3.2. Geochemical Analysis of the Stone Materials
The concentrations of the major and minor oxides of the monument samples are presented in
Table 3. According to the classification of plutonic rocks of Middlemost [
30], most of the rocks are granites and quartz monzonites. Only the PU2P sample falls in the Syenite area.
For the geochemical analysis, bulk elemental composition and descriptive statistics for the studied samples are provided in
Table 3 and
Table 4. Some correlations between elements as oxides are also shown in
Figure 4.
The samples presented an average concentration of SiO
2 of 68.3% and Al
2O
3 in the order of 16.9%. Both show an inverse trend among them (
Figure 4a), as expected, due to their contents in feldspars and micas. The average content of Na
2O is in the order of 6.2%, due to the content in plagioclase. Na
2O shows an inverse correlation with K
2O (
Figure 4b), with a content in the order of 7.8%. Potassium is found mainly in micas and potassium feldspars, so, as expected, granites with higher potassium feldspars will contain less plagioclase. There is also a positive trend between K
2O and Rb
2O (
Figure 4c). Potassium and rubidium have similar charges, and substitutions may exist between them. Fe
2O
3 has an average concentration of 2.9%, which may be associated with biotite and Fe-oxides, while oxides, such as MgO, MnO and CaO, have concentrations of less than 1.5%. There is a significant positive correlation between oxides of Fe
2O
3 and MgO (
Figure 4d), fundamentally due to the content in biotite and with ZnO related mainly to Fe-Zn oxides (
Figure 4e). These elements are interesting as the oxidation of Fe and Mn due to chemical and biological processes usually results in visible chemical weathering effects such as stains and coatings [
14,
25]
Other elements, such as P
2O
5, are present in all samples with an average of 0.4% and showing a positive correlation with CaO (
Figure 4f), probably due to the formation of apatite, carbonatoapatite and hydroxyapatite [
31] in granites. However, samples PU4A, NO5, PU2P and Ria04 did not show the same trend and may be related to stains or coatings caused by an external origin or, alternatively, low concentrations in K- and alkali feldspars, and some accessory minerals of granite rocks (
Figure 4f). The samples with the highest concentration of calcium (Ria3, Ria4, Rib3, PU2P) also presented concentrations of barium oxide, related to substitutions between the Ca and Ba cations. Zr, Sr and Ba are trace elements whose contents tend to increase as the alteration of the granites in the monuments progresses, because they are poorly leached when compared with other more mobile elements, such as Ca, Na, K or Mn [
23,
32]. In addition, we cannot discard an external origin of one part of the CaO, at least in some monuments, as it has been demonstrated that such an element can come from dissolution of the lime of joint mortars in granite buildings [
23]. Furthermore, only sample NO 5 recorded SO
3 (1.2%). It is likely that the content in the other samples is very low and cannot be detected by this technique.
Several relative chemical weathering indices have been used for the quantification of sample weathering (
Table 5 and
Table 6), in an attempt to assess the weathering in the studied buildings and their correlation with the variables reported above. The calculations required for the assessment of such indices can be seen in
Table 5, and more details are provided elsewhere [
23,
32]. While absolute chemical weathering indices allow for the assessment of the degree of weathering, comparing the studied samples with a fresh (non-weathered) rock specimen, relative indices provide a quick assessment of chemical weathering by comparing rock samples affected by chemical weathering in different degrees [
23]. Correlations were also established between weathering indices (
Figure 5) to check their reliability to provide a picture of the decay of the granites. Weathering indices have been proposed for the quantification of weathering in the natural environment and weathering rates after exposure to artificial weathering tests [
23,
31,
32,
33,
34,
35,
36,
37,
38,
39,
40,
41,
42]. They usually provide scattered data, depending on the type of weathering (chemical or not), the rock heterogeneity and the rate of weathering. They can be either absolute or relative. Relative indices also allow for determining the weathering degree of different samples from the same rock type by comparing the proportion of some mobile (e.g., CaO, Na
2O,K
2O) and immobile oxides (e.g., Al
2O
3, Fe
2O
3, Si
2O), and they are expressed as ratios [
31,
32,
33,
34,
35,
36,
37,
38,
39,
40,
41,
42]. Alternative indices consider the loss on ignition (LOI) or the water content, reflecting the increase in weathering caused by the hydration of clays, and they are considered as sensitive to weathering under humid conditions [
20,
42,
43,
44,
45,
46,
47,
48]. On the contrary, absolute indices are based on the assumption that the fresh rock composition is known, being less applicable in building stone [
23,
31,
32,
33,
34,
35,
36,
37,
38,
39,
40,
41,
42,
43,
44,
45,
46,
47]. In general, the use of weathering indices on building stones is scarce, but they show an active potential for studying chemical weathering in all types of stone materials [
23,
43,
44,
45,
46,
47]. In this study, we tested the use of some indices that provided reliable results in granites from historical buildings in NW Spain [
23]. Such indices, their formulas, acronyms and authorship are provided in
Table 5. After assessing the indices, we concluded that the indices CIA, MWPI, SA, W
P, SI-Ti Index, Kr, ACN, AKR Rc are the most suitable to evaluate the weathering of the studied samples. It is clear that the correlation between two indices can be due, in part, to the fact that they use one or some oxides. However, as different indices use different oxides, we consider that their correlation provides some indications of decay and not only due to differences in the composition of granites. Indeed, comparing the correlations observed between the different indices for the area of Barbanza, with another similar study carried out with two different granites used in historical buildings of A Coruña (NW Spain), they are completely different [
23]. We consider this as evidence that the prevalence of weathering is the cause of the differences, more than differences in the granite composition, although we also compared these two variables below (indices vs. granite type).
The correlation of such indices vs. the studied variables was also observed. Taking the selected weathering indices, we compare them with the different considered variables. We consider that there is a linear correlation when the coefficient is above 0.5. Comparing the different weathering indices with the distance of the buildings to the shore, we do not observe any clear trend. The CIA index alone provides a positive correlation coefficient with the distance, higher for buildings located far from the shore. As there are several factors that can be correlated with this distance (e.g., effect of the sea-salt spray, precipitation), we divided the buildings into two different groups to refine the two variables. A group of buildings called Area 1 corresponds to those located at less than 500 m from the shore, while buildings in Area 2 correspond to those located more than 500 m away. Looking at the indices of the two groups, some of them provide clear linear correlations between the distance and the indices (see
Table 7). This is the case for the two groups for CIA, AKR and WPI for Area 1. In general, the lower the value of the index, the higher the weathering degree of the granite [
23]. The indices that show a correlation with distance for Area 1 show a negative correlation, indicative of more weathering degree at higher distances inland from the shoreline. As in the Barbanza Peninsula, there is a correlation of rainfall and humidity with distance to the sea, which could be the main cause of more decay in buildings far from the sea shore, at least until a distance of 500 m.
We do not observe the same result for Area 2 (distance above 500 m from the sea shore). Several more indices seem to fit better for Area 2, showing a clear correlation for MWPI, WP, VR, Si-Ti Index, Kr, ACM and ALK (
Table 7 and
Figure 6). However, half of them show a negative trend with distance while the other half show a positive trend. As most of them use the same elements, with different equations (see
Table 5), we can consider that they provide evidence of a lower weathering degree in areas far from the sea. Looking at some of the plots (
Figure 6), it can even be observed that there is a change in tendency from samples located at 1 km distance from the sea shore. So, it could be possible that the effect of sea salts can reach a higher distance (up to 1 km), with this area (between some hundreds of meters and 1 km distance to the sea) being where the effect of sea salt plus more humidity cause a more clear effect on granite, enhancing chemical weathering. Particularly, more samples should be considered to clarify this observation.
Considering other possible correlations, we checked the correlation coefficient between the altitude above the sea level and weathering indices. No significant correlations are observed for any samples. However, we divided the buildings again into two groups: Area 1 (<30 m above the sea level) and Area 2 (>30 m) (
Table 8). For Area 1, a correlation was observed for indices WPI, MWPI and ACR. Both WPI and AKN showed a strong negative correlation. This is indicative of a higher weathering degree at more elevated altitudes. We consider that the other indices did not provide a clear correlation because of the small number of samples considered.
For Area 2, a significant correlation is observed for CIA, MWPI, Si-Ti Index, Kr and AKR (
Figure 7). The MWPI shows a contrary trend to Area 1, while both CIA and AKR show a positive trend. In other words, there is a decrease in weathering with altitudes. Thus, it seems that there is a significant correlation between the weathering degree and altitude, which is different for the two areas. We can explain this because, as showed, buildings located close to the sea shore correspond to those located at lower altitudes, with one exception [
26]. Thus, again, it is observed that the buildings located in a strip area some hundreds of meters from the sea shore and at lower altitudes (between 10 and 30 m) correspond to more weathered granites. In addition, in the Barbanza Peninsula, there is a significant correlation of altitude with total annual precipitation. Thus, precipitation (and humidity) is not the sole factor that enhances weathering; some precipitation (and humidity) ranges, summed to the sea-salt spray effect, may be involved in the increasing weathering rates of the granites. It is possible that in this strip of land, a mixture of sea spray and rainfall causes the dissolution of salts in rainfall, and this water is less acidic. As granite minerals, such as quartz, feldspars and micas, are resistant to dissolution and hydrolysis at acidic pH but not basic [
49], the chemical weathering can be increased in this area. However, given the scarce number of samples used for this study, more research is needed to shed more light on this.
Looking at other possible variables, we considered the same ones related above (distance to the sea and altitude) with the type of granite, but we did not observe significant correlations. We also built three groups of samples considering the size of the population where the building is located. This is because historical buildings located in large populations can be exposed to higher air pollution concentrations. However, we did not observe any clear correlation with weathering indices, although it seems that it is possible that some indices can be different depending on this variable (
Figure 8). Considering the biological colonization of surfaces, we built three groups of samples and compared weathering indices. As occurs with air pollution, there is no clear trend, although some indices could show some weak correlation.
Concluding, we did not observe any clear correlation between the staining of granite surfaces in buildings due to weathering because of the type of granite, biological colonization or deposition of air pollutants, while it seems that there may be a correlation with weathering that should be studied in more detail.
3.3. Microscopic SEM-EDS
The SEM-EDX of the samples from Pazo de Goians (B09AP), Pazo de la Merce (PU3AP) and Church of Ribasieira (PO3) is shown in
Figure 9,
Figure 10 and
Figure 11. From
Figure 3, it is possible to see the location of the coating corresponding to each sample.
The most abundant minerals are identified from the micrographs as quartz, micas, K-feldspars and plagioclase. These minerals present surface alterations that show the chemical weathering of the rocks (
Figure 9a,f). Plagioclase frequently presents altered surfaces, either via the dissolution or precipitation of minerals (
Figure 9b,e). For its part, quartz shows characteristic microtextures, such as conchoidal fracture, straight and arcuate step (produced by fracturing processes), precipitation and dissolutions (
Figure 9c). Through this technique, the presence of microorganisms in the studied samples is also evident. (
Figure 9d,g,h,i)
The presence of Cl, S, P and C is related to the mineral neoformation, among which chloride salts (halite and calcium chloride), sulfates (possibly gypsum), amorphous calcium phosphates and carbonates (calcite, ankerite) stand out in samples BO9P and PU3AP (
Figure 9j,k and
Figure 10a–c). Both buildings are very close to the sea, thus evidencing the provenance and impact of the marine environment. Salt deposition, surface efflorescence and salt-rich coatings are expected in areas close to the sea [
20,
21,
22,
48,
50]. Such salts will contribute to the deterioration of the monumental stone [
1,
2,
3,
20,
21,
22,
48,
50,
51]. A recent work demonstrated that they decay granite in buildings in the Barbanza Peninsula [
26]. The more spherical morphologies could be related to air pollution. In the BO9AP sample, the element Te was identified, possibly related to the same salts (
Figure 9c). It is interesting and unexpected to find this element in this building, because it is located 500 m from the seaside. This is the area where the chemical weathering indices showed higher weathering.
Cl and S salts were not identified in the PO3 sample (
Figure 11). On the contrary, the percentages of C, Si and P in abundance reveal the biological action, some identified by means of this technique, such as fungi and thallus lichens. This building is located in Area 2, 4.5 km from the sea shore at an altitude of 167 m. Despite this, the chemical weathering indices did not show significant weathering at this distance, but visible chemical weathering features of rock-forming minerals are also visible under the SEM. Thus, despite the weathering indices probably not being resolutive enough, the SEM provided evidence of weathering. Due to the presence of biological colonization and the high humidity and average annual precipitation, we cannot discard the biological weathering as responsible for these effects [
14]. Indeed, Ti and Fe were identified under the SEM, previously reported in FRX, possibly related to oxides (
Figure 9l).