The Potential for Natural Stones from Northeastern Brazil to Be Used in Civil Construction

: Natural stones (limestones, granites, and marble) from mines located in northeastern Brazil were investigated to discover their potential for use in civil construction. The natural stones were characterized by chemical analysis, X-ray diffraction, differential thermal analysis, and optical microscopy. The physical-mechanical properties (apparent density, porosity, water absorption, compressive and ﬂexural strength, impact, and abrasion) and chemical resistance properties were also evaluated. The results of the physical-mechanical analysis indicated that the natural stones investigated have the potential to be used in different environments (interior, exterior), taking into account factors such as people’s circulation and exposure to chemical agents.


Introduction
Since ancient times, natural stones have been used as natural stone tiles, structural pieces in architectural works, and monuments. Natural rocks play an important role in the maintenance and repair of monuments and historic buildings with national or World Heritage significance [1,2]. It is estimated that, since 2013, the world natural stones industry has grown at a rate of approximately 4.0% per year [3,4]. Brazil stands out in the mining and commercializing of natural stones, reaching fourth position worldwide, and is responsible for 7% of the sector's production [5][6][7]. In Brazil, the mining of natural stones in the northeast region produces 1.63 million tons per year, representing about 20% of the national production; most mining occurs in the states of Bahia and Ceará [3].
Among Brazilian natural stones, those that stand out are marble, limestone, sandstone, quartzite, slate, gneiss, and granite; these are used extensively in civil construction. These natural stones can be used as dimensional stones, tiles, kitchen countertops, and internal flooring. The enormous interest in using these stones in the civil construction sector is due to their aesthetic properties, durability, and physical and mechanical characteristics [8,9]. However, these stones may show gradual and slow degradation under natural conditions. Acidic waters, environments with polluted air, and contact with chemical components can cause physical changes and modify the surface roughness, leading to granular disintegration and increased porosity [10,11]. Consequently, this affects mechanical behavior and some aesthetic properties, such as color or polished finish [12,13]. Thus, the study of the physical-mechanical, chemical and microstructural properties of these materials is essential

Materials
The natural stones used in this study are from mines located in Ceará state (Brazil) and were provided by companies. Table 1 summarizes the stone type, commercial name, identification used in this work, location of extraction, and supplier companies. The blocks of limestone, granite, and marble, approximately 300 mm × 300 mm × 300 mm, were received without failures, cracks, or alteration zones. Due to their lamellar characteristics, the limestone blocks from the Cariri region (LCC and LCA) presented an average thickness of 80 mm. The samples were cut with a diamond saw blade, and the surface was polished in an automatic polishing machine at a speed of 400 rpm. Abrasive stones of silicon carbide (numbers 400 and 600) were employed.

Characterization of Natural Stones
The chemical composition of the natural stones was determined by X-ray fluorescence spectroscopy (model EDX 720, Shimadzu, Kyoto, Japan). The mineralogical phases were identified by X-ray diffraction (model XRD 6000, Shimadzu, Kyoto, Japan), with CuKα (40 kV/30 mA), 2θ interval between 5 • -60 • and 0.02 • step. The diffraction peaks observed were indexed using the Search Match ® Program and the JCPDS database. The thermal behavior was evaluated by differential thermal analysis (model RB-3000, Instrumentec BP, Campinas, Brazil), where the samples were heated to 12.5 • C.min −1 from room temperature to 1000 • C. All DTA experiments were conducted under a compressed-air atmosphere.
The morphology and petrographic features of the natural stones were investigated by an optical stereo microscope. The images were acquired using an microscope (Zeizz, Axiotech) coupled to a digital camera (ColorView Soft Imaging System-II) and analySIS ® software (Version 2.3). This technique was used to identify the minerals and to characterize their microstructure.

Physical-Mechanical Properties
The physical properties of the natural stones (apparent density-AD, apparent porosity-AP, and water absorption-WA) were determined following ASTM C-97/C97M [32]. For this, 12 samples with dimensions of 50 mm × 50 mm × 50 mm were used. The experiments to determine the thermal expansion coefficient (CTE) were recorded using a dilatometer RB-3000 by BP Engineering. These tests were performed on three cylindrical samples (15 × 50 mm in length), previously dried in an oven at 60 • C for 48 h, under identical experimental conditions. The samples were heated at 5 • C.min −1 from 25 • C to 300 • C. In order to stabilize the equipment, it was kept at 30 • C for 30 min. The CTE values were calculated in 2 temperature ranges: 30 • C to 50 • C and 40 • C to 100 • C.
The mechanical compression strength experiments were accomplished in a universal testing machine (INSTRON EMIC, model DL 1000). Five samples with dimensions of 70 mm × 70 mm × 70 mm were used, and all experiments were performed with a 200 ton load cell and a test speed of 1.0 mm.min −1 following the ASTM C-170/C170M standard [33]. The rupture modulus was determined from the 3-point flexural strength test on a universal testing machine (EMIC, DL 1000), with a distance between the support points of 180 mm and a test speed of 1.0 mm.min −1 . The experiments were carried out on five samples with dimensions of 200 mm × 100 mm × 50 mm following ASTM C-99/C99M and C-880/C880M [34,35].
The impact resistance was determined from the hard body impact test. The experiment consisted of letting a steel ball with a mass of 1 kg fall on the test pieces with dimensions of 200 mm × 200 mm × 30 mm. The steel ball was released at an initial height of 0.2 m. The initial height was increased by 0.05 m until the rupture of the specimen. For each experimental condition, five samples were used. The result was obtained using the average drop height for which there was a sample rupture. The abrasion resistance values were determined by the abrasion wear test on an Amsler abrasimeter (Contenco, I-4212). Five samples with dimensions of 70 mm × 70 mm × 40 mm were used. Dry sand was used as an abrasive at a flow of 72 cm 3 min −1 . After 500 and 1000 turns, the wear was measured according to the ASTM C-241/C241M standard [36].

Resistance to Chemical Attack
Resistance to chemical attack analysis was carried out to assess the behavior of natural stones in the presence of chemical substances present in daily use. Five chemical reagents were used: ammonium chloride (NH 4 Cl), sodium hypochlorite (NaClO), citric acid (C 6 H 8 O 7 ), hydrochloric acid (HCl), and potassium hydroxide (KOH). The tests were performed according to the ISO 10545-13 standard [37]. Table 2 shows the concentrations of each reagent and the exposure time of the samples during the experiment. Six samples of each natural stone with dimensions of 100 mm × 100 mm × 20 mm were used for the experimental conditions of the chemical resistance test ( Table 2). All samples were polished, weighed, and analyzed for surface brightness with the Gloss Meter (Horiba, Gloss Checker 16.310). Then, one of the six samples was separated to be used as a blank. In the other five samples, PVC tubes with a diameter of 75 mm were fixed on the surface, and 50 mL of the reagents was introduced. The tubes were sealed with PVC films to avoid evaporation of the reagents. After a specific exposure time (Table 2), the PVC tubes were removed. The exposed surfaces were washed with running water, dried naturally, and subjected to new brightness measures (average of 10 readings). After exposure to the different substances, the samples were weighed again to assess the loss of mass.

Chemical Composition, Mineralogical Phases, and Thermal Behavior of Natural Stones
The chemical analysis, mineralogical phases, and the differential thermal analysis (DTA) curves measured from the limestones investigated in this study are listed in Table 3, Figure 1, and Figure 2, respectively. CaO and MgO were the major components detected in the limestones LCA, LCB, LBC, and LBM. This result is in accordance with X-ray diffraction (XRD) patterns, where the mineralogical phases calcite (JCPDS 47-1743) and dolomite (JCPDS: 89-5862) were identified in these samples (see Figure 1). The LCC, LMC, and LBSM samples presented only the calcite phase (JCPDS . This result agrees with the chemical analysis, since MgO contents were not identified in these samples (see Table 3). Based on the MgO content, the LCC, LMC, and LBSM samples were classified as calcitic limestones (MgO < 1.1%); the LCA, LBC, and LCB samples as magnesian limestones (1.1% < MgO < 2.1%); and the LBM sample as dolomitic limestone (2.1% < MgO < 10.8%) [38]. The presence of low levels of SiO 2 , Al 2 O 3 , Fe 2 O 3 , TiO 2 , and SO 3 detected on the limestones is probably related to impurities such as clays, feldspars, micas, quartz, and sulfides, among others [39].
In some DTA curves of the limestones (Figure 2), more precisely in the LCC, LMC, and LBSM samples, an endothermic peak was observed at approximately 950 • C. This thermal event is related to the decomposition of calcite (CaCO 3 → CaO + CO 2 ). In the LCA, LCB, LBC, and LBM samples, two endothermic peaks were observed; the first peak (~800 • C) is probably associated with the decomposition of magnesium carbonate (MgCO 3 → MgO + CO 2 ) (see Figure 1) and the second peak (~950 • C) is associated with calcite decomposition.    Table 4 and Figures 3 and 4 show the chemical analysis, the mineralogical phases, and the DTA curves measured from the granite samples. Quartz (JCPDS 46-045), feldspars, and mica (JCPDS 83-1808) were the main mineralogical phases identified in the GBC, GRF, GRD, GGG, GVV, GJPT, and GBSP samples (see Figure 3). In general, the granite matrix consisted of the feldspars orthoclase (potassium feldspar, JCPDS 31-0966) and plagioclase (calcium-sodium feldspar, JCPDS 41-1486). These results agree with the chemical composition, which indicated high levels of K 2 O, Na 2 O, and CaO oxides in these materials. Fe 2 O 3 and SO 3 contents were also detected, which indicates the probable presence of sulfides. The reddish and pink color presented by some granites was probably associated with hematite inclusions in the feldspar [40]. The presence of feldspar is significant for the mechanical behavior of granites, because in these materials, the resistance is closely associated with the content of feldspar, quartz, and mica and their granulometry [41,42]. As the quartz content increases, the mechanical strength of granite tends to increase. However, the high content of feldspars causes a reduction in mechanical properties due to these being easily cleavable minerals that have a tendency to suffer weathering. The presence of mica also compromises the strength of the stones because of the low hardness presented by this mineral (Mohs hardness 2-4) [43,44]. A small amount of kaolinite (JCPDS 78-2110) was identified in the GGG and GRD samples as a secondary mineral. In granite, the presence of kaolinite usually occurs as an impurity but can also be formed from the wear of feldspars if the removal of cations occurs intensely [45].   In the DTA curves of the granite samples ( Figure 4), it was possible to identify the presence of two endothermic events (~200 • C and~573 • C, respectively). The first event is probably associated with the elimination of free water, and the second with the polymorphic transformation of quartz (α → ß) [7]. The GGG sample had a small endothermic band at 600 • C and an endothermic peak around 780 • C. These events are possibly related to the decomposition of kaolinite and calcite present in this material, respectively.
The chemical analysis of the marble investigated in this study is listed in Table 5. The mineralogical phases and DTA curve are shown in Figure 5a-b. It was possible to observe that dolomite (JCPDS: 89-5862) was the predominant mineralogical phase (Figure 5a). Traces of the calcite phase were also identified (JCPDS 47-1743). These results are in accordance with the chemical analysis (Table 5), in which the dolomitic limestone character was confirmed by the high content of MgO detected. The presence of SiO 2 may be related to impurities such as quartz and feldspars. High levels of SiO 2 (above 7%) are essential to give the material the compatible mechanical strength for use as natural stone tiles for civil construction. Two endothermic events were also observed in the MBA sample; the first thermal event (~800 • C) is associated with the decomposition of magnesium carbonate, and the second (~950 • C) with the decomposition of calcium carbonate [6].  It has a massive structure with a microcrystalline bioclastic texture. It also has scattered fossil structures with elongated shapes of approximately 0.5 cm, whose composition is 99% carbonates (formed by very fine calcium compounds). It has a predominantly light to dark beige color, with yellow tones. LCA is similar to LCC. It presents a whitish-gray color when it dries, and when wet, it becomes bluish gray. When dispersed, it presents beige lines, with orientation in the form of waves, which likely indicates the conditions of origin. LCB has a massive structure with fine to medium monocrystals. It has a microcrystalline texture, with a composition of 95% carbonates in the form of concentric aggregates of inorganic origin, with very fine CaO compounds predominating and 3% MgO. Quartz is present with a 3% content, in the form of sub-rounded clasts with an average size of 0.7 cm. LMC and LBC show similar characteristics to LCB, which indicates the same mineral formation. The three stones are light beige and have dark pigmentation with elongated shapes of up to 0.5 cm, originating from different fossils. LBM and LBSM have similar structures to LBC. LBM is composed of magnesian carbonates (above 10%) and is calcitic (greater than 80%). The quartz content approaches 4%. It has a light beige color, with dark pigmentation originating from different fossils. LBSM has 95% calcium carbonates and approximately 2% quartz. It has a whitish-gray color, with variegated pigments also originating from other fossils.
The presence of bioclast in the limestones can significantly influence the physical and mechanical behavior of this rock. Studies [46] observed a bioclast contribution to limestone rock porosity, with micropores embedded in the bioclasts. Despite a relatively minor contribution to the total porosity, these micropores can negatively influence the mechanical strength of the material. Studies [47,48] also observed that bioclasts could contribute with macropores (50-300 µm) and with micropores (5-10 µm) in other calcareous porous rocks, and that under stress, the development of the conjugate shear bands were accompanied by damage in the form of Hertzian fractures emanating from bioclast contacts, with grain crushing and pore collapse. Naeem et al. [49] observed that the strength of the rock increased with increased calcite content but linearly decreased with a rise in the number of bioclasts. Figure 7 shows optical stereomicrographs obtained from the surface of granites (GBC, GRF, GRD, GGG, GVV, GJPT, and GBSP). In general, the crystals observed in the granite samples were larger than 500 µm. However, in GBC, GBSP, and GGG samples, small crystals (<200 µm) were immersed in the matrix. These samples present medium to coarse and isotropic granulation, with an atypical granular phaneritic texture, characterized by prismatic quartz crystals with trapezoidal terminations immersed in a feldspar matrix containing Muscovites. This differs from other granites in that it is hololeucocratic and white with porphyry of feldspar and quartz. GRF is a coarse, equigranular granite with a sieno-granitic composition. It shows no signs of deformation and is essentially composed of potassium feldspar, quartz, and biotite. Potassium feldspar occurs in the form of crystals that easily reach 1 cm, exhibiting various shades of green. Quartz occurs in the form of sub-rounded to hexagonal crystals and varies from 0.2 to 0.5 cm. When quartz is in contact with K-feldspars, it acquires a pseudo-green color. Biotite occurs in lamellar clusters and may exceed 0.5 cm in thickness. GRD is classified as biotite-magnetite granite. It presents an intense red color, is medium to coarse-grained, and is composed predominantly of mega crystals of potassium feldspar (approximately 53%) and subordinately by smaller crystals of quartz (18%), plagioclase (12%), and biotite lamellas (5%). GGG is a coarse granite that presents an intense fracturing on the entire surface, mainly in the quartz crystals. In feldspar crystals, both potassium and plagioclase, the microcracks and interstices on the surface hide the distinct characteristics of these two formations.
GVV is a sieno-granitic rock. It has a greenish to slightly yellowish color and medium to coarse grain. It consists essentially of quartz, feldspar, biotite, and amphibole. Potassium feldspar occurs in the form of crystals that easily reach 1 cm, exhibiting various shades of green. Quartz occurs in the form of sub-rounded to hexagonal crystals and varies from 0.2 to 0.5 cm. The biotite, reddish-brown in color, occurs in lamellar clusters and may exceed 0.5 cm in thickness. GJPT is a sieno-granitic rock with an ocher brown color. The color is likely due to weathering (oxidation and hydration of minerals containing ferrous iron). They have reddish-brown grain edges that accentuate the color of the rock. It has a milky white color and a fine to medium granular texture and shows some medium-grained crystals. It is essentially formed by feldspars, quartz, and gray-colored micas. Table 6 lists the results of the analysis of the physical properties (apparent density (AD), apparent porosity (AP), and water absorption (WA)), mechanical properties (compressive strength (CS), flexural strength (FS), impact resistance (IHB), and abrasion wear test (AW)) and the coefficients of thermal expansion measures in the limestone, granite, and marble samples. Table 7 contains a comparative and qualitative analysis of these results, classifying the investigated natural stones as low, regular, medium, good, or excellent for use as natural stone tiles in civil construction.

Physical-Mechanical Properties
All limestones studied showed AD values between 2.25 g/cm 3 and 2.52 g/cm 3 . According to ASTM C568/C568M [50], the limestones for application in civil construction can be classified into three categories: low density (1.76 g/cm 3 to 2.16 g/cm 3 ), medium density (2.16 g/cm 3 to 2.56 g/cm 3 ), and high density (>2.56 g/cm 3 ). Therefore, in agreement with ASTM C568/C568M, the limestones investigated in this study are classified as medium density. Among the limestones, the LCA stood out for presenting the lowest values of porosity (1.88%) and water absorption (0.75%). On the other hand, LMC showed the highest values, with 13.74% and 6.11% apparent porosity and water absorption, respectively. Vigroux et al. [51] evaluated six types of limestone from different regions of France and found porosity values of 11.2-25%. Studies by Eslami et al. [52] report limestone with porosity values of 10-35%. According to some studies [53][54][55], limestones for application as natural stone tiles in civil construction must have less than 4% apparent porosity. However, the applicability of limestones as natural stone tiles depends on the type and location of application. Thus, there are reports of materials used as natural stone tiles with a porosity range of 10% to 45% [56][57][58][59].
The water absorption of the studied granites was in the range of 0.1-0.3%. These values are compatible with those found in commercial granites (WA between 0.1 and 0.8%) [30]. In general, the results of the physical properties of granite and marble exceeded the average values suggested by ASTM C615/C615M [60]. All granites presented AP values <1%, WA values < 0.4%, and an AD (dry) > 2.56 g/cm 3 . The GGG sample presented better experimental results (AD = 2.69 g/cm 3 , AP = 0.35%, and WA = 0.14%), while the GRD presented values closer to the limit established by the standard (AD = 2.60 g/cm 3 , AP = 0.81%, and WA = 0.31%). Granites are also suitable for application in conditions with exposure to the environment since they have less than 3% porosity [21,44,61,62].
Comparing the WA results of limestone, granite, and marble with the criteria established by ISO 13006 [63], it is noted that granite and marble would be comparable, in terms of WA, with porcelain tiles (WA ≤ 0.5%). In comparison, LCC and LCB samples have WA equivalent to that of stoneware (0.5% < WA < 3.0%), and LCA, LMC, LBC, LBM, and LBSM samples have WA equivalent to semi-stoneware (3.0% < WA < 6.0%). This shows that, although limestones have high porosities for natural stones, they present water absorption values that allow use as natural stone tiles [63].
Regarding the thermal expansion coefficients, the limestones present values in the range of 4.5-6.0 × 10 −6 • C −1 (30 • C-50 • C). Harvey [64] determined the expansion coefficient of 39 various limestone specimens and observed that the expansion coefficient varied from 1.9 to 6.1 × 10 −6 • C in the range from room temperature to 100 • C. The determination of the thermal expansion coefficient of natural stones is of great importance for civil construction applications, as the rocks undergo volume variation (expansion or contraction) when subjected to temperature fluctuations. Thus, this coefficient is used for calculations in the dimensioning of expansion joints for slabs, panels, and tiles. Ideally, for this application, the material has the lowest thermal coefficients to avoid the appearance of cracks during the expansion and contraction process, which would reduce the mechanical strength of the material.
For granites, the range of variation of the expansion coefficient is wider (7.7-11 × 10 −6 • C −1 ); this is probably related to differences in microstructural characteristics, such as quartz content, feldspar content (constitutes the matrix), and presence of other phases. It is known that the degree of thermal expansion in natural stones depends mainly on their mineralogical composition (especially the content of quartz and calcite) and orientation of the crystal. Each mineral has different thermal expansion values, and some are anisotropic. For example, quartz has the highest thermal expansion coefficient (α) perpendicular to the c axis (α = 13 × 10 −6 • C −1 ) and the smallest parallel to it (7.7 × 10 −6 • C −1 ) [65]. Granites with high plagioclase content generally have low α values because this mineral has minimal volume expansion. Also, the thermal expansion depends on factors such as porosity, grain size, and temperature. The expansion coefficient decreases with increasing porosity and increases with increasing temperature, grain size, or quartz content [66,67]. The thermal expansion coefficient for commercial granites ranges from 5.0 to 10.28 × 10 −6 • C −1 [68].  Table 7. Comparative and qualitative analysis of the physical-mechanical properties of limestones, granites, and marble. In general, granites have a microstructure composed of crystalline grains (quartz and mica) immersed in a feldspar matrix (see Figure 7), which gives, in addition to low porosity, a structure of grains and matrix interrelated with each other by strong primary connections. Among the granites, the GJPT sample was the one with the lowest compressive strength, probably due to the presence of micro-cracks as a result of its high thermal expansion coefficient (11 ×10 −6 • C −1 ).

Properties
The limestones presented laminar characteristics and porosity between 6.7% and 13.7%. The LCA sample showed promising results in strength compression and flexural strength-(30.2 ± 4.8) MPa and (24.7 ± 2.3) MPa, respectively-compared to other limestones. This behavior is also directly related to the low porosity (1.88% ± 0.10%) presented by this sample [69,70], which was the smallest among the studied limestones. The compressive strength of the limestones in this study varies between 16.5 and 34.3 MPa. LMC was the sample that showed the lowest CS value, with (16.5 ± 6.5) MPa, while LBM and LBSM exhibited the highest CS values, with (34.3 ± 3.0) MPa and (34.1 ± 3.9) MPa, respectively. These values are similar to those found in studies by [51,52]. Granites have compressive strength values between 95.2 and 221.6 MPa. These values are within the range observed in commercial granites (CS varies between 60 and 292 MPa) [71]. Considering ASTM C170 and ASTM C99 [33,34], all limestones showed compressive and flexion strength values above the limit established for limestone with average density (CS ≥ 28 MPa and FS ≥ 3. The impact strength value obtained for granites ranges between 0.53 m and 0.75 m. As seen in Figure 7, the granite microstructure is formed by a glassy matrix of feldspar with quartz crystals. However, it is common for these granite microstructures to present microcracks around quartz crystals and small cracks resulting from differences in the coefficients of thermal expansion between the glass and crystalline phases. Thus, the impact strength is compromised, as the microstructure cannot absorb the impact's energy and the cracks formed end up spreading quickly. In limestones, the high porosity facilitates crack formation since the pores act as stress concentrators. However, these pores also act as barriers to the spread of cracks; the character of these laminar materials provides a series of plans for propagating the crack, which consumes the impact energy and minimizes damage to the material as a whole. Resistance to abrasion wear is closely related to factors such as mineral hardness, weathering degree, and porosity. As the porosity increases, the functional area to withstand abrasive stress is smaller. Among the limestones, the LCA sample obtained the lowest porosity (1.88%) and consequently the best abrasion resistance (1.09 mm). For granites, GBSP (AP = 0.30%) was the one with the highest abrasion resistance (0.11 mm). The ASTM standard does not define values for Amsler wear in limestones, granites, and marbles. According to the NBR 15.844 standard [72], the maximum wear limit for a 1000 m course must be ≤1.0 mm for the natural stone to be considered good quality. In all limestones, wear was greater than 1 mm. Their application should be avoided on floors, especially those intended for high traffic of people or heavy equipment. All granite and marble had values below the minimum limit established by the aforementioned standard; thus, these stones can be used on floors in areas with heavy traffic. The best performance was observed for GBC (0.51 mm), GGG (0.48 mm), and GBSP (0.43 mm), as these can maintain the polished surface layer over time, even when subject to high friction loads.

Resistance to Chemical Attack
The analysis of resistance to chemical attack is essential since several substances can compromise the useful life of the parts. In general, the limestone samples suffered the greatest changes in color and texture after exposure to chemical reagents. This behavior is possibly related to the high porosity values (1.88-13.74%) compared to granites and marbles. The laminar structure also favors the infiltration of substances that affect mechanical and aesthetic aspects. In addition, carbonates (the main constituents of limestones) are sensitive to attack by acidic solutions [73].
The magnitude of the acid attack varied significantly according to the sample and the reagent used. Citric and hydrochloric acids were the reagents that most affected the surface of limestones and granites. Figures 8 and 9 show the limestone and granite sample images before and after exposure to citric acid (C 6 H 8 O 7 ) and hydrochloric acid (HCl). Feldspars are sensitive to hydrochloric acid, which explains the more pronounced changes observed in granites when exposed to this acid. Table 8 shows the brightness values measured in the standard samples (without acid attack) and the samples exposed to the reagents. It is noted that the limestones showed an initial brightness significantly lower than that of granites; this was expected due to the high porosity of limestone.
When limestones and granites were exposed to ammonium chloride, most samples showed a loss of brightness of 5% to 20%. No significant change was observed in the color or occurrence of mineral wear (corrosion). In marble, this loss of shine was more significant, around 27%. However, the LMC and LBSM samples exhibited a 100% loss in brightness when exposed to all reagents, except for potassium hydroxide. This is probably related to the high content of (CaO) presented by these samples (see Table 3). Sodium hypochlorite, which simulates bleaches and disinfectants, caused a loss of brightness in the limestones between 13% and 40%. In granites, sodium hypochlorite caused a low loss of gloss (between 2% to 7%), and no changes in color or surface corrosion were observed.
On the other hand, the marble showed slight corrosion and a loss of 3% in the gloss, which was not enough to modify its color. Citric acid caused deterioration in all limestones, in some granites (GRF, GGG, GVV, and GJPT), and in marble. The GBC, GRD, and GBSP samples showed no change in color, and the loss of brightness was in the range of 5% to 7%. Except for GBC and GRD samples, all-natural stones showed corrosion when exposed to hydrochloric acid; however, no color changes were observed. On the other hand, potassium hydroxide did not cause a color change in the pieces, reducing the brightness by less than 10% for granites. Figure 10 presents the qualitative analysis of mass losses of the natural stones after exposure to chemical agents. All granites did not show mass loss when exposed to NH 4 Cl and NaClO. The same was not observed for the limestone LMC, LCB, and LCC samples, which were the most affected for these reagents. HCl and KOH most frequently caused a mass loss in the samples, proving that these severely attack natural stones.
The mass losses resulting from exposure to KOH confirm that granites and limestones are attacked by this alkalizing agent, despite the small change in brightness produced by KOH in granites. Generally, alkalis attack silicates present in granite stones, which justifies their action on granites and the loss of mass observed. The mass loss of limestones and marble observed after exposure to KOH is probably related to an alkali-carbonate reaction, where alkalis, such as Na and K, under certain pH conditions, attack carbonates forming hydroxides, which recrystallize, causing tensions and material damage (including loss of mass) [73]. This reaction occurs more intensely in dolomites or dolomitic limestones.

Conclusions
The results obtained from chemical analyses, XDR, and DTA carried out on the selected samples indicated that the studied limestones can be classified according to the MgO content as calcitic limestone (LCC, LMC, and LBSM), magnesian limestone (LCA, LBC, and LCB), and dolomitic limestone (LBM). The investigated granites (GBC, GRF, GRD, GGG, GVV, GJPT, and GBSP) are constituted of feldspars (potassium feldspar and calcium-sodium feldspar) quartz, and mica. The GGG and GRD samples also showed a small amount of kaolinite, which may be present as an impurity or feldspar wear product. The reddish color presented by granites GRD and GJPT is probably associated with hematite inclusions in the feldspar matrix. The high MgO content detected in the MBA sample indicates that the investigated marble can be classified as dolomitic calcareous.
The analyzed limestones had high porosity (1.88-13.74%) and suffered the greatest number of changes in color and texture after exposure to chemical reagents. After analyzing the physical properties, mechanical properties, and thermal expansion coefficient, it was possible to infer that the studied limestones can be used as natural stone tiles in low-traffic indoor environments without exposure to strong acidic or basic solutions. Among the studied limestones, LCA had the best mechanical and water absorption properties. On the other hand, the studied marbles and granites can be used as natural stone tiles for indoor and outdoor environments, even in high-traffic areas. They can also be used as tables, benches, and sinks as well as in other similar applications. However, care must be taken when exposing these stones to acids and alkalis, especially hydrochloric acid, since feldspar (the main constituent of granites) is sensitive to this acid. For application in outdoor conditions, GBC, GGG, and GBSP stones stand out for presenting the best results in water absorption and mechanical resistance.