Paper for the Localised Damage 96 conference in Fukuoka, Japan, June 1996. Organised by the Wessex Institute of Technology

Fatigue fracture mechanism of fibre reinforced injection moulded polyamide.

J.J. Horst

Laboratory for Mechanical Reliability
Faculty of Industrial Design Engineering
Delft University of Technology
Leeghwaterstraat 35
2628 CB Delft.


A fatigue mechanism of bridged cracks is proposed, based on some novel experimental techniques: Measurements of creep during fatigue, through thickness strength profiles of specimens that had been subjected to a fixed number of fatigue cycles. SEM fractography was used in the normal way on broken specimens, but also on specimens that were fatigued, but not up to failure, and consequently cryogenically broken.

Damage starts with void initiation at fibre ends, the voids consequently grow along the fibre and merge into larger "cracks". For this growth process especially the tensile strength of the fibre - matrix interface is important. However not one complete crack exists, as the crack walls remain connected at a number of spots. Damage is not homogeneously distributed over the sample, but is confined to disk-like areas, similar in appearance to crazes found in amorphous plastics.

1 Introduction

Injection moulded thermoplastics reinforced with short glass or carbon fibres (SFTP's) are being used increasingly in load bearing applications. This because of weight, cost, corrosion resistance and ease of production. This is by the injection moulding process, which makes freedom of design and integration of functions possible.

SFTP's have a characteristically high degree of anisotropy. The anisotropy, caused by fibre orientation, and formed during injection moulding, brings about strong variations in properties: e.g. heat conductivity, elastic modulus, tensile strength and fatigue behaviour. This makes the calculation of properties of products complex.

To be able to predict the fatigue lifetime of products, fatigue behaviour and Ultimate Tensile Strength (UTS) were investigated. A correlation between UTS and the maximum fatigue stress to attain a certain lifetime was found for GFPA.[1,2] The reason for this correlation is not completely understood. Mechanisms in tensile and fatigue experiments are different, such that a similar stressing of the matrix material with different fibre orientations, in both tensile and fatigue experiments, cannot be assumed. Therefore investigation of the fatigue mechanism has started.

The mechanism in tensile experiments is well known, consisting of crack initiation and growth, accompanied by fibre pull-out. In the current investigations this mechanism could be confirmed.

1.1 Theory

The mechanism in fatigue is generally considered to consist of the following four stages [3]:
1 Initiation of local damage due to cyclic deformation, generally at the locations of highest stress intensity, the fibre ends.[4]
2 Initiation of microcrack.
3 Stable crack growth due to cyclic loading. Local modes of crack extension depend on local fibre orientation, matrix ductility and the degree of interfacial adhesion (Lang).[5]
4 Fast (instable) crack growth in the last load cycle, which should be comparable to failure in a tensile test.

Dally [6] reports for a system with the much more ductile PE matrix, and almost no fibre - matrix bonding, an entirely different mechanism. Massive debonding reduces the glass fibres from reinforcement to unbonded inclusions, giving rise to a sharp drop in modulus. The greater strains are accommodated by the matrix without failure. This process of general degradation rather than a dominant crack was also reported by Mandell [7] for unnotched specimens, while Dibenedetto et al.[8] also observed a similar fatigue mechanism in compact tension specimens of graphite fiber reinforced PA 6.6, conditioned to equilibrium water content.

In this paper it will be shown that the fatigue mechanism of crack growth as described above for dry as moulded GlassFibre Reinforced PolyAmide 6 (GFPA 6) cannot be applied to conditioned GFPA 6. The mechanism Dally found for GFPE does not fit to the experimental results either. A modified fatigue failure mechanism is proposed.

2 Experimental

The material used was Polyamide 6 containing 30%wt. of glassfibres; Akulon K224-G6, provided by DSM, the Netherlands. Square plates of 100x100mm2 and 2 and 5.75mm thickness were injection moulded from this. The mould has a line gate, to obtain a straight flow front. For fatigue and tensile experiments non standard dog-bone type specimens were milled from the plates, using a Roland PNC-3000 Computer Aided Modelling Machine. Specimens were conditioned by exposing them to laboratory air for at least 1 year, giving a water content of approximately 1.5%.
The fatigue experiments were carried out on a servo-hydraulic MTS 810 bench. The load frequency used was 1Hz, to avoid temperature increases of more than 3 degrees K due to hysteretic heating. Earlier experiments [1] showed the high sensitivity of the fatigue lifetime of this particular material to the test frequency.
The minimum to maximum load ratio R was 0.1.
During the fatigue experiments the displacement of the grips was monitored, enabling the calculation of the elastic modulus and creep. Thus we were able to get a maximum of information from one fatigue test.
Tensile experiments were executed on the same type of specimens, with a cross-head speed of 50mm/min, resulting in a strain speed of 143 %/min. Tests were carried out in an environmental chamber at a temperature of 23C and at a relative air humidity of 50%.
The assessing of fatigue damage was done by making strength profiles:[2,9,10] Miniature test bars are milled from the specimens after subjecting these for a certain number of cycles to a fatigue load. Expected lifetime of these specimens can be accurately predicted using the creep speed - lifetime correlation.[1,2,9] From these miniature samples, with a width of 2 mm, 85micron thick slices are cut parallel to the surface, using a Leitz microtome. The fracture strength and fracture strain are measured for each slice, using a miniature tensile test machine. Plotting strength or fracture strain to the distance to surface gives the required profile.
SEM micrographs of fracture surfaces were made using a JEOL JSM- 840A after gold coating in a Balzers SCD 040. To reveal the structure inside the specimen during the fatigue process, some specimens were first fatigued for a certain percentage of their expected lifetime, and consequently fractured after immersing them for 5 minutes in liquid nitrogen.

3 Results

3.1 Fatigue experiments

It had already been shown [2,10] that especially the creep speed Vc (increase in F>F?F5F@FFuFtFBFvFFFFC~FFAr~ rl>t~u 6&jFM=w. dlt|d>,t96, t,6,j;(u6,j26*jj+*,j uP@tjvFMM]M XEU؃><u<w uMM]MXEU؁h" v v zPzPg)B~tj }+F~uv v hZ b tdždžv v hZ *F @F BzPhzPhjP zPzPPz t<u. u6<<wjPF<whh jj}>.t><t+>.t8wj6&hh u@ tHuSj6&hhnj6&h)h>.u*j6&h>h! Ճ>.uj6&hh >.t,f tHtY#<tMF )SwhjD F6&j(v vj4 =t~ )u,6<<wF *<whj뮋<w#>.t tHt<tj6&hPh!<wGFV uVhV‹&RPX<wj6&hih&>.tj 6&hh[>.u{6&jjjj>أ6& PF+.@P, FPv >%PD hz6->t6ljvjF uPj uv փ>أ6&g PF+.@Pv FPv >%P h6%>أ6&PF+.@PFPv >%Ph66lh'6lh&Hu 6lh(j<wFFF6^F+FF6&hvvv un6&jhPh0Zhj= wJ6&jhj6*jFPj(m6&FPjjj@u6&h,hrhMM]MؐEU؁~vvS~Cs3uf]MؐEUظW^ s++f]MؐEU؁~vvS C~ r%3f]MؐEUVWY+U ]׊_^f]MˌؐEUشG *f]MˌؐEU؋Vg *f]MؐEUVWF uZ ~@:\&G r3_^f]MؐEUW~}:u*,APt GG t״; r3_f]MؐEU؁~vvS ~A r3f]MؐEU؁W~v v S7 vvSo ~VB r3_f]MؐEUVWv~ NjN1;F u;v;sNOǨtI_^f]M 'jnox.H w d  ( Q$` * U/L }%EG9X~YeZ fXEU؃^F^?tG^? tSj .FV t^&MM]MUF^GF؋ u~2r6&jMjsPh0FFF]XEU؃^F&^&?:u-APFHt~u+MM]MU6lj t GvjF tPj uvP؋GƋ]XEU؃FFF^?t? uMM]MUV^F ^ +ɉN^^<\u |\uF3? t6?\t?:u@^ F@^^?.uvFF^?uŋ^ ^9s^] XEU؃vFPNQPPV~uv;FsvhMM]MUV^vFvVF^F ? t^F^?u~t^G^]U^?t ^?t~Pv~Pn~PP~QPP~t,~h u~PPj@uPhq=qufjjj S|=sL^?u$SPjvhvv^?uSPj@vv]XEU؃ V^vFPFPFPFP~uCF9Ft;vP~w^:u^+^v~tv+v^@^MM]MXEU؃ vFPNQQNQ,~tFFFMM]MXEU؁J|Ph6*hPj6*h\Pj 6*hPh|P=~|Ph|PQQPv@PPQQQ_Ɔ\PPPj|Pt}th=PhPNPh@@u@P t9NPh@@u#P tPv,PjMM]MU ?#uF^?u]UƆ~h" L6*v lPhlP6*jPjPv~Ps~u~ t~Pv V^dž4Hdž6F8dž:l<>+ɉB@džHJFDLNNPdžR~\^`bdžddžf+ɉvtNln+ɉVT> sPh= s!8jhjh0dž|5Pjj uvō4P|h" |]XEUvhj6V6T uvjhPh0 MM]MXEU؃ TVFVF @TVvv6*FV6*F +RPv vvFvvFVTVFMM]MXEUWVvF -tHt~9F tV @=utu6< uVjuPjt@>.uVjtPj| u@VjtjuW뒋F =~twtKtL,rt t!++IVjtP t@WVju7 P uB@9VjVj )^_MM]M XEU؃WVvVjvt FF -tBHu79F tATVFVT\VVFVTVF<w FV uWVju<__؎F&G"؉^FG‹&RP VjtFV‹&RP<w >:Ru  tE <t1uhBjvhvhjBPL PWc F= uvhWWW ~tVjPjF =~uwt te++jvhvhjjj FPjj t0<FPwWjj* tv^wu VjVjz^_MM]M XEU؃WVvVjv FF -tCHudF 9tmTVFVT]VVFVTV>F6*jFhj2 Vh6*jGhj2{Vjsh <w! FV uAVju<__؎F&G"؍G‹&RP VjtFV‹&RP <w >:Lu  tE 6E uhBj, vhvhjBPl PW F= uvhWWW ~tVj3Pj<F =~uwt t[++`vhvhjjj PjjS t)6<<wWjj*`u <w VjVj X^_MM]M UvjjPvr~tvhjjjD ]XEU؁ F -tHt29F tP9F uBv8vjwhhjjF =~twt@u,nu,t++vjwhhh ~+Pv:[[Ɔzh" vjwhhhhhhXhhho>tzPhzPhhPvjjjjjhzPhvjiPGF t(vjuwVoVFPF+FFjFN+NN++FVPj+F+FPvvjV렋F =~twt t++jV~ u+Pj^_MM]M XEU؁VFjzP vVvvv~t;^?t3S tv  tv j4hjh1=tIF~tq~t^?ucv Ph@e@u2v j4j5v h1=u3zP ,vv7v j4hv h1=t͍zP cvve^ ?\u\t::u!RPAVcs tv j7Z }V80 tv j6j8jh4=u+^MM]MUTVFVT[VvjdxPhhhjhxPv t0vjdxP vj(vj* PjjFFFVTVF]UdžtvhvPv;vj uvxPh@@uz uPh@Pj uZPxPh@@uCz tt 9>PtVNVjdhZVjdhhjjVjfhj  PjjVjfF PNQjj djbbdfv jjjvPjNF+ffVFPVjjjF)FvfF)FvjgHF VjdhhhjjhPF jVjfhjjjshPhhVv hP u79NusN6*hhh4Vjdhh@hPhhq6*hPjF ue9Nt?NPJv RVj2hjj0b6*hhhVjdhh*jv Vj2j3jj0Vjfj jjj6*hPjvFPFV F#Vjfhj^F&FjPFF9FvvVjfh6Hjj=u;uVjfhjjjVjfj jjjVjf9Pjjjjv =FVRPjj^&Gt&w j&w jRP^&w jjjGPjjjjjhj^&|!&w Ğh&@@P^&G@@P&w^&Gu&w GPj  ^&Gj ^&GnN^&wWZF =~uJwuu,bt,tZ, t++VU u!VjdhBhhBjLhBhh uVjfhjjj VjdhBhhBhh thhhB(F ~tVjdhBhchBhh uVjfh PPPF }FFHhZhhVjVjPB^_MM]M U>t 6x]U>t 6~PhZ%~Ph[~P荆~PhA6*~P6H u6*6-=u+H6*~Ph6*~P6H]UP:j tV~u9<u<9GtՃ~thZ j;Fuw u u؋G늋G"؎&6*hFPj@ tg‹&RP‹&RPFPP6&j}j~Ph1wYuB+]UXvFPj(L FPFP`FPPFPFP t7vhjjjFVvFP vhjvvd ]XEU؁ WVv~ -t Hu9>t 9>PtVa6<P t Vj5Vjdhj(jj/VjehhjjAVjyhhjj\ZHVjdPjVjfPjF =~uww!uuL,btt%u, t\++L~tHv?~t5Vjehhh ~+FPVH[[Vjf_PvgTVFVTZVVjevPh-VjyhhhhhhvPV tCVjevPVjoPjwVjfPj Vj(VjPjjCFVTVUVjehhhVjyhhkhhjwhhh t hhhhEhhhvPhZvP%H>ZuhZvPHhZjVh5h3R uVj}h6jjoVj{j3jjjvPj tdVjehhhVjyhhhhhhhhhrhhhPVhhPhL uP^Vjehhh VjyhhVjdhj)~hhhhg>uhhh>ZuhZP hZH>u hhh'<whhhhZhvVji{P6Hjjjj]>t>u<w>u <w V~ u+P! \^_MM]M XEU؃WVv~ -tHt59>uLV1DVjdhjjjIVjwhhjj֋F =~twt\u,bt ,t++Vjdhhh ~+FVjwhhh ~+ FvVv[[lVjdhj)Vjwhhhc jjhoP Fj hhh!*Fj2v:~uVjVj ^_MM]M UTVFVTZVvjexPht vjyhhhhhhxPvE tTvjexPWvj Pj vjf Pj\vj(vjTPjjFFFVTVF]XEU؁VvF -t$Hu99F t P9F tQVH6<<w!xz u[dVjdhj(jj"Vjdxz‹&RP"6<<whhhg)!hhh!hh$hhdhhhh@$6<<whhhg)j"Vjehhjj!"Vjehh-"Vjyhhjj"Vjyh hZxz‹&RPi$Ğx&GH6<<whZ " t VjijVj{j26<<whZ "Pjj"<w$Vj}h6jj7< u>.rbVj"Pj"Vjf"Pj"Vj{#Pj #Vjd#Pj#Vjy##Pj+#Vji?#PjG# >.tVjeO#PjW#Vjp#Pj F =~uwuuD,btt"td, tY++;~tv~tVjehhh# ~+vPVZ[[Vjf PvtVNTVjehhh $ u>Vjyhhhhh&$ t hhhhhhh|P>dunh|P thZ|P$HD6<<w$rt u)hZt‹&RP$<w>ZuhZhh$|PhZOjVh:h3 tQVj(VdP'Pjj [Vj}h6jj,%Vj{j3jjjq'rPj uVjehhh[%Vjyhh&hhjR&hhh% t hh!hhhhh|P!Vhh|Ph~ uVj&h|PV$ t,>duhZ|PJ"hZBH|PRލjP<GPVjehhh.&Vjyhh<&Vjdhj)'>uhhh_(VjdhD'v6<<w&v<whhhhZhrVjiP6HjjPj]'Vj*'r^MM]M XEU؁WVvF -tHuF 9tVN<wR'rt u VjVjdt‹&RP`'<w(Vjw<w!Vjdhjjj'Vjwhhjj+(< u >.smVj'Pj'Vjw'Pj'Vjd"Pj":F =~u*w tu@++><Vjwhhh(nVjdhj)#uh$9(Vjwhjjju! tuhhQ!hhvPhp@tuhhS ; tVjqjrjj03}tuj@n@@PE t Phh%b&{&Vj{^_MM]M UvjH);Ftvjvvjjj9,]UvjFF~tNvvvj2]UWVv6*jyFPj *VjF*F uƆLdžJtjLPj{LPFP)Ft)jLP)LPjLLPFP)Ft)jLP)LPjLPFPOF*PLP*LPf%P6*j{LPjdLPJVVFPWjY+WVjF,PQ+Hja*Fj)+FV t3j*F̉VWRP*F̉VWvvЍLPJ*ejnF̉V t4WvP*F̉VWvvЍLPJWvv̚ WjjjLPJvvjjWLPJFҚ;uj4+FPjPVR,Ht WHWjVFP^_]XEU؃WVv~ F Hu-uHu-uHuHHu#-u"-uHuHu{Hu-Tu-yu-s-wZHHu-s-wC-uVjVj0jPjj