Durability assessment of natural stone
Durability assessment of natural stone (in The stone cycle and the conservation of historic buildings, J. Cassar, M. G. Winter, B. R. Marker, E. N. Bromhead, J. W. N. Smith, D. G. Toll, N. R. G. Walton, D. C. Entwisle and T. A. Dijkstra)
Quarterly Journal of Engineering Geology and Hydrogeology (November 2013) 46 (4): 377-390
Durability is one of the most significant aspects of the behaviours of natural stone in the real environment. As it is not a fundamental property, it cannot be measured in situ or by using any simple laboratory test. Therefore, durability is assessed by using several approaches such as (1) accelerated laboratory durability tests (freeze-thaw cycling, wetting-drying, salt crystallization resistance), (2) complex environmental testing in test cabinets, and (3) exposure site testing. In this review paper, the advantages and drawbacks of these approaches are discussed and compared with other approaches such as practical experience and/or indirect assessments of selected physical properties, which have been found empirically to have a strong influence upon the durability of materials. Based on current information, accelerated laboratory durability tests (although the most widespread type of tests) exhibit serious limitations in terms of oversimplified test methods. Complex environmental testing provides more reliable results, but is less accessible for common testing owing to experiments' long duration and high operational costs. Similar drawbacks are applicable for in situ testing. Practical experience combines the value of the material"s exposure in real-life conditions and the duration of interactions; but it is highly impractical when the sampling of experimental materials is required. Extrapolation of the durability of natural stone from the results of some common laboratory tests (porosimetry approach for complex evaluation of pore space; mechanical properties under various degrees of saturation; response of a natural stone exposed to water (i.e. swelling behaviour)) seems to be an appropriate alternative to previous approaches.