composites and CF/EP composite