5.1. Mineralogy and Microstructure
The weathering layers observed on the tuff surfaces have diverse color, morphology, and extension, and the mineralogical and microscopic analyses reveal the most significant components being C particles, gypsum, and swelling clay minerals, from both Eger Castle and Sirok Castle.
Soiling and black crusts contribute to the most typical discolorations, produced by the deposition and cementation at different degrees of blackish C particles (Figure 4
). These crusts frequently have mid-high levels of Cl and a variable extension. In Eger Castle, they reach the greatest thickness, exceeding 100 µm. This is consistent with the more polluted environment of a larger town like Eger. There, motor vehicle exhaust emissions–typically composed of spherical C particles, coated with NOx
, VOC, sulfates, metals, etc. [41
]–are supposedly more abundant, even more so considering that a major road runs just 20–30 m from the castle walls. Likewise, a railway runs a few meters away, between the road and the castle, so that the steam and, later, diesel locomotives in transit since 1908 represent a further source of soot.
The occurrence of gypsum is determined at different degrees, from a general sulfation of the tuff surfaces to the presence of discrete discernible crystals (Figure 4
). Their grain size is the largest in Eger Castle. The XRD results indicate that gypsum is the secondary phase most frequently detected (Figure 5
). Nevertheless, it is often finely dispersed on the stone surface, without forming compact aggregates of coarse-grained crystals. This last observation is explained by the low air concentration of SO2
(about 7 µg/m3
, Table 1
), typical of a rural environment [44
], and the lack of a major intrinsic source of Ca. CaO concentration in the studied tuffs is typically below 3% [33
], so that the plausible source is external and represented by lime-mortar joints, from which CaCO3
dissolves and then reprecipitates. The role of mortars in gypsum formation on volcanic rocks has been already postulated in the literature [43
], although some studies indicate the possible contribution of the rock-forming minerals, e.g., calcic plagioclases [15
]. Generally, the possible role of wind-blown carbonate grains, from loess or calcite-rich rocks, needs also to be considered [48
The deterioration risk associated with weathering crusts with a mix of organic and inorganic components is the formation of surface layers with different physico-mechanical and chemical properties in respect to the substrate. As such, they differently respond to environmental changes and stresses, producing strongly localized mechanical fatigue at the interface with the underlying host rock. The stresses generated by cyclic salt crystallization alone are acknowledged as a major decay cause [50
As for the clay minerals, specifically montmorillonite, their crystallization seems generally more widespread on the exposed surfaces. This arises after comparing the XRD signals from the host rock and the surface (Figure 5
) and the examination of the outermost layers under the microscope: there, the alteration degree of feldspar phenocryst fragments is higher, and the matrix, typically having very low crystallinity, shows several domains of devitrification to cryptocrystalline aggregates (Figure 6
a). Indeed, the clay minerals derived from the hydrolysis of the feldspars and, to a greater extent, the glassy matrix. Volcanic glass is thermodynamically unstable and alters more rapidly than the associated minerals and, when it has a silicic composition, montmorillonite is the most common product [54
]. Enrichments in swelling clay minerals may enhance stone vulnerability during wetting/drying alternating phases, producing damaging cycles of dilation/contraction [55
]. Differential erosion may also increase [7
]. Figure 6
b shows a characteristic surface topography of the tuff from the historical walls of Eger Castle (c.f. Figure 2
): the surface outline follows the coarse crystals, pumice, and lithic clasts standing out in relief in respect to the weaker matrix, which is further stressed by the action of clay minerals.
Finally, it is worth mentioning a weathering pattern observed more frequently at Sirok Castle, i.e., orange patinas mainly constituted of goethite (Figure 5
). This phase might derive from the leaching of Fe-bearing minerals and subsequent Fe mobilization [43
], especially from biotite–which releases Fe2+
from the octahedral layers–or the basaltic-andesitic lithoclasts [29
]. Another source might be the glassy matrix that, during hydration, is subjected to the strong oxidation of the Fe contained therein [54
]. In fact, an enrichment in Fe oxides and hydroxides was observed nearby biotites and in the matrix, spreading out following the pore channels and penetrating into the pumice clasts (Figure 6
c). This was also in agreement with the higher biotite concentration in Sirok tuff, estimated by XRD (Table 2
), which exceeds 50% in some samples.
5.2. Trace-Element Chemical Composition
The ICP-MS analyses of the non-silicate fraction of crusts, patinas, and deposits on the tuff surfaces, as well as of the host rock, allow outlining the concentration patterns of trace elements. The most interesting data are related to heavy metals (Figure 7
). Generally, compared to the bulk stone, the weathered surfaces are noticeably richer in As, Co, Cr, Cu, Fe, Mn, Mo, Ni, Pb, Sb, Sn, V, and Zn. The absolute maximum concentrations for some illustrative elements are: As 40 ppm, Cr 28 ppm, Cu 48 ppm, Ni 28 ppm, Pb 89 ppm, V 135 ppm, Zn 79 ppm. The lowest and highest average concentrations are both detected at Eger Castle, on the historical and restored walls, respectively. As for the balance of trace-element concentration between the surface and the host rock, the historical walls of Eger Castle report the lowest decrease of heavy metals in the inner layers, i.e., less than 30% on average: for instance, Fe is 10,027 ppm on the surface and 8579 ppm underneath, instead Pb is 5.6 ppm versus 4.3 ppm. On the contrary, the other tuffs record mean differences between 45% (Eger Castle) and 55% (Sirok Castle), pointing out that the accumulation of those trace elements is mostly superficial.
Heavy metals represent air pollution markers and are useful for understanding the environmental context. The anthropogenic sources may be numerous, e.g., transports, industry, construction, electricity and heat production, petroleum refining, incinerators [59
]. However, in this case, the only actual influential source was assessed to be transports. Road transport is associated with diverse major emissions, mostly coming from the wear of car brakes and tires and composed of Cu, Zn, Pb and, secondarily, Cr and Ni, although including a variety of other elements (As, Cd, Hg, Sb, Se, V) and sources (motor fuel and oil, asphalt abrasion) [59
]. In the case of Eger Castle, the contribution of rail transport also needs to be accounted for, in view of the steam and diesel engines and the abrasion of metallic train elements and rails releasing Pb, Zn, Cd, Cu, Cr, and Fe mainly [62
These observations are only partly concordant with the experimental results. In fact, the concentration of heavy metals determined at Eger Castle can be higher than the levels of Sirok Castle, consistent with the different localization of the monuments in respect to the surrounding pollution sources. However, those somehow narrow geochemical differences do not indicate two much different environmental contexts. Moreover, within Eger Castle, so in the same higher-pollution scenario, stone chemical alteration can even be the least intense, depending on the tuff variety, as already pointed out. Pb concentration can be taken as a representative example: the average values recorded, ranging from about 5 to 20 ppm, are very far from the 100s and 1000s ppm detected in black crusts from Budapest and other European cities [49
]. In other terms, the boundary environmental conditions of the castles of Eger and Sirok and air quality may be less significant than the lithological constraints in defining the accumulation of metallic pollutants.
Therefore, metal mobility and its dependence on the rock properties must be taken into account too. The tuff from the historical walls of Eger Castle, if compared to the others, has a significantly higher open porosity (by a factor of 1.4 to 1.6) given almost totally by large pore-size classes (i.e., capillary pores and macropores [11
]), those mainly affected by water infiltration and movement. These characteristics allow for a boosted pollutant absorption and deeper migration inwards, even in liquid solution when the surface gets wet. That explains the low heavy-metal concentrations on the surface and the relative geochemical balance with the inner layers. Finally, that tuff is also the softest and most prone to disintegration, so that, during the enhanced surface recession (Figure 2
), the accumulation of airborne particulate may reset frequently.
The results of the N2
adsorption tests provide evidence that weathering produces measurable changes even in the pore network, and most importantly may lead to a decrease in microporosity. The microporosity of the surface layers is 3/5 that of the host rock for the tuff from the historical walls of Eger Castle, and about 4/5 for Sirok tuff. The microporosity range measured goes from 11% (host rock) down to 1.8% (weathered surface). The pore-size distributions point out another trend, namely the general decrease of pore size in the weathered surface: while the finest micropores apparently are almost unaffected, the quantity of those with a diameter larger than 5–10 nm decreases significantly (Figure 8
These data indicate that the deposition of exogenous pollutants and surface crystallization of secondary phases lead to micropore filling, or that the weathering crusts may reach a higher density than the deteriorated underlying host rock. The grain size of the weathering-related components is constrained by the lower limit of 5–10 nm that marks the main pore-size changes. When an increase of BET surface area is also recorded, that translates to a higher chemical reactivity, i.e., larger surfaces accessible to decay processes [67
]. Nevertheless, a microporosity decrease may also result in hindering further absorption of water vapor and dissolved pollutants; in fact, microporosity is strictly associated with hygroscopic condensation, which can occur in micropores even for values of relative air humidity much lower than 100%–the smaller the pore, the lower the humidity of condensation [68
The aforementioned considerations refer only to the investigated size range of micropores, which have a different weight depending on the tuff variety. A previous study of Eger Castle, in fact, reports a weathering-related increase of larger pores, involved in possible enhanced processes driven by liquid water [7
]. One example is the increased accumulation of heavy metals in the historical walls, mentioned previously.
Finally, it is worth mentioning the only data that do not conform to the general trend, obtained on the tuff of the restored walls of Eger Castle. Its weathered surface has a porosity increased by a factor of 1.2, compared to the host rock. This might indicate a less advanced stage of compaction, influenced by, among other things, the shorter exposure time from the restorations and the relative lowest open porosity [11