epoxy composites

epoxy composites

COMPOSITES SCIENCE AND TECHNOLOGY Composites Science and Technology 64 (2004) 2271–2278 www.elsevier.com/locate/compscitech The effect of water immers...

695KB Sizes 1 Downloads 15 Views

COMPOSITES SCIENCE AND TECHNOLOGY Composites Science and Technology 64 (2004) 2271–2278 www.elsevier.com/locate/compscitech

The effect of water immersion ageing on low-velocity impact behaviour of woven aramid–glass fibre/epoxy composites Krystyna Imieli nska

a,*

, Laurent Guillaumat

b

a

b

Materials Science and Engineering Department, Gdansk University of Technology, Narutowicza 11/12, 80-952 Gdansk, Poland Laboratoire Materiaux Endommagement Fiabilite et Ingenierie des Procedes, ENSAM, Esplanade des Arts et Metiers, 33405 Talence Cedex, France Received 29 October 2003 Available online 12 April 2004

Abstract Two different woven glass–aramid-fibre/epoxy laminates were subjected to water immersion ageing followed by instrumented low velocity impact testing. The hybrid aramid–glass reinforcement consisted of 10 plies of woven aramid–glass-fibre fabric or alternatively aramid-fibre fabric with glass-fibre fabric interlayers. The impacted plates were retested statically in compression to determine residual strength for assessment of damage tolerance. The maximum water absorption (4.1–4.4%) and water diffusion coefficients were found to be only slightly dependent on reinforcement configuration. The delamination threshold load and impact energy absorption were not significantly affected by the absorbed water. Due to low fibre–matrix adhesion, the prevailing failure modes at low impact energy were fibre/matrix debonding and interfacial cracking. The compression strength suffered significant reductions with water absorbed (28%) and impact (maximum 42%). The least sensitive to impact damage were wet samples of interlaminated composite. The experimental results of residual compression strength have been compared with predictions based on a simple, empirical model. Ó 2004 Elsevier Ltd. All rights reserved. Keywords: A. Aramid fibre; B. Durability; B. Hygrothermal effect; B. Impact behaviour; D. Scanning electron microscopy

1. Introduction Glass- and aramid-reinforced polymer laminates are commonly used as light-weight materials in a wide variety of marine applications including sporting equipment as well as military structures. However, the shortcoming of glass-reinforced plastics along with other fibre-reinforced polymer laminates is that their mechanical properties in the translaminar (through thickness) direction are relatively low [1]. Owing to the weak bonds between the plies, even small energies imparted by out-of-plane loads (low velocity impact) can result in damage, which, although hardly detectable, causes considerable strength losses in tension and, especially, in compression [2–5]. Changes in material properties can also result when a

*

Corresponding author. E-mail addresses: [email protected] (K. Imieli nska), l-guillaumat @lamef.bordeaux.ensam.fr (L. Guillaumat). 0266-3538/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.compscitech.2004.03.002

composite structure is exposed to moisture. The irreversible material degradation includes chemical changes (e.g., relaxation and oxidation of the matrix material), debonding at fibre/matrix interfaces and continuous cracks. These changes affect the overall material properties, e.g., elastic modulus, hygrothermal expansion and diffusion coefficients [6]. Susceptibility to low velocity impact damage of composite materials has been well documented [2–5,7,8]. Studies have also been conducted to evaluate the mechanical response and creep behaviour of polymeric composites subjected to environmental conditions such as humidity and temperature [9–11]. However, there is little information available on the role of service environment on the impact resistance/tolerance of polymeric composites. Karasek et al. [12] have evaluated the influence of temperature and moisture on the impact resistance of epoxy/graphite fibre composites. They found that only at elevated temperatures did moisture have a significant effect on damage initiation energy and that

2272

K. Imielinska, L. Guillaumat / Composites Science and Technology 64 (2004) 2271–2278

the energy required to initiate damage was found to decrease with temperature. Impact damage resistance and tolerance of two high performance polymeric systems was studied after exposure to environmental aging. Specimens aged in nitrogen for 18 months had equivalent damage to those aged in air for only 2 months. For cross-ply laminates, the post-impact tensile strength values fell significantly (by maximum 70–75% of original composite strength) depending on ageing time, environment and impact velocity [13]. Sala [14] found that barely visible impact damage, BVID, due to the impact of 1 J/mm (for 2.2-mm laminate thickness) increased the moisture saturation level from 4.8% to 6% for aramid fibre-reinforced laminates and enhanced the absorption rate. In the case of carbon fibre composite, there was no effect of BVID on moisture absorption curves. Hybrid, fibrous – glass, aramid, carbon fibre laminates have rarely been characterised in terms of moisture absorption [11] and impact [15–17]. In the absence of epoxy matrix, Kevlar fibres exhibit very high impact resistance in terms of maximum force and energy dissipation. The impact resistance of two-ply laminates: Kevlar/Kevlar/epoxy and hybrid carbon/Kevlar/epoxy was not improved when compared with brittle carbon/ carbon/epoxy laminate [15]. In the construction of small, high performance vessels, Kevlar fibres are often used in combination with glass and carbon reinforcements as a low weight, impact-resistant replacement of glass fibres. Consequently, there is a need to provide more experimental data on the behaviour of laminates reinforced with hybrid fibres, so as to explain the role of aramid fibres, especially in terms of impact and wet environment resistance. In the present work the combined effect of water immersion and low velocity impact behaviour has been studied for epoxy resin composites reinforced with woven aramid and glass fibres. The aim was to assess impact damage tolerance, depending on the type of glass–aramid reinforcement and laminate conditioning.

2. Materials and methods Laminates were fabricated from woven orthogonal balanced fibre fabrics: aramid–glass (REA 390S), ara-

mid (RA320H5) or E glass (STR 66-110). Aramid fibre fabric was supplied by SP Systems (UK) and glass fibre fabric by Krosno, Poland. Standard pro-adhesive treatment of aramid and glass fibres for use with epoxy resins was provided by the suppliers. Laminates consisted of 10 plies of fabric impregnated with epoxy resin by pressure bagging technique, widely used in marine industry. The resin was SP 115, typical Diglycidyl Ether of Bisphenol – A (DGEBA) cured with amine SP 115 hardner supplied by SP Systems. Detailed description of components and curing conditions of laminates is given in Table 1. The approximate fibre volume fraction was Vf ¼ 45–50%. Accelerated water immersion tests were performed on small pieces of laminates 15 mm  25 mm. Before the test samples were conditioned in an oven at 70 °C for 24 h, specimens were immersed in a container with distilled water at 70 °C for 8 weeks, and regular (every 24–72 h) weight gain measurements were performed. Water uptake was calculated as weight gains related to the weight of the dry specimen. Five composite specimens were tested and results averaged. Impact behaviour was assessed using dropping-mass tower [18]. An electric motor with a magnet raised the mass to the pre-determined drop height (0.5 and 0.75 m) – 4.3 kg. The plates (120 mm  100 mm) were placed on two steel opposite supports with a 100-mm span. Two flexible sheets made of metal with a rubber end were used just to maintain the plate. The pressure applied on the sample was very low compared to the incoming energy. These boundary conditions allow rotation and vertical movement. The contact load between the striker and the composite specimen versus time was measured by an accelerometer and a piezoelectric captor, which were attached to the dropping mass. Two laser captors logged the striker displacement and the out-of-plane displacement of the centre of the composite structure versus time. The front and back surfaces of the sample were examined optically. Internal damage was highlighted by backlighting the impact site and its size measured using an image analyser. Compression-after-impact (CAI) tests evaluated the damage tolerance of the composite, which refers to the ability to tolerate a specified amount of damage [5]. The compression specimens (120 mm  100 mm) were supported in an anti-buckling guide [3] that

Table 1 Description of laminates Type of reinforcement

Matrix

Number of plies thickness, h

Aramid–glass A–E ‘‘hybrid’’ woven aramid–glass fabric REA 390S

Epoxy resin SP 115 (hardener SP115), 3:1 Curing conditions: 23 °C, 2.5 h in vacuum (pressure bag), post-curing: 4 h in 60 °C

10h ¼ 4:1 mm

Aramid/glass A/E ‘‘interlayer’’ woven aramid fabric RA320H5 (SP Systems) interlaminated with woven E glass fabric STR 66-110 (Krosno, Poland) [A/E/A/E/A]s

10h ¼ 4:8 mm

K. Imielinska, L. Guillaumat / Composites Science and Technology 64 (2004) 2271–2278

gave edge support to the specimen but did not stop localised buckling of the damaged area. All tests were performed at a crosshead speed of 2 mm/min. The compression strength of undamaged specimens was also measured.

2273

posite consisting of hygroscopic aramid fibres impregnated with the resin. The effect of water immersion ageing on the microstructural integrity of the composites is illustrated in

5

3.1. Water-absorption behaviour Water uptake curves, showing only minor scatter, are shown in Fig. 1. The total moisture content at equilibrium for ‘‘hybrid’’ aramid–glass/fabric epoxy is only slightly smaller than that for ‘‘interlayer’’ composite. The shape of the curves is consistent with Fickian diffusion. The diffusion coefficients [6,9] are shown in Table 2, indicating only a slightly different behaviour of the two composites with a somewhat higher diffusion coefficient in the interlayer composite. This can be accounted by the external surfaces of ‘‘interlayer’’com-

Moisture uptake M,%

3. Results and discussion 4

"Hybrid" A-E/epoxy

3 2

"Interlayer" A/E/epoxy

1

scatter+/-0,2% 0 0

10

20

30

40

1/2

time, hours

Fig. 1. Water uptake in aramid–glass/epoxy laminates immersed in water at 70 °C.

Table 2 Water immersion test results Material

Maximum moisture contents

Diffusion coefficient (107 mm2 /s)

‘‘Hybrid’’ aramid–glass A–E/epoxy ‘‘Interlayer’’ aramid/glass/epoxy [A/E/A/E/A]s

4.1 4.4

3.34 4.83

Fig. 2. (a–c) SEM micrographs of the aramid/glass/epoxy (interlayer) composite in (a) ‘‘dry’’, (b–d) ‘‘wet’’, saturated (70 °C) condition. Arrows show the debonding and interfacial cracks between the aramid (white arrow) and glass (black arrow) fibres, (d) interfacial crack between individual aramid fibres.

2274

K. Imielinska, L. Guillaumat / Composites Science and Technology 64 (2004) 2271–2278

Fig. 2, which shows fine cracks and fibre/matrix debonding. White arrow in Fig. 2(c) points to the aramid fibre/matrix interfacial crack shown in detail in Fig. 2(d). Poor wetting of aramid fibre by epoxy resin is evident; it may easily produce fibre/matrix debonding even in the absence of water, as a result of thermal expansion mismatch during curing and thermal post curing of the laminate. However, since the cracks are also present in glass fibre regions (black arrow in Fig. 2(c)), it appears that it is rather hygrothermal mismatch of the fibre–matrix that is at the origin of interfacial cracking of the composite. 3.2. Low velocity impact behaviour Impact characteristics, load and deflection with respect to time as well as load versus deflection, for

6000

32J

"hybrid" dry

Force, N

5000 4000

"interlayer" dry

3000

"interlayer" wet

2000

"hybrid" wet

1000 0 -2

0

2

(a)

4

6

8

10

Time, ms 32J

Deflection, mm

0

hybrid dry

-2 -4

interlayer dry -6

interlayer wet -8

hybrid wet

-10 -2

0

2

(b)

4

6

8

3.3. Residual compression strength, damage tolerance

10

Time, ms 6000

hybrid

5000

32J

dry

Force. N

wet 4000 3000 2000

interlayer dry, wet

1000 0 -10

(c)

-8

-6

-4

-2

both dry and wet samples, are shown in Fig. 3. The usual [7,19] profile of load–time curve was observed with the damage threshold load identified by the sudden load drop due to loss of stiffness from unstable damage development (arrow). For both types of reinforcement (‘‘hybrid’’ and ‘‘interlayer’’) the load versus time curves are almost identical for ‘‘dry’’ and ‘‘wet’’ conditions. Only a slight difference in the initial slope of load–time curves for ‘‘dry’’ and ‘‘wet’’ samples indicates that the stiffness of the samples remained almost unchanged following water immersion. The impact compliance increased slightly for wet samples (Fig. 3(a)) due to matrix plasticisation and consequently bigger deflection of wet samples resulted (Fig. 3(b)). Load–deflection plots illustrating energy absorption potential of the samples showed that in wet condition (Fig. 3(c)) elastic energy absorbed (the surface area under lower portion of the curves) is a little lower than for dry samples, which is indicative of softening of the wet material. The area enclosed in force–deflection loop is only slightly larger for the ‘‘hybrid’’ wet samples, which indicates that energy dissipated for plastic deformation and damage is not significantly changed due to the absorbed water. Fig. 4 show the front and back faces (a–b, e–f) and back-face damage highlighted by backlighting (c–d, g–h) of the laminates in ‘‘wet’’ and ‘‘dry’’ conditions at incident impact energy 32 J and undamaged samples. The projected damage area is slightly more extensive for ‘‘interlayer’’ reinforcement and ‘‘wet’’ samples (Fig. 5). Fig. 6 illustrates impact damage in ‘‘dry’’ (a,b) and wet (c,d) aramid–glass/epoxy composites (incident impact energy 26 J) with delamination near impact site (a,b) as well as complex impact damage (matrix cracks combined with fibre/matrix debonding) in lower (c,d) part of the sample.

0

Deflection, mm

Fig. 3. Impact characteristics of aramid–glass/epoxy ‘‘interlayer’’ and ‘‘hybrid’’ composite in ‘‘dry’’ and ‘‘wet’’ condition. Incident impact energy 32 J.

The concept of damage tolerance embraces basic elements of the maximum amount of load or IKE a structure can take without experiencing damage, and the minimum or residual load-bearing capability the damaged structure can retain [3,20–22]. The first requirement is established by identification of the threshold load (Fig. 3) and, in order to satisfy the second requirement, the plots of compression strength after impact with respect to impact damage area need to be analysed (Fig. 7). The strength retention factors (residual compressive strengths of damaged specimens normalised with those for undamaged specimens, rr =r0 ) have been estimated as a function of impact energy (Fig. 8). Following the successful applications of a simple model [23] of residual strength for carbon- and glassfibre composites [24,25] the threshold energy and the

K. Imielinska, L. Guillaumat / Composites Science and Technology 64 (2004) 2271–2278

2275

400

Damage area, mm

2

hybrid dry hybrid wet

300

interlayer wet interlayer dry 200

100

0 0

5

10

15

20

25

30

35

IKE, J

Fig. 5. Damage area as a function of IKE (incident impact energy).

the methods to determine the threshold values, c0 and U0 , and the exponents a; b have been given in [5]. The threshold exponents a and b represent the rate of strength degradation beyond the threshold values. Both the threshold energy and the threshold damage area were similar in the different fibre configurations. The rates of strength degradation a and b beyond the threshold values dominate the entire strength characteristics. For interlaminar composites b ¼ 0:19 (dry samples), b ¼ 0:15 (wet samples), for the hybrid b ¼ 0:26 (both dry and wet). This result indicates a better retention of CAI strength for the interlayer composite. The experimental points in Figs. 7 and 8 generally compare well with the predictions (except hybrid dry samples at lower impact energies, where the scatter is quite significant).

4. Discussion

Fig. 4. Aramid–glass/epoxy ‘‘hybrid’’ (a–d) and ‘‘interlayer’’ (e–h) composites after impact of incident energy 32 J. Front face (a,e), back face (b,f) internal damage highlighted by backlighting of the back surface for ‘‘dry’’ (c,g) and ‘‘wet’’ (d,h) condition.

threshold damage width/area are predicted on the basis of a power-law relationship:  b rr  c0 a Uo ¼ ¼ ; r0 c U where r0 is the original strength, unaffected by cracks or damage smaller than the threshold value, c0 , rr is the residual strength after impact damage and c is the corresponding damage area or damage width. U0 represents the threshold impact energy that the material can withstand without strength degradation. The details of

Fig. 4 shows that the projected impact damage area is less extensive for wet samples. This must not lead to the conclusion that wet samples suffered less damage than the dry, since the nature of damage was different. The projected damage area cannot be the only criterion of damage assessment. Fig. 9 shows the ‘‘wet’’ hybrid sample in backlighting, prior to impact. No visible defects can be observed. However, the interfacial damage resulting from the action of water is present (as illustrated in Fig. 2), yet it was nor extensive or dense enough to scatter the light. It is suggested that at the moment of impact the interfacial defects propagated and combined with matrix cracks and delaminations (Fig. 6(b) and (c)). In wet samples the impact energy was absorbed in the propagation of existing defects, consequently, the delaminations were less extensive. This result needs to be analysed in combination with the data on compression strength retention factor as a function of incident impact energy (Fig. 8). Slight difference in the

2276

K. Imielinska, L. Guillaumat / Composites Science and Technology 64 (2004) 2271–2278

hybrid dry

hybrid dry interlayer dry

150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0

1,0

interlayer dry interlayer wet

interlayer wet hybrid wet

0

100

predictions interlayer, hybrid wet

hybrid wet

200

Damage area, mm

300

400

2

Fig. 7. Residual compression strength with respect to damage area for ‘‘hybrid and ‘‘interlayer’’ composite: results and prediction.

behaviour of dry and wet samples favours the wet samples. This can be explained by more extensive delaminations produced in dry samples (more detrimental to compression strength) as compared to interfacial damage prevailing in wet samples. Our results of compression strength after impact can be compared with the data of Kim and Sham [26] who considered the effect of glass fibre surface treatment on impact behaviour of glass/epoxy composites. Their threshold coefficient b was 0.25–0.38 and a ¼ 0:14–0:2,

Strength retention factor

Residual compression strength, MPa

Fig. 6. Impact damage in ‘‘dry’’ (a,b) and wet (c,d) aramid–glass/epoxy composites. Incident impact energy 26 J. Delamination near impact site (a,b), complex impact damage (matrix cracks and fibre/matrix debonding) in lower (c,d) part of the sample.

0,8

0,6

predictions interlayer, hybrid dry hybrid dry

0,4

hybrid wet interlayer dry

0,2

interlayer wet

0,0 0

10

20

30

40

IKE, J

Fig. 8. Compression strength retention factor as a function of a impact energy for ‘‘wet’’ and ‘‘dry’’ samples – results and predictions: ‘‘hybrid’’ and ‘‘interlayer’’ composites.

depending on fibre surface treatment, comparable with the results obtained in this study. However, their [5] threshold impact energy was much higher (15.9–22.6 J) than the approximate value assumed in the present study – 5 J – adopted following experimental observations for all materials, due to low fibre/matrix adhesion. The strength retention factor obtained in [5] for glass fibre/vinyl ester composites following impact (30 J) of

K. Imielinska, L. Guillaumat / Composites Science and Technology 64 (2004) 2271–2278

2277

8. The results of residual compression strength compare well with predictions of an empirical model of Caprino [23]. Acknowledgements The authors thank Mr. Rafal Wojtyra from the Ocean Technology and Shipbuilding Faculty Gdansk University of Technology for producing the materials and to Mr. Piotr Wiercinski an undergraduate of Gdansk University of Technology (Poland), for his participation in the experiments during his ErasmusSocrates training at Universite Bordeaux 1. Fig. 9. ‘‘Wet’’ hybrid sample in backlighting, prior to impact.

References dry samples was 80% as compared with 71% for dry samples of ‘‘interlayer’’ composites studied in this work.

5. Conclusions Two different woven glass–aramid fibre/epoxy laminates were subjected to water immersion ageing followed by low velocity impact. The impacted plates were retested statically in compression to determine residual strength for the assessment of damage tolerance. The following results were found: 1. Water immersion ageing affected microstructural integrity of the two composites causing numerous internal defects. 2. No important effect has been found of aramid–glassfibre configuration (woven hybrid fabric of glass and aramid fibres or aramid fibre fabric interlaminated with layers of glass fibre fabric) on moisture absorption and impact tests characteristics. 3. Threshold damage (fibre/matrix debonding) load was very low (1000 N) and independent of laminate lay-up or conditioning. 4. Due to low fibre–matrix adhesion, the prevailing impact failure modes were fibre/matrix debonding and interfacial cracks for both ‘‘wet’’ and ‘‘dry’’ materials. 5. Impact damage area was slightly less extensive in wet samples, which is suggested to be the result of the propagation of interfacial damage present in wet samples prior to impact, which absorbed impact energy and inhibited the delamination formation. 6. The compression strength of the two materials suffered 28% reduction due to water absorption (undamaged condition) and maximum 42% with the impact of incident energy 32 J (wet samples). 7. Wet samples of interlayer composite were less sensitive to impact than wet hybrid samples, which resulted in the higher (minimum) compression strength retention factor: 0.77 and 0.63, respectively.

[1] Mouritz AP, Gallagher J, Goodwin AA. Flexural strength and interlaminar shear strength of stiched GRP laminates following repeated impacts. Compos Sci Technol 1997;57:509– 22. [2] Lopresto V, Scarponi C, Briotti G. Influence of material thickness on the response of carbon-fabric/epoxy panels to low velocity impact. Compos Sci Technol 1999;59:2279–86. [3] Davies GA, Hitchings D, Zhou G. Impact damage and residual strength of woven fabric glass/polyester laminates. Composites 1996;27A:1147–56. [4] Hawyes VJ, Curtis PT, Sout C. Effect of impact damage on the compressive response of composite laminates. Composites A 2001;32:1263–70. [5] Hirai Y, Hamada H, Kim JK. Impact response of glass woven fabric composites – I. Effect of fibre surface treatment. Compos Sci Technol 1998;58:91–104. [6] Gellert EP, Turley DM. Seawater immersion ageing of glass-fibre reinforced polymer laminates for marine applications. Composites A 1999;30:1259–65. [7] Schoeppner GA, Abrate S. Delamination threshold loads for low velocity impact on composite materials’. Composites 2000;31:9. [8] Richardson MOW, Wisheart MJ. Review of low velocity impact properties of composite materials. Composites A 1996;27: 1123. [9] Imieli nska K, Wojtyra R. ’Water absorption behaviour of vinyl ester laminates reinforced with aramid and glass fibres’. Kompozyty 2003 [in Polish]. [10] Akay M, Kong AM, Stanley A. Influence of moisture on the thermal and mechanical properties of autoclaved and oven-cured Kevlar-49/epoxy laminates. Compos Sci Technol 1997;57:565–71. [11] Wood CA, Bradley W. Determination of the effect of seawater on the interfacial strength of an interlayer E-glass/graphite/epoxy composite by in situ observation of transverse cracking in an environmental SEM. Compos Sci Technol 1997;57:1033. [12] Karasek MC, Strait LH, Amateau MF. Effect of temperature and moisture on the impact behaviour of graphite/epoxy composites: parts I and II. J Mater Res Technol 1995;17(1):3–15. [13] Parvatareddy H, Wilson Tsang PH, Dillard DA. Impact damage resistance and tolerance of high performance polymeric composites subjected to environmental aging. Compos Sci Technol 1996;56:1129–40. [14] Sala G. Composite degradation to fluid absorption. Composites B 2000;31:357. [15] Jang BZ, Chen LC, Wang CZ, Lin HT, Zee RH. Impact resistance and energy absorption mechanisms in hybrid composites. Compos Sci Technol 1989;34:305–35.

2278

K. Imielinska, L. Guillaumat / Composites Science and Technology 64 (2004) 2271–2278

[16] Imieli nska K, Wojtyra R, Castaings M. Impact resistance and damage tolerance of hybrid: carbon, glass, kevlar/epoxy laminates. In: XVIth Physical Metallurgy and Materials Science Conference on Advanced Materials & Technologies, AMT’2001, Gda nsk-Jurata, Poland; September 16–20, 2001. [17] Harris B. Impact properties of glass-fibre carbon fibre hybrid composites. Composites 1975:197–201. [18] Guillaumat L. Reliability of composite structures – impact loading. Comput Struct 2000;76:163–72. [19] Davies GAO, Hitchings D, Wand J. Prediction of threshold impact energy for onset of delamination in quasi-isotropic carbon/ epoxy composite laminates under low-velocity impact. Compos Sci Technol 2000;60:1–7. [20] Davies GAO, Zhang X, Zhou G, Watson S. Numerical modelling of impact damage. Composites 1994;25:342–50.

[21] Davies GAO, Zhang X. Impact damage prediction in carbon composite structures. Int J Impact Engrg 1995;16(1):149–70. [22] Zhang X. Impact damage in composite aircraft structures – experimental testing and numerical simulation. Proc. Inst. Mech. Eng. 1998:212–59. [23] Caprino G. On the prediction of residual strength for notched laminates. J Mater Sci 1983;18:2269–73. [24] Kim JK, Mackay DB, Mai YW. Drop-weight impact damage tolerance of CFRP with rubber modified epoxy matrix. Composites 1993;24:458–94. [25] Caprino G, Teti R. A consistent engineering method for the treatment of impact in composite materials. Compos Sci Technol 1992;53:395–412. [26] Kim J-K, Sham M-L. Impact delamination failure of wovenfabric composites. Compos Sci Technol 2000;60:745–61.