Modelo de penetração de íons sulfato em concreto: abordagem baseada em ensaios de resistividade e migração iônica

Abstract

Sulfate attack is a complex degradation phenomenon affecting cementitious matrices, and there is no consensus in the literature regarding models for evaluating this type of degradation. This study proposes a penetration model for application in external sulfate attack studies, based on resistivity and ionic migration data. The research methodology was divided into three stages. In the first stage, the migration test was adjusted for one concrete mix, varying the electrodes and sample thickness in sodium and magnesium sulfate solutions. Based on the results, graphite electrodes and 25 mm thick samples were selected. In the second stage, the input parameters for the model were determined, including a damage factor to account for the mechanical response of the interaction between sulfate ions and the cementitious matrix. This stage was conducted for six concrete mixes exposed to sodium and magnesium sulfates, both in ionic migration and diffusion tests (under immersion), as well as for two reference mixes. The diffusion samples were monitored monthly for resistivity and ultrasonic pulse velocity, and tested for compressive strength, water absorption by immersion, sulfate profiles, and damage analysis using SDT. For the migration test samples, the electrical conductivity of the solutions was monitored, followed by sulfate profile determinations. The model equation was proposed to estimate sulfate penetration for other concrete mixes, based on the water/cement ratio and the concrete’s electrical resistivity after the curing period. In the third stage, the proposed model was validated for two additional concrete mixes exposed to both sulfate types, and for literature data. These samples were subjected to diffusion in sulfate solutions, with monthly monitoring of resistivity and ultrasonic pulse velocity, as well as sulfate profile measurements at different ages. The empirical sulfate penetration depths were compared with the values predicted by the model, showing errors ranging from approximately 2% to 80%. The maximum absolute difference was 8 mm for concrete mixes tested in this stage and 14 mm for the literature data. The study highlights the limitations and potential applications of the proposed model.