Producing superhydrophobic/oleophobic coatings on Cultural Heritage building materials

Preparation of superhydrophobic product

The reagents employed were: (1) Aerosil OX50 (hereinafter SiNPs), which consists of fumed silica NPs (primary size of 40 nm and surface area of 50 m²·g⁻¹) supplied by Evonik. (2) Dynasylan F8815 (hereinafter F8815) is a water-soluble fluoroalkylsilane. It has a fluoroalkyl chain, which provides hydrophobic and oleophobic properties, an aminoalkyl chain which confers solubility in water, and two hydroxyl groups for siloxane polymerization.

The superhydrophobic product was prepared by means of a simple method: (1) SiNPs were dispersed in deionized water (2% w/v), and the dispersion was ultrasonically agitated for 10 min. (2) 2% v/v of F8815 was added dropwise, and the product was magnetically stirred for 30 min. The product was always applied 1 h after its preparation in order to maintain reproducibility of the results. The product showed stability in closed vessel for 1 month. After this time, a slight precipitate, which can be re-dispersed by mechanical agitation, was observed.

Application of superhydrophobic product

The sol was applied on four different building materials: limestone, granite, concrete and wood. The samples size was 4×4×2 cm for granite and limestone, 3.5×3.5×3.5 cm for the concrete and 3.5×3.5×0.5 cm for the wood. The four materials under study are commonly employed in modern buildings. The selected limestone (commercially available as Capri limestone) has an open porosity around 9%, and it presents a homogeneous structure, composed of micritic matrix containing oolites and oncolites, having a 90% of calcite and a 10% of silicate minerals. The granite (white pearl granite) is a biotitic granite with an open porosity around 0.5%, and the main components are SiO₂ (64%) and Al₂O₃ (19%). The rest of components (Fe₂O₃, MgO, Na₂O and K₂O) are found in percentage lower than 10%. A type I white Portland cement and a calcareous sand (ɸ 1–6 mm) were employed to fabricate the concrete samples. The water:cement ratio and the cement:sand ratio (w/w) were 1:2 and 1:5 (w/w), respectively. The specimens were cured under water at 21±2°C for 28 days, according with standard UNE-EN 12390-2 [29], being their open porosity around 10%. Finally, the pine wood samples (hereinafter Wood) present an open porosity around 58% and it is composed of cellulose (50%), hemicellulose (25%) and lignin (25%). All the surfaces were used as received without any modification. The sol was applied by brushing on all the substrates, except wood. Specifically, three applications, with an elapsed time between applications of 30 s were performed. The wood samples were immersed into the sol for 5 min. After application, the samples were let overnight in laboratory conditions (20°C and 40% RH), to allow the sol–gel transition. Then, the samples were heated into an oven at 100°C, for 1 h, in order to remove the solvent (water). Once the treated samples were dried and cold, the following experiments were carried out on two replicates for each surface under study. In order to calculate the uptake of product, the samples were weighed before and immediately after being treated. After drying, the samples were re-weighed to calculate the dry matter. The penetration depth of the product inside the building materials under study was evaluated. Specifically, the samples were cut in order to obtain cross-sections. Then, water was dropped on the cross-sections, corresponding the non-wetted area to the penetration depth of the product.

Characterization of the treated surfaces and effectiveness

Static and dynamic CAs were evaluated by using OCA15 Plus video-measuring equipment, from DataPhysics Instruments, according to the following method: droplets of distilled water (5 μL) were applied with a syringe at five different points on each surface under study, and SCA values were evaluated. Then, 2.5 μL were added, obtaining the advancing CA (ACA) values. Finally, 2.5 μL were removed and the receding CA (RCA) values were measured. In addition, the oleophobic properties of the coatings were evaluated by placing 5 μL droplets of olive oil at five different points on each surface and measuring the SCA values.

As an additional experimental test of the water-repellent behavior, a water jet was projected onto each treated surface under study from a distance of 20 cm. The experiment was recorded by a digital camera (Sony Cyber-shot model DSCP200) at 30 frames/s.

The self-cleaning properties of the treated surfaces and their untreated counterparts were tested according to a previously described procedure [30]. Specifically, methylene blue powder was applied to the surfaces of the samples. Next, the samples were placed under a water column in order to simulate the effect of rain. The water column fell over the surface of the samples from a height of 50 cm, with a flow of 45 mL·s⁻¹. These experiments were registered with the previously described camera. In order to evaluate the self-cleaning effectiveness, the color of the samples was measured before and after the test by using a ColorFlex spectrophotometer from HunterLab. Illuminant D65 and observer CIE10° were selected as measurement parameters. The CIE L*a*b* scale was employed [31]. These experiments were registered with the previously described camera.

In order to evaluate the changes produced in the topography of the treated samples, atomic force microscopy (AFM) images of the surfaces were obtained by using an AFM Dulcinea model from Nanotec Electrónica, operating in tapping mode. The root mean square (RMS) roughness values were evaluated from 5 μm×5 μm images. Measurements were recorded in five points on the surface.

Finally, the color change caused by the application of the product was evaluated by means of a ColorFlex spectrophotometer from HunterLab. Illuminant D65 and observer CIE10° were selected as measurement parameters. The CIE L*a*b* scale was employed [31].

Evaluation of durability

The lasting properties of the treated samples were tested by performing two different tests. The adhesion between coating and substrate is an important issue that determines the lasting properties of the treatment, since a coating with poor adhesion will be easily removed due to environmental conditions (rain, wind). Thus, the adhesion of the coatings to the surface samples was evaluated according to a previous test developed of our group [28]. Specifically, an adhesive tape was applied on the treated surfaces and it was subsequently removed. Static and dynamic CAs for water were measured after 1, 2, 5, 10, 15 and 20 attachment-detachment cycles.

The resistance to water impact was evaluated by placing the treated samples under a water column in order to simulate the effect of rain. The water column fell over the surfaces of the samples from a height of 50 cm with a flow of 45 mL·s⁻¹. The total amount of water falling onto the treated surfaces was 2500 L·m⁻². After the test, the samples were dried in an oven at 60°C and the static and dynamic CA values for water were evaluated.

Application of superhydrophobic product

Table 1 shows the average uptake, dry-matter and penetration depth values for the applied samples.

As observed in Table 1, all the substrates showed product uptake values with a clear relation with their open porosity values. In the case of wood, this value was significantly higher than the other substrates due to the application method (immersion). After drying, all the samples showed a significant loss of weight due to the solvent evaporation. Regarding to penetration depth, the measured value was also related with the porosity of the substrates. In the case of wood, this value could not be measured because the product was applied by immersion.

Characterization of the treated surfaces and effectiveness

Figure 1 shows the SCA, ACA and RCA values for the treated samples under study (left), and images of 25 μL water droplets on the building materials surfaces under study are shown on the right. The untreated surfaces were hydrophilic, showing SCA values of 35°, 60° and 10° for limestone, granite and concrete samples, respectively. In the case of the wood samples, the water droplet was quickly absorbed and thus, the SCA could not be evaluated. All the treated surfaces showed superhydrophobic properties with SCA values of around 160°. Moreover, ACA and RCA presented very close values, the CAH value being around 3° in all cases. As previously discussed [11], [22], 40 nm-SiO₂ NPs are able to produce a surface roughness typical of a Cassie-Baxter regime, providing repellent properties [32], [33], [34]. In addition, the fluoroalkylsilane is able to decrease surface energy due to the action of fluoroalkyl chains [35], [36], [37]. The combination of both effects (roughness and low surface energy) gives rise to superhydrophobic properties. In order to demonstrate the water repellent properties of the treated surfaces, a video was recorded (see Supplementary Information, Video S1). The video shows a water jet impacting on the four treated surfaces. The water jet is completely repelled for all the building material surfaces evaluated.

Fig. 1:

Static (SCA), Advancing (ACA) and Receding (RCA) Contact Angle values obtained on the surface of the treated building material samples under study. In order to stablish a comparison, SCA of the untreated samples were also included. Photographs of water droplets on the treated building material surfaces evaluated are included: (a) limestone, (b) granite, (c) concrete, and (d) wood. The water droplets were dyed with methylene blue to have a better contrast.

Alfieri et al. [38] also employed fluoroalkylsilane in order to produce hydrophobic coatings, but they added TiO₂ NPs with a size of around 4 nm in order to confer photoactive properties. In this case, repellence was not evaluated, and the SCA values were lower than those obtained in the present work. This different behavior is associated with the size of NPs integrated into the fluoroalkylsilica gel. The 4 nm-size NPs (ten times less than the 40 nm-size NPs employed in the present synthesis) cannot produce the required roughness to promote a Cassie-Baxter regime and an SCA above 150°.

In our work, the superhydrophobic behavior observed suggests that a self-cleaning effect can be induced in these surfaces [19], [39]. Thus, a test was carried out as previously described in the experimental section. Results are shown in Table 2.

In the case of the untreated samples, the dye was dissolved by the water droplets, and the hydrophilic nature of the surfaces promoted the absorption of the solution by the pores of the different substrates. Thus, a blue stain was permanently maintained on the building materials under study. Thus, significant ΔE* values were obtained for all the untreated samples, as observed in Table 2. On the other hand, the dye was completely removed from the treated surfaces by the droplets rolling off. This behavior was confirmed by the colorimetric test, being all the ΔE* values lower than 0.6. As an example, a video showing this self-cleaning effect on the limestone samples is shown in the Supplementary Information (Video S2).

The addition of a fluorinated compound reduces the surface energy to values lower than those produced by organic components, giving rise to oleophobic properties. The surface tension of olive oil (32 mN·m⁻¹, at 25°C) [40] is significantly lower than that of water (72 mN·m⁻¹). Thus, it is necessary to use halogen compounds in order to decrease the coating surface energy below the surface energy in the olive oil-coating interphase [35], [37], [41], [42]. Therefore, the oleophobic behavior of the treated surfaces under study was evaluated by measuring the SCA values of olive oil droplets. Figure 2 shows the SCA values of oil droplets deposited on the untreated substrates and their treated counterparts. In addition, photographs of the olive oil droplets on the treated surfaces are included. The untreated surfaces showed oleophilic behavior with SCA values of 10, 15 and 21° for limestone, granite and concrete, respectively. In the case of wood, oil droplets were rapidly absorbed and, therefore, they could not be measured. Regarding to the treated substrates, all of them showed highly oleophobic properties, with SCA values above 100°. Photographs of the olive oil droplets on the treated surfaces, along with the SCA values, can be seen in Fig. 2. All the surfaces showed highly oleophobic properties, with SCA values above 100°.

Fig. 2:

SCA values of the oil droplets deposited on the untreated substrates and their treated counterparts. Photographs of the olive oil droplets deposited on the surfaces of the treated building material samples under study: (a) limestone, (b) granite, (c) concrete, and (d) wood.

Since the superhydrophobicity is a combination of decreasing surface energy and creating a roughness characteristic of a Cassie-Baxter regime, a study of the topography modification of the surfaces under study was carried out by AFM measurements. Figure 3 shows AFM results for the treated samples under study and their untreated counterparts.

Fig. 3:

3D AFM images for the untreated (top) and treated (middle) surfaces. (a) limestone, (b) granite, (c) concrete, (d) wood. The root mean square (RMS) roughness values are also included. 2D roughness profiles are included below.

All the untreated samples showed completely heterogeneous surfaces with a random roughness (high RMS values are observed). On the contrary, the treated substrates showed homogeneous surfaces composed by densely-packed, uniform, NPs aggregation. It confirms that SiO₂ NPs were integrated into the fluorinated silica gel, producing the uniform packing observed by AFM. All the surfaces showed a significant decrease in the RMS values after being treated because the coating is able to cover the irregularities of the surfaces excepting the granite samples. In this case RMS value is increased after the application of the coating. This can be due to the low porosity of granite (0.5%), which promotes the accumulation of product on the surface. This can be explained as a consequence of the low product penetration in granite (see Table 1).

Figure 3 (bottom) presents the 2D AFM roughness profiles. As previously discussed by Bhushan and Her [43], the space between peaks is important in order to trap air pockets, producing a Cassie-Baxter regime. The treated surfaces show, for all the subtrates, distance between peaks with fairly uniform values between 150 and 250 nm, whereas the untreated surfaces showed heterogeneous distances between peaks, bigger than 1000 nm. These results suggest that the water droplets will be less able to penetrate into the treated surfaces, hence the contact area between droplet and surface will be minimized. This promotes high contact angle values and repellence behavior, as previously suggested [11], [22], [43], [44]. In the case of the oleophobic behavior, a mushroom-like roughness is required to produce superoleophobicity [42], [45]. Thus, our product promotes high SCA, but it cannot produce oil repellence.

A practical limitation of the superhydrophobic products to be applied in Cultural Heritage will be encountered if a change in the color of the building material is produced (physical-chemical compatibility). Therefore, the variations in the colorimetric coordinates (ΔL*, Δa* and Δb*) of the substrates induced by the product were obtained, and total color difference (ΔE*) was calculated. The results are shown in Table 3 and photographs of the untreated building materials and their treated counterparts are shown in Fig. 4.

Fig. 4:

Photographs of untreated (left) and treated (right) substrates, together with the calculated values of total color difference. (a) limestone, (b) granite, (c) concrete, (d) wood.

As observed in Table 3, the application of the superhydrophobic product on the building materials under study caused negligible color modification, as can be demonstrated by ΔE* values lower than or equal to 3. According with Delgado-Rodrigues and Grossi [6], color variation lower than 3 are perfectly acceptable in Cultural Heritage field. Regarding to the changes produced in the individual color coordinates (ΔL*, Δa* and Δb*), it was observed that the values obtained for Δa* and Δb* were negligible. In the case of ΔL*, a slight increase to positive values was observed, which indicates a slight lighter than that observed for the untreated surfaces. This value was slightly higher in the case of granite due to their lower porosity that allows the accumulation of the product on the surface, as previously discussed.

Evaluation of durability

The durability of the superhydrophobic surfaces is a critical requirement for their application on building materials which are exposed to outdoor conditions. Thus, the lasting-effectiveness of the coatings is an important parameter to evaluate. Adhesion of the coatings to the substrates was evaluated by a tape test. Figure 5 shows evolution of the SCA and CAH values through attachment-detachment cycles for the building material surfaces under study.

Fig. 5:

Evolution of the SCA (solid lines) and CAH (dashed lines) values with the tape test for the surfaces under study.

The obtained results showed that all the treated surfaces presented long-lasting mechanical properties. The superhydrophobic behavior was maintained in all the building materials under study during the 20 attachment-detachment cycles tested. Only small fluctuations in the SCA and CAH values were observed, and they can be related to measurement errors.

Finally, a durability test associated with rain resistance was performed (see Fig. 6). The obtained results did not show significant changes in the SCA values of the treated samples. In the case of the CAH values, a slight increase can be observed in the surfaces, except for the granite samples. However, the hysteresis is always maintained below 5° (required value to maintain superhydrophobic properties). Thus, we can conclude that all the treated building materials present long-lasting performance against rain water, demonstrating its possible application for outdoor conditions.

Fig. 6:

SCA (left) and CAH (right) values of water droplets on the treated surfaces, before and after the rain test.

The long-lasting properties presented by our coating can be explained as a consequence of its effective grafting to the four building materials evaluated, produced by condensation reactions between silanol groups (Si–OH) present in the fluoroalkylsilane, which allow it to polymerize and, moreover, to condense with Si–OH present in the SiO₂ NPs, giving rise to a durable nano-particulate coating [11], [22]. These silanol groups can condense with the silanol groups of the quartz minerals present in high and low proportion in the granite and the limestone, respectively. The Si–OH can also condense with the silanol groups of the concrete, and with the hydroxyl groups (OH) of the cellulose of wood. In addition, the high penetration of the product into the pore structure of the substrates (see Table 1) contributes significantly to improve the durability of the superhydrophobic materials because in case of removing the superficial coating, the product into the building material can preserve the superhydrophobic properties. Additionally, the polymerization inside the pore structure can take place between substrate and product in the case of the building materials containing hydroxyl groups (granite, concrete and wood). In the case of the limestone, the simple presence of the product inside the substrate can improve the durability.

In addition, the high resistance to rain water demonstrate the high stability of the obtained Cassie-Baxter regime. On a Cassie-Baxter surface the water droplets are placed on the top of roughness peaks, and air-pockets are created between water and solid surface [4]. However, if a critical pressure is reached, the water droplets eventually touch the roughness valleys [46] and thus, the surface can be transformed to a Wenzel regime [47]. In this regime, the repellence properties are lost. In the specific resistance rain simulation (see Fig. 6) carried out in this work, repellence is maintained, confirming the high resistance of the coatings under study.