The materials used in this study with their composition are listed in Table 1.
One hundred extracted, sound, and caries-free mandibular molar teeth of similar occlusal size (± 1 mm) were selected. Upon collection, adhering soft tissues were removed under running water and the teeth were stored in a 0.5% chloramine T solution at 4 °C for a period not exceeding 2 months. The size of each tooth was measured from buccolingual and mesiodistal directions with a digital caliper (Mitutoyo Corp., Tokyo, Japan). The mean dimensions were 10.3 (± 0.5) in buccolingual and 11.3 (± 0.6) in mesiodistal directions. The teeth were mounted on an acrylic block (diameter 2.5 cm) at the cement-enamel junction using auto-polymerized acrylic resin (Palapress; Heraus Kulzer, Wehrheim, Germany). Ninety teeth received a similar coronal preparation. Two operators performed all teeth preparations and restorations. Ten teeth were left intact and served as control.
Tooth preparation and restorative procedures
MOD cavities with removed palatal cusp preparations were fabricated on ninety mandibular molars. The removed palatal cusps were at the level of the isthmus floor. The preparation was made having a flat cavity floor with 5 mm of occlusal reduction (Fig. 1). The remaining buccal wall thickness was around 3 mm. The margins were placed 1–1.5 mm above the cement-enamel junction (CEJ). Preparation was achieved with flat-end parallel carbide bur (H21LR.314.010, Brasseler, Savannah, GA, USA) and round-end diamond bur (850–014 M SSWhite, Lakewood, NJ, USA) at high speed and under water cooling. According to Bijelic-Danova et al., this MOD preparation was named as flat-box type of preparation and represented a situation that was commonly seen after removal of an old complex amalgam restoration .
After completing the cavity preparation, the tooth surfaces were prepared for bonding with a one-step adhesive system (G-Premio Bond, GC Corp., Tokyo, Japan) according to the manufacturer instructions. The teeth were then restored via two approaches as direct and indirect restorations for the purpose of imitate clinical techniques.
For these groups, a translucent model (Memosil 2, Heraeus Kulzer GmbH, Hanau, Germany) of tooth crown before preparation was used to assist standardized restoration fabrication. In order to study the influence of thickness ratio of SFC-core to the thickness of the veneering PFC (occlusally), more groups were constructed having a SFC-core with various thicknesses (0, 1, 2, 3, 4 mm), remaining the thickness of the bi-structure restoration being 5 mm. The thickness of SFC-core before polymerization was controlled by the use of scaled periodontal probe, as the material was horizontally applied on the flat cavity floor.
Group 1: 0-mm SFC-core + 5-mm PFC
Group 2: 1-mm SFC-core + 4-mm PFC
Group 3: 2-mm SFC-core + 3-mm PFC
Group 4: 3-mm SFC-core + 2-mm PFC
Group 5: 4-mm SFC-core + 1-mm PFC
Direct composite restorations were manually made by buildup of PFC (G-aenial Anterior). The PFC pastes were packed into the space created between the index and the prepared cavity (with or without SFR-core), followed by curing through a hand-light curing unit (Elipar TM S10, 3M ESPE, Seefeld, Germany) from all directions and for 40 s per increment (wavelength of the light was between 430 and 480 nm and light intensity was 1600 mW/cm2). The light curing tip was in close contact (1–2 mm) with the resin composite surface. The missing axial walls for all groups (bi-structure) were built up with PFC composite (1 mm).
Group 6: 5-mm Cerasmart 270
Group 7: 2-mm SFC-core + 3-mm Cerasmart 270
Group 8: 5-mm e-max CAD
Group 9: 2-mm SFC-core + 3-mm e-max CAD
Groups (bi-structure) made of SFC as core material (2 mm) leave a space (3 mm occlusally; 1 mm proximally and lingually) for the veneering materials (manual buildup and CAD/CAM-fabricated) to be extended over the whole restoration surfaces.
For CAD/CAM-fabricated groups, a photoimpression of the prepared cavity (with or without SFR-core) was taken, and then, restoration was designed and milled (CEREC, Sirona Dental Systems Inc., Long Island City, NY) of Cerasmart 270 and e-max CAD blocks. Before cementation, the inner surface of all restorations was acid-etched by 9.6% hydrofluoric acid (Pulpdent Corporation, Watertown, MA, USA) for 60 s followed by washing and air-drying. The CAD/CAM-fabricated restorations were then cemented using a multi primer (G-Multi Primer, GC, Tokyo, Japan) and dual-cure resin cement (G-CEM linkForce, GC, Tokyo, Japan), followed by light curing from all directions using a hand-light curing unit (Elipar TM S10) for 20 s per segment. The light source was placed in close contact with the crown surface.
Prior to testing, all restorations were polished using abrasive polishing points (Jiffy Polishers, Ultradent, South Jordan, UT, USA) and stored in water for 48 h at 37 °C.
Fracture load test
A static compressive load was applied to the restored teeth with a universal testing machine (Lloyd model LRX, Lloyd Instruments Ltd., Fareham, UK) at a speed of 1 mm/min. The loading was applied vertically between the triangular ridge of the buccal and lingual cusps (Fig. 1) using a round-shaped metallic tip (Ø 5 mm). The loading event was registered until restoration fracture (final drop in the load-deflection curve). Fracture patterns of each loaded restorations were visually examined and classified to three typical behavior: catastrophic fracture of restoration and tooth structure, fracture of only restoration, and chipping or delamination of veneered restorative material from SFC-core.
The representative fractured restorations were selected and examined by scanning electron microscopy (SEM, JSM 5500, Jeol Ltd., Tokyo, Japan). Prior to observation, all the specimens were cleaned by alcohol and then coated with a gold layer using a sputter coater in vacuum evaporator (BAL-TEC SCD 050 Sputter Coater, Balzers, Liechtenstein). The analysis was started from the edge of the fractured restoration specimen, from the upper loading part to the inner surface and ending at the SFC-core.
The data were statistically analyzed with SPSS version 23 (SPSS, IBM Corp.) using analysis of variance (ANOVA) at the p < 0.05 significance level followed by a Tukey HSD post hoc test to determine the differences between the groups.