Suppressing the Photocatalytic Activity of TiO2 Nanoparticles by Extremely Thin Al2O3 Films Grown by Gas-Phase Deposition at Ambient Conditions

This work investigated the suppression of photocatalytic activity of titanium dioxide (TiO2) pigment powders by extremely thin aluminum oxide (Al2O3) films deposited via an atomic-layer-deposition-type process using trimethylaluminum (TMA) and H2O as precursors. The deposition was performed on multiple grams of TiO2 powder at room temperature and atmospheric pressure in a fluidized bed reactor, resulting in the growth of uniform and conformal Al2O3 films with thickness control at sub-nanometer level. The as-deposited Al2O3 films exhibited excellent photocatalytic suppression ability. Accordingly, an Al2O3 layer with a thickness of 1 nm could efficiently suppress the photocatalytic activities of rutile, anatase, and P25 TiO2 nanoparticles without affecting their bulk optical properties. In addition, the influence of high-temperature annealing on the properties of the Al2O3 layers was investigated, revealing the possibility of achieving porous Al2O3 layers. Our approach demonstrated a fast, efficient, and simple route to coating Al2O3 films on TiO2 pigment powders at the multigram scale, and showed great potential for large-scale production development.


Introduction
The brilliant white color and high photostability of nanoparticulate titanium dioxide (TiO 2 ) make it an excellent white pigment in the paint, plastic, and paper industries [1]. In particular, TiO 2 has been widely used as a white pigment in oil paint since the 20th century, replacing lead white-the most important white pigment in history [2]. However, the high photocatalytic activity of TiO 2 under UV light irradiation causes the inevitable degradation of surrounding materials, which consequently changes the appearance and severely decreases the lifetime of the paintings. Depending on the surrounding materials, the photocatalytic degradation can occur via different regimes [3]. For instance, photocatalytic reactions can either promote polymerization of the organic binder, creating cross-linking that leads to embrittlement [4], or decompose the binder into volatile organic components, resulting the mitigation of photocatalytic activity of TiO 2 by thin Al 2 O 3 films deposited by ALD was also investigated. Hakim et al. demonstrated that an Al 2 O 3 layer obtained for 50 ALD cycles at 170 • C (GPC of 0.2 nm) could suppress the high photocatalytic activity of P25 TiO 2 [40]. It is worth noting that, for pigment application, the thickness of the coating layer is of crucial importance: the layer must be thick enough to efficiently suppress the photocatalytic activity of TiO 2 , but thin enough to conserve the gloss and brightness of TiO 2 . This requires an optimal thickness, which is normally in the range of a few nanometers [17]. Therefore, reducing the coating thickness while ensuring its photocatalytic suppression ability is highly desirable.
This work reports on the suppression of photocatalytic activity of various types of TiO 2 powders (i.e., anatase, rutile, and P25 TiO 2 ) by Al 2 O 3 films deposited via an ALD-like process using TMA and H 2 O as precursors, which was carried out at room temperature and atmospheric pressure in a home-built fluidized bed reactor. This gas-phase deposition process enabled the control of coating thickness at the sub-nanometer level, allowing us to investigate the dependence of photocatalytic suppression ability of Al 2 O 3 on the film thickness. The surface morphology and film thickness, composition, and crystallinity of the coating layers were characterized by transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS) and thermogravimetric analysis (TGA), and X-ray diffraction spectroscopy (XRD), respectively. The results showed that highly conformal Al 2 O 3 films with a thickness as thin as 1 nm were obtained, which efficiently suppressed the photocatalytic activity of TiO 2 powders without affecting their bulk optical properties. Furthermore, the influence of high-temperature annealing on the properties of the Al 2 O 3 layers was investigated, revealing the possibility to achieve porous Al 2 O 3 layers, which could be useful for other applications.

Reaction Mechanism of Al 2 O 3 ALD Using TMA and H 2 O: A Brief Overview
The high reactivity of TMA facilitates the deposition of Al 2 O 3 in a broad range of temperature on various types of substrates and materials with any geometries, including flat surfaces [21], high-aspect-ratio structures [41][42][43], porous media [44][45][46][47], nanoparticles [29,32,34,40], fibers [48], carbon nanotubes [49,50], graphene [51][52][53], polymers [54], and biomaterials [55]. The reaction mechanism in ALD of Al 2 O 3 using TMA and H 2 O has been intensively investigated in the past decades, both theoretically and experimentally [21,[56][57][58][59][60][61][62][63][64][65][66][67]. Accordingly, the surface chemical reactions that lead to the deposition of Al 2 O 3 in the ideal case can be divided into two half-reactions. During the exposure to TMA, the first half-reaction takes place and proceeds as: where the -OH represents the functional groups (i.e., hydroxyl groups), which are formed during the exposure of the substrate to air, or surface pretreatment. After all of the -OH groups have reacted with TMA, the reactions achieve saturation, forming a -CH 3 terminated surface, which is ready for the second half-reactions to take place during the exposure to H 2 O, described as: These reactions convert the -CH 3 into CH 4 and create a new surface terminated by -OH groups, which serves two purposes: blocking further reactions with H 2 O, which causes saturation, and providing an active surface for the chemical reactions in the next cycle. A typical GPC in the range of 0.1-0.3 nm is obtained for TMA/H 2 O ALD depending on experimental conditions such as temperature range, pressure range, and reactor type. The sequential exposures are repeated to deposit Al 2 O 3 films, which translates into ability to control the film thickness at atomic level.
Nevertheless, the TMA/H 2 O ALD chemistry is rather complex, which involves-apart from the ligand-exchange reactions described above-dissociation reactions that can take place in both the half-reactions [21,68]. In the first half-reaction, TMA can react with unsaturated Al-O pairs of Al 2 O 3 layers, creating a -CH 3 terminated surface described as [68]: In the second half-reaction, dissociation reactions can occur between H 2 O and the oxygen on Al 2 O 3 surface, forming -OH surface functional groups, described as [21]: The reversed reaction of the dissociation, the dehydroxylation, can also take place that consequently reduces the concentration of the OH groups, especially at high temperatures [21]: Although the dissociation reactions are not often mentioned in the literature when discussing the surface chemistry in TMA/H 2 O ALD, these reactions are believed to contribute considerably to growth. Especially, the dissociative chemisorption of TMA in the first half-cycle is the key mechanism that is used to interpret the nucleation on hydroxyl-free surfaces [59,60,[69][70][71][72], even at room temperature [70]. In addition, at high deposition temperatures, the desorption of -OH groups and decomposition of TMA can occur simultaneously, which will strongly affect the growth rate of Al 2 O 3 ALD [57,73]. Nevertheless, the ligand-exchange reactions are considered the dominant reactions during the deposition and have been most studied.
The GPC in ALD regime is generally temperature-independent, which is commonly known as "ALD window" [22]. However, ALD of Al 2 O 3 using TMA/H 2 O is found to be temperature-dependent [57,61,63,65]. Rahtu et al. [57] demonstrated a slight increase of GPC with increasing temperature from 150 to 250 • C, followed by a drop of GPC with further increase of temperature. The lower GPC at low temperature was attributed to the slow kinetics of the chemical reactions between the precursors, whereas the desorption of hydroxyl groups at high temperatures led to the drop of GPC. Most recently, the decrease of GPC at low deposition temperatures has been thoroughly investigated by Vandalon and Kessels [65]. Using in-situ vibrational sum-frequency generation technique, the authors investigated the reaction mechanism between TMA and H 2 O in a broad temperature range (i.e., 100-300 • C), which revealed that the low GPC is caused by the incomplete removal of -CH 3 groups by H 2 O at low temperature. The unreacted -CH 3 groups decline the chemisorption of TMA in the subsequent cycles, which consequently decreases the growth rate. The study also demonstrate that the persistent -CH 3 groups are not accumulated with increasing the number of cycles, which explains the low carbon impurity in Al 2 O 3 layers grown by ALD at low temperatures [65]. Nevertheless, the experimental data obtained by Ylivaara et al. have demonstrated that the low-temperature deposition of Al 2 O 3 using TMA and H 2 O is inherently associated with a considerable amount of impurities, especially hydrogen and carbon, which increase with decreasing deposition temperature [74]. The hydrogen impurity arises from the unreacted hydroxyl groups, which has recently been verified by Guerra-Nunez et al. [66]. The degree of the GPC decrease at low temperatures is not well determined, and strongly depends on experimental conditions. For instance, contrary to the considerable drop of GPC at low temperatures observed by Vandalon and Kessels (i.e., the growth rate drops by approximately 50% when reducing the temperature from 250 to 100 • C), Groner et al. previously showed a slight decrease of GPC with decreasing temperature, even to near room temperature (i.e., 33 • C) [61]. The discrepancy between the studies can be attributed to the different ALD conditions, especially the process pressure range and reactor types.

Morphology of Al 2 O 3 -Coated TiO 2
We will first focus on the coating and performance of anatase TiO 2 powders. The morphology of the Al 2 O 3 films on anatase TiO 2 for different exposure cycles is shown in Figure 1, demonstrating the deposition of uniform Al 2 O 3 films, even at a film thickness of about 1 nm (Figure 1a). The coating thickness increases linearly with the number of cycles, with a GPC of approximately 0.3 nm (Figure 1d). The obtained GPC is higher than the GPC reported for ALD of Al 2 O 3 on flat substrates, which is typically in the range of 0.1-0.2 nm [61,65]. However, this GPC value is lower than the GPC obtained for the Al 2 O 3 ALD on particles (i.e., 0.5 nm) reported by Liang et al. [75]. Nevertheless, previous work from our group demonstrated that the growth of Al 2 O 3 at room temperature and atmospheric pressure in FBR follows the chemical vapor deposition (CVD) mode, in which the GPC increases with increasing dosing time (i.e., without a self-saturating regime) [34]. With a GPC of 0.3 nm obtained for the examined conditions, a controlled deposition at sub-nanometer level of highly conformal Al 2 O 3 films is totally achievable. In fact, this higher GPC is beneficial for the fast and scalable production that is required for practical applications. increases with increasing dosing time (i.e., without a self-saturating regime) [34]. With a GPC of 0.3 nm obtained for the examined conditions, a controlled deposition at sub-nanometer level of highly conformal Al2O3 films is totally achievable. In fact, this higher GPC is beneficial for the fast and scalable production that is required for practical applications.  Figure 2c). The deconvolution of O 1s spectrum revealed the presence of a considerable amount of -OH groups (BE = 530.8 eV) and physisorbed H2O (BE = 532.3 eV) [76,77]. The physisorbed H2O can be removed by applying a heat treatment in air at 170 °C for 1 h, as indicated by Figure 2b. We note that no distinguishable difference was observed for the growth of the Al2O3 on the TiO2 without and with heat treatment. However, the existence of -OH groups on the surface is important to the inception of TMA chemisorption. The C 1s spectrum exhibits different states of C contamination (Figure 2d [76,77]. The physisorbed H 2 O can be removed by applying a heat treatment in air at 170 • C for 1 h, as indicated by Figure 2b. We note that no distinguishable difference was observed for the growth of the Al 2 O 3 on the TiO 2 without and with heat treatment. However, the existence of -OH groups on the surface is important to the inception of TMA chemisorption. The C 1s spectrum exhibits different states of C contamination (Figure 2d), including C=O (BE = 288.6 eV), O-C (BE = 286.2 eV) and C=C (BE = 284.7 eV) [76,78]. These carbon compounds could arise from adsorbed species on the surface of the powders and/or the carbon tape used to immobilize the TiO 2 particles.
2.3 eV) [76,77]. The physisorbed H2O can be removed by applying a heat treatment or 1 h, as indicated by Figure 2b. We note that no distinguishable difference was obse th of the Al2O3 on the TiO2 without and with heat treatment. However, the existence n the surface is important to the inception of TMA chemisorption. The C 1s spectrum t states of C contamination (Figure 2d), including C=O (BE = 288.6 eV), O-C (BE = 2 (BE = 284.7 eV) [76,78]. These carbon compounds could arise from adsorbed specie of the powders and/or the carbon tape used to immobilize the TiO2 particles.  [78][79][80]. No noticeable change was observed for the BE of Al 2p with increasing coating thickness. However, the peaks of TiO 2 (i.e., BE = 529.4 eV for O 1s, BE = 463.9 eV for Ti 2p 1/2 , and BE = 458.2 eV for Ti 2p 3/2 ) are gradually attenuated with increasing Al 2 O 3 thickness. For the TiO 2 coated with 5 nm Al 2 O 3 , the TiO 2 peaks are almost vanished, which suggests that the photoelectrons of TiO 2 excited by the X-rays are effectively shielded by the Al 2 O 3 film. The disappearance of TiO 2 features in XPS spectra is evidence of uniform Al 2 O 3 coating on a large scale, in addition to the evidence provided by the TEM micrographs shown in Figure 1. Furthermore, the gradual attenuation of C 1s at BE = 288.6 eV (C=O) and BE = 286.2 eV (O-C) relatively compared to the peak at BE = 284.7 eV (C=C) with increasing Al 2 O 3 thickness can be explained that these carbon-containing species are on the surface of TiO 2 . This, however, does not rule out the possibility that the carbon content is accumulating in the growing film, which cannot be avoided due to the deposition at room temperature. d that these carbon-containing species are on the surface of TiO2. This, however, d the possibility that the carbon content is accumulating in the growing film, which ca due to the deposition at room temperature. ure 3. XPS spectra of anatase TiO2 coated with 3 nm thick (a,c,e,g-on the left side) and 5 nm th ,f,h-on the right side) Al2O3 films. a photocatalytic suppression layer for pigment applications, the transparency of the hich is crucially important to maintain the bulk optical properties of TiO2 such as th d high brightness, is highly desirable. The optical absorption spectra shown in F that the absorption of TiO2 remained unaffected upon the coating with Al2O3 with d knesses. From the absorption spectra, an optical bandgap of 3.2 eV was determined onds to that of anatase TiO2 [81]. The characterization using X-ray diffraction sho ous state of the Al2O3 layers, even after annealing at 500 °C for 12 h ( Figure S1). As a photocatalytic suppression layer for pigment applications, the transparency of the coating layer, which is crucially important to maintain the bulk optical properties of TiO 2 such as the white color and high brightness, is highly desirable. The optical absorption spectra shown in Figure 4 indicate that the absorption of TiO 2 remained unaffected upon the coating with Al 2 O 3 with different film thicknesses. From the absorption spectra, an optical bandgap of 3.2 eV was determined, which corresponds to that of anatase TiO 2 [81]. The characterization using X-ray diffraction showed the amorphous state of the Al 2 O 3 layers, even after annealing at 500 • C for 12 h ( Figure S1).  shows TGA plots of the TiO2 powders coated with Al2O3 films with different thicknesses. The first stage (i.e., T ≤ 120 °C) describes the desorption of water when the powders are heated from 25 °C, showing an increasing amount of water with Al2O3 thickness. The second stage demonstrates the desorption of hydroxyl groups, which increases remarkably with increasing the coating thickness. Following the calculation method proposed by Pratsinis et al., from the TGA plots shown in Figure 5a, the density of -OH groups was estimated by following the common assumption that all the -OH groups are on the surface [82,83]. For uncoated TiO2, a density of 4 -OH/nm 2 was obtained, which is close to the value reported for anatase TiO2 [83]. On Al2O3-coated TiO2, the density of -OH groups increased linearly with Al2O3 thickness (Figure 5b): a density of 9.6 -OH/nm 2 was obtained for the 1 nm thick Al2O3 coated powder, and increased to 95.3 -OH/nm 2 for the TiO2 coated with 5 nm Al2O3. As this number is not reasonable for the assumption that all the -OH groups are on the surface of the powder (i.e., it is theoretically and practically impossible to have 95 OH -groups on 1 nm 2 ), the calculated values suggest that most of the -OH groups are located inside the coating layer. This is caused by the incomplete consumption of -OH groups, which leads to hydrogen impurity in the film [66]. The existence of -OH strongly affects the density of Al2O3 [61]. Nevertheless, Groner et al. [61] demonstrated that the Al2O3 films grown at low temperatures exhibited excellent electrical properties despite containing high -OH concentrations. As we will demonstrate later, despite having high densities of -OH groups, ultrathin Al2O3 films deposited at room temperature can provide excellent photocatalytic suppression ability.  Figure 5a, the density of -OH groups was estimated by following the common assumption that all the -OH groups are on the surface [82,83]. For uncoated TiO 2 , a density of 4 -OH/nm 2 was obtained, which is close to the value reported for anatase TiO 2 [83]. On Al 2 O 3 -coated TiO 2 , the density of -OH groups increased linearly with Al 2 O 3 thickness (Figure 5b): a density of 9.6 -OH/nm 2 was obtained for the 1 nm thick Al 2 O 3 coated powder, and increased to 95.3 -OH/nm 2 for the TiO 2 coated with 5 nm Al 2 O 3 . As this number is not reasonable for the assumption that all the -OH groups are on the surface of the powder (i.e., it is theoretically and practically impossible to have 95 OH -groups on 1 nm 2 ), the calculated values suggest that most of the -OH groups are located inside the coating layer. This is caused by the incomplete consumption of -OH groups, which leads to hydrogen impurity in the film [66]. The existence of -OH strongly affects the density of Al 2 O 3 [61]. Nevertheless, Groner et al. [61] demonstrated that the Al 2 O 3 films grown at low temperatures exhibited excellent electrical properties despite containing high -OH concentrations. As we will demonstrate later, despite having high densities of -OH groups, ultrathin Al 2 O 3 films deposited at room temperature can provide excellent photocatalytic suppression ability.

Dependence of Photocatalytic Suppression Ability of Al2O3 on Film Thickness
The photocatalytic suppression ability of the room-temperature grown Al2O3 was investigated by the photodegradation of rhodamine B (RhB) solution under the irradiation of sunlight generated by a solar simulator. The ability to control the growth at sub-nanometer level allowed us to study the interdependence of the photocatalytic suppression ability and the thickness of Al2O3. Figure 6 shows the photocatalytic activity toward the photodegradation of RhB of TiO2 powders coated with Al2O3 films with different thicknesses. For comparison, the self-degradation of RhB (i.e., without TiO2) caused by the UV light and the photocatalytic activity of uncoated TiO2 were also investigated. Prior to the UV irradiation, the solution was continuously stirred in the dark (i.e., light-off stage) for 30 min to obtain adsorption/desorption equilibrium of RhB and uniform suspensions, which were collected after certain time-intervals to determine the concentration of the residual RhB. The results showed that during the light-off stage, the concentration of RhB in the solution without TiO2 powders (blank RhB) remained unchanged. However, a small drop of RhB concentration was observed for the solutions with TiO2 powders (both uncoated and coated with Al2O3). This drop is due to the adsorption of a fraction of RhB molecules on the surface of the particles. Shortly after that (i.e., after 2 min) the concentration of RhB reached the steady state. In the light-on stage, a rapid decomposition of RhB in the solution containing uncoated TiO2 was observed (the circles in Figure 6a), and after 25 min of irradiation, the RhB was completely decomposed, indicative of the high photocatalytic activity of the TiO2.

Dependence of Photocatalytic Suppression Ability of Al 2 O 3 on Film Thickness
The photocatalytic suppression ability of the room-temperature grown Al 2 O 3 was investigated by the photodegradation of rhodamine B (RhB) solution under the irradiation of sunlight generated by a solar simulator. The ability to control the growth at sub-nanometer level allowed us to study the interdependence of the photocatalytic suppression ability and the thickness of Al 2 O 3 . Figure 6 shows the photocatalytic activity toward the photodegradation of RhB of TiO 2 powders coated with Al 2 O 3 films with different thicknesses. For comparison, the self-degradation of RhB (i.e., without TiO 2 ) caused by the UV light and the photocatalytic activity of uncoated TiO 2 were also investigated. Prior to the UV irradiation, the solution was continuously stirred in the dark (i.e., light-off stage) for 30 min to obtain adsorption/desorption equilibrium of RhB and uniform suspensions, which were collected after certain time-intervals to determine the concentration of the residual RhB. The results showed that during the light-off stage, the concentration of RhB in the solution without TiO 2 powders (blank RhB) remained unchanged. However, a small drop of RhB concentration was observed for the solutions with TiO 2 powders (both uncoated and coated with Al 2 O 3 ). This drop is due to the adsorption of a fraction of RhB molecules on the surface of the particles. Shortly after that (i.e., after 2 min) the concentration of RhB reached the steady state. In the light-on stage, a rapid decomposition of RhB in the solution containing uncoated TiO 2 was observed (the circles in Figure 6a For the Al2O3-coated TiO2, the photocatalytic activity was reduced, which exhibited a strong dependence on the thickness of the Al2O3 layer. This thickness dependence can be quantitatively estimated by the kinetics of the photodegradation reaction, which can be described by first-order kinetics [84]: where C0 and C are the initial concentration and the concentration at the time t, respectively, kapp is the apparent first-order kinetic constant that represents the reaction rate. The kinetic plots describing the degradation of RhB are shown in Figure 6b, from which the kapp value for each reaction was extracted and presented in Table 1. Accordingly, the photocatalytic activity of TiO2 (kapp ≈ 160 × 10 −3 min −1 ) was reduced approximately 18 times (to kapp ≈ 8.87 × 10 −3 min −1 ) by coating an Al2O3 layer for 1 cycle of TMA and H2O exposures (i.e., the calculated thickness of about 0.3 nm). Nevertheless, the Al2O3 at this thickness cannot entirely suppress the photocatalytic activity of TiO2. This could be due to (1) the insufficient thickness that can block the transport of photogenerated charges from the TiO2 to the surface, and/or (2) the lack of continuity of the coating layer. With increasing film thickness, kapp value decreased rapidly, which dropped to far below 1 × 10 −3 min −1 for Al2O3 films with a thickness of 1 nm or thicker (thickness a ≥ 1 nm). The results showed that an Al2O3 layer with a thickness of 1 nm efficiently suppressed entirely the catalytic activity of the TiO2. The degradation of RhB observed for the TiO2 coated with Al2O3 films with thickness a ≥ 1 nm is identical to that of the self-degradation of RhB solution without the powder (Figure 6b and Table 1). A comparison with other coating processes and materials that have been used to mitigate the photocatalytic activity of TiO2 pigments is given in Table 2. It is can be seen that gas-phase deposition methods such as ALD and CVD have been employed. However, the reported processes were carried out at much higher temperatures, i.e., above 150 °C for Al2O3 and 175 °C for SiO2. In addition, thicker layers are needed to efficiently suppress the photocatalytic activity of TiO2, which is also the case for the layer obtained by wetchemistry methods. This indicates that our approach has significant advantages in providing a fast and efficient coating method that can be carried out at room temperature, which increases the feasibility for further development to large-scale production. For the Al 2 O 3 -coated TiO 2 , the photocatalytic activity was reduced, which exhibited a strong dependence on the thickness of the Al 2 O 3 layer. This thickness dependence can be quantitatively estimated by the kinetics of the photodegradation reaction, which can be described by first-order kinetics [84]: ln(C 0 /C) = k app ·t, or C = C 0 ·exp(−k app ·t), (6) where C 0 and C are the initial concentration and the concentration at the time t, respectively, k app is the apparent first-order kinetic constant that represents the reaction rate. The kinetic plots describing the degradation of RhB are shown in Figure 6b, from which the k app value for each reaction was extracted and presented in Table 1. Accordingly, the photocatalytic activity of TiO 2 (k app ≈ 160 × 10 −3 min −1 ) was reduced approximately 18 times (to k app ≈ 8.87 × 10 −3 min −1 ) by coating an Al 2 O 3 layer for 1 cycle of TMA and H 2 O exposures (i.e., the calculated thickness of about 0.3 nm). Nevertheless, the Al 2 O 3 at this thickness cannot entirely suppress the photocatalytic activity of TiO 2 . This could be due to (1) the insufficient thickness that can block the transport of photogenerated charges from the TiO 2 to the surface, and/or (2) the lack of continuity of the coating layer. With increasing film thickness, k app value decreased rapidly, which dropped to far below 1 × 10 −3 min −1 for Al 2 O 3 films with a thickness of 1 nm or thicker (thickness a ≥ 1 nm). The results showed that an Al 2 O 3 layer with a thickness of 1 nm efficiently suppressed entirely the catalytic activity of the TiO 2 . The degradation of RhB observed for the TiO 2 coated with Al 2 O 3 films with thickness a ≥ 1 nm is identical to that of the self-degradation of RhB solution without the powder (Figure 6b and Table 1). A comparison with other coating processes and materials that have been used to mitigate the photocatalytic activity of TiO 2 pigments is given in Table 2. It is can be seen that gas-phase deposition methods such as ALD and CVD have been employed. However, the reported processes were carried out at much higher temperatures, i.e., above 150 • C for Al 2 O 3 and 175 • C for SiO 2 . In addition, thicker layers are needed to efficiently suppress the photocatalytic activity of TiO 2 , which is also the case for the layer obtained by wet-chemistry methods. This indicates that our approach has significant advantages in providing a fast and efficient coating method that can be carried out at room temperature, which increases the feasibility for further development to large-scale production. As demonstrated by the thermal analysis shown in Figure 5, the room-temperature-grown Al 2 O 3 films contain a high density of -OH groups, which can desorb at high temperatures. Therefore, a calcination at 500 • C for 12 h was applied to investigate the influence of the -OH desorption on the suppression ability of Al 2 O 3 . The results show that the calcination of Al 2 O 3 -coated TiO 2 resulted in an enhanced photocatalytic activity ( Figure S2), in which k app values increased nearly an order of magnitude (Table 1). This indicates that the -OH desorption reduced the suppression ability of the Al 2 O 3 layer. We speculate that the desorption of -OH groups resulted in the formation of porous Al 2 O 3 that allows photogenerated electrons to transport to the surface. In addition, the densification, which my introduce cracks during calcination, can also take place and increase the porosity of the Al 2 O 3 [74]. This is supported by the XPS spectra of the calcined Al 2 O 3 /TiO 2 shown in Figure S3. Accordingly, the two XPS peaks of the Ti 2p core-levels that were attenuated after coating with a 5-nm Al 2 O 3 film were partially recovered after the calcination, which suggests that the photoelectrons generated by the X-rays could travel through the calcined Al 2 O 3 layer. For uncoated TiO 2 , the results show that the -OH desorption reduced the photocatalytic activity ( Figure S2), indicated by the drop of k app from 160 × 10 −3 to 100 × 10 −3 min −1 (Table 1). This can be explained by the fact that in TiO 2 photocatalysis, hydroxyl groups act as hole traps, which causes two important effects that enhance the photocatalytic activity: enhancement of charge separation and formation of hydroxyl radicals [76,[88][89][90].

Photocatalytic Suppression Ability of Al 2 O 3 on P25 and Rutile TiO 2
Analogously to the deposition on anatase TiO 2 , uniform Al 2 O 3 films were also achieved on rutile and P25 TiO 2 powders ( Figures S4 and S5), which enabled the study on photocatalytic suppression ability of the Al 2 O 3 films on P25 and rutile TiO 2 nanoparticles. The results are demonstrated in Figure 7. The photocatalytic activity of rutile TiO 2 (k app ≈ 4.85 × 10 −3 min −1 ) is much lower than that of anatase TiO 2 (k app ≈ 159.88 × 10 −3 min −1 ) under identical conditions, which is also known from the literature [90,91]. After coating the rutile particles with 1 nm Al 2 O 3 , the photocatalytic activity is strongly suppressed, which is confirmed by the low k app value obtained (0.28 × 10 −3 min −1 ). In contrast, the photocatalytic activity of P25 TiO 2 (k app ≈ 231 × 10 −3 min −1 ) was found to be higher than that of anatase TiO 2 . Nevertheless, by coating with an Al 2 O 3 layer with a thickness of approximately 1 nm, the photocatalytic activity of P25 was also strongly suppressed (k app ≈ 2.01 × 10 −3 min −1 ). These results have further demonstrated the high photocatalytic suppression ability of the ultrathin Al 2 O 3 films deposited under mild conditions. nm Al2O3 film were partially recovered after the calcination, which suggests that the photoelectrons generated by the X-rays could travel through the calcined Al2O3 layer. For uncoated TiO2, the results show that the -OH desorption reduced the photocatalytic activity ( Figure S2), indicated by the drop of kapp from 160 × 10 −3 to 100 × 10 −3 min −1 (Table 1). This can be explained by the fact that in TiO2 photocatalysis, hydroxyl groups act as hole traps, which causes two important effects that enhance the photocatalytic activity: enhancement of charge separation and formation of hydroxyl radicals [76,[88][89][90].

Photocatalytic Suppression Ability of Al2O3 on P25 and Rutile TiO2
Analogously to the deposition on anatase TiO2, uniform Al2O3 films were also achieved on rutile and P25 TiO2 powders ( Figures S4 and S5), which enabled the study on photocatalytic suppression ability of the Al2O3 films on P25 and rutile TiO2 nanoparticles. The results are demonstrated in Figure  7. The photocatalytic activity of rutile TiO2 (kapp ≈ 4.85 × 10 −3 min −1 ) is much lower than that of anatase TiO2 (kapp ≈ 159.88 × 10 −3 min −1 ) under identical conditions, which is also known from the literature [90,91]. After coating the rutile particles with 1 nm Al2O3, the photocatalytic activity is strongly suppressed, which is confirmed by the low kapp value obtained (0.28 × 10 −3 min −1 ). In contrast, the photocatalytic activity of P25 TiO2 (kapp ≈ 231 × 10 −3 min −1 ) was found to be higher than that of anatase TiO2. Nevertheless, by coating with an Al2O3 layer with a thickness of approximately 1 nm, the photocatalytic activity of P25 was also strongly suppressed (kapp ≈ 2.01 × 10 −3 min −1 ). These results have further demonstrated the high photocatalytic suppression ability of the ultrathin Al2O3 films deposited under mild conditions.

Experimental Section
The deposition was carried out in a home-built fluidized bed reactor operating at atmospheric pressure and room temperature, as described elsewhere [34]. Briefly, the system consisted of a glass

Experimental Section
The deposition was carried out in a home-built fluidized bed reactor operating at atmospheric pressure and room temperature, as described elsewhere [34]. Briefly, the system consisted of a glass column (26 mm in internal diameter and 500 mm in height) placed on top of a single motor Paja PTL 40/40-24 vertical vibration table to assist the fluidization of nanoparticles. Anatase (mean diameter d ≈ 270 nm) and rutile (mean diameter d ≈ 300 nm) TiO 2 pigment powders were provided by Taihai TiO 2 Pigment Co. (Panzhihua, China); aeroxide P25 TiO 2 (mean diameter d ≈ 21 nm) were purchased from Evonik Industries (Hanau, Germany). Semiconductor grade trimethylaluminum (TMA) was provided by Akzo Nobel HPMO (Amersfoort, The Netherlands) in a 600 mL WW-600 stainless steel bubbler. Both the TiO 2 powders and the TMA precursor were used as received. Pressurized nitrogen (99.999 vol %) was used as the carrier gas. For each experiment on anatase and rutile TiO 2 , 7.0 g of powders was used. An optimized N 2 gas flow of 0.5 L min −1 was introduced through the distributor plate placed at the bottom of the glass column to fluidize the powders. A coating cycle consisted of alternating exposures of the powders to TMA precursor (2 min) and deionized water vapor (2 min), separated by a purging step (5 min) using N 2 . The coating on P25 followed the conditions described elsewhere [34]. The temperature inside the bed was monitored by a type-K thermocouple inserted in the column, which showed a small variation during the deposition (i.e., 27 ± 3 • C), possibly due to the heat release from the chemical reactions.
As-coated Al 2 O 3 /TiO 2 powders were suspended in ethanol and transferred to regular transmission electron microscopy (TEM) grids (3.05 mm in diameter). TEM images were taken at several locations on the grids using a JEOL JEM1400 transmission electron microscope (JEOL, Peabody, MA, USA) operating at a voltage of 120 kV and a current density of 50 pA cm −2 .
X-ray photoelectron (XPS) characterizations were carried out using a ThermoFisher K-Alpha system (ThermoFisher Scientific, Waltham, MA, USA) using Al Kα radiation with photon energy of 1486.7 eV. A sufficient amount of powder was immobilized on carbon tape before loading into the XPS chamber. Survey scans were acquired using a 200 µm spot size, 55 eV pass energy and 0.1 eV/step with charge neutralization. The peaks positions were calibrated according to the C 1s peak at 284.7 eV.
A Mettler Toledo TGA/SDTA 851e thermogravimetric analyzer (Mettler Toledo B.V., Tiel, The Netherlands) was used for studying the thermal behavior of the synthesized powders. An amount of 30 mg of powders was used for each TGA measurement. The TGA curves were recorded while (1) heating up the powders from 25 to T 1 = 120 • C with a ramping rate of 10 • C min −1 in a N 2 flow of 100 mL min −1 ; (2) maintaining at 120 • C for 10 min; (3) ramping up to 500 • C with a ramping rate of 20 • C min −1 ; and (4) finally maintaining at this temperature for 10 min. Steps (1) and (3) are commonly used to determine the density of physisorbed water and chemisorbed hydroxyl groups on the surface, respectively [83].
The Al 2 O 3 /TiO 2 was transferred onto a Si wafer with 300 nm of SiO 2 thermal oxide, which was to eliminate the influence of the substrate (Si) signal in the X-ray diffractograms of the powders. The X-ray diffractograms were obtained by a PANalytical X'pert Pro diffractometer (PANalytical, Almelo, The Netherlands) with Cu Kα radiation, secondary flat crystal monochromator and X'celerator RTMS Detector system. The angle of interest 2θ was measured from 10 • to 90 • with fine steps of 0.001 • .
Optical absorption measurements were performed using a PerkinElmer-Lambda 900 spectrometer (PerkinElmer, Spokane, WA, USA) equipped with an integrated sphere device. The powder was suspended in ethanol, which was transferred onto a quartz substrate and dried in air. The spectra were acquired in the wavelength range of 600-200 nm, with fine steps of 1 nm.
The photocatalytic activity of the Al 2 O 3 -coated TiO 2 powders was evaluated by the photodegradation of RhB solution. For each test, 30 mg of the powders were added to 30 mL RhB solution (concentration of 8 mg L −1 ) and continuously stirred in the dark for 30 min to obtain a uniform suspension. Thereafter, the suspension was exposed to UV radiation generated by a mercury lamp with a power of 45 W for different exposure times. The set-up allowed us to assess up to 10 samples simultaneously, which ensured that all samples were irradiated under the same conditions, such as light intensity, exposure time, and temperature. The suspension was then centrifuged to separate the powders from the solution. Finally, the solution was analyzed by UV-visible spectrophotometry (HACH LANGE DR5000 UV-vis spectrometer, Hach-Lange GmbH, Düsseldorf, Germany) to determine the residual concentration of the RhB in solution, which was used to evaluate the catalytic activity suppression of the Al 2 O 3 layers.

Conclusions
We have demonstrated the deposition and investigated the photocatalytic suppression ability of ultrathin Al 2 O 3 films on different types of TiO 2 powders: anatase (mean diameter d ≈ 270 nm), rutile (mean diameter d ≈ 300 nm) and P25 (mean diameter d ≈ 21 nm). The deposition was carried out using an ALD-like process, in which the TiO 2 powders were alternatively exposed to TMA and H 2 O at room temperature and atmospheric pressure in a FBR. This enabled the deposition of ultrathin, conformal and uniform Al 2 O 3 films with the thickness control at sub-nanometer level. The deposition at room temperature resulted in amorphous Al 2 O 3 layers containing a high concentration of hydroxyl groups, which was caused by the incomplete reactions between the precursors at low temperature. Nevertheless, the films exhibited excellent photocatalytic suppression ability, which showed that an Al 2 O 3 layer with a thickness of 1 nm could efficiently suppress the photocatalytic activities of the TiO 2 powders. In addition, the optical absorption of TiO 2 was not affected by the Al 2 O 3 coating. Upon calcination at a high temperature, photocatalytic suppression ability of the Al 2 O 3 was decreased, possibly due to the formation of porous Al 2 O 3 layer, which created pathways for charge carriers to transport to the surface. Our approach of using a FBR operating at atmospheric pressure is a fast, efficient, simple method for depositing ultrathin conformal Al 2 O 3 films that meet the requirements for coating pigments, which can be further developed for large-scale production.
Supplementary Materials: The following are available online at www.mdpi.com/2079-4991/8/2/61/s1, Figure  S1: XRD patterns of uncoated TiO 2 and Al 2 O 3 -coated TiO 2 , with or without annealing at 500 • C for 12 h. Figure  S2: The comparison of photocatalytic activity of Al 2 O 3 /TiO 2 (anatase) powders before and after annealing at 500 • C for 12 h. Figure S3: XPS spectra of C 1s, Al 2p, O 1s and Ti 2p core-levels of anatase TiO 2 coated with 5 nm Al 2 O 3 as-deposited and after annealing at 500 • C for 12 h in air. Figure S4: TEM micrographs of rutile TiO 2 coated with Al 2 O 3 films deposited for 3 cycles, 5 cycles, 10 cycles and 17 cycles. Figure S5: High-resolution TEM micrographs of P25 TiO 2 nanoparticles coated with Al 2 O 3 for 4 cycles, 7 cycles and 15 cycles.