Reliability, failure probability, and strength of resin-based materials for CAD/CAM restorations

ABSTRACT Objective: This study investigated the Weibull parameters and 5% fracture probability of direct, indirect composites, and CAD/CAM composites. Material and Methods: Discshaped (12 mm diameter x 1 mm thick) specimens were prepared for a direct composite [Z100 (ZO), 3M-ESPE], an indirect laboratory composite [Ceramage (CM), Shofu], and two CAD/CAM composites [Lava Ultimate (LU), 3M ESPE; Vita Enamic (VE), Vita Zahnfabrik] restorations (n=30 for each group). The specimens were polished, stored in distilled water for 24 hours at 37°C. Weibull parameters (m= modulus of Weibull, σ0= characteristic strength) and flexural strength for 5% fracture probability (σ5%) were determined using a piston-on-three-balls device at 1 MPa/s in distilled water. Statistical analysis for biaxial flexural strength analysis were performed either by both one-way ANOVA and Tukey's post hoc (α=0.05) or by Pearson's correlation test. Results: Ranking of m was: VE (19.5), LU (14.5), CM (11.7), and ZO (9.6). Ranking of σ0 (MPa) was: LU (218.1), ZO (210.4), CM (209.0), and VE (126.5). σ5% (MPa) was 177.9 for LU, 163.2 for CM, 154.7 for Z0, and 108.7 for VE. There was no significant difference in the m for ZO, CM, and LU. VE presented the highest m value and significantly higher than ZO. For σ0 and σ5%, ZO, CM, and LU were similar but higher than VE. Conclusion: The strength characteristics of CAD/ CAM composites vary according to their composition and microstructure. VE presented the lowest strength and highest Weibull modulus among the materials.


INTRODUCTION
Resin composites are widely used as direct posterior restorative materials due to good aesthetic, biological and physico-mechanical properties 16,22,24 . Resin composites used for direct restorations can present high elastic modulus and fracture strength. In addition, these materials can have a subcritical crack growth susceptibility to slow crack growth 15,16,24 . However, they have some inherent drawbacks such as polymerisation shrinkage, shrinkage stress, and limited depth of cure. Despite the advances in material technology, bulk fracture persists as one of the major causes of direct composite restoration failure in clinical trials 12,24 .
To overcome some of these limitations, indirect resin composites were introduced 14 . The polymerization protocols of these materials are optimized and often include the use of powerful lights for an extended period of time to achieve high degrees of conversion 4,14,21 . Furthermore, some can be heated to temperatures above the composite's glass transition temperature to allow an increase in polymer chain mobility, leading to additional cross-linking and stress relief 4,14,25 . Nonetheless, fabrication of indirect composites involves additional laboratory procedures that potentially increase the time required to deliver the restoration to the patient and operational cost. These disadvantages introduction of CAD/CAM composite blocks 13 . CAD/ 2016;24(5):447-52 handling and is less labour intensive, increasing 5,13 . Several resin-based CAD/CAM composites blocks are available with different compositions and network (Enamic) and nanoparticle and nanoclusterthese two materials exhibit smoother milled margins compared to ceramic-based blocks 3 , they present materials 3,11,20 .
The measure of structural performance for brittle dental materials is often estimated by their fracture strength values. However, structural performance cannot be directly predicted by strength data, since the latter is a contingent and not an inherent material property 17 . Weibull statistical analysis is largely applied for describing the scatter in fracture strength measurements of brittle materials. It is related to fracture probability and can describe the reliability of materials (e.g. stress required to fracture a given percentage of specimens). In addition, it can be scaled-up considering different specimen size (bulk volume or surface area under stress) 8,17 . The two-parameter Weibull distribution is based on two distinct parameters: Weibull modulus and characteristic strength. The Weibull modulus that describes the variation in the strength or asymmetric strength distribution as a result of 17 . Thus, low Weibull modulus values indicate higher scattering for fracture strength data and hence lower reliability. Since it is inversely related to the standard deviation in a normal distribution, high Weibull modulus relates to higher reliability of materials. A higher Weibull modulus equates to a more homogeneous which results in higher structural reliability and lower failure probability 8,16,17,19 .
The objective of this study was to evaluate the CAD/CAM resin-based composites. The hypothesis to be tested is that CAD/CAM materials present higher strength features as compared with direct and indirect composite materials.

MATERIAL AND METHODS
The materials studied are listed in Figure 1. Disk shaped specimens (12 mm diam x 1 mm thick) were prepared as follows: ZO and CM: specimens were prepared by condensing ZO and CM composites into a cylindrical metal mould. A glass slide was used to extrude excess material. ZO specimens were polymerised using a LED light unit (Monitex BlueLEX TM GT-1200, New Taipei City, Taiwan) at 600 mW/cm 2 for 40 seconds per quadrant. For CM, the uncured discs were placed in a laboratory curing unit (Solidilite V, Shofu, Kyoto, Kyoto Prefecture, Japan) containing four halogen light bulbs of 150 W each and cured for 5 minutes.
LU and VE: a digital impression (CEREC Omnicam, Sirona, Charlotte, NC, USA) was obtained from a metal cylinder (12 mm in diameter x 14 mm in height) and used as a template to mill the blocks (CEREC inLab MC XL milling unit, Sirona, NC, USA). The cylindrical blocks were sliced into 1.1 mm thick discs (IsoMet 1000 Precision Cutter, Buehler, Bergneustadt, Cologne, Germany).
All the specimens were polished down (Buehler Phoenix Beta Grinder Polisher, Bergneustadt, Cologne, Germany) using a series of sand paper disc of decreasing grits (MicroCut discs, Germany: P1200, P2500 and P4000, 100 rpm / 30 seconds specimens were measured at four points diagonally of 1 mm and 12 mm diameter. Specimens were subsequently stored for 24 hours in distilled water at 37°C prior to testing. each material was tested using a piston-on-threeballs device according to Ornaghi, et al. 17 (2012) BI ) was calculated according to Equations 1 to 3, where P is the fracture load (in N); b is the thickness of the specimen at fracture origin (in mm); r1 is the radius of the support ball circle (5 mm); r2 is the radius of loaded area (0.6 mm); and  where Pf is the fracture probability. The 0 corresponds to the strength at the failure probability of 63.2%. Based on Equation 5% was also determined, which corresponds to the strength at the more clinically relevant failure probability of 5%. The Weibull parameters were calculated based on the maximum likelihood method, according to the ASTM C 1239 2 intervals. Fractured surfaces of randomly selected specimens were examined using optical and scanning electron microscopy (Stereoscan 400, Wetzlar, Giessen, Germany).
were analyzed by one-way ANOVA and multiple comparisons were performed using Tukey's post hoc were calculated using Pearson's correlation test.

RESULTS
The s BI obtained for ZO, CM, and LU were similar, a positive correlation between s BI and the number  of fragments. For ZO and CM, 40 and 43% of the specimens fractured into four pieces and for LU and VE, 57 and 83% of the discs produced three and two fragments, respectively. Figure 2 and 3 show the fracture surface of selected specimens. Table 2 shows the Weibull parameters and the respective 95% confidence intervals. The Weibull plots are shown in Figure 4. There was no 0 5% , ZO, CM, and LU were similar, but higher than VE.
There was an average positive correlation BI and the number of fragments for all the materials tested (Table 1).

DISCUSSION
The hypothesis was rejected, since LU presented similar strength characteristics as CE and ZO. Table 1 shows that the biaxial strength of VE was Similar trends have been previously reported 3,10 . VE is composed of with a urethane dimethacrylate and triethylene glycol dimethacrylate copolymer. This combination increases the ability to endure mechanical loading by experiencing more elastic deformation before failure 3 . This can possibly be the reason why VE fractured in fewer fragments as compared with the other materials tested (Table 1).
Brittle materials, such as ceramics and composites, are very sensitive to the presence of 15,19 Group m  and are often characterized as scratches, nonpores, and other innate defects. In direct restorative materials, they can be introduced during the manufacturing process, clinical handling, fabrication 16 . The fractographic analysis showed that, typically, the fractures were ( Figure 2 and 3). Similar critical defects have been previously observed on fracture surfaces porcelains and resin composites 18,19 . Interestingly, the Weibull modulus (m) for LU was slightly higher than that obtained for ZO and CM. As the procedures involved to fabricate LU specimens were automated, it may be suggested that the handling of these materials had a minor effect in adding defects to the pre-0 ) corresponds to the stress level for a 63.2% probability of failure. This location parameter is dependent on the 0 obtained for Z0, CM, and LU were similar (Table  2) and at least 65% higher than the one obtained for VE. Nonetheless, it presented the highest m, resulting in a material with the utmost reliability and predictability for failure among the materials tested. The existence of the polymer phase within the ceramic framework spread the plasticity upon loading, hence increasing the crack resistance with crack extension. This could give a higher observable Weibull modulus, rather than toughening itself, resulting in a material with high m and moderate 0 7,9,11 . 5% ), LU presented the highest value, followed by CM and ZO respectively ( Table 2). The higher strength for LU may be related to its microstructure, which silica clusters (consisting of 20 nm silica and 4-11 nm zirconia particles); non-agglomerated/nonaggregated 20 nm silica nanoparticles; and nonagglomerated/non-aggregated 4 to 11 nm zirconia nanoparticles 1 . The wider granulometric distribution 16,24 .

CONCLUSION
The strength characteristics of CAD/CAM composites vary according to their composition and microstructure. VE presented the lowest strength and highest Weibull modulus among the materials. LU and CM presented similar strength parameters compared with the direct restorative material tested.