To the knowledge of the authors, no data for monolithic zirconia are existent comparing the different flexural strength test methods. For the discussion of our data we will also consider studies with other ceramic materials.
The tested null hypothesis was that the flexural strength test methods would show no impact on strength values. This has to be rejected. It could be clearly discriminated between biaxial and 4-point tested groups, which presented the lowest flexural strength findings. In 4-point flexural strength testing, a larger area of the material is involved in stress application. Thus, the probability of crack initiation, reflected by lower flexural strength, is higher than for the 3-point test method [10
]. In our data, 4-point tested groups were in all but one group statistically different from 3-point tested groups. This is in accordance with the data of a previous study, where veneering ceramics for zirconia have been tested [26
In another study with zirconia core materials, biaxial strength testing resulted in higher and statistically different values compared to those, measured by 3-point flexural strength testing [27
]. Except for one group, our data showed the highest mean biaxial flexural strength values, but we could only observe a clear discrimination between 3-point and biaxial testing in two groups. The high values in the biaxial test can be related to the negligible effect of undesirable edge failures [28
]. Jin et al.
] used all of the three test methods on different ceramic types. Although no zirconia material was included in this study, the authors also identified the lowest values for 4-point flexural strength measurement and could not clearly differentiate between biaxial and 3-point testing. A ranking order of the different flexural strength test methods can be described as the following: biaxial > 3-point > 4-point flexural strength measurement. Regarding the influence of the test method on the Weibull modulus it was observed that only wet polished, 3-point tested zirconia C and wet polished, 4-point tested zirconia Z had a higher Weibull modulus compared to other groups. Neither surface roughness nor SEM pictures can explain these findings.
The second tested hypothesis was, that the specimen preparation methods do not influence the flexural strength and the reliability of monolithic zirconia. The inherent reliability of the material can be described with the Weibull modulus. A high Weibull modulus is associated with a statistically higher reliability. It was found that specimen preparation after sintering showed significantly higher flexural strength values whereas no general trend for the Weibull moduli was observed. The Weibull modulus for 3-point tested Zirconia C polished after sintering was even higher, than for those polished before sintering. Thus the first part of the second hypothesis could not be confirmed. Regarding the reliability, the other part of the hypothesis is accepted within flexural strength test-groups and zirconia material. In contrast to that, a statistically lower Weibull modulus was noticed in wet polished 4-point tested groups for Zirconia C and Z. This indicates that polishing may induce flaws, but the probability of critical flaw detection increases, when a larger volume of the material is involved in the testing method. As failure is initiated by the largest flaw or element, this effect is also referred to as the weakest link hypothesis [31
]. The authors believe that surface scratches shown in the SEM pictures, functioned as crack origin and led to significantly lower strength values of the 4-point tested specimen. The values for Weibull moduli ranged between (5.1 and 16.5), which is in accordance with the values reported for Y-TZP core materials tested in other studies [25
]. In the present literature, contradictory findings are reported for flexural strength of zirconia materials, dependent on different surface alteration methods after the sintering process. Several authors described a decrease in flexural strength and Weibull modulus of zirconia, when the surface underwent a grinding process [14
]. On the other hand, an increase in mean flexural strength and a decreased reliability was observed with the use of fine-grained diamond burs [35
] and with wet hand grinding [36
] although in this study the Weibull distribution was not discussed. When corund-blasting as a surface method is conducted, an increase of strength is also reported associated with a reduction of the Weibull modulus [18
]. The increase in flexural strength can be explained with the phase transformation in zirconia materials. The amount of the monoclinic phase was found to rise, when the surface was treated after sintering [18
]. This well-known transformation toughening mechanism of zirconia materials can lead to compressive surface stresses and thus elevate the flexural strength [38
]. In contrast to that, in the present study no higher fraction of monoclinic zirconia on the surface of wet polished specimen was found. Thus the polishing procedure must have removed the layer of monoclinic zirconia. This effect was also observed by Guazatto et al.
]. The above discussion of possible origins of the flaws is supported by Anderson-Darling (AD) goodness of fit estimates in Table 4
. Smaller AD values indicate a better fit of the data by the assumed distribution. The authors were unable to find a single distributional assumption of flexural strength values fitting optimally for all test configurations. However, both Weibull and normal assumptions provided reasonable and useful approximations to compare all test groups. This conclusion is supported by a strong negative association between Weibull moduli (m) and COV (rho = −0.972, p
< 0.001). Weibull modulus estimates ranged from 5.1 to 16.5 whereas those of COV from 0.077 to 0.223 indicating that both Weibull and normal assumptions were applicable to all test groups.
The third hypothesis was that different zirconia materials do not influence flexural strength. This has to be accepted, as a general trend as the zirconia materials could not be identified. It was observed that the zirconia materials used in this study act differently dependent on flexural strength test and treatment methods. Literature shows [19
] that D contains a significantly larger grain size compared to zirconia C and Z. However, no association between grain size and all tested parameters was found in our study.