1

Introduction

Currently, most all-ceramic restorations have a high crystalline content infrastructure that provides resistance and a veneering ceramic that offers improved esthetics . However, clinical studies have reported that ceramic restorations experience chipping, cracking, or delamination of the porcelain . Failure of the veneering porcelain is a complex mostly associated to thermal and mechanical properties of the ceramics and geometry and dimensions of the restoration .

In addition to porcelain chipping, multilayer ceramic structures may present a poor bonding between layers with potential for delamination . The material under tension, the thermal and mechanical interactions between ceramic layers, the restoration design, and the bonding to a substrate may influence the fracture resistance and the failure mode of multilayer structures . In addition, compression forces can generate tensile stresses in the intaglio surface of the restoration , which corresponds to the fracture origin of clinically failed crowns .

The development of high crystalline content dental ceramics, particularly the yttria partially-stabilized tetragonal zirconia polycrystal (Y-TZP) because of its high fracture toughness, and the introduction of the CAD/CAM (computer aided design/computer aided manufacturing) technology increased the use of metal-free restorations in dentistry. CAD/CAM systems are mainly based on three steps: scanning and digitalization of tooth preparation, computer processing of structural design, and milling of prefabricated blocks. This technique may minimize internal flaws and imperfections resulting from other fabrication methods, improving reliability and offering greater quality control .

The CAD-on technique (Ivoclar) was developed as an attempt to overcome the susceptibility of porcelain chipping from veneered zirconia restorations. This ceramic system produces ceramic restorations with a Y-TZP framework veneered by a lithium disilicate-based glass-ceramic, which are fabricated by CAD/CAM and fused together with a glass. As a recent method, the literature is still scarce and little is known about the multilayer structure obtained from this technique. Thus, this study aimed to evaluate: (1) the reliability of monolithic and multilayer ceramic structures bonded to a dentin analog e subjected to compressive load and (2) the mode of failure produced by loading such ceramic structures, testing the hypotheses that there is no significant difference in the Weibull modulus ( m ) among the evaluated ceramic structures and that radial crack is the predominant failure origin for the ceramic structures.

2

Materials and methods

Four ceramic structures were fabricated (n = 30) as follows:

• –

  CAD-on: Trilayer structure (1.8 mm thick) made of a Y-TZP (YZC – IPS e.max ZirCAD, Ivoclar Vivadent, Schaan, Liechtenstein) framework (1 mm thick) and a 0.7 mm thick lithium disilicate-based glass-ceramic veneer (LDC- IPS e.max CAD, Ivoclar Vivadent, Schaan, Liechtenstein) fused together with a glass (IPS e.max CAD Crystall./Connect, Ivoclar Vivadent, Schaan, Liechtenstein) of about 0.1 mm thick. The YZC and LDC blocks were ground (Ferdimat CA51H, São José dos Campos, Brazil) using a diamond stone (Tyrolit TN 634709, Cabreuva, Brazil) into a cylindrical shape, which were cut in slices (discs) by a metallographic cutter (Strues Minitron, Copenhagen, Denmark) using a diamond disc under water cooling. The disc-shaped specimens were polished (Strues Abramin, Copenhagen, Denmark) up to 1200-graded SiC sandpapers under water cooling. YZC discs were sintered (Zirkonofen 600/V2, ZirkonZahn, Gais, South Tyrol, Italy) according to manufacturer’s instructions (Ivoclar Vivadent). The YZC and LDC discs were united using the fusion ceramic (IPS e.max CAD Crystall./Connect). The capsule containing the powder and liquid of the fusion ceramic was mixed by vibration (Ivomix, Ivoclar Vivadent, Schaan, Liechtenstein) for 10 s. The capsule was opened and the material was applied to the LDC disc surface and, immediately, bonded to the YZC disc. The trilayer structure was placed under a load of 750 g and the excess of fusion ceramic was removed before sintering. The fusion process and crystallization of LDC were conducted simultaneously (Programat EP5000, Ivoclar Vivadent, Schaan, Liechtenstein) following manufacturer’s instructions.

• –

  YLD: Bilayer structure (1.8 mm thick) fabricated with a YZC framework (1 mm thick) veneered with a glass-ceramic (LDT- IPS e.max Ceram, Ivoclar Vivadent, Schaan, Liechtenstein). The YZC was fabricated as described for the CAD-on group. A thin layer (about 0.1 mm) of IPS e.max ZirLiner (Ivoclar Vivadent, Schaan, Liechtenstein) was applied on the YZC and fired (Programat EP5000) according to manufacturer’s instructions. The YZC discs were placed into a silicone matrix (Zetaplus, Zhermack SpA, Badia Polesine, Italy) and a LDT layer (0.7 mm thick) was applied over it using the conventional layering technique. The excess moist from LDT was removed using sonic vibration and absorbent paper, and the ceramic was sintered according to manufacturer’s instructions.

• –

  YZW: Monolithic structure (1.8 mm thick) of Y-TZP (Zenostar Zr Translucent, Wieland Dental, Rosbach vor der Höhe, Germany). A cylindrical pattern was scanned (Cerec InLab, Sirona Dental Company, Bensheim, Germany) and data was transferred to the milling machine (InLab MC X5, Sirona Dental Company, Bensheim, Germany) that milled the large YZW disc-shaped blocks to obtain several YZW cylinders. They were cut in slices with a diamond disc under water cooling using a metallographic cutting machine (Strues Minitron, Copenhagen, Denmark). The disc-shaped specimens were sintered (Zirkonofen 600/V2) according to manufacturer’s instructions (Ivoclar Vivadent).

• –

  LDC: Monolithic structure (1.8 mm thick) of lithium disilicate-based glass-ceramic (IPS e.max CAD) fabricated as described for the CAD-on group. LDC structures were crystallized (Programat EP5000) according to manufacturer’s instructions.

All ceramic structures were bonded to a fiber-reinforced epoxy resin-based (G10- NEMA G10, International Paper, Hampton, SC, USA) structure (dentin analog), which presents an elastic modulus (E = 18.6 GPa) similar to dentin (E = 18 GPa). Resin bond to G10 is similar to dental structure . G10 is provided in rod shape (25 mm diameter), which was cut (Strues Minitron, Copenhagen, Denmark) in slices (4 mm) using a diamond disc (BuehlerTM IsoMetTM Diamond Wafering Blades, USA) under water cooling.

The bonding surface of G10 was sandblasted with 110 μm aluminum oxide particles (Renfert do Brasil, Ribeirão Preto, Brazil) for 20 s, with a pressure of 2 bar from a 10 mm distance. Surface was sonically cleaned in distilled water for 5 min, then water-sprayed and air-dried. The adhesive system (Primers A and B, Ivoclar Vivadent, Schaan, Liechtenstein) was applied according to manufacturer’s instructions on the treated surface of G10. The ceramic bonding surface was treated with silane (Monobond S, Ivoclar Vivadent, Schaan, Liechtenstein) for 3 min and air-dried for 10 s. The dual resin cement with phosphate monomer (MDP) (Multilink N, Ivoclar Vivadent, Schaan, Liechtenstein) was mixed and applied on the treated bonding surfaces of G10 and ceramic that were united and placed under a 750 g load for 1 min. Luting cement excess was removed with a brush (cavibrush, FGM, Joinvile, SC, Brazil) and cement was light-activated for 120 s with a light curing unit (Bluephase N, Ivoclar Vivadent, Schaan, Liechtenstein; 1200 mW/cm ² power).

Specimens were tested in 37 °C distilled water under compression load applied by a round (6 mm diameter) metal piston in a universal testing machine (EMIC DL 2000, São José dos Pinhais, PR, Brazil) with cross-head speed of 0.5 mm/min ( Fig. 1 ) until the sound of first crack, which was acoustically detected using an amplified microphone connected to the testing machine. Sound waves were recorded by a specific software (Audacity Sound Editor, Free Software Foundation, Boston, USA). Detection of the first evident sound corresponded to the popped critical crack and the drop in the loading curve to where the respective load (L _f ) was recorded.

Compressive load application in 37 °C distilled water.

Cracked samples were examined using light microscopy and transillumination with blue light, which assisted on identification of initial crack and other fractographic features. Failures were classified according to the type of crack, as follows: radial crack, cone crack, and cone + radial crack. The failure process was recorded as follows: cracking (no fracture) or fracture (chipping or catastrophically).

As the distribution of size and location of failures is related to the variability of experimental data, Weibull distribution was used to statistically analyze the data and estimate the Weibull modulus ( m ) and the characteristic fracture load (L ₀ ) with 95% confidence intervals (95% CI). As the data (L _f values) did not have a normal distribution it was statistically analyzed using Kruskal–Wallis and Student-Newman–Keuls tests ( α = 0.05).

2

Materials and methods

Four ceramic structures were fabricated (n = 30) as follows:

• –

  CAD-on: Trilayer structure (1.8 mm thick) made of a Y-TZP (YZC – IPS e.max ZirCAD, Ivoclar Vivadent, Schaan, Liechtenstein) framework (1 mm thick) and a 0.7 mm thick lithium disilicate-based glass-ceramic veneer (LDC- IPS e.max CAD, Ivoclar Vivadent, Schaan, Liechtenstein) fused together with a glass (IPS e.max CAD Crystall./Connect, Ivoclar Vivadent, Schaan, Liechtenstein) of about 0.1 mm thick. The YZC and LDC blocks were ground (Ferdimat CA51H, São José dos Campos, Brazil) using a diamond stone (Tyrolit TN 634709, Cabreuva, Brazil) into a cylindrical shape, which were cut in slices (discs) by a metallographic cutter (Strues Minitron, Copenhagen, Denmark) using a diamond disc under water cooling. The disc-shaped specimens were polished (Strues Abramin, Copenhagen, Denmark) up to 1200-graded SiC sandpapers under water cooling. YZC discs were sintered (Zirkonofen 600/V2, ZirkonZahn, Gais, South Tyrol, Italy) according to manufacturer’s instructions (Ivoclar Vivadent). The YZC and LDC discs were united using the fusion ceramic (IPS e.max CAD Crystall./Connect). The capsule containing the powder and liquid of the fusion ceramic was mixed by vibration (Ivomix, Ivoclar Vivadent, Schaan, Liechtenstein) for 10 s. The capsule was opened and the material was applied to the LDC disc surface and, immediately, bonded to the YZC disc. The trilayer structure was placed under a load of 750 g and the excess of fusion ceramic was removed before sintering. The fusion process and crystallization of LDC were conducted simultaneously (Programat EP5000, Ivoclar Vivadent, Schaan, Liechtenstein) following manufacturer’s instructions.

• –

  YLD: Bilayer structure (1.8 mm thick) fabricated with a YZC framework (1 mm thick) veneered with a glass-ceramic (LDT- IPS e.max Ceram, Ivoclar Vivadent, Schaan, Liechtenstein). The YZC was fabricated as described for the CAD-on group. A thin layer (about 0.1 mm) of IPS e.max ZirLiner (Ivoclar Vivadent, Schaan, Liechtenstein) was applied on the YZC and fired (Programat EP5000) according to manufacturer’s instructions. The YZC discs were placed into a silicone matrix (Zetaplus, Zhermack SpA, Badia Polesine, Italy) and a LDT layer (0.7 mm thick) was applied over it using the conventional layering technique. The excess moist from LDT was removed using sonic vibration and absorbent paper, and the ceramic was sintered according to manufacturer’s instructions.

• –

  YZW: Monolithic structure (1.8 mm thick) of Y-TZP (Zenostar Zr Translucent, Wieland Dental, Rosbach vor der Höhe, Germany). A cylindrical pattern was scanned (Cerec InLab, Sirona Dental Company, Bensheim, Germany) and data was transferred to the milling machine (InLab MC X5, Sirona Dental Company, Bensheim, Germany) that milled the large YZW disc-shaped blocks to obtain several YZW cylinders. They were cut in slices with a diamond disc under water cooling using a metallographic cutting machine (Strues Minitron, Copenhagen, Denmark). The disc-shaped specimens were sintered (Zirkonofen 600/V2) according to manufacturer’s instructions (Ivoclar Vivadent).

• –

  LDC: Monolithic structure (1.8 mm thick) of lithium disilicate-based glass-ceramic (IPS e.max CAD) fabricated as described for the CAD-on group. LDC structures were crystallized (Programat EP5000) according to manufacturer’s instructions.

All ceramic structures were bonded to a fiber-reinforced epoxy resin-based (G10- NEMA G10, International Paper, Hampton, SC, USA) structure (dentin analog), which presents an elastic modulus (E = 18.6 GPa) similar to dentin (E = 18 GPa). Resin bond to G10 is similar to dental structure . G10 is provided in rod shape (25 mm diameter), which was cut (Strues Minitron, Copenhagen, Denmark) in slices (4 mm) using a diamond disc (BuehlerTM IsoMetTM Diamond Wafering Blades, USA) under water cooling.

The bonding surface of G10 was sandblasted with 110 μm aluminum oxide particles (Renfert do Brasil, Ribeirão Preto, Brazil) for 20 s, with a pressure of 2 bar from a 10 mm distance. Surface was sonically cleaned in distilled water for 5 min, then water-sprayed and air-dried. The adhesive system (Primers A and B, Ivoclar Vivadent, Schaan, Liechtenstein) was applied according to manufacturer’s instructions on the treated surface of G10. The ceramic bonding surface was treated with silane (Monobond S, Ivoclar Vivadent, Schaan, Liechtenstein) for 3 min and air-dried for 10 s. The dual resin cement with phosphate monomer (MDP) (Multilink N, Ivoclar Vivadent, Schaan, Liechtenstein) was mixed and applied on the treated bonding surfaces of G10 and ceramic that were united and placed under a 750 g load for 1 min. Luting cement excess was removed with a brush (cavibrush, FGM, Joinvile, SC, Brazil) and cement was light-activated for 120 s with a light curing unit (Bluephase N, Ivoclar Vivadent, Schaan, Liechtenstein; 1200 mW/cm ² power).

Specimens were tested in 37 °C distilled water under compression load applied by a round (6 mm diameter) metal piston in a universal testing machine (EMIC DL 2000, São José dos Pinhais, PR, Brazil) with cross-head speed of 0.5 mm/min ( Fig. 1 ) until the sound of first crack, which was acoustically detected using an amplified microphone connected to the testing machine. Sound waves were recorded by a specific software (Audacity Sound Editor, Free Software Foundation, Boston, USA). Detection of the first evident sound corresponded to the popped critical crack and the drop in the loading curve to where the respective load (L _f ) was recorded.

Compressive load application in 37 °C distilled water.

Cracked samples were examined using light microscopy and transillumination with blue light, which assisted on identification of initial crack and other fractographic features. Failures were classified according to the type of crack, as follows: radial crack, cone crack, and cone + radial crack. The failure process was recorded as follows: cracking (no fracture) or fracture (chipping or catastrophically).

As the distribution of size and location of failures is related to the variability of experimental data, Weibull distribution was used to statistically analyze the data and estimate the Weibull modulus ( m ) and the characteristic fracture load (L ₀ ) with 95% confidence intervals (95% CI). As the data (L _f values) did not have a normal distribution it was statistically analyzed using Kruskal–Wallis and Student-Newman–Keuls tests ( α = 0.05).