8 7. DETERIORATION PATTERNS

Conditions required for surface and subsurface crystallization of sodium chloride, leading to flaking or powdering of porous material, have been clearly described by Lewin (1982). These rely on an equilibrium between the rates of capillary migration and evaporation, and Lewin postulated that the necessary condition for this type of decay is the development of a steady state at the exposed surface. Since capillary migration is not the only mechanism involved in bringing water and salts into a porous material, a steady state is hard to imagine given the variability in environmental conditions that may lead to condensation and moisture-retention phenomena. These may play a substantial role in the amount of moisture present in the material which in turn may enhance the aging of the salt deposits influencing the deterioration pattern (Arnold and Zehnder 1985, Delgado Rodrigues 1991).

The two most common weathering patterns that can be found on any type of porous material are flaking (or contour scaling) and powdering (or sanding). Although the patterns in different materials may be similar and salts have always been reported, mechanisms that produce deterioration may vary from material to material and among salts of different solubililities. For example, in the case of clay-bearing sandstones, such as the ubiquitous brownstone in New York City, it has been shown that the deterioration may be attributed to the effect that soluble salts have on the normal hygric dilatation of these sandstones. While both hygric and hydric dilatation have been proved reversible in short-term cycling (Snethlage and Wendler 1997), the presence of soluble salts changes this behavior significantly and irreversibly (Wendler and R�ckert-Th�mling 1992). When salts such as sodium chloride, magnesium sulfate, and calcium nitrate are present, a stone will contract rather than expand in a moist atmosphere during cycling between 35% and 90% RH. This behavior has been attributed to the effect ions have in modifying the thickness of the double layer formed on the surface of clay minerals. Furthermore, salts concentrate in the area that retains moisture longer—i.e., the zone of maximum moisture content. From numerical model calculations, this area has been shown to be located 1–5 cm from the surface, depending on the porosity of the stone (Wendler 1991, Snethlage and Wendler 1997). This zone is mechanically stressed, leading to disrupture and eventual powdering with the detachment of surface flakes. The presence of less-soluble salts, such as gypsum, exerts a mechanical wedge action through precipitation in interstitial pores, contributing to the irreversibility of the dilatation process (Wendler and R�ckert-Th�mling 1992).

The mechanism resulting in flaking and sanding of porous calcareous stones appears to be different, particularly for stones exposed to air pollution. Acid attack, including carbonic acid in unpolluted rain, results in the migration of Ca++ ions to the surface, inducing a relative hardening by formation of a gypsum or a calcium carbonate crust. The underlying Ca++-depleted area loses mechanical strength and powders, following a mechanism similar to that described for tooth decay (Cussler and Featherstone 1981). This result was clearly described by W. Domaslowski (1982) and repeatedly observed in the field for more porous stones (Charola and Koestler 1985–86; Gisbert et al. 1996). However, whether the above mechanism is operative or whether a thin adherent dark gypsum layer forms on the surface is strongly dependent on the texture of the limestone, as indicated by R. Kozlowski et al. (1990). The presence of soluble salts, such as halite, can induce both flaking and sanding, depending on environmental conditions (Amadori et al. 1990). The laboratory studies of M. Amadori et al. showed that the overall porosity increased and that this salt tended to concentrate in smaller pores in the interior of the stone, as confirmed by the experiments of Rodriguez-Navarro and Doehne (1999). The presence of clays in limestones, particularly if concentrated along bedding planes, will induce delamination and scaling. This phenomenon is enhanced by the presence of salts that will contribute to increased moisture absorption (Rodriguez-Navarro et al. 1997).

The formation of gypsum crusts on low-porosity calcareous stones, such as marbles and compact limestones, has been addressed by Camuffo et al. (1982, 1983, 1987). A gypsum crust is formed by attack from acidic pollutants, and it is suggested that differences in thermal expansion between gypsum and the calcite crystals favor disruption of stone surfaces. The crystallization of gypsum between grains and in fissures provides a mechanical wedge action that leads to flaking. Soluble salts, such as halite, can also induce some flaking, but their presence results mainly in sanding (Amadori et al. 1990). Sugaring, as it is more commonly called in the case of marbles, is strongly influenced by the anisotropic behavior of the calcite under thermal stresses (Zezza et al. 1985). The presence of clays in compact limestones may give rise to different weathering patterns depending on their distribution within the limestone matrix. If present in layers, flaking and scaling will result, but if present in discrete pockets, an alveolar pattern may develop (Aires-Barros et al. 1998).

In the case of granites, flaking has been correlated to the crystallization of gypsum originating in mortars, particularly in areas where the stone remains moist for a longer time (Rivas Brea et al. 1994). On the other hand, sanding has been attributed to the presence of soluble salts as well as stresses induced during tooling of surfaces (Silva et al. 1994). Further studies have shown that flaking can occur in both fairly sound and weathered granites, while sanding occurs only in the latter. While deterioration is considered a physical process, influenced by temperature and hydric behavior among other factors, increased weathering corresponds to increased chemical changes such as leaching of key elements and oxidation-reduction reactions (Delgado Rodrigues 1996).