Tempering of glass produces a state of compressive stress in surface regions which can enhance the resistance to crack initiation and growth. The objective of this study was to determine the influence of tempering on the sizes of surface cracks induced within the tempered surfaces of opaque porcelain-body porcelain discs, with contraction coefficient differences (alpha O-alpha B) of +3.2, +0.7, 0.0, -0.9, and -1.5 ppm/degrees C. We fired the discs to the maturing temperature (982 degrees C) of body porcelain and then subjected them to three cooling procedures: slow cooling in a furnace (SC), fast cooling in air (FC), and tempering (T) by blasting the body porcelain surface with compressed air for 90 s. We used body porcelain discs as the thermally compatible (delta alpha = 0) control specimens. We measured the diameters of cracks induced by a microhardness indenter at an applied load of 4.9 N at 80 points along diametral lines within the surface of body porcelain. The mean values of the crack diameters varied from 75.9 microns (delta alpha = -1.5 ppm/degrees C) to 103.3 microns (delta alpha = +3.2 ppm/degrees C). The results of ANOVA indicate that significant differences in crack dimensions were controlled by cooling rate, contraction mismatch, and their combined effect (p less than 0.0001). Multiple contrast analysis (Tukey's HSD Test) revealed significantly lower (p less than 0.05) crack sizes for tempered specimens compared with those of fast-cooled and slow-cooled specimens.(ABSTRACT TRUNCATED AT 250 WORDS)

Delayed failure of metal-ceramic restorations due to static fatigue can occur when residual tensile stress is present in porcelain, even in the absence of intra-oral forces. Fixed-partial-denture (FPD) specimens and semicircular arch specimens with gapped cross-arch segments were employed to characterize the potential of two incompatible metal-ceramic systems for producing delayed crack development and to determine the relative sensitivity of these test designs as monitors of incompatibility stresses which resulted from thermal contraction differences between a nickel-chromium alloy and three experimental porcelains. The arch specimens were judged to be more suitable for analysis of residual stresses because of the larger magnitude of gap changes at each procedural change. However, the FPD specimens exhibited earlier evidence of delayed crack growth in porcelain when the thermal contraction coefficient of the metal exceeded that of the porcelain by either 1.7 × 10 -6 /°C or 2.2 × 10 -6 /°C. For these two states of incompatibility, the agreement between experimental gap values for the arch specimens and the gap values predicted from composite strip equations was excellent.