Mesoscopic and macroscopic investigation of a dolomitic marble subjected to thermal damage

Abstract

Thermal loading is an important factor that could lead to the weakening and deterioration of rock materials. Understanding the thermal properties of rocks and their evolution under different high temperatures is important in the post-fire-hazard evaluation and cultural heritage conservation. Yet it is challenging to understand the evolution of thermally-induced changes in rock properties and to quantitatively study degrees of thermal damage when samples are limited. This study investigates the effects of high temperatures (i.e., 200 °C, 400 °C, 600 °C, 800 °C, and 1000 °C) on a dolomitic marble using combined mesoscopic and macroscopic testing techniques. The test results show that increasing marble temperature led to a deterioration of physical properties (i.e., increasing open porosity and weight loss; but decreasing P-wave velocity) and mechanical properties (i.e., increasing axial strain corresponding with the peak stress; but decreasing uniaxial compressive strength, Young’s modulus, and brittleness). There existed a threshold temperature of 600 °C, which marks different thermal damage mechanisms. Below the threshold, the rock deterioration was mainly caused by physical changes such as crack propagation and grain breakage, which can be characterized by mesoscopic parameters (i.e., linear crack density and mineral grain size distribution). On the contrary, when the temperature was higher than the threshold, the deterioration was caused by chemical changes, including mineral decomposition and re-crystallization, which was indicated by the changes in mineral compositions and relative atomic mass calculation. Based on the experimental results (e.g., mineralogical and physico-mechanical changes) and obtained relationships between the parameters in mesoscale and macroscale, a novel scheme for thermal damage evaluation is proposed to estimate thermally-induced changes in macroscopic parameters (e.g., Young’s modulus) based on the corresponding mesoscopic parameters (e.g., particle size distribution and linear crack density).