Influence of domestic and environmental weathering in the self-cleaning performance and durability of TiO2 photocatalytic coatings

(1)

Contents lists available atScienceDirect

Building and Environment

journal homepage:www.elsevier.com/locate/buildenv

In

ﬂuence of domestic and environmental weathering in the self-cleaning

performance and durability of TiO

2

photocatalytic coatings

Erika Iveth Cedillo-González

a,b,∗

, Virginia Barbieri

b

, Paolo Falcaro

c

, Leticia M. Torres-Martínez

d

,

Isaías Juárez-Ramírez

d

, Laura Villanova

e

, Monica Montecchi

b,f

, Luca Pasquali

b,f,g

,

Cristina Siligardi

b

a_{Universidad Autónoma de Nuevo León, Facultad de Ciencias Químicas, Departamento de Ingeniería Química, San Nicolás de los Garza, NL 66455, Mexico} b_{Dipartimento di Ingegneria}_{“Enzo Ferrari”, Università degli Studi di Modena e Reggio Emilia, Via P. Vivarelli 10/1, Modena 41125, Italy}

c_{Graz University of Technology, Institute of Physical and Theoretical Chemistry, Stremayrgasse 9/Z2, 8010 Graz, Austria}

d_{Universidad Autónoma de Nuevo León, Facultad de Ingeniería Civil, Departamento de Ecomateriales y Energía, Ciudad Universitaria, San Nicolás de los Garza, NL}

66455, Mexico

e_{Graz University of Technology, Institute of Computational Biotechnology, Petersgasse 14, 8010 Graz, Austria} f_{IOM-CNR, s.s. 14, Km. 163.5 in AREA Science Park, 34149 Basovizza, Trieste, Italy}

g_{Department of Physics, University of Johannesburg, PO Box 524, Auckland Park, 2006, South Africa}

A R T I C L E I N F O

Keywords: Mesoporous Dense TiO2release Weathering Microstructure Self-cleaning

A B S T R A C T

Weathering of photocatalytic TiO2coatings represents an important issue for the successful application of TiO2 -based self-cleaning materials. Photocatalytic efficiency of the as-prepared materials is crucial for commerciali-zation; however, changes in the coating performance due to weathering become a critical factor for practical applications. Moreover, chemical durability should be considered as weathering can promote the release of photocatalyst nanoparticles, which can pollute the environment and be hazardous for human health. In this study, two photocatalytic TiO2coatings with different microstructures (namely compact and mesoporous) were exposed to chemical treatments to simulate domestic and environmental weathering. Results show that dense TiO2coatings with a slow photocatalytic activity are suitable for domestic applications as minimum leaching of photoactive material was observed. Conversely, once exposed to chemical solutions commonly present in do-mestic environments, the initially highly active mesoporous TiO2coatings showed a dramatic drop of the self-cleaning performance and a significant release of nanoparticles in the surrounding environment. It is expected that the results reported here will be of particular relevance for the construction sector, as the manuscript discloses important knowledge for the development of TiO2-based self-cleaning materials once exposed to indoor or outdoor environments.

1. Introduction

Since the discovery of photo-splitting of water in a TiO2 anode photochemical cell in 1972 [1], the ﬁeld of photocatalysis has con-stantly grown. The BCC AVM069B report states that the total market for photocatalyst products is estimated to be valued≈ $2.9 billion by 2020 [2]. Anatase TiO2is the most studied photocatalyst because of its non-toxicity (it is chemically and biologically inert) and excellent energy conversion eﬃciency [3]. Furthermore, it is inexpensive and easy to obtain [3]. When TiO2is exposed to UV light, holes (h+) and excited electrons (e-) are generated. Holes are trapped by adsorbed hydroxyl groups and induce the formation of hydroxyl radicals (OH•) [4]; these chemical species play a fundamental role in the degradation of organics

[5–7]. Electrons react with molecular oxygen to form superoxide anion (O2−), which further reacts with H+ to produce HO2•, resulting in photo-oxidation reactions [8]. In the absence of adsorbed pollutants, hydrophilic surface Ti(III) species are generated via the reduction of Ti (IV) by the electrons, and oxygen vacancies are produced via the oxi-dation of the bridging O2−to oxygen by the holes. Subsequently, hy-droxyl ions can be preferentially adsorbed into the TiO2oxygen va-cancies, thus an increased hydrophilicity of the TiO2surface is induced [9]. Both the photo-induced hydrophilicity and the photocatalytic de-gradation of pollutants allow for the exploitation of TiO2-based self-cleaning materials in the construction industry [8,10,11]. Currently, many of such light-responsive materials are commercially available [8,9] and the interest in the photocatalysis is conﬁrmed by the

https://doi.org/10.1016/j.buildenv.2018.01.028

Received 10 August 2017; Received in revised form 30 December 2017; Accepted 15 January 2018

∗_{Corresponding author. Universidad Autónoma de Nuevo León, Facultad de Ciencias Químicas, Departamento de Ingeniería Química, San Nicolás de los Garza, NL 66455, Mexico.}

E-mail address:erika.cedillogn@uanl.edu.mx(E.I. Cedillo-González).

(2)

increasing number of published papers (> 16800 papers from 2008 to 2017 [12]); however, for the practical application of such a functional coating, crucial information such as the chemical durability of TiO2 photocatalytic coatings [13–22] and the influence of weathering on their self-cleaning performance [13,15,17,18,23] are needed. Un-fortunately, the influence of weathering on the photocatalytic proper-ties (e.g. efficiency and chemical durability) is poorly investigated. Furthermore, driven by weathering, TiO2nanoparticles can be released from the coating; this can represent a source of pollution for the en-vironment [24] and a potential hazard for human health [25].

In this study, the eﬀects of indoor and outdoor weathering on the chemical durability and degradation eﬃciency of TiO2coatings were evaluated for dense [8,17,26] and mesoporous coatings [27–31]. While our interest towards dense coating was driven by their commercial availability, mesoporous coatings were chosen as their accessible sur-face area allows for a higher photocatalytic activity [32].

TiO2coatings prepared via sol-gel were deposited on Si substrates. These substrates were preferred over glass substrates because: 1) silicon does not contain Na+_{that was found to decrease the photocatalytic} activity [13,33] and 2) silicon weakly absorbs in the infrared region of interest (c.a. 4000-600 cm−1) allowing for the monitoring of the vi-brational bands of the model pollutant used. Experimental procedures similar to those reported in our previous work [13] were used. Two sets of sol-gel TiO2coatings were tested: 1) TiO2coatings with a dense (or compact) morphology [26] and 2) mesoporous TiO2coatings prepared by evaporation-induced self-assembly (EISA) [34] using non-ionic sur-factants as templating agents [35–37].

In the current study, X-ray Diﬀraction (XRD), Field Emission Scanning Electron Microscopy (FEG-SEM), Energy-dispersive X-ray Spectroscopy (EDS), FTIR Spectroscopy, X-ray Photoemission (XPS), contact angle and adhesion measurements were used to shade light on the eﬀects of weathering on dense and mesoporous TiO2coatings. We posit that the results reported here will be useful for the practical ap-plication of TiO2-based self-cleaning coatings in indoor or outdoor en-vironments.

2. Materials and methods

2.1. Synthesis of dense and mesoporous TiO2coatings

For each set of coatings (dense and mesoporous), 16 samples were prepared. Silicon substrates (c.a. 20 × 20 mm) were cleaned with a piranha solution (3 parts of concentrated H2SO4 and 1 part of 30% H2O2solution). Dense TiO2coatings were synthesized using the meth-odology reported by Costacurta et al. [26]. The TiO2precursor sol was prepared by mixing titanium isopropoxide (Sigma-Aldrich), absolute ethanol (Carlo Erba) and HCl 1M (37% INCOFAR) in the molar ratio 1:100:0.06. The sol was stirred for 1 h before coating. The Si substrates were dip-coated using a withdrawal speed of 85 mm min−1at a relative humidity of 14–15%. This procedure was repeated and 4 layers were deposited. Before the deposition of the next layer, each coating was thermally treated for 10 min @ 110 °C and subsequently for 30 min @ 250 °C. After four depositions, the obtained coatings were thermally treated at 500 °C for 3 h to induce the growth of the anatase as a pre-ferential crystal phase.

Mesoporous TiO2coatings were prepared using a procedure similar to that reported by Crepaldi et al. [34]. In the original procedure, mesoporous titanium oxide coatings were synthetized by evaporation-induced self-assembly (EISA) using a triblock non-ionic surfactant (i.e. Pluronic F127) as templating agent [38]. In this process, the pre-ferential evaporation of the solvent (typically an alcohol) enriches the deposited solution with non-volatile surfactant and inorganic species allowing for the organization of the surfactant in micelles that assemble within the coating in an ordered fashion [34–36]. The decomposition of the surfactant template by thermal treatment allows for the formation of aﬁnal mesoporous structure [37]. In the modiﬁed procedure adopted

here, the precursor solution was prepared by the slow addition of TiCl4 (FLUKA Analytical) to a mixture of absolute ethanol (Carlo Erba) and the tri-block copolymer Pluronic®F-127 (SIGMA). Water (18 MΩ) was added drop-by-drop after stirring for 5 min. The molar ratios were Ti-Cl4:EtOH:H2O = 1:40:12 and TiCl4/Pluronic F-127 = 0.006. Films were dip-coated using a withdrawal speed of 85 mm min−1and 40% relative humidity. After deposition, the coatings were exposed to water vapour for 30 s. The obtained mesoporous coatings were maintained at 200 °C for 24 h to stabilize the micelle arrangement. After this thermal treatment, the temperature was increased up to 500 °C and then maintained for 3 h to crystallize the anatase as predominant phase.

2.2. Domestic and environmental weathering treatments

Indoor weathering was simulated exposing the coatings to two diﬀerent model aqueous solutions that are often present in domestic environments: a 5% isopropanol (Carlo Erba) solution (comparable to a commercial glass-cleaner [39]) and a detergent solution made of 2.5% propylene glycol propyl ether (Aldrich) and 2.5 wt% of sodium dodecyl-benzene-sulfonate (Aldrich) (comparable to a commercial detergent solution [40]). Outdoor weathering was simulated exposing coatings to two diﬀerent solutions: a 18 MΩ deionized water (to simulate rain [17]) and an aqueous acid solution made of HNO3(INCOFAR) and H2SO4 (INCOFAR) at pH 2–3 (acid rain solution according to the EPA Method 1320 [41]). To simulate a prolonged exposure, TiO2coatings were kept in the solutions for 7 days at 37 ± 1 °C. In this study, it was assumed that coatings are not subjected to mechanical cleaning (e.g. rubbing); indeed, when a self-cleaning TiO2 coating is exposed to rain or an aqueous solution (cleaning agents), the hydrophilic characteristics of TiO2allow for the formation of a water coating over the surface [42]. To apply industrial test conditions (TOYOTA Rista Coat Performance tests [43]), an additional ageing test was carried out exposing the coatings to boiling water for 1 h.

After treatments, all coatings were dried at room temperature and stored in a closed vessel until use.

2.3. Coating characterization

Grazing incidence XRD was performed using a Bruker D8 Advance diffractometer with Cu Kα1 radiation coupled with a VANTEC detector. XRD patterns were collected using the same sample before and after chemical treatments. Samples were mounted with the same position, on the same sample holder, into the same XRD diffractometer and in-vestigated using the same measurement conditions. Typically, these are considered conditions sufficient for a comparison between diffraction patterns that are collected from the same substrate subjected to treat-ment [44,45]. Microstructure, coating thickness and particle size were investigated using a FEI Nova NanoSEM 450 FEG-SEM, in immersion lens mode. Thickness was measured by placing each sample horizon-tally into the FEG-SEM chamber and then applying a 90-degree twist to investigate the sample along the axes as previously reported in Ref. [46]. XPS measurements of representative mesoporous samples were taken with a dual anode Vacuum Generators source (model XR3), op-erated at 15 kV, 18 mA and delivering Mg Kα photons (photon energy 1253.6 eV). The used electron analyser is a PHI double pass cylindrical mirror analyser (CMA) model 15-255G. It was operated at a constant pass energy of 50 eV. The Mg Kα satellite lines and a Shirley type background were subtracted from the spectra before analysis. The spectra werefitted with Voigt components for quantitative analyses. The initial water contact angle (WCAi) of all the samples (i.e. the WCA of the samples before and after the chemical weathering in the dark - no UV light) was measured using an OCA Contact Angle System (See Table 1). FTIR spectra for coating characterization were recorded using a Thermo Electron Nicolet 380 instrument. Coating adhesion of the as-prepared and treated samples was determined by scratch test using a CSM Open Platform with a Micro Scratch Tester. A diamond 10μm

(3)

Rockwell tip was used and a load of 400 mN was gradually applied along 6 mm on the surface of the sample. Three scratch measurements were performed on each sample and the average value of the critical load (corresponding to the load necessary to detach the coating from substrate) was calculated.

2.4. Chemical durability

The following analyses were performed to determine the TiO2 re-leased from the coating into the model solutions. TiO2 losses from treated coatings (area = 4 cm2_{) were determined by Inductively} Coupled Plasma Optical Emission Spectrometry (ICP/OES) using a Perkin Elmer ICP OES 2100 DV instrument at 323 nm. Initial amounts of TiO2for dense and mesoporous coatings were calculated by exposing the coating to a digestion solution and measuring the Ti concentration by ICP/OES. The TiO2 concentration present in the model solutions after 7 days of contact with the coatings at 37 °C was calculated from the Ti concentration measured by ICP. The release of TiO2 after weathering was calculated from the ICP results using the following equation: % TiO2released = (μg TiO2in the model solution/μg TiO2in the as-prepared coating) × 100. Additional analyses were performed to determine if the release of the TiO2from the substrates was induced by a loss of adhesion and fragments of TiO2films or particles were released from the coating into the model solution. Mesoporousfilms were used for this task because more influenced by weathering. In particular, the detergent-treated mesoporous coating wasfirst exposed to weathering by the detergent solution at the aforementioned conditions; then, the solution wasfiltered (pore size = 2.5 μm) to remove coating fragments and then analysed by ICP. Thus, the concentration of Ti from TiO2 particles with size < 2.5μm released into the solution was measured. 2.5. Photocatalytic tests

Photocatalytic activity of the as-prepared and treated TiO2coatings was evaluated by assessing the degradation of stearic acid (SA) under UV irradiation at 365 nm (Vilber VL-215.LC lamp, 365/254 nm, in-tensity of 1350/1100μW cm−2). 100μL of an 8.8 × 10−3M SA (Merck) ethanolic solution was poured on each sample to form afilm composed exclusively of the pollutant model. Degradation over time was mon-itored by FTIR. Each spectrum was acquired in transmission mode using a Vertex 70-BRUKER spectrometer, averaging 32 scans between 4000 and 400 cm−1 with a 4 cm−1spectral resolution. Photocatalytic e ffi-ciency was measured by integrating the area under the C-H vibrational bands (stretching) of stearic acid in the 3000-2700 cm−1range at dif-ferent UV exposure times. The photocatalytic tests of the as-prepared and treated coatings were performed in duplicates; no significant dif-ference was observed. The averaged efficiency of the duplicates was used to calculate degradation rates. Each weathering treatment was carried out using freshly prepared coatings, this simulated the use of newly installed TiO2surfaces for construction purposes.

The trends in SA photodegradation were modelled with the pseudo ﬁrst order kinetic model proposed by Sawunyama et al. [47]. This model was preferred over the commonly used zero-order model [3] because it

provided with a more accuratefit to the collected data. The pseudo-first order model reported by Sawunyama et al. describes the random sur-face reactivity between a TiO2coating and a non-homogeneous Lang-muir-Blodget (LB) coating of stearic acid during the photocatalytic process. The underlying equation is given by [SA]t= [SA]iexp(- kobst), where [SA]tis the concentration of stearic acid at time t, [SA]iis the initial concentration of stearic acid, and kobsis the pseudofirst order rate constant. Rate constants were obtained by plotting ln [SA]t/[SA]i vs time. As the weathering solutions promoted leaching of TiO2from both dense and porous coatings (seeTable 2), the analysis was per-formed assuming inhomogeneous active sites [47].

3. Results and discussion

3.1. Characterization of the as-prepared and treated coatings

Fig. 1displays the X-ray diffraction patterns of (a) dense and (b) mesoporous TiO2coatings, before and after weathering. With the ex-ception of the detergent-treated mesoporous sample, in all the other treatedfilms, TiO2in anatase phase was identified. Additional peaks observed in the XRD diffraction patterns were found to correspond to Si and SiO2from the substrate. The presence of SiO2was attributed to the use of strongly oxidizing piranha solution employed during the cleaning of the silicon wafers. Diffraction patterns shown inFig. 1a demonstrate that dense anatase coating are retained on the substrate after all weathering treatments. However, X-ray investigation on mesoporous-detergent treated TiO2sample (Fig. 1b), did not present any diffraction peak of anatase indicating a negligible amount TiO2after exposure to the detergent solution. Only the Si or SiO2peaks were present in such a pattern. Based on a visual inspection, traces of the coating were de-tected only at the edges of the substrate.

FEG-SEM micrographs were performed on as-prepared dense and mesoporous coating. Electron microscopy investigation was performed on the same coatings after exposure to detergent-treatment. The mea-sured thicknesses of the as-prepared dense and mesoporous coatings were 121 ± 4 nm and 117 ± 10 nm, respectively.Fig. 2a shows that the dense coating consisted of a continuous surface of agglomerated TiO2 nanoparticles with a diameter of c.a. 41 ± 6 nm. Pores of 18 ± 3 nm were present as a result of the agglomeration of the na-noparticles, but it can be inferred that these pores were not inter-connected.Fig. 2b shows the presence of a“worm-like” mesoporous structure composed of TiO2 nanoparticles of 21 ± 2 nm and pore channels with a diameter of about 17 ± 2 nm. The mesostructure was less ordered than what is usually reported for TiO2coatings prepared by the EISA method. However, the morphology was found to be very si-milar to the mesostructure recently reported by Li et al. [32] for a mesoporous TiO2coating where meso-periodicity was present. Thus, comparing the porosity fromFig. 2a and b the higher surface area of the mesoporous sample was conﬁrmed. With FEG-SEM we found that weathering did not aﬀect the morphology of the dense TiO2coating. Indeed,Fig. 2c displays a representative micrograph of the detergent-treated (the most aggressive treatment in terms of coating removal)

Table 1

WCAiand adhesion values of the TiO2coatings.

Sample WCAi(°) Critical load (mN)

Dense Mesoporous Dense Mesoporous As-prepared 39 ± 0 34 ± 1 112 ± 1 137 ± 1 Water 34 ± 1 29 ± 1 112 ± 1 136.3 ± 2 Acid rain 50 ± 1 33 ± 1 113 ± 1 128 ± 6 Glass cleaner 55 ± 0 42 ± 1 115 ± 2 142 ± 4 Detergent 31 ± 1 17 ± 2 118 ± 1 138 ± 4 Ageing test 55 ± 1 35 ± 1 116 ± 0 131 ± 1 Table 2

Percentage of the originally deposited TiO2released after weathering.

Treatment TiO2released (%)

Dense Mesoporous Water 3.7 11.1 Acid rain 9.5 50.0 Glass-cleaner 3.3 12.5 Detergent 13.3 90.0a_{; 70.1}b Ageing test 3.4 11.1 a_Without_ﬁltration. b_{Filtered solution.}

(4)

dense sample; a morphology very similar to the as-prepared coating was observed. On the other hand, depending on the zone of the sample, weathering by the detergent solution seems to have a major influence over the mesoporous TiO2coating. The detergent solution did not affect the morphology in the edges of the sample (Fig. 2d), while the centre of the coating showed a dramatic change of the morphology: porosity was no longer detected (Fig. 2e). The EDS spectra performed on the centre of the mesoporous sample (Fig. 2f) did not reveal Ti (only C and Si were found), thus the detrimental effect of this weathering treatment was

demonstrated. On the contrary, the EDS spectra collected on the as-prepared sample showed the presence of Ti (See Supplementary Ma-terial). By combining the results obtained from XRD, FEG-SEM and EDS, it was concluded that the detergent treatment caused a substantial removal of the TiO2mesoporous coating.

To promote an eﬃcient self-cleaning performance, TiO2must show good photo-oxidation properties and high hydrophilicity. This second characteristic allows for the rain to wash away dust and organic con-taminants from the TiO2surface [42].Table 1shows that weathering

Fig. 1. X-ray patterns of the (a) dense and (b) mesoporous TiO2coatings before and after chemical treatments simulating diﬀerent common weathering conditions.

Fig. 2. FEG-SEM images of as-prepared (a) dense and (b) mesoporous coatings; the detergent-treated (c) dense and (d and e) mesoporous coatings (edge and centre of the sample, respectively) and the (f) EDS spectra of the detergent-treated mesoporous coating (centre of the sample).

(5)

applied in the absence of UV light (i.e. dark environment) resulted in a modified initial hydrophilicity for the different coatings. Mesoporous samples always showed lower WCAi values than dense coatings in-dicating that a porous anatase film has a higher affinity for water. Further discussions related to the self-cleaning performance are re-ported in Section3.4.

3.2. Coating adhesion

Table 1presents the load value that is required to mechanically detach the coating from the Si substrate (critical loads). For as-prepared mesoporous coatings, the critical load was found to be 1.22-fold higher than the one measured on dense coatings. The values of the critical loads of the treated dense coatings revealed that the mechanical ad-hesion did not change significantly after indoor weathering, since a maximum adhesion difference of 4.6% was measured between the as-prepared and treated coatings. Regarding the mechanical adhesion of mesoporous coatings, weathering promoted a maximum difference between the as-prepared and treated coatings of 6.6%. In case of de-tergent solution, the mesoporous coating was found on the edges of the sample (vide supra); on these regions (where TiO2survived to weath-ering) a critical load similar to that of the as-prepared coating was measured.

From these tests, it was concluded that weathering did not sig-niﬁcantly change the mechanical adhesion of the coatings; then we speculate that the removal of TiO2mesoporous coatings was chemically driven.

3.3. TiO2release

From the ICP measurements, it was found that the initial quantities of TiO2deposited in the dense and mesoporous coatings were 12.5 and 4.3μg cm−2, respectively. Percentages of TiO2 release (including coating residues) after indoor and outdoor weathering were calculated by ICP analyses and results are shown inTable 2. For both coatings, the highest release was measured after treatment with detergent and acid rain solutions.Table 2shows that the TiO2losses from treated meso-porous coatings were 3.0- to 6.7-fold higher than the corresponding losses from dense coatings. Remarkably, 70% of the originally de-posited TiO2was present in the model solution (filtered solution) after weathering of mesoporous coatings. These results suggest that the chemical durability is influenced by the high surface area of the me-soporousfilm. We believe that the model solutions can diffuse through the pores of the mesoporous coating and interact with a large volume of TiO2present in the sample. Instead, in a dense sample the model so-lutions can only interact with the outer TiO2surface. In general, it can be concluded that TiO2nanoparticles migrated to the model solution after chemical treatment. Currently, TiO2 is considered non-toxic; however, the release of nanoparticles is of potential concern for both human health and environment. Hence, the here performed durability tests can be of importance for the practical application of mesoporous TiO2coated materials.

3.4. Photocatalytic activity of dense and mesoporous TiO2coatings

Fig. 3shows the photocatalytic degradation of stearic acid during irradiation at 365 nm and room temperature.Fig. 4shows the ln [SA]t/ [SA]ivs time; from these plots, degradation rate constants were ob-tained (Fig. 5). Both rate constants and R2_{values obtained from the} Sawunyama model (Fig. 4) are in good agreement with those reported for dense TiO2ﬁlms [17,48]. Rate constants are signiﬁcantly higher in mesoporous TiO2coatings.

Fig. 3shows that the photocatalytic eﬃciency of the as-prepared dense coating was lower than its mesoporous counterpart. As presented inFig. 5, the rate of the mesoporous coating (191.8 × 10−4min−1) is 8-fold higher than that of the dense sample (24.0 × 10−4min−1). This

Fig. 3. Photocatalytic eﬃciency of (a) dense and (b) mesoporous TiO2coatings before

and after the chemical treatments.

Fig. 4. Calculation of the pseudo-ﬁrst order rate constant (kobs) applying the Sawunyama

model for (a) dense and (b) mesoporous TiO2coatings.

Fig. 5. SA photocatalytic degradation rates for the as-prepared and treated (a) dense and (b) mesoporous TiO2coatings.

(6)

result was expected as a consequence of the larger surface area deriving from the higher porosity of mesoporous samples. An analysis ofFigs. 3a and 5a revealed that weathering was beneﬁcial for the dense TiO2 coating. Aged and acid rain-treated dense samples presented the best photocatalytic performance showing degradation rates that were 2.7-fold greater than the as-prepared sample. In the case of acid-treated coating, Yu et al. [49] previously demonstrated that treatments with H2SO4of TiO2coatings deposited over soda-lime glasses enhance their photocatalytic activity due to (i) the reduction of Na contamination -from glass substrates- and to (ii) an increase of the adsorbed hydroxyl content on the TiO2surface. In this work, the Na contamination con-tribution from substrates can be excluded as Si wafers were used as substrates; then, the higher activity was related to the increase of the adsorbed hydroxyl content on the dense TiO2surface and the conse-quently higher concentration of OH• for SA degradation when the system is irradiated with UV light. In the case of the aged dense sample, the improvement in eﬃciency was also attributed to a high availability of hydroxyl radicals as a result of exposure to boiling water for 1 h.

The water-treated dense coating showed a good performance; in-deed, its SA degradation rate was found to be 1.8-fold greater than that of the as-prepared dense coating. The water-treated dense sample was expected to show a degradation rate comparable to that of the aged coating because both samples were immersed into water. However, complete degradation was reached after 480 min irradiation; thus a reduced degradation rate was observed for the water treated-dense coating (Fig. 5a). This behaviour was attributed to a long treatment time (i.e. 7 days) that promoted the inhibition of the photo-catalyst due to the saturation of the TiO2surface with–OH [5,7]; such a saturation is achieved because water and SA molecules compete for the allocation of TiO2active sites [5,49].

The saturated state was conﬁrmed by the WCAivalues, which were lower for the water-treated dense sample (34.2° ± 0.5°) than for the acid rain-treated (50.1° ± 1.2°) and aged (54.5° ± 0.5°) samples.

Degradation rates of dense coatings after exposure to detergent and isopropanol treatments were found to be 1.7-fold and 1.5-fold higher than that of the as-prepared coating, respectively. This improved e ﬃ-ciency was correlated with the exposure to diluted aqueous solutions of cleaning agents that allows for an abundant source of –OH groups. However, considering the values of the TiO2losses after weathering (Table 2), it is not advisable to expose dense TiO2coatings (derived from similar recipes or with similar characteristics to those of the coating investigated here) to environments with detergent solutions, in order to prevent the release of nanoparticles to the environment.

Figs. 3b and 5b show that weathering can be either detrimental or beneﬁcial for the photocatalytic activity of mesoporous TiO2coatings. Indoor weathering by cleaning agents has the highest detrimental ef-fect, leaving the coating with a decreased or negligible photocatalytic activity. The acid rain and aging for mesoporous samples induced si-milar degradation rate values (139.8 and 164.8 × 10−4 min−1, re-spectively). It is worth noticing that these values are slightly lower than the degradation rate determined for the as-prepared sample (191.8 × 10−4 min−1). However, simulated rain (water-treated coating) enhanced the degradation rate of SA in mesoporous TiO2 coatings (477.1 × 10−4 min−1), resulting in a 2.8-fold degradation eﬃciency with respect to the as-prepared coating. This increased ac-tivity was correlated with the much higher availability of–OH groups (vide supra). Another advantage of this treatment is the low release of nanoparticles from the mesoporous coating (seeTable 2).

FromFigs. 3b and 5b it is clear that the cleaning agents have the most negative eﬀect on the activity of mesoporous TiO2coatings. Based on the TiO2release values presented in Section3.3, the low activity was expected for the detergent-treated sample, but not for the isopropanol-treated one, which only lost 12.5% of the originally deposited TiO2. In the case of the isopropanol treatment, the absence of activity was at-tributed to chemical deactivation of the coating and not the removal of the same, according to the results presented inTable 2. It was reported

that the contact of gaseous isopropyl alcohol (IPA, the main component of the glass cleaner solution) with TiO2at 300 K leads to the formation of three adsorbed surface species: (i) strongly bonded non-dissociated IPA species on Ti+δsites; (ii) strongly dissociated adsorbed IPA ac-cording to the following reaction:

CH3CH(OH)CH3+ S(1)+ S(2)→ CH3CH(O-S(1)ads)CH3+ S(2)-HadsEq. (1) where S(1)and S(2)represent two different TiO2adsorption sites and (iii) weakly hydrogen-bonded IPA species on Ti+δsites [50]. Thus, the low activity of the coating was correlated with the irreversible ad-sorption of IPA over the mesoporous TiO2active sites, preventing the interaction of SA with the remaining TiO2surface. The FTIR spectra in the 4000-2400 cm−1range of the as-prepared and isopropanol-treated mesoporous sample (before SA deposition) are shown inFigure S2(See supplementary material). The spectrum of the isopropanol-treated sample was measured 12 h after completion of the entire weathering treatment. If compared with the as-prepared sample, the C-H stretching modes of the adsorbed IPA are visible in the isopropanol-treated coating. Further confirmation of the adsorption of a non-polar specie on the TiO2 surface was revealed by the increase in the WCAiafter the treatment (SeeTable 1). As both dense and mesoporous coatings are composed of TiO2but only differ in their microstructure, the spectra of both isopropanol-treated coatings were compared in order to determine if IPA was also adsorbed on the dense surface. As no traces of CH vi-brational modes were found in the spectrum of the dense coating (not shown here), it was concluded that the microstructure of the dense TiO2 coating with lower surface area leads to a negligible adsorption of IPA. IPA degradation induced by TiO2was reported [51]; therefore, it is likely that TiO2coatings require treatments with UV irradiation before SA deposition in case a mesoporous TiO2structure is used. It is worth to mention that this pre-treatment with UV light after isopropanol-treat-ment and before SA deposition was not performed in the present study, since this work focused on the effects of domestic weathering of TiO2 coatings. Restoration of the original properties of photocatalytic TiO2 coatings after indoor weathering will require further investigations.

The mesoporous detergent-treated coating presented partial and slow photocatalytic activity during theﬁrst 90 min of UV irradiation (Figs. 3 and 5). After this time, the coating did not signiﬁcantly degrade SA. This behaviour was correlated to a combination of factors: 1) the remaining photocatalytic TiO2from the original coating, 2) the loss of surface area (SeeFig. 2e) and 3) the smaller amount of hydroxyl radi-cals produced.

3.5. XPS analysis of the as-prepared and detergent-treated mesoporous coating

To confirm the presence of TiO2after treatment and validate the aforementioned hypotheses, XPS analysis was performed.Fig. 6shows the XPS spectra of Ti2pand Si2plevels measured on the as-prepared and detergent-treated mesoporous samples.Fig. 6a confirms that traces of TiO2 remained on the substrates after treatments. Indeed, the XPS spectra show a strong reduction of the Ti peaks after treatment with the detergent model solution. As XPS is sensitive to the surface,Fig. 6b shows that the substrate contributions (Si peak at 99 eV) increased when the coating was exposed to weathering with detergent solutions. The peak at 104 eV is related to SiO2.Fig. 6confirms that the detergent treatment promotes erosion of the TiO2coating. This erosion reduces the amount of photocatalyst available. Deactivation after 90 min of ir-radiation can be related to consumption of hydroxyl radicals, since the replenishment of hydroxyl groups (in this case from air humidity) are necessary for photocatalysis to proceed, and the limited surface area of the remaining coating is likely to hamper the interaction between hu-midity and the coating, leading to deactivation of thefilm.

(7)

have slow activity but can be subjected to several indoor and outdoor weathering conditions with low risk of release of nanoparticles. Conversely, highly active mesoporous TiO2 coatings have lower che-mical durability than their dense counterparts; their exposure to weathering is detrimental to their self-cleaning performance and pro-motes the release of TiO2particles from the coating. These character-istics are listed in Table 1S (See Supplementary Material), where comparisons with a TiO2self-cleaning glass derived from a commercial nano-dispersion (ø = 50 nm) [13], (named“nanoparticled coating”) are made.Table 1S shows that the nanoparticled and dense TiO2coatings have higher resistance to weathering than the mesoporous one. This result was related to the characteristic porosity and high surface area of the latter, as it allows for a strong interaction between the model so-lutions and the TiO2surface during weathering.

4. Conclusions

Results show that the mechanical adhesion of both dense and me-soporous TiO2coatings are not affected by indoor (domestic) and out-door (environmental) weathering. However, weathering affects other characteristics such as the initial contact angle, the TiO2concentration, the photocatalytic efficiency, the degradation rate and the surface composition (adsorption of OH−and isopropanol). For the mesoporous sample, weathering by detergent solution also affects the micro-structure of the sample. Without any simulated weathering treatment, the photocatalytic efficiency of the as-prepared dense coating is lower than that of its mesoporous counterpart. Weathering has a positive ef-fect on the efficiency of the dense coating; however, on the mesoporous sample, weathering can be either detrimental or beneficial for the photocatalytic activity. Results suggest that the use of dense TiO2 coatings can reduce the release of nanoparticles; however, a slower photo-degradation activity with respect to mesoporous coatings should be taken into account. Conversely, highly active mesoporous TiO2 coatings, exposed to chemicals used in domestic environments, are subjected to a decrease of their self-cleaning properties; from these coatings TiO2particles are released into model solutions. Finally, by comparing the results obtained here with those already reported for TiO2 coatings prepared using a commercial nano-dispersion, it was concluded that dense coatings have higher resistance to weathering than coatings with higher porosity. This fact was attributed to the larger surface area that promotes a strong interaction between model solu-tions and the accessible surface of TiO2coatings.

Conﬂicts of interest None.

Acknowledgements

Authors would express their acknowledgement to the ﬁnancial support made by the ARACNE - Laboratorio Integrato Sviluppo

Tecnologie Avanzate Materiali Innovativi per Costruzioni Ecosostenibili through the Italian Program Bando regionale“Dai distretti produttivi ai distretti tecnologici”.

Appendix A. Supplementary data

Supplementary data related to this article can be found athttp://dx. doi.org/10.1016/j.buildenv.2018.01.028.

References

[1] A. Fujishima, K. Honda, Electrochemical photolysis of water at a semiconductor electrode, Nature 238 (1972) 37–38.

[2] M. Gagliardi, Photocatalysts: Technologies and Global Markets: AVM069B, BCC Research, 2015, https://www.bccresearch.com/market-research/advanced-materials/photocatalysts-technologies-markets-report-avm069b.html, Accessed date: 6 June 2017.

[3] A. Mills, A. Lepre, N. Elliott, S. Bhopal, I.P. Parkin, S.A. O'Neill, Characterisation of the photocatalyst Pilkington Activ™: a reference ﬁlm photocatalyst? J. Photochem. Photobiol. Chem. 160 (2003) 213–224.

[4] O. Akhavan, Lasting antibacterial activities of Ag–TiO2/Ag/a-TiO2nanocomposite thinﬁlm photocatalysts under solar light irradiation, J. Colloid Interface Sci. 336 (2009) 117–124.

[5] C. Raillard, V. Héquet, P. Le Cloirec, J. Legrand, Kinetic study of ketones photo-catalytic oxidation in gas phase using TiO2-containing paper: eﬀect of water vapor, J. Photochem. Photobiol. Chem. 163 (2004) 425–431.

[6] T.N. Obee, R.T. Brown, TiO2photocatalysis for indoor air applications: eﬀects of humidity and trace contaminant levels on the oxidation rates of formaldehyde, toluene, and 1,3-butadiene, Environ. Sci. Technol. 29 (1995) 1223–1231. [7] J.M. Coronado, M.E. Zorn, I. Tejedor-Tejedor, M.A. Anderson, Photocatalytic

oxi-dation of ketones in the gas phase over TiO2thinﬁlms: a kinetic study on the inﬂuence of water vapor, Appl. Catal. B Environ. 43 (2003) 329–344.

[8] J. Chen, C. Poon, Photocatalytic construction and building materials: from funda-mentals to applications, Build. Environ. 44 (2009) 1899–1906.

[9] A. Mills, S.-K. Lee, A web-based overview of semiconductor photochemistry-based current commercial applications, J. Photochem. Photobiol. Chem. 152 (2002) 233–247.

[10] K. Demeestere, J. Dewulf, B. De Witte, A. Beeldens, H. Van Langenhove, Heterogeneous photocatalytic removal of toluene from air on building materials enriched with TiO2, Build. Environ. 43 (2008) 406–414.

[11] P. Krishnan, M.-H. Zhang, Y. Cheng, D.T. Riang, L.E. Yu, Photocatalytic degradation of SO2using TiO2-containing silicate as a building coating material, Construct. Build. Mater. 43 (2013) 197–202.

[12] Scopus search with“TiO2” and “Photocatalysis” keywords. May 2017, (2017).

[13] E.I. Cedillo-González, R. Riccò, M. Montorsi, M. Montorsi, P. Falcaro, C. Siligardi, Self-cleaning glass prepared from a commercial TiO2nano-dispersion and its pho-tocatalytic performance under common anthropogenic and atmospheric factors, Build. Environ. 71 (2014) 7–14.

[14] A. Chabas, T. Lombardo, H. Cachier, M.H. Pertuisot, K. Oikonomou, R. Falcone, M. Verità, F. Geotti-Bianchini, Behaviour of self-cleaning glass in urban atmosphere, Build. Environ. 43 (2008) 2124–2131.

[15] A. Chabas, L. Gentaz, T. Lombardo, R. Sinegre, R. Falcone, M. Verità, H. Cachier, Wet and dry atmospheric deposition on TiO2coated glass, Environ. Pollut. 158 (2010) 3507–3512.

[16] M.-Z. Guo, T.-C. Ling, C.-S. Poon, Nano-TiO2-based architectural mortar for NO removal and bacteria inactivation: inﬂuence of coating and weathering conditions, Cement Concr. Compos. 36 (2013) 101–108.

[17] N.P. Mellott, C. Durucan, C.G. Pantano, M. Guglielmi, Commercial and laboratory prepared titanium dioxide thinﬁlms for self-cleaning glasses: photocatalytic per-formance and chemical durability, Thin Solid Films 502 (2006) 112–120. [18] J. Olabarrieta, S. Zorita, I. Peña, N. Rioja, O. Monzón, P. Benguria, L. Scifo, Aging of

photocatalytic coatings under a waterﬂow: long run performance and TiO2 nano-particles release, Appl. Catal. B Environ. 123 (2012) 182–192.

[19] M. Piispanen, J. Määttä, S. Areva, A.-M. Sjöberg, M. Hupa, L. Hupa, Chemical

(8)

resistance and cleaning properties of coated glazed surfaces, J. Eur. Ceram. Soc. 29 (2009) 1855–1860.

[20] A. Calia, M. Lettieri, M. Masieri, Durability assessment of nanostructured TiO2 coatings applied on limestones to enhance building surface with self-cleaning ability, Build. Environ. 110 (2016) 1–10.

[21] P. Munafò, E. Quagliarini, G.B. Goﬀredo, F. Bondioli, A. Licciulli, Durability of nano-engineered TiO2self-cleaning treatments on limestone, Construct. Build. Mater. 65 (2014) 218–231.

[22] L. Graziani, E. Quagliarini, F. Bondioli, M. D'Orazio, Durability of self-cleaning TiO2 coatings onﬁred clay brick façades: eﬀects of UV exposure and wet & dry cycles, Build. Environ. 71 (2014) 193–203.

[23] M. Baudys, J. Krýsa, M. Zlámal, A. Mills, Weathering tests of photocatalytic facade paints containing ZnO and TiO2, Chem. Eng. J. 261 (2015) 83–87.

[24] Y. Ju-Nam, J.R. Lead, Manufactured nanoparticles: an overview of their chemistry, interactions and potential environmental implications, Sci. Total Environ. 400 (2008) 396–414.

[25] C. Ostiguy, B. Soucy, G. Lapointe, C. Woods, L. Ménard, M. Trottier, Health Eﬀects of Nanoparticles Report R-589 2nd Ed., Montreal, (2008).

[26] S. Costacurta, G.D. Maso, R. Gallo, M. Guglielmi, G. Brusatin, P. Falcaro, Inﬂuence of temperature on the photocatalytic activity of Sol−Gel TiO2ﬁlms, ACS Appl. Mater. Interfaces 2 (2010) 1294–1298.

[27] L. Pinho, M. Rojas, M.J. Mosquera, Ag–SiO2–TiO2nanocomposite coatings with enhanced photoactivity for self-cleaning application on building materials, Appl. Catal. B Environ. 178 (2015) 144–154.

[28] A.A. Ismail, H. Bouzid, Synthesis of mesoporous ceria/titania thinﬁlms for self-cleaning applications, J. Colloid Interface Sci. 404 (2013) 127–134.

[29] T. Sreethawong, S. Ngamsinlapasathian, S. Yoshikawa, Positive role of in-corporating P-25 TiO2to mesoporous-assembled TiO2thinﬁlms for improving photocatalytic dye degradation eﬃciency, J. Colloid Interface Sci. 430 (2014) 184–192.

[30] J. Krýsa, M. Baudys, M. Zlámal, H. Krýsová, M. Morozová, P. Klusoň, Photocatalytic and photoelectrochemical properties of sol–gel TiO2ﬁlms of controlled thickness and porosity, Catal. Today 230 (2014) 2–7.

[31] J. Rathouský, V. Kalousek, V. Yarovyi, M. Wark, J. Jirkovský, A low-cost procedure for the preparation of mesoporous layers of TiO2eﬃcient in the environmental clean-up, J. Photochem. Photobiol. Chem. 216 (2010) 126–132.

[32] R. Li, M. Faustini, C. Boissière, D. Grosso, Water capillary condensation eﬀect on the photocatalytic activity of porous TiO2in air, J. Phys. Chem. C 118 (2014) 17710–17716.

[33] Y. Paz, A. Heller, Photo-oxidatively self-cleaning transparent titanium dioxideﬁlms on soda lime glass: the deleterious eﬀect of sodium contamination and its preven-tion, J. Mater. Res. 12 (1997) 2759–2766.

[34] E.L. Crepaldi, G.J.de A.A. Soler-Illia, D. Grosso, F. Cagnol, A. François Ribot, C. Sanchez, Controlled formation of highly organized mesoporous titania thinﬁlms: from mesostructured hybrids to mesoporous nanoanatase TiO2, J. Am. Chem. Soc.

125 (2003) 9770–9786.

[35] J.M. Bertolo, A. Bearzotti, P. Falcaro, E. Traversa, P. Innocenzi, Sensoristic appli-cations of self-assembled mesostructured silicaﬁlms, Sens. Lett. 1 (2003) 64–70. [36] P. Innocenzi, L. Malfatti, T. Kidchob, P. Falcaro, M.C. Guidi, M. Piccinini,

A. Marcelli, Kinetics of polycondensation reactions during self-assembly of mesos-tructuredﬁlms studied by in situ infrared spectroscopy, Chem. Commun. 0 (2005) 2384–23-86.

[37] P. Falcaro, D. Grosso, H. Amenitsch, P. Innocenzi, Silica orthorhombic mesos-tructuredﬁlms with low refractive index and high thermal stability, J. Phys. Chem. B 108 (2004) 10942–10948.

[38] P. Innocenzi, L. Malfatti, T. Kidchob, P. Falcaro, Order−Disorder in self-assembled mesostructured silicaﬁlms: a concepts review, Chem. Mater. 21 (2009) 2555–2564. [39] Colgate-Palmolive Company, Ajax Expert®Glass and Multi-surface Cleaner RTU

MSDS, (2004).

[40] Colgate-Palmolive Company, All Purpose Cleaner Fabuloso-lavender MSDS, (2010). [41] M. US EPA, OSWER, ORCR, SW-846 test method 1320: multiple extraction

proce-dure, https://www.epa.gov/hw-sw846/sw-846-test-method-1320-multiple-extraction-procedure, (1986) , Accessed date: 7 June 2017.

[42] A. Fujishima, X. Zhang, Titanium dioxide photocatalysis: present situation and fu-ture approaches, Compt. Rendus Chem. 9 (2006) 750–760.

[43] TOYOTA Rista Coat Performance tests, 2005. URL:http://www.ristacoat.cn/en/ index.html. (Accessed 4 April 2016).

[44] D. Grosso, G.J.A.A. De Soler-Illia, F. Babonneau, C. Sanchez, P.A. Albouy, A. Brunet-Bruneau, A.R. Balkenende, Highly Organized Mesoporous Titania Thin Films. [45] P. Falcaro, S. Costacurta, G. Mattei, H. Amenitsch, A. Marcelli, M.C. Guidi,

M. Piccinini, A. Nucara, L. Malfatti, T. Kidchob, P. Innocenzi, Highly ordered “de-fect-free” self-assembled hybrid ﬁlms with a tetragonal mesostructure, J. Am. Chem. Soc. 127 (2005) 3838–3846.

[46] E.I. Cedillo-Gozález, R. Riccò, S. Costacurta, C. Siligardi, P. Falcaro, Below room temperature: how the photocatalytic activity of dense and mesoporous TiO2 coat-ings is aﬀected, Appl. Surf. Sci. 435 (2018) 769–775.

[47] P. Sawunyama, L. Jiang, A. Fujishima, K. Hashimoto, Photodecomposition of a Langmuir−Blodgett ﬁlm of stearic acid on TiO2ﬁlm observed by in situ atomic force microscopy and FT-IR, J. Phys. Chem. B 101 (1997) 11000–11003. [48] P. Falcaro, G. Zaccariello, V. Stoyanova, A. Benedetti, S. Costacurta, Temperature

matters: an infrared spectroscopic investigation on the photocatalytic eﬃciency of titania coatings, Sci. Adv. Mater. 6 (2014) 1330–1337.

[49] J.C. Yu, J. Yu, J. Zhao, Enhanced photocatalytic activity of mesoporous and or-dinary TiO2thinﬁlms by sulfuric acid treatment, Appl. Catal. B Environ. 36 (2002) 31–43.

[50] L. Sivachandiran, F. Thevenet, P. Gravejat, A. Rousseau, Isopropanol saturated TiO2 surface regeneration by non-thermal plasma: inﬂuence of air relative humidity, Chem. Eng. J. 214 (2013) 17–26.

[51] S.A. Larson, J.A. Widegren, J.L. Falconer, Transient studies of 2-propanol photo-catalytic oxidation on titania, J. Catal. 157 (1995) 611–625.