This study evaluated the effectiveness of the plasma arc curing (PAC) unit for composite curing. To compare its effectiveness with conventional quartz tungsten halogen (QTH) light curing units, the microhardness of two composites (Z100 and Tetric Ceram) that had been light cured by the PAC or QTH units, were compared according to the depth from the composite surface. In addition, linear polymerization shrinkage was compared using a custom-made linometer between composites which were light cured by PAC or QTH units. Measuring polymerization shrinkage for two resin composites (Z100 and Tetric Ceram) was performed after polymerization with either QTH or PAC units. In the case of curing with the PAC unit, the composite was light cured with Apollo 95E for two (Group 1), three (Group 2), six (Group 3) or 2 x 6 (Group 4) seconds. For light curing with the QTH unit, the composite was light cured for 60 seconds with Optilux 500 (Group 5). The linear polymerization shrinkage of composites was determined in the linometer. Two resin composites were used to measure microhardness. Two-mm thick samples were light cured for three seconds (Group 1), six seconds (Group 2) or 12 (2 x 6) seconds (Group 3) with Apollo 95E or they were conventionally light cured with Optilux 500 for 30 seconds (Group 4) or 60 seconds (Group 5). For 3 mm thick samples, the composites were light cured for six seconds (Group 1), 12 (2 x 6) seconds (Group 2) or 18 (3 x 6) seconds (Group 3) with Apollo 95E or they were conventionally light cured with Optilux 500 for 30 seconds (Group 4) or 60 seconds (Group 5). Twenty samples were assigned to each group. The microhardness of the upper and lower surfaces was measured with a Vickers hardness-measuring instrument under load. The difference in microhardness between the upper and lower surfaces in each group was analyzed by paired t-test. For the upper or lower surfaces, one-way ANOVA with Tukey was used. For Tetric Ceram, the amount of polymerization shrinkage was lower when cured with the Apollo 95E for two or three seconds than when cured for six and 12 (2 x 6) seconds, or for 60 seconds with Optilux 500 (p<0.05). For Z100, the amount of linear polymerization shrinkage was lower when cured with the Apollo 95E for two, three and six seconds than for 12 (2 x 6) seconds with Apollo 95E or for 60 seconds with the Optilux 500 (p<0.05). The results of the microhardness test indicated that there was no statistically significant difference in microhardness between groups for the upper surface. However, for the lower surface, when the composites were light cured with Apollo 95E for three seconds as recommended by the manufacturer, microhardness of the lower surface was usually lower than that of the upper surface and did not cure sufficiently. Conclusively, when compared with conventional QTH unit, the PAC unit, Apollo 95E did not properly cure the lower composite surface when the layer thickness exceeded 2 mm. In addition, three seconds of curing time, which the manufacturer recommended, was insufficient for optimal curing of composites.