d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) 252–258

Available online at www.sciencedirect.com

jo ur n al homep age : w ww.int l .e lsev ierhea l th .com/ journa ls /dema

Simulation of cumulative damage associated with long termcyclic loading using a multi-level strain accommodatingloading protocol

Moustafa Nabil Aboushelib ∗

Dental Biomaterials Department, Faculty of Dentistry, Alexandria University, Egypt

a r t i c l e i n f o

Article history:

Received 4 December 2011

Received in revised form 7 May 2012

Accepted 16 October 2012

Keywords:

Fractography

Fracture

Cyclic loading

Crack

Zirconia

a b s t r a c t

Objective. To assess step by step the associated cumulative damage introduced in zirconia

veneered restorations after long term cyclic loading using a new multi-level strain accom-

modating loading protocol.

Methods. 40 zirconia veneered crowns received thermal and cyclic loading (3.5 million cycles

at maximum load of 25 kg representing 70% of the critical load of the veneer ceramic). The

used loading protocol allowed for reproduction of the combined damping action of the peri-

odontal ligament, food substance, jaw deformation, and free movement of the mandibular

joint. Speed of load application and release was obtained from the chewing cycle of adult

patients. Principles of fractographic analysis were used to study the behavior and origin of

critical crack and associated structural damage.

Results. The multi-level strain damping effect prevented generation of cone cracks and con-

tact damage under the loading indenter commonly associated with fracture strength tests.

29 specimens (73%) survived 3.5 million cycles without fracture, 9 specimens (22%) demon-

strated cohesive fracture of the veneer ceramic and limited axial fracture of the framework

was observed in two specimens (5%). Of all fractured specimens, 2 restorations (5%) failed

after 500,000 cycles while the rest survived at least 3 million cycles before fracture was

observed. Fractographic analysis revealed initial wear and abrasion below the loading area,

subsurface micro-cracking of the glass matrix followed by slow crack growth that traveled

in a stepping pattern till deflection at zirconia veneer interface.

Significance. Cyclic loading using multi-level strain accommodating model can reproduce

clinical failure. With exception to manufacturing errors, zirconia veneered restoration sur-

vived a simulated 7-year service time without fracture.

emy

as close as possible the loading mechanism observed in the

© 2012 Acad

1. Introduction

Fracture strength tests of anatomically shaped restorationsare commonly used in dental literature to assess their load

∗ Correspondence address: Dental Biomaterials Department, Faculty of

Tel.: +201090020505; fax: +2035600466.E-mail address: info@aboushelib.org

0109-5641/$ – see front matter © 2012 Academy of Dental Materials. Puhttp://dx.doi.org/10.1016/j.dental.2012.10.009

of Dental Materials. Published by Elsevier Ltd. All rights reserved.

bearing capacity. While the aim of these tests is to mimic

Dentistry, Champolion street, Alexandria, Egypt.

oral cavity, they remain far from that goal. Several key pointswere previously addressed in order to reproduce more realis-tic results as the shape of the loading indenter, presence of

blished by Elsevier Ltd. All rights reserved.

2 9

asNoi

ocfriotcobbifcivc

litrtotirbeaioiiaia[

fttafls[iasun

ca

d e n t a l m a t e r i a l s

stress distribution cushion, loading angle and speed, repre-enting the periodontal ligament, and presence of water [1].evertheless, neither the failure load nor the facture patternsbserved in fracture strength tests mimicked those observed

n clinical failure [2].Two important factors are responsible for the previously

bserved differences. First, it has to be noted that under clini-al conditions, the restoration is subjected to average chewingorces which have been accurately estimated in differentegions of the mouth [3]. These forces change continuouslyn magnitude and direction following the dynamic movementf the mandible [4]. Most importantly, these forces are dis-ributed over the occlusal table of the restoration and pointontact with the antagonist cups is prevented by the presencef the food substance. In case of clenching or intentional heavyiting, the loaded restoration is protected from excessive loadsy the vertical contacts obtained from neighboring teeth dur-ng maximum inter-cuspation [5,6]. Secondly, maximum peakorces are observed only for a fraction of the second in everyhewing cycle and immediately followed by quick relief. Evenn presence of an unexpected hard object, jaw closure pre-enting mechanism is immediately activated to prevent pointontact, otherwise traumatic fracture is observed [7].

Under laboratory conditions, a continuously increasingoad, controlled by the crosshead speed of the loading device,s delivered by the loading indenter over a selected loca-ion of the restoration until fracture is observed. The loadedestoration is actually squeezed between two solid fixtures,he indenter from one hand and the attachment unit on thether hand. With a continuously increasing load over a rela-ively longer loading time, the stress breaking function of thentermediate cushion, if placed between the indenter and theestoration, or the artificially made periodontal ligament wille lost as both would have been compressed beyond theirlastic limit ending in point contact with the indenter. Asxial loading continues, marked surface damage under load-ng indenter is observed ending in sudden explosive failuref the restoration at unrealistically very high loads [8]. Know-

ng that clinically failed dental restorations are usually brokenn two or three intact pieces without marked occlusal damagend at much lower loads [9], fracture strength values observedn the laboratory could only be used for screening purposesnd not to reflect insight on the expected clinical performance10].

Clinical failure, with exception of traumatic and impactractures, is a function of several parameters that interact allogether leading to a slow process of damage accumulation inhe fractured restoration. Cyclic loading at an estimated aver-ge load and under the influence of a chemically and thermallyuctuating environment starts to weaken the restoration bytressing its internal structure especially at interface regions11]. While the influence of a single individual load cycle is def-nitely negligible, repeated cycles for hundreds of thousandsnd even millions have a marked deteriorating effect. Survivaltatistics can predict expected failure for a percentage of pop-lation, restorations in this case, if given loading parameters,

umber of cycles, and some material properties [12].

Luckily enough, the brittle nature of ceramics allow pre-ise recording of all fracture events on the fractured surfacend when the process is relatively slow; the fractured surface

( 2 0 1 3 ) 252–258 253

becomes a time map of the entire process. Fractographic anal-ysis of fractured specimen can predict with high accuracy thesize and location of critical crack, presence of water, relativenumber of loading cycles, entire pathway of the crack frommoment of origin to later deflections, presence of regionalstresses, and most importantly the load at failure if the frac-ture toughness of the material was previously determined [13].

In the present study, general guidelines were addressedin order to present a useful model for a laboratory fracturestrength test of anatomically shaped restorations. First, longterm cyclic loading at sub-critical loads was applied for rela-tively a high number of cycles. Following a review of literaturean average chewing load of 25 kg was selected in this studyrepresenting almost 70% of the maximum biting force in theanterior region of the mouth [3,14,15]. A suggested servicetime of seven years was selected as a plateau level indicatingsuccess of the restoration if no fracture was observed. It wasestimated that a normal adult would perform an average of500,000 loading cycles/year. Thus a 7-year service time wouldrequire 3.5 million cycles [16].

Secondly, stress concentration was prevented at any siteof the restoration using a multi-level strain accommodatingprotocol which allowed for distribution of the loading strain(compression deformation associated with load application)over different components:

- 0.6 mm artificially created periodontal ligament using heavyconsistency polyether impression material (Impregum, 3MESPE, St. Paul, MN, USA) injected in the supporting socketbefore seating the root portion of the resin dies. This layerpermitted macro-movement of the loaded specimens giv-ing room for limited change in angle direction between therestoration and the loading indenter [17].

- 0.15 mm thick cement film thickness which prevented sharpcontact points between the inner surface of the zirconiaframework and the supporting tooth. These misfit contactscould lead to generation of internal Hoope stresses dur-ing loading resulting in barrel type fracture. Moreover, acontinuous and even cement film thickness would providebetter stress distribution under the cemented restorationsespecially at sharp angles as cusp tips or occluso-axial lineangles.

- 2 cm heavy spring coil inserted between the loading inden-ter and the load cell. This spring acted as a piston andallowed the absorption of impact forces delivered by theloading indenter during every load cycle thus protectingthe restoration from sudden rise in loading forces and sub-sequent surface contact damage. Additionally, the springmaintained continuous contact with the loaded restorationwhich prevented excessive abrasion, considering the highnumbers of selected loading cycles.

- 0.5 mm high tensile strength (17.3 MPa) neoprene rub-ber sheet (Rubber sheet roll, Shippenburg, PA, USA) wasinserted between the loading indenter and the surface ofthe restoration to prevent sharp contacts and to representthe consistency of the food substance.

A third key point to consider is the anatomy of the loadingcycle. Dynamic tracings of the mandibular movements werepreviously analyzed and time–space–speed plots have been

l s 2 9 ( 2 0 1 3 ) 252–258

Fig. 1 – Schematic diagram illustrating test setup used toload the specimens.

254 d e n t a l m a t e r i a

accurately presented [18]. Unfortunately, representing thesecomplicated movements require advanced chewing simula-tors which are not available for the majority of dental studies.Nevertheless, a simple tear drop shaped loading cycle thataccounts for speed of load application and removal (0.9 cycle/s)could be incorporated into any loading device. A slow load-ing frequency (40–45 Hz) was chosen for this study to insurefull compression of the heavy spring and full applicationof the selected load. The test started with a short thermo-cycling program (1000 cycles between 6 and 55 ◦C) followedby 10,000 load cycles at 5 kg to allow for proper seating of theloaded restoration before commencing with the determinedload (accommodation phase).

Finally, simulation of chemically active environment iscrucial for biomechanical degradation of material properties.Water acts as a lubricant, cooling medium, hydrolytic agent,and its rule in the process of slow crack growth cannot beover looked. Presence of water at sharp crack tip facilitatesadvancement of crack front compared to failure observedunder inert environment [19].

The aim of the presented work was to assess accumulatingdamage and nature of fracture process during cyclic loading ofbilayered zirconia restorations in an attempt to mimic clinicalfailure and to understand failure process.

2. Materials and methods

2.1. Fabrication of bilayered zirconia restorations

A maxillary central incisor received a full crown preparationaccording to the following criteria: 1.5 mm incisal reduction,1.2 mm axial reduction, and 0.9 mm round chamfer finish line.Resin dies (duplicates) were obtained using polyether impres-sion material (Impregum Penta, 3M ESPE, Seefeld, Germany)and posterior composite resin restorative material (Z250, 3MESPE). The resin dies were scanned and standardized zirconiaframeworks were prepared using a CAD/CAM system (Cercon,Degudent, Hanau Wolfgang, Germany). The sintered copingswere veneered using press-on ceramic material (Ceram Press,Dugodent) according to the anatomy of the unprepared cen-tral incisor. All restorations were cemented on the dies using aresin cement (Panavia F 2.0, Kuraray, Osaka, Japan) which waslight polymerized using high intensity light emitting diodeunit (Blue Phase C9, Ivoclare Vivadent, Shaan, Liechtenstein).

2.2. Cyclic loading and biomechanical degradation

The cemented restorations received cyclic loading (3.5 millioncycles) at sub-critical load (25 kg) starting with the accom-modating loading phase (5 kg applied for 10,000 cycles). Theload was applied in sinusoidal pattern representing a slowchewing rate at 40 cycles/min [20]. A stainless steel inden-ter (3 mm spherical tip) that provided 0.8 mm2 contact areawith the loaded restoration was changed every 50,000 cycles.Generation of cone cracks was prevented using the previ-

ously discussed multi-level strain accommodating protocol.The suggested loading protocol was performed using a com-puter controlled custom made pneumatic testing machinewhich was composed of three loading stations each equipped

with a pneumatic driven piston and an electrical timer con-trolling speed of load application. Pressure gauges controlledmaximum delivered force which was calibrated using a digitalload cell for every loading station, Fig. 1.

A thermally controlled water bath (37 ◦C) was incorporatedto ensure continuous immersion of every loaded speci-men in the following mediums which were changed every100,000 cycles: water, table vinegar (10% acetic acid, pH 2.4),0.1 mol buffered sodium hydroxide (pH 13), and 6% alcohol(pH 7.4) [21].

2.3. Assessment of accumulating structural damage

The fracture surface of each broken restoration was imme-diately ultrasonically cleaned, dried, gold sputter coated,and examined under scanning electron microscope (XL30,Phillips, Eindhoven, the Netherlands). Small fragments andsmall veneer chips, separated from the restorations demon-strated basically mirror surfaces and were thus not includedin fractographic analysis. In case a restoration survived 3.5million cycles without fracture, the test was stopped andthe restoration was axially sectioned using a diamond coatedsaw and a precision cutting machine (MicraCut 120, Metkon,Germany). Cut sections were examined at different magnifi-cations and angles to evaluate structural damage. Grindingdamage produced during sectioning procedure was character-ized by parallel lines in a single flat plane which were much

different from the stepping pattern observed for slow crackgrowth landmarks.

d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) 252–258 255

Table 1 – Failure type and percentage of tested restorations.

Number Percentage Cycles Observation

29 73% 3,500,000 Survived with minor surface wear below loading point.After sectioning, radial cracks and Hoop stress failure were observedin seven specimens.

9 22% 3,000,000 Four clinically restorable cohesive fracture of the veneer ceramic fromintact zirconia framework.Five extensive un-restorable cohesive fracture of the veneer ceramic.Limited axial fracture of the zirconia framework due to structuraldefects.

3

2tvo2s(

esofttv

mi(fd

vitpnHF

4

AiuAcie

latam

Fig. 2 – (A) Digital image, 18×, of gold sputtered fracturedspecimen demonstrating wear and abrasion at the loadingsite and subsurface critical crack. (B) SEM image, 120×, ofthe previous specimen demonstrating stepping pattern ofthe slow crack growth followed by mirror, mist, and hackle

2 5% 500,000

. Results

9 specimens (73%) survived 3.5 million cycles without frac-ure, 9 specimens (22%) demonstrated cohesive fracture of theeneer ceramic and limited fracture of the framework wasbserved in two specimens (5%). Of all fractured specimens,

restorations (5%) failed after 500,000 cycles while the resturvived at least 3 million cycles before fracture was observedTable 1).

Fractographic analysis revealed how cumulative damagended in fracture of the specimens. First, initial wear and abra-ion were observed below the loading area (Fig. 2A) in the formf mechanical abrasion of the surface of the veneer ceramic,ollowed by subsurface micro-cracking of the glassy matrix ofhe veneer ceramic (Fig. 2B), and eventually slow crack growthhat traveled in a stepping pattern till deflection at zirconiaeneer interface.

The cohesive facture observed in the majority of the speci-ens had specific characteristic features starting by minimal

ncisal damage in the form of minor chipping or abrasionFig. 3A), initial crack origin located below the loading sur-ace (Fig. 3B), and multiple arrest lines propagating in cervicalirection (Fig. 3C).

After careful sectioning of intact specimens which sur-ived 3.5 million cycles, additional three types of cracks weredentified; radially propagating vertical crack originating fromhe zirconia resin interface and directed towards the loadingoint (Fig. 4A), radially propagating crack originating at zirco-ia veneer interface and traveling in occlusal direction, andoope stress fracture due to contact misfit on the axial wall,ig. 4B and C.

. Discussion

fter seven years simulation of clinical function, the major-ty of zirconia restorations survived 3.5 million loading cyclesnder the influence of a chemically active environment.mazingly, the survival rate of the tested specimens wasomparable to a previous clinical [22,23] and laboratory stud-es [24,10] keeping in mind that only fracture was taken as anvaluation parameter.

Fractographic analysis of the veneer ceramic under theoading site revealed step by step how cumulative damage

ssociated with long term cyclic loading ended in final frac-ure. Initially, signs of wear, abrasion, and minor chippingt the incisal edge were observed followed by subsurfaceicro-cracking of the glassy matrix, Fig. 2A and B. A similar

pattern of the final critical crack.

pattern was observed after examination of clinically frac-tured all-ceramic crowns which also revealed presence ofsigns wear of the veneer ceramic, edge chipping, and arrestlines which occurred as secondary events associated withpropagation of crack front [25,26].

Sectioning of intact specimens after completion of cyclicloading, revealed how fracture proceeded from the estab-

lished surface and subsurface damage below the loadingpoint, Fig. 3A. With continuation of cyclic loading, the regionlocated between the loading point and the underlying frame-work becomes subjected to complex stresses (compression,

256 d e n t a l m a t e r i a l s 2 9 ( 2 0 1 3 ) 252–258

Fig. 3 – (A) SEM image, 30×, demonstrating signs or wearand abrasion at the incisal edge, initial micro-cracking ofthe veneer ceramic, and localized edge chipping. (B) SEMimage, 20×, of a cut specimen demonstrating intact loadingsurface of the veneer ceramic and subsurface slow crackgrowth which was the most common failure mode of thetested specimens. (C) SEM image, 38×, of the previousspecimen demonstrating the stepping pattern of the slowcrack growth after 3.5 million cycles. Notice proximity ofthe last arrest lines indicating acceleration of slow crackgrowth before final fracture of the specimen.

Fig. 4 – (A) SEM image, 750×, revealing radial crackascending from zirconia resin interface causing fracture ofthe framework. (B) Cut section of a specimendemonstrating sharp contact point between zirconiaframework and the supporting tooth. (C) Digital image ofthe previous specimen demonstrating Hoop stress failure atthe sharp contact point.

2 9

tpatitfttl

fadofermoftf[

trcbisld

cucspipsldft

rfiifbwdca

tip

r

[13] Lawn B, Sanjit B, Bush QDZ. Failure modes in ceramic-based

d e n t a l m a t e r i a l s

ension, and shear) leading to weakening of materialroperties. Eventually, a critical crack develops, Fig. 3B,nd the crack front reaches a state of equilibrium betweenhe amount of induced damage, applied load at crack tip, andnternal material resistance to crack propagation (fractureoughness). With subsequent loading cycles the complexorces moves deeper in the material and the crack frontravels a further step to reach a new balance state. Repeti-ion of this process leaves behind the characteristic arrestines, Fig. 3C.

The last event of fracture is rapid deflection of the crackront leading to cohesive fracture of the veneer which maylso occur at core veneer interface. Presence of structuralefects (milling defects, sharp scratches, or material flaws)r entrapped residual stresses in this region would facilitateaster crack movement. A point worth mentioning is that thestimated failure stress of clinically fractured glass ceramicestorations was in range of 31–40 MPa [27] which is in good

atch with the estimated tensile strength of different typesf veneer ceramics [28]. On the contrary, the estimated load atailure of the veneer ceramic under one cycle load to failureesting was much higher, 52 MPa, indicating over estimation ofailure load associated with laboratory fracture strength tests7].

Kim in 2009 indicated that during axial loading penetra-ion of cone cracks into the occlusal surface of the loadedestoration was the predominant fracture mode of the veneereramic. Generation of cone cracks requires over loading of therittle veneer with a hard indenter which is not observed clin-

cally [4]. Cone cracks were not observed in any of the testedpecimens. Thanks to the multi-level strain accommodatingoading protocol which prevented generation of peak stressesuring loading of the brittle veneer.

Under a wet environment, presence of water at sharprack tip directly influences crack size and the load at fail-re knowing that fracture toughness of the material remainsonstant. In a previous study, specimens tested under oil hadmaller crack sizes and required higher fracture loads com-ared to specimens tested under water [29]. This observation

s expected to have a stronger effect when cyclic loading iserformed for a longer period. The crack size observed in thistudy was relatively larger in dimension compared to one cycleoad to failure experiments performed in the lab, indicatingegradation in mechanical properties under the influence ofatigue [7]. These data will be presented in the second part ofhis study.

Cyclic loading of different types all-ceramic restorationsesulted in significant reduction of their fracture strength. Thisnding was observed for glass ceramic crowns [11,30], glass

nfiltrated restorations [31], as well for polycrystalline aluminarameworks [32], but zirconia frameworks were not influencedy cyclic loading at higher loads which is in contradictionith the finding of the present investigation [24,33,34]. Thisifference could be directly related to the limited number ofycles used in these studies and to the absence of chemicalttach.

Oxide ceramics are sensitive to presence of water at crack

ip which could lead to reduction in surface energy thus facil-tating further crack growth. Moreover, water could exert aumping pressure at crack tip during every loading cycle,

( 2 0 1 3 ) 252–258 257

have a lubricating effect, and wash-off material debris alltogether enhancing slow crack growth [35]. It was interest-ing to observe three additional types of cracks leading tofracture of the zirconia framework; radial crack originat-ing at zirconia resin interface, radial crack originating atzirconia veneer interface, and Hoop stress crack at the axialwall, Fig. 4.

5. Conclusion

The proposed long term cyclic loading using multi-level strainaccommodating model can reproduce clinical failure. Withexception to manufacturing errors, zirconia veneered restora-tion survived a simulated 7-year service time without fracture.

e f e r e n c e s

[1] Kelly J. Clinically relevant approach to failure testing ofall-ceramic restorations. The Journal of Prosthetic Dentistry1999;81:652–61.

[2] Kelly JR. Approaching clinical relevance in failure testing ofrestorations. The Journal of Prosthetic Dentistry1999;81:652–61.

[3] Morneburg TR, Proschel PA. Measurement of masticatoryforces and implant loads: a methodologic clinical study. TheInternational Journal of Prosthodontics 2002;15:20–7.

[4] Kim JW, Kim JH, Janal MN, Zhang Y. Damage maps ofveneered zirconia under simulated mastication. Journal ofDental Research 2008;12:1127–32.

[5] Geckili O, Bilhan H, Mumcu E, Tuncer N. The influence ofmaximum bite force on patient satisfaction and quality oflife of patients wearing mandibular implant over dentures.The Journal of Oral Implantology 2009;14:123–6.

[6] Tartaglia GM, Testori T, Pallavera A, Marelli B, Sforza C.Electromyographic analysis of masticatory and neckmuscles in subjects with natural dentition, teeth-supportedand implant-supported prostheses. Clinical Oral ImplantsResearch 2008;19:1081–8.

[7] Aboushelib MN, Feilzer AJ, Kleverlaan CJ. Bridging the gapbetween clinical failure and laboratory fracture strengthtests using a fractographic approach. Dental Materials2009;25:383–91.

[8] Lawn B, Deng Y, Thompson V. Use of contact testing in thecharacterization and design of all-ceramic crownlike layerstructures: a review. The Journal of Prosthetic Dentistry2001;86:495–510.

[9] Quinn JB, Quinn GD, Kelly JR, Scherrer SS. Fractographicanalyses of three ceramic whole crown restoration failures.Dental Materials 2005;21:920–9.

[10] Coelho PG, Silva NR, Bonfante EA, Guess PC, Rekow ED,Thompson VP. Fatigue testing of two porcelain–zirconiaall-ceramic crown systems. Dental Materials 2009;25:1122–7.

[11] Attia A, Kern M. Influence of cyclic loading and luting agentson the fracture load of two all-ceramic crown systems. TheJournal of Prosthetic Dentistry 2004;92:551–6.

[12] Mitov G, Lohbauer U, Rabbo MA, Petschelt A, Pospiech P.Investigations of subcritical crack propagation of theEmpress 2 all-ceramic system. Dental Materials2008;24:267–73.

layer structures: a basis for materials design of dentalcrowns. Journal of the American Ceramic Society2007;90:1671–83.

l s 2

258 d e n t a l m a t e r i a

[14] Curtis AR, Wright AJ, Fleming GJ. The influence of simulatedmasticatory loading regimes on the bi-axial flexure strengthand reliability of a Y-TZP dental ceramic. Journal ofDentistry 2006;34:317–25.

[15] Conserva E, Menini M, Tealdo T, Bevilacqua M, Pera F, RaveraG, et al. Robotic chewing simulator for dental materialstesting on a sensor-equipped implant setup. TheInternational Journal of Prosthodontics 2008;21:501–8.

[16] Teixeira EC, Piascik JR, Stoner BR, Thompson JY. Dynamicfatigue and strength characterization of three ceramicmaterials. Journal of Materials Science Materials in Medicine2007;18:1219–24.

[17] Rees JS, Jacobsen PH. Elastic modulus of the periodontalligament. Biomaterials 1997;18:995–9.

[18] Grigoriadis A, Johansson RS, Trulsson M. Adaptability ofmastication in people with implant-supported bridges.Journal of Clinical Periodontology 2011;38:395–404.

[19] Kohorst P, Dittmer MP, Borchers L, Stiesch-Scholz M.Influence of cyclic fatigue in water on the load-bearingcapacity of dental bridges made of zirconia. ActaBiomaterialia 2008;4:1440–7.

[20] Throckmorton GS, Buschang BH, Hayasaki H, Phelan T. Theeffects of chewing rates on mandibular kinematics. Journalof Oral Rehabilitation 2001;28:328–34.

[21] Mirmohammadi H, Aboushelib MN, Salameh Z, KleverlaanCJ, Feilzer AJ. Influence of enzymatic and chemicaldegradation on zirconia resin bond strength after differentsurface treatments. American Journal of Dentistry2010;23:327–30.

[22] Sailer I, Feher A, Filser F, Gauckler LJ, Luthy H, Hammerle CH.Five-year clinical results of zirconia frameworks forposterior fixed partial dentures. International Journal ofProsthodontics 2007;20:383–8.

[23] Sailer I, Feher A, Filser F, Luthy H, Gauckler LJ, Scharer P,et al. Prospective clinical study of zirconia posterior fixedpartial dentures: 3-year follow-up. QuintessenceInternational 2006;37:685–93.

[24] Att W, Stamouli K, Gerds T, Strub JR. Fracture resistance ofdifferent zirconium dioxide three-unit all-ceramic fixedpartial dentures. Acta Odontologica Scandinavica2007;65:14–21.

9 ( 2 0 1 3 ) 252–258

[25] Scherrer SS, Quinn GD, Quinn JB. Fractographic failureanalysis of a Procera AllCeram crown using stereo andscanning electron microscopy. Dental Materials2008;24:1107–13.

[26] Scherrer SS, Quinn JB, Quinn GD, Kelly JR. Failure analysis ofceramic clinical cases using qualitative fractography. TheInternational Journal of Prosthodontics 2006;19:185–92.

[27] Taskonak B, Yan J, Mecholsky Jr JJ, Sertgoz A, Kocak A.Fractographic analyses of zirconia-based fixed partialdentures. Dental Materials 2008;24:1077–82.

[28] Aboushelib MN, de Jager N, Kleverlaan CJ, Feilzer AJ.Microtensile bond strength of different components of coreveneered all-ceramic restorations. Dental Materials2005;21:984–91.

[29] Taskonak B, Griggs JA, Mecholsky Jr JJ, Yan JH. Analysis ofsubcritical crack growth in dental ceramics using fracturemechanics and fractography. Dental Materials 2008;24:700–7.

[30] Attia A, Abdelaziz KM, Freitag S, Kern M. Fracture load ofcomposite resin and feldspathic all-ceramic CAD/CAMcrowns. The Journal of Prosthetic Dentistry 2006;95:117–23.

[31] Borges GA, Caldas D, Taskonak B, Yan J, Sobrinho LC, deOliveira WJ. Fracture loads of all-ceramic crowns under wetand dry fatigue conditions. Journal of Prosthodontics2009;18:649–55.

[32] Blatz MB, Oppes S, Chiche G, Holst S, Sadan A. Influence ofcementation technique on fracture strength and leakage ofalumina all-ceramic crowns after cyclic loading.Quintessence International 2008;39:23–32.

[33] Zaghloul HH, Younis JF. Marginal fit of implant-supportedall-ceramic zirconia frameworks. The Journal of OralImplantology 2012, ahead of publications.

[34] Wolfart S, Ludwig K, Uphaus A, Kern M. Fracture strength ofall-ceramic posterior inlay-retained fixed partial dentures.Dental Materials 2007;23:1513–20.

[35] Chevalier J, Deville S, Fantozzi G, Bartolome JF, PecharromanC, Moya JS, et al. Nanostructured ceramic oxides with a slowcrack growth resistance close to covalent materials. NanoLetters 2005;5:1297–301.