A novel TiO2–SiO2 nanocomposite converts a very friable stone into a self-cleaning building material

Applied Surface Science 275 (2013) 389–396 Contents lists available at SciVerse ScienceDirect Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc A novel TiO2 –SiO2 nanocomposite converts a very friable stone into a self-cleaning building material Luís Pinho, Farid Elhaddad, Dario S. Facio, Maria J. Mosquera ∗ Departamento de Química-Física, Facultad de Ciencias, Campus Universitario Río San Pedro, Universidad de Cádiz, 11510 Puerto Real, Cádiz, Spain a r t i c l e i n f o Article history: Received 2 October 2012 Accepted 23 October 2012 Available online 5 November 2012 Keywords: Stone Non-ionic surfactant TiO2 –SiO2 nanocomposite Self-cleaning agent Consolidant Salt-resistant product a b s t r a c t A TiO2 –SiO2 nanocomposite material was formed inside the pore structure of a very friable carbonate stone by simple spraying of a sol containing silica oligomers, titania particles and a non-ionic surfactant (n-octylamine). The resulting nanomaterial provides an effective adhesive and crack-free surface layer to the stone, and gives it self-cleaning properties. In addition, it efficiently penetrates into the pores of the stone, significantly improving its mechanical resistance, and is thus capable of converting an extremely friable stone into a building material with self-cleaning properties. Another important advantage of the nanocomposite is that it substantially improves protection against salt crystallization degradation mechanisms. In the trial described, the untreated stone is reduced to a completely powdered material after 3 cycles of NaSO4 crystallization degradation, whereas stone treated with this novel product remains practically unaltered after 30 cycles. For comparison purposes, two commercial products (with consolidant and photocatalytic properties) were also tested, and both alternative materials produced coatings that crack and provide less mechanical resistance to the stone than this product. These results also confirm the valuable role played by n-octylamine in reducing the capillary pressure responsible for consolidant cracking, and in promoting silica polymerization inside the pores of the non-reactive pure carbonate stone. © 2012 Elsevier B.V. All rights reserved. 1. Introduction Natural stone of diverse types is employed as construction material around the world, for reasons of esthetic appeal and elegance but mainly for its durability. Demand for natural stone is, therefore, usually limited to the more durable varieties, such as granites, marbles and some sandstones. Another type of stone, pure carbonates, presents an exceptionally bright white color, which is much appreciated by consumers as a building material for floors, walls and external facades. However, this natural rock has low mechanical resistance and is easily stained, thus inhibiting it commercial application. Therefore, the development of a treatment product specifically intended to enhance the robustness and durability of carbonate stone, and with self-cleaning properties, should be of considerable interest for architecture and construction. Since the early discovery of the self-cleaning properties of titanium dioxide [1], it has been considered to be the most efficient, stable and cheap photocatalytic material available [2,3]. In recent years, the application TiO2 to very widely different substrates, such as textiles [4–7], plastics [8–12] and glasses [13–16], has been ∗ Corresponding author. Tel.: +34 956016331; fax: +34 956016471. E-mail address: mariajesus.mosquera@uca.es (M.J. Mosquera). 0169-4332/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.apsusc.2012.10.142 widely reported. However, its application to various types of stone has been much more limited [17–21]. It is commonly employed as an aqueous dispersion of titania particles [17,19,20]. The results obtained for these products on stone are not wholly satisfactory, for two reasons: (1) a cracked coating is formed on stone [17], and (2) the titania is easily removed from the stone surface [18]. Nowadays, most commercial products applied for the protection of stonework and other building materials contain alkoxysilane monomers or oligomers [22]. These species polymerize, in situ, inside the pore structure of the stone, by a classical sol–gel process; this improves properties of the product such as its mechanical resistance or its hydrophobic behavior. Two main reactions take place during sol–gel transition: (1) hydrolysis of alkoxy groups to create silanols; (2) polymerization by condensation of silanol groups from the products. In addition, condensation also occurs between silanol groups from the products and those present in the siliceous mineral surface of the stone. The advantages of these products, widely reported in the literature, are: (1) the low viscosity of the monomer or oligomer facilities penetration deeper into the pore structure of the stone; (2) environmental moisture is sufficient to produce the hydrolysis process; and (3) stable siloxane polymers are created, of a composition similar to the siliceous minerals of the stone. However, a well-known drawback of these products, which is characteristic of all the sol–gel materials, is their tendency to crack 390 L. Pinho et al. / Applied Surface Science 275 (2013) 389–396 during their drying process [23]. It is obvious that a cracked substance cannot protect the treated stone very effectively. In earlier work, our research group has developed surfactant-synthesized nanomaterials specifically designed for protecting and restoring various types of stone and other building materials [24–29]. The inclusion of the surfactant provides an efficient means of preventing cracking of the gel by reducing the capillary pressure; this effect is the result of two factors: (1) a coarsening of the gel network; and (2) decreased solvent surface tension. Adopting to this strategy, we have prepared consolidants [24–28], hydrophobic products [25–28] and stain-resistant materials [28]. In a recent paper [29], we describe the addition of titanium dioxide particles to a silica oligomer, in the presence of the surfactant, in order to produce a self-cleaning product for stonework and other building materials. In this paper, we report the application of a sol containing titania particles and silica oligomer, to an extremely low-compaction and friable dolostone with an exceptionally bright white color. This dolostone is very prone to severe disaggregation, and can easily break into several pieces on routine manipulation. Consequently, this esthetically attractive stone is totally unusable as a building material. The aim of the work described is to convert the dolostone into a self-cleaning building material with adequate mechanical resistance. If achieved, this would make the dolostone a significant new product in the market for building stone. Specifically, the present study is focused on evaluating the effectiveness on the dolostone under study for: (1) providing selfcleaning properties; (2) improving mechanical resistance; and (3) adherence to the substrate. We also investigate the durability of the stones treated, by applying a standard crystallization test. The effectiveness shown by this nanocomposite product as a consolidant and self-cleaning product has also been compared, in additional tests, with a commercial photocatalytic product and a commercial consolidant which were also evaluated on the same dolostone. Another well-known and serious drawback of existing commercial siloxane products [22–30] is their very limited effectiveness as a consolidant of pure carbonate stones not containing siliceous minerals; there are two reasons for this deficiency: (1) carbonates slows the sol–gel transition [31]; and (2) carbonate salts do not have active OH groups on their surface which could react with alkoxysilanes included in the products [32]. Hence chemical bonds between the stone and the consolidant product are not created. However, the significant increase in mechanical resistance observed in the dolostone under study after application of the nanocomposite synthesized in our laboratory demonstrates the useful contribution made by n-octylamine in the interaction between a non-siliceous stone and this siloxane product, which is also discussed in the present paper. 2. Experimental The titania–silica nanocomposite was prepared from a starting sol containing TES40 WN (Wacker Chemie AG, GmbH), P25 particles (Evonik AEROXIDE® TiO2 P25) and n-octylamine (Aldrich). According to the technical data sheet, TES 40 WN (hereafter TES40) is a mixture of monomeric and oligomeric ethoxysilanes. The average chain length is approximately 5 Si O units. P25 has an average primary particle size of 21 nm and a specific surface area (BET) of 50 ± 15 m2 g−1 . The Sol was prepared by mixing TES40 with P25 particles in the presence of n-octylamine under ultrasonic agitation (125 W cm−3 ) for 10 min. The proportion of n-octylamine and P25 to TES40 was 0.36% v/v and 2% w/v, respectively. The formulation has been designated as UCATiO2 via the procedure devised at the University of Cadiz (the number indicates the % w/v of P25 included in the material). After synthesis, the product was applied to the dolostone samples under study. For comparison, a popular commercial consolidant, Tegovakon V100 (hereafter TV100), supplied by Evonik, has also been applied. TV100 is a solvent-free one-component consolidant consisting of partially pre-polymerized TEOS and dioctyltin dilaurate (DOTL) catalyst. A commercial photocatalytic coating, E503 supplied by Nanocer, was also applied. According to the material data sheet, E503 is a TiO2 -containing water-based sol with 7500–10,000 ppm of the oxide. Prior to their application to the stone, the rheological properties of the products were studied, using a concentric cylinder viscosimeter (model DV-II+ with UL/Y adapter) from Brookfield. Experiments were performed at a constant temperature of 25 ◦ C maintained by recirculated water from a thermostatic bath. A shear stress versus shear rate flow curve was generated.The stone selected is a very friable dolostone composed of magnesium carbonate (99%). It presents evident disaggregation problems due to its geological formation and/or natural aging processes. This stone was selected particularly because of its high degree of whiteness, which makes it a suitable candidate for self-cleaning treatment application. For all the experiments carried out, the stone samples were cut in the form of 5 × 5 × 2 cm slabs. The sols under study were applied by spraying onto the upper surface of the samples, in 5 periods of 5 s, during a total time of 25 s. The stone samples were then dried under laboratory conditions until reaching constant weight. Uptake of products and their corresponding dry matter by the stone samples was calculated. The samples corresponding to untreated stone and their treated counterparts were characterized by the procedures described below, after constant weight was reached. All the results reported correspond to average values obtained from three stone samples. A JEOL Quanta 200 scanning electron microscope (SEM) was used to visualize changes in the morphology of the stones after coating. Surface fragments of treated stone specimens and their untreated counterpart were visualized. The chemical bonds in the treated samples under study were analyzed by Fourier transform infrared spectrophotometry (FTIR). The spectra were recorded in powder using a FTIR-8400S from Shimadzu (4 cm−1 resolution) in the region from 4000 to 650 cm−1 . Experiments were carried out in attenuated total reflection mode (ATR). FTIR spectra of powdered fragments of untreated and treated samples were obtained. The adherence of the coating to the stone surface was evaluated by performing a peeling test using Scotch® MagicTM tape (3M). The test was carried out according to previously reported methods [29,33,34]. The changes in stone surface morphology were visualized by SEM working in low-vacuum mode, and energy-dispersive X-ray spectroscopy (EDX) spectra were recorded in order to elucidate the variations in surface composition after the test. The improvement in mechanical properties in treated stone was evaluated means of the Standard procedure Vickers hardness test, using a Universal Centaur RB-2/200 hardness tester. The loading was 30 kg during 30 s, with a preload time of 15 s. 10 measurements were made for each stone specimen. Vickers hardness (VH) was calculated according the following equation: VH = 1.8544 · W d2 (1) where W is the load over the surface area of the indentation; d is the indentation diagonal. The improvement in mechanical properties was also measured using the drilling resistance measuring system (DRMS), by SINT Technology. Drill bits of 4.8 mm diameter were employed with a rotation speed of 600 rpm and penetration rate of 5 mm/min. For each specimen, 10 holes were carried out. L. Pinho et al. / Applied Surface Science 275 (2013) 389–396 391 Table 1 Properties of the treated stone specimens and their untreated counterpart. Product Untreated TV100 E503 UCATiO2 Uptake (%w/w) Dry matter (%w/w) Material removed by peeling (mg) Vickers hardness (kP/mm2 ) E* – – 34.5 ± 14.9 49.33 ± 15.21 – 0.53 ± 0.01 0.30 ± 0.04 1.9 ± 1.1 54.17 ± 3.37 16.00 ± 3.39 0.40 ± 0.06 0.12 ± 0.11 15.9 ± 1.4 – 1.05 ± 0.79 0.44 0.30 2.3 62.72 6.96 ± ± ± ± ± 0.20 0.15 0.1 5.32 0.69 Data correspond to average values. Standard deviations are also included. We also evaluated the possible disadvantage of this material associated with changes in stone color induced by the treatment. This effect was determined using a solid reflection spectrophotometer, Colorflex model, from Hunterlab. For each specimen, color changes were tested on five points on the surface. The conditions used were: illuminant C and observer 10◦ . CIELa*b* color space was used and variations in color were evaluated using the parameter: total color difference (E*) [35]. The effectiveness of the materials under study as self-cleaning coatings for stone surfaces was evaluated by using a test adapted from the literature [29,36,37]. First, 1 ml of a solution of 1 mM methylene blue (Panreac) in ethanol was deposited on treated stone specimens and their untreated counterparts. Next, stone samples were irradiated with UV light working at 365 nm in a Vilber Lourmat CN15.CL chamber with 2 tubes of 15 W. The distance between the samples and the tubes was approximately 20 cm. Color variations, recorded as a function of irradiation time, were determined using the same procedure described above. The parameter: total color difference (E*) was again evaluated. In order to evaluate the durability of the products, an aging test by sodium sulphate crystallization was carried out. In this case, the sols were applied to all the 6 faces of the specimen. This test was performed according to the standard UNE-EN 12370 [38]. Immersion of specimens in the salt solution was substituted by capillary rise in order to make the test more aggressive. Each cycle consisted of three steps: (1) partial immersion in salt solution during 2 h; (2) drying at 100 ◦ C during 10 h; and (3) cooling of the samples under laboratory conditions. After 30 cycles, the samples were immersed in distilled water during 24 h to remove the salts deposited inside the pores and next they were washed with normal water. Samples were then dried at 100 ◦ C during another 24 h, and the weight lost was determined. 3. Results and discussion A rheological study of the sols was carried out, and viscosity data were obtained immediately after stirring the dispersions. E503, TV100 and UCATiO2 sols showed Newtonian behavior at the shear range evaluated. The viscosity was calculated as the slope of the shear rate versus shear stress curve. In all the cases, the linear regression coefficients were above 0.99. The value obtained for the UCATiO2 sol was only slightly higher (5.82 mPa s) than that Fig. 1. Scanning electron microscopy micrographs of the coatings on the dolostone under study. L. Pinho et al. / Applied Surface Science 275 (2013) 389–396 392 Si-OH, Si-N CO 32- CO 32- transmitance (a.u.) UNTREATED E503 TV100 UCATiO2 CH3,CH2 CH2 C-N 0 3000 1500 Si-O-Si 1250 Si-O-Si 1000 750 wavenumber (cm-1) Fig. 2. FTIR spectra for the samples under study. corresponding to the commercial TV100 sol (5.25 mPa s). This result suggests that UCA product should penetrate into a stone substrate to a depth similar to that corresponding to the commercial consolidant. Penetration is a key factor for achieving a suitable consolidant effect on the stone. E503 shows a much lower value (0.282 mPa s) because it is mainly constituted by water. The product uptake and dry matter values obtained after application of the products to the dolostone samples are shown in Table 1. For TV100 and UCA products these values are similar; however, E503 shows a significantly lower dry matter value because its solvent content is very high. Fig. 1 shows SEM micrographs of stone specimens treated with the products under study. TV100 creates a dense and vitreous coating on the stone surface. The microporous nature of this product, characterized using nitrogen physisorption in our previous papers [24–27], is responsible for the formation of this coating. E503, in turn, seems to create a thin and easy detachable surface layer. Both commercial coatings present fractures. In the case of the UCATiO2 material, a crack-free, homogeneous and coarse coating on the stone surface is observed. The formation of this crack-free coating has previously been explained as a consequence of the role played by the surfactant n-octylamine, which reduces the capillary pressure while the gel is drying, as explained in the Section 1 [24–29]. FTIR spectra of the treated stone samples and their untreated counterpart are shown in Fig. 2. All the samples present the carbonate peaks, which are characteristic of all carbonate stones including the dolostone used in this work. Specifically, the peaks at 1420, 880 and 730 cm−1 correspond to asymmetric, out-of-plane and inplane bending vibration modes of the anion CO3 2− , respectively [39]. The spectra corresponding to the untreated dolostone and the stone sample treated with E503 do not present any additional peak. We relate the similarity between treated and untreated dolostone spectra with the low dry matter value obtained for E503. In the case of the dolostone samples treated with TV100 and UCATiO2, we observe some additional peaks in the spectra. In particular, there are two peaks corresponding to typical siloxane vibrations, located at 800 and 1070 cm−1 , corresponding to bending and stretching vibrations, respectively [40]. This confirms the presence of silica gels in the surface of the two treated stones. However, the intensity of these two peaks is different, being significantly lower in the stone treated with the commercial product. This confirms the poor effect of commercial siloxane products on pure carbonate stones reported in the literature [22]. The higher intensity of Si O Si bonds observed in the stone treated with our product can be attributed to the role played by n-octylamine, which may accelerate the condensation process in the carbonate stone [28,29]. This hypothesis is also confirmed by other additional Si O Si peak presented by the stone treated with the UCA product. Specifically, this sample showed a peak adjacent to the stretching vibration, located at 1161 cm−1 . Demjén et al. [32] associated this double band (1070–1161 cm−1 ) to the formation of highmolecular weight siloxane chains with a ladder-type structure. We also found this double band in silica xerogels prepared in presence of n-octylamine [41]. In the case of the TV100 product, this additional band was not observed and subsequently, a cyclic, lowmolecular weight polymeric siloxane structure could have been formed, according to these authors. These authors found this double band in CaCO3 treated with amino functional silanes. They concluded that the primary amine group promotes the formation of high- molecular weight siloxane polymers due to its catalytic effect on polycondensation. The peak at 970 cm−1 shown in the spectra of the UCA product deserves particular attention. It can be attributed, a priori, to Si OH stretching [42]. However, the absence of a broad band at 3750–3250 cm−1 , associated with hydrogen-bonded silanol groups with absorbed molecular water, suggests that silanol groups are not present in this coating. Consequently, we think that the 970 cm−1 peak could be attributed to Si N stretching vibration, which is located in the same region [41,43]. In addition, Han et al. [44] confirmed that Si N interaction occurred between methylamine and zeolite. According to these authors, the strong hydrogen bonding interaction results in the H atom of the amine group attacking the Si O framework to form Si O· · ·H N bond, which leads to the formation of Si N bonds in the zeolites. Similarly, we think hydrogen bonding created between n-octylamine and the silica oligomer could also generate silica nitration. In the stone treated with the UCA product, we also observed a band related to the asymmetric vibration of the CH2 group at 2974 cm−1 and to the symmetric vibration of the CH3 and CH2 groups at 2928 and 2852 cm−1 , respectively. All of these are associated with alkyl chains, according the literature [32]. These peaks could be associated with the surfactant residues or even with ethoxy groups from non-hydrolyzed oligomers. In the spectrum of this stone sample, we also observe a band at 1300 cm−1 , corresponding to CH2 twist vibration, which could be also attributed to ethoxy groups [45]. The spectrum of the stone treated with TV100 did not show these bands. Finally, the slight peak at 1365 could be attributed to amine C N stretching [24]. A noticeable feature is the absence of a peak associated with titania in the spectrum of the UCA coating. It is often reported that the bands observed in the range 900–1000 cm−1 may be associated with Ti OH and Si O Ti species [46]. Thus, these peaks could appear in our spectrum due to co-condensation between silica oligomer and titania particles. However, they are not visible because these bands are probably obscured by the peak attributed to Si OH and Si N located at the same wavenumber [29,43]. Since one significant drawback of commercial products applied on stone has been associated with a reduction in photocatalytic efficiency during long-term use, due to the elimination of TiO2 from the stone surface [18,29], we have investigated the degree of adhesion of the coatings applied on stone, by performing a peeling test, adapted from the literature. Fig. 3 shows the micrographs obtained by SEM and the EDX analyses carried out on tested and non-tested areas. Table 1 shows the weight lost by the untreated stone and its treated counterparts after testing. On comparing titanium content present in the surface of the stones, as expected, the greatest content is observed for the stone L. Pinho et al. / Applied Surface Science 275 (2013) 389–396 393 drilling resistance force (N) 40 UNTREATED TV100 UCATiO2 30 20 10 0 0 5 10 15 20 25 30 penetration depth (mm) Fig. 4. Drilling resistance force measurements for the treated dolostone under study. Fig. 3. Scanning electron microscopy micrographs of the dolostone under study after the peeling test. On the left side of each micrograph, the non-tested surface is presented, together with the corresponding EDX spectrum. On the right side, the corresponding tested surface and spectrum are presented. sample treated with UCATiO2 (Fig. 3). E503 is a water-based product with extremely low titanium dioxide content (0.75–1%). This may explain why the titanium peak is very low for the stone treated with this commercial product. After complete drying TV100 is a silicon dioxide gel and therefore, presents no EDX peaks for titanium, and only silicon peaks are present in its spectrum. Regarding to the SEM micrographs obtained, we observe clear differences between samples due to the products applied. The stone treated with TV100 shows a dense coating, typical of a microporous material. The removal of a very small amount of material from the sample surface is observed after peeling, with some stone mineral being observed in the micrograph. This is confirmed by observing a slight decrease of Si in the EDX analysis after the peeling test. Concerning the mass of the amount removed for TV100, only a slight loss of mass is observed (see Table 1). The micrograph for E503 shows a severely cracked, friable and easily detachable surface. In this case, a significant amount of material is removed from the stone surface (see Table 1). Since this product is largely eliminated from the stone surface, we consider that it has little potential use for stone with low cohesion, such as the dolostone under study. Concerning the changes in titanium content after peeling, no significant differences can be observed, probably because its content is too low to observe these changes. In the case of UCATiO2, the micrograph obtained does not show any changes in the stone surface. Moreover, the EDX results confirm that no significant surface mass removal occurred after the tests. In respect of the amount of mass removed, the data obtained show only a slight loss of mass (see Table 1), similar to the result obtained for TV100. These findings confirm that TiO2 has been integrated into the silica matrix, which has been capable of adhering firmly to the stone. Thus, we can conclude that the inclusion of the photocatalyst in a mesoporous silica coating is an interesting solution for keeping particles tightly adhered to the treated surface, providing long-term wear resistance. An important objective of this work is to improve the robustness of the stone under study, so that it can be utilized as a viable building material. Therefore, we have evaluated changes in two of the mechanical properties of the stone, i.e. hardness and drilling resistance, after application of each product. E503 was not considered in this evaluation because it is not a consolidant product. Changes in Vickers hardness of the stone surface after treatment are given in Table 1. We observed increases in the surface hardness after application of the two products under study, with this increase being significantly greater in the samples treated with UCATiO2. Drilling resistance results are shown in Fig. 4. TV100 increases the stone’s resistance in the surface zone (to a depth of 5 mm). However, after application of the UCA product, the drilling resistance is increased by a factor of 6 in the full depth evaluated of 30 mm. These results demonstrate that n-octylamine is playing a significant role in enhancing the consolidant effectiveness of siloxane, since silicon-based products are known to be ineffective as consolidants in pure carbonate stones [22,30]. This confirms the role played by the n-octylamine favoring the interaction between the siloxane and the apolar carbonate stone, as we previously observed for a pure limestone [29]. Earlier, Demjén et al. [32] had reported that amino-functional silanes adhere strongly to the surface of CaCO3 . They discussed this effect in the following terms: the primary amine group promotes adhesion due to its catalytic effect on siloxane polycondensation. We think the primary amine group of the n-octylamine could play a similar role. It is confirmed by the FTIR spectrum previously discussed, in which the siloxane peaks observed the for UCA product are of significantly higher intensity 394 L. Pinho et al. / Applied Surface Science 275 (2013) 389–396 Fig. 5. On the left side, the evolution of total color difference (for methylene blue stains) on treated and untreated samples of dolostone is presented. On the right side, we present photographs after the self-cleaning test of the surfaces of stone samples treated with the coatings under study. The samples were previously stained with methylene blue and then irradiated with UV light ( = 365 nm) for more than 800 h. (For interpretation of the references to color in text, the reader is referred to the web version of the article.) than the peaks corresponding to TV100 (without octylamine). It is also corroborated by the presence of a double Si O Si peak, which is assigned to the formation of high molecular weight polysiloxanes [32]. We also evaluated changes in color following the treatments applied to the dolostone studied. Total color difference values (E*) are shown in Table 1. Due to the low dry matter deposited on the stone surface, E503 produced low values of E*, below the perception threshold (E* < 3) [35]. UCATiO2 produced color changes near the generally accepted threshold value (E* ≤ 5), even for the most restrictive applications (ancient building restoration) [47]. In the case of the TV100 product, this produced an unacceptable color change (E* = 16) on the dolostone under study. We think the UCA product significantly inhibits color change in this extremely white carbonate stone because the titania particles have a notable whitening effect [29]. We have investigated the self-cleaning properties of the products on the dolostone tested by carrying out a photo-degradation test of stains previously deposited on the stone surface. Methylene blue (MB) was used as the staining agent. The evolution of total color differences under UV light over time was recorded and results are shown in Fig. 5 for the two coatings with self-cleaning properties under study. TV100 was not included in this test because it has no self-cleaning properties. The untreated stone showed a gradual reduction in the stain; after more than 800 h of exposure its final E* value was 19.07 (a reduction of 34%). The MB bleaching under visible/UV light has previously been reported, and is associated with the progressive absorption of light by the dye in the 350–520 nm range [48,49]. In the case of the stones treated with the two products with selfcleaning properties under study, we observe a similar behavior in MB degradation. Specifically, two different rates of degradation at different times can be distinguished clearly in the profiles. Very rapid MB bleaching occurs in the first 72 h, accounting for around 70% of the total color variation recorded. Next, a slower rate of degradation is observed over the longer term. The final E* value obtained was 4.23 for the stone treated with the UCA product, whereas stone treated with E503 showed a slightly higher final E* value (6.67). This variation between products may be associated with higher TiO2 content (2%) of the UCATiO2 product compared with that of E503 (0.75–1%). The two findings reported above clearly confirm that the photocatalytic action of the titania particles produces most of the total degradation effect on the stain in the first few hours of exposure (the first part of the curve). In the case of the UCATiO2 product, we think that the second stage with a much slower rate of degradation may be caused by the silica/titania coating reducing the capacity of the MB to penetrate into the limestone pore structure. This hypothesis is supported by the degradation profiles of pure P25 particles on a limestone, which we previously published [29]. That study revealed that it is possible to obtain faster MB degradation with pure TiO2 particles. P25 particles induced a single rate of Fig. 6. On the left side, the weight loss of the dolostone samples treated with TV100, UCATiO2 and their untreated counterpart during the salt crystallization test is represented. On the right side, the photographs show the treated stone samples and their untreated counterpart after completion of the salt crystallization test (untreated, 4 cycles; TV100, 15 cycles; UCATiO2, 30 cycles). L. Pinho et al. / Applied Surface Science 275 (2013) 389–396 degradation, with the stain being almost totally degraded in the first 6 h. Fig. 5 also shows photographs of the stone surfaces under study before and after completion of the test. Before the start of the test, it can be observed that the untreated stone surface shows an MB stain significantly similar to that of the surface treated with E503. This is probably due to the insufficient cohesion of E503, which allows the staining MB solution to penetrate deep in the stone, as it does in the untreated sample. In the case of UCATiO2, a more intense initial blue color is observed in the surface stone because UCATiO2 creates a crack-free, homogeneous coating preventing the MB penetration into the stone pore structure. We also consider that the self-cleaning effect produced by the UCATiO2 coating is partly induced by the coarse texture of its gel network. As Yamauchi et al. have reported [50], the photocatalytic degradation of MB is clearly enhanced by the addition of TiO2 particles to a silica mesoporous structure, in contrast to the effect observed when the particles are integrated in a dense microporous matrix. The explanation of this enhancement effect given by these authors is that mesopores accelerate the diffusion of MB toward the reaction sites (i.e. toward the titania particles). As discussed earlier in this paper, the UCATiO2 coatings also create a mesoporous coating on the stone. Finally, we subjected the stone samples to a salt crystallization test in order to evaluate the durability of the treatment. Fig. 6 shows the weight loss of the treated dolostone samples and its untreated counterparts during the test. On the right side, photographs are also shown of the dolostone treated with the products under study and their untreated counterpart. The ‘before and after’ of these photographs illustrate the condition before subjecting the samples to the salt crystallization test and to their condition when the test was finished for each sample (corresponding to the last point in the graphic presented in Fig. 6). As observed in the pictures, the untreated stone sample has completely disintegrated after the 4th cycle. Coincidently, TV100 has lost almost 50% of its mass by the 4th cycle. The samples treated with UCATiO2 remain almost unaltered until the 20th cycle, and even after the 30th cycle show a fairly good behavior. In the case of the untreated stone, the results obtained confirm that it is a very friable stone, totally unsuitable for use as a building material. Regarding the samples treated with TV100, we think that the poor mechanical properties achieved for the stones treated (see Fig. 4) are responsible for its limited durability. As previously discussed, the low mechanical resistance could be due to the combined effect of the following factors: (1) extensive cracking that occurs for TV100 coatings (Fig. 1); (2) poor interaction between a siloxanebased product and carbonate stones; this was confirmed in the present paper by the FTIR spectra obtained; and (3) the presence of micropores in a stone generates high sodium sulfate crystallization pressures, which give rise to significant damage to the porous substrate [51]. Therefore, the microporosity of TV100 could also increase its susceptibility to sodium sulfate crystallization damage. In the case of UCATiO2, a homogeneous, crack-free coating that adheres well to the dolostone is created with greater in-depth mechanical resistance. This explains why UCATiO2 gives very good durability on the dolostone substrate. Again, we think n-octylamine is making a significant contribution to enhancing the durability, since it not only prevents cracking but also facilitates the interaction with the carbonate stone, as discussed extensively in previous paragraphs. The TiO2 particles could also increase the durability of the treatment when subjected to crystallization salts, as Miliani et al. reported for silica consolidants modified with TiO2 particles applied on a sandstone [47]. As a final remark, it may be of interest that the dolostone under study, protected with the products synthesized in 395 our laboratory, is going to be commercialized under an exploitation patent [52]. 4. Conclusions We have transformed an extremely weakly compacted and friable dolostone, currently unsuitable for application in building, into a suitable building material by means of a simple and low-cost procedure. This consists specifically of spraying onto the stone surface a newly developed sol. This sol contains silica oligomers, titania particles and n-octylamine, and by this treatment, a TiO2 –SiO2 nanocomposite consolidant material is formed inside the pore structure of the stone. We have demonstrated that this nanomaterial is capable of: (1) creating a crack-free coating, which adheres firmly to the stone surface and ensures that the conservation and self-cleaning properties of the coating have a long-term effect; (2) significantly increasing the mechanical resistance of the stone; (3) providing proven self-cleaning properties to the stone; and (4) making the stone suitably durable to degradation by salt crystallization. Moreover, we have demonstrated that two commercial products tested (a consolidant and a self-cleaning agent) produce coatings that crack and generate less mechanical resistance for the stone than our product. These results obtained in this paper confirm the valuable role played by n-octylamine in: (1) creating a mesoporous structure which reduces the capillary pressure which is responsible for cracking and increases the self-cleaning properties of the coating; and (2) promoting silica polymerization in the pores of the non-reactive pure carbonate stone. Acknowledgments We are grateful for financial support from the Spanish Government/FEDER-EU (Project MAT2010-16206 and Project Regenera (Innpacto subprogram), and the Government of Andalusia (project TEP-6386 and Group TEP-243). We also thank the company Tino Stone S.A. for financial support under a research contract. L.P. thanks the Fundação Ciência e Tecnologia for his predoctoral grant (SFRH/BD/43492/2008). References [1] A. Fujishima, K. Honda, Nature 238 (1972) 37–38. [2] A. Fujishima, T. Rao, D. Tryk, Journal of Photochemistry and Photobiology C 1 (2000) 1–21. [3] K. Hashimoto, H. Irie, A. Fujishima, AAPPS Bulletin 17 (2007) 12–28. [4] A. Bozzi, T. Yuranova, J. Kiwi, Journal of Photochemistry and Photobiology A 172 (2005) 27–34. [5] K. Qi, W. Daoud, J. Xin, C. Mak, W. Tang, W. Cheung, Journal of Materials Chemistry 16 (2006) 4567–4574. [6] M. Uddin, F. Cesano, D. Scarano, F. Bonino, G. Agostini, G. Spoto, S. Bordiga, A. Zecchina, Journal of Photochemistry and Photobiology A 199 (2008) 64–72. [7] D. Wu, M. Long, ACS Applied Materials and Interfaces 3 (2011) 4770–4774. [8] K. Tennakone, C. Tilakaratne, I. Kottegoda, Journal of Photochemistry and Photobiology A 87 (1995) 177–179. [9] S. Lam, A. Soetanto, R. Amal, Journal of Nanoparticle Research 11 (2009) 1971–1979. [10] R. Fatteh, A. Ismail, R. Dillert, D. Bahnemann, Journal of Physical Chemistry C 115 (2011) 10405–10411. [11] K. Nakata, M. Sakai, O. Tsuyoshi, T. Murakami, K. Takagi, A. Fujishima, Langmuir 27 (2011) 3275–3278. [12] H. Chen, C. Liang, H. Huang, J. Chen, R. Vittal, C. Lin, K. Wu, K. Ho, Chemical Communications 47 (2011) 8346–8348. [13] Y. Paz, Z. Luo, L. Rabenberg, A. Heller, Journal of Materials Research 11 (1995) 2842–2848. [14] T. Watanabe, Thin Solid Films 351 (1999) 260–263. [15] J. Chen, C.S. Poon, Building and Environment 44 (2009) 1899–1906. [16] B. Xi, L.K. Verma, J. Li, C.S. Bathia, A.J. Danner, H. Yang, H.C. Zeng, ACS Applied Materials & Interfaces 4 (2012) 1093–1102. [17] I. Poulios, P. Spathis, A. Grigoriadou, K. Delidou, P. Tsoumparis, Journal of Environment Science and Health, Part A Environmental Science 34 (1999) 1455–1471. 396 L. Pinho et al. / Applied Surface Science 275 (2013) 389–396 [18] K. Rao, M. Subrahmanyam, P. Boule, Applied Catalysis B: Environmental 49 (2004) 239–249. [19] E. Quagliarini, F. Bondioli, G. Goffredo, A. Licciuli, P. Munafò, Journal of Cultural Heritage 13 (2012) 204–209. [20] E. Quagliarini, F. Bondioli, G. Goffredo, C. Cordoni, P. Munafò, Construction and Building Materials 37 (2012) 51–57. [21] M.F. La Russa, S.A. Ruffolo, N. Rovella, C.M. Belfiore, A.M. Palermo, M.T. Guzzi, G.M. Crisci, Progress in Organic Coatings 74 (2012) 186–191. [22] G. Wheeler, Alkoxysilanes and the Consolidation of Stone, The Getty Conservation Institute, Los Angeles, 2005. [23] G.W. Scherer, Journal of the American Ceramic Society 73 (1990) 3–14. [24] M.J. Mosquera, D.M. de los Santos, A. Montes, L. Valdez-Castro, Langmuir 24 (2008) 2772–2778. [25] M.J. Mosquera, D.M. de los Santos, L. Valdez-Castro, L. Esquivias, Journal of Non-Crystalline Solids 354 (2008) 645–650. [26] M.J. Mosquera, D.M. de los Santos, T. Rivas, P. Sanmartín, B. Silva, Journal of Nano Research 8 (2009) 1–12. [27] M.J. Mosquera, D.M. de los Santos, T. Rivas, Langmuir 26 (2010) 6737–6745. [28] J.F. Illescas, M.J. Mosquera, Journal of Physical Chemistry C 115 (2011) 14624–14634. [29] L. Pinho, M.J. Mosquera, Journal of Physical Chemistry C 115 (2011) 22851–22862. [30] A.P. Ferreira Pinto, J. Delgado Rodrigues, Journal of Cultural Heritage 9 (2008) 38–53. [31] C. Danehey, G. Wheeler, S.H. Su, The influence of quartz and calcite on polymerization of methyltrimethoxysilane, in: Proceedings of the 17th International Congress on Deterioration and Conservation of Stone, Lisbon, 1992, pp. 1043–1052. [32] Z. Demjén, B. Purkánszky, E. Földes, J. Nagy, Journal of Colloids & Interface Science 190 (1997) 427–436. [33] L. Ling, I. Phang, G. Vancso, J. Huskens, D. Reinhoudt, Langmuir 25 (2009) 3260–3263. [34] M. Drdácký, J. Lesák, S. Rescic, Z. Slízková, P. Tiano, J. Valach, Materials & Structures 45 (2012) 505–520. [35] R.S. Berns, Billmeyer and Saltzman’s Principles of Color Technology, WileyInterscience, New York, 2000. [36] T. Tatsuma, S. Tachibana, M. Tetsuya, D. Tryk, A. Fujishima, Journal of Physical Chemistry B 103 (1999) 8033–8035. [37] ISO. 10678 Fine ceramics (advanced ceramics, advanced technical ceramics) Determination of photocatalytic acitivity of surfaces in an aqueous medium by degradation of methylene blue. ISO, 2010. [38] UNE-EN 1999, Natural Stone Test Methods. Determination of Resistance to Salt Crystallization, AENOR, Madrid, Spain, 1999. [39] A. Sdiri, T. Higashi, T. Hatta, F. Jamoussi, N. Tase, Environmental Earth Sciences 61 (2010) 1275–1287. [40] L. Tellez, J. Rubio, F. Rubio, E. Morales, J.L. Oteo, Journal of Materials Science 38 (2003) 1773–1780. [41] J.F. Illescas, M.J. Mosquera, ACS Applied Materials & Interfaces 4 (2012) 4259–4269. [42] G. Orcel, J. Phalippou, L.L. Hench, Journal of Non-Crystalline Solids 88 (1986) 114–130. [43] M. Sekine, S. Katayama, M. Mitomo, Journal of Non-Crystalline Solids 134 (1991) 199–207. [44] A.J. Han, J. Guo, H. Yu, Y. Zeng, Y.F. Huang, H.Y. He, Y.C. Long, ChemPhysChem 7 (2006) 607–613. [45] P. Innocenzi, Journal of Non-Crystalline Solids 316 (2003) 309–319. [46] C.M. Whang, S.C. Yeo, Y.H. Kim, Bulletin of the Korean Chemical Society 22 (2001) 1366–1370. [47] C. Miliani, M.L. Velo-Simpson, G.W. Scherer, Journal of Cultutral Heritage 8 (2007) 1–6. [48] A. Mills, J. Wang, Journal of Photochemistry and Photobiology A 127 (1999) 123–134. [49] M. Mrowetz, W. Balcerski, A.J. Colussi, M.R. Hoffmann, Journal of Physical Chemistry B 108 (2004) 17269–17273. [50] N. Suzuki, X. Jiang, L. Radhakrishnan, K. Takai, K. Shimasaki, Y. Huang, Y. Yamauchi, Bulletin of the Chemical Society of Japan 84 (2011) 812–817. [51] C. Rodríguez-Navarro, E. Doehne, E. Sebastián, Cement and Concrete Research 30 (2000) 1527–1534. [52] M.J. Mosquera, L. Pinho, Spanish Patent No. P201100741. (June 24, 2011).