- Email: [email protected]
epoxy laminates
COMPOSITES SCIENCE AND TECHNOLOGY Composites Science and Technology 66 (2006) 665–675 www.elsevier.com/locate/compscitech
Mode I delamination fatigue properties of interlayer-toughened CF/epoxy laminates Masaki Hojo a,*, Satoshi Matsuda b, Mototsugu Tanaka a, Shojiro Ochiai c, Atsushi Murakami b a
Department of Mechanical Engineering and Science, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan b Department of Mechanical System Engineering, University of Hyogo, Himeji 671-2201, Japan c International Innovation Center, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan Received 10 November 2004; accepted 18 July 2005 Available online 13 October 2005
Abstract Mode I delamination fatigue crack growth behavior was investigated in carbon fiber (CF)/epoxy laminates with two kinds of interlayer/interleaf. One was with heterogeneous interlayer with fine polyamide particles, T800H/3900-2, and the other was with interleaf of new types of thermoplastic resin, ionomer, UT500/111/ionomer. Tests were carried out using double cantilever beam specimens. For T800H/3900-2 laminates, the crack path shifted from the heterogeneous interlayer region (Stage I) to the interlayer/base lamina interface (Stage II) with the increase of the crack length. The delamination fatigue crack growth resistance of the T800H/3900-2 laminates was about 3 times higher than that of the reference CFRP for Stage I. This ratio decreased to 1.5 times for Stage II. For UT500/111/ionomer laminates, the high level of the fatigue crack growth resistance was kept without respect to the crack length. The threshold value of UT500/111/ionomer laminates was about 3 times higher than base UT500/111 laminates. The mechanism of the difference of the toughening mechanism for these interlayer/interleaf laminates was discussed on the bases of fractrographic observation and mechanism consideration. 2005 Elsevier Ltd. All rights reserved. Keywords: Polymer–matrix composites; Fatigue crack growth; Delamination; Interlaminar fracture; Interlayer; Interleaf
1. Introduction Although two decades have passed since the importance of the interlaminar fracture was recognized [1,2], interlaminar strength is still one of the design limiting factors in structural composite laminates [3]. A large number of research works have been carried out on static interlaminar properties of toughened laminates [4,5]. Current topics are the effects of through the thickness reinforcement such as stitching, Z-pin and three-dimensional fabric on the interlaminar strength [6–9]. However, only little has been reported on delamination properties of interlayer/inter-
*
Corresponding author. Tel.: +81 75 753 4836; fax: +81 75 771 7286. E-mail address: [email protected] (M. Hojo).
0266-3538/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.compscitech.2005.07.038
leaf-toughened laminates under fatigue loading [10–12]. Since new generation commercial aircraft such as A380 and B787 are reported to increase application of composite materials to structural parts, the understanding of fatigue properties is urgently required for interlayer/interleaftoughened laminates. One of the established ways to increase the interlaminar strength is to replace resin at prepreg interface to a tougher system [13–15]. This method is called as ‘‘interleaf’’ or ‘‘interlayer’’. A full commercial product, T800H/3900-2, has heterogeneous interlayer including fine thermoplastic particles, and has already been used in the primary structures of Boeing 777 [16,17]. Though the mode II interlaminar fracture toughness for T800H/3900-2 indicated excellent properties, the mode I fracture toughness decreased with the increment of the crack length [10]. We
666
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
have only carried out detailed researches on mode II delamination fatigue crack growth properties [11]. Here, the threshold value increased only 1.5 times, and this increase was much smaller than that of static interlaminar fracture toughness (2.4 times for initial and 3.9 times for propagation values) [10,11]. However, only little has been reported on the fatigue properties under mode I loading [10], and the mechanism of delamination fatigue crack growth is still not understood. The static results of T800H/3900-2 under mode I loading indicate necessity of new interleaf materials with higher ductility and adhesion strength to the base CFRP lamina [18]. A new type of thermoplastic resin, ionomer, has high ductility and good adhesion strength to epoxy resin [19]. In addition, ionomer forms the toughened interphase at the ionomer/base lamina interface [20]. We have already reported that both mode I and II fracture toughness of ionomer-interleaved CFRP laminates greatly increased from the base laminates [18,20]. Specially, the higher mode I fracture toughness did not decrease with the increment of the crack length [20]. The objective of this study is to clarify the effect of the heterogeneous interlayer and the ionomer interleaf on the mode I delamination fatigue crack growth behavior in CFRP laminates. The toughening mechanisms were discussed on the bases of mechanism consideration and the microscopic observation. 2. Experimental procedure 2.1. Materials and specimens For the case of laminates with heterogeneous interlayer, unidirectional laminates of nominal thickness of 3 mm, (0)16 were made from carbon fiber/epoxy prepregs of Toray T800H/3900-2 by using an autoclave. The details of the mechanical properties of the unidirectional T800H/3900-2 laminates are given in a separate paper [11]. Fig. 1(a) schematically shows the transverse section of T800H/3900-2 laminates. The thickness of the heterogeneous interlayer was about 30 lm. This interlayer with fine polyamide particles was inserted in every ply interface between the base
Carbon fiber
2.2. Fracture toughness test and fatigue test The energy release rate, G, was calculated using the modified compliance calibration method [23,24]. The stress intensity factor, K, was calculated using the relation between G and K as follows [25,26]: G ¼ HK 2 ;
Interlayer
Carbon fiber (CF)
T800H3900-2
Distribution of resin
} Epoxy Interphase of epoxy } (Mixture and ionomer) } Ionomer } Interphase } Epoxy
CF/epoxy
a
ð1Þ
where H is a function of elastic moduli (Eij, mij). The H value for T800H/3900-2 was calculated as 6.83 · 1011 Pa1.
30 m
Polyamide particle
CF/epoxy lamina. Starter slits were introduced into the laminates by inserting single 7.5 lm thick polyimide film at midplane during molding. For the case of ionomer-interleaved laminates, unidirectional laminates of nominal thickness of 3 mm, (0)24 were fabricated from carbon fiber/epoxy prepregs of Toho UT500/111 by using a hot press. The details of the laminates and ionomer are explained in a separate paper [20]. The ionomer film was inserted only at the midplane of the laminates during molding. The thickness of the ionomer film was 25 and 100 lm. These laminates are indicated as UT500/111/ionomer (25 or 100 lm). For reference, the base laminates without interleaf films, UT500/111, were also fabricated. In each laminate, a polyimide film of 13 lm thickness was inserted during molding at the midplane as a starter slit. Fig. 1(b) depicts the schematic structure of the transverse section near the interfaces. It is known that there is an interphase region of one- or two-carbon fiber thickness next to the ionomer region, where epoxy and ionomer are mixed [20]. Double cantilever beam (DCB) specimens (width B = 20 mm, length L = 140 mm, and nominal thickness 2h = 3 mm) were used for the tests of both laminates under static and fatigue loadings. Fig. 2 shows the DCB specimen and aluminum blocks for load introduction. Special loading apparatus with universal joints were used for tests [21,22]. The side surfaces of the specimens were polished with abrasive papers and diamond paste (1 and 6 lm). Then, white brittle paint was coated on both side surfaces of the test specimen to enhance the crack length measurement. The length of the initial crack was about 20–25 mm.
b
UT500/111/ionomer
Fig. 1. Schematics of transverse section of laminates.
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
667
Polyimide film
20
10
5
3
10
Aluminum block
5
Fiber direction
5 20-25 140
Fig. 2. DCB specimen with aluminum blocks (dimensions are in mm).
3. Results and discussion 3.1. T800H/3900-2 laminates 3.1.1. Interlaminar fracture toughness The open marks in Fig. 3 show the relation between the fracture toughness and the increment of the crack length for T800H/3900-2 laminates. Similar to the former report [10], the fracture toughness decreased with the increase of the crack length, and then leveled off where the increment of the crack length, Da, was larger than 35–40 mm. These unique R-curves give the initial values of the fracture toughness, GIc as 0.71 kJ/m2, and the saturated propagation value, GIR, as about 0.28 kJ/m2. The solid marks in
0.8
Fracture toughness, GIc, GIR (kJ/m2)
Those for UT500/111/ionomer and UT 500/111 were calculated as 5.94 · 1011 and 5.83 · 1011 Pa1, respectively. The tests were carried out in a computer-controlled servohydraulic testing system (Shimazu 4880, 9.8 kN) with original software [21,25]. Load cell of 490 N capacity was attached to the testing machine. Mode I interlaminar fracture toughness test was first carried out for all laminates to obtain basic static properties. The crosshead speed was controlled to be 0.5–1 mm/min [24]. For fatigue tests, the stress ratio, R, of the minimum load to the maximum load was kept constant to be 0.1 and 0.5 for T800H/3900-2 laminates. That for UT500/ 111/ionomer (25 lm) and UT 500/111 laminates was kept constant to be 0.2 and 0.5. The load-shedding tests were carried out for all laminates. The maximum crack opening displacement was kept constant for T800H/3900-2 laminates because the change of the crack growth rate was not simple. Load shedding was controlled by the attached computer for UT500/111/ionomer and UT 500/111 laminates. The normalized gradient of the energy release rate (1/G) dG/da was kept under 0.1 mm1. The frequency of the stress cycle was 10 Hz. The crack length during fatigue tests was computed from the measurement of the compliance [21,25].
T800H/3900-2 T800H/3631
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
10 20 30 40 50 Increment of crack length, a (mm)
60
Fig. 3. Relation between fracture toughness and increment of crack length for T800H/3900-2 laminates.
Fig. 3 indicate the results of T800H/3631 laminates as reference base laminates. The GIc value was 0.18 kJ/m2 and the GIR value at Da = 50 mm was 0.25 kJ/m2. Then, although the initial values, GIc, of T800H/3900-2 were about four times higher than that of reference T800H/ 3631, the propagation value, GIR, was almost identical for both laminates. 3.1.2. Fatigue delamination The results of the fatigue tests for T800H/3900-2 also show unique behavior. Fig. 4 shows the change of the crack propagation rate, da/dN, with the increment of the crack length for the load-shedding test under R = 0.5. Though the applied load and the corresponding energy release rate decreased with the increment of the crack length, da/dN first decreased till Da = 6 mm, then increased between Da = 6 and 8 mm, and decreased again after Da = 8 mm. Fig. 5 compares the macroscopic fracture surface under static and fatigue loadings. The fracture surface was divided into white and black parts. Fig. 6 shows the corre-
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
Crack propagation rate, da/dN (m/cycle)
668
10
-6
10
-7
10
-8
10
-9
T800H/3900-2 (R=0.5)
Fig. 5. Comparison of macroscopic fracture surfaces under static and fatigue loadings for T800H/3900-2 laminates. -10
10
0
2 4 6 8 10 Increment of crack length, a (mm)
12
Fig. 4. Change of crack propagation rate with increment of crack length for T800H/3900-2 laminates.
sponding microscopic fracture surfaces using SEM (scanning electron microscope). Higher ductility of resin and roughness were observed in the white parts. On the other hand, typical interlaminar fracture surfaces with carbon fiber and brittle epoxy resin were observed in the black parts. Then, the unique crack growth behavior of T800H/3900-2 laminate in Figs. 3 and 4 is attributed to the transition of the crack path from the toughened interlayer region (Stage I) to the untoughened interlayer/base lamina interface
(Stage II). It is interesting to note that this transition under fatigue loading was completed only after Da = 8 mm in Figs. 4 and 5. Thus, the transition from Stage I to II occurred for much shorter Da under fatigue loading than that under static loading. Fig. 7 compares the optical micrographs of the side surface of the specimen under static and fatigue loadings. While the crack path under fatigue loading was rather straight passing through polyamide particles, this path under fatigue loading often went around polyamide particles. The smaller cyclic damage zone is probably responsible for this difference. Then, the driving force for the stage transition is suggested to be larger under fatigue loading. This can be the reason for the shorter transition length under fatigue loading.
Fig. 6. Scanning electron micrographs of fracture surfaces for T800H/3900-2 laminates. (a-1) Static, white part (Da = 0.1 mm), (a-2) static, black part (Da = 9.3 mm), (b-1) fatigue, white part (Da = 0.4 mm), (b-2) fatigue, black part (Da = 6.1 mm).
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
669
Fig. 7. Optical micrographs of side surface of specimen at interlayer region for T800H3900-2 laminates: (a) static fracture, (b) fatigue fracture.
Fig. 8. Relation between crack propagation rate and maximum energy release rate under R = 0.1 for T800H/3900-2 laminate.
Fig. 9. Relation between crack propagation rate and maximum energy release rate under R = 0.5 for T800H/3900-2 laminate.
Figs. 8 and 9 show the relation between da/dN and the maximum energy release rate, Gmax, which corresponding to the maximum load, under R = 0.1 and 0.5. All of the fatigue tests were started from the end of the starter slit without introducing the precracks, and different marks were used for each specimen. As was suggested in Fig. 3, the crack growth behavior was very complicating. The specimen indicated by open circles in Fig. 9 corresponds to the same specimen for Fig. 3. The data of this specimen suggest that there exist two power-law relations for Stage I and II. Since almost all specimens indicate the transition from Stage I to II from the beginning of the fatigue tests, the dd/dN just after starting the tests were regarded to represent the data for Stage I. Finally, the power-law relation for Stage I is shown by solid lines in Figs. 8 and 9.
It is also difficult to get the clear power-law relations for Stage II. SEM observation at lower magnification indicated bridged fibers. This brings apparent increase of the fatigue crack growth resistance for Stage II, and da/dN– Gmax relation deviates from the true Stage II relation [27,28]. SEM observation indicated that the fiber bridging started just after the transition from Stage I to II was finished. Since this transition gradually grows in the width direction, it is very difficult to define the true Stage II relation without the fiber bridging effect. Then, we tentatively defined the Stage II relation from the data just after the transition to Stage II was finished throughout the width direction. These points are indicated by solid symbols, and the power-law relations are shown by the dashed lines in Figs. 8 and 9. The power-law relation for the
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
DK eq ¼ DK 1c K cmax ; where the stress-ratio-effect parameter, c (0 < c < 1), indicates the relative contribution of the maximum load to the cyclic load in determining the crack propagation rate. These c values for Stage I and II are 0.30 and 0.50, respectively. The c value for Stage I is similar to that for ductile AS4/PEEK (0.27) [30], and that for Stage II is comparable to that for T800H/3631 (0.46) [29]. These facts agree well with the crack paths for the Stage I and II. In Stage I, the polyamide particles give much higher ductility as shown in Fig. 6, and the mode I fatigue fracture mechanism is controlled by the cyclic load similar to that of the AS4/ PEEK where the matrix is ductile thermoplastic. In Stage II, the fracture surface is similar to normal CF/epoxy laminates, and the rather brittle epoxy is responsible for the higher contribution of the maximum load. 3.2. UT500/111/ionomer laminates 3.2.1. Interlaminar fracture toughness Static fracture toughness tests were carried out for UT500/111/ionomer (25 and 100 lm) and the base UT500/111 laminates to obtain the base properties using the same laminates as those for fatigue tests. Fig. 10 shows the relation between the mode I fracture toughness and the increment of crack length (R-curves) for all laminates. Here, the data points at Da = 0 indicate GIc values at the NL points. The GI values for the initial stage of the crack propagation sharply increased by about 2–3 times where Da is smaller than 5 mm for all laminates. Where Da is longer than 10 mm, GIR values leveled off though GIR values indicate scatter for all laminates. Whereas GIc and GIR values of UT500/111/ionomer (25 lm) laminates were 3.2
UT500/111/Ionomer
2
Mode I
2
Mode I fracture toughness, GIc , GIR (kJ/m )
reference laminate, T800H/3631, is also indicated in Fig. 9 [29]. Since the crack growth behavior is so complicated with the stage transition, it is rather difficult to compare the fatigue crack growth resistance by comparing the threshold values. Then, the growth resistance was compared at da/dN = 109 m/cycle. Fig. 9 indicates that the growth resistance of T800H/3900-2 (Stage I and Stage II) is 3.0 and 1.6 times higher than that of T800H/3631, respectively. This increase in Stage I is slightly lower than the increase of static GIc (4 times higher). Though the Stage I resistance is 0.64 times of that of AS4/PEEK with thermoplastic matrix [30], these results assure higher resistance than the reference T800H/3631 laminates even in Stage II, and indicate substantial increase in Stage I. It is interesting to note that the threshold values of T800H/3900-2 under mode II loading increased only about 1.5 times from that of the reference T800H/3631. This increase was much smaller than the increase of GIIc (2.4 times) and GIIR (3.9 times) [11]. We are proposing the following equivalent stress intensity range, DKeq, as a controlling parameters for fatigue crack growth of composite laminates under carious R values [21,25]:
1.5
1
0.5 Ionomer 100µm Ionomer 25µm Base
0
0
10 20 30 Increment of crack length,
40
Fig. 10. Relation between fracture toughness and increment of crack length for base UT500/111 and UT500/111/ionomer.
times and 3.5 times higher than those of the base laminates, GIc and GIR of UT500/111/ionomer (100 lm) laminates were 3.9 times and 5.5 times higher than those of the base UT500/111 laminates. 3.2.2. Fatigue delamination For the fatigue tests of UT500/111/ionomer (25 lm) laminates, the GImax-constant test [27] was first carried out, and the results were shown in Fig. 11. The change of the crack propagation rate is within 3 times, and the global change of the crack propagation rate is not observed. This suggests that the transition of the crack path which is observed in T800H/3900-2 is not expected for UT500/ 111/ionomer (25 lm) laminates. Fig. 12 shows the relation between the crack propagation rate and the maximum energy release rate in the GImaxdecreasing tests for UT500/111/ionomer (25 lm) and the
Crack propagation rate, da/dN (m/cycle)
670
10
-7
10
-8
10
-9
UT500/111/ionomer
Mode I
G
10
-10
10
-11
2
Imax
=250 (J/m )
R=0.5
0
2
4
6
8
10
12
14
16
Increment of crack length, Fig. 11. Change of crack propagation rate with increment of crack length for UT500/111/ionomer laminates.
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
671
Fig. 12. Relation between crack propagation rate and maximum energy release rate for base UT500/111 and UT500/111/ionomer laminates.
base UT500/111 laminates. As was expected from Fig. 11, normal power-law relations were observed even for the ionomer-interleaved laminates under R = 0.2 and 0.5 da/dN deviates from the power-law relations where da/dN is lower than 109 m/cycle, and there existed the growth threshold. The threshold value of the maximum energy release rate, GImax, for UT500/111/ionomer (25 lm) laminates (120 J/m2 under R = 0.2, 180 J/m2 under R = 0.5) is 3.5–3.3 times higher than those for the base laminates (34 J/m2 under R = 0.2, 54 J/m2 under R = 0.5). This ratio is similar to that for T800H/3900-2 under fatigue loading in Stage I. Moreover, this ratio is almost the same as that of GIR for UT500/111/ionomer (25 lm) laminates. The great advantage for this ionomer-interleaved laminates is no transition to the weaker stage. The exponents of the power function for UT500/111/ionomer (25 lm) laminates (16 under R = 0.5) were about two times larger than those for the base laminates (8 under R = 0.5). It has been reported that the relative increase of the fatigue crack growth resistance from the base laminates to CFRP laminates with thermoplastic matrix such as the AS4/PEEK was lower than that of the fracture toughness [21,31]. Table 1 shows the comparison of the ratio of the fatigue crack growth resistance to the fracture toughness, GImaxth/GIc, among T800H/3900-2, UT500/111, UT500/ 111/ionomer and AS4/PEEK laminates [30]. The GImaxth/ GIc values for T800H/3900-2, UT500/111, UT500/111/ ionomer and AS4/PEEK laminates were 0.35, 0.34, 0.36
and 0.25, respectively. Then, the GImaxth/GIc values for the interlayer or interleave laminates are higher than that of laminates with thermoplastic matrix. Stress-ratio effect parameter, c, was calculated for each crack propagation rate using Eq. (1) to discuss the stress ratio dependency on the fatigue crack growth. Fig. 13 shows the results. The c values for both laminates (0.5– 0.7) were relatively high, which indicates that the contribumode I
1 UT500/111/ionomer(25 m) UT500/111(base) 0.8
0.6
0.4
0.2
0
-9
-8
-7
10 10 10 Crack propagation rate, da/dN (m/cycle)
Fig. 13. Stress-ratio effect parameter for base UT500/111 and UT500/111/ ionomer laminates.
Table 1 Comparison of delamination threshold to fracture toughness Laminates
T800H/3900-2
Base UT500/111
UT500/111/ionomer (25 lm)
AS4/PEEK
R = 0.5 GImaxth(J/m2) GIc (J/m2) GImaxth/GIc
250 710 0.35
54 160 0.34
180 505 0.36
300 1200 0.25
672
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
Fig. 14. Lower magnification SEMs of fracture surfaces for UT500/111/ionomer laminates. (a-1) Static fracture, (a-2) static fracture, (b-1) fatigue, da/ dN = 107 m/cycle, base lamina side, (b-2) fatigue, da/dN = 107 m/cycle, ionomer side.
tion of maximum stress is large in the fatigue fracture behavior under mode I loading for both laminates [31]. 3.2.3. Microscopic observation and mechanism consideration Fig. 14 presents the lower magnification SEMs of the fracture surfaces of the UT500/111/ionomer (25 lm) laminates under static and fatigue loadings, respectively. The ionomer region was left like small ‘‘islands’’ on the fracture surfaces, which indicates the existence of ionomer bridging [32]. Traces of pull-out fibers are observed at the ionomer region, and the fibers are observed at the rest of the fracture surface. This indicates that the crack path is ionomer interleaf/base lamina interface. Fig. 15 shows the SEMs of higher magnification for UT500/111/ionomer (25 lm) and the base UT500/111 laminates. In this study, the fatigue fracture surfaces of both laminates under R = 0.2 and R = 0.5 were investigated. No significant effect of the stress ratio, R, was observed on the fatigue fracture surfaces for each laminates. Thus, the fatigue fracture surfaces under R = 0.5 were chosen for the discussion. For both laminates, little difference was observed under static and fatigue fracture. Moreover, the effect of the growth rate under fatigue loading was also minimal. For the base UT500/111 laminates, fiber/matrix interfacial fracture was dominant. The areal ratio of the interfacial fracture was about 70%. The remaining resin part indicates rather brittle fracture.
The areal ratio of the interfacial fracture for UT500/111/ ionomer (25 lm) laminates was also about 70% both under static and fatigue loadings. For the Stage I of T800H/39002, large drawing of resin was only observed under static loading, and there was a substantial difference between the fracture surfaces of static and fatigue fracture. Similar results were obtained for AS4/PEEK with thermoplastic matrix (Fig. 6). However, no significant difference was observed for this ionomer-interleaved laminates under static and fatigue loadings. This means that the microscopic fracture mechanisms of the resin under static and fatigue loadings are rather similar. Large drawing of resin, which was observed only under static loading for T800H/3900-2 (Stage I) and AS4/PEEK [21], was observed both under static and fatigue loadings. Since the fracture path was at ionomer interleave/base lamina interface, the resin consists of ionomer and epoxy, and this resin indicates higher ductility than base matrix epoxy resin. The similarity of the fracture surfaces under static and fatigue loadings agrees with the higher c values in Fig. 13 (c = 0.5–0.7), which indicates that the contribution of maximum load is higher than that of cyclic load in delamination fatigue fracture. The higher ratio of interfacial fracture also contributes to this tendency. Section 3.1 indicates that the fracture toughness and the fatigue crack growth resistance under mode I loading de-
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
673
Fig. 15. SEMs of fracture surfaces for base UT500/111 and UT500/111/ionomer laminates (continued). (a-1) Static fracture, base UT500111, (a-2) static fracture, UT500111/ionomer, (b-1) fatigue, da/dN = 107 m/cycle, base UT500/111, (b-2) fatigue, da/dN = 107 m/cycle, UT500/111/ionomer, (c-1) fatigue, da/dN = 1010 m/cycle, base UT500/111, (c-2) fatigue, da/dN = 1010 m/cycle, UT500/111/ionomer.
creases with the increment of crack length for the interlayer-toughened T800H/3900-2 laminates. The toughened region is usually limited within the resin-rich layer between the base lamina for the conventional interleaved laminates and the interlayer-toughened laminates. Thus, the crack path easily deviates from the toughened region to the untoughened region because the crack path can only be arrested by fibers (Fig. 16(a)). On the contrary, the high fracture toughness and the high fatigue crack growth resistance under mode I loading were maintained with the increment of crack length for the ionomer-interleaved laminates. It is known that the toughened region invades the base lamina
with forming the interphase for the ionomer-interleaved laminates (Fig. 1(b)). Thus, the crack path is always inside the toughened region for the ionomer-interleaved laminates (Fig. 16(b)) [20]. This is the reason why the high fracture toughness and the high fatigue crack growth resistance under mode I loading were maintained with the increment of crack length for the ionomer-interleaved laminates. Only one point we should note for this ionomer-interleaved laminate is that only this concept can be used for the future structures and the real application is still the subject of the future researches because the glass transition temperature of ionomer is rather low.
674
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
Fig. 16. Schematic models of crack paths under mode I loading: (a) interlayer-toughened laminates, (b) ionomer-interleaved laminates.
4. Conclusions
Acknowledgment
Mode I delamination fatigue crack growth properties of two types of interlayer/interleaf CFRP laminates were investigated in the present study. The results are summarized as follows: The laminates with heterogeneous interlayer with fine polyamide particles, T800H/3900-2, indicated the transition of the crack path from the toughened interlayer region (Stage I) to the untoughened interlayer/base lamina interface (Stage II). The corresponding relation between the crack growth rate and the maximum energy release rate also showed transition of the power-law relation from Stage I to II. The crack growth resistance for toughened Stage I was 3.0 times higher than the reference laminates. Though the growth resistance decreased in Stage II, it was still 1.6 times higher than that of the reference laminates. This transition occured for much shorter increment of the crack length than that under static loading. The laminates with a new thermoplastic resin (ionomer) interleaf, UT500/111/ionomer, keep higher crack growth resistance without respect to the increment of the crack length under fatigue loading. The threshold value was 3.3–3.5 times higher than the base laminates. This increase was similar to that of the propagation values of the static fracture toughness. Though the crack path was at the interleaf/base lamina interface, the resin of this region consists of ionomer and epoxy resin, forming toughened interphase. Thus, the crack path was still within the toughened region without respect to the crack length.
Authors thank Prof. T. Aoki of University of Tokyo for his heplful discussion, and Mr. H. Ito. of All Nippon Airways and T. Ando of Graduate Student of Kyoto University for carrying out experiments. Authors also thank Dr. A. Kitano ot the Tray Industries and Mr. M. Ando of Toho Tenax for supplying laminates. References [1] OÕBrien TK. Characterization of delamination onset and growth in a composite laminate. In: Reifsnider KL, editor. Damage in composite materials, ASTM STP 775; 1982. p. 140–67. [2] Standard tests for toughened resin composites. NASA RP 1092; 1983. [3] Paul K, Kelly L, Venkayya V, Hess T. Evolution of U.S. military aircraft structures technology. J Aircraft 2002;39(1):18–29. [4] Friedrich K. Application of fracture mechanics to composite materials. Amsterdam: Elsevier; 1989. [5] Sela N, Ishai O. Interlaminar fracture toughness and toughening of laminated composite materials: a review. Composites 1989;20(5): 423–35. [6] Jain LK, Mai Y-W. On the effect of stitching on mode I delamination toughness of laminated composites. Compos Sci Technol 1994;51: 331–45. [7] Yan W, Liu H-Y, Mai Y-W. Numerical study on the mode I delamination toughness of Z-pinned laminates. Compos Sci Technol 2003;63:1481–93. [8] Kamiya R, Chou TW. Strength and failure behavior of stitched carbon/epoxy composites. Metall Mater Trans A 2000;31:899– 909. [9] Mouritz AP, Bannister MK, Falzon PJ, Leong KH. Review of applications for advanced three-dimensional fibre textile composites. Composites A 1999;30:1445–61.
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675 [10] Kageyama K, Kimpara I, Ohsawa I, Hojo M, Kabashima K. Mode I and mode II delamination growth of interlayer toughened carbon/ epoxy (T800H/3900-2) composite system. In: Martin RH, editor. Composite materials: fatigue and fracture – fifth volume, ASTM STP 1230; 1995. p. 19–37. [11] Matsuda S, Hojo M, Ochiai S. Mesoscopic fracture mechanism of mode II delamination fatigue crack propagation in interlayertoughened CFRP. JSME Int J Ser A 1997;40(4):423–9. [12] Ogihara S, Takeda N, Kobayashi S, Kobayashi A. Effects of stacking sequence on microscopic fatigue damage development in quasiisotropic CFRP laminates with interlaminar-toughened layers. Compos Sci Technol 1999;59:1387–98. [13] Bradley WL. Relationship of matrix toughness to interlaminar fracture toughness. In: Friedrich K, editor. Application of fracture mechanics to composite materials. Amsterdam: Elsevier; 1989. p. 159–87. [14] Sela N, Ishai O, Banks-Sills L. The effect of adhesive thickness on interlaminar fracture toughness of interleaved laminates specimens. Composites 1989;20(3):257–64. [15] Singh S, Partridge IK. Mixed mode fracture in an interleaved carbonfibre/epoxy composite. Compos Sci Technol 1995;55:319–27. [16] Odagiri N, Kishi H, Nakae T. T800H/3900-2 toughened epoxy prepreg system: toughening concept and mechanism. In: Proceedings of the American society for composites, 6th technical conference; 1991. p. 43–52. [17] Odagiri N, Kishi H, Yamashita M. Development of TORAYCA prepreg P2302 carbon fiber reinforced plastic for aircraft primary structural materials. Adv Compos Mater 1996;5:249–52. [18] Hojo M, Matsuda S, Ochiai S, Murakami A, Akimoto H. The role of interleaf/base lamina interphase in toughening mechanism of interleaf-toughened CFRP. In: Proceedings of the twelfth international conference on composite materials; 1999 [CDROM]. [19] Akimoto H, Kanazawa T, Yamada M, Murakami A. Impact behaviors of ethylene based ionomers in tensile and penetration mode. In: Proceedings of the fifth Japan international SAMPE symposium; 1997. p. 67–72. [20] Matsuda S, Hojo M, Ochiai S, Murakami A, Akimoto H, Ando M. Effect of ionomer thickness on mode I interlaminar fracture toughness for ionomer toughened CFRP. Composites A 1999;30:1311–9.
675
[21] Hojo M, Ochiai S, Gustafson C-G, Tanaka K. Effect of matrix resin on delamination fatigue crack growth in CFRP laminates. Eng Fract Mech 1994;49(1):45–7. [22] Hojo M, Matsuda S, Fiedler B, Kawada T, Moriya K, Ochiai S, et al. Mode I and II delamination fatigue crack growth behavior of alumina fiber/epoxy laminates in liquid nitrogen. Int J Fatigue 2002;24:109–18. [23] Hojo M, Tanaka K, Kageyama K. Pre-standardization study on mode I interlaminar fracture toughness test for CFRP in Japan. Composites 1995;26(9):243–55. [24] JIS K 7086-1993. Testing methods for interlaminar fracture toughness of carbon fiber reinforced plastics. Japanese Standards Association, Tokyo; 1993. [25] Hojo M, Tanaka K, Gustafson C-G, Hayashi R. Effect of stress ratio on near-threshold propagation of delamination fatigue cracks in unidirectional CFRP. Compos Sci Technol 1987;29:273–92. [26] Sih GC, Paris PC, Irwin GR. On cracks in rectilinearly anisotropic bodies. Int J Fracture Mech 1965;1:189–203. [27] Hojo M, Ochiai S, Aoki T, Ito H. New simple and practical test method for interlaminar fatigue threshold in CFRP laminates. In: Proceedings of European conference on composites testing and standardization (ECCM-CTS 2); 1994. p. 553–61. [28] Martin RH, Murri GB. Characterization of mode I and mode II delamination growth and thresholds in AS4/PEEK composites. In: Garbo SP, editor. Composite materials: Testing and design. 9th vol., ASTM STP 1059; 1990. p. 251–70. [29] Tanaka K, Tanaka H, Tsuji T, Yamagishi K. Effect of stress ratio on mode I propagation of interlaminar fatigue cracks in CFRP. J Soc Mater Sci Jpn 1995;44:960–6. [30] Hojo M, Ochiai S, Aoki T, Ito H. Mode I fatigue delamination for CF/PEEK laminates using maximum-energy-release-rate constant tests. J Soc Mater Sci Jpn 1995;44:953–9. [31] Hojo M, Matsuda S, Ochiai S. Delamination fatigue crack growth in CFRP laminates under mode I and II loadings – effect of mesoscopic structure on fracture mechanism. In: Proceedings of international conference on fatigue of composites; 1997. p. 15– 26. [32] Matsuda S, Hojo M, Ochiai S, Moriya K, Murakami A, Usui Y, et al. J Soc Mater Sci Jpn 1998;47:1125–30.
Mode I delamination fatigue properties of interlayer-toughened CF/epoxy laminates Masaki Hojo a,*, Satoshi Matsuda b, Mototsugu Tanaka a, Shojiro Ochiai c, Atsushi Murakami b a
Department of Mechanical Engineering and Science, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan b Department of Mechanical System Engineering, University of Hyogo, Himeji 671-2201, Japan c International Innovation Center, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan Received 10 November 2004; accepted 18 July 2005 Available online 13 October 2005
Abstract Mode I delamination fatigue crack growth behavior was investigated in carbon fiber (CF)/epoxy laminates with two kinds of interlayer/interleaf. One was with heterogeneous interlayer with fine polyamide particles, T800H/3900-2, and the other was with interleaf of new types of thermoplastic resin, ionomer, UT500/111/ionomer. Tests were carried out using double cantilever beam specimens. For T800H/3900-2 laminates, the crack path shifted from the heterogeneous interlayer region (Stage I) to the interlayer/base lamina interface (Stage II) with the increase of the crack length. The delamination fatigue crack growth resistance of the T800H/3900-2 laminates was about 3 times higher than that of the reference CFRP for Stage I. This ratio decreased to 1.5 times for Stage II. For UT500/111/ionomer laminates, the high level of the fatigue crack growth resistance was kept without respect to the crack length. The threshold value of UT500/111/ionomer laminates was about 3 times higher than base UT500/111 laminates. The mechanism of the difference of the toughening mechanism for these interlayer/interleaf laminates was discussed on the bases of fractrographic observation and mechanism consideration. 2005 Elsevier Ltd. All rights reserved. Keywords: Polymer–matrix composites; Fatigue crack growth; Delamination; Interlaminar fracture; Interlayer; Interleaf
1. Introduction Although two decades have passed since the importance of the interlaminar fracture was recognized [1,2], interlaminar strength is still one of the design limiting factors in structural composite laminates [3]. A large number of research works have been carried out on static interlaminar properties of toughened laminates [4,5]. Current topics are the effects of through the thickness reinforcement such as stitching, Z-pin and three-dimensional fabric on the interlaminar strength [6–9]. However, only little has been reported on delamination properties of interlayer/inter-
*
Corresponding author. Tel.: +81 75 753 4836; fax: +81 75 771 7286. E-mail address: [email protected] (M. Hojo).
0266-3538/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.compscitech.2005.07.038
leaf-toughened laminates under fatigue loading [10–12]. Since new generation commercial aircraft such as A380 and B787 are reported to increase application of composite materials to structural parts, the understanding of fatigue properties is urgently required for interlayer/interleaftoughened laminates. One of the established ways to increase the interlaminar strength is to replace resin at prepreg interface to a tougher system [13–15]. This method is called as ‘‘interleaf’’ or ‘‘interlayer’’. A full commercial product, T800H/3900-2, has heterogeneous interlayer including fine thermoplastic particles, and has already been used in the primary structures of Boeing 777 [16,17]. Though the mode II interlaminar fracture toughness for T800H/3900-2 indicated excellent properties, the mode I fracture toughness decreased with the increment of the crack length [10]. We
666
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
have only carried out detailed researches on mode II delamination fatigue crack growth properties [11]. Here, the threshold value increased only 1.5 times, and this increase was much smaller than that of static interlaminar fracture toughness (2.4 times for initial and 3.9 times for propagation values) [10,11]. However, only little has been reported on the fatigue properties under mode I loading [10], and the mechanism of delamination fatigue crack growth is still not understood. The static results of T800H/3900-2 under mode I loading indicate necessity of new interleaf materials with higher ductility and adhesion strength to the base CFRP lamina [18]. A new type of thermoplastic resin, ionomer, has high ductility and good adhesion strength to epoxy resin [19]. In addition, ionomer forms the toughened interphase at the ionomer/base lamina interface [20]. We have already reported that both mode I and II fracture toughness of ionomer-interleaved CFRP laminates greatly increased from the base laminates [18,20]. Specially, the higher mode I fracture toughness did not decrease with the increment of the crack length [20]. The objective of this study is to clarify the effect of the heterogeneous interlayer and the ionomer interleaf on the mode I delamination fatigue crack growth behavior in CFRP laminates. The toughening mechanisms were discussed on the bases of mechanism consideration and the microscopic observation. 2. Experimental procedure 2.1. Materials and specimens For the case of laminates with heterogeneous interlayer, unidirectional laminates of nominal thickness of 3 mm, (0)16 were made from carbon fiber/epoxy prepregs of Toray T800H/3900-2 by using an autoclave. The details of the mechanical properties of the unidirectional T800H/3900-2 laminates are given in a separate paper [11]. Fig. 1(a) schematically shows the transverse section of T800H/3900-2 laminates. The thickness of the heterogeneous interlayer was about 30 lm. This interlayer with fine polyamide particles was inserted in every ply interface between the base
Carbon fiber
2.2. Fracture toughness test and fatigue test The energy release rate, G, was calculated using the modified compliance calibration method [23,24]. The stress intensity factor, K, was calculated using the relation between G and K as follows [25,26]: G ¼ HK 2 ;
Interlayer
Carbon fiber (CF)
T800H3900-2
Distribution of resin
} Epoxy Interphase of epoxy } (Mixture and ionomer) } Ionomer } Interphase } Epoxy
CF/epoxy
a
ð1Þ
where H is a function of elastic moduli (Eij, mij). The H value for T800H/3900-2 was calculated as 6.83 · 1011 Pa1.
30 m
Polyamide particle
CF/epoxy lamina. Starter slits were introduced into the laminates by inserting single 7.5 lm thick polyimide film at midplane during molding. For the case of ionomer-interleaved laminates, unidirectional laminates of nominal thickness of 3 mm, (0)24 were fabricated from carbon fiber/epoxy prepregs of Toho UT500/111 by using a hot press. The details of the laminates and ionomer are explained in a separate paper [20]. The ionomer film was inserted only at the midplane of the laminates during molding. The thickness of the ionomer film was 25 and 100 lm. These laminates are indicated as UT500/111/ionomer (25 or 100 lm). For reference, the base laminates without interleaf films, UT500/111, were also fabricated. In each laminate, a polyimide film of 13 lm thickness was inserted during molding at the midplane as a starter slit. Fig. 1(b) depicts the schematic structure of the transverse section near the interfaces. It is known that there is an interphase region of one- or two-carbon fiber thickness next to the ionomer region, where epoxy and ionomer are mixed [20]. Double cantilever beam (DCB) specimens (width B = 20 mm, length L = 140 mm, and nominal thickness 2h = 3 mm) were used for the tests of both laminates under static and fatigue loadings. Fig. 2 shows the DCB specimen and aluminum blocks for load introduction. Special loading apparatus with universal joints were used for tests [21,22]. The side surfaces of the specimens were polished with abrasive papers and diamond paste (1 and 6 lm). Then, white brittle paint was coated on both side surfaces of the test specimen to enhance the crack length measurement. The length of the initial crack was about 20–25 mm.
b
UT500/111/ionomer
Fig. 1. Schematics of transverse section of laminates.
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
667
Polyimide film
20
10
5
3
10
Aluminum block
5
Fiber direction
5 20-25 140
Fig. 2. DCB specimen with aluminum blocks (dimensions are in mm).
3. Results and discussion 3.1. T800H/3900-2 laminates 3.1.1. Interlaminar fracture toughness The open marks in Fig. 3 show the relation between the fracture toughness and the increment of the crack length for T800H/3900-2 laminates. Similar to the former report [10], the fracture toughness decreased with the increase of the crack length, and then leveled off where the increment of the crack length, Da, was larger than 35–40 mm. These unique R-curves give the initial values of the fracture toughness, GIc as 0.71 kJ/m2, and the saturated propagation value, GIR, as about 0.28 kJ/m2. The solid marks in
0.8
Fracture toughness, GIc, GIR (kJ/m2)
Those for UT500/111/ionomer and UT 500/111 were calculated as 5.94 · 1011 and 5.83 · 1011 Pa1, respectively. The tests were carried out in a computer-controlled servohydraulic testing system (Shimazu 4880, 9.8 kN) with original software [21,25]. Load cell of 490 N capacity was attached to the testing machine. Mode I interlaminar fracture toughness test was first carried out for all laminates to obtain basic static properties. The crosshead speed was controlled to be 0.5–1 mm/min [24]. For fatigue tests, the stress ratio, R, of the minimum load to the maximum load was kept constant to be 0.1 and 0.5 for T800H/3900-2 laminates. That for UT500/ 111/ionomer (25 lm) and UT 500/111 laminates was kept constant to be 0.2 and 0.5. The load-shedding tests were carried out for all laminates. The maximum crack opening displacement was kept constant for T800H/3900-2 laminates because the change of the crack growth rate was not simple. Load shedding was controlled by the attached computer for UT500/111/ionomer and UT 500/111 laminates. The normalized gradient of the energy release rate (1/G) dG/da was kept under 0.1 mm1. The frequency of the stress cycle was 10 Hz. The crack length during fatigue tests was computed from the measurement of the compliance [21,25].
T800H/3900-2 T800H/3631
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
10 20 30 40 50 Increment of crack length, a (mm)
60
Fig. 3. Relation between fracture toughness and increment of crack length for T800H/3900-2 laminates.
Fig. 3 indicate the results of T800H/3631 laminates as reference base laminates. The GIc value was 0.18 kJ/m2 and the GIR value at Da = 50 mm was 0.25 kJ/m2. Then, although the initial values, GIc, of T800H/3900-2 were about four times higher than that of reference T800H/ 3631, the propagation value, GIR, was almost identical for both laminates. 3.1.2. Fatigue delamination The results of the fatigue tests for T800H/3900-2 also show unique behavior. Fig. 4 shows the change of the crack propagation rate, da/dN, with the increment of the crack length for the load-shedding test under R = 0.5. Though the applied load and the corresponding energy release rate decreased with the increment of the crack length, da/dN first decreased till Da = 6 mm, then increased between Da = 6 and 8 mm, and decreased again after Da = 8 mm. Fig. 5 compares the macroscopic fracture surface under static and fatigue loadings. The fracture surface was divided into white and black parts. Fig. 6 shows the corre-
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
Crack propagation rate, da/dN (m/cycle)
668
10
-6
10
-7
10
-8
10
-9
T800H/3900-2 (R=0.5)
Fig. 5. Comparison of macroscopic fracture surfaces under static and fatigue loadings for T800H/3900-2 laminates. -10
10
0
2 4 6 8 10 Increment of crack length, a (mm)
12
Fig. 4. Change of crack propagation rate with increment of crack length for T800H/3900-2 laminates.
sponding microscopic fracture surfaces using SEM (scanning electron microscope). Higher ductility of resin and roughness were observed in the white parts. On the other hand, typical interlaminar fracture surfaces with carbon fiber and brittle epoxy resin were observed in the black parts. Then, the unique crack growth behavior of T800H/3900-2 laminate in Figs. 3 and 4 is attributed to the transition of the crack path from the toughened interlayer region (Stage I) to the untoughened interlayer/base lamina interface
(Stage II). It is interesting to note that this transition under fatigue loading was completed only after Da = 8 mm in Figs. 4 and 5. Thus, the transition from Stage I to II occurred for much shorter Da under fatigue loading than that under static loading. Fig. 7 compares the optical micrographs of the side surface of the specimen under static and fatigue loadings. While the crack path under fatigue loading was rather straight passing through polyamide particles, this path under fatigue loading often went around polyamide particles. The smaller cyclic damage zone is probably responsible for this difference. Then, the driving force for the stage transition is suggested to be larger under fatigue loading. This can be the reason for the shorter transition length under fatigue loading.
Fig. 6. Scanning electron micrographs of fracture surfaces for T800H/3900-2 laminates. (a-1) Static, white part (Da = 0.1 mm), (a-2) static, black part (Da = 9.3 mm), (b-1) fatigue, white part (Da = 0.4 mm), (b-2) fatigue, black part (Da = 6.1 mm).
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
669
Fig. 7. Optical micrographs of side surface of specimen at interlayer region for T800H3900-2 laminates: (a) static fracture, (b) fatigue fracture.
Fig. 8. Relation between crack propagation rate and maximum energy release rate under R = 0.1 for T800H/3900-2 laminate.
Fig. 9. Relation between crack propagation rate and maximum energy release rate under R = 0.5 for T800H/3900-2 laminate.
Figs. 8 and 9 show the relation between da/dN and the maximum energy release rate, Gmax, which corresponding to the maximum load, under R = 0.1 and 0.5. All of the fatigue tests were started from the end of the starter slit without introducing the precracks, and different marks were used for each specimen. As was suggested in Fig. 3, the crack growth behavior was very complicating. The specimen indicated by open circles in Fig. 9 corresponds to the same specimen for Fig. 3. The data of this specimen suggest that there exist two power-law relations for Stage I and II. Since almost all specimens indicate the transition from Stage I to II from the beginning of the fatigue tests, the dd/dN just after starting the tests were regarded to represent the data for Stage I. Finally, the power-law relation for Stage I is shown by solid lines in Figs. 8 and 9.
It is also difficult to get the clear power-law relations for Stage II. SEM observation at lower magnification indicated bridged fibers. This brings apparent increase of the fatigue crack growth resistance for Stage II, and da/dN– Gmax relation deviates from the true Stage II relation [27,28]. SEM observation indicated that the fiber bridging started just after the transition from Stage I to II was finished. Since this transition gradually grows in the width direction, it is very difficult to define the true Stage II relation without the fiber bridging effect. Then, we tentatively defined the Stage II relation from the data just after the transition to Stage II was finished throughout the width direction. These points are indicated by solid symbols, and the power-law relations are shown by the dashed lines in Figs. 8 and 9. The power-law relation for the
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
DK eq ¼ DK 1c K cmax ; where the stress-ratio-effect parameter, c (0 < c < 1), indicates the relative contribution of the maximum load to the cyclic load in determining the crack propagation rate. These c values for Stage I and II are 0.30 and 0.50, respectively. The c value for Stage I is similar to that for ductile AS4/PEEK (0.27) [30], and that for Stage II is comparable to that for T800H/3631 (0.46) [29]. These facts agree well with the crack paths for the Stage I and II. In Stage I, the polyamide particles give much higher ductility as shown in Fig. 6, and the mode I fatigue fracture mechanism is controlled by the cyclic load similar to that of the AS4/ PEEK where the matrix is ductile thermoplastic. In Stage II, the fracture surface is similar to normal CF/epoxy laminates, and the rather brittle epoxy is responsible for the higher contribution of the maximum load. 3.2. UT500/111/ionomer laminates 3.2.1. Interlaminar fracture toughness Static fracture toughness tests were carried out for UT500/111/ionomer (25 and 100 lm) and the base UT500/111 laminates to obtain the base properties using the same laminates as those for fatigue tests. Fig. 10 shows the relation between the mode I fracture toughness and the increment of crack length (R-curves) for all laminates. Here, the data points at Da = 0 indicate GIc values at the NL points. The GI values for the initial stage of the crack propagation sharply increased by about 2–3 times where Da is smaller than 5 mm for all laminates. Where Da is longer than 10 mm, GIR values leveled off though GIR values indicate scatter for all laminates. Whereas GIc and GIR values of UT500/111/ionomer (25 lm) laminates were 3.2
UT500/111/Ionomer
2
Mode I
2
Mode I fracture toughness, GIc , GIR (kJ/m )
reference laminate, T800H/3631, is also indicated in Fig. 9 [29]. Since the crack growth behavior is so complicated with the stage transition, it is rather difficult to compare the fatigue crack growth resistance by comparing the threshold values. Then, the growth resistance was compared at da/dN = 109 m/cycle. Fig. 9 indicates that the growth resistance of T800H/3900-2 (Stage I and Stage II) is 3.0 and 1.6 times higher than that of T800H/3631, respectively. This increase in Stage I is slightly lower than the increase of static GIc (4 times higher). Though the Stage I resistance is 0.64 times of that of AS4/PEEK with thermoplastic matrix [30], these results assure higher resistance than the reference T800H/3631 laminates even in Stage II, and indicate substantial increase in Stage I. It is interesting to note that the threshold values of T800H/3900-2 under mode II loading increased only about 1.5 times from that of the reference T800H/3631. This increase was much smaller than the increase of GIIc (2.4 times) and GIIR (3.9 times) [11]. We are proposing the following equivalent stress intensity range, DKeq, as a controlling parameters for fatigue crack growth of composite laminates under carious R values [21,25]:
1.5
1
0.5 Ionomer 100µm Ionomer 25µm Base
0
0
10 20 30 Increment of crack length,
40
Fig. 10. Relation between fracture toughness and increment of crack length for base UT500/111 and UT500/111/ionomer.
times and 3.5 times higher than those of the base laminates, GIc and GIR of UT500/111/ionomer (100 lm) laminates were 3.9 times and 5.5 times higher than those of the base UT500/111 laminates. 3.2.2. Fatigue delamination For the fatigue tests of UT500/111/ionomer (25 lm) laminates, the GImax-constant test [27] was first carried out, and the results were shown in Fig. 11. The change of the crack propagation rate is within 3 times, and the global change of the crack propagation rate is not observed. This suggests that the transition of the crack path which is observed in T800H/3900-2 is not expected for UT500/ 111/ionomer (25 lm) laminates. Fig. 12 shows the relation between the crack propagation rate and the maximum energy release rate in the GImaxdecreasing tests for UT500/111/ionomer (25 lm) and the
Crack propagation rate, da/dN (m/cycle)
670
10
-7
10
-8
10
-9
UT500/111/ionomer
Mode I
G
10
-10
10
-11
2
Imax
=250 (J/m )
R=0.5
0
2
4
6
8
10
12
14
16
Increment of crack length, Fig. 11. Change of crack propagation rate with increment of crack length for UT500/111/ionomer laminates.
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
671
Fig. 12. Relation between crack propagation rate and maximum energy release rate for base UT500/111 and UT500/111/ionomer laminates.
base UT500/111 laminates. As was expected from Fig. 11, normal power-law relations were observed even for the ionomer-interleaved laminates under R = 0.2 and 0.5 da/dN deviates from the power-law relations where da/dN is lower than 109 m/cycle, and there existed the growth threshold. The threshold value of the maximum energy release rate, GImax, for UT500/111/ionomer (25 lm) laminates (120 J/m2 under R = 0.2, 180 J/m2 under R = 0.5) is 3.5–3.3 times higher than those for the base laminates (34 J/m2 under R = 0.2, 54 J/m2 under R = 0.5). This ratio is similar to that for T800H/3900-2 under fatigue loading in Stage I. Moreover, this ratio is almost the same as that of GIR for UT500/111/ionomer (25 lm) laminates. The great advantage for this ionomer-interleaved laminates is no transition to the weaker stage. The exponents of the power function for UT500/111/ionomer (25 lm) laminates (16 under R = 0.5) were about two times larger than those for the base laminates (8 under R = 0.5). It has been reported that the relative increase of the fatigue crack growth resistance from the base laminates to CFRP laminates with thermoplastic matrix such as the AS4/PEEK was lower than that of the fracture toughness [21,31]. Table 1 shows the comparison of the ratio of the fatigue crack growth resistance to the fracture toughness, GImaxth/GIc, among T800H/3900-2, UT500/111, UT500/ 111/ionomer and AS4/PEEK laminates [30]. The GImaxth/ GIc values for T800H/3900-2, UT500/111, UT500/111/ ionomer and AS4/PEEK laminates were 0.35, 0.34, 0.36
and 0.25, respectively. Then, the GImaxth/GIc values for the interlayer or interleave laminates are higher than that of laminates with thermoplastic matrix. Stress-ratio effect parameter, c, was calculated for each crack propagation rate using Eq. (1) to discuss the stress ratio dependency on the fatigue crack growth. Fig. 13 shows the results. The c values for both laminates (0.5– 0.7) were relatively high, which indicates that the contribumode I
1 UT500/111/ionomer(25 m) UT500/111(base) 0.8
0.6
0.4
0.2
0
-9
-8
-7
10 10 10 Crack propagation rate, da/dN (m/cycle)
Fig. 13. Stress-ratio effect parameter for base UT500/111 and UT500/111/ ionomer laminates.
Table 1 Comparison of delamination threshold to fracture toughness Laminates
T800H/3900-2
Base UT500/111
UT500/111/ionomer (25 lm)
AS4/PEEK
R = 0.5 GImaxth(J/m2) GIc (J/m2) GImaxth/GIc
250 710 0.35
54 160 0.34
180 505 0.36
300 1200 0.25
672
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
Fig. 14. Lower magnification SEMs of fracture surfaces for UT500/111/ionomer laminates. (a-1) Static fracture, (a-2) static fracture, (b-1) fatigue, da/ dN = 107 m/cycle, base lamina side, (b-2) fatigue, da/dN = 107 m/cycle, ionomer side.
tion of maximum stress is large in the fatigue fracture behavior under mode I loading for both laminates [31]. 3.2.3. Microscopic observation and mechanism consideration Fig. 14 presents the lower magnification SEMs of the fracture surfaces of the UT500/111/ionomer (25 lm) laminates under static and fatigue loadings, respectively. The ionomer region was left like small ‘‘islands’’ on the fracture surfaces, which indicates the existence of ionomer bridging [32]. Traces of pull-out fibers are observed at the ionomer region, and the fibers are observed at the rest of the fracture surface. This indicates that the crack path is ionomer interleaf/base lamina interface. Fig. 15 shows the SEMs of higher magnification for UT500/111/ionomer (25 lm) and the base UT500/111 laminates. In this study, the fatigue fracture surfaces of both laminates under R = 0.2 and R = 0.5 were investigated. No significant effect of the stress ratio, R, was observed on the fatigue fracture surfaces for each laminates. Thus, the fatigue fracture surfaces under R = 0.5 were chosen for the discussion. For both laminates, little difference was observed under static and fatigue fracture. Moreover, the effect of the growth rate under fatigue loading was also minimal. For the base UT500/111 laminates, fiber/matrix interfacial fracture was dominant. The areal ratio of the interfacial fracture was about 70%. The remaining resin part indicates rather brittle fracture.
The areal ratio of the interfacial fracture for UT500/111/ ionomer (25 lm) laminates was also about 70% both under static and fatigue loadings. For the Stage I of T800H/39002, large drawing of resin was only observed under static loading, and there was a substantial difference between the fracture surfaces of static and fatigue fracture. Similar results were obtained for AS4/PEEK with thermoplastic matrix (Fig. 6). However, no significant difference was observed for this ionomer-interleaved laminates under static and fatigue loadings. This means that the microscopic fracture mechanisms of the resin under static and fatigue loadings are rather similar. Large drawing of resin, which was observed only under static loading for T800H/3900-2 (Stage I) and AS4/PEEK [21], was observed both under static and fatigue loadings. Since the fracture path was at ionomer interleave/base lamina interface, the resin consists of ionomer and epoxy, and this resin indicates higher ductility than base matrix epoxy resin. The similarity of the fracture surfaces under static and fatigue loadings agrees with the higher c values in Fig. 13 (c = 0.5–0.7), which indicates that the contribution of maximum load is higher than that of cyclic load in delamination fatigue fracture. The higher ratio of interfacial fracture also contributes to this tendency. Section 3.1 indicates that the fracture toughness and the fatigue crack growth resistance under mode I loading de-
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
673
Fig. 15. SEMs of fracture surfaces for base UT500/111 and UT500/111/ionomer laminates (continued). (a-1) Static fracture, base UT500111, (a-2) static fracture, UT500111/ionomer, (b-1) fatigue, da/dN = 107 m/cycle, base UT500/111, (b-2) fatigue, da/dN = 107 m/cycle, UT500/111/ionomer, (c-1) fatigue, da/dN = 1010 m/cycle, base UT500/111, (c-2) fatigue, da/dN = 1010 m/cycle, UT500/111/ionomer.
creases with the increment of crack length for the interlayer-toughened T800H/3900-2 laminates. The toughened region is usually limited within the resin-rich layer between the base lamina for the conventional interleaved laminates and the interlayer-toughened laminates. Thus, the crack path easily deviates from the toughened region to the untoughened region because the crack path can only be arrested by fibers (Fig. 16(a)). On the contrary, the high fracture toughness and the high fatigue crack growth resistance under mode I loading were maintained with the increment of crack length for the ionomer-interleaved laminates. It is known that the toughened region invades the base lamina
with forming the interphase for the ionomer-interleaved laminates (Fig. 1(b)). Thus, the crack path is always inside the toughened region for the ionomer-interleaved laminates (Fig. 16(b)) [20]. This is the reason why the high fracture toughness and the high fatigue crack growth resistance under mode I loading were maintained with the increment of crack length for the ionomer-interleaved laminates. Only one point we should note for this ionomer-interleaved laminate is that only this concept can be used for the future structures and the real application is still the subject of the future researches because the glass transition temperature of ionomer is rather low.
674
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675
Fig. 16. Schematic models of crack paths under mode I loading: (a) interlayer-toughened laminates, (b) ionomer-interleaved laminates.
4. Conclusions
Acknowledgment
Mode I delamination fatigue crack growth properties of two types of interlayer/interleaf CFRP laminates were investigated in the present study. The results are summarized as follows: The laminates with heterogeneous interlayer with fine polyamide particles, T800H/3900-2, indicated the transition of the crack path from the toughened interlayer region (Stage I) to the untoughened interlayer/base lamina interface (Stage II). The corresponding relation between the crack growth rate and the maximum energy release rate also showed transition of the power-law relation from Stage I to II. The crack growth resistance for toughened Stage I was 3.0 times higher than the reference laminates. Though the growth resistance decreased in Stage II, it was still 1.6 times higher than that of the reference laminates. This transition occured for much shorter increment of the crack length than that under static loading. The laminates with a new thermoplastic resin (ionomer) interleaf, UT500/111/ionomer, keep higher crack growth resistance without respect to the increment of the crack length under fatigue loading. The threshold value was 3.3–3.5 times higher than the base laminates. This increase was similar to that of the propagation values of the static fracture toughness. Though the crack path was at the interleaf/base lamina interface, the resin of this region consists of ionomer and epoxy resin, forming toughened interphase. Thus, the crack path was still within the toughened region without respect to the crack length.
Authors thank Prof. T. Aoki of University of Tokyo for his heplful discussion, and Mr. H. Ito. of All Nippon Airways and T. Ando of Graduate Student of Kyoto University for carrying out experiments. Authors also thank Dr. A. Kitano ot the Tray Industries and Mr. M. Ando of Toho Tenax for supplying laminates. References [1] OÕBrien TK. Characterization of delamination onset and growth in a composite laminate. In: Reifsnider KL, editor. Damage in composite materials, ASTM STP 775; 1982. p. 140–67. [2] Standard tests for toughened resin composites. NASA RP 1092; 1983. [3] Paul K, Kelly L, Venkayya V, Hess T. Evolution of U.S. military aircraft structures technology. J Aircraft 2002;39(1):18–29. [4] Friedrich K. Application of fracture mechanics to composite materials. Amsterdam: Elsevier; 1989. [5] Sela N, Ishai O. Interlaminar fracture toughness and toughening of laminated composite materials: a review. Composites 1989;20(5): 423–35. [6] Jain LK, Mai Y-W. On the effect of stitching on mode I delamination toughness of laminated composites. Compos Sci Technol 1994;51: 331–45. [7] Yan W, Liu H-Y, Mai Y-W. Numerical study on the mode I delamination toughness of Z-pinned laminates. Compos Sci Technol 2003;63:1481–93. [8] Kamiya R, Chou TW. Strength and failure behavior of stitched carbon/epoxy composites. Metall Mater Trans A 2000;31:899– 909. [9] Mouritz AP, Bannister MK, Falzon PJ, Leong KH. Review of applications for advanced three-dimensional fibre textile composites. Composites A 1999;30:1445–61.
M. Hojo et al. / Composites Science and Technology 66 (2006) 665–675 [10] Kageyama K, Kimpara I, Ohsawa I, Hojo M, Kabashima K. Mode I and mode II delamination growth of interlayer toughened carbon/ epoxy (T800H/3900-2) composite system. In: Martin RH, editor. Composite materials: fatigue and fracture – fifth volume, ASTM STP 1230; 1995. p. 19–37. [11] Matsuda S, Hojo M, Ochiai S. Mesoscopic fracture mechanism of mode II delamination fatigue crack propagation in interlayertoughened CFRP. JSME Int J Ser A 1997;40(4):423–9. [12] Ogihara S, Takeda N, Kobayashi S, Kobayashi A. Effects of stacking sequence on microscopic fatigue damage development in quasiisotropic CFRP laminates with interlaminar-toughened layers. Compos Sci Technol 1999;59:1387–98. [13] Bradley WL. Relationship of matrix toughness to interlaminar fracture toughness. In: Friedrich K, editor. Application of fracture mechanics to composite materials. Amsterdam: Elsevier; 1989. p. 159–87. [14] Sela N, Ishai O, Banks-Sills L. The effect of adhesive thickness on interlaminar fracture toughness of interleaved laminates specimens. Composites 1989;20(3):257–64. [15] Singh S, Partridge IK. Mixed mode fracture in an interleaved carbonfibre/epoxy composite. Compos Sci Technol 1995;55:319–27. [16] Odagiri N, Kishi H, Nakae T. T800H/3900-2 toughened epoxy prepreg system: toughening concept and mechanism. In: Proceedings of the American society for composites, 6th technical conference; 1991. p. 43–52. [17] Odagiri N, Kishi H, Yamashita M. Development of TORAYCA prepreg P2302 carbon fiber reinforced plastic for aircraft primary structural materials. Adv Compos Mater 1996;5:249–52. [18] Hojo M, Matsuda S, Ochiai S, Murakami A, Akimoto H. The role of interleaf/base lamina interphase in toughening mechanism of interleaf-toughened CFRP. In: Proceedings of the twelfth international conference on composite materials; 1999 [CDROM]. [19] Akimoto H, Kanazawa T, Yamada M, Murakami A. Impact behaviors of ethylene based ionomers in tensile and penetration mode. In: Proceedings of the fifth Japan international SAMPE symposium; 1997. p. 67–72. [20] Matsuda S, Hojo M, Ochiai S, Murakami A, Akimoto H, Ando M. Effect of ionomer thickness on mode I interlaminar fracture toughness for ionomer toughened CFRP. Composites A 1999;30:1311–9.
675
[21] Hojo M, Ochiai S, Gustafson C-G, Tanaka K. Effect of matrix resin on delamination fatigue crack growth in CFRP laminates. Eng Fract Mech 1994;49(1):45–7. [22] Hojo M, Matsuda S, Fiedler B, Kawada T, Moriya K, Ochiai S, et al. Mode I and II delamination fatigue crack growth behavior of alumina fiber/epoxy laminates in liquid nitrogen. Int J Fatigue 2002;24:109–18. [23] Hojo M, Tanaka K, Kageyama K. Pre-standardization study on mode I interlaminar fracture toughness test for CFRP in Japan. Composites 1995;26(9):243–55. [24] JIS K 7086-1993. Testing methods for interlaminar fracture toughness of carbon fiber reinforced plastics. Japanese Standards Association, Tokyo; 1993. [25] Hojo M, Tanaka K, Gustafson C-G, Hayashi R. Effect of stress ratio on near-threshold propagation of delamination fatigue cracks in unidirectional CFRP. Compos Sci Technol 1987;29:273–92. [26] Sih GC, Paris PC, Irwin GR. On cracks in rectilinearly anisotropic bodies. Int J Fracture Mech 1965;1:189–203. [27] Hojo M, Ochiai S, Aoki T, Ito H. New simple and practical test method for interlaminar fatigue threshold in CFRP laminates. In: Proceedings of European conference on composites testing and standardization (ECCM-CTS 2); 1994. p. 553–61. [28] Martin RH, Murri GB. Characterization of mode I and mode II delamination growth and thresholds in AS4/PEEK composites. In: Garbo SP, editor. Composite materials: Testing and design. 9th vol., ASTM STP 1059; 1990. p. 251–70. [29] Tanaka K, Tanaka H, Tsuji T, Yamagishi K. Effect of stress ratio on mode I propagation of interlaminar fatigue cracks in CFRP. J Soc Mater Sci Jpn 1995;44:960–6. [30] Hojo M, Ochiai S, Aoki T, Ito H. Mode I fatigue delamination for CF/PEEK laminates using maximum-energy-release-rate constant tests. J Soc Mater Sci Jpn 1995;44:953–9. [31] Hojo M, Matsuda S, Ochiai S. Delamination fatigue crack growth in CFRP laminates under mode I and II loadings – effect of mesoscopic structure on fracture mechanism. In: Proceedings of international conference on fatigue of composites; 1997. p. 15– 26. [32] Matsuda S, Hojo M, Ochiai S, Moriya K, Murakami A, Usui Y, et al. J Soc Mater Sci Jpn 1998;47:1125–30.