*3.3. Water Vapor Permeability*

The results of the water vapor permeability test, as expressed by the water vapor diffusion resistance coefficient (μ), show a μ decrease for all hydrophobic treatments on limestone and mortar specimens, reducing the breathability of the substrates (Table 7). The reduction in water vapor permeability is an inevitable consequence of the water repellence properties of polymer film; however, the lowest possible decrease is pursued [43].



This reduction is more significant in limestone specimens, whereas it is almost negligible in mortar specimens. In fact, an increase of μ of 227% (HSila/Sil), 129% \*(HSil) and 74% (HNST) is observed on treated limestone specimens, when compared to untreated ones, whereas a minimal μ increase (<4%) is observed on the treated mortar specimens.

After artificial aging, only the HNST treatment applied on limestone maintains reasonably higher μ values (increase of 7%) when compared to untreated specimens, whereas HSil and HSila/Sil treatments still induce a drastic μ increase (46% and 188%, respectively). In the case of mortar specimens, the higher increase of μ was observed with HSila/Sil treatment (9%), and only a slight μ decrease in the case of HSil and HNST treatments (<2%). In general, the hydrophobic treatment that illustrates the most suitable behavior to water vapor permeability was HNST, with a moderate reduction of the μ on both substrates, even after artificial aging.

#### **4. Discussion**

The variation of the moisture transport properties of the substrates treated with the hydrophobic products is presented in Figure 4. In general, it can be observed that the hydrophobic products induce a decrease of the capillary water absorption coefficient (C), this decrease being more relevant with HNST treatment on limestone specimens. However, after artificial aging, HSila/Sil and HSil treatments show a higher durability, with a higher decrease of the C, if compared to specimens treated with HNST. Concerning mortar specimens, all treatments maintain a similar water absorption coefficient (considerably lower than untreated specimens), even after artificial aging.

**Figure 4.** Percentage variation of (**a**) capillary water absorption coefficient (C), (**b**) drying index (DI) and (**c**) diffusion resistance of the water vapor coefficient (μ) of treated specimens, before and after artificial aging, when compared to untreated specimens.

The different moisture transport properties of mortar and limestone can be attributed to the higher open porosity (30%) of the mortar when compared to the studied limestone (11%). Additionally, mortar specimens generally have a higher volume of coarse pores (>100 μm) when compared to this type of compact Moleanos limestone [40].

Regarding the difference in the effectiveness and durability of the hydrophobic products, it is worth noting that (monomer) silane molecules are considerably smaller (10 to 15 Å) when

compared to (oligomeric) siloxane molecules (25 to 75 Å) [44]. Thus, the longer Si–O chains of the siloxanes compared to the silane ones have a lower penetration depth in the compact limestone. Additionally, organo-modified siloxanes have an extremely high reactivity, which can hinder their in-depth penetration [2]. The silane has a potentially deeper penetration in the treated surface, with a higher reduction of the hydrophilicity and, thus, improved hydrophobic effectiveness. Furthermore, concerning the silane, it is generally assumed that the larger the molecule of the alkyl group linked to the silicon atom (which constitutes the structure of this compound), the higher the water repellency of the silane [8,45]. On the other hand, the long siloxane chains are more affected by environmental agents and weathering, undergoing degradation processes which can reduce their effectiveness as hydrophobic products [14].

The optimal performance obtained with the nanostructured product based on SiO2 and TiO2 (HNST) on the limestone can probably be attributed to the chain arrangement of the TiO2-SiO2 nanoparticles, due to the creation of the Si–O–Ti bond [23]. In fact, the copolymerization of the TiO2 and silane within the silica network can give rise to the formation of homogeneous organic–inorganic hybrid xerogel [24]. Additionally, the nanosize (<0.1 μm) of the silicon titanium oxide particles can match the dimension of pore network of the limestone. On the other hand, the low durability of HNST treatment can be attribute to the photocatalytic oxidation (of the organic radicals) and thermal degradation of the SiO2-TiO2 composite, which can weaken the adhesion and thus, durability of the coating [38].

Concerning the drying index, the hydrophobic products induced an increase of the DI on limestone specimens, with higher variation in the case of the HSil and HSila/Sil treatments. After aging, HSila/Sil significantly increase the DI of the treated limestone specimens, whereas HSil and HNST maintain values similar to unaged specimens. Greater variations are observed in mortar specimens; in fact, the HNST treatment induces an increase of the DI, and a drastic decrease is observed after artificial aging, indicating the probable degradation of the HNST treatment. On the other hand, the HSila/Sil and HSil treatments decrease the DI on mortar specimens, with a lower decrease of the latter after artificial aging compared with the HNST treatment. As reported in other works [46], the silane/siloxanes products can modify the pore netwok of the treated material by increasing the volume of capillary pores, probably also due to an air entraining effect of liquid siloxane. This feature can justify the improved resistance to freeze-thaw cycles, and thus, durability of HSila/Sil and HSil treatments compared to HNST.

Furthermore, the difference in the moisture transport properties of treated limestone and mortar can be attributed also to the higher roughness of the mortar, compared to flatter surface of the limestone, which accounts for better adhesion of both fissured and crack free SiO2-TiO2 films to the substrate [37].

All treatments induce an increase of the μ in limestone specimens. More specifically, specimens treated with HSila/Sil and HSil show the highest μ increase, which significantly decreases after artificial aging in the case of the HNST treatment. The HNST treatment shows a lower μ after artificial aging compared with unaged specimens. Considering the mortar specimens, it can be concluded that μ variation is extremely low with all treatments, both before and after artificial aging. The HSila/Sil is the only treatment which induced a slight increase of the μ after artificial aging.

Ultimately, it would be expectable that a substrate with higher DI would have also a higher μ (i.e., lower WVP). For treated limestone specimens this trend is confirmed; however, specimens treated with HSila/Sil have a higher μ/DI ratio (even after artificial aging) compared to the other treatments. A different behavior is observed with treated mortar specimens: Aged and unaged specimens with HNST treatment show the trend mentioned above, whereas unaged HSil treatment and aged HSila/Sil treatment show a DI decrease and a μ increase.

#### **5. Conclusions**

In this paper, the effectiveness and durability of three commercially available hydrophobic products (a silicon and titanium dioxides-based nanostructured dispersion—HNST; a silane/oligomeric siloxane—HSila/Sil; and a siloxane—HSil) when applied to a Moleanos limestone and on a cement-based mortar, were analyzed. The alteration of the moisture transport properties (water absorption by capillarity and under low pressure, drying kinetics and water vapor permeability) of the treated substrates, prior to and after artificial aging tests, was evaluated.

Results show that the effectiveness and durability of the water-repellent treatment is influenced both by the type of hydrophobic product and by the treated substrate.

Although the products were applied at different concentrations, following the recommendations of the producers, all treatments induce a significant decrease of the values of the capillary water absorption and water absorption under low pressure on the mortar specimens. A lower decrease was observed in the limestone specimens. This difference is attributed to the higher open porosity of the mortar specimens compared to limestone specimens, thus allowing a deeper penetration of the hydrophobic products, which increases the water-repellency of the treated mortar. After artificial aging, all hydrophobic treatments show a significant durability to the type and duration of the artificial aging cycles considered in this work, maintaining a reasonably low water absorption in both mortar and limestone specimens, when compared to untreated specimens. The HNST treatment shows a slightly greater loss of efficacy in terms of water capillary absorption and water vapor permeability after artificial aging, mostly when applied on limestone. Therefore, it can be considered less durable.

The treatments induce also a variation of the drying index, which increases with all treatments on limestone specimens (even after artificial drying) and a general decrease on mortar specimens, except for the HNST treatment before aging, which slightly increases the DI. On the other hand, the HNST treatment shows a lower drying index after artificial aging.

HSila/Sil and HSil treatments significantly reduce the water vapor permeability of limestone specimens, whereas the HNST treatment induces a smaller decrease, with values similar to those of untreated specimens after artificial aging. The WVP of the treated mortar specimens was not significantly affected by the hydrophobic treatments, even after aging tests.

These observations confirm that the hydrophobic products are generally more effective and durable on mortar specimens, rather than on low-porosity limestone, as in the case of the studied Moleanos limestone.

When pondering the variation of all the moisture transport properties, the hydrophobic product based on siloxane (HSil) has the best performance on cement-based mortar; in fact, the molecular structure of siloxanes matches to the higher porosity of this substrate. On the other hand, although it has a lower durability compared to the other treatments, HNST has the best performance when applied on Moleanos limestone, with a significant decrease of the capillary water absorption and a low variation of the drying index and of the WVP. This behavior can result from the combination of the low porosity and micro-sized pores of the stone with the surface deposition of a nanostructured layer. Additionally, the presence of TiO2 confers antibacterial activity against the microorganism growth and pollutant absorption.

The HSila/Sil treatment significantly decreases its water-repellent properties. However, it hinders the drying process and the breathability of the substrates, even after artificial aging. Thus, based on the results of this study, its use is not recommended in either limestone or mortar.

Further tests (e.g., optimization of the protocol; artificial aging cycles with UV light, pollutants and/or biological colonization; FTIR analysis of the hydrophobic products; morphological analysis by SEM-EDS; contact angle measurements of the treated substrates, among others) are ongoing to correlate the effect of the physical-chemical aging on the effectiveness of hydrophobic products, and ultimately, to verify the durability to more prolonged weathering action (from lab to real natural scale tests) of the hydrophobic products.

**Author Contributions:** Conceptualization, G.B.; Methodology, I.F.-C. and R.V.; Investigation, C.E.; Writing—Original draft preparation, G.B.; Writing—Review and editing, I.F.-C. and R.V.; Supervision, I.F.-C.; Project administration, I.F.-C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Portuguese Foundation for Science and Technology (FCT), grant number PTDC/ECI-EGC/30681/2017 (WGB\_Shield – Shielding building' facades on cities revitalization. Triple resistance for water, graffiti and biocolonization of external thermal insulation systems).

**Acknowledgments:** The authors acknowledge the companies CIN, Saint-Gobain Weber and NanoPhos for the supply of the hydrophobic products.

**Conflicts of Interest:** The authors declare no conflict of interest.
