1

Introduction

For restoring molar teeth with single crowns, yttria-stabilized zirconia cores veneered with dental porcelains are highly esthetic alternatives to conventional metal-ceramics. Zirconia ceramics can be processed with CAD/CAM (computer aided design/computer aided manufacturing) or CAM technologies, and their suitability as high-strength substructure materials has been proven under in vitro and in vivo conditions over the past years . Furthermore, zirconia ceramics provide higher fracture toughness, a smaller range of strength variation, and higher structural reliability than glass ceramics . The number and sequence of fabrication steps (for example, wax modeling versus CAD) depend on the chosen system and material, but all zirconia substructures finally undergo particular CAM processing. After milling, these frameworks have to be veneered with feldspathic or glass ceramics by means of the layering, the press, or the digital veneering technique. Although the application of full-contour zirconia restorations is currently discussed as an alternative to commonly veneered restorations , esthetically superior results can only be achieved by applying veneering materials with mechanical properties inferior to those of the frameworks. Because the veneering glass ceramic is the weakest part in this system, clinically observed failures are mainly restricted to the veneer layer . As the veneering porcelain is directly exposed to chewing, clenching, and moisture, fatigue mechanisms and stress corrosion further weaken the veneer and finally result in cracks or chippings .

These brittle breakdowns are typical failures of ceramic materials. Although chippings may also present a problem with porcelain-fused-to-metal (PFM) restorations, particularly the chipping failures with all-ceramic zirconia restorations are discussed . Various influencing factors have been reported, such as the support and thickness of the veneer , the morphology of the circular finishing line , the adhesive forces between substructure and veneering , the mismatch of coefficients of thermal expansion (CTE) , or the firing protocol during the veneering process . Optimization of the zirconia substructure design has been proven as a considerable factor in reducing chipping failures , and coping modifications are still a topic of current investigations . The main issues of Y-TZP-based restorations seem to be the structural integrity of the veneering porcelain on the one hand and its support by the zirconia substructure on the other hand; therefore, the influence of the application technique and firing regime of the veneering material in combination with the zirconia substructure design should be further investigated to decrease failure rates.

Laboratory tests, such as the finite element analysis, may help to predict the fracture behavior of specific material combinations . But failure types and patterns are notably influenced by clinical variables, such as an individual crown design with its occlusal variations, a patient’s chewing behavior, and functioning in an oral environment. These variables may have different effects on loading, force distribution, and aging. Chewing simulations that imitate the clinical situation with dynamic loading and thermal cycling may help examine specimen behavior under clinically approximated conditions . Material breakdowns during simulation can be compared with clinically observed failures, and fractographic methods can be applied for further investigations of failed ceramic restorations . Chippings and fractures are mostly initiated by flaws inside the material or defects in the marginal areas or on the occlusal surface . Even in cases without any catastrophic failures (fracture) during oral application, aging and deterioration effects might occur, which weaken the ceramic structure, thus reducing strength and fracture resistance. In these cases a subsequent static fracture test may help locate initiated weak points.

The hypothesis tested in this study was that different substructure designs (simple core or anatomically reduced design), veneer application techniques (layering, press, or digital veneering technique), and firing regimes (normal or slow cooling) influence the number and dimension of failures in zirconia-based all-ceramic crowns during simulated oral service and affect the fracture resistance after fatigue testing.

2

Materials and methods

The tooth 46 (Morita, Dietzenbach, Germany) was prepared for a single crown according to the directives for zirconia all-ceramic restorations. A circular and occlusal anatomical reduction of 1.5–2.0 mm was carried out with a preparation angle of 4°. The finishing line resulted in a 1 mm deep circular shoulder with rounded inner angles at an isogingival height of the tooth cervix. Sharp inner edges and undercuts were eliminated. This prepared tooth was then multiplied resulting in 48 identical polymethylmethacrylate (Palapress Vario, Heraeus-Kulzer, Hanau, Germany) teeth. Their roots were covered with a 1 mm thick layer of polyether material (Impregum, 3M Espe, Seefeld, Germany) to simulate periodontal tooth mobility . For achieving a constant layer, the roots were dipped in a wax bath, which was replaced by polyether in a second fabrication process, as described in previous studies . We positioned the teeth in resin blocks (Palapress Vario, Heraeus-Kulzer) and made polyether impressions (Permadyne, 3M Espe) and working dyes of class IV dental stone (Fuji Rock, GC-Corporation, Tokyo, Japan). 48 substructures for the molar crowns were fabricated with yttria-stabilized zirconia (Lava, 3M Espe) using the CAD/CAM technique according to the manufacturer’s instructions. Six groups were defined ( n = 8/group) that finally showed the same crown shape but differed in substructure design, veneer application technique, and firing regime:

• SCLT, simple core; veneered in layering technique (reference group)

• ARLT, anatomically reduced core; veneered in layering technique

• SCLT_L, simple core; veneered in layering technique (long-term cooling)

• SCPT, simple core; veneered in press technique

• ARPT, anatomically reduced core; veneered in press technique

• ARDV, anatomically reduced core; veneered with digital veneering technique.

The substructures were divided into two groups:

• –

  simple core (SC) with an overall thickness of 0.5 mm, resulting in a varying veneer thickness

• –

  anatomically reduced (AR) core, in which the crown dimension was reduced all around by 1 mm, resulting in a substructure thickness between 0.5 and 1 mm to ensure optimal support and constant thickness of the veneer layer.

The maximal thickness of the entire restoration was 2 mm. Thus, the veneering thickness depended on the substructure design. Veneering was done using the layering technique (Lava Ceram, 3M Espe), the press technique (IPS e.max ZirPress, Ivoclar Vivadent, Schaan, Liechtenstein), or the digital veneering technique (exp. material, 3M Espe) according to the manufacturers’ instructions. Details on the materials used are provided in Table 1 . For the digital veneering technique, the CAD/CAM-manufactured veneer and the Y-TZP core were connected with fusion porcelain. Long-term cooling (SCLT_L) was performed according to standard regimes, only the cooling phase was prolonged to 6 min. All crowns were cemented using a self-adhesive resin-based cement (RelyX Unicem Aplicap, 3M Espe; 4 × 20 s, Bluephase C8, 800 mW/cm ² , Ivoclar Vivadent).

48 identical crowns were fabricated based on tooth 16 (Morita) of CoCr-alloy (Wirobond LFC, Bego, Bremen, Germany) and a veneering porcelain (VM 15, Vita, Bad Säckingen, Germany), which ought to serve as antagonists during the entire chewing simulation. The all-ceramic crowns and their antagonists were adjusted in a three point contact in a dental articulator (Artex CN, Amann Girrbach, Pforzheim, Germany) and transferred to the chewing simulator (EGO, Regensburg, Germany) using a bite registration. According to an idealized occlusion concept, the distobuccal and the two lingual cusps were loaded by the opposing mesiopalatinal cusp of the antagonistic crown via tripodized contacts in the central fossa. Further antagonistic contact areas were located at the mesiobuccal and distobuccal cusp tips.

Thermal cycling (TC: 2 × 3000 × 5°/55°; 2 min each cycle) and mechanical loading (ML: 1.2 × 10 ⁶ ; 50 N; 1.6 Hz) was performed. Parameters are based on literature data on zirconia and ceramic restorations expressing that chewing simulation using these parameters might simulate a maximum of five years of oral service . During the simulation time, all crowns and their antagonists were monitored, appearing failures of the specimens were documented (type, number of mechanical cycles) and failed crowns were excluded from the further simulation process. Location (mesial, distal, buccal or lingual direction) and extension area or length of the occurring failure mode were determined by means of a light microscope (M420, Wild, Heerbrugg, Switzerland). Scanning electron microscopy (SEM; magnification: 20–600×; working distance: 20.4 mm; voltage: 5–10 keV; low vacuum; Quanta FEG 400, FEI Company, Hillsboro, USA) was used for fractographic failure analyzing. This way, overview and detailed micrographs were produced.

Molar crowns without any failure during TCML were subsequently loaded with a testing machine (Zwick 1446, Ulm, Germany; v = 1 mm/min) until failure. The force was applied using a steel ball ( d = 12 mm) and a folded tin foil (4 × 0.25 mm, Dentaurum, Ispringen, Germany), between crown and antagonist prevented force peaks. Crowns were optically examined after fracture testing, and failure modes were divided into chipping of the veneer or combined fracture of the veneer and core.

Calculations and statistical analysis were carried out using SPSS 19.0 for Windows (SPSS Inc., Chicago, IL, USA). Mean values and standard deviations (SD) were calculated and analyzed by means of one-way analysis of variance (ANOVA) and the Bonferroni multiple comparison test for post hoc analysis. The level of significance was set to α = 0.05.

2

Materials and methods

The tooth 46 (Morita, Dietzenbach, Germany) was prepared for a single crown according to the directives for zirconia all-ceramic restorations. A circular and occlusal anatomical reduction of 1.5–2.0 mm was carried out with a preparation angle of 4°. The finishing line resulted in a 1 mm deep circular shoulder with rounded inner angles at an isogingival height of the tooth cervix. Sharp inner edges and undercuts were eliminated. This prepared tooth was then multiplied resulting in 48 identical polymethylmethacrylate (Palapress Vario, Heraeus-Kulzer, Hanau, Germany) teeth. Their roots were covered with a 1 mm thick layer of polyether material (Impregum, 3M Espe, Seefeld, Germany) to simulate periodontal tooth mobility . For achieving a constant layer, the roots were dipped in a wax bath, which was replaced by polyether in a second fabrication process, as described in previous studies . We positioned the teeth in resin blocks (Palapress Vario, Heraeus-Kulzer) and made polyether impressions (Permadyne, 3M Espe) and working dyes of class IV dental stone (Fuji Rock, GC-Corporation, Tokyo, Japan). 48 substructures for the molar crowns were fabricated with yttria-stabilized zirconia (Lava, 3M Espe) using the CAD/CAM technique according to the manufacturer’s instructions. Six groups were defined ( n = 8/group) that finally showed the same crown shape but differed in substructure design, veneer application technique, and firing regime:

• SCLT, simple core; veneered in layering technique (reference group)

• ARLT, anatomically reduced core; veneered in layering technique

• SCLT_L, simple core; veneered in layering technique (long-term cooling)

• SCPT, simple core; veneered in press technique

• ARPT, anatomically reduced core; veneered in press technique

• ARDV, anatomically reduced core; veneered with digital veneering technique.

The substructures were divided into two groups:

• –

  simple core (SC) with an overall thickness of 0.5 mm, resulting in a varying veneer thickness

• –

  anatomically reduced (AR) core, in which the crown dimension was reduced all around by 1 mm, resulting in a substructure thickness between 0.5 and 1 mm to ensure optimal support and constant thickness of the veneer layer.

The maximal thickness of the entire restoration was 2 mm. Thus, the veneering thickness depended on the substructure design. Veneering was done using the layering technique (Lava Ceram, 3M Espe), the press technique (IPS e.max ZirPress, Ivoclar Vivadent, Schaan, Liechtenstein), or the digital veneering technique (exp. material, 3M Espe) according to the manufacturers’ instructions. Details on the materials used are provided in Table 1 . For the digital veneering technique, the CAD/CAM-manufactured veneer and the Y-TZP core were connected with fusion porcelain. Long-term cooling (SCLT_L) was performed according to standard regimes, only the cooling phase was prolonged to 6 min. All crowns were cemented using a self-adhesive resin-based cement (RelyX Unicem Aplicap, 3M Espe; 4 × 20 s, Bluephase C8, 800 mW/cm ² , Ivoclar Vivadent).

48 identical crowns were fabricated based on tooth 16 (Morita) of CoCr-alloy (Wirobond LFC, Bego, Bremen, Germany) and a veneering porcelain (VM 15, Vita, Bad Säckingen, Germany), which ought to serve as antagonists during the entire chewing simulation. The all-ceramic crowns and their antagonists were adjusted in a three point contact in a dental articulator (Artex CN, Amann Girrbach, Pforzheim, Germany) and transferred to the chewing simulator (EGO, Regensburg, Germany) using a bite registration. According to an idealized occlusion concept, the distobuccal and the two lingual cusps were loaded by the opposing mesiopalatinal cusp of the antagonistic crown via tripodized contacts in the central fossa. Further antagonistic contact areas were located at the mesiobuccal and distobuccal cusp tips.

Thermal cycling (TC: 2 × 3000 × 5°/55°; 2 min each cycle) and mechanical loading (ML: 1.2 × 10 ⁶ ; 50 N; 1.6 Hz) was performed. Parameters are based on literature data on zirconia and ceramic restorations expressing that chewing simulation using these parameters might simulate a maximum of five years of oral service . During the simulation time, all crowns and their antagonists were monitored, appearing failures of the specimens were documented (type, number of mechanical cycles) and failed crowns were excluded from the further simulation process. Location (mesial, distal, buccal or lingual direction) and extension area or length of the occurring failure mode were determined by means of a light microscope (M420, Wild, Heerbrugg, Switzerland). Scanning electron microscopy (SEM; magnification: 20–600×; working distance: 20.4 mm; voltage: 5–10 keV; low vacuum; Quanta FEG 400, FEI Company, Hillsboro, USA) was used for fractographic failure analyzing. This way, overview and detailed micrographs were produced.

Molar crowns without any failure during TCML were subsequently loaded with a testing machine (Zwick 1446, Ulm, Germany; v = 1 mm/min) until failure. The force was applied using a steel ball ( d = 12 mm) and a folded tin foil (4 × 0.25 mm, Dentaurum, Ispringen, Germany), between crown and antagonist prevented force peaks. Crowns were optically examined after fracture testing, and failure modes were divided into chipping of the veneer or combined fracture of the veneer and core.

Calculations and statistical analysis were carried out using SPSS 19.0 for Windows (SPSS Inc., Chicago, IL, USA). Mean values and standard deviations (SD) were calculated and analyzed by means of one-way analysis of variance (ANOVA) and the Bonferroni multiple comparison test for post hoc analysis. The level of significance was set to α = 0.05.