Synthesis of TiO₂/SBA-15 Nanocomposites by Hydrolysis of Organometallic Ti Precursors for Photocatalytic NO Abatement

Abstract

The development of advanced photocatalysts for air pollution removal is essential to improve indoor air quality. TiO₂/mesoporous silica SBA-15 nanocomposites were synthesized using an organometallic decoration method, which leverages the high reactivity of Ti precursors to be hydrolyzed on the surface water groups of silica supports. Both lab-made Ti(III) amidinate and commercial Ti(IV) amino precursors were utilized to react with water-rich SBA-15, obtained through a hydration process. The hydrated SBA-15 and the TiO₂/SBA-15 nanocomposites were characterized using TGA, FTIR, ¹H and ²⁹Si NMR, TEM, SEM, N₂ physisorption, XRD, and WAXS. This one-step TiO₂ decoration method achieved a loading of up to 51.5 wt.% of approximately 9 nm anatase particles on the SBA-15 surface. This structuring provided excellent accessibility of TiO₂ particles for photocatalytic applications under pollutant gas and UV-A light exposure. The combination with the high specific surface area of SBA-15 resulted in the efficient degradation of 400 ppb of NO pollutant gas. Due to synergistic effects, the best nanocomposite in this study demonstrated a NO abatement performance of 4.0% per used mg of TiO₂, which is 40% more efficient than the reference photocatalytic material TiO₂ P-25.

1. Introduction

Indoor air pollution is a major societal concern in our industrial societies. Indeed, the toxic NO_x (NO + NO₂) compounds present in ambient air caused by human activities (e.g., road traffic, industrial plants, and fossil energy combustion) [1] are an environmental threat to human health. In particular, the WHO (World Health Organization) recently revised the air quality guidelines of NO₂ to lower thresholds (from 40 μg/m³ in 2005 to 10 μg/m³ in 2021 for long-term exposure and from 200 to 25 μg/m³ for short-term exposure) [2].

Photocatalysis driven by metal oxide semiconductors has been employed for decades to oxidize and mineralize atmospheric pollutants and reduce the overall pollution levels of NO_x, either in outdoor or indoor conditions [3]. This method relies on the ability of a photocatalytic semiconducting oxide (SCO), such as zinc oxide (ZnO) or titanium dioxide (TiO₂), to promote, under ambient air and solar light exposure, electron/hole pairs that conduce to the formation of free radicals involved in pollutant degradation reactions [4,5]. One of the most studied materials applied in this field is TiO₂, particularly the now archetypical Evonik P-25, which contains a mixture of both anatase (around 80 wt.%) and the high-temperature rutile phase [6,7]. Although the presence of the most photocatalytically active anatase phase is a mandatory requirement, some other physico-chemical parameters can be tuned to enhance the photocatalytic properties of this material. Two parameters are of particular importance to achieve improved photocatalyst properties. First, the specific surface of the SCO has to be increased to allow higher adsorption of pollutant species on its surface. This is generally achieved through the use of nanosized metal oxides, where the surface-to-volume ratio is maximized [8,9]. However, downsizing the particle size in excess can be counterproductive for catalysis if the crystalline domains of the SCO become too small [10]. Indeed, the recombination rate of electron/hole pairs in the SCO is largely driven by the density of point defect in the structure of the material, and highly disordered systems, or quasi-amorphous structures, exhibit low electron/hole pair lifetime due to fast recombination effects [11].

The optimized material for photocatalysis should, therefore, be the result of a compromise between low dimensionality, associated with high specific surface levels, and large crystalline domains where a low recombination rate of excitons is achieved. This delicate balance is difficult to reach with the bare TiO₂ compound only, but one of the possible strategies to overcome this problem is to associate a high specific surface material directly in strong interaction with a well-crystallized TiO₂ anatase phase [12]. A close interface between materials involved in photocatalytic nanocomposites is also a key parameter in reducing the charge carriers’ recombination due to surface defects [13,14]. Indeed, SBA-15 is a well-known mesoporous silica material (mesopore size from 3 to 6 nm large) exhibiting a very high specific surface (up to 1000 m²/g) due to its structuration [15]. Thanks to its porous structure, SBA-15 has been used as an efficient NO_x physisorption material in surface photovoltage (SPV) devices for NO_x gas sensing [16]. SBA-15 is also a material of choice that has already been studied as catalyst support for selective catalytic reduction (SCR) of NO_x found in oxygen-rich exhaust gases at high temperatures [17,18]. The combination of porous SBA-15, which has efficient adsorption properties for gaseous or liquid species, with the photocatalytically active TiO₂ semiconductor has been proposed due to the interesting expected synergism within this hybrid material [8,19,20,21,22,23]. Different chemical protocols are suggested for synthesizing this composite according to the interest of having or not the SCO located inside, outside (or both locations) the pores of the silica matrix. In a one-step protocol, colloidal TiO₂ nanocrystals (prepared from Ti(IV) isopropoxide) are mixed together with the molecular precursors used to build the porous silica matrix (tetraethyl orthosilicate and triblock copolymer Pluronic P123) and followed by an hydrothermal annealing at around 373 K [24]. This technique leads to the efficient incorporation of atomic titanium into the silica matrix and very small TiO₂ particles (mean size 7 nm) within the pores of the SBA-15. From the photocatalysis point of view, these low-crystallinity TiO₂ particles may not be suitable for enabling a low recombination rate of photo-excited charge carriers. Another method that is often employed is the SBA-15, which is a hard template that can be decorated in a second step with TiO₂ prepared from the solution impregnation of a Ti precursor. Landau [25,26] has described a method called CSD (chemical solution decomposition) where a Ti(IV) alkoxide precursor [Ti(OnBu)₄)] is maintained inside the pores of the SAB-15 matrix in an autoclave under pressure. After thermal decomposition, the TiO₂ loading is up to 30 wt.% for one run and can be doubled if the procedure is repeated. In that case, the TiO₂ particles are exclusively located in the mesopores of the silica matrix. Several other examples in the literature describe the use of Ti(IV) precursor for the decoration of mesoporous silica matrixes [19,21,22,27,28,29]. One important parameter for mastering the size of the TiO₂ nanoparticles is the use of a high-reactivity precursor in order to facilitate the nucleation and growth of nanocrystals under mild conditions [22,30]. In this study, we have prepared a rather unusual precursor based on a Ti(III) amidinate composition [31] that should offer higher reactivity than the usual Ti(IV) ones. This precursor has been used for the first time to decorate hydrated SBA-15 to yield TiO_x nanoparticles located within and outside the mesoporous silica grains. Our chemical process exploits the high reactivity of this type of organometallic precursor toward hydroxyl groups and water molecules adsorbed on silica surface, to directly generate either a crystalline or amorphous metal oxide phase [32,33,34]. The same procedure was used with a less reactive Ti(IV) precursor (tetrakis(ethylmethylamido)titanium). The two different hybrid nanocomposites have been characterized and their photocatalytic performances were compared for 400 ppb NO abatement under UV-A exposure in humidified air. The use of Ti(III) precursor with hydrated SBA-15 allowed to prepare a photocatalytic material showing a 40% improvement in NO degradation compared to bare P-25 photocatalyst. This result highlights the importance of synergetic mechanisms (specific surface of the matrix and crystalline structure of TiO₂) for photocatalysis reactions and the key role of the metal precursor to yield nanomaterials with optimized performances.

2. Results and Discussion

2.1. SBA-15 Hydration Methods and Characterization

Our team has previously developed a technique that takes advantage of the presence of molecular water on nanostructured supports on which metalorganic precursors react locally to give rise to the growth of nanosized metal oxide particles [32,33,34]. It is, therefore, important to master and maximize the amount of water in SBA-15 to allow the reaction with the metalorganic precursor to occur. Two methods have been set up in order to control the hydration of the SBA-15 surface. The first method (method A), described by Y. Belmoujahid [35], consists of placing the silica powder in a closed glass vessel containing water saturated with NaCl salt at room temperature (RT) and allowing a constant relative humidity of 75% RH. The water intake saturation in SBA-15 is obtained after around 4 h.

The second method (method B) is rather straightforward and consists of dispersing the silica powder into distilled water placed in a closed glass vessel and letting it boil under reflux at 100 °C for 2 h [36].

The amount of water (wt.%) in the SBA-15 samples was measured with TGA analysis after they were dried for 1 h under a primary vacuum at RT (Figure 1). The water present in the “as-received” SBA-15 powder was evacuated at around 80 °C and corresponded to 8.5% of the total weight. The sample treated with hydration method A (saturated NaCl solution) contained 23.8 wt.% of water, whereas the sample obtained by method B (boiling water at reflux) contained 55 wt.% of water at the same temperature.

The FTIR analyses of the different silica powders confirm this evolution (Figure S1). The siloxane (–Si–O–Si–) band appears as a broad strong peak centered at 1082 cm⁻¹. The bands at 3426 and 1610 cm⁻¹ are attributed to the stretching and bending vibrations of the hydroxyl groups and the adsorbed water molecules, respectively. The structural bands of the mesoporous silica are not modified by the adsorption of water.

There are only a few differences between the FTIR spectra of the SBA-15 as-received and that hydrated by method A. Conversely, the spectrum of the sample treated with method B exhibits a higher intensity of the band at 1610 cm⁻¹. The higher loading of water in the sample treated with hydration method B should be more adapted to the next step, where the hydrolysis of the organometallic precursor may lead to the generation of TiO_x nanoparticles on SBA-15.

Figure 2 shows the ¹H NMR MAS (magic angle spinning) spectra of mesoporous SBA-15 silica in the solid state according to the different hydration methods. This figure shows a predominant signal in the region around 4.8 ppm (4.92, 4.99, and 4.71 ppm for the SBA-15 as-received, in method A, and in method B, respectively). This resonance is typical of water molecules in weak interaction with the surface of silica. Part of this signal may also be associated with OH surface groups where their H are rapidly exchanged with those of water. A higher-frequency chemical shift indicates a greater proportion of physisorbed water molecules and/or of hydroxyl OH surface groups, whereas a lower-frequency shift indicates the formation of larger free-water clusters. Their relative populations also have an impact on the linewidth of the resonances because the water in the clusters has greater local mobility than physisorbed molecules and OH groups, leading to sharper resonances for the former. The ¹H spectra show that the initial silica already contains a significant amount of water. Hydration method A leads to a slightly higher proportion of physisorbed water and/or OH surface groups. The use of boiling water (method B) leads to the incorporation of a large quantity of free-water clusters and possibly a reduction in the amount of OH surface groups.

The ²⁹Si MAS NMR spectra (Figure S2) are characterized by three signals typical of Q⁴ framework silica sites ((SiO)₄Si) (at ~−111 ppm), Q³ silanol sites ((SiO)₃–Si-OH) (at ~−101 ppm) and Q² ((SiO)₂–(Si-(OH)₂) species (at ~−92 ppm). The relative intensity of Q³ and Q² resonances increases in the ²⁹Si CPMAS NMR (Figure S3) compared to MAS experiments as a result of the greater transfer of magnetization from neighboring H than for Q⁴. As with ¹H NMR, the ²⁹Si NMR spectra of the as-received and hydrated SBA-15 are very close (Figures S2 and S3 green and blue spectra). On the other hand, the spectra of silica hydrated in boiling water (method B) show some differences (Figures S2 and S3 red spectra), with a higher relative proportion of Q⁴. The population decrease of Q² for the sample prepared with method B is slightly lower than for Q³ (Figure S2). Low-frequency displacement is also observed in the sample treated with method B and is greater for Q⁴ (1.6 ppm in CPMAS and 1.0 ppm in MAS) than for Q³ (0.6 ppm in CPMAS) and Q² (~0.2 ppm in CPMAS). These results indicate a slight but undoubtful modification of the SBA-15 surface in the presence of boiling water, with a lower proportion of OH on the surface. Therefore, the exposition of silica to boiling water leads to the formation of Si–O–Si bonds by dehydration condensation between the hydroxyl groups themselves.

Small-angle XRD pattern and N₂ sorption isotherms of the SBA-15-based materials according to the different hydration methods are reported in Figures S4 and S5, respectively. The corresponding textural properties are presented in

Table 1. Textural properties of the different SBA-15 samples.

SBA-15 Sample	SBET (m2/g)	Vmicro (cm3/g)	Vmeso (cm3/g)	Dp (nm)
As-received SBA-15	1041	0.16	0.99	6.2
(SBA-15)/A	721	0.08	0.77	6.2
(SBA-15)/B	470	0.04	0.81	7.2

S_BET is the specific surface area obtained using the BET model, V_micro is the microporous volume extrapolated from the t-plots, V_meso = V_tot − V_micro is the mesopore volume where V_tot is the total pore volume obtained at P/P₀ = 0.9, and D_p is the mean pore diameter obtained using the BdB model on the desorption branch.

.

Small-angle XRD patterns of SBA-15 samples show three diffraction peaks (100), (110), and (200) at the same position, which are characteristic of the 2D hexagonal pore (with P6mm hexagonal symmetry typical for SBA-15 mesoporous materials) with the same unit cell a_hex of 10.6 nm. No significant evolution of the structure of the silica was noticed, regardless of the hydration treatments used. The same observation was also confirmed by the absence of any change in the microstructure of the different SBA-15 revealed by SEM observations (Figure S6).

Whatever the hydration method, the SBA-15 isotherms are all of type I and IV according to the IUPAC classification [37]. As a result, the as-received SBA-15 material has a high specific surface area (S_BET = 1041 m²/g) and large pore volume (V = 1.15 cm³/g, V meso + V micro) of 6.2 nm diameter. The SBA-15 synthesized in these conditions simultaneously exhibits mesopores and connections between the main mesopores walls through micropores [38]. It is noteworthy that the hydration method particularly influences the microporous volume (pore size < 2 nm). Indeed, the latter decreases when the amount of water increases, particularly with hydration method B (Table 1, Figure 1). In this case, V_micro has decreased by a factor of four compared to the as-received SBA-15 sample. The other textural parameters (V_meso and D_p) were less influenced by the hydration method used. The results strongly suggest that hydration treatments induce a phenomenon of silica dissolution/redeposition condensation within micropores, as reported by Galarneau et al. [39]. For the sample hydrated with method B, D_p increased from 6.2 to 7.2 nm, which was due to a decrease in the wall thickness because the unit cell a_hex is the same, indicating the same pore-to-pore distance. This evolution is in good agreement with ²⁹Si NMR, which indicates the condensation of hydroxyl groups upon hydration with method B and the formation of more Si–O–Si bonds.

2.2. Bare TiO₂ Powder Syntheses

The synthesis of TiO_x-yH₂O nanopowders was obtained by the controlled hydrolysis under an argon atmosphere of a metalorganic Ti precursor dissolved in an organic solvent (Figure S7). In this procedure, moisture slowly diffuses from the argon gas phase into the organic solvent containing the metalorganic precursor and induces the breaking of nitrogen to metal bonds to eventually produce metal oxide nanoparticles [40]. For the purpose of this study, we tested two different precursors with distinct titanium oxidation numbers. The tetrakis(ethylmethylamido)titanium (TEMAT; see 1 in Scheme 1) is commonly used in the atomic layer deposition (ALD) technique for TiO₂ films [28,41] and demonstrates a Ti center with an oxidation state of IV. In addition, an even more reactive metalorganic precursor for water has been envisaged. We synthesized and studied the hydrolysis in an organic solution of the (N-N′ diisopropylacetamidinato) Ti (III) precursor (Ti-Amd; see 2 in Scheme 1). The presence of a reduced oxidation number for the Ti center is expected to bring a better reactivity of the compound to water. The hydrolysis process, set up under an argon atmosphere at RT, is achieved through the slow diffusion of water from the gas phase into the organic solvent reacting with the Ti precursor, as described in Scheme 1. According to the reaction, 2 molar equivalents of water are necessary for a stoichiometric reaction. In the experimental conditions, an amount of 4 molar equivalents of water is employed in order to saturate the gas phase with water vapor and ensure the full hydrolysis of the dissolved precursors. In these conditions, the hydrolysis reaction takes place over two days.

The samples are characterized by wide-angle X-ray scattering (WAXS), a very sensitive technique to determine the local structure of materials with very small crystalline domains (Figure 3). Whatever the precursor composition 1 or 2, after calcination at 150 °C, the samples have no discernible crystalline domains and correspond to barely amorphous structures (Figure 3a). The pair distribution function (PDF) of the sample obtained from precursor 2 and calcined at 350 °C shows the very first coherence domains corresponding to a crystalline phase (Figure 3b). The structures of TiO₂ anatase (COD 7206075) and TiO₂ brookite (COD 9004140) were refined against the PDF obtained from the WAXS measurement of this sample (Figure 3d,e). Refinement results show that the local structure (r < 10 Å) is comparable to that of brookite (Figure 3d), while the average structure (r > 10 Å) is close to anatase (Figure 3e). The size of the coherent crystal domains calculated for spherical grains corresponds to a diameter of ca. 3.6 nm. Reversely, the oxide sample obtained from 1 still remains amorphous at this temperature.

This result clearly shows the higher reactivity of 2 toward hydrolysis, giving rise to an easier growth of crystalline domains at the moderate temperature of 350 °C.

As the calcination temperature increases (500 °C), both samples prepared with TiO₂/Ti-amd or TiO₂/TEMAT unambiguously show the anatase structure of TiO₂, as shown in Figure 3c and Figure S8. According to PDF analysis, TiO₂/TEMAT exhibits coherent domains of 17.5 nm (±2.3 nm), whereas TiO₂/Ti-Amd has a slightly higher crystallite size of 19.8 nm (±2.8 nm), which is consistent with the earlier presence of the very first crystalline domains obtained at 350 °C with this precursor. The presence of the anatase phase after calcination at 500 °C is also confirmed by X-ray diffraction analysis of the different powders (Figure S9). According to the Scherrer equation applied on the (101) peak measured at 2 theta 25.5°, the crystallite size is around 18.4 nm and 21.0 nm for 1 and 2, respectively.

Figure S10a,c shows the TEM images of the TiO_x-yH₂O powders obtained after hydrolysis of precursors 1 and 2 at RT. The powders have a rather ill-defined microstructure that may be formed by the agglomeration of amorphous particles. Calcination of the as-prepared samples has been performed to remove the water molecules and allow the crystallization of the TiO₂ phase. After calcination at 350 °C, there is no change in the structure of TiO_x grains obtained from 1 (Figure S10b), whereas a nano-structuration (ca. 4 nm grain size, determined on around 100 particles) already appears for the sample prepared with 2 (Figure S10d). After calcination at 500 °C under air, the structuration of the powder is clearly nanocrystalline for the two samples where similar nanosized grain size of ca. 13.5 nm (σ = 3.6 nm) and 13.1 nm (σ = 3.9 nm) are measured for TiO₂ obtained from 1 and 2, respectively (Figure 4).

2.3. Synthesis of TiO₂/SBA-15 Nanocomposites

The decoration of SBA-15 with TiO₂ is performed only on the silica samples prepared with the hydration method B because it leads to a higher amount of water in the silica structure. Indeed, our strategy is to maximize the loading of TiO₂ on SBA-15 by using the hydrated sample that contains a higher amount of adsorbed water.

The mass of the hydrated SBA-15 in our experimental protocol is 200 mg and, since it contains 55 wt.% water (Figure 1), this corresponds to 110 mg of water (i.e., 6 mmol) and 90 mg of SiO₂ per sample. We use an amount of 1.3 mmol of Ti precursor 1 or 2 to react with the hydrated SBA-15, which corresponds to 4.6 molar equivalent of water relative to titanium. Therefore there is a large excess of water compared to the employed Ti precursor and similar to the ratio used for the synthesis of bare TiO_x powders. This excess of water (a stoichiometric amount of 2 molar equivalents according to Scheme 1) is necessary to ensure the complete hydrolysis of the titanium precursor on the silica surface. Assuming the hydrolysis reaction is complete, 1.3 mmol of TiO₂ (62 mg of Ti, 103 mg of TiO₂) will form on the silica, which corresponds to a theoretical mass ratio of Ti equal to 32% (or a mass ratio of 51.5% for TiO₂).

The ICP-MS analyses of the two nanocomposite samples prepared with precursors 1 and 2 are presented in

Table 2. ICP-MS analyses of Ti content present inside the TiO₂/SBA-15 nanocomposites obtained from 1 and 2 and calcined at 500 °C.

Precursor	Theoretical Ti Content (wt.%)	ICP Measured Ti Content (wt.%)
TEMAT (1)		30.09
	32	30.27
		Mean: 30.18
Ti-Amd (2)		31.94
	32	31.17
		Mean: 31.55

.

Interestingly, the experimental amounts of Ti in TiO₂/SBA-15 samples obtained from precursors 1 and 2 (30.2 and 31.5 wt.%, respectively) are very close to the theoretical value previously calculated. This result indicates that, whatever the composition of the Ti precursor, the hydrolysis process is very efficient and leads to the total transformation of the precursor on the silica matrix surface.

Low magnification (×500) SEM images of the nanocomposites calcined at 500 °C are shown in Figure S11. They reveal that the TiO₂ aggregates are exclusively linked to the silica blocks and are distributed on the silica matrix surface.

Figure 5 shows the high magnification SEM image (×5000) of the nanocomposites acquired in the chemical contrast mode (back-scattered electrons). The images confirm that the TiO₂ nanoparticles form aggregated structures over the silica matrix. However, although it appears homogeneous at low magnification (Figure S11), the structuration at high magnification reveals a rather erratic dispersion of TiO₂ aggregated nanoparticles on the porous silica. Some areas of SBA-15 bundles appear to be devoid of TiO₂ particles, while in other locations, TiO₂ nanoparticles are clustered together. This TiO₂ dispersion over the silica matrix is the same for the two Ti precursors studied. In addition, no free TiO₂ aggregate is located apart from the SBA-15 matrix. A SEM-EDS analysis (Figure 6) of the different components of the nanocomposites confirms this particular repartition of TiO₂ nanoparticles over the SBA-15 support.

TiO₂ nanoparticles are concentrated in specific zones of the silica matrix. There are locations where the silica matrix is not decorated with TiO₂ whereas the photocatalytic oxide is concentrated into agglomerated structures. This peculiar repartition is very different from what we have previously observed on the decoration of WO₃-2H₂O nanoplatelets with ZnO nanoparticles by using a similar hydrolysis method. In the former case, ZnO appears as a very homogeneous nanostructured layer distributed all over the surface of the WO₃-2H₂O support [33]. This result was achieved thanks to the availability of one of the structural water molecules present in the support to hydrolyze the Zn precursor. In the case of the SBA-15 support, it seems that the water added by the boiling process is not homogeneously distributed over the silica. Instead, water-rich areas, such as clusters or droplets of water, could be randomly distributed over the SBA-15 surface. This assumption is also supported by the ¹H MAS NMR analyses that reveal the presence of water clusters in the sample hydrated by method B (boiling water). The Ti precursor in contact with these water reservoirs reacts to locally produce a high density of TiO_x-yH₂O structures. The interesting point is that these water reservoirs remain tightly linked to the substrate because all the TiO_x structures are attached to the silica support in a similar way. This mechanism is supported by TEM imaging of the nanocomposites (Figure 7). In these images, the TiO₂ nanoparticles appear irregularly dispersed on the silica surface and do not form a continuous layer. Similar TiO₂ grapes are observed for the nanocomposites prepared with precursor 1 or 2. This analysis confirms that despite the aggregated structure of TiO₂ nanoparticles, they are mainly present on the surface of SBA-15 grains and not dispersed homogeneously into the pores (Figure 7c,d).

The mean grain size of TiO₂ particles formed on SBA-15 supports is close to 9 nm whatever the employed Ti precursor (9.8 and 9.1 nm for precursor 1 and 2, respectively). This value is clearly lower than that of bare TiO₂ nanoparticles obtained by hydrolysis in solution, which is close to 13 nm (13.5 and 13.1 nm for precursor 1 and 2, respectively). This result underlines the important role played by the support in the hydrolysis and growth mechanism of the oxide particles. The growth of TiO₂ crystals occurs mainly at the SBA-15 surface, particularly on the micropores where water clusters are present and not inside the mesopores. The interaction between the silica support and TiO₂ induces the particles’ stabilization and their limited size. This low dimensionality of the anatase phase is also evidenced by the XRD diffractograms of the nanocomposites, which show all the peaks relating to the crystalline anatase structure but with rather large half-height peaks that reflect the limited size of the crystalline domains (Figure S12 vs. Figure S9).

The main physico-chemical characteristics of the SBA-15 as-received, TiO₂ oxides, and TiO₂/SBA-15 nanocomposites are summarized in

Table 3. Main physico-chemical characteristics of the samples tested for photocatalysis.

Sample	TiO2 Content (wt.%)	SBET (m2/g)	TiO2 Crystal Size (nm)
			TEM	XRD
TiO2 P-25	100	50	21.0	31.0
TiO2 TEMAT	100	51	13.5	18.4
TiO2 Ti-Amd	100	21	13.1	21
As-received SBA-15	-	1041	-	-
SBA-15 hydrated B	-	470	-	-
TiO2 TEMAT/SBA-15	50.4	297	9.8	11.7
TiO2 Ti-Amd/SBA-15	52.6	251	9.3	11.9

.

Based on the previous findings, the decoration of hydrated SB1-15 with TiO_x NPs, prepared from the hydrolysis of Ti-Amd precursor, is schematically depicted in Figure 8.

2.4. NO Abatement Tests

NO degradation percentages were calculated from Equation (1) (see Section 3.6 Evaluation of the Photocatalytic Activity) for 20 mg of photocatalytic dispersion applied to the surface of a glass substrate (length × width: 10 × 5 cm²). The photocatalytic runs were performed on three main types of samples: nanocomposites (TiO₂/SBA-15 obtained from 1 and 2), bare TiO₂ powders (P-25, TiO₂ powder calcined at 500 °C and obtained from 1 and 2) and physical mixtures of SBA-15 and TiO₂ powders to assess the role of the silica matrix. However, the TiO₂/SBA-15 nanocomposites and the SBA-15 mixed with TiO₂ powders are not 100% TiO₂ by weight, unlike the bare oxides (TiO₂ synthesized from 1 and 2, and TiO₂ P-25). The effective quantity of TiO₂ deposited on the glass surface was therefore considered according to ICP-MS quantification (Table 2) in order to obtain the NO degradation results for the same surface density of TiO₂, i.e., 1 mg of TiO₂ per 50 cm² of the substrate surface (0.02 mg/cm²). The NO degradation results obtained under artificial UV-A light for bare TiO₂, TiO₂/SBA-15 nanocomposites, and SBA-15 physically mixed with bare TiO₂ are shown in Figure 9. The photocatalytic activity of SBA-15 alone was also assessed using the same protocol and found to be equal to zero.

For bare oxides, the highest percentage of NO degradation was obtained with TiO₂ P-25 (2.7%). The efficiency of bare oxides obtained from precursors 1 (1%) and 2 (1.3%) was around twice as low. The photocatalytic efficiency of TiO₂ depends on several factors, including crystal structure, crystalline phase, crystallinity, particle size, and surface area, and the balance between these factors determines the overall activity of a given photocatalyst [42]. The air depollution performance of TiO₂ P-25 has been widely reported in the literature and is mainly attributed to the synergistic effect between the anatase and rutile crystalline phases, which reduces photogenerated charge recombination [43,44]. The physico-chemical characterization presented above showed that crystalline anatase was the only phase detected for TiO₂ nanoparticles obtained from 1 and 2 (after calcination at 500 °C, Figure S9), which could explain the lowest activity compared to TiO₂ P-25. Moreover, TiO₂ from 1 had a mean particle size of 13.5 nm and a specific surface area of 51 m²/g compared with values of 13.1 nm and 21 m²/g, respectively, for TiO₂ nanoparticles from 2 (Table 3). The two TiO₂ anatase powders prepared in this study demonstrated the same performance for NO abatement (around 1%), which is expected for metal oxides that have similar structural properties.

For the same TiO₂ surface density (1 mg of TiO₂ per 50 cm² of substrate surface), the TiO₂/SBA-15 nanocomposites showed similar and even superior performance than TiO₂ P-25. Note that photocatalytic activity is highly dependent on experimental conditions [45], making it difficult to compare the performance of the nanocomposites tested with other materials in the literature. For this reason, we compared the NO degradation percentages obtained with the TiO₂/SBA-15 nanocomposites with those of the reference photocatalyst TiO₂ P-25 tested under the same experimental conditions. The best NO degradation (4.0%) was obtained with TiO_{2 Ti-Amd}/SBA-15 (around 1.5 times higher than that of TiO₂ P-25), while TiO_{2 TEMAT}/SBA-15 had the same NO degradation percentage than TiO₂ P-25. These results clearly highlighted the following: (1) the advantage of decorating mesoporous SBA-15 silica with TiO₂ nanoparticles compared to the corresponding bare oxides as the performance is increasing by a factor of 3 (1% vs. 2.7% and 1.3% vs. 4.0%), and (2) the higher efficiency of TiO_{2 Ti-Amd}/SBA-15 compared to bare TiO₂ P-25. The higher photoactivity of TiO₂/SBA-15 nanocomposites compared to bare TiO₂ may be related to three synergistic factors that made NO pollutant molecules more accessible to the active sites and thus improve the photocatalytic reaction: (1) smaller TiO₂ particle size (~12 nm for nanocomposites and ~20 nm for bare oxides, from XRD analysis; Table 3), (2) larger surface area (around 250–300 m²/g for nanocomposites and around 20–50 m²/g for bare oxide, Table 3), and (3) well dispersed TiO₂ nanoparticles over the surface of the silica matrix. The positive effect of these factors and the higher photocatalytic activity of mesoporous silica–titania compared to bare TiO₂ were notably highlighted by other authors [46,47,48].

The last series of results allows us to compare the photocatalytic activity of the physical mixing of TiO₂ powders and SBA-15 to the chemically synthesized nanocomposites (with identical TiO₂/SBA-15 mass ratio). Interestingly, the physical mixing also showed an improvement in NO degradation compared to the corresponding bare TiO₂ powders. However, degradation percentages were lower than those of TiO₂/SBA-15 nanocomposites. A higher UV absorption intensity in the mixed oxides due to the presence of SBA-15 could be responsible for this increase. A stronger UV absorption intensity implies that more TiO₂ particles can be activated by light and, therefore, more photogenerated charges can be promoted, leading to enhanced photocatalytic activity. The role of the received light irradiation on photocatalytic activity was highlighted by several authors in the literature. Notably, Li and Kim [47] partially attributed the higher photo-oxidation of benzene of TiO₂-xSiO₂ composites compared to bare titania to greater UV absorption intensity. Alonso-Tellez et al. [49] demonstrated that the higher irradiance received by the UV100 TiO₂ coating and the deeper penetration of light within this coating than in the case of the TiO₂ P-25 coating led to increased conversion and mineralization of the methylethylketone pollutant. This effect is even more pronounced in the case of the chemically prepared nanocomposites of this study, especially for the one prepared with the precursor Ti-Amd.

This suggests a more efficient UV-A sensitization of TiO₂ particles when they are well dispersed and closely attached to the hydrated SBA-15 support, thanks to the use of a highly reactive Ti precursor.

3. Materials and Methods

3.1. Synthesis of SBA-15 Powder

The ordered mesoporous silica SBA-15 was prepared at IS2M (Mulhouse, France) according to the protocol described by Zhao et al. [15]. In a 250 mL polypropylene flask, 4 g of triblock copolymer P123 (Aldrich, Saint Louis, MO, USA) were dissolved in HCl aqueous solution (19.5 mL of 37 wt.% HCl and 127 mL of distilled H₂O). The flask was placed in a water bath at 40 °C with magnetic stirring (500 rpm) for about 3 h to allow a complete dissolution of P123. Then, 8.62 g of Tetraethylorthosilicate TEOS (Aldrich, Saint Louis, MO, USA) were added by maintaining the stirring and the temperature conditions. The molar composition of the gel was 1 TEOS: 0.017 P123: 5.68 HCl: 197 H₂O. The solution was stirred at 40 °C for 24 h and then the bottle was placed without any stirring in an oven for 24 h at 90 °C. The precipitated solid (as-made SBA-15) was recovered by filtration on a Büchner funnel, washed with 200 mL of distilled water, and dried for 48 h at 70 °C. To eliminate the porogen agent (P123) and thus release the porosity of the material, the as-made SBA-15 was calcined under air in a muffle furnace at 300 or 500 °C for 4 h (the heating time from ambient to final temperature was 6 h). The mesoporous silica was hydrated after synthesis in two different ways. The first method (method A) consisted of hydrating the mesoporous silica in a controlled relative humidity atmosphere for 4 h. A 75% relative humidity is created in a closed glass enclosure using an aqueous saturated NaCl (Aldrich) in Milli-Q water solution. The second method (method B) consisted in boiling the mesoporous silica directly in Milli-Q water (purification system) for 2 h at 100 °C. Typically, around 500 mg of SBA-15 mesoporous silica was brought to reflux in 100 mL of Milli-Q water [36]. Afterward, the water was removed using centrifugation, and the silica paste was dried for 1 h under a vacuum at RT.

3.2. Synthesis of Titanium Tris-Amidinate Precursor

The titanium amidinate (Ti-Amd) precursor was synthesized according to an adaptation of the protocol described by Gordon et al. [31]. All reactions and manipulations were conducted under a nitrogen atmosphere using either a glovebox or a standard Schlenk line technique. The used solvents were dried and collected using a solvent purification system. For the synthesis of the lithium amidinate, a solution of methyllithium (1.6 M in Et₂O, 40 mL, 0.063 mol) in Et₂O was added dropwise using a bromine bulb to a solution of 1,3-diisopropylcarbodiimide (6.9 g, 0.055 mol) in 80 mL of Et₂O at −30 °C. The mixture was warmed to RT and stirred for 4 h. The solvent was then removed under reduced pressure. The resulting white solid was transferred to the glovebox and extracted with pentane (washing three times with 30 mL), then dried under reduced pressure. A solution of the obtained solid in ether was added dropwise to a solution of titanium chloride tetrahydrofuran TiCl₃.(THF)₃ (7.8 g, 0.054 mol) in 50 mL of Et₂O. The reaction mixture was stirred for 12 h in the glovebox. The resulting solution was filtered through a pad of Celite^® on a glass frit to give a brown solution. The concentration of the filtrate and its cooling to −33 °C in the glovebox afforded a brown solid.

3.3. Synthesis of Bare TiO₂ Powder

TEMAT (tetrakis(ethylmethylamido)titanium IV) was received from Sigma-Aldrich supplier. Ti-Amd (N-N′ diisopropylacetamidinato)titanium III) precursor was prepared in the laboratory according to the protocol described in Section 3.2. The hydrolysis of the organometallic precursors was carried out in a proprietary designed glass reactor (Figure S7), comprising two separated and concentric reservoirs: one containing the dissolved precursor in anisole, and the second placed in the center of the vessel in order to receive the water (4 molar equivalent compared to titanium precursor) introduced from the septum placed above and in the center of the airtight vessel lid. This reactor allowed the slow hydrolysis of the titanium precursor in organic solvent by diffusion of the water from the gas phase to the solvent one, the reaction occurring at the gas-solvent interface. A white TiO_x-yH₂O powder was collected after 2 days, and the powder was purified by washing it in toluene and centrifuged before drying it under a vacuum at RT. The samples were then calcined at various temperatures from 150 to 500 °C for 2 h in air.

3.4. Synthesis of TiO₂/SBA-15 Nanocomposites

The nanocomposites of TiO₂/SBA-15 were synthesized as follows: a volume of 400 µL of Tetrakis(ethylmethylamido)titanium (TEMAT) (99%, STREM, Newburyport, MA, USA), or 613 mg of Ti amidinate precursor, was added to a suspension of 200 mg of hydrated SBA-15 within 12 mL of anisole. Hence, the amount of water introduced in the reaction medium through the hydrated silica corresponds to around 4.6 molar equivalent (i.e., very similar to the amount used to prepare bareTiO₂ by water moisture presented in Section 3.3). The mixture was stirred for 24 h at RT. At the end of the reaction, the resulting solution was centrifuged (15 min, 3000 rpm) and then transferred to the glovebox. The precipitate was repeatedly washed three times with anisole and centrifuged. At the end of the washing procedure, the TiO_x/SBA-15 was dried under a vacuum for 2 h in order to fully remove the anisole solvent. Thermal annealing was performed under air at various temperatures from 150 to 500 °C to achieve fully oxidized and crystallized TiO₂/SBA-15 nanocomposites.

3.5. Characterization Techniques

Thermogravimetric analysis (TGA) was performed using a Setaram thermobalance (Setaram Engineering, Caluire-et-Cuire, France) with a ramp of 10 °C/min in the 30–600 °C range under ambient air.

NMR experiments were recorded on a Bruker Avance 400 III HD spectrometer (Bruker Corp., Billerica, MA, USA) operating at magnetic fields of 9.4 T. Samples were packed into 4 mm zirconia rotors and were rotated at a frequency of 10 kHz at 295 K. ¹H MAS were performed with the DEPTH pulse sequence and a recycle delay of 5 s. ²⁹Si MAS were acquired with the UDEFT pulse sequence [50] with a recycle delay of 80 s. Ramped cross-polarization (CP) ¹H → ²⁹Si MAS spectra were recorded with a recycle delay of 2 s and a contact time of 2 ms. Chemical shifts were referenced to TMS. Fourier transform infra-red spectroscopy (FTIR) spectra were obtained by analyses of SBA-15 powders mixed in a KBr pellet using a Perkin–Elmer 100 spectrometer (Perkin–Elmer, Waltham, MA, USA). The spectra were registered between 4000 and 350 cm⁻¹ wavenumbers.

Total scattering (WAXS) measurements were performed on a Malvern Panalytical Empyrean III diffractometer (Malvern, Worcestershire, UK) in transmission geometry, and equipped with a Mo anode (λ = 0.71 Å) and a Galipix3D detector. The maximum reached q value is q_max = 16.6 Å⁻¹, thus corresponding to a maximum angle of 2θ_max = 143°. Powder specimens were placed in 1 mm diameter capillaries. For each specimen measurement, an empty capillary contribution was subtracted. Pair distribution functions (PDF) were extracted using PDFgetX3 software v2.1.2. Least square refinements of experimental PDF were performed using WAXS_toolbox software v1.0 suite [51], based on the diffpy-cmi package [52]. Crystalline structural models of TiO₂ phases were taken from the Crystal Open Database [53]. Powder X-ray diffraction patterns were obtained on a Seifert XRD 3000 TT X-ray diffractometer (Seifert X-ray, Deutschland, Germany), with Cu-Kα radiation, fitted with a diffracted-beam graphite monochromator. Data were collected in the 2θ configuration between an angle of 10 and 70°.

Physisorption measurements were made with the Micromeritics ASAP 2020 analyzer (Micromeritics, Norcross, GA, USA). The samples were degassed at 200 °C under high vacuum for 12 h before analyses. A quantity of 200 mg of powder was used to carry out the measurement. The specific surface area of the samples was determined using the linear part of the Brunauer, Emmet, and Teller (BET) plot, and the pore size was estimated using the desorption branch of the isotherms and the Broekhoff–De Boer (BdB) model. The microporous volume was extrapolated from the linear part of the t-plots.

Transmission electron microscopy (TEM) images were obtained with a Jeol 1400 transmission electron microscope at 120 kV (Jeol Ltd., Akashima, Tokyo, Japan). Field emission scanning electron microscopy (FESEM) images were obtained using a Jeol JSM-6700F microscope operating at 10 kV.

3.6. Evaluation of the Photocatalytic Activity

TiO₂-decorated mesoporous silica powders hydrated using method B with each precursor (1 for TEMAT and 2 for Ti-Amd at 500 °C) were dispersed in water (10 mg/mL) with sonication for 30 min. Each TiO₂/SBA-15 nanocomposite dispersion was then applied to the surface of the glass substrate (10 cm × 5 cm = 50 cm²) using a glass pipette. The quantity of the dispersion deposited on the surface was 20 mg for each sample. The same dispersion preparation and deposit protocols were followed for bare oxides (TiO₂ obtained from 1 and 2 at 500 °C as described in Section 3.3, and TiO₂ P-25). Dispersions of SBA-15 mixed (physical mixing) with bare oxides (TiO₂ obtained from 1 and 2, and TiO₂ P-25) were also prepared and applied to the surface of the glass substrate. The weight percentage of TiO₂ added to these mixtures was chosen to be the same as the one found in TiO₂/SBA-15 nanocomposites (ICP-MS results, Table 2).

Photocatalytic degradation of NO was evaluated using the experimental set-up described by Hot et al. [54] and Castelló Lux et al. [4], to which the reader is kindly referred for a more detailed description of the methodology. It is adapted from standard ISO 22197-1 [55]. The main points of the protocol to keep in mind as as follows: (1) continuous polluted air (at a flow rate of 1.5 L/min) was injected through the by-pass to check that the desired concentration was reached (400 ppb NO), (2) the polluted air was injected into a borosilicate reactor containing the photocatalytic sample to be tested in the dark for 10 min, (3) the photocatalysis was activated by switching on a UV-A light (Narva LT-T8 Blacklight blue 18 W 073 fluorescent tube) placed above the surface of the sample for 20 min (1 W/m² on the sample surface measured with a Gigahertz–Optik radiometer (UV-A detector: UV-3717 model, 315–400 nm; Gigahertz–Optik GmbH, Türkenfeld, Germany), and (4) before the end of the test, the polluted air was returned through the by-pass to check the concentration. Relative humidity and temperature were kept constant at 50% and 20 °C, respectively. The reactor was placed in a light-tight box to isolate the sample from room light and to receive only artificial illumination when the light was switched on. The NO concentration was measured by a chemiluminescence analyzer (model AC32M, Envea SA, Poissy, France).

The photo-oxidation of NO was assessed with Equation (1):

$$\mathrm{N}\mathrm{O} \mathrm{d}\mathrm{e}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} (\%)=100\times \frac{{\left[\mathrm{N}\mathrm{O}\right]}_{\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l}}-{\left[\mathrm{N}\mathrm{O}\right]}_{\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}}{{\left[\mathrm{N}\mathrm{O}\right]}_{\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l}}}$$

(1)

where [NO]_initial was the concentration in ppb measured by the analyzer at the exit of the reactor before light activation once the steady state was established, and [NO]_final was the concentration in ppb measured at the exit of the reactor after the light activation and averaged over the last 10 min.

Photocatalytic tests were conducted on glass samples coated with TiO₂ P-25 oxide, TiO₂/SBA-15 nanocomposite, and SBA-15 mixed with TiO₂ P-25 oxide. Each test was repeated twice.

4. Conclusions

In this paper, we have investigated a novel metalorganic method for the preparation of bare TiO₂ and TiO₂/SBA-15 nanocomposite powders aimed at the photocatalytic abatement of NO pollutant gas (400 ppb). The photocatalytic activity of the different materials prepared in this study is compared, at a laboratory reactor scale, with that of commercial TiO₂ P-25 powder. The mechanical combination of TiO₂ powders with mesoporous SBA-15 silica results in improved photocatalytic abatement of NO compared to bare TiO₂ powders. This improvement is likely due to a more efficient UV diffusion pathway toward the TiO₂ grains. However, chemically prepared TiO₂/SBA-15 nanocomposites demonstrate even higher photocatalytic activity compared with the simple physical mixing of TiO₂ powders with SBA-15 and bare TiO₂ oxides. Specifically, the nanocomposite prepared from Ti-Amd precursor, which is more reactive than TEMAT under hydrolysis conditions, exhibits the best activity for NO abatement with a value of 4.0% (NO abatement/mg of TiO₂ deposited on the glass substrate surface), which represents a 40% increase compared to TiO₂ P-25. Thus, maximizing the interaction and relationship between porous silica and TiO₂ nanoparticles is helpful to achieve a higher synergistic effect for the photocatalytic degradation of NO pollutants. Future work will focus on incorporating Au nanoparticles onto chemically prepared TiO₂/SBA-15 nanocomposites to further enhance the catalytic activity thanks to the plasmonic properties of the noble metal nanoparticles in the visible domain of the spectrum.