If the samples break, this means that the resin is not crosslinked, or not enough. Was the corresponding formulation respected, or did you forget to add the crosslinker in sufficient amount ? Did you cure it at sufficient temperature or during a sufficient time ? The mould should not have influence as far as it doesn't stick to the material, but the photos you provide show that the result is not that bad.

Alain

I assume that the recipe is correct and that your samples (for fracture toughness tests) have a relatively high thickness. If this is true, please, take into account that being the cross-linking reaction extremely exothermic the transient heat transfer phenomena can cause very high frozen-in residual stresses.

To avoid this shortcoming thinner samples should be used. Tacking into account that KIc for thermosetting resins is of order of 0.1-1 MPam^0.5 and the Sr (sigma to rupture) =60-70 MPa it results that you may use samples with thickness of order of 1 mm. I'd take another precaution by degassing the resin before cure cycle to avoid bubbles in the sample.

The residual stress are unavoidable. The origin is twofold.

1-The kinetics of crosslinking reactions are thickness-dependent, due to the transient heat transfer phenomena that develop as consequence of the heat of reaction.

2-Upon cooling from the cure temperature above Tg to below it each point of the sample suffer different "thermal histories".

Please see the following paper to appreciate the importance of the phenomena in play. I can anticipate that the residual stresses may be of order of magnitude of the stress to rupture of cross-linked resins, namely 50 to 70 MPa.

References

- Grassia L., D’Amore A. (2013) , “Calculation of the shrinkage-induced residual stress in a viscoelastic dental restorative material” Mechanics of Time-Dependent Materials Volume 17, Issue 1, February 2013, Pages 1-13

- L. GRASSIA, D'AMORE A (2005). Residual stresses in Amorphous Polymers. MACROMOLECULAR SYMPOSIA, ISSN: 1022-1360

Thank you so much for your answers !!! @D'Amore, I will refer to the paper you suggested.  @Celzard, I have added a post-curing step to my process (in addition to a 24hr room temperature cure) to ensure that my samples are fully crosslinked.