Hybrid composites based on polyethylene and carbonfibres. Part 6: Tensile and fatigue behaviourCitation for published version (APA):Peijs, A. A. J. M., & Kok, de, J. M. M. (1993). Hybrid composites based on polyethylene and carbon fibres. Part6: Tensile and fatigue behaviour. Composites, 24(1), 19-32. https://doi.org/10.1016/0010-4361(93)90260-F

DOI:10.1016/0010-4361(93)90260-F

Document status and date:Published: 01/01/1993

Document Version:Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can beimportant differences between the submitted version and the official published version of record. Peopleinterested in the research are advised to contact the author for the final version of the publication, or visit theDOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, pleasefollow below link for the End User Agreement:

www.tue.nl/taverne

Take down policyIf you believe that this document breaches copyright please contact us at:

openaccess@tue.nl

providing details and we will investigate your claim.

Download date: 12. Jan. 2020

Hybrid composites based on polyethylene and carbon fibres

Part 6: Tensile and fatigue behaviour

A.A.J.M. PEIJS and J.M.M. DE KOK

(Eindhoven University of Technology, The Netherlands)

Received May 1992; accepted 2 June 1992

The tensile and fatigue behaviour of unidirectional carbon-high-performance polyethylene/epoxy hybrid composites has been studied, including the effect of hybrid design and surface treatment of the high-performance polyethylene (HP- PE) fibres. Results indicated that the tensile behaviour of carbon-HP-PE hybrids in both monotonic and fatigue testing can be interpreted, adopting the conventional 'constant strain' model for hybrid composites. Deviations from this constant strain model, so-called hybrid effects, were observed in monotonic tensile testing for those hybrid systems with the highest degree of fibre dispersion, incorporating either untreated or treated HP-PE fibres, whereas only the latter displayed synergistic fatigue performance. Hybrid effects under tensile loading conditions were in reasonable agreement with calculations accounting for statistical effects and stress concentrations as determined by finite element analyses.

Key words: composite materials; hybrid; polyethylene fibre; carbon fibre; epoxy; hybrid effect; tension; fatigue; adhesion

Advanced carbon fibre-reinforced composites possess desirable characteristics of high specific strength and stiffness, as well as superior long-term properties including fatigue resistance. However, due to their brittle nature, these materials are rather susceptible to impact damage. Consequently, material design concepts for enhanced damage tolerance of these materials are of great interest. One such concept is hybridization with high-performance polyethylene (HP-PE) fibres I-7. Depending on the volume fractions of the components, hybrid design and interface properties, it is possible to tailor the properties of such hybrids to specific applications requiring either structural performance 1,2,7, damage tolerance ~,3,6 or energy absorption 4.5.

Besides damage tolerance under impact conditions, hybridization also offers interesting possibilities with respect to crack arrest in composites. It has already been recognized for many years that the incorporation of strips into a laminate, which introduces macroscopic variations in stiffness and fracture toughness, results in a material construction with crack-arrest capability 8~9.

The incorporation of a low modulus fibre of high extensibility into a composite with high modulus fibres provides a mechanism of stopping or deflecting cracks at a microscopic level. It is generally recognized that such

0010-4361/93/010019-14 0

crack-arresting effects on a microscopic scale are of major importance with respect to the uniaxial tensile behaviour of unidirectional (UD) hybrid composites. Uniaxial tensile failure of hybrids has been studied quite extensively since the mid 1970s. Many of these studies identified a phenomenon called 'synergistic strengthening' or the 'hybrid effect'. In general, hybrid effects are defined as a positive or negative deviation from the rule of mixtures (ROM). However, in the case of uniaxial tensile loadings ROM behaviour is useless with respect to the understanding of the strength of hybrid composites. In such a case the hybrid effect is generally defined as an enhancement of the first failure strain of the low elongation (LE) fibre-reinforced component. The first macroscopic fracture in a hybrid occurs in the component with the lowest failure strain, being the carbon fibre composite in hybrid systems based on carbon and glass, aramid or HP-PE fibres. This enhancement of carbon fibre failure strain when tested in a hybrid was first reported by Hayashi j° for the system carbon-E-glass. Subsequent studies by several other researchers on hybrid composites also reported first failure strain enhancement under uniaxial tensile loadingl 1-15.

One of the early explanations for these hybrid effects was proposed by Bunsell and Harris H, who considered

1993 Butterworth-Heinemann ktd

COMPOSITES. VOLUME 24. NUMBER 1 . 1993 19

residual thermal strains as an important parameter. However, although these effects do occur in carbon-glass hybrid systems, in all cases these thermal strains are too small or even negative (in the case of carbon-aramid 14 and carbon-HP-PE 2) to have a significant effect on the enhancement of the carbon failure strain. Another, more important parameter with respect to the degree of strain enhancement is the degree of dispersion of the different fibres. For hybrids with constant fibre volume ratios, the largest hybrid effects have been observed for composites with the highest degree of fibre dispersion 11,15.16

In contrast with the amount of papers published on hybrid effects under monotonic tensile loading, little attention has been paid to the potential benefits of such crack-arresting effects on the performance of UD hybrid composites under fatigue loading conditions. Nevertheless, since a second fibre of higher extensibility can act to arrest a crack in brittle fibre composites, hybridization may also be favourable with respect to improved fatigue life.

Phillips ~7 showed for carbon-glass cloth an increase in tensile fatigue strength which is proportional to the amount of carbon fibres. Hofer et el. 18,19 studied the effect of stacking sequence on fatigue behaviour of (0°), ( 4- 45*) and (0", 4- 45*) hybrid laminates based on carbon and glass fibres. They showed that interply hybridization with alternating carbon-glass plies was superior to sandwich hybrids. Their work also included the influence of a number of environmental conditioning treatments on the fatigue performance of such hybrids. Grimes 2° studied fatigue testing of angle-ply hybrids with various compositions. Bunsell and Harris 21 used acoustic emission (AE) to monitor fatigue cycling of intimately mixed hybrids. More recently a group of researchers from the University of Bath ( U K ) 22-24 reported on repeated tension and tension-compression fatigue behaviour of UD and (+45 °, 0 °) hybrid laminates at various stress ratios. They noted that the fatigue stress for a given lifetime of UD carbon-Kevlar" hybrids varied linearly with composition 22. Since the tensile strength of such composites varies with Kevlar ® content at a lower rate than predicted by the ROM, the authors stated that the fatigue ratio, being the fatigue stress divided by the monotonic tensile strength, exhibited a positive synergistic or hybrid effect. Another related study on carbon-glass hybrids 24 stated that in the case of UD carbon-glass hybrids, both fatigue stress and fatigue ratio displayed a positive deviation from the ROM. A study by Marom et aL 25 dealt with flexural fatigue behaviour of UD carbon-aramid hybrid composites. They reported a positive hybrid effect on flexural fatigue stress for hybrids with a carbon fibre-reinforced core and an aramid fibre-reinforced skin. However, the degradation rate of the S I N curve itself for this hybrid composite---i.e., the slope of the S I N curve---was given by a ROM relationship.

Although there is only a limited amount of research devoted to the fatigue behaviour of hybrid composites, even less information is available on the influence of the interface on the fatigue performance of composites. The few studies in this area generally state that improved interracial bonding results in improved fatigue performance. Sih and Ebert 26 reported a linear increase in fatigue life and fatigue stress with shear strength for

2.00

0-

£

A

1.50

1.00

0.50

B

0.00

I\B T

j " \ - \

J J

J j " f

I

2O I I I

40 60 80 100

Volume ~ HP-PE

Fig. 1 Tensile s t rength vs. composi t ion of ca rbon-HP-PE hybrid composi tes : O, untreated HP-PE; 0 , treated HP-PE 2

UD glass/epoxy composites subjected to flexural loadings. Improved fatigue performance has also been reported for the system HP-PE/epoxy after fibre surface treatment 27.

The objective of the present study was to find conditions for synergistic strengthening in carbon-HP-PE hybrid composites under uniaxial tensile and fatigue loadings, including the influence of hybrid design and surface treatment of the HP-PE fibres.

THEORY

Constant strain model

The tensile strength of hybrid composites does not obey the ROM since the LE fibres are expected to break when the failure strain is reached. Based on such a 'constant strain' assumption, Manders and Bader 15 and Chou and Kelly 28 constructed a strength diagram of the type shown in Fig. 1 for the system carbon-HP-PEL In such a diagram, point A denotes the tensile strength of the LE

fibre composite, in this case the carbon composite, and point D that of the high elongation (HE) fibre composite, being the HP-PE composite. The strength of the hybrids is given by the two straight lines AC and CD. The line AE represents the stress in the hybrid at which failure in the LE phase is expected. At HE fibre volume fractions below point C, failure of the LE phase leads to instant failure of the hybrid composite since there is insufficient HE fibre to sustain the load after failure of the LE fibre component. The strength is then given by:

(~rhybn d = O'LEma x VLE "}- ELEmaxEHE VHE ( l )

where VLE and VnE are the volume fractions of the LE and HE fibres respectively, e is the strain and E is Young's modulus. With increasing amounts of HE fibres a transition from single fracture to a multiple fracture

20 C O M P O S I T E S . N U M B E R 1 . 1993

mode occurs when there are sufficient HE fibres to sustain the load after failure of the LE fibres at a stress level given by line CE. The ultimate strength of the hybrid (line CD) is then given by:

O'hybn d = O'HEma x VHE (2)

Since the strength of a material and in particular that of fibre-reinforced composites is a 'process' rather than a material property, the constant strain model is an oversimplification, yielding a lower bound of hybrid strength.

S t a t i s t i c a l m o d e l s

The requirement for statistical models arises from the variations in fibre strength resulting from a random distribution of flaws. Several theoretical models have been proposed to explain the experimentally observed deviations from the constant strain model--i.e., hybrid effects. Most models invoke a statistical distribution of strength, in which failure of the weakest LE fibre results in a crack that is bridged by the surrounding HE fibres, allowing the stronger LE fibres to reach their full potential. A statistical analysis of the strength of hybrid composites was originally developed by Zweben ~4. Based on models originally developed by Rosen 29 and Zweben 3° for single fibre composites, he presented a statistical analysis of hybrid composites consisting of a two- dimensional array of alternating LE and HE fibres in a matrix. His analysis recognizes a number of important material properties that influence the failure process in hybrid composites; firstly, the statistical failure strain characteristics of fibres and secondly, the stiffness ratio of both components as a result of different fibre moduli and cross-sectional areas. Besides Zweben, several other authors attempted to model the tensile strength of hybrids using statistical models (Manders and Bader 3t, Fukuda 32, Harlow 33 and Faribotz et al.34). Fukuda 32 modified the Zweben model, and gives the hybrid failure strain as:

E,.h = [UL6hp~(kh ~ - l ) ] - ' /~ (3)

where N is the total number of fibres in the composite, L is the specimen length, 8h is the so-called ineffective length, kh is the stress concentration factor (SCF) in the n e a r e s t LE fibre in the hybrid composite and p and q are the Weibull parameters of the LE fibre obtained from the Weibuli distribution function. Although the Weibull distribution function is not a perfect description of fibre strength, it is mathematically convenient and has been proven to describe the tensile strength of brittle fibres such as glass and carbon fairly well.

In the case of non-hybrid composites the failure strain is calculated as

ez = [2NLrp2(k q - I)] ,/2q (4)

being the lower bound of the failure strain; k and & are the SCF and ineffective length for the LE composite respectively. The factor two appears because there are twice as many LE fibres in the alI-LE composite than in the hybrid. Since the hybrid effect is the enhancement of the initial failure strain which corresponds to the failure strain of the LE fibre, the hybrid effect may be calculated as:

Y

C

PE

C

PE

c (0)

] PE(1)

I c (2) ] PE (3)

] c (41

Fig. 2 Model for the analysis of hybrid composite

g =E2.= [~.(k.~- 1)] -,/2. f e2 [2EKk~- l ) ] (5)

To calculate the hybrid effect R~, only the values for &, 8h, k, kh and q have to be known. Thus the hybrid effect Rc depends on the SCFS, the ineffective lengths and the Weibull parameter of the LE fibre.

St ress c o n c e n t r a t i o n s a n d i n e f f e c t i v e l e n g t h

To predict the mechanical properties of hybrid composites, knowledge of the redistribution of stress near an initial fibre breakage is necessary. Hedgepeth 35 was the first to report such a stress redistribution around fibres in a UD single fibre composite. In his analysis the SCF in the fibre adjacent to a broken fibre was derived using a shear-lag theory. In the case of hybrid composites the effect of a fibre breakage on the stress distribution is even more complicated. Stress concentrations in hybrid composites have been calculated by Zweben ~4 in which the SCF was calculated by modifying his analytical method for conventional composites 3°. Another study was adopted by Fukuda and Chou who studied the SCF in an intermingled hybrid sheet containing only one discontinous fibre 36.

Various models have been proposed to describe the ineffective length in hybrid composites and many differ in the precise definition of &. Rosen 29, for example, defined the ineffective length as the length from a fibre end in which the average axial stress is smaller than ~b (~0.9) times the stress which would exist in infinite fibres. Zweben 14,3° and Fukuda 32 gave approximate solutions for the ineffective length, in which the actual fibre stress distribution is replaced by an equivalent step function.

This paper examines the load redistribution in a hybrid composite sheet using numerical methods. A micromechanical analysis using the finite element program MARC is performed on a hybrid UD laminate having an initial LE fibre fracture. Results of such an analysis can be incorporated into statistical strength models as described above and given by Equations (3), (4) and (5). The hybrid model is shown in Fig. 2 with LE and HE fibres in alternating positions, resembling an intermingled fibre arrangement. Tensile load is applied along the fibre direction and the problem was considered as one of plane strain. Principal directions of the orthotropic plate were aligned in two lines of symmetry

COMPOSITES. NUMBER 1 . 1993 21

II II I i

Ih r

I I I [

i x Fig. 3 Finite element mesh

Table 1. Material parameters of fibres and matrix

r

L

- - - m -

~ - m = .

Material Ex, Eyy (GPa) (GPa)

V x y V~z Gx~ G~z (GPa) (GPa)

HS-carbon 230 20 HP-PE 80 2 Epoxy 3.2

0.013 0.23 20 8 0.010 0.23 0.8 0.7

0.2 1.23

and therefore only one quadrant of the plate was considered. Fig. 3 shows the mesh used, consisting of a total of 700 elements, of which 420 elements are of the fibre. A special crack-tip element was used to connect the broken fibre end with the matrix, in order to describe the singularity at that point. Perfect adhesion and equal fibre diameters for the LE (carbon) and HE (HP-PE) fibres are assumed. The interfibre distance is chosen such that the total fibre volume equals 50%. Material parameters being considered are listed in Table 1.

Fig. 4 gives the longitudinal stress in the fibres at the plane of symmetry (x = 0) at a strain level of 1.5% for the plain carbon (Fig. 4(a)) and hybrid composite (Fig. 4(b)). In analytical models a constant longitudinal stress over the cross-section of the fibre is assumed. However, the numerical results indicate that for fibres close to a broken fibre this is incorrect. In these fibres the SCF varies along the thickness of the fibre. Therefore in our analysis the SCFS are taken as an average.

The SCFS and ineffective lengths of the broken carbon fibres are listed in Table 2. In our case we used Rosen's definition for the ineffective length, i.e., the load transfer length necessary to reach an arbitrary stress of 0.9 times the undisturbed stress. The numerical results also show

that in a hybrid composite of carbon and HP-PE, the SCF in the HP-PE fibre next to a broken carbon fibre is higher than that in a composite consisting of only HP-PE fibres, indicating that, after fracture of the carbon fibres, HP-PE fibres are more sensitive to fracture in a hybrid than in a plain HP-PE composite. The calculated SCFs in the fibres next to a broken fibre in plain carbon and HP-PE composites are in both cases about 1.33 and are similar to the SCF value calculated by Hedgepeth 35. However, in a hybrid composite we have to focus on the SCFS in the LE fibres, which is 1.08 for the nearest carbon fibre in a carbon-HV-aE hybrid composite with an initial carbon fibre fracture. A good agreement between the numerical results shown here and the analytical SCFs as calculated by Fukuda is obtained if a stiffness ratio (EHE AHE/ELE ALE) of 0.35 is used, which is a typical value for HP-PE and HS-carbon.

It can be shown that the high peak stress that is observed in the nearest HP-PE fibre (fibre 1 in Fig. 4(b)) is related to the low shear modulus of this highly anisotropic fibre. Similar analyses for a hybrid system based on carbon and (isotropic) glass fibres resulted in a lower peak stress in the nearest glass fibre, although the average SCF in this glass fibre was comparable to the SCF of fibre 1 as listed in Table 2 and calculated for the carbon-HP-PE system.

22 COMPOSITES. NUMBER 1 . 1993

a

D

6 --

LI --

2 --

Ik Ix

,%

- 1 r 1 ' ~ r . . . . . . . . I I J 1 I

I I I ~ I I I ; I I i I I I I I I I I I I I I I I I

c,) [,,.J c2) k.J c3) I..-J c,,)

I I

I t ,

t/1

0

b

Fig. 4

6 P

m

N

m

I q I \ \ I I I x I I

" - I I I . . . . . "3 I I f "t Lj .' c,) I , , I (1) ~,,. j (3) L - J (q)

Longi tud ina l f ibre stress at x = 0: (a) plain carbon compos i te (e = 1.5%); and (b) ca rbon -HP/PE hybrid compos i te (e = 1.5%)

- I ! I I

T a b l e 2. S t r e s s c o n c e n t r a t i o n s a n d I n e f f e c t i v e l e n g t h

Fibre no.

(0) (1) (2) (3) (4)

Stress concentrations Plain carbon Plain HP-PE Hybrid

Ineffective/ength &/d &/d

- 1 .32 1 .07 1 .04 1 .03 - 1 .33 1 .05 1 .03 1 .02 - 1 .77 1 .08 1 .04 1 .03

14 21

Using the data of Table 2, we now can calculate the enhancement of the failure strain using Equation (5). A value of RE of 1.12 and 1.19--i.e., a hybrid effect of 12% and t 9 % - - results from analysis on bundle and single filament level respectively, using different Weibull moduli for a bundle of carbon fibres and for single carbon fibres (20 and 7 respectively14,3~).

Fatigue modelling

The fatigue behaviour of a material is generally represented by an SIN relation. The simplest form of this empirical relation is generally semi-logarithmic:

S = a + b logN (6)

where S represents the maximum value of the cyclic stress, N the number of cycles and a and b material constants. Generally, the properties of a hybrid composite are considered to be an average of those of the parent composites. Consequently, the fatigue behaviour of a hybrid is also often weighted by a ROM relation. In virtually all fatigue studies on hybrid composites the appearance of a hybrid effect was related to this ROM behaviour ~7-25. However, similar to the tensile behaviour of hybrid composites, in tension-tension fatigue ROM

behaviour is also less meaningful and thus a 'constant strain' type of behaviour should be adopted rather than ROM. In a hybrid composite, a fatigue damage process will occur in both types of fibre at strain levels close to the failure strain of the LE fibre. Since fibre fracture during fatigue in a composite is strongly governed by the applied strain level, as noted by Talreja 37, this type of fatigue damage will manifest itself mainly in the LE fibre. By selecting strain instead of stress as an independent variable, Talreja suggested a basic pattern of the fatigue life diagrams for composites which include the basic fatigue damage mechanisms in UD composites such as fibre fracture, interfacial debonding, matrix cracking and interfacial shear failure.

Based on a constant strain model, both hybrid and LE fibre composites should exhibit the same e/log N curve (Fig. 5(a)). Deviation from such a constant strain fatigue diagram can now be regarded as synergistic or hybrid effects as result of the 'crack arrest' phenomenon as described earlier. This strain/life diagram can easily be converted into the expected stress/life behaviour using

COMPOSITES. N U M B E R 1 . 1993 23

E

tn

E 3 E

a Log N

/ Teflon disc

I I

Stainless steel disc

Fig. 6 Assembly of rings forming a mandrel for winding UD com- posite rings

VHE

b Log N

Fig. 5 Constant strain based fatigue life dmgrams: (a) strain vs. cycles; and (b) stress vs. cycles

the stiffness of the hybrid, which follows a ROM relation 38 and the linear elasticity theory (Fig. 5(b)). It can be shown that the slopes for the hybrid composites in the constructed S I N curves are less than that for a plain LE

composite. When considering such a constant strain fatigue model, the results of Dickson et al. 24 for carbon- glass hybrids can easily be explained. The authors stated that their hybrid composites showed an 'unexpectedly' lower slope than that of plain carbon laminates. In fact, the conclusions of Dickson et al. concerning synergistic effects are a misinterpretation of a constant strain fatigue model and the use of an e/log N diagram to elucidate synergistic effects. They interpreted similarities in strain/ life behaviour for both plain carbon composite and carbon-glass hybrids as an indication for synergistic effects.

A further implication of such a constant strain fatigue model is that, generally, in a plot of fatigue stress vs. composition, the ROM should not be used to interpret data with respect to synergistic effects but a diagram similar to that of Fig. 1 should be adopted.

EXPERIMENTAL DETAILS

Materials

UD composites were fabricated using a common epoxy resin from Ciba-Geigy (Araldite LY556/HY917/DY070) reinforced with HP-PE fibres supplied by Allied Signal

Fig 7 Winding frame and open-ended mould for manufacture of UD composites

(Spectra ~ 1000, 650 denier) and a surface-treated high- strength carbon fibre (XAS/3K) from Courtaulds plc. To study the effect of improved adhesion of HP-PE on the performance of the hybrids, composites incorporating untreated and chromic-acid treated fibres 39,4° were used.

Two hybrid geometries have been studied: layered sandwich constructions with an HP-PE core and carbon skins and tow-to-tow intermingled hybrids with a high degree of dispersion. As a reference plain carbon and HP- PE composites were also prepared. In the HP-PE composites as well as in the hybrid composites, both untreated and chromic-acid treated HP-PE fibres were used. All composites had a volume percentage of fibres of approximately 55% and in the case of hybrids a carbon-HP-PE fibre ratio of 60/40 was selected.

As well as UD laminates, filament-wound (NOL) rings were also manufactured. These rings were prepared by winding HP-PE and/or carbon fibres unidirectionally on a mandrel of stainless steel discs (Fig. 6). To e n s u r e demoulding, the discs were covered with release tape. All rings had an inner diameter of 145 mm and a width of 6 mm. The thickness of the rings was approximately 1.2 mm. Layered sandwich hybrid rings were manufactured by alternately winding carbon and HP-PE yarns. Intermingled hybrids were fabricated by simultaneous winding of carbon and HP-PE yarns. Both types of hybrid had equal numbers of HP-PE and carbon tows. Consequently, the total cross-sectional area of the fibres in all composites is the same, minimizing the scatter in strength within one type of composite. After winding, the rings were cured for 4 h at 80"C while rotating in an air- circulating oven and subsequently post-cured for 12 h at 110*C.

UD laminates were fabricated by wet winding of fibres on a frame (Fig. 7). After winding, the whole set-up was degassed in a vacuum oven at a pressure of 400 mm Hg at 60"C to obtain a void-free composite. Laminates were

24 C O M P O S I T E S . N U M B E R 1 . 1993

Table 3. Tensile test data for unidirectional composites

Material Modulus, E (GPa) Tensile strength (MPa) Failure strain (%) Hybrid effect (%)

Single fibre composite Carbon 126 1750 1.41 HP-PE

(untreated) 40 910 3.10 HP-PE

(treated) 42 1070 3.30

Hybrid composite Sandwich

(untreated) 91 1280 1.41 Sandwich

(treated) 92 1300 1.43 Intermingled

(untreated) 88 1360 1.53 Intermingled

(treated) 87 1460 1.69

0

1.4

8.5

19.8

prepared by positioning the frame in the open-ended mould and placing the whole assembly in a hot-press. Again a release film was used between mould and composite. The laminate thickness was controlled by metal stop strips of the required thickness, while during hot-pressing the excess resin was squeezed out through the open ends of the mould. During curing at 80°C for 4 h, a pressure of 4 bar was maintained. After curing, the mould was removed and the laminates were post-cured for 12 h at 110°C.

In the case of sandwich hybrids, the laminates were manufactured in a two-step process. First, the impregnated HP-PE fibres were wound on the frame, subsequently degassed and partly cured for 15 min at 80°C in a hot-press. After B-staging, carbon fibres were wound on this HP-PE 'prepreg laminate' following the same procedure as described above. This two-step process resulted in a three-layer hybrid construction with the HP-PE fibre component sandwiched between two carbon fibre-reinforced plies and a uniform thickness of the HP-PE core with few carbon fibres migrating into the HP-PE laminate. This manufacturing procedure resulted in nearly void-free, high quality, flat UD laminates of dimensions 160 x 300 mm. Laminate thicknesses for the single fibre composites and hybrid composites were 0.4 and 0.8 mm respectively.

Tensile specimens of dimensions 12.5 x 185 mm were cut from these composite plates using a diamond cutting- wheel. Aluminium end-tabs of 30 mm length were adhesively bonded to the specimens, leaving a test length of 125 mm.

Testing

Mechanical properties of the hybrids were determined by tensile tests at room temperature. Tests were performed on a Zwick 1474 universal testing machine equipped with an extensometer. Each specimen was loaded monotonically to failure at a cross-head speed of 0.5 mm min-L

The accumulation of damage during testing was monitored using an acoustic emission system (PAC 8900 Locan AT, Physical Acoustics Corp) including a 40 dB pre-amplifier (PAC 1200A). AE was monitored with a piezoelectrical transducer (PAC R 15) coupled to the specimen using vacuum grease. The AE system allowed the AE event to be analysed with respect to different characteristic parameters such as events, counts, peak amplitude, duration, rise time and energy.

Tension-tension fatigue tests were carried out on a Zwick Rel servo-hydraulic machine at a frequency of 5 Hz and a stress ratio, R(O'mm/O-max) , of 0.1. The system was operating under load control, applying a harmonic tensile stress with a constant amplitude. To avoid clamping problems, such as shear failure within the tabs in the case of plain HP-PE composites due to their low shear strength, fatigue tests were performed on filament- wound NOL rings using a split-disc loading device (ASTM D 2291). To minimize wear of the composite a PTFE-based lubricant was used. Damage development during fatigue using modulus reduction, ultrasonic C- scan and AE was also evaluated on laminated plates of plain carbon and hybrid compositions.

RESULTS AND DISCUSSION

Tensile behaviour

The results for the four hybrid systems and their parent composites as obtained from the tensile tests are presented in Table 3 and are an average of at least five test results. The experimental accuracy of all quoted data is within 10%. In all hybrids, no multiple fracture was observed; i.e., the stress/strain curves were all linear up to total fracture of the hybrid. From the failure strain data, it can be seen that an enhanced first failure strain was observed in the case of intermingled hybrids, whereas for sandwich hybrids no strain enhancement was detected. These observations are in good agreement with previous work on carbon glass ~5. Values for hybrid effects found in the present work are also in good agreement with values quoted in Reference 39 despite the fact that

COMPOSITES. NUMBER 1 . 1993 25

"O t_

.Q

-1-

40

30

20

10

0 100

F i b r e ? u n d l e

20 40 60 80

Volume 96 HP-PE

Fig. 8 Hybrid effectsvs, hybnd composition: ©, untreated HP-PE2; 0 , treated HP-PE2; V, sandwtch (untreated); V, sandwich (treated); A, intermingled (untreated); &, intermingled (treated)

different materials were used. Furthermore, it should be noted that the experimental values for the hybrid effect are in reasonable agreement with the theoretical predictions based on statistical models including SCF effects 32. Fig. 8 shows the hybrid effect vs. hybrid composition, including the data from Reference 2. The dotted and full lines are the theoretical predictions based on Fukuda's model using either fibre bundle or single filament Weibull parameters for the carbon fibre.

From this part of the study it can be concluded that, under monotonic tensile loading, hybrid effects are observed for tow-to-tow intermingled hybrids incorporating both untreated and treated HP-PE fibre bundles, whereas no hybrid effects are observed for the sandwich constructions.

Acoustic emission

Acoustic emission monitoring was used for two reasons: firstly to gain more insight into the onset of damage in the different types of hybrid compared with plain carbon composite; and secondly, to discriminate between different failure processes in hybrid composites such as carbon fibre breakage, HP-PE fibre breakage, matrix cracking, debonding and delamination. In Fig. 9 the cumulative plots of events vs. strain are shown for both untreated and treated HP-PE composites. The AE events, observed during tensile loading up to failure of the composite, increased in number as the test progressed. For composites incorporating untreated fibres, initial emissions occurred almost immediately upon loading, with a relative gradual increase in event rate with increasing load. Composites incorporating treated fibres showed an onset of emissions starting at about 20% of maximum strain, with a rapid increase at approximately 35% of maximum strain. Obviously, improved adhesion has a positive effect on the onset of damage in HP-PE composites.

Fig. l0 gives the total activity plots for the plain carbon and hybrid composites when loaded up to about 5 kN. From these plots it can be concluded that the onset of damage is approximately the same for the plain carbon composite and the intermingled hybrids. However, in particular, the sandwich hybrid with untreated fibres was consistently noisier than the other composites,

E

> UJ

10000

8000

6000

4000

2000

0 4

Treated

reated

L / I I I 1 2 3

Strain (g)

Fig. 9 Cumulative AE plots for HP-PE/epoxy composites

¢D >

UJ

10000

8000

6000'

4000'

2000'

4O G p ,/ l / l®

/ ,,y

50 100 150 200 250

Time (s)

Fig. 10 Cumulattve AE plots for plain carbon and hybrid compo- sites: (1) plain carbon; (2) sandwich (untreated); (3) sandwich (treated); (4) mtermingled (untreated); and (5) intermingled (treated)

presumably due to extensive debonding within the HP-PE component and/or delaminations at the carbon/HP-PE interface. It seems that a reduction of the total carbon/ HP-PE interface scale accelerates the initiation of damage processes such as debonding and interlaminar shear failure.

Amplitude distribution analysis of AE signals has shown itself useful in discriminating between different fracture mechanisms in composite materials 41,42. Generally, at the initial stages of tensile testing, low-amplitude emissions are associated with matrix crazing or cracking, whereas fibre fracture is characterized by high-amplitude emissions.

Fig. 11 shows the distribution of the peak amplitudes of the AE signals at various load levels for composites incorporating treated HP-PE fibres; at low loads (0-65%) a concave-shaped curve is observed with most amplitude events at approximately the threshold level of 30 dB, indicating debonding. At higher loads (65-100%), a

26 COMPOSITES. NUMBER 1 . 1993

150

0

400 65-80~

0 300'

0 20 40 60 80 100

Amplitude (dB)

Fig. 11 Amplitude distnbution AE characteristics of treated HP- PE/epoxy composite at different load levels

typical case of a double or even triple-peaked amplitude distribution occurs as would be expected in the normal case of fibre fracture in combination with matrix crazing and/or interface failure. Composites with untreated HP-

PE fibres showed even at high load levels an amplitude distribution signature similar to that of treated HP-PE composites at low load levels (0-65%) with no clear double-peak and only a few high amplitude events in the range 70-80 dB. Apparently, the number of fibre fractures in this type of composite is less than that in the case of treated fibres. Presumably, the large debonding length for untreated HP-PE fibres prevents extensive multiple fibre fractures within one filament, since no fibre fracture will occur over this ineffective length. Mercx e t

al . 4° reported for a similar chromic acid treatment an increase in pull-out strength from 0.3 MPa to 1.9 MPa. This increase in pull-out strength (with a factor of 6) is indicative of a decrease in ineffective length after surface treatment, enlarging the probability of multiple fibre fractures and leading to higher effectiveness of the fibre with respect to composite strength 43.

4ooL 80-100t

> LU

6 0 - 8 0 t

..3" ~,a~, 0

500~'- 40-60g

° J i lOOl 20-40g

I I 0 20 40 60 80 100

Amplitude (dB)

Fig. 12 Amplitude distribution AE characteristics of carbon/epoxy composite at different load levels

Amplitude distribution acoustic emission analysis for carbon/epoxy composites is shown in Fig. 12. Here also a typical case of a double-peaked amplitude distribution is observed. However, in contrast to HP-PE signatures, the low-amplitude peak shifts with increasing load completely to higher amplitude levels. Between 60-80% of maximum load a distinct second peak or shoulder is formed at approximately 50 dB. This becomes much stronger at higher loads suggesting that at these load levels a significant amount of fibre fracture occurs. At the higher load levels (80-100%) the low-amplitude peak totally disappears, indicating that fibre fracture is now the dominating fracture process. This peak shifting also indicates that less debonding and delamination is taking place in the carbon fibre composite than in the HP-PE

composites, where at high load levels low-amplitude signals were still predominant. This assumption was confirmed by scanning electron micrographs of fracture surfaces, which showed that even in the case of treated HP-PE fibre composites, numerous fibre pull-outs occurred 2, whereas the carbon fibre composite failed in a

C O M P O S I T E S . N U M B E R 1 . 1 9 9 3 27

100

0

c

m 200

90-I00~

6 0 - 8 0 ~

o 0-403

300

I I 20 qO 60 80 100

Ampl i tude (dB)

Fig. 13 Amplitude distribution AE characteristics of carbon-HP- PE intermingled hybrid composite at different load l e v e l s

rather brittle mode with a negligible amount of debonding and pull-out. It is interesting to note that the high-amplitude events associated with carbon fibre fracture (50 dB) are lower than that for HP-PE (70-80 dB), indicating that the latter fibres exhibit higher energy acoustic emissions. This observation is in accordance with the work-to-break of both fibres. Since the tenacity of both fibres is about the same, the fibre fracture energy is expected to relate to the failure strain, which is for HP- PE about twice that of carbon fibre.

Amplitude distribution signatures of the various hybrid configurations were all more or less alike (Fig. 13) and indicated that in all hybrid composites, in the latter part of the test, debonding and/or delamination is a more prominent failure process than in the case of plain carbon composites, since even at high load levels a distinct low-amplitude signal is evident. However, similar to the plain carbon composites there is some indication of the development of a second peak at 50 dB, suggesting carbon fibre fracture at higher load levels.

2.50

2 .00

~ 1.so

E 1.00

E ×

:~ 0 .50 o

0 . 0 0 ~ ~ i ~ 0 1 2 3 4 5 6

Log N

Fig. 14 Stress/life curves for parent composites: + , carbon; ©, untreated HP-PE; Q, treated HP-PE

Since there are no significant 70-80 dB signals it seems i that no HP-PE fibre fracture is taking place which is not surprising, since these fibres have a failure strain of more " than twice that of carbon.

Fatigue behaviour

Fig. 14 shows that the S/log N curves for plain carbon and HP-PE composites incorporating untreated and chromic-acid treated fibres. Although there is some reduction in static composite strength as a result of a reduction in tensile strength of the H P - P E fibres after chromic acid treatment 4°, the S/log N curves for plain HP-

PE composites clearly demonstrate the advantage of improved adhesion on the fatigue performance of these composites. The slopes of the S/log N curves of untreated and treated HP-PE composites are approximately 230 MPa and 77 MPa per decade respectively. The slope for the treated HP-PE/epoxy composite is even flatter than that for plain HS-carbon composite, illustrating the excellent fatigue resistance of this type of composite. After surface treatment the fatigue performance of HP-PE composites becomes more fibre-dominated rather than interfacial-dominated.

Since the testing of filament-wound rings may have some disadvantages with respect to the non-uniformity of the strain in the ring, the possibility of stress concentrations at the edges of the split discs and frictional wear of the composite, comparative experiments were carried out on UD laminates fabricated from prepreg (Fibredux 913C/ XAS-5-34). Fatigue performance of these high-quality laminates was approximately the same as that of the carbon composite rings, with even a small advantage for the rings with respect to better fatigue performance, presumably due to a better stress introduction and/or size effects. Consequently, the rings were considered as representative test samples.

The fatigue data for the various hybrid composites are presented in Figs 15(a) and 15(b). Two important observations can be made from the S/log N curves of the hybrids. Firstly, it should be noted that the slopes of the S/log N curves for all the hybrids are lower than that of the plain carbon composite and, secondly, the slopes for the hybrids are all more or less the same, with a positive exception for the intermingled hybrid incorporating

2 8 . C O M P O S I T E S . N U M B E R 1 . 1 9 9 3

2.50

,_, 2.00

L3 o~ 1 .50 o~

E 1.00

:~ 0 . 5 0

0 .00

Ill

2.50

I I I I I

1 2 3 4 5

Log N

2 . 0 0 g

~ 1.50 m

E 1.00 E X

:~ 0.50

0.00 , i i , , 1 2 3 4 5

b Log N

F=g 15 Stress/life curves for hybrid composites: (a) sandwich hybr ids--V, untreated HP-PE; T, treated HP-PE; and (b) intermin- gled hybr ids~A, untreated HP-PE; &, treated HP-PE

treated HP-PE fibres. The fatigue parameters a and b for the various composites, adopting the semi-logarithmic S/ N relation (Equation (6)), are given in Table 4. Using the experimentally obtained fatigue parameters for a plain carbon composite, Equation (I) for the calculation of the 'constant strain' fatigue stress of the hybrid and R O M for the modulus, one can determine the theoretical values for the fatigue parameters of the hybrid. In the case of a viscoelastic fibre such as HP-PE we have to consider strain-rate effects. Therefore, in our calculations we used a value of 100 GPa for the modulus of HP-PE fibre at 5 H Z 7, instead of the 80 GPa as listed in Table 1. Predicted values for the slope b of the hybrid composite, expecting constant strain fatigue behaviour, are in good agreement with experimentally obtained values, except for the intermingled hybrid with treated HP-PE which exhibits a significantly lower slope than predicted by the constant strain model. Furthemore, it should be stressed that the degree of scatter in the hybrids, especially in the hybrids with untreated fibres, is smaller than that for the plain carbon composites. These observations are in contrast with earlier work on ca rbon -Kev la f 22 and carbon- glass 24, where more scattering in the fatigue life of the hybrids was observed than in that of the parent composites.

We can also compare the behaviour of the carbon and hybrid composites using a plot similar to that of Fig. 5(b). By extrapolating the data to o-,,a~=0 deviations from the constant strain model are observed, since in

Table 4. Experimental and predicted values of fatigue constants (Equation (6))

Material Intercept, a Slope, b (GPa) (GPa)

Single fibre composite Carbon 2.21 5 - 0.157 HP-PE

(untreated) 1.81 6 - 0.230 HP-PE

(treated) 1.397 - 0.077

Hybrid composite Sandwich

(untreated) 1.406 - 0.089 Sandw ich

(treated) 1.558 - 0.095 Intermingled

(untreated) 1.574 - 0.114 Intermingled

(treated) 1.438 - 0.051

Predicted (constant strain model) 1.638 - 0.116

2.50

2.00 #_

1.50

E 1.00 E ×

:~ 0.50

0.00

o n

Hybrids \ \ x~ \~ ~ ' ~ - " " ~ "~'~ ~Jl ntermingled hybrid

0 5 10 15 20 25 30

Log N

Fig 16 Extrapolated stress/hfe curves indicating synergistic fati- gue life for the intermingled hybrid with treated HP-PE

such a model all lines should have the same intersection with the x-axis (o-.~ax = 0). From the extrapolated curves in Fig. 16 it can be seen that all curves intersect the x-axis at about log N = 15, except the curve for the intermingled hybrid with treated HP-PE which intersects at approximately log N=28 , illustrating the synergistic fatigue life for this type of composite. The authors hasten to add that extrapolation of the fatigue data up to O'max = 0 and ignoring the existence of a fatigue limit has no physical relevance. However, in this context the point of intersection is only used as a convenient way to compare the slopes of the S/log N curves.

When we convert the stress data to strain data, by using the relevant moduli, and plot an e/log N diagram, the slopes for the hybrids can directly be compared with that of the plain carbon composite (Fig. 17). With respect to the interpretation of such diagrams, it should be noted

C O M P O S I T E S . N U M B E R 1 . 1993 29

2.00

~" 1.50

c ~_

1.00 E E X

0.50

~ .~bon

~ - - - ~ - - - ~ - - . ~ In terming led hybrid .... C~ .... ~ ~ .......

0.00 i t , , i 0 1 2 3 q 5 6 7

Log N

Fig. 17 Strain/l i fe curves indicat ing synergistic fatigue life for the intermingled hybrid wi th treated HP-PE

that only small changes in slope result in large effects in fatigue life since a logarithmic scale is used.

From the data listed in Table 4 and Fig. 16 as well as Fig. 17 it can be concluded that all hybrids follow the constant strain fatigue model, except the intermingled hybrid with treated HP-PE fibres which exhibits synergistic or hybrid effects with respect to fatigue performance.

Fatigue damage characterization

To investigate the effect of the HP-PE fibres and the surface treatment of these fibres on the crack propagation in the hybrids, damage development was studied using modulus reduction measurements, ultrasonic C-scan and AE. For experimental convenience, in preference to the filament-wound rings, damage characterization was performed on flat laminates, similar to the test specimens used in the tensile test. Since we are particularly interested in the damage evolution of the hybrids relative to that of plain carbon composites, we only considered these composites. When comparing the damage state in the hybrid composites with that in a plain carbon composite, all tests have to be carried out at equal strain levels. In these experiments an initial maximum strain level (e-max) of 1.10% and R = 0.1 were used.

The essential first step in fatigue characterization of composites is to measure changes in the stiffness properties of the laminate. However, similar to what is generally observed for UD composites 18, all materials show excellent fatigue resistance since the damage which occurs during fatigue (such as debonding, delamination and fibre fracture) has no or little influence on the stiffness. Consequently, differences in intermittently measured moduli are within experimental error and moduli remain essentially constant over their fatigue life.

To study the effect of improved adhesion of the HP-PE

fibres on the accumulation of debonding and delamination in the hybrid composites, ultrasonic C-scan was used. Fig. 18(a)--(c) shows some of the extreme results: plain carbon, sandwich hybrids with untreated HP-PE and intermingled hybrids incorporating treated HP-PE respectively. All samples were fatigued at the same initial strain level under load control. The plain carbon

0 cycles

200 cycles

400 cycles

8

0 cycles

400 cycles

800 cycles

b

0 cycles

1000 cycles

2000 cycles

¢

Fig. 18 Ultrasonic C-scans at different number of cycles of: (a) plain carbon composite; (b) untreated sandwich hybrid; and (c) treated intermingled hybrid

laminate showed little detectable damage in the form of delaminations. Only after 400 cycles is a small edge delamination observed. In all hybrids, especially in the sandwich construction with untreated HP-PE (Fig. 18(b)), there is clear evidence of extensive debonding and delamination. However, in the case of intermingled hybrids with treated fibres there is no sign of damage, even at higher cycle numbers (Fig. 18(c)).

Table 5 gives the number of acoustic emission events until fracture for the plain carbon and hybrid composites using a threshold level of 40 dB. The threshold was chosen at 40 dB to reduce background noise and to concentrate on the high amplitude signals, which are indicative of fibre fracture. The lowest number of 60 dB AE signals was measured for the intermingled hybrid with treated HP-PE, suggesting relatively little severe fatigue damage in this type of composite. The highest number of low-amplitude (40 dB) signals manifested itself in the hybrids with untreated HP-PE fibres, presumably due to extensive debonding and delamination. These findings are in accordance with the fatigue life diagrams, which showed the best fatigue

30 COMPOSITES. NUMBER 1 .1993

Table 5. Number of AE events during fatigue

Material 40dB 60 dB

Carbon 1300 600 Sandwich hybrid

(untreated) 6600 800

~ Sandwich hybrid (treated) 2400 500

intermingled hybrid (untreated) 4600 1400

Intermingled hybrid (treated) 11 00 1 20

per fo rmance for the in termingled hybr ids with t rea ted HP-PE.

CONCL USlONS

The mechanica l p roper t i e s o f UD hybr id compos i tes based on ca rbon and HP-PE fibres have been s tudied under bo th m o n o t o n i c and cyclic tensile loading. Two different hybr id conf igura t ions have been studied; c o r e shell sandwich hybr ids and tow- to - tow in termingled hybrids . Both hybr id types were manu fa c tu r e d incorpora t ing ei ther un t rea ted or chromic-ac id t rea ted HP-PE fibres, resul t ing in a to ta l o f four hybr id systems evaluated . Resul ts ind ica ted that the existence of synergist ic or hybr id effects depends on bo th hybr id design and the interfacial b o n d s t rength o f the HP-PE fibres. U n d e r mono ton i c tensile loading, posi t ive devia t ions f rom the cons tan t s train behav iou r were observed for bo th types o f in termingled hybr id , with the highest hybr id effect for the in termingled system with surface- t rea ted HP-PE fibres. N o hybr id effects were observed for the sandwich hybr ids . Obvious ly , the poss ibi l i ty o f crack arrest due to the presence o f HP-PE fibres, prevent ing rap id crack extension f rom ini t ial ly failed ca rbon fibres, d iminishes when these fibres are not highly d ispersed t h r o u g h o u t the ca rbon fibre composi te .

A E studies on hybr id and pa ren t compos i tes dur ing m o n o t o n i c tensile load ing also indica ted bet ter pe r fo rmance for the in termingled hybr ids with respect to the onset o f d a m a g e and showed no evidence o f HP-PE fibre b reakage in the hybr id .

Fini te e lement mic romechan ica l ca lcula t ions were pe r fo rmed on a hybr id mode l to calcula te stress concen t ra t ions in carbon-HP-PE hybr id composi tes , which can be used to calcula te hybr id effects using stat is t ical models . Us ing such a model , hybr id effects could be descr ibed reasonab ly well.

A fatigue model , based on the conven t iona l "constant strain" model for the tensile s t rength o f hybr id composi tes , is a d o p t e d and descr ibes the fat igue behav iour o f all hybr ids fair ly well. Synergist ic effects under fat igue load ing cond i t i ons - - i . e . , posi t ive devia t ions f rom the expected fat igue behav iou r based on such a m o d e l - - w e r e observed for the in termingled hybr ids with t rea ted HP-PE. These hybr ids also showed the lowest level o f fat igue damage as charac te r ized by ul t rasonic C-scan and acoust ic emission.

A CKNO WL EDGEMENTS

The au thors are indeb ted to A n d r e w Rus t idge and Inge Sedeijn for pe r fo rming the acoust ic emiss ion exper iments and to F red Re in ink for car ry ing ou t the finite e lement analysis . This work was suppo r t ed by the Du tch Min is t ry o f Economic Affairs under g ran t I O P - P C B P 2.2.

REFERENCES

1 Peijs, A.A.J.M and Lemstra, P.J. 'Hybrid composites based on polyethylene and carbon fibres. Part 1: Compressive and impact behaviour' in Integration of Fundamental Polymer Science and Technology, vol 3 edited by Lemstra and Kleintjes (Elsevier Applied Science, London, UK, 1989) pp 218 227

2 Peijs, A.A.J.M., Catsman, P., Govaert, L.E. and Lemstra, P.J. "Hybrid composites based on polyethylene and carbon fibres. Part 2: Influence of composition and adhesion level of polyethylene fibres on mechanical properties' Composites 21 No 6 (1990) pp 513-521

3 Peijs, A.A.J.M., Ven_derbosch, R.W. and Lemstra, P.J. 'Hybrid composites based on polyethylene and carbon fibres. Part 3: Impact resistant structural composites through damage management" Composites 21 No 6 (1990) pp 522 530

4 Peijs, A.A.J.M. and Venderbosch, R.W. 'Hybrid composites based on polyethylene and carbon fibres. Part 4: Influence of hybrid design on impact strength' J Mater Sci Lett 10 (1991) pp 112~ 1124

5 Peijs, A.A.J.M. and Van Klinken, E.J. "Hybrid composites based on polyethylene and carbon fibres. Part 5: Energy absorption under quasi-static crash conditions' J Mater Sct Lett l l (1992) pp 520-522

6 Peijs, A.A.J.M., Catsman, P. and Venderbosch, R.W. 'Impact resistant and damage tolerant hybrid composite structures based on carbon and polyethylene fibres' in Composite Structures 6 edited by I.H. Marshall (Elsevier Apphed Science, London, UK, 1991) pp 585 598

7 Govaert, L.E., d'Hooghe, E.L.J.C.L. and Peijs, A.A.J.M. 'A micromechanlcal approach to viscoelasticity of unidirectional hybrid composites' Composztes 22 No 2 (1991) pp 113 119

8 Eisenman, J.R. and Kaminski, B.E. Engng Fracture Mech 4 (1972) pp 907~ 13

9 Sendeckyj, F.P. 'Concepts for crack arrestment in composites' Fracture Mechanws of Composites, ASTM STP 593 (American Society for Testing and Materials, 1975) pp 21~226

10 Hayashi, T. 'On the improvement of mechanical properties of composites by hybrid composition" m Proc 8th lnt Reinforced Plastws Con/(British Plastics Federation, Brighton, UK, 1972) pp 149 152

11 Bunsell, A.R. and Harris, B. "Hybrid carbon and glass fibre composites" Cornposltes $ No 4 (1974) pp 157 164

12 Phillips, L.N. 'On the usefulness of glass-fibre~:arbon hybrids" Brtttsh Plastics Federation Congre,~s (Brighton, UK 1976) paper 21

13 Aveston, J. and Sillwood, J.M. "Synergistic fibre strengthening in hybrid composites" J Mater Set l l No 10 (1976) pp 1877 1883

14 Zweben, C. 'Tensile strength of hybrid composites' J Mater Sci 12 No7(1977) pp 1325 1337

15 Manders, P.W. and Bader, M.G. 'The strength of hybrid glass/ carbon fibre composites Part 1: Failure strain enhancement and failure mode" J Mater Set 16 (1981) pp 2233-2245

16 Marom, G., Fischer, S., Tuler, F.R. and Wagner, H.D. "Hybrid effects m composites, conditions for positive or negative effect versus rule-of-mixture behaviour" J Mater Sei 13 No 7 (1978) pp 1419 1426

17 Phillips, L.N. 'The hybrid effect~loes it exist '" Composites 7 (1976) pp 7 8

18 Hofer, K.E., Rao, N. and Stander, M. "Fatigue behaviour of graphite/glass/epoxy hybrid composites' 2nd Int Conf on Carbon Fibres (The Plastics Institute, London, UK, 1974) pp 201 212

19 Hofer, K.E., Stander, M. and Bennet, L.C. 'Degradation and enhancement of the fatigue behavlour of glass/graphite/epoxy hybrid composites after accelerated aging' Polymer Engng Sci 18 (1978) pp 12(~127

20 Grimes, G.C. "Structural design significance of tension-tension fatigue data on composites" 4th Con/" Composite Materials Testing

COMPOSITES. NUMBER 1 .1993 31

and Design, ASTM STP 617 (American Society for Testing and Materials, 1977) pp 106-119

21 Bunsen, A.R. and Harris, B. 'Hybrid carbon/glass composites' in Proc 1st lnt Conf on Composite Materials (Metallurgical Society of AIME, Geneva and Boston, April 1975) pp 174-190

22 Fernando, G., Dickson, R.F., Adam, T., Reiter, H. and Harris, B. 'Fatigue behaviour of hybrid composites. Part 1: Carbon/Kevlar hybrids' J Mater Sci 23 (1988) pp 3732-3743

23 Adams, T., Fernando, G., Diekson, R.F., Reiter, H. and Harris, B. "Fatigue life predictions for hybrid composites' Int J Fatigue 11 No 4 (1989) pp 233~37

24 Dickson, R.F., Fernando, G., Adam, T., Reiter, H. and Harris, B. 'Fatigue behaviour of hybrid composites, Part 2: Carbon glass hybrids' J Mater Sci 24 (1989) pp 227 233

25 Marom, G., Hard, H., Neumann, S., Friedrieh, K., Sehulte, K. and Wagner, H.D. 'Fatigue behaviour and rate dependent properties of aramid fibre/carbon fibre hybrid composites' Composites 20 No 6 (1989) pp 537-544

26 Sih, G.C. and Ebert, L.J. 'Effect of the interface on fatigue performance of unidirectional fibre glass composites' Composttes Sci & Techno128 (1987) pp 137-161

27 Peijs, A.A.J.M., Rustidge, A.D., de Kok, J.M.M. and Rijsdijk, H.A. 'Influence of the interface on the performance of polyethylene fibre reinforced composites' in Interfacial Phenomena m Composite Materials '91 edited by I. Verpoest and F. Jones (Butterworth-Heinemann Ltd, Oxford, UK, 1991) pp 34-37

28 Chou, T.W. and Kelly, A. 'Mechanical properties of composites' Ann Rev Mater Sci 10 (1980) pp 229-259

29 Rosen, B.W. 'Tensile failure of fibrous composites' J AIAA 2 No 11 (1964) pp 1985-1991

30 Zweben, C. 'Tensile failure of fibrous composites" J AIAA 6 No 12 (1968) pp 2325~331

31 Manders, P.W. and Bader, M.G. 'The strength of hybrid glass/ carbon fibre composites. Part 2: A statistical model' J Mater Sci 16 (1981) pp 2246-2256

32 Fuknda, H. 'An advanced theory of the strength of hybrid composites' J Mater Sci 19 (1983) pp 974-982

33 Harlow, D.G. Proc Roy Soc London A389 (1983) pp 67 100

34 Fariborz, S.J., Yang, C.L. and Harlow, D.G. J Composite Mater 19 (1985) pp 334-357

35 Hedgepeth, J.M. 'Stress concentrations in filamentary structures NASA Technical Note D-882 (1961)

36 Fukuda, H. and Chou, T.W. 'Stress concentrations around a discontinuous fibre in a hybrid composite sheet' Trans Japan Soc Composite Mater 7 (1981) pp 37-42

37 Talreja, R. "Fatigue of composite materials: damage mechanisms and fatigue-life diagrams' Proc Roy Soc London A378 (1981) pp 461-475

38 Kretsis, G. 'A review on the tensile, compressive, flexural and shear properties of hybrid fibre-reinforced plastics' Composites 18 No 1 (1987) pp 13 23

39 Laflizeskey, N.H. and Ward, I.M. 'A study of the adhesion of drawn polyethylene fibre/polymeric resin system" J Mater.Sci 18 (1983) pp 533 544

40 Mercx, F.P.M., Benzina, A., van Langeveld, A.D. and Lenmtra, P.J. "Surface modification of oriented high-modulus polyethylene structures; 1. Acid treatment' J Mater Sci (in press)

41 Valentin, D. "Failure mechanism discrimination in carbon fibre- reinforced epoxy laminates' Composites 14 No 4 (1983) pp 345- 351

42 Short, D. and Summerscales, J. 'Amplitude distribution acoustic emission signatures of unidirectional fibre composite hybrid materials' Composites 15 No 3 (1984) pp 20(L206

43 De Kok, J.M.M., Rijskijk, H.A. and Peijs, A.A.J.M. 'Influence of the interface on the longitudinal tensile strength of polyethylene fibre reinforced composites' in Interracial Phenomena m Composite Materials "91 op. cit. pp 28(~281

AUTHORS

A.A.J.M. Peijs and J.M.M. de Kok are with the Centre for Polymers and Composites, Eindhoven Univeristy of Technology, PO Box 513, 5600 MB Eindhoven, The Netherlands, Enquiries should be addressed to A.A.J.M. Peijs.

3 2 C O M P O S I T E S . N U M B E R 1 . 1 9 9 3