Bioweathering of Egyptian Nubian sandstone and Theban limestone: three months insight by experimental incubation

This study focuses on Nubian sandstone and Theban limestone (Pomeyrol, 1968), which are significant ancient architectural and sculptural materials widely used during the Medium Kingdom and New Kingdom periods (Klemm & Klemm, 2001). Nubian sandstone can be defined geologically as a clastic sequence primarily composed of sandstones intercalated with mudstones and characterized by the absence of marine fossils. Geographically, it extends beyond the border of Egypt into Libya and Sudan, spanning several degrees of latitude and longitude (Pomeyrol, 1968). The Nubian Sandstone is exposed on both edges of the Nile Valley in southern Egypt, the northern Red Sea coast, and the middle of Sinai. It should be known that due to the challenge of stratigraphic correlation, there is no formal stratigraphic nomenclature for the Nubian sandstone formations and their broad extension (El-Shayeb et al., 2013). Furthermore, sandstone has been utilized in numerous Egyptian antiquities, both as a building material and in the production of statues, stelae, sphinxes, and rams since the Middle Kingdom era, during the reign of Mentuhotep II. Its use reached its zenith in the New Kingdom era and continued into the Ptolemaic and Roman eras, such as in the Hathor temples in Dendera. For instance in sculpturing, sandstone has been used in carving statue of Amenhotep I (18th dynasty) from Deir El-Bahari (Harrell, 2012). The Nubian Sandstone exposed in the Nile Valley escarpments and used for architectural purposes is formed as a siliceous and rarely calcified quartz sandstone (Klemm & Klemm, 2001). It comprises several successive lithofacies. Feldspathic quartz arenite is the most common, while sublithic quartz arenite and calcareous quartz arenite are less frequently encountered (Temraz & Khallaf, 2016). The Nubian Sandstone, which was used as a building material in ancient times, provided two types of sandstones that were unusually hard: iron oxide-cemented ferricrete and quartzcemented quartzite. These sandstones were utilized in buildings and numerous ornamental applications where enhanced strength was required (Emberling & Williams, 2021; Harrell & Mohamed, 2021). The scope of previous research on the impact of microbiological damage to ancient monuments, including Egyptian sandstone such as the Karnak Temple, has been restricted to the effects of fungi and their biological agents, as well as other biological damage products (Demkina et al., 2010; Emberling & Williams, 2021; Geweely & Afifi, 2011; Scheerer et al., 2009).

The use of Theban limestone in architecture during the Middle and New Kingdoms was not limited to the Theban region. Petrographic and geochemical analyses have shown that the limestone used in these periods originated from two main mining districts: the Dibabiya (E. Gebelein) region, located approximately 30 km south of Thebes, and the Tura-Maasara region, located approximately 600 km north. Therefore, the term “Theban limestone” used thereafter refers to limestone that was quarried from the Theban region, but not necessarily from the same region where the architecture was built. In the vicinity of Luxor, Theban limestone ranges from marly and dense to chalky and porous variety. The latter, exposed on the east bank of the Nile, was quarried already in the times of the Old Kingdom, used to build 27 pyramids (Klemm & Klemm, 2001). Only one small-size limestone quarry is known locally, on the West Bank, i.e. the ‘Hatshepsut quarry’ near Qurna, which yielded a low-grade porous limestone that was used only in the New Kingdom (Karlshausen & De Putter, 2020; Bradley & Middleton, 1988; Dakal & Cameotra, 2012; Sterflinger & Piñar, 2013), where our limestone samples were collected. Several studies and archaeological expeditions have provided information on the geology of Theban limestone and the decay of some Egyptian limestone sculptures in open sites and during storage in a museum environment including biological factors and their undesirable aspects (Bradley & Middleton, 1988; Dakal & Cameotra, 2012; Sterflinger & Piñar, 2013). However, most of these studies have focused on fungi rather than bacteria, particularly in terms of species identification and their effects on limestone. In contrast, sandstone has been the subject of relatively few studies (Chen et al., 2021; El-Derby et al., 2016; El-Gohary, 2015; Lewin & Winkler, 1974). The main types of bacteria affecting stones include autotrophic and heterotrophic bacteria and the former have a great influence on the primary weathering of minerals and rocks (Fitzner & Heinrichs, 2001). The decay of some Egyptian limestone sculptures occurs during storage in a museum environment. Fitzner’s classification scheme of weathering forms identifies various microbiological colonization such as pitching, etching, pore filling, biofilm formation (Cappitelli et al., 2020; Demkina et al., 2010; Warscheid, 2000).

Samples of Nubian sandstone were obtained from an ancient quarry (Nag el-Falilih) located approximately 38.5 km to the south of Gebel Silsila East (Fig. 1). This quarry which belongs to “Umm Barmil” geological Formation is located at 24°20.04’ N, 32°55.27’ E (24.3340000N, 32.9211667E). The petrology of this sandstone was described (Harrell, 2016): very fine-to mainly fine-grained; massive to indistinct tabular cross-bedding; light yellowish-to pinkish brown (total feldspar = 3.5%). Samples of Theban limestone were collected from area Dra’ Abu El-Naga about 1.7 km to the northeast of Hatshepsut Temple on the west bank of the Nile at 25°44’ 54’’ N, 32°34’ 44’’E (25.7483333N, 32.5788889E) (Fig. 1). Two representative samples of Nubian Sandstone (S3 and S6) and one sample of Theban limestone (TL) have been chosen for research. Samples were cut into small cubes using a grinding machine and crushed, then sieved to the required size (0.5-2 mm) for experimental bio-weathering via bacteria (Pseudomonas fluorescens).

Figure 1.

The chemical composition of sandstones (Tab. 1) was determined in a commercial laboratory (Bureau Veritas Analytical Laboratories, Vancouver, Canada). Samples were subjected to lithium borate/tetraborate fusion and digestion of the melting product with nitric acid. Inductively coupled plasma atomic mass spectrometry (ICP-MS) measurement allowed to determine bulk elemental composition. Analytical accuracy as estimated from 3 measurements of standard (STD SO-19) was assessed at the 95% confidence limits.

Polarized light microscopy (PLM) and scanning electron microscopy (SEM-EDX) were used for mineralogical characterization of the Nubian sandstone and Theban limestone samples. Standard petrographic as well as polished thin sections (35–40 μm in thickness) were used for the petrographic analysis. The modal composition of the samples was determined using JMicroVision software (Roduit, 2007). A series of microphotographs of thin sections and BSE images were analyzed using the point count method. In addition, for sandstone samples, Feret max (the longest Feret diameter) was determined for at least 500 grains per sample using JMicroVision software (Roduit, 2007). The results were then processed using the Gradistat grain size distribution and statistics package (Blott & Pye, 2001). The descriptive parameters were evaluated according to the Folk and Ward method (Folk & Ward, 1957). The degree of roundness of sandstone framework grains was determined using the chart for visual evaluation (Powers, 1953). Prior to further analyses, the sandstone samples were gently crushed in an agate mortar and then sieved (using a 63 μm sieve) to separate the framework grains from the matrix and cement. Thermal analysis (DSC-TG) was performed using a Perkin-Elmer STA6000 calorimeter equipped with an Al₂O₃ sample pan, in the temperature range of 40-999 °C at a constant heating rate of 15 °C/min, in an N₂ atmosphere. PYRIS software was used for data interpretation. X-ray diffraction (XRD) analysis was performed using a Bruker D8 Advance diffractometer operating at 40 kV and 40 mA, with CuKα radiation. All measurements were performed in the 2θ range from 5° to 75°, with scan times of 0.01° and 2 seconds. Results were compared to standard reference patterns using the ICDD database (ICDD, 2003).

This study investigated the contribution of heterotrophic bacteria, Pseudomonas fluorescens (strain DSM 50091 purchased from the Deutsche Sammlung von Mikroorganismen und Zellkulturen; Braunschweig, Germany), to the dissolution of sandstones and limestone as representatives of sedimentary rocks (Fig. 2). The reason for choice of Pseudomonas fluorescens bacterium was following: (i) this strain has been widely recognized for colonization of rocks and minerals via biofilm formation (Abu Quba et al., 2023; Vieira & Melo, 1995), (ii) this strain produces siderophores known to enhance the mineral dissolution (Kalinowski et al. 2000; Reichard et al., 2007), (iii) this strain was identified on monuments in locations close to that of studied area (Khalil et al., 2022; Sakr et al., 2018). Bacteria were cultivated according to the provider recommendation and grown in the Luria-Bertani medium (non-selective nutrient culture medium) for 24 h (Bertani et al., 2004). Afterwards, the cells were harvested and cleaned three times with sterile 5% NaCl and resuspended in a growth medium composed of (per L⁻¹) 4 g of succinic acid (C₄H₆O₄), 1 g of ammonium sulfate ((NH₄)₂SO₄), 0.2 g of disodium hydrogen phosphate (Na₂HPO₄), and 6 g of Tris buffer (C₄H₁₁NO₃). The initial pH of the growth medium was adjusted to 7.0 ± 0.1 using 5 M NaOH.

Figure 2.

Solid samples with a particle size of 0.5–2 mm, all glassware and polypropylene materials were sterilized (VWR, Vapour Line Lite Autoclave) at 121 °C for 20 min prior to the experiments (Fig. 2). Next, 1 g of each sample was placed in Erlenmeyer flask and filled with the following weathering solutions: i) demineralized water (H₂O) as a control unaggressive conditions; ii) a sterile growth medium (MED) simulating abiotic weathering conditions; and iii) a growth medium inoculated with the bacteria Pseudomonas fluorescens (BAC) simulating biotic weathering. Incubation flasks were kept on an orbital shaker set at 100 rpm at temperature 30 °C. Consecutive sampling intervals were fixed at 14, 28, 60 and 90 days. Leachate was filtered using 0.22 μm PTFE filters and acidified with Suprapur 65 wt.% HNO₃ (Merck). The concentrations of elements in the leaching solutions were analyzed by a certified laboratory using inductively coupled plasma mass spectrometry (ICP-MS, Nexion 300D Perkin Elmer). Quality control included systematic analysis of certified reference materials.

At the end of the incubation experiment, the spectrophotometric (HACH DR9300) measurements were done at the wavelength range 330–430 nm for siderophore detection (Perez et al., 2019), whereas its concentration was assessed based on absorbance measured at 405 nm. The presence of two specific shoulder peaks at the wavelength range 330–430 nm confirmed siderophore to be present in the solution, whereas siderophore concentration was determined according to the Beer-Lambert law and molar extinction coefficient of free pyoverdine (Tseng et al., 2006). Since various siderophore complexes were expected in the studied incubation setup, the value for a free pyoverdine was chosen for calculation for the simplicity and comparative reasons.

Studied sandstones are primarily composed of SiO₂ reaching 92.6 wt.% and 85.7 wt.% for S3 and S6 sandstone, respectively (Tab. 1). The S3 sandstone is enriched in Al₂O₃ (3.12 wt.%) relative to S6 sandstone (0.96 wt.%). Sandstone S6 is characterized by a notably higher quantity of Fe₂O₃ (nearly 8 wt.%) relative to sandstone S3 (1.2 wt.%). Other components including MgO, CaO, Na₂O, K₂O, Ti₂O generally do not exceed 1 wt.% with an exception observed for CaO (1.6 wt.%) in case of sandstone S6 (Table 1).

Sandstone samples S3 and S6 are predominantly composed of quartz (Fig. 3; 62.9-58.0 vol. %), accompanied by minor amounts of alkali feldspars (Fig. 4; 8.6-5.5 vol. %) and lithic grains (4.8-5.8 vol. %). Accessory minerals (0.2-0.1 vol. %) such as rutile, zircon, apatite, tourmaline, amphibole, pyroxene, monazite, chlorite and opaques (ilmenite, magnetite, galena) are also present. Quartz is observed as either strained or unstrained monocrystalline grains. Polycrystalline grains are extremely rare. The samples contain hatch-twinned microcline or heterogeneous perthite. Fine-grained leucocratic crystalline rocks are composed of quartz and feldspars. Opaques are mainly anhedral magnetite or, less commonly, ilmenite (Fig. 4). Their crystals occupy pore spaces. Galena is a very rare phase. It occurs as tiny grains (up to a few microns) in a clay matrix.

Figure 3.

Figure 4.

Microscopic examination shows that sandstones S3 and S6 are dominated by the sand fraction (93.69-8.9%). They are fine to medium grained in texture and moderately well sorted. The content of the silt fraction is very limited (6.4-1.1%). Framework grains are mostly angular or subangular to less commonly subrounded, contacts concavo-convex or tangential (Fig. 3). The main components of the framework form relatively large, sand-sized grains, while accessories are much smaller, with diameters below 50 μm.

The matrix, interstitial to the framework grains, is composed of clay (Fig. 3) with a few small clastic grains embedded. It occupies 8.2-11.7 vol. % of the rock. Kaolinite is the main component of matrix (Figs. 4 and 5) and occurs as pore-filling aggregates. They are composed of irregularly oriented flakes or occur in the form of a concertina (book-like) morphology (Fig. 5). The aggregates are locally enriched in microcrystalline goethite/lepidocrocite or hematite, heterogeneously distributed among the kaolinite flakes (Fig. 5). In addition to the Fe-bearing phases, tiny grains of calcite (Fig. 4) are found distributed among the kaolinite lamellae. The Fe-bearing phases also occur as pore lining cement. Syntaxial quartz overgrowths (Fig. 6) are the next and most important type of cement. The rock is highly porous (15.5-10.2 vol. %) and the voids have irregular shapes and sizes ranging from a few μm to about 100 μm. According to the Pettijohn classification (Pettijohn et al., 1973) for clastic rocks both sandstone samples are subarcose.

Figure 5.

Figure 6.

Chemically, limestone is mainly composed of CaO (50.2 wt.%) and SiO₂ (7.1 wt.%), whereas other components are present in the quantities not exceeding 1 wt.% (Tab. 1). Loss on ignition was 40.7 which is evident for limestones.

The rock is mostly calcium carbonate mud (Fig. 7). It consists of coccoliths (Fig. 6) up to 10 μm in size. Foraminifera and ostracode bioclasts up to 300 μm in size are scattered throughout the rock (Fig. 7). Some of the microfossils have thin encrustations of Fe-bearing minerals (Fig. 7). Euhedral dolomite crystals (Figs. 5 and 6) up to 10-20 microns are scattered throughout this matrix. A small number of clastic grains (quartz, alkali feldspar; Fig. 4) up to 10 μm in diameter occur in the rock. Accessories include rutile (up to 25μm), barite and Fe-Cu sulfide (up to 1μm), apatite (biogenic, up to 5μm). The most common accessory minerals, however, are the relatively large opaque minerals (Fig. 7). These are anhedral to rarely euhedral grains (up to 100 μm) consisting of iron oxides (magnetite), Fe-Ti (ilmenite/Timagnetite), partly weathered. The rock is very porous (21.5% by volume). According to the classification (Dunham, 1962; Folk, 1959) it is an fossiliferous micrite/mudstone.

Figure 7.

Sandstone behavior varied from a one incubation conditions to another (Figs. 7, 8) and was specific to the elements initially bound in the samples. Slightly higher bacterial influence on Si release was observed for S3 sandstone reaching up to 392 mg kg⁻¹ (BAC), whereas its release under abiotic conditions was 342 mg kg⁻¹ and 209 mg kg⁻¹ for MED and H₂O, respectively. There was a less pronounced difference in case of S6 sandstone for which the Si release was up to 516 mg kg⁻¹, 540 mg kg⁻¹ and 392 mg kg⁻¹ for BAC, MED and H₂O, respectively. The amount of Al released under BAC conditions was up to 7 mg kg⁻¹ and 10 mg kg⁻¹ for S3 and S6 sandstone respectively. Likewise, the Fe leaching under biotic conditions was up to 6.1 mg kg⁻¹ and 8.9 mg kg⁻¹ under biotic conditions with S3 and S6 sandstone respectively. Its release under abiotic conditions oscillated at the level 0.2-8.9 mg kg⁻¹ and 0.7-2.4 mg kg⁻¹ taking into account both H₂O and MED conditions (Fig. 8).

Figure 8.

The pH changes recorded in individual treatments were displayed in Table 2. Siderophore concentration in biotic incubation was assessed at 45.2 μM L⁻¹ and 75.5 μM L⁻¹ for S3 and S6 sandstones, respectively (Tab. 2).

The Si release from limestone revealed values up to 420 mg kg⁻¹, 342 mg kg⁻¹ and 269 mg kg⁻¹ for BAC, MED and H₂O conditions, respectively (Fig. 9). The Ca release being the most important element from the view of the studied sample was released in the quantity of up to 381 mg kg⁻¹ for BAC, 2370 mg kg⁻¹ for MED and 202 mg kg⁻¹ for H₂O (Fig. 9). An immersion of limestone in the weathering solution resulted in the release of Fe at the level of 8.1 mg kg⁻¹, 1.9 mg kg⁻¹ and 1.1 mg kg⁻¹ in BAC, MED and H₂O, respectively. The difference between biotic and abiotic conditions was notable for Al that reached up to 10.7 mg kg⁻¹ in BAC, 2.5 mg kg⁻¹ in MED and 3.3 mg kg⁻¹ in H₂O (Fig. 9).

Figure 9.

The pH changes recorded in individual treatments involving TL were displayed in Table 2. Siderophore concentration in biotic incubation was assessed at 90.6 μM L⁻¹ (Tab. 2).

The behavior of dimension stones during immersion in the weathering solution varied from one incubation condition to another and the elements dissolution pattern was specific to the host mineral. Among the three samples studied, limestone appeared to be more significantly impacted by the presence of bacteria as compared to its influence on sandstone and expectedly its overall reactivity was higher than that observed for sandstones.

First of all, bacteria applied in biotic incubation is known to produce Fe-chelating siderophores even if complexation affinity towards other elements has also been demonstrated (Rachid & Ahmed, 2005). The Fe leaching at the end of biotic incubation was alike for both sandstones (6.1 mg kg⁻¹ for S3 and 8.9 mg kg⁻¹ for S6 for both corresponding to < 0.1%). The Fe leaching showed an increasing trend during 60 days of S3 incubation and depleted afterwards, whereas the trend was rather stable throughout the experiment with S6 sandstone. The most plausible reason for observed depletion in Fe leaching is lower concentration of siderophore in S3 incubation as compared to incubation containing sandstone S6 (Tab. 2). In addition, given the difference in chemical composition of sandstones (Fe₂O₃ 1.2 wt.% vs 8 wt.% for S3 and S6, respectively) combined with comparable quantitative Fe leaching under biotic conditions (Fig. 8), bacterial impact can be assigned to the structural location of the Fe-bearing pore lining cement where Pseudomonas fluorescens likely attached. Bacterial attachment on the porous solids has widely been reported (Cheng et al., 2019; Fletcher, 1985; Redman et al., 2004; Tuson & Weibel, 2013). Biotic experiments show that moderate enhancement (factors up to 1.5 and 5.7 for S3 and S6 respectively) in Fe release relative to the abiotic incubation occurred (Fig. 8). This indicates that initiation of Fe solubility was primarily driven by the Fe scarcity conditions that triggered bacteria to excrete siderophores. Looking at the elements of main interest (Si, Al and Fe), the release under biotic incubation generally did not exceed 0.1% neither in case of S3 nor S6 sandstone. This finding is in good agreement with previous studies investigating sandstone exposure to siderophore-producing bacteria that demonstrated dissolution not exceeding 1.3% (Si, Al, Fe) (Potysz & Bartz, 2023). However, it must be emphasized that this experimental study did not consider the amount of elements associated with biomass, while previous study has proven the role of bacterial cells in element capturing (Potysz & Bartz, 2023). Bacteria notably enhanced Al leaching (factors 2.2 for S3 and 7.0 for S6; Fig. 8) especially at the early stages of the experiment. Likely, kaolinite was the Al donor which is in accordance with previous findings proving the role of microbially-derived chelating agents on kaolinite dissolution and the progress of dissolution to be primarily driven by the presence of bacteria rather that surface reactivity itself (Ams et al., 2002; Grybos et al., 2011; Maurice et al., 2001). Furthermore, the impact of bacteria in terms of Ca and Mg release showed the release reaching 7.1% (Ca) and 2.1% (Mg) for S3 sandstone and the amounts of 2.5% (Ca) and 12.1% (Mg) were leached from S6 sandstone at the end of the BAC experiment (90 days). However, it must be emphasized that this high relative leaching was not solely driven by bacteria as chemical processes played predominant role (Fig. 8). Thus, Ca and Mg leaching was attributed to host minerals rather than by the conditions of exposure. These elements originated from microcrystalline carbonates allocated between the kaolinite structures. Susceptibility of microcrystalline carbonates to dissolution has previously been discussed (Chigira & Oyama, 2000).

Furthermore, there was a sharp discontinuity in Si and Al leaching showing relatively low Al leaching from sandstones as compared to that of Si (Fig. 8). Both elements are incorporated in kaolinite in similar proportions. However, looking at the leached kaolinite, it was impoverished in Al rather than in Si (Fig. 6). This discrepancy may imply that the Si content in the solution also originated from a mineral other than kaolinite, likely detrital quartz or syntaxial quartz overgrowths. Furthermore, larger Al leaching from S6 as compared to S3 confirms Al origin from kaolinite as S6 sandstone was far richer in Al-bearing clay matrix as compared to S3 sandstone.

In addition to the presence of microorganisms, consideration needs to be given to the chemical weathering in ultrapure water (H₂O) and sterile growth medium (MED). As mentioned, the remarkable leaching under chemical incubation conditions was noted for Ca (34.1% for S3 and 20.3% for S6) and Mg (7.7% for S3 and 27% for S6), recorded at the end of MED incubation. Undoubtedly, the disparity in Ca and Mg leaching is related to different content of carbonates occurring as intergrowths within the matrix. This result also suggests that the presence of biofilm on the surface of microcrystalline carbonates played a role of passivation layer inhibiting progress of dissolution as opposed to chemical conditions where the dissolution was not prevented. This effect was clearly evident throughout the experiment during which the Ca release was enhanced by (MED vs. BAC) factor 4.8 and 8.2 for S3 and S6 sandstone, respectively (Fig. 8).

Of particular interest is water-mediated leaching as such conditions are to be widely encountered in the natural environment (Malmström et al., 2000). As previously demonstrated, sandstones wetting leads to leaching of matrix and cement minerals, consequently increasing sandstones porosity and affecting further leaching (Lin et al., 2005; Potysz & Bartz, 2023). Even if the overall leaching in H₂O incubation was low, the Ca was found to be the most leached as compared to other elements (Fig. 8) due to the reason discussed above.

Looking at various leaching trends, it also needs to be noted that there were pH variations observed in the leachates throughout the incubation experiment. Generally, there was a pH shift observed towards higher pH values with respect to initially fixed pH of the solutions. Among the studied treatments, BAC conditions were found to have the highest pH values (Tab. 2) as compared to other treatments (MED and H₂O). This behavior of samples indicate the most intensive buffering during BAC incubation which is likely due to dissolution of solids (sandstones and limestones) being alkaline in nature (Jage et al., 2001; Nkoh et al., 2024; Sheldon et al., 2003). In addition, it could have also been aligned to the presence of bacteria Pseudomonas fluorescens known to retard acidification of the environment (Nkoh et al., 2024). Of a high importance is also the fact that increased bacterial population can be found under higher pH conditions (Ganeshan & Manoj Kumar, 2005) which likely resulted in enhanced leaching of some elements.

Looking at limestone behavior under experimental conditions, rock generally revealed higher susceptibility to dissolution than that of sandstones (Figs. 8 and 9). Relative values calculated for elements of interest revealed leaching not exceeding 1.5% with an exception observed for Mg leached in higher proportions (factor up to 42.1). First of all, bacteria (BAC) was found to enhance Si, Al and Fe release in a more extensive way than it took place in the case of sandstones; element release from limestone was enhanced under biotic conditions (BAC) by factors 13 (Al), 3.7 (Fe) and 1.2 (Si) compared to abiotic conditions (MED). The release of Fe in the studied incubation conditions revealed following trend BAC (0.3%) > MED (0.03%) > H₂O (<0.01%). There was a low volumetric proportion of Fe-bearing phases (likely FeOOH) inside the microfossils. Therefore, the microstructures colonized by the bacteria (BAC) could have easily released Fe, yet magnetite (Fig. 4) could have served as additional Fe source. Given a lower content of Fe₂O₃ in limestone as compared to sandstones (Tab. 1), there are two reasons for its susceptibility to biotic dissolution. Iron deficient conditions in BAC incubation with limestone triggered bacteria to siderophore excretion in a higher quantity (Saha et al., 2013) as confirmed by spectrophotometric analysis and calculation of siderophore concentration (Tab. 2). Consequently, it led to higher Fe release (up to 0.3%) than that observed for sandstones (Fig. 8). The Si release from limestone revealed the following trend according to the incubation conditions tested BAC (1.2%) > MED (1.0%) > H₂O (0.6%). We suggest that Si originated mostly from small amount of quartz (Fig. 4A) or opal, chalcedony, disseminated in pelitic form among carbonates. These mineral phases are the main non-carbonate components of the Theban Limestone. On the other hand, there is no clear information on Al release, however we hypothesize that dispersed clay minerals were Al donors. Therefore, they can also be a source of silicon, but they play a subordinate role compared to quartz. Even if such minerals have not been detected by the applied mineralogical methods likely due to methodological restrictions associated with the detection limits, these minerals have been reported in the literature (Cherblanc et al., 2016). Impact of bacteria (BAC) was evident enhancing Al release by factor up to 13 as compared to abiotic leaching (MED and H₂O) which is in accordance with previous findings showing increased Al leaching due to the presence of Pseudomonas sp. (Berthelin & Belgy, 1979). Likewise, it has been demonstrated that organic ligands enhance Al leaching by factor 3-500 than the water does (Keller et al., 1971). Furthermore, Mg leaching from limestone could be assigned to dolomite dissolution. In fact, it is an accessory mineral as compared to volumetrically prevailing calcite (Figs. 5 and 6). Maximum Mg leaching was observed in MED (16.2% at the end of the experiment), not BAC incubation (0.6% at the end of the experiment). Similar trend was observed for Ca leaching reaching 0.6% vs. 0.01% for MED and BAC, respectively (Fig. 9). Given these values, we suggest that bacteria Pseudomonas fluorescens preferentially colonize and dissolve larger euhedral dolomite crystals rather than microcrystalline calcite. However, in both cases bacterially-mediated leaching was lower than that observed in abiotic incubation (MED, H₂O) implying protective function of biofilm (De Belie, 2010) and potentially preferential colonization of individual minerals by microbial biofilms (Jones & Bennett, 2014). As previously reported, microbial attachment and productions of extracellular polysaccharides may decrease dissolution rates by 40-70% (Davis et al., 2007); in our case these factors were even more notable. It is worth to note that both, studied sandstones and limestones revealed inhibited Ca and Mg leaching under BAC conditions. In spite of an obvious chemical and mineralogical difference between these samples, it can be a proof of passivating effect caused by the bacteria towards liberation of these elements from solids to the liquid phase. In addition, the role of Ca and Mg in changing biofilm structure has been reported (Goode & Allen, 2011; He et al., 2016; Körstgens et al., 2001; Wu et al., 2022). These elements play a role in extracellular polymeric substances excretion and strengthening biofilm structure (Wu et al., 2022). In case of studied solids, the release of Ca and Mg could have also modified biofilm structure and effectively prevent progress of Ca and Mg leaching by bacteria. The discrepancy on the impact of Pseudomonas fluorescens bacterium on sandstone (Fig. 8) versus limestone (Fig. 9) weathering remains incompletely resolved at the present time. However, our experimental study proves that behaviour of bacteria in terms of siderophore excretion and biofilm formation depends on specificity of solid present in the weathering system. Our results suggest that deeper study considering the impact of biofilm on progress of bioweathering would be interesting to pursue.

Laboratory leaching tests reproducing bioweathering conditions encountered on-site allow to mimic potential behavior of dimension stone under exposure to chemical and biotic factors.

However, it must be emphasized that lab-scale processes are more intense in terms of elements leaching kinetics as compared to those occurring naturally (White & Brantley, 2003). The main reasons for discrepancies observed between lab-scale and on-site scenarios are following: (i) the contact between solid and liquid can change the leaching rates in favor to lab-conditions, (ii) humidity conditions including dry-wet weathering cycles occur in the field, while complete solid immersion is applied at the lab scale, (iii) experimental weathering proceeds under controlled temperature, while temperature fluctuations occur during on-site weathering, and (iv) lab-scale weathering involves solid exposure to individual weathering factor, whereas numerous weathering factors act at the same time onsite.

Experimental bioweathering performed for studied sandstones and limestones considered sample exposure to constant temperature of 30 °C. However, according to previous literature reports, even narrow temperature fluctuations can intensify of weathering due to occurrence of microcracks and intra-granular pores resulting in increased rock permeability (Jaber et al., 2024). As demonstrated by Alameen et al., (2024) generally sandstone dissolution increases as temperature rises regardless of pH conditions. However, influence of the temperature on leaching kinetics can vary from one element to another (Gong et al., 2012) mainly due to different sensitivity of host minerals to dissolution under certain temperature.

Furthermore, our experimental setup relied on stone immersion in the weathering fluids, whereas on-site conditions are characterized by humidity fluctuations (Sancho et al., 2003). Overall, humidity periods enhance element dissolution, whereas more intense secondary phases precipitation can be expected during dry periods (Sun & Zhang, 2019). Secondary phases formation can limit surface area exposed to weathering on a one hand and can increase salt-mediated weathering on the other (Yan et al., 2022). Dry conditions promote cracking and subsequently increase surface area exposed to weathering (Hua et al., 2015; Wang et al., 2021). Of a high importance is also a period of exposure to specific weathering factor. As demonstrated by Mitchell & Sass (2024) longer exposure to moisture can result in higher cracking relative to solids exposed to dry conditions. (Wells, et al. 2008) reported sharp increase of weathering rates as the time of exposure increases. Likewise, Niu et al. (2023) also highlighted that reaction time largely increase the stone damage. Furthermore, variety of biotic factors act simultaneously in real-time scenarios, therefore this aspect requires further consideration (see section 4.4.). Above-mentioned examples prove how complex weathering is and therefore the interplay of various weathering factors require specific consideration.

Comparative approach to dissolution susceptibility of studied samples was made applying theoretical estimates on time required for reaching complete sample deterioration. These estimates were made taking into account the dissolution rates of Si and Ca. These elements were chosen for calculation, because of its abundance in the sandstone and limestone samples, respectively. Assuming that Si leaching achieved after 90 days of incubation would proceed in a stable manner, approximately 215-574 years would be required for dissolution S3 and S6 sandstones regardless of conditions applied. In contrast, time calculated for limestone dissolution was assessed at 45-1890 years. It must be emphasized that this is simplified approach since the real weathering situation would involve many other factors shifting progress of the dissolution reactions (Li et al., 2008). However, such estimate proves overall higher limestone susceptibility to dissolution than sandstones in terms of initiation of dissolution reaction.

Given the susceptibility of dimension stones to deterioration factors, proper maintenance practices including sealing, cleaning, and addressing any damage promptly are required to extend their durability. Based on the findings of the present experimental study, several recommendations can be given for the preservation of structures or sculptures made from studied materials (sandstones and limestones). The guidelines can be categorized into two main approaches. Firstly, active conservation strategy including: (i) routine removal of microbial biofilms from stone surfaces is essential to prevent substantial deterioration (maintenance); this can be done using gentle chemical cleaning methods to avoid damaging the stone (Bosch-Roig et al., 2015); (ii) biological consolidation as a novel method that employs bacterial activity to consolidate deteriorated stone surfaces being particularly effective for carbonate stones (Jroundi et al., 2010, 2020). Secondly, implementation of preventive conservation strategy including: (i) environmental control (especially in museums) via maintaining optimal microclimate conditions (temperature, humidity, and light) in order to reduce bacterial growth on stone surfaces, (ii) regular monitoring to early detect signs of bacterial colonization and deterioration could enable timely intervention (Osman, 2019; Jroundi et al., 2020), c) protective coatings can serve as a barrier against bacterial colonization and environmental pollutants. These strategies are crucial for preserving the integrity and durability of archaeological stones used in both sculptures and architectural objects.

The relevance of studies dedicated to analysis of geochemical stability of dimension rocks is crucial for implementation of protective methods for conservation (Scrivano et al., 2018). Proper sealing, cleaning and addressing any damage could promptly extend the lifespan of stones. This study has shown that examining all major constituents of rocks provides insight into overall reactivity of such materials. Laboratory incubation favors dissolution rates, however the general mechanisms involved in the natural weathering systems may be alike to those studied in vitro simulations (Sitzia et al., 2021; Trudgill & Viles, 1998). However, predominant weathering conditions encountered in real-time scenarios might be much more variable. Thus, further need for comparative evaluation of rock dissolution persists for bioweathering to be fully understood (Potysz & Bartz, 2022). Various microorganisms affect the stability of rocks and minerals (phases) to a different extent in a natural weathering systems (Uroz et al., 2011), therefore consideration should be given to other microbial species in a single and mixed culture variants (Gaylarde & Baptista-Neto, 2021; Gleeson et al., 2006). Furthermore, accurate understanding of microbial metabolites is essential for quantifying mass transfer in terms of direct and indirect microbial contribution (Bosma et al., 1997; Widder et al., 2016). Contact and non-contact experiments allow to differentiate the effect of biofilm formation on rock reactivity from that of metabolites themselves. In other words, contact experiments enable microorganisms to attach to solid surface and form biofilm, whereas in non-contact experiments the attachment is prevented and only microbial metabolites may act due to solid encapsulation in the dialysis bag (Perez et al., 2019). In addition, it must be emphasized that dimension rocks are characterized by different degrees of surface roughness (Ribeiro & Paraguassú, 2008) due to rock intrinsic properties, but also due to preparation being specific to target application of stone. Consequently, surface roughness determines the range of space available for microbial attachment and may affect overall abundance of microbial community on the surface. Microorganisms attached to the solid surface may increase concentration of weathering agents (e.g. metabolites) beneath the biofilm on one hand and may prevent exposure of solid to dissolution on the other. All above-mentioned aspects remain an open debate, therefore an integrated geochemical approach quantifying habitat complexity and rock susceptibility to weathering is still required.