Two series of composites were tested in the following systems: universal testing machine (UTM) using glass rods as bonding substrate, UTM/acrylic rods, 鈥渓ow compliance device鈥? and single cantilever device (鈥淏ioman鈥?. One series had five experimental composites containing BisGMA:TEGDMA in equimolar concentrations and 60, 65, 70, 75 or 80 wt%of filler. The other series had five commercial composites: Filtek Z250 (3M ESPE), Filtek A110 (3M ESPE), Tetric Ceram (Ivoclar), Heliomolar (Ivoclar) and Point 4 (Kerr). Specimen geometry, dimensions and curing conditions were similar in all systems. 蟽pol was monitored for 10 min. Volumetric shrinkage (VS) was measured in a mercury dilatometer and elastic modulus (E) was determined by three-point bending. Shrinkage rate was used as a measure of reaction kinetics. ANOVA/Tukey test was performed for each variable, separately for each series.
For the experimental composites, 蟽pol decreased with filler content in all systems, following the variation in VS. For commercial materials, 蟽pol did not vary in the UTM/acrylic system and showed very few similarities in rankings in the others tests system. Also, no clear relationships were observed between 蟽pol and VS or E.
The testing systems showed a good agreement for the experimental composites, but very few similarities for the commercial composites. Therefore, comparison of polymerization stress results from different devices must be done carefully.