The effect of finish line preparation and layer thickness on the failure load and fractography of ZrO2 copings

Sven Reich, DMD, PhD,a Anselm Petschelt, DMD, PhD,b and Ulrich Lohbauer, MSc, PhDc University of Leipzig, Leipzig, Germany; University of ErlangenNuremberg, Erlangen, Germany
Statement of problem. To prevent tooth weakening or pulp irritation, there is a need for a minimally invasive method of preparing single anterior crowns. Restoration dimensions for reduced coping thicknesses or less invasive finish line preparations are required. Purpose. The purposed of this in vitro was to study investigate the fracture performance of high-strength zirconia copings, compare knife-edge margins with chamfer finish lines, and examine the effect of reducing the layer thickness from 0.5 mm to 0.3 mm. Material and methods. Y-TZP zirconia copings were manufactured on brass dies of a maxillary central incisor. Forty copings, with 2 layer thicknesses (0.5 and 0.3 mm), and 2 finish line preparations (knife edge and chamfer; n=10) were cemented using a conventional glass ionomer cement and stored in distilled water at 37C for 24 hours. The copings were vertically loaded until fracture using a universal testing machine. Data were analyzed by 2-way ANOVA (=.05). Fractographic examination was performed using scanning electron microscopy and confocal laser scanning microscopy. Results. A significantly higher mean failure load was measured for knife-edge (0.5 mm, 1110 175 N; 0.3 mm, 730 160 N) versus chamfer (0.5 mm, 697 126 N; 0.3 mm, 455 79 N) preparations (P<.001), and for 0.5-mm versus 0.3-mm thickness layers (P<.001). Conclusions. Knife-edge preparations present a promising alternative to chamfer finish lines; the fracture load required for knife-edge preparations was 38% greater than that required for chamfer preparations, regardless of coping thickness. Reducing the thickness of a single crown coping from 0.5 to 0.3 mm resulted in a 35% reduction in fracture load required for either preparation type. (J Prosthet Dent 2008;99:369-376)

Clinical Implications

In terms of the mechanical resistance of zirconia copings against fracture, knife-edge preparations present a promising alternative to chamfer finish lines. Within the limits of this in vitro study, knife-edge preparations in combination with zirconia core material should be considered when a minimally invasive preparation design in the anterior region is indicated.

Currently, dental prosthetic treatments follow principles based on preserving sound tissue,1 generally requiring the removal of limited amounts of sound tooth structure,
a

including axial reduction and the beveling of finish lines. To provide a restoration with the necessary retention and resistance, the axial walls must be slightly tapered and cement must be

placed in compression.2-5 Modern adhesive technologies and high strength ceramic materials with enhanced fracture toughness may facilitate the development of minimally invasive

Associate Professor, Department of Prosthodontics, University of Leipzig. Professor and Chair, Operative Dentistry and Periodontology, University of Erlangen-Nuremberg. c Materials Science Lab Supervisor, Operative Dentistry and Periodontology, University of Erlangen-Nuremberg.
b

Reich et al

370
preparation techniques.6 Zirconia (ZrO2) has recently been introduced as a promising metal-free core structure for fixed prostheses.7,8 One recent study reported that, since 1998, 15 studies have demonstrated a 90% or greater success rate using zirconia-based prostheses.9,10 Zirconia with a 0.5-mm layer thickness served as a framework for primarily posterior 3- to 5-unit fixed partial dentures. Failures were attributed to cracking or crazing of the veneering porcelain, but bulk fractures were uncommon.9 Tetragonal zirconia polycrystal (TZP) is commonly used as the core material in which the metastable tetragonal phase is stabilized by compounding 3 mol% yttrium oxide (3Y-TZP).9 TZP uses the tetragonal-tomonoclinic phase transformation of zirconia particles under load, resulting in enhanced fracture toughness.8,11 Numerous studies have reported superior fracture strength and toughness data for TZP.9,11-13 The underlying mechanisms that allow for additional energy dissipation at developing crack tips are stress-induced transformation of the crystalline phase, followed by a 3-5% volume expansion and nucleation of microcracks.11,13 The development of computeraided design/computer-aided manufacturing (CAD/CAM) technology has focused on precise and consistent manufacturing of zirconia ceramics with high strength and toughness. CAD/CAM technology relies on exact dimensional predictions to compensate for sintering shrinkage, and is an economical and highly reproducible method for manufacturing complex and individual geometries out of a green or presintered ceramic material.7 The clinical success of zirconiabased posterior prostheses reaffirms the structural potential and processing possibilities of this material.9 The superior mechanical properties of zirconia allow clinicians to reconsider established preparation guidelines for the design of single anterior teeth copings, and consider variations such as reducing the coping thickness from 0.5 mm to 0.3 mm and changing finish line preparations from chamfer (CHA) to minimally invasive knifeedge (KNE) margins. In vitro research evaluating the influence of processing variables on fracture resistance of all-ceramic restorations has revealed highly divergent failure loads of 450 to 1600 N for zirconia single crown copings, depending on coping thickness, marginal design, and applied luting agent.14-16 In general, higher fracture loads can be produced by increasing the coping thickness or by using adhesive luting materials instead of retentive cementation.4,5,17 However, conventional cementation is recommended by the manufacturers.17 Improved resin adhesion has been reported when using tribochemical silica coating and silanization of the zirconia frameworks, or when using phosphate-modified resin luting agents.18 Fracture resistance evidence can also be provided by microscopic fracture analysis. Fractography, in this context, represents an analytical tool for assessing fractured surfaces for the purpose of locating fracture origins and elucidating fracture patterns.19,20 Reasons for fracture can be assessed from different crack modes such as cone, radial, or fatigue crack pathways.21,22 One minor influence on fracture resistance is the design of the finish line.23 However, marginal integrity and the degree of seating of a restoration is dependent upon the finish line of the preparation. The configuration of the gingival margins dictates the shape and bulk of the restorative material. The preferred gingival finish line for metal ceramic restorations is the CHA preparation.2 For all-ceramic bilayer restorations, the CHA and shoulder preparation are recommended.2,7 These finish line configurations are reported to transfer a minimum of masticatory stress from the coping into the veneering porcelain, which in turn might help to reduce clinical failures due to crazing, cracking, or chipping of the veneering porcelain.2,23-26

Volume 99 Issue 5
In contrast, the KNE preparation provides the most acute marginal finish line by preserving a maximum of sound tissue. Thin, tapered margins present disadvantages such as difficult processing accuracy and susceptibility to clinical chipping fractures.2 However, to prevent tooth weakening or pulp irritation, it might be necessary to perform minimally invasive treatment procedures. As a consequence, restoration dimensions with reduced coping thicknesses or less invasive finish lines, including KNE, are needed.2,9 For example, circumferential or partly KNE preparations are indicated for lingual surfaces of mandibular posterior teeth, for teeth with convex axial surfaces, and for inclined tooth surfaces.2,27 One indication for KNE finish line crown preparations is the severely decayed, endodontically treated anterior tooth with a restored post-and-core foundation. The use of a protective ferrule at the coronal surface has been proposed to improve the integrity of an endodontically treated tooth.27 Therefore, the crown should extend 1-2 mm beyond the tooth-core junction, which might be most suitably realized by a KNE compared to a CHA or shoulder preparation. Furthermore, for young vital teeth, a KNE preparation might be a less invasive alternative to a CHA margin. The purpose of this study was to evaluate whether high-strength zirconia core materials were suitable for CHA or KNE finish line preparations, and how reducing the layer thickness influenced the resulting fracture resistance of in vitro prepared copings. This study tested a twofold null hypothesis, that neither the layer thickness nor the finish line design significantly affected the fracture load of cemented zirconia copings.

MATERIAL AND METHODS

This study used a model human maxillary central incisor (tooth 11; KaVo Dental, Biberach, Germany) and CAD/CAM technology to prepare either KNE or CHA margins. For

The Journal of Prosthetic Dentistry

Reich et al

May 2008
the KNE preparation, the circumferential reduction was 1.2 mm at the midcoronal area, the incisal reduction was 1.5 mm, and the width of the incisal plateau was 1.2 mm. The internal angles were rounded with a radius of 0.8 mm, and the axial walls were slightly tapered to 4 degrees convergence. Circumferential reduction was measured using a digital caliper (Mitutoyo Corp, Kawasaki, Japan) with an accuracy of 0.01 mm, while finish line angles were optically detected on the master dies under scanning electron microscopy (SEM) using an image-processing software (DISS 5; Point Electronic GmbH, Halle, Germany). Silicone impressions (DUOSIL D; Shera Werkstoff-Technologie GmbH, Lemforde, Germany) of the master die for the KNE preparations (MDKE) were made and duplicated using acrylic resin (Alpha Die; Schuetz Dental GmbH, Rosbach, Germany). The finish line of the duplicate was then modified from a KNE into a CHA finish line by reducing the axial walls but maintaining the convergence angle (4 degrees) of the preparation (master die chamfer: MDC). Thus, 2 master dies with different finish lines and different marginal areas were generated, as shown in Figure 1. To duplicate the master dies, 3-dimensional (3-D) CAD data were generated from replicas cast from gypsum type IV (Fujirock EP, white; GC Europe NV, Leuven, Belgium). The replicas were optoelectronically scanned using a commercial CAD scanner (Lava Scanner; 3M ESPE, Seefeld, Germany), the data were transferred to the CAM procedure (Lava Form; 3M ESPE), and 20 copies of each MDKE and MDC were milled out of brass. The brass dies were optically scanned and 10 copings with a final layer thickness of either 0.3 mm or 0.5 mm were CAD/CAM manufactured out of yttrium-stabilized tetragonal zirconia polycrystals (Lava Frame; 3M ESPE).8 Due to the use of presintered blanks, oversized geometries were milled to compensate for sintering shrinkage of approximately 20% by volume.9 Maximum homogeneity of grain sizes and porosity is required to ensure a constant linear shrinkage and, thus, to calculate the final fit of a coping.8 The presintered copings were placed in a furnace (Lava Therm; 3M ESPE), dried, and sintered according to the manufacturers instructions. The automatic sintering program began after a 3.5-hour drying period, and the furnace was heated to 1500C. The sintering process, including drying time, lasted for approximately 11 hours. The CHA 0.5 preparation (CHA 0.5 = copings with a wall thickness of 0.5 mm on chamfered supporting dies) matched the manufacturers recommendations regarding the CHA finish line, convergence angle of 4 degrees, and a minimum zirconia thickness of 0.5 mm. The brass dies were fixed axially in a PMMA supporting plate (Plexiglas; Evonik Degussa GmbH, Duesseldorf, Germany). The final copings were cemented onto the dies using conventional glass ionomer cement (Ketac Cem; 3M ESPE) by applying a 10 N static load for 10 minutes. The luted specimens were stored for 24 hours in distilled water at 37C. Fracture loads were measured in a universal testing machine (Zwick Z2.5; Zwick, Ulm, Germany) using a 2 kN load cell. Vertical load was applied with a crosshead speed of 0.5 mm/min perpendicular to, and on the central part of, the incisal plateau of the copings until catastrophic fracture occurred (Fig. 2). Tin foil, 0.3 mm thick (Dentaurum, Ispringen, Germany), was placed between the load piston and

371

B
1 Standardized brass dies showing: A, knife-edge versus B, chamfer preparation types (original magnification x15).

Reich et al

372

Volume 99 Issue 5

2 Test apparatus showing axial loading in universal testing machine using tin foil (KNE 0.5). the coping for consistent load transfer. Fracture load data, in N, were analyzed with a 2-way ANOVA (=.05). The fractured zirconia fragments were examined using SEM and confocal laser scanning microscopy (CLSM). The specimens were gold sputtered for 120 seconds in high vacuum (SCD 050; Balzers Instruments, Balzers, Liechtenstein). The SEM (ISI-SR-50; Leitz, Wetzlar, Germany) was operated under 20 kV acceleration voltage. The backscatter electron mode was used for improved surface visualization. The CLSM (TCS SL; Leica Microsystems, Bensheim, Germany) was used in reflecting and fluorescing mode in order to display surface morphologies and artifacts in 3-D, and to perform macroscopic and microscopic failure analysis. To observe microcracks, an aqueous solution of Rhodamine B isothiocyanate (concentration approximately 0.01%; maximum absorption, 540 nm; maximum emission, 625 nm; Merck, Darmstadt, Germany) was applied onto the contact zone of the fractured zirconia copings. An argon ion laser was used at 514 nm excitation wavelength, and emissions were detected using a DD 458/514 band pass filter. Confocal z-sections were made either at 29.8m intervals (Objective: N Plan, x2.5 magnification, numerical aperture (NA) = 0.07; Leica Microsystems) or at 1.6-m intervals (Objective: HC PL Fluotar, x10 magnification, NA = 0.9; Leica Microsystems).

RESULTS
Means and standard deviations are reported in Table I. ANOVA revealed significant effects for both finish line (F=60.425, P<.001) and layer thickness (F=49.248, P<.001), but no significant interaction effect (F=2.409, P>.05), as shown in Table II. Failure load was significantly higher

Table I. Design parameters and failure load data (mean and SD) in N
CHA 0.5
Preparation type Layer thickness in mm Mean failure load (SD) chamfer 0.5 697 (126)

CHA 0.3
chamfer 0.3 455 (79)

KNE 0.5
knife edge 0.5 1110 (175)

KNE 0.3
knife edge 0.3 730 (160)

Table II. Statistical analysis of interaction effects between finish line and layer
thickness (2-way ANOVA). Dependent variable: failure load

Source
Finish line Layer thickness Finish line vs. layer thickness

Type III Sum of Squares

1187409 967764 47339

df
1 1 1

Mean Square
1187409 967764 47339

F
60.4 49.2 2.4

P
<.001 <.001 =.129

The Journal of Prosthetic Dentistry

Reich et al

May 2008
for KNE than CHA preparations, and significantly higher for 0.5-mm than for 0.3-mm thicknesses. The highest fracture load, 1110 175 N, was measured for the KNE copings with a 0.5-mm layer thickness (KNE 0.5), whereas the lowest fracture load, 455 79 N, was measured for the 0.3-mmthick CHA preparation (CHA 0.3). Reducing the layer thickness from 0.5 to 0.3 mm resulted in a 35% reduction in fracture load for CHA 0.5/0.3 and a 34% reduction for KNE 0.5/0.3. When comparing KNE with CHA margins with the same layer thickness, a 38% reduction in the interpreparational fracture load was observed, favoring the KNE preparation. All 40 fractured copings were subjected to macroscopic and microscopic fractographic examination. Figure 3 illustrates typical macroscopic fracture patterns. Both CHA and KNE preparations predominantly fractured into 3 fragments (62.5% of the fractured copings). Closer examination of the contact region between the incisal plateau and the load piston under fluorescent light showed a stained CLSM image with radial and circumferential microcracking, representative of 29% of the fractured copings (Fig. 4) (Table III). In 71% of the investigated situations (n=31), no analyzable fracture patterns were detected. In the majority of specimens (59.3%), SEM located the fracture origin at the intaglio side of the incisal plateau of the copings (Fig. 5, A) (Table III). Fractographic patterns typical of a brittle failure mode were observed on the fracture surfaces (Fig. 5, B). Figure 4 also shows evidence of cone cracking (circumferential microcracks). Figure 6 shows a section close to the gingival margins of the CHA preparation (CHA 0.5) that exhibits fracture patterns indicative of shear stress.

373

3 Typical macroscopic fracture patterns of axially loaded copings in 3 fragments (original magnification x15).

4 Confocal laser scanning microscopy image of incisal plateau after fracture, exhibiting typical microcracks (CHA 0.5, original magnification x25).

Table III. Quantitative analysis of data from microscopic examination of fractured copings
Macroscopic appearance (n=40) Microscopic findings (n=31) Fracture origin site (n=27) 2 fragments: 12.5% 3 fragments: 62.5% >3 fragments: 25%

Radial and circumferential cracking mode: 29% Intaglio surface: 59.3% (n=7 CHA; n=9 KNE)

No analyzable fracture patterns: 71% No explicit allocation: 40.7% (n=5 CHA; n=6 KNE)

Reich et al

374

Volume 99 Issue 5

5 Microscopic determination of fracture origin at surface of incisal plateau of coping. A, Coping overview (original magnification x50). B, Magnification exhibiting fracture origin and typical mist and hackle regions (original magnification x200).

6 Confocal laser scanning microscopy image of fractured coping (CHA 0.5). Section close to gingival margin (bottom right part) exhibits fracture patterns indicative of shear stress (original magnification x100).

DISCUSSION
The twofold null hypothesis was rejected because both layer thickness and finish line design had a significant effect on the in vitro fracture performance of cemented zirconia copings. In general, KNE preparations outperformed copings with CHA margins. The fractured zirconia copings exhibited failure loads between 455 and 1110 N, depending upon coping thickness and finish line. It is difficult

to compare failure loads found in the literature to those found in this study, due to different experimental variables. Potiket et al14 reported a fracture load of 381 N for 0.6-mm-thick zirconia copings placed on noncarious human central incisors and located the fracture origins within the supporting natural tooth. Another study reported fracture loads between 697 and 1607 N for 0.4-mm-thick posterior zirconia copings placed on dental composite resin dies.15 However, here

the authors distinguished between the values for fracture initiation (697 N) and fracture end (1607 N). Other studies of complete premolar or molar crowns reported fracture load values between 980 and 1400 N, depending on the material and luting procedure used.4,5 Bernal et al5 found superior fracture performance using adhesive luting, but no influence as a result of margin geometry. One reason for the inferior fracture performance of the CHA prepa-

The Journal of Prosthetic Dentistry

Reich et al

May 2008
rations as compared to KNE finish lines found in this study could be the luting agent. It has been reported that ceramic restorations that are resinbonded to dentin are more fracture resistant than retentively cemented restorations.4,17 Bernal et al5 reported significantly stronger all-ceramic crowns after adhesive luting as compared to luting with zinc phosphate or glass-ionomer. The authors further investigated the effect of finish line preparation but found no differences in fracture resistance when using a resin luting agent. Finite element analysis showed that there was effective stress transfer between a stiff coping and the supporting structure, such as dentin.23 Adhesive luting materials demonstrated superior performance as compared to zinc phosphate cements, due to a lower elastic modulus of the adhesive luting agent in the range of 1 magnitude (Eadhesive = 1.2 GPa, Ece= 13 GPa).23,25 The copings under ment investigation were luted with a glass ionomer cement that exhibits minor self-adhesive properties.4,5,17 As a consequence, an increased susceptibility to creep and interfacial delamination under vertical load results in stress concentration, particularly at the gingival margins of CHA preparations. Figure 6 shows fracture patterns due to shear stress in the gingival region of a CHA preparation. In addition to vertical loading, the induced shear stress could account for the reduced fracture performance of CHA 0.5 and CHA 0.3 preparations (Table I). Bindl et al15 observed extreme differences between the fracture initiation (697 N) and catastrophic fracture forces (1607 N) due to increased toughness and extensive crack propagation of the cemented copings. Crack propagation within a coping is accelerated by high shear stresses under vertical loading, reducing the catastrophic fracture forces. Postmortem fractographic examination of bilayer all-ceramic crown fragments has shown that the site of failure origin is located within the core material.19 Competing damage modes have been identified, such as cone cracking and quasiplastic yielding at the occlusal surface within the antagonist contact zone, and radial cracking originating at the cementation surface of a restoration.19,21 Damage accumulation and subsequent failures that arise from radial cracks are reported to be the most deleterious, because they develop from the intaglio surface and therefore are seldom detected by clinicians.26 The authors proposed that cone cracking was responsible for chipping fractures and that quasiplasticity drove microcrack coalescence and accelerated wear at the occlusal surfaces, while radial cracking resulted primarily in bulk fractures.26 Radial cracks are reported to contribute to long-term creep within the restorative complex, and, thus, accelerate fatigue due to slow crack growth.22 In accordance with the literature, Figure 5 shows that the fracture origin is located at the intaglio surface of a zirconia coping24; this was the situation for 59.3% of the copings investigated in the current study (n=27; Table III). The brittle fracture origin can be located by identifying fractographic patterns such as mist and hackle regions or by locating various arrest or Wallner lines. These patterns are footprints of a developing crackfront during fracture; for example, rib-shaped Wallner lines represent a wavelike contour in the axis of principal tension originating at discontinuities in the crack plane.20 In contrast, this study found little evidence of cone cracking. Figure 4 shows an example of circumferential cracking, which was detected in 29% of the specimens (n=31; Table III). However, Figure 5 does not support cone-shaped fracture patterns. Nonveneered copings were examined in this study. Lawn et al26 described radial cracking as the predominant failure mode in a bilayer test configuration examining different ceramic layers on a dentin-like polycarbonate substrate. The influence of ceramic core layer thickness and of the elastic constants of the supporting materials are discussed.21,26 While Wakabayashi and Anusavice21 showed increasing fracture resistance with an increasing core-to-veneer thickness ratio for veneered all-ceramic systems, Lawn et al26 reported less dependency on the ratio than on the net thickness of a crown. Based on fractographic examination of the zirconia copings (n=27; Table III), no differences in damage mode were observed between CHA 0.5 and CHA 0.3 or KNE 0.5 and KNE 0.3 preparations, respectively. These observations support the idea that reducedthickness zirconia copings (0.3-0.5 mm) fracture solely due to radial cracking, which has been explained in the literature by a dependency of the incidence of radial cracking mode on ceramic layer thickness.26 However, in a clinical situation, core materials are veneered, resulting in an increased overall thickness of the all-ceramic restoration; this increased thickness might prevent further radial cracking. Therefore, it might be feasible to reduce the thickness of the coping from 0.5 to 0.3 mm without adversely affecting the structural integrity of the prosthesis.26 Prosthesis fractures are sensitive to the mechanical strength and hardness of the selected restoration materials. For bilayer all-ceramic restorations, attention should be given to strengthening the core ceramic.26 The close correlation of fracture initiating site to fracture resistance is attributed to the elastic constants of the system under load. Stiffer ceramic cores, such as zirconia, provide superior protection for the underlying dentin as well as for the veneering porcelain.25 The absolute values of the fracture loads measured here might be overestimated from a clinical standpoint, due to the use of supporting brass dies in the experimental design. Brass exhibits a superior elastic modulus (Ebrass = 90 GPa) as compared to dentin (Edentin = 18 GPa). Therefore, a stiffer support of the ceramic copings and a reduced elastic mismatch in the system zirconia (E zirconia = 205 GPa), cement (Ecement

375

Reich et al

376
= 5 GPa), and supporting brass die is expected.25 Based on the investigated failure loads, reducing the zirconia coping thickness from 0.5 to 0.3 mm seems feasible and appropriate for further clinical investigation. Because in vitro fracture initiation is located on the intaglio surface of the incisal plateau, it could be that the strength of a single anterior crown is influenced less by gingival margin geometry than by the adhesive or retentive support of the substrate. These results demonstrate the fracture performance for 1 luting agent and 1 all-ceramic system, and cannot be generalized beyond the materials evaluated. Furthermore, nonveneered single crown copings were investigated. Different results would be expected from fatiguing or aging veneered structures. Within the limits of the study, the application of a KNE instead of a CHA finish line could be considered as an alternative to established preparation techniques, and might be preferred when minimally invasive preparations are indicated. However, prior to any clinical recommendation, confirmatory in vivo studies are required.

Volume 99 Issue 5
REFERENCES
1. Tyas MJ, Anusavice KJ, Frencken JE, Mount GJ. Minimal interventive dentistry a review. FDI Commission Project 1-97. Int Dent J 2000;50:1-12. 2. Shillingburg HT, Hobo S, Whitsett LD, Jacobi R, Brackett SE. Principles of tooth preparation. In: Fundamentals of fixed prosthodontics. 3rd ed. Chicago: Quintessence; 1997. p. 119-37. 3. Oilo G, Tornquist A, Durling D, Andersson M. All-ceramic crowns and preparation characteristics: a mathematic approach. Int J Prosthodont 2003;16:301-6. 4. Leevailoj C, Platt JA, Cochran MA, Moore BK. In vitro study of fracture incidence and compressive fracture load of all-ceramic crowns cemented with resin-modified glass ionomer and other luting agents. J Prosthet Dent 1998;80:699-707. 5. Bernal G, Jones RM, Brown DT, Munoz CA, Goodacre CJ. The effect of finish line form and luting agent on the breaking strength of Dicor crowns. Int J Prosthodont 1993;6:286-90. 6. Raigrodski AJ. Contemporary materials and technologies for all-ceramic fixed partial dentures: a review of the literature. J Prosthet Dent 2004;92:557-62. 7. Tinschert J, Natt G, Hassenpflug S, Spiekermann H. Status of current CAD/CAM technology in dental medicine. Int J Comput Dent 2004;7:25-45. 8. Suttor D. Lava zirconia crowns and bridges. Int J Comput Dent 2004;7:67-76. 9. Denry I, Kelly JR. State of the art of zirconia for dental applications. Dent Mater 2007 [Epub ahead of print]. 10.Sailer I, Feher A, Filser F, Luthy H, Gauckler LJ, Scharer P, et al. Prospective clinical study of zirconia posterior fixed partial dentures: 3-year follow-up. Quintessence Int 2006;37:685-93. 11.Claussen N. Microstructural design of zirconia-toughened ceramics (ZTC). In: Claussen N, Ruhle M, Heuer AH, editors. Science and technology of zirconia II, advances in ceramics. Vol 12. Columbus: The American Ceramic Society, Inc; 1984. p. 325-51. 12.Guazzato M, Albakry M, Ringer SP, Swain MV. Strength, fracture toughness and microstructure of a selection of all-ceramic materials. Part II. Zirconia-based dental ceramics. Dent Mater 2004;20:449-56. 13.Evans AG. Perspective on the development of high-toughness ceramics. J Am Ceram Soc 1990;73:187-206. 14.Potiket N, Chiche G, Finger IM. In vitro fracture strength of teeth restored with different all-ceramic crown systems. J Prosthet Dent 2004;92:491-5. 15.Bindl A, Luthy H, Mormann WH. Thinwall ceramic CAD/CAM crown copings: strength and fracture pattern. J Oral Rehabil 2006;33:520-8. 16.Rekow ED, Harsono M, Janal M, Thompson VP, Zhang G. Factorial analysis of variables influencing stress in all-ceramic crowns. Dent Mater 2006;22:125-32. 17.Ernst CP, Cohnen U, Stender E, Willershausen B. In vitro retentive strength of zirconium oxide ceramic crowns using different luting agents. J Prosthet Dent 2005;93:551-8. 18.Blatz MB, Sadan A, Kern M. Resin-ceramic bonding: a review of the literature. J Prosthet Dent 2003;89:268-74. 19.Quinn JB, Quinn GD, Kelly JR, Scherrer SS. Fractographic analyses of three ceramic whole crown restoration failures. Dent Mater 2005;21:920-9. 20.Frechette VD. Failure analysis of brittle materials. Advances in ceramics. Vol 28. Westeville: American Ceramic Society; 1990. p. 7-42. 21.Wakabayashi N, Anusavice KJ. Crack initiation modes in bilayered alumina/ porcelain disks as a function of core/veneer thickness ratio and supporting substrate stiffness. J Dent Res 2000;79:1398-404. 22.Lohbauer U, Petschelt A, Greil P. Lifetime prediction of CAD/CAM dental ceramics. J Biomed Mater Res 2002;63:780-5. 23.Proos KA, Swain MV, Ironside J, Steven GP. Influence of core thickness on a restored crown of a first premolar using finite element analysis. Int J Prosthodont 2003;16:474-80. 24.Vult von Steyern P, Carlson P, Nilner K. All-ceramic fixed partial dentures designed according to the DC-Zirkon technique. A 2-year clinical study. J Oral Rehabil 2005;32:180-7. 25.Kelly JR. Clinically relevant approach to failure testing of all-ceramic restorations. J Prosthet Dent 1999;81:652-61. 26.Lawn BR, Pajares A, Zhang Y, Deng Y, Polack MA, Lloyd IK, Rekow ED, Thompson VP. Materials design in the performance of all-ceramic crowns. Biomaterials 2004;25:2885-92. 27.Sorensen JA, Engelman MJ. Ferrule design and fracture resistance of endodontically treated teeth. J Prosthet Dent 1990;63:52936. Corresponding author: Dr Ulrich Lohbauer Dental Clinic 1 - Operative Dentistry and Periodontology Glueckstr. 11 91054 Erlangen GERMANY Fax: +49 9131 853 3603 E-mail: lohbauer@dent.uni-erlangen.de Acknowledgements The authors acknowledge 3M ESPE for the contribution of materials, and thank Mrs Gudrun Amberger and Mr Herbert Broenner for technical assistance. Copyright 2008 by the Editorial Council for The Journal of Prosthetic Dentistry.

CONCLUSIONS
Within the limitations of this study, the following conclusions were drawn: 1. KNE preparations present a promising alternative to CHA finish lines based on the 38% increase in fracture load measured for KNE finish lines relative to CHA finish lines, regardless of coping thickness. 2. Regardless of the margin preparation, a reduction in the layer thickness of a single crown coping from 0.5 to 0.3 mm resulted in a 35% reduction in fracture resistance.

The Journal of Prosthetic Dentistry

Reich et al