Scribd Logo
Upload

Welcome to Scribd!

  • Hutchinson 1

    Uploaded by

    Chaitanya Desai
    8 views8 pages

    Original Title

    hutchinson1

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    Download as pdf
    8 views8 pages

    Hutchinson 1

    Uploaded by

    Chaitanya Desai

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    Download as pdf
    of 8
    ‘Acta meal Vol. 37, No. pp. 99-916, 199 Proud im Great Brian A ight mend mor-61039 $3.004000 (Copyraht © 1989 Perancn Pre pe EFFECTS OF NON-PLANARITY ON THE MIXED MODE FRACTURE RESISTANCE OF BIMATERIAL INTERFACES ‘A. G. EVANS! and J. W. HUTCHINSON? "Materials Department, Univeristy of Calforia, Senla Barbara, CA. 93106 and "Division of Applied ‘Sciences, Harvard University, Cambridge, MA 02138, USA. “ (Received 3 ne 1988) ‘Abstract—The effects of non-planarity onthe fracture resistance locus of interfaces bas boro invetigated ‘sing a simple model of contacting facets along the crack surface. The conlacts resist the motion of the ‘rack surface by means of friction and locking and Wereby modi) the energy release rate Gat the crack front, The modified G" governs the effet of the contacting facets on the overall interface fracture resistance, G. The tends in G, with phase angle of loading are found to be influenced largely By & ‘non-dimensional parameter that determines the length of the contact zone. Tht parameter iain tira, Aependent on the amplitude of the undulations on the fracture interface as well ats atin facture Résamt—Nous avons étudié les eles de la non-planité sur les eux de résistance & la rupture des Interfaces en ublisant un model simple de Tacetes en contact Te long de la surface dela Basure. Les Contacts freinent le mouvement dela surface de la Fsture par fietion et blocage, ett modifint done lavitese de perte d'nergie Gen tte de ssure. Cette vitesse G! modifergi efet des facets en contact Sur la résisance 4 la ruptue interfacial totale, G Les variations de G, avec angle de phase dela mise cen charge sont fortement influences par un paramétre sans dimension qui détermine ls longueur de la tone de contact. Ce paramere, pe conte, depend de Tamplitude des ond interface de rupture ‘comme de sa résistance inrinseque a la rupture ‘Zesammenfassung—De* EinluD der Unebenbet auf den Ort des Bruchwiderstandes von Grenaflichen ‘wurde mit einem einfachen Modell kontaktierender Faceten entlang der Bruchflche untersucht. De Kontakte widerstchen der Bewegung er RiGoberNiche durch Reibung und Verankerung und verdndern ater die Fresetzungsrate der Energie Gan der RiGfront. De modifierte Rate G'beherrcht den Ein er Kontaktierenden Facetten auf den gesamien Bruchwiderstand der Grenzfiche ,-Die Abhinpgkeiten von G' von dem Phasenwinkel der Belastung werden rdGtenels durch einen dimensionsiosen Parameterbecinfult, der die Lange der Kontaktzone bestimmt. Dieser Parameter hing! wiederum von der Amplitoge der Wellungen auf der Bruch-Grenzféche und deren intrinsischem Bruchonersiand ab. 1, INTRODUCTION Many important interface fracture problems involve mixed mode (shear and opening) displacements along the crack surfaces, as exemplified by thin film deco- hesion [1-3] and fiber debonding in composites (4) Subject to such displacements, interface fracture must be influenced by non-planarity of the interface and by the phase angle of loading, = tan“ (Ky/K;), where K, and Kare the Mode Iand Mode II stress intensity factors. Typical interfaces are non-planar [5] (Fig. 1) and crack surface contact either at undulations or at facets along the interface crack can have an effect on the overall fracture resistance of the interface, G,, especially at large phase angles. Such phase angle effects are illustrated in Fig. 2, which indicates the results of a fracture test on Al,O, bonded with Ti. ‘The upper, interface failure was caused by applied loading. with ¥ 0. The lower crack in the Al,O, parallel to the interface formed subsequently, caused 09 by the residual stress in the Ti acting as a thin film. For this case, ¥ = n/4 [6] at the interface crack, whereupon the crack is diverted into the Al,O, rather than propagating along the interface. Trends in G, with phase angle of loading are predicted in the present study for the case of a faceted interface. The simplest model of the process that provides insightful preliminary conclusions, illus- trated in Fig. 3, consists of kinks along the crack surface. When the crack surfaces contact atthe kinks, the stress intensities A° at the crack front differ from the applied values to an extent governed by the kink angle, the kink amplitude and the friction coefficient. ‘The trends in X* with these variables provide one contribution to the increase in interface fracture resistance with phase angle of loading, as elaborated in the present article, Other possible influences on G,, such as crack front deflections, plasticity etc. are not considered in this study EVANS and HUTCHINSON NON-PLANARITY AND FRACTURE RESISTANCE o Fig 1. Nonsplonse metal eetamic imeelaen tcourtess AM" Rie) (a) Imetfve faces tm the system" Nb ALO, (6) Interface “undulations am the system Ti ALO:. with imterphases of TiAl and TiAl Cracking ossuts along the ‘ALO, TIAL interface Fig. 2 The results ofa facture es! on a Ti Al,O, system, ‘The upper interface crack formed fst ypon lnading. The lower crack im the ALLO, formed sufsequenty. becasue of ‘residual sess inthe Ti teourtesy M. Rue) Fig. 3. The crack kink model used to analyz effects of crack surface contact, The loading inthis cates K,>0 and Kio 2. THE BASIC MODEL, ‘The mode! is developed fora thin imerface between elastically homogeneous bodies with identical elastic properties. However. the general trends should be applicable to bimaterial interfaces, albeit that com plex stress intensities should then be used (7). The basic geomerry involves a single kink (Fig. ¥a)] at angle f. along the crack surfaces at a distance. d, from the erack front. This geometric choice simulates cach contact along a muliply faceted snterface ‘When the phase angle of loading allows contact at the kink, Coulomb friction is assumed to obtain with « friction angle e where sis the fiction coefficient. Otherwise, elasticity ‘exists throughout, An approximation forthe effect of contact on the crack tip field invokes an inclined force, F. acting on the surfaces ofa planar crack [Fig 3e)]. with the telination angle w governed by @ and the kink angle (assuming that the faces of the facet are either sliding or on the verge of sliding) where wonp-o Q The detivafion which follows assumes 0-< B < x/2.as in Fig. 3. For kink angles in the range 2/2 < <2. all equations continue to be valid if ¢ is replaced by ran 0 EVANS and HUTCHINSON: NON-PLANARITY AND FRACTURE RESISTANCE =@ corresponding to the switch in direction of the fretion force. The force is represented as 2 wnitormly distributed traction acting over a characteristic ‘microstructural length, 2d [Fig. (c)]. This geometric choice allows the effect of contact to be expressed in terms of normal and shear forces, P and Q. re- spectively, acting over ed PeFsinw, Q=Feosw Ga) Q4iP = Fem. ab ‘The uniformly distributed normal force causes a crack opening. wat x= —d uw = {Ml — vin ]Pe(e) (4a) where [8] Viti = ize r4eIn meyer) IRE DT RED 029 x4 yt vieor) ‘while the uniformly distributed shear force causes a relative shear displacement, r: atx =~ 4a. a where Gis the shear modulus and vis Poisson's ratio. ‘The crack surface displacements caused by contact are related to the forces by 4u—9) *e (4b) ” Onw 0 uta Fe such that the contact induced contribution 10 the Stress intensities atthe crack tp. Kare iy + ik) V Tad Oe where Wao land the subscripts 1 and II refer to the opening and shear modes. respectively ‘The corresponding ctack surface displacements at the contact site (x = ~d) dictated by the remote loads ve ct iv = (MU WG) FARK + IK) where K, and Ki, ate the stress intensities associated ‘with the applied loads 3. CONTACT AND LOCKING CONDITIONS The existence of crack surface contact and of frictional locking are influenced by the relative mage ritudes of the phase angle of loading, ¥. the friction langle, 6. and the kink angle, B. Contact along the kink only occurs when cos B +r sin B > 0. Hence, by defining & characteristic stress intensity Kas © Kt = ~(Ky cos + K,sin ) 0°) contact of the facet faces occurs when Kr>0 (108) which, since tan ¥ = Ky/R;, is equivalent to K-P<¥<2n—p. (106) Furthermore, coniact along the straight crack sur- faces occurs when K<0. ay ‘The basic contact conditions are mapped in Fig. 4 Fer present purposes, the condition, K,<0, és not meaningful. The analysis should only be used for positive K;. A separate analysis would be needed for negative Mode 1 Frictional /ocking occurs when the tangential force along the kink, T, satisfies the inequality a where N is the normal force. This condition can be ‘expressed in terms of the forces P and Q using N= Qcosh+Psinf = Rel(Q+ iP] (13a) T=Q sinB ~ Pcos f= —Im(Q + iPye |, (130) ‘Then, by noting that N sing — Tos @ = Im{(W ~ iT}e*] = Im[iQ + Pe") Teun as) 1" Cac crm rere on pontine) _ Fig. 4A map ofthe sliding and locking conditions an pated with the crack asperty. For present purposes. only Ky 0 i relevant ae the locking condition becomes Iml(Q + iP)e**-">0 (15a) or, because locking requires that w= r = 0, (3b), (5) and (7) give Iml(Ky + ike" P< 0. (15) Consequently. since Ky iKy= [Klee (16) locking occurs when sin +89 —m/2)>0 (7a) A2~P+O 0, range of loading exits in which K,>0 and the facets are frictionally locked as mapped in Fig. 4 4. INTERFACE FRACTURE RESISTANCE The above conditions of contact and frictional locking provide insights concerning the analysis of the interface fracture resistance, For purposes of analysis, the spectrum of contacts must be simulated by means of a contact model. Two such models are presented, each providing a different perspective of the contact phenomenon. One model considers a Single row of contacts with friction having a full spectrum of facet angies located at a fixed distance from the crack front. The fracture resistance can then bbe estimated. in principle. by summing over the number of rows within the contact zone. However. 10 accomplish thi. an estimate of the zone size and of interaction between rows is needed. For this reason, 4 second zone model is developed that examines simple contact condition at the facets. but more rigorously addresses the zone size and interactions that occur along the zone 4.1. Single row model Contact behavior at facets within a row is governed by the net displacements of the facet surfaces. derived from equations (5) and (7) a5 (a icy = a [Ferntree 22 ree~]. a When sliding contact occurs, fw = ¢ and wm of, equivalently, Rew + ic)e %} =0. Thus, by (18). the resultant force across the facet faces is Vedi K*icos ro) Inserting this expression forthe force F into equation (6) and adding the stress intensities [rom the applied EVANS and HUTCHINSON: NON-PLAN RITY AND FRACTURE RESISTANCE loads gives Kit IK} = Ky 41K, +h(e™K4 00s 6 (20) where (c)= (OiR(6) is the function plotied in Fig. 5. Consequently, the tip stress intensities are j= K+ MOsin(p ~ OK e086 Kiy= Ky + h(c)e0s( — 9)K cos @ ay and the strain energy release rate at the tip is Gall —2G MKF +(KYF. 22) Noting that the effective energy release rate associ ‘ated with the applied loads is G = (0-9) 2G1K} + Ki the crack tip energy release rate becomes @ once x pores sin(f ~ 0)-+ Kycosth cos 2 Thus, by (9) 4G St H6.8..0 Usin 8 + cos tan vyisin(B — 6) ay + cost = oyanw)) = tert + ano) inf + cos f anv eS cos! gt + tan) where AG = GG" is the reduction in G. ie. the shielding, caused by contact at the facet. (Recall that this equation, as well as all others above in: fer os| mie Fig, 5 Comparison of the fnetions Ai) fom the present pproximate model and Ae) fom the exact soluion fora Tmirocrack ahead ol macrocrack EVANS and HUTCHINSON: NON-PLANARITY AND FRACTURE RESISTANCE, in =e x_ Yee A" a 0 & 3 a ws . a a Fan vee B c) oa ven ones) Cy Fig. 6 (8) Reduction in crack tip release ‘of pase angle of loading for various kink ote bience of fetion (M'= DS, @ =O}. (b) Elect of Friction (@=45) on reduction in near tip energy release rate 05). volving @, ate limited to 0 < B 3n/4. ‘The interface fracture resistance is estimated based on consideration of a row of contacts with uniform distnbvtion of kink angles parallel to the front, ‘ranging from 0 to x. (Kink angles with B < Oor B > « are unlikely because alterative, noninterfacial, frac- ture paths would normally oblain a facets baving B<0 or B>x.,) For loading combinations with O<¥ A/2, sliding contact occurs for n-¥a /2 Curves of (AG)/G as a function of y obtained from Gla.) are plotted in Fig. 7 for several values of the friction angle $. These curves were obtained using += 05. The effect of different choices for h ranging ‘over all realistic possiblities (ef. Fig. $) is shown in Fig. 8. These results can be used to illustrate some of the issues involved in determining the extent of crack shielding. 11s firstly evident that there is only a moderate effect of the friction coefficient on crack shielding. ‘within the usual range 0 < @ < m4, because locking constitutes the greatest impediment to crack surface displacement. It i also noted from Fig. $ that h remains quite large ($0.2) down to quite small values of «( ~ 0.05). Consequently. by simply sum- «aaa rer Fig. §. The variation in crack shielding (AG) with phase angle caused by a) row of contacts, Tor» range of ain @tané =I). EVANS and HUTCHINSON: NON-PLANARITY AND FRACTURE RESISTANCE ming (8G) on parallel rows of nominally identical facets all governed by equation (31), the net shielding ‘must be sensitive t0 the number of rows and hence, the size of the contact zone. These insighis suggest that a zone model which emphasizes the locking ‘characteristics and explicitly incorporates the zone size should provide a more reasonable prediction of the effects of contact on the interface fracture re- sistance. Such a model is presented below. 42. Zone model ‘A-2one mode! is developed for the simplified contact conditons depicted in Fig. 9 corresponding 10 B= 0 and ¢ =0 (no friction). Then, Kj= Kad the ‘crack opening depends on X, only. as given by (7) = 80 = PK) RE (32) where ri the distance from the erack tip. Contacts exist overa zone length Z that satisfies the condition uLy=H where H is the height of the interface step (Fig. 9), such that L = (2 EMM ~ KE. @) Within the contact zone, the shear stresses and displacements are elastic and analogous to those associated with a linear array of microcracks (Fig. 9). Furthermore. in order to evalaute Kj. the ‘microcracks can be simulated by a continuous linear spring mode! (Fig. 9) in which the stresses ¢ and Grace Settee 18) Zone Configuration Db). Mieroerack Zone ©) Else Spring Contiguraton Fig. 9. A schematic illustrating the zone model used to ‘etermine tends in G, with phase ongle of loading EVANS and HUTCHINSON: NON-PLANARITY AND FRACTURE RESISTANCE Table 1 meri property for the e soGre ” 1m Tsuen G isplacements rare related by Bell =v) In| /singnD a aE Dy} a4) where / is the spacing between facet (microcrack) centers and 2 is the facet length (Fig. 9). This is the exact result for the additional shear displacement duc to the presence of an infinite linear array of micro- ‘cracks subject to remote shear stress r parallel to the cracks [11,12]. The linear spring model has been. solved by Budianshy er al. [13], For the present linear spring (34), the result of interest is KivKu= ka) as) where wn ** iifi-sintxd 2) and the function k() is given in Table | (as 1-4) in Reef, {13}, Plots of Kj, Ky vs Dil from (35) are shown in Fig. 10 for various numbers of microcracks NGL = N1). The result for one microcrack isthe exact result from Ref. [9] Using Kj, Ky and (35), the relation between the energy release rates is obtained as 1G) tan wy TH any 36) GG= an Further progress is achieved by combining equations (33) and (36) t0 give REHAG' GUI + tan" y) 8 330 = FG" Inf sintxD 2) sal ‘The basic trends in the fracture behavior of the interface with the phase angle of loading can be estimated by sclecting the value of at y =O as a reference value. 24. and setting G' equal 10 the fracture resistance of the interface Gy, such that (EWG) EH IG) 35 XR iniismaD ayy =O The quantity 2, contains the baste information con: cceming the interface and is thus a material parameter. ‘The results contained in equations (35), (37), (38) and (39) can be combined to provide an expression for the crack shielding as 616 = Y= Kn 98616 + 40) Specific trends in AG/G with ¥ for various 2 deter- ‘mined using equation (40) are plotted in Fig. 11 Ww may be ascertained from Fig. 11 that two fegimes enist; one at 3,31. and the other when 3,2 10 For the former. contact has the maximum ‘effect on crack shielding. such that. Ki, = 0 and any Taian For the latter. there is essentially no shielding when ¥

    You might also like