R-curve behaviour is often used for evaluating crack growth resistance in quasibrittle materials but few studies have focused on polycrystalline graphite. In this study, R-curve behaviour in three commercial grade nuclear graphites, of
varying structure and properties, is compared using an optical method, a theoretical compliance method, and, a potential drop (PD) technique to measure crack length. Two graphites are coarse-grained and the third a fine grained graphite. Both 3-point bend and compact tension specimens are used.
The fine-grain graphite shows lowest resistance, with the coarse-grained materials displaying similar R-curves. The compliance method is simplest but assumes the material is linear elastic, producing similar R-curves to the optical
method. The PID method seriously underestimates the crack length due to crack face bridging, causing the R-curves to show a falling behaviour. The shortfall in the PID measurements presents a novel way of physically measuring
the bridging zone length. The graphites display similar variation in the apparent bridging zone length despite the difference in grain size.
Higher resistance in the coarser material results from increased crack path deflection, coupled with stronger grain bridging traction within the bridging zone.
The bridging zone length is longer in compact tension specimens than in 3-point bend specimens, explaining partly why higher fracture energy is recorded in compact tension. In oxidised graphite, the crack growth resistance reduces but the coarser materials still show useful resistance. The fine-grain graphite shows a tendency toward flat R-curve behaviour with increasing oxidation. The mechanisms responsible for toughening in non-oxidised graphite prevail in oxidised material but diminish with increasing porosity and loss of binder phase.
A preliminary study of the nano-indentation behaviour of the nuclear graphites reveals a similar response in all the materials. Anelastic hysteresis in the loadunload
response is found with little residual deformation. There is also evidence of a creep effect during the dwell period at the maximum load. Comparison of the average indentation modulus with the bulk flexural and sonic modulus indicates that the nano-indentation technique is unable to sense bulk
modulus changes such as occur with oxidation or the forming process. Instead, the method is susceptible to localised structural inhomogeneities.