The error of tensile strength tests and an approach for improvement.

Abstract

The deviation and spread of values measured by tensile strength tests should be significantly reduced by optimized axially pull-off-device. Factors like geometry, roughness, surface texture and fit of the specimens, the luting procedure (mixing ratio, compacting pressure) were standardized and the precision of the axially load transmission during the tensile force application optimized. Copings of CoCr-alloy were luted using a provisional zinc-oxide eugenol and a glass-ionomer cement (n = 20 per group each) on 4°, respectively 8° titanium implant abutment analogs. 24 h after luting the copings were pulled off with the improved (IPD) and the previous non improved device (NIPD). Means, std. dev. and t-tests were calculated. The stress- strain curves produced by the IPD showed a clear and linear relationship between stress and strain. The curves of the NIPD were ragged and chaotic. For both luting agents the means of the retentive force were considerable higher (p > 0.000) with IPD. However, the standard deviation was unimproved for the zinc oxide eugenol cement: IPD: 27%; NIPD: 27%, but significantly improved with glass ionomer: IPD: 16%, NIPD: 28%. The results from luting agents, which react like a solid body, may benefit from the optimized pull-off-device. Despite the fact that many factors in this investigation have been controlled, unknown flow characteristics of non-Newtonian-fluids like luting agents may have a considerable influence on the standard deviation of pull-off tests.