PSI - Issue 78

Riccardo Liberotti et al. / Procedia Structural Integrity 78 (2026) 1919–1926

1923

to stiffness, similar conclusions can be drawn, as the previously described specimens fall into the same three categories, except for the type H specimen, which in this case exhibits a high stiffness, comparable to that of specimens B, E, F, I, and L. So if sample H is compared to E and F, it is observed that the presence of sand reduces its strength but the presence of pozzolanic cement, which is also in E and F, confers a high stiffness. The post peak behaviour of the samples is observed thanks to the test which is in control of displacement. Anyway, all the samples ranging from A to I exhibit a fragile behaviour, which can be inferred from the steep slope of the post-peak descending branch. The only exception is represented by the type L specimens, which are identical to those of type I, namely alabastrine gypsum, but with the addition of longitudinally oriented sisal fibres. The presence of these fibres allows the specimen, upon brittle failure, to retain post-peak load-bearing capacity and undergo large displacements, thus providing the material with significant ductility. This is observed by comparing type I and L plots. Due to the handmade nature of the specimens, a high degree of variability was observed in some cases. For instance, in mix F, the stiffness varied significantly between the two samples. In the case of types D and G, the variability extended beyond stiffness, affecting both the peak load and the overall shape of the force–displacement curve. This suggests that these two compositions may also present issues related to the manufacturing process. For this reason, future work will involve extending the experimental campaign to include more specimens for each type, in order to ensure a more rigorous statistical analysis. In addition, further specimens for some of the aforemen tioned mixtures were subjected to mechanical testing after consolidation with an innovative agent based on TEOS functionalised with graphene fibres. In particular, the tetraethoxysilane (TEOS) is a well-established alkoxysilane-based consolidant widely employed in the conservation of silicate-based materials. In the present paper this treatment is designed, thanks to the use of graphene particles, to enhance the mechanical perfor mance and penetration capacity of the commercial interventions. In this second phase of the experimental campaign, still ongoing, the three kinds of mix-design were selected taking into account also the aforemen tioned historical-iconographical studies: type A-gypsum and sand, type B-gypsum and black pozzolana and type E-lime, pozzolanic cement and pozzolana. The results of the additional three-point bending tests are reported in Fig. 3 comparing as-built and consolidated prototypes.

Fig. 1. On the left, the red pozzolana is finely sifted both in the form of powder and of fine sand before the raw earth put in place wet and ambient cured formworks. On the right, end of a test on a fine-grained (alabastrine) gypsum sample is reported.