scieee Open visual document viewer

Strength and reliability of WC-Co cemented carbides: Understanding microstructural effects on the basis of R-curve behavior and fractography

Tarragó, José María; Coureaux, D.; Torres Hernández, Yadir; Jiménez-Piqué, Emilio; Schneider, Ludwig; Fair, J.; Llanes Pitarch, Luis Miguel

Abstract

Strength and reliability of WC-Co cemented carbides (hardmetals) are dependent on effective fracture toughness as well as on nature, size and distribution of processing flaws. Regarding toughness, they exhibit a crack growth resistance (R-curve) behavior, derived from the development of a multiligament bridging zone at the crack wake. Accordingly, successful implementation of fracture mechanics requires consideration of tangency criterion, between applied stress intensity factor and R-curve, in addition to fractographic inspection. It is the aim of this study to evaluate the strength behavior of a series of experimental WC-Co grades on the basis of R-curve failure criteria. Results indicate that microstructural effects on the strength of hardmetals may be satisfactorily rationalized following the referred criterion. The analysis includes consideration of nature and distribution of fracture origins, found to be more diverse and wider, respectively, for the harder fine-grained grades. This experimental evidence, together with the fact that these hardmetals exhibit steeper rising R-curves than tougher coarse-grained ones, leads to lower reliability for the former. This investigation documents and validates the great relevance of R-curve behavior for optimizing the mechanical performance of WC-Co cemented carbides on the basis of microstructural design.

Full text

ACCEPTED MANUSCRIPT S eng h and eliabili y o WC-Co cemen ed ca bides: Unde s anding mic os uc u al e ec s on he basis o R-cu e beha io and ac og aphy J. M. Ta agó1,2,ϕ, D. Cou eaux1,3, Y. To es4, E. Jiménez-Piqué1,2, L. Schneide 5, J. Fai 6 and L. Llanes1,2 1CIEFMA, Depa men o Ma e ials Science and Me allu gical Enginee ing, EEBE, Uni e si a Poli ècnica de Ca alunya, 08019 Ba celona, Spain 2Ba celona Resea ch Cen e in Mul iscale Science and Enginee ing, Uni e si a Poli ècnica de Ca alunya, 08019 Ba celona, Spain 3Facul ad de Ingenie ía Mecánica, Uni e sidad de O ien e, 90500 San iago de Cuba, Cuba. 4Depa amen o de Ingenie ía y Ciencia de los Ma e iales y del T anspo e, Uni e sidad de Se illa, 41092 Se illa, Spain 5Sand ik Machining Solu ions, Co en y CV4 0XG, UK 6Sand ik Hype ion, Co en y CV4 0XG, UK Φ Cu en add ess: Sand ik Hype ion, 08107 Ma o elles, Spain ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT Abs ac S eng h and eliabili y o WC-Co cemen ed ca bides (ha dme als) a e dependen on e ec i e ac u e oughness as well as on na u e, size and dis ibu ion o p ocessing laws. Rega ding oughness, hey exhibi a c ack g ow h esis ance (R-cu e) beha io , de i ed om he de elopmen o a mul iligamen b idging zone a he c ack wake. Acco dingly, success ul implemen a ion o ac u e mechanics equi es conside a ion o angency c i e ion, be ween applied s ess in ensi y ac o and R-cu e, in addi ion o ac og aphic inspec ion. I is he aim o his s udy o e alua e he s eng h beha io o a se ies o expe imen al WC-Co g ades on he basis o R-cu e ailu e c i e ia. Resul s indica e ha mic os uc u al e ec s on he s eng h o ha dme als may be sa is ac o ily a ionalized ollowing he e e ed c i e ion. The analysis includes conside a ion o na u e and dis ibu ion o ac u e o igins, ound o be mo e di e se and wide espec i ely o he ha de ine-g ained g ades. This expe imen al e idence, oge he wi h he ac ha hese ha dme als exhibi s eepe ising R-cu es han oughe coa se-g ained ones, leads o lowe eliabili y o he o me . This in es iga ion documen s and alida es he g ea ele ance o R-cu e beha io o op imizing he mechanical pe o mance o WC-Co cemen ed ca bides on he basis o mic os uc u al design. Keywo ds WC-Co cemen ed ca bides, s eng h, eliabili y, ac og aphy, R-cu e beha io ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT In oduc ion B i le ac u e is one o he mos common ailu e mechanisms in he applica ion o cemen ed ca bides as o ming ools and s uc u al componen s, pa icula ly unde he combined ac ion o mechanical and con ac loads [1]. In a la ge numbe o cases ac u e s ems om he p opaga ion o p e-exis ing laws ei he in insic o he manu ac u ing and/o pos - ea men p ocesses (e.g. Re . [2–6]) o induced unde se ice condi ions (e.g. Re s. [7–9]). Simila o o he b i le ma e ials, ensile s eng h o cemen ed ca bides depends on he size o he majo law and exhibi s wide sca e due o he a iabili y associa ed wi h de ec dis ibu ion. Acco dingly, s eng h is size-dependen as he p obabili y o inding a majo law inc eases wi h specimen size. Wi hin his con ex , Weibull s a is ics [10] ha e been equen ly ecalled as a sui able app oach o a ionalize he up u e s eng h and eliabili y o ha dme als (e.g. Re s. [11,12]). An ex ensi e numbe o s udies ha e been de o ed o analyze he in luence o he mic os uc u e on he ac u e s eng h o ha dme als (e.g. Re . [1,4,13–18]). Gu land and Ba dzil documen ed ha he s eng h ini ially inc eases wi h he binde mean ee pa h (λCo) un il i eaches a maximum a λCo alues a ound 0.3-0.6 μm, and hen i s a s dec easing [13]. This beha io was associa ed wi h a b i le-duc ile ansi ion. Fo small λCo alues, ac u e mechanisms a e go e ned by he b i le ac u e o he WC con inuous phase. The e o e, he dec ease o he lexu al s eng h wi h dec easing λCo was ela ed o la ge WC amoun s and o highe con igui y alues o he ca bide phase. On he opposi e side o he cu e, he s eng h was domina ed by he ex ensi e plas ic de o ma ion o he binde phase. Thus, ac u e s eng h was in e sely p opo ional o λCo due o a dec ease o he cons ain deg ee exe ed by he ca bide phase on he binde , when inc easing λCo alue. This mic os uc u e-s eng h pe spec i e was sha ed by Fang [17], who plo ed ac u e s eng h agains ha dness and documen ed ha he o me goes h ough a peak a in e media e alues o he la e . In line wi h Gu land’s ea lie discussion, Fang pos ula es ha s eng h o ha dme als is based on he compe i i e ac ion played be ween ha dness and ac u e oughness. Howe e , Gu land’s heo y could no explain he high ac u e s eng h dispe sion measu ed in cemen ed ca bides. This pe cep ion changed wi h he implemen a ion o linea elas ic ac u e mechanics (LEFM) o a ionalize he b i le beha io o ha dme als [2,3,14]. Wi hin his ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT con ex , he ac u e oughness is he ma e ial p ope y ha , combined wi h he c i ical p e-exis ing law sizes and he applied s ess, de e mines he ac u e s eng h o cemen ed ca bides. LEFM app oach conside s a ully b i le beha io wi h a ailu e p ocess gi en by ex ension o c i ical laws, wi hou s able subc i ical g ow h be o e ac u e [1]. Howe e , as widely epo ed in li e a u e, cemen ed ca bides exhibi subc i ical c ack g ow h unde mono onic loads. I yields (and esul s om) he de elopmen o a c ack-b idging mul iligamen mechanism a he c ack wake [19,20]. This mechanism en ails he de elopmen o a ising c ack g ow h esis ance cu e [21]. The au ho s o his pape ecen ly epo ed a ele an in luence o mic os uc u e on he R-cu e cha ac e is ics o ha dme als [22]. I is he pu pose o his in es iga ion o e alua e he mic os uc u e-s eng h ela ions o a se ies o expe imen al WC-Co g ades on he basis o R-cu e ailu e c i e ia. In doing so, special a en ion is paid o explaining he impac o R-cu e beha io on he s eng h and eliabili y o cemen ed ca bides. Ma e ials and expe imen al aspec s The in es iga ed ma e ials we e a se o expe imen al WC-Co cemen ed ca bide g ades supplied by Sand ik Hype ion. Eigh g ades co esponding o di e en combina ions o binde con en and ca bide mean g ain size we e s udied. Specimen codes and key mic os uc u al pa ame e s - binde con en (Vw %), mean g ain size (dWC), ca bide con igui y (CWC), and binde mean ee pa h (λCo) - a e lis ed in Table I. In he ine g ades, conc e ely 3UF, 6UF, 10UF, 11M and 15UF, a small amoun o C 3C2 was added as g ain g ow h inhibi o . Mean g ain size was measu ed ollowing he linea in e cep me hod, using ield emission scanning elec on mic oscopy (FE-SEM) mic og aphs aken in a JEOL-7001F uni . Ca bide con igui y and binde mean ee pa h we e deduced acco ding o empi ical ela ions om open li e a u e [5,23,24]. Mechanical cha ac e iza ion includes ha dness (HV30), lexu al s eng h (σR) and ac u e oughness (KIc). Ha dness was measu ed using a Vicke s diamond py amidal inden e and applying a load o 294N. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT In all he o he s cases, es ing was conduc ed using a ou -poin bending ully a icula ed es jig wi h inne and ou e spans o 20 and 40 mm, espec i ely. Flexu al s eng h es s we e pe o med on an Ins on 8511 se ohyd aulic machine, and 10 o 15 specimens o 45x4x3 mm3 we e es ed pe g ade. The su ace subjec ed o he maximum ensile loads was polished o mi o -like inish and he edges we e cham e ed o educe hei e ec as s ess aise s. Flexu al s eng h esul s we e analyzed using Weibull s a is ics. F ac u e oughness was de e mined using 45x10x5 mm3 single edge p e-c acked beam (SEPB) specimens, wi h a no ch leng h- o-specimen wid h a io o 0.3, ollowing he p ocedu e p oposed by To es e al. [25]. A de ailed ac og aphic analysis on b oken specimens was ca ied ou by means o FE-SEM in o de o de ec c i ical laws ha p omo ed ailu e. Subsequen ly, he size o iden i ied c i ical laws was measu ed using he image analysis ee-a ailable so wa e ImajeJ. Resul s and discussion In luence o he mic os uc u e on he lexu al s eng h Ha dness, ac u e oughness, ac u e s eng h and Weibull cha ac e is ic s eng h (σ0) and modulus (m) alues o he in es iga ed cemen ed ca bides a e gi en in Table II. The Weibull plo s o s udied ma e ials a e shown in Fig. 1. Ve y in e es ing, he ha de g ades (i.e. 3UF, 6UF and 10UF) exhibi a bi- modal s eng h dis ibu ion. The e o e, he Weibull modulus o hese g ades is no epo ed. On he o he hand, he o he ma e ials exhibi a clea ly uni-modal dis ibu ion. A ac og aphic analysis by means o FESEM was conduc ed in o de o iden i y he c i ical laws ha p omo ed ailu e. Some examples o he di e en ypes o c i ical laws e idenced o s udied ha dme als a e shown in Fig. 2. The la ge sca e measu ed o he ha de g ades can be di ec ly associa ed wi h wo di e en de ec popula ions, consis ing o po es and inclusions o he low s eng h alues, and abno mal coa se WC g ains o he op pa o he Weibull dis ibu ion. I is also in e es ing o poin ou ha he highe s eng h de e mined alues co espond o he specimens wi h he smalle de ec sizes, in line wi h he heo e ical amewo k es ablished by LEFM [3,26]. In addi ion, ga he ed in o ma ion om Fig. 1 and 2 indica es ha unde he conside a ion o de ec s o di e en na u e bu simila equi alen size, de ec s such as po es, ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT agglome a es o inclusions, imply a g ea e se e i y han in insic mic os uc u al he e ogenei ies (i.e. abno mal coa se g ains). This is in ag eemen wi h ideas gi en in li e a u e o o he ce amic ma e ials [27]. Fu he mo e, i he Weibull analysis o 6UF g ade is pe o med uniquely in he specimens ac u ing om mic os uc u al he e ogenei ies (i.e. he uppe pa o he Weibull dis ibu ion), he Weibull modulus would ise om 7 up o 23. This is a clea indica ion ha an exhaus i e con ol o c i ical laws na u e esul s in ele an s eng h and eliabili y enhancemen s in cemen ed ca bides. On he o he hand, c i ical de ec popula ion o he oughe g ades is cha ac e ized by an absence o po es and inclusions, oge he wi h a endency o simply consis on mic os uc u al he e ogenei ies such as abno mally coa se WC pa icles. In Fig. 3, ac u e s eng h is plo ed as a unc ion o binde mean ee pa h. In acco dance wi h p e ious in es iga ions [13,28], lexu al s eng h inc eases wi h λCo un il i eaches a maximum a alues a ound 0.25 μm and hen i dec eases o la ge λCo alues. I can he e o e be concluded ha he ma e ial wi h he highes s eng h is nei he he oughes no he ha des , concep s ha end o me ge in indus ial a eas. The e o e, a comp omise be ween ac u e oughness and ha dness is equi ed o achie e an op imum pe o mance o he ma e ial. I is also in e es ing o ema k ha s eng h dispe sion e idenced o in es iga ed cemen ed ca bides g ades wi h λCo o e 0.25 μm is signi ican ly lowe han ha obse ed o g ades wi h smalle binde mean ee pa hs. Two main ac o s may explain his signi ican di e ence in s eng h sca e . Fi s , as p e iously indica ed, and p obably he main eason, he Weibull cu e o he s udied ha dme als wi h a sho λCo exhibi a bi-modal beha io (Fig. 1). Meanwhile Weibull plo s o he o he s udied ha dme als a e clea ly uni-modal. Second, as p e iously commen ed in he in oduc ion, cemen ed ca bides exhibi a ising c ack g ow h esis ance beha io ha impa s eliabili y and damage ole ance o ha dme als ools and componen s [29–32]. The e ec i e eliabili y enhancemen de i ed om he c ack b idging mechanism is a unc ion o bo h ac u e oughness and ac u e s eng h o he composi e ma e ial [22]. S eng h eliabili y inc eases wi h he mic os uc u al size, de ined as dWC + λCo, un il i eaches a pla eau ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT o dWC + λCo leng hs o a ound 1 μm [22]. Consequen ly, i is easonable o expec lowe Weibull modulus o ha dme als wi h sho binde mean pa hs han o coa se ma e ials. Es ima ion o he c i ical law size The c i ical law sizes o in es iga ed ma e ials we e es ima ed acco ding o LEFM and R-cu e angency c i e ia by assuming hem o be in e nal ci cula laws ( o embedded ci cula c acks, Y has a alue o 1.12 [33]). On he one hand, LEFM ailu e c i e ion ela es he ac u e s eng h (σR) wi h he c i ical law size (ac ) acco ding o he exp ession [34]: (1) whe e Y is an un-dimensional c ack geome ic ac o . On he o he hand, when conside ing an R-cu e beha io , he angency c i e ia o uns able c ack p opaga ion p e ails. This c i e ia de ines ha he ma e ial ca as ophically ails when he ollowing wo condi ions a e sa is ied [32]: (2) (4) whe e Kapp is he applied s ess in ensi y ac o , KR is he c ack- ip esis ance s ess in ensi y ac o and a is he c ack leng h. Kapp and a may be ela ed o he applied s eng h (σapp) acco ding o LEFM basic equa ion [34]: (3) The c ack g ow h esis ance cu e o in es iga ed ha dme als can be es ima ed om an equa ion ecen ly published by he au ho s o desc ibe R-cu e beha io o cemen ed ca bides [21,22]: (4) ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT whe e K is he c i ical c ack- ip s ess in ensi y ac o equi ed o c ack ini ia ion, Δa is he subc i ical c ack g ow h leng h and is a c ack leng h no malizing pa ame e . The e o e, in acco dance wi h ins abili y c i e ion o R-cu e ma e ials, he size o he c i ical laws can be de e mined as ollows: (5) Acco ding o angency c i e ion, ising R-cu e ma e ials ac u ing om he p opaga ion o “sho ” c acks a e no able o de elop he whole oughening capabili y, and ac u e a e ec i e oughness le els (Ke ) lowe han he in insic ac u e oughness measu ed o long c acks [21,22,32,35,36]. Fu he mo e, R- cu e ma e ials also exhibi s able subc i ical c ack g ow h ex ension be o e ca as ophic ailu e. Reached e ec i e oughness le els and subc i ical c ack g ow h leng hs can be de e mined acco ding o angency c i e ion and hey a e a unc ion o bo h R-cu e shape and ini ial law [32,36]. This p ocess is exempli ied in Fig. 4 o he 6UF and 15C in es iga ed cemen ed ca bides. The a ea o he iden i ied de ec s du ing he ac og aphic analysis was expe imen ally measu ed and compa ed o he es ima ed ones. In doing so, de ec s we e conside ed o be ci cula shaped and he e o e, an equi alen c i ical adius (ac ) is epo ed. C i ical law sizes expe imen ally measu ed and es ima ed using he LEFM ( ull black squa es) and angency (whi e ci cles) c i e ia a e plo ed in Fig. 5 agains he binde mean ee pa h. No e ha in he case o he expe imen ally measu ed c i ical de ec sizes, a leng h ange is gi en whe e he bo om and op alues co espond o he smalle and la ge iden i ied laws, espec i ely. No de ec s we e iden i ied du ing he ac og aphic examina ion o he ha dme al g ade wi h he longes binde mean ee pa h (22M) and consequen ly hey a e no shown in he g aph. Absolu e di e ences be ween he es ima ed law sizes es ima ed ollowing he LEFM and he angency ailu e c i e ia a e a he small o he ha de g ades, bu hey a e ai ly la ge o he oughe cemen ed ca bides and inc easing wi h binde mean ee pa h. As i is clea ly shown in Fig. 5, a eally good ag eemen be ween he c i ical de ec sizes expe imen ally measu ed and hose es ima ed ollowing LEFM and R-cu e app oaches is a ained o he ha de g ades, e en hough ob ained ac ,R-cu e leng hs a e sligh ly lowe . Howe e , o he oughe g ades, ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT es ima ed ac ,R-cu e alues clea ly ma ch wi h he sizes expe imen ally measu ed in he iden i ied c i ical laws, whe eas he applica ion o he LEFM heo y esul s in a law size o e es ima ion. This ac e lec s he need o aking in o conside a ion he subc i ical c ack g ow h ex ension and he e ec i e de eloped oughness o a ionalizing he ac u e beha io o cemen ed ca bides [22]. On he one hand, low Ke alues imply la ge es ima ed size di e ences be ween bo h c i e ia. On he o he hand, la ge subc i ical c ack g ow h alues lead o an o e es ima ion o c i ical law sizes by LEFM. Mic os uc u al e ec s on ailu e p obabili y As p e iously commen ed, he ac u e beha io o cemen ed ca bides is o s ochas ic na u e [2– 4,11,12,14]. The e o e, p obabilis ic ailu e mechanics a e essen ial o p ope designing wi h his b i le- like ma e ial [37,38]. Acco ding o Weibull dis ibu ion s anda ds [39], he ange o ailu e p obabili ies co e ed by he expe imen al assessmen o he Weibull cu e inc eases wi h he dimension o he sample (i.e. he numbe o es ed specimens). Howe e , due o he cos s o he specimens he numbe o es ed specimens is usually a he small [40]. In his in es iga ion, he sample size was limi ed be ween 10 o 15 specimens, and he e o e he ange o conside ed p obabili ies o ailu e was kep om a 5% o a 95%. Ne e heless, o designing pu poses o ools and componen s, he measu ed cha ac e is ic s eng h usually needs o be ex apola ed o eally low p obabili ies o ailu e and o di e en e ec i e loaded olumes [40]. In cemen ed ca bides his ex apola ion is no s aigh o wa d, as hey exhibi a R-cu e beha io ha may imply a de ia ion om Weibull s a is ics [11,22,40]. Mo eo e , i has been obse ed in his in es iga ion, in some cases he ac u e o cemen ed ca bides exhibi a mul i- modal law popula ion whe e a Weibull dis ibu ion is no expec ed o occu [40]. The e o e, a deep knowledge o he ac u e p ocess o cemen ed ca bides is demanded o he p ope design o hea ily s essed ools and componen s. Acco ding o Weibull heo y [10,41], he s eng h o an indi idual specimen (σi) o a sample can be deduced o a gi en ailu e p obabili y (P ) by using he ollowing equa ion: ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT [34] G.R. I win, F ac u e, in: Handbuch De Physics, Sp inge -Ve lag, Be lin, 1958. [35] J.J. K uzic, R.L. Sa e , M.J. Ho mann, R.M. Cannon, R.O. Ri chie, The u ili y o R-cu es o unde s anding ac u e oughness-s eng h ela ions in b idging ce amics, J. Am. Ce am. Soc. 91 (2008) 1986–1994. [36] M.E. Launey, R.O. Ri chie, On he ac u e oughness o ad anced ma e ials, Ad . Ma e . 21 (2009) 2103–2110. [37] D. Rubes, B. Smoljan, R. Danze , Main ea u es o designing wi h b i le ma e ials, J. Ma e . Eng. Pe o m. 12 (2003) 220–228. [38] R. Danze , T. Lube, P. Supancic, R. Damani, F ac u e o ce amics, Ad . Eng. Ma e . 10 (2008) 275–298. [39] ASTM C 1239-00: S anda d p ac ice o epo ing uniaxial s eng h da a and es ima ing Weibull dis ibu ion pa ame e s o ad anced ce amics, USA, 2000. [40] R. Danze , P. Supancic, J. Pascual, T. Lube, F ac u e s a is ics o ce amics – Weibull s a is ics and de ia ions om Weibull s a is ics, Eng. F ac . Mech. 74 (2007) 2919–2932. [41] J.B. Wach man, Mechanical p ope ies o ce amics, 2nd ed., John Wiley & Sons, New Yo k, 1996. [42] J.M. Ta agó, G. Fa gas, L. Ise n, S. Do lo, E. Ta es, C.M. Mülle , E. Jiménez-Piqué, L. Llanes, Mic os uc u al in luence on ole ance o co osion-induced damage in ha dme als, Ma e . Des. 111 (2016) 36–43. [43] J.M. Ta agó, S. Do lo, J. Es e e, L. Llanes, In luence o he mic os uc u e on he he mal shock beha io o cemen ed ca bides, Ce am. In . 42 (2016) 12701–12708. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT Lis o ables Table I. Mic os uc u al pa ame e s o in es iga ed cemen ed ca bides. Table II. Ha dness, lexu al s eng h, Weibull cha ac e is ic s eng h, Weibull modulus and ac u e oughness alues o in es iga ed cemen ed ca bides. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT Lis o igu es Figu e 1: F ac u e s eng h alues in he o m o Weibull dis ibu ion o s udied ha dme al g ades. The in es iga ed cemen ed ca bide g ades wi h he lowe binde mean ee pa h exhibi s a: (a) bi-modal Weibull dis ibu ion, whe eas he o he g ades ha e (b) uni-modal Weibull cu es. Figu e 2: FESEM mic og aphs o ailu e ini ia ion si es; namely: (a) ca bide agglome a e in he 3UF g ade, (b) po e in he 10UF g ade, (c) abno mal coa se g ain in he 10C g ade and, (d) binde less agglome a e o coa se WC pa icles in he 15C g ade. Figu e 3: F ac u e s eng h as a unc ion o he binde mean ee pa h. Figu e 4: Es ima ed c ack esis ance cu es o he 6UF and 15C in es iga ed ha dme als ( ull lines). Uns able c i e ion is eached when angency condi ions be ween he R-cu e and he applied s ess in ensi y ac o (dash-do lines) a e ul illed. Dashed lines indica e he subc i ical c ack g ow hs leng hs and e ec i e oughness le els eached when angency c i e ia is accomplished (i.e. uns able c ack g ow h). F ac u e is es ima ed o he expe imen ally measu ed lexu al s eng h alues o 3287MPa and 2570MPa o 6UF and 15C, espec i ely. Figu e 5: Es ima ed and expe imen ally de e mined c i ical laws sizes. Figu e 6: F ac u e s eng h as a unc ion o ac u e oughness (coo dina e axis) and p obabili y o ailu e (o dina e axis). ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT FIGURES (a) (b) Figu e 1: F ac u e s eng h alues in he o m o Weibull dis ibu ion o s udied ha dme al g ades. The in es iga ed cemen ed ca bide wi h he lowe binde mean ee pa h (a) exhibi s a bi-modal Weibull dis ibu ion whe eas he o he g ades (b) ha e uni-modal Weibull cu es. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT (a) (b) (c) (d) Figu e 2: FESEM mic og aphs o ailu e ini ia ion si es; namely: (a) ca bide agglome a e in he 3UF g ade, (b) po e in he 10UF g ade, (c) abno mal coa se g ain in he 10C g ade and, (d) binde less agglome a e o coa se WC pa icles in he 15C g ade. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT Figu e 3: F ac u e s eng h as a unc ion o he binde mean ee pa h. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT Figu e 4: Es ima ed c ack esis ance cu es o he 6UF and 15C in es iga ed ha dme als ( ull lines). Uns able c i e ion is eached when angency condi ions be ween he R-cu e and he applied s ess in ensi y ac o (dash-do lines) a e ul illed. Dashed lines indica e he subc i ical c ack g ow hs leng hs and e ec i e oughness le els eached when angency c i e ia is accomplished (i.e. uns able c ack g ow h). F ac u e is es ima ed o he expe imen ally measu ed lexu al s eng h alues o 3287MPa and 2570MPa o 6UF and 15C, espec i ely. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT Figu e 5: Es ima ed and expe imen ally de e mined c i ical laws sizes. ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT Figu e 6: F ac u e s eng h as a unc ion o ac u e oughness (coo dina e axis) and p obabili y o ailu e (o dina e axis). ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT TABLES Vw .% dWC (MPa) Cwc (MPa) λCo (μm) 3UF 3 0.37 ± 0.22 0.74 ± 0.42 0.09 ± 0.05 6UF 6 0.40 ± 0.21 0.61 ± 0.12 0.12 ± 0.06 10UF 10 0.39 ± 0.19 0.46 ± 0.06 0.16 ± 0.06 11M 11 1.12 ± 0.71 0.38 ± 0.07 0.42 ± 0.28 10C 10 2.33 ± 1.38 0.31 ± 0.11 0.68 ± 0.48 15UF 15 0.47 ± 0.22 0.36 ± 0.02 0.24 ± 0.11 15C 15 1.70 ± 1.08 0.27 ± 0.07 0.77 ± 0.54 22M 22 1.64 ± 0.75 0.19 ± 0.04 1.13 ± 0.56 Table I. Mic os uc u al pa ame e s o in es iga ed cemen ed ca bides. ACCEPTED MANUSCRIPT