Non-destructive Analysis of the Displacement and Strain in Composite Restorations

Method: A cylindrical cavity (diameter: 2.3 mm, depth: 2 mm) was prepared into 6 cylindrical radiolucent composite blocks (Gradia Direct Anterior, GC). Half of the cavities received a thin sprayed layer of silicon oil to prevent bonding (‘non-bonded’), while the other cavities were treated with Clearfil SE Bond (Kuraray) (‘bonded’). All cavities were restored using a flowable composite (G-ænial Flo, GC), and scanned (Phoenix Nanotom s, GE) before and after polymerization (Bluephase i20, Ivoclar-Vivadent) at a voxel size of 2.5 µm.

Reconstructed images (NRecon, Skyscan) of the uncured and cured samples were imported in commercially available analytical software, namely CTAn (SkyScan), and in a custom-made software program based on Elastix (http://elastix.isi.uu.nl/). To compare the uncured and cured images, non-rigid image registration was performed using a B-spline deformation mesh with the grid spacing set at 50 µm. The displacement field was visualized by means of vectors. The mean displacement and displacement relative to both horizontal axes and the vertical axis were calculated. The resulting strain could be visualized by means of a color-coded tomographic image set.

Result: Mean displacement was 25.28 ± 4.75 µm and 31.19 ± 1.58 µm for the ‘non-bonded’ and ‘bonded’ cavities, respectively. Both cavities showed a mainly downward displacement direction. At the cavity bottom, the vertical displacement component tended to decrease, while the horizontal component increased. The strain at the border was the highest at the bottom of both the cavities.

Conclusion: Internal displacement, and thus polymerization shrinkage, as well as strain could be visualized using the µCT analysis protocol. The highest negative strain (contraction) was found at the bottom of the cavity for both the ‘non-bonded’ and ‘bonded’ cavities.