The TiO₂/kaolin composite powders were obtained by the impregnation of TiO₂ commercial suspension (in contents of 3 and 10wt.%) into a kaolin clay support. The commercial TiO₂ suspension used for this purpose (80 wt.% anatase and 20 wt.% rutile; grain size < 100 nm; content of dry matter 30.0 ± 1.0 wt. % and pH 7) was obtained from Degussa company, Germany. The process of impregnation was carried out by mechanical activation in two different conditions of grinding. Firstly, the impregnation was performed in an attritor mill for 90 min at the speed of 1500 rpm. The second way of impregnation was carried out in a planetary mill during a period of 180 min at the speed of 200 rpm. The material/ball ratio was 1/5 in both procedures, while the pH value was in the range of 9 to 9.5. This pH value was maintained with the NaCO₃ and NaOH solutions, with concentrations of 0.67M and 2.25M, respectively. The obtained powders were dried at 105^oC for 24 hours.

According to the TiO₂ content and the condition of impregnation, the obtained TiO₂/kaolin composite powders were labeled as E0A; E1A; E2A; E0P; E1P; E2P. The initial letter E indicates the kaolin clay used as support, numbers 0, 1 and 2 indicate the content of TiO₂ (0- without TiO₂, 1- 3.wt%TiO₂ and 2- 10wt.% TiO₂) while the last letter A / P indicate the attritor/ planetary mill conditions, respectively.

The obtained TiO₂/kaolin composite powders were used to make suspensions that later acted as protective coatings on mineral substrates. The composite suspension was formed by suspending 0.5g of TiO₂/kaolin composite powder into demineralized water (100 ml) using di-ammonia-hydrogen citrate as a dispersing agent. The suspension was stirred at 300 rpm, for 1hour. Ultrasonic bath (30 min) was used in order to prevent possible agglomeration.

Spray technique was applied for the deposition of the suspension onto the surface of the substrates under the following conditions: spraying pressure 6.5MPa, distance of the spray device from the sample 90cm, angle of spraying 45^o, and diameter of nozzle 1.3mm. The coated substrates were afterwards dried at RT/24h.

Three types of mineral substrates were used as the medium for the application of the obtained suspensions as protective coatings. The first one, considered a porous system, was a clay roofing tile (CRT) produced in industrial conditions (Company NEXE Group) by extrusion, the second substrate was produced in laboratory conditions by pressing (P= 73MPa) using 60 wt.% of kaolin clay (35 wt% kaolinite) and 40 wt.% of fly ash obtained from the thermal power plant REK Bitola, Republic of Macedonia. The used fly ash influenced the increase of porosity of the mineral substrate (26) and it was considered as a high porous system in this investigation. The last substrate was a window glass (WG) used as a representative non-porous material. The substrates used in the investigation without coating are assigned as the reference using suffix R (CRT-R, CFA-R, WG-R), while the coated substrates were labelled by the suffix of the applied suspension (CRT-E2A, CRT-E2P, CAF-E2A, CAF-E2P, WG-E2A, WG-E2P).

The particle size distribution of the TiO₂/kaolin composite powders was determined by Malvern Instruments, zeta-nanoseries, NanoZS under the following conditions: refraction index of the investigated suspensions (n=1.55), light absorption (a=0.3) and pH = 9. Additionally, a Malvern Mastersizer 2000 instrument was used for a more precise determination of the micro-size distribution of the particles from 0.02 to 2000 mm. For this measurement, the sample particles were dispersed in water with sodium pyrophosphate added with agitation and sonication until a stable dispersion was obtained.

The phase composition of the TiO₂/kaolin composite powders was determined by X-ray diffraction (Philips PW 1710). The investigations were done under the following conditions: monochromatic CuKα radiation with λ=1.5418 Ȧ in the 5-55^o of 2θ range, scan rate 0.02^o, 0.5 s per step.

The measurement of the contact angle values (Surface Energy Evaluation System, Advex Instruments, Brno, Czech Republic) was carried out with two experimental fluids: (distilled water and glycerol) before and after the water rinsing procedure and adhesion test. Droplets of the appropriate experimental fluid (cca 5 μl in volume) were gently deposited on the coated mineral substrate by micro syringe. The initial contact angle measurements (θ_ci, after 1s) were performed at five different points for each of the three specimens of the investigated mineral substrates. Each droplet deposited onto the surface of the mineral substrate was measured five times.

The surface roughness (Surtronic 25, Taylor Hobson) of the reference mineral substrates and of the coated samples, before and after water rinsing procedure and adhesion test, was evaluated based on the Ra parameter which represents the average roughness values obtained with a 4mm linear probe length. The obtained data were calculated according to ISO 4287 standard.

Before and after the water rinsing procedure and adhesion test, the Vickers micro-hardness values of the reference and of the coated samples were measured by Vickers micro-hardness technique (Microhardness tester model HVS 1000A, ZZV Precision Toll Supply) applying 0.3kg load.

The water absorption values by capillarity of the reference and the coated substrates were evaluated according to standard SRPS U. M8.300, 1985 – EN (27). The test samples (surface area of 60cm²) were dried until the constant mass (m₀) was reached, lateral sides of the samples were sealed with silicon, and placed in a vessel with demineralized water on the coated side. The samples were weighed (m_i) at determined time intervals (5min). The amount of water absorbed by the sample per unit area Q_i (kg/m²) at time t_i (s) was calculated as follows [1]:

where:

S is the area of the sample in contact with water.

The capillary water absorption coefficient (A) was defined as the slope of the linear section of the curve obtained by plotting the mass change per area (Q_i) vs. the square root of time (t_i^1/2).

The photocatalytic activity of the synthetized powders (E0A; E1A; E2A; E0P; E1P; E2P) and that of the coated mineral substrates (before and after water rinsing and adhesion test) was investigated by monitoring the Rhodamine B (RhB) concentration change under the UV/VIS irradiation according to the procedures described in detail in (22). The photocatalytic activity was estimated based on the efficiency of the RhB degradation. The photocatalytic activity A (%) was calculated according to the following relation [2]:

where:

C₀ presents the RhB solution concentration for the sample in the dark at defined time.

C presents the RhB solution concentration for the sample under UV/VIS light irradiation at defined time.

The RhB concentration was measured by UV-VIS spectrophotometer (EVOLUTION 600 spectrophotometer).

The assessment of the durability of the coated surfaces was carried out by using water rinsing test (as essential for self-cleaning of the coating) and tape adhesion test.

Water rinsing durability test was applied in order to examine the stability of three mineral substrates coated with the suspensions made of E2A / E2P nanocomposite powders in severe conditions (rain rinsing). Namely, the laboratory simulated rain rinsing procedure (19) was realized using the equipment which provides constant tap water flow (0.04 l/s) through a pipe system (nozzle diameter of 0.90mm) and water drop (height of 50 cm) on the sample (set at an angle of 45^o). The duration of the test was 30 min.

A modified adhesion test according to the procedure for assessment of the porous models (15) was applied on the chosen substrates. A pressure sensitive tape (Scotch Tape, 3M) of defined dimensions was applied on the coated surface. The relative mass loss (Δm/S) after the application of the tape test was reported; where Δm presents the difference between the mass of the tape measured after and before the test, while S presents the surface covered with the tape. The lower relative mass loss (Δm/S) indicates a better adhesion of the coating to the substrate surface.

The functional properties (photocatalytic activity and self-cleaning efficiency) of the coated samples were measured before and after the water rinsing and tape adhesion tests during UV/VIS irradiation. The Rhodamine B (10ppm) was used as the model pollutant forphotocatalytic activity assessment, while the results of the contact angle measurements were used for the estimation of self-cleaning properties.

The results of the photocatalytic activity of the obtained TiO₂/kaolin composite powders (E0A; E1A; E2A; E0P; E1P and E2P) obtained by monitoring the photocatalytic degradation of RhB during the period of 210min by UV/VIS irradiation are presented in Figure 1.

It is evident that composite powder E2A has the highest photocatalytic activity value (65%), while the composite powder E2P has the lowest one (55%).

Regarding the TiO₂ content impregnated into the kaolin clay support, the samples impregnated with 10% TiO₂ showed significantly higher photocatalytic activity than the samples impregnated with 3% TiO₂, while the samples treated under the same conditions without impregnated TiO₂ showed no photocatalytic activity under the UV/VIS irradiation (28).

Depending on the type of mill, during the process of impregnation, the TiO₂/kaolin composite powders obtained in an attritor mill showed higher photocatalytic activity (for 10%) than those impregnated in a planetary mill. Based on those preliminary photocatalytic activity results, the TiO₂/kaolin composite powders E2A and E2P were additionally characterized and used for the creation of photocatalytic suspensions which were further applied on the chosen mineral supports (28).

Particle size distribution of the most photocatalytic active TiO₂/kaolin composite powders i.e. E2A and E2P is presented in Figure 2.

It is evident that E2A powder (10wt.% TiO₂/attritor mill) possesses a three modal particle size distribution with the following maximal particle diameters: 50nm, 255nm and 1700nm, while the E2P powder (10wt.% TiO₂/ planetary mill) has bimodal particle size distribution with maximal particle diameters of 217 nm and 1138nm. The small fraction of the particles with the average diameter size smaller than 100 nm were obtained only in the case of the impregnation in an attritor mill, which is evidently the reason for the better photocatalytic behavior of this powder in comparison with the powder obtained by the impregnation in a planetary mill. Also, it was noticed that both powders had particles distributed in micro-size range, Figure 3. The micro-sized distribution of the powders could present some difficulties regarding their stability, but since the attempt of this work was to develop a coating with appropriate compatibility to the used mineral substrate, the micro-size of the powder particles did not present a great disadvantage.

The phase composition of both E2A and E2P powders are shown in Figure 4. Mainly, kaolinite phase appears as the dominant phase followed by quartz and Na₂CO₃ used in the milling procedure. The presence of anatase is evident in both samples, while the rutile phase of TiO₂ was not identified. Based on the obtained diffraction patterns, it can be concluded that the different way of TiO₂ impregnation did not modify the initial structure of the clay mineral and of the TiO₂phase.

The values of average capillary water absorption coefficient presented in Table 1 were used in order to estimate the water absorption value of the different mineral substrates by capillary phenomenon for both the reference samples and coated samples. As expected, the CFA-R (sample based on clay material and fly ash) substrate showed the highest value of capillary water absorption coefficient, due to the highest porosity due to the presence of fly ash (26,18). The glass (WG) substrates as non-porous materials, showed a negligible capillary water absorption.

After the deposition of the coating on to the mineral substrates, the average capillary water absorption coefficient decreased for all the examined substrate samples (19). This decrease was low for the CRT sample, but more pronounced for the CFA substrate. No significant difference between the samples coated with deferent suspensions was noticed. The WG substrate samples after the coating procedure still possessed a negligible capillary water absorption coefficient.

The comparison of the surface properties (surface roughness and Vickers micro-hardness), before and after the application of both coatings (E2A / E2P), provided the opportunity to investigate the surface changes which evidently had a great influence on the functional properties of the coated samples.

Surface roughness values (Ra parameter) were used for evaluating the surface roughness of the reference and coated substrates. The values obtained for Ra are presented in Table 2. After the deposition of the coatings, a slight increase of the surface roughness value was noticed for the WG substrates. Evidently, the initial higher roughness of the CFA-Rand CRT-R substrates made an appropriate deposition of the coatings difficult and consequently, the obtained difference of the Ra values between the reference and the coated samples was negligible (18).

The results of the Vickers micro-hardness (HV) test presented in Table 3, indicate an insignificant improvement of the micro-hardness after the application of both coatings to the substrates. This could be mainly attributed to the higher amount of crystalline particles on the coated substrates (19).

The results of the average initial contact values q_ci of the reference substrates (CRT-R, CFA-R and WG-R) and of the coated substrates (CRT-E2A, CRT-E2P, CFA-E2A, CFA-E2P, WG-E2A and WG-E2P) are presented in Table 4. The values of the q_ci of the reference substrates are lower than 90^o, which confirms that all the substrates are hydrophilic. After the deposition of the coatings (E2A/E2P) on the substrates, the contact angle generally decreased increasing the surface hydrophilicity of the substrates (18,19). The value of the q_ci decreased for almost 60% when the WG substrates were coated. It is worth noticing that the contact angle measurements for CRT and WG substrate samples were performed with water as a working liquid, while for the CFA samples they were performed with glycerol. Due to the high values of capillary absorption of the CFA substrate, the applied drops of water instantly penetrated into the substrate so that it was impossible to measure the contact angle values by using water as a working fluid. The coating obtained by the impregnation in a attritor mill showed higher hydrophilicity (lower contact angle) for all the studied substrates than the coating obtained by the impregnation in a planetary mill.

The self-cleaning properties of the coated substrates assessed by the measurement of the initial contact angle q_ci are presented in Figure 5. The evident decrease of the initial contact angle values during UV/VIS irradiation is consistent with other similar studies (18). This fact proved the existence of the self-cleaning effect of the applied coatings (E2A and E2P) on both types of substrates, porous and non-porous.

The results of the evaluation of photocatalytic activity based on the RhB degradation efficiency, with UV/VIS irradiation, are presented in Fig. 6. The obtained results showed a significant photocatalytic activity for the coatings E2A/ E2P applied to the CRT and WG substrates, but a negligible activity for the CFA substrate (only 3.5% for E2P and 5% for E2A after 24h of UV/VIS irradiation). The present phenomenon was directly correlated to the surface properties and porosity of the substrates (Tables 1, 2 and 3). Namely, the high porosity and high capillary water absorption of the CFA substrate enables the entering of the active suspension deep into the material, disabling the activation by light inside the material, leaving only the small remaining amount of the suspension deposited on the surface of the material capable of photocatalytic reacting. The evident low photocatalytic activity of the coatings applied on the CFA substrate made this substrate unsuitable for further investigations regarding the durability tests. The glass substrates (WG) as nonporous samples showed almost the same photocatalytic activity for both coatings (E2A/E2P) during the period of 24 hours. In the case of the CRT substrates, the difference between two types of coatings is evident. Namely, up to 3.5h of UV/VIS irradiation, both coatings applied to this substrate had the same activity. However, after 24h of UV/VIS irradiation, the activity was 42% higher in the case of the coating composed of E2A powder. As standard UNI 11259:2008 (29) and literature reference (30) state (where Rhodamine B is used as a model pollutant), a photocatalytic material can be considered active if its efficiency reaches 20% after 4 hours and 50% after 26h of UV irradiation. According to this, it can be concluded that both coatings (E2A and E2P) on the WG substrate can be considered as photocatalytic materials because after 4h they showed the activity of 48% and 47%, respectively. In addition, after just 24 h they showed the activity of 90 % and 88 %, respectively. The E2A coating deposited on the CRT substrate showed the activity of 13% after 4 hours, but after 24 h the test was stopped. However, after only 24h its efficiency was 43% with the tendency to rise thus proving its good catalytic property.

The results of the modified tape adhesion test presented in Table 5 indicate good adhesion durability due to a low mass loss of the coated substrates in the case of the CRT and CFA substrates. For the WG samples a more significant mass loss was detected than in the case of the two samples specified above.

The comparisons of the surface roughness values, Ra parameter, for the three types of samples before and after the application of the tape adhesion test and water rinsing durability test are presented in Table 2. No significant changes were found, except in the case of the WG coated samples after the rinsing test, where the coating was almost completely removed. Additionally, a negligible decrease of the micro-hardness value after the applied durability tests for both types of suspensions was identified, Table 3.

The results of the θ_ci of the coated CRT, as the function of UV/VIS irradiation time before and after durability (water rinsing and tape adhesion) tests are presented in Figure 7

(a and b). It is evident that there was a significant increase of θ_ci for both E2A and E2P coatings, applied on CRT substrate, after the durability test. The increase of θ_ci was more pronounced after the water rinsing procedure for both coatings had been applied to the CRT substrate, having its maximum at the start of the experiment (without UV/Vis irradiation).

As a result of very low resistance towards rinsing with water of both E2A and E2P coatings applied on WG substrate, (Table 2), the θ_ci and photocatalytic degradation efficiency values were measured only after tape adhesion durability test, Figure 7 (c and d). Obviously, there was a slight increase of θ_ci value after the tape adhesion test (without the influence of UV/Vis irradiation), but during UV/Vis irradiation a significant decrease of θ_ci value was evident.

The increase of the θ_ci value of both CRT and WG coated substrates after the applied durability tests (without the influence of UV/VIS irradiation) could be the result of the physical removal of the deposited coatings and of the changes of the surface morphology during the performing durability tests. The existence of a decreasing trend line of the θ_ci value as the function of UV/VIS irradiation after the durability tests of the analyzed coatings suggests that a considerable self-cleaning efficiency still exists.

The results of photocatalytic degradation efficiency of Rhodamine B with UV/VIS irradiation before and after the appropriate durability tests of the coated CRT and WG substrates are presented in Figure 8. It is evident that up to 2.5h of UV/Vis irradiation, the photocatalytic activity was almost the same before and after the durability tests (water rinsing and tape adhesion tests) for the coated CRT substrates. Nevertheless, by increasing the time of UV/Vis irradiation up to 24h, the photocatalytic activity decreased down to 29% for the CRT-E2A sample after both durability tests. Even more prominent decrease was evident for the CRT-E2P sample, i.e. up to 17,5% after tape adhesion test and 15.5% after water rinsing test. In the case of the WG-E2A and WG-E2P samples, a more significant decrease of photocatalytic activity values was noticed after 24h of UV/Vis irradiation: up to 54% and 47% respectively, after tape adhesion test, Figure 8 (c and d).

According to the presented results, it can be concluded that both coatings exhibited relatively good durability when they were applied to CRT and WG substrates. The coating composed of E2A powder and applied on WG substrate showed better photoactivity after durability tests, suggesting good compatibility and stability with this substrate, which is in good correlation with other reported studies (29,30).