Bimetallic TiO2 Nanoparticles for Lignin-Based Model Compounds Valorization by Integrating an Optocatalytic Flow-Microreactor

The challenge of improving the activity of TiO2 by modifying it with metals and using it for targeted applications in microreactor environments is an active area of research. Recently, microreactors have emerged as successful candidates for many photocatalytic reactions, especially for the selective oxidation process. The current work introduces ultrasound-assisted catalyst deposition on the inner walls of a perfluoro-alkoxy alkane (PFA) microtube under mild conditions. We report Cu-Au/TiO2 and Fe-Au/TiO2 nanoparticles synthesized using the sol–gel method. The obtained photocatalysts were thoroughly characterized by UV–Vis diffuse-reflectance spectroscopy (DRS), high-resolution scanning electron microscopy (HR-SEM), high-resolution transmission electron microscopy (HR-TEM), X-ray diffraction analysis (XRD), Fourier transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS), and N2 physisorption. The photocatalytic activity under UV (375 nm) and visible light (515 nm) was estimated by the oxidation of lignin-based model aromatic alcohols in batch and fluoropolymer-based flow systems. The bimetallic catalyst exhibited improved photocatalytic selective oxidation. Herein, four aromatic alcohols were individually investigated and compared. In our experiments, the alcohols containing hydroxy and methoxy groups (coniferyl and vanillin alcohol) showed high conversion (93% and 52%, respectively) with 8% and 17% selectivity towards their respective aldehydes, with the formation of other side products. The results offer an insight into ligand-to-metal charge transfer (LMCT) complex formation, which was found to be the main reason for the activity of synthesized catalysts under visible light.


Introduction
TiO 2 is used as a photocatalyst for the degradation of impurities, a component of selfcleaning coatings, cosmetics, and for the production of hydrogen in water decomposition reactions [1,2]. Despite the numerous application possibilities, however, its industrial use is very limited. TiO 2 due to wide band gap is capable of absorbing only UV radiation, which constitutes only 3 to 5% of solar radiation. The consequence of this is the fact that TiO 2 still does not have satisfactory photocatalytic efficiency under visible light irradiation [3,4]. The use of UV lamps as a radiation source, due to the high energy consumption, significantly increases the cost of the process, which is an important factor limiting the broader use of this method in removing pollutants on an industrial scale. Consequently, many studies focused on the visible response of TiO 2 photocatalysts by: (i) doping with metal or non-metal ions, (ii) hetero-junctions creation with other semiconductors [5,6], (iii) sensitization [7], (iv) the formation of a surface complex with energy transfer [8,9] or (v) surface modification with noble metals nanoparticles [10]. By doping TiO 2 with transition-metal ions (Fe, Cu, Co, Mn, and Ni) was reported to be effective for the enhancement of the photocatalytic activity by changing the Upon the deposition of catalysts on the walls of the microreactor, the synthesize bimetallic materials revealed better photo reactivity in regard to both BnOH specific con version rates, an absolute contrast trend with the batch experiments ( Figure 2a). After 3 of light experiment, the specific conversion rate of Cu-Au/TiO2 prepared using coppe nitrate precursor (CuN-Au/TiO2) in the batch system was 14 µ mol/m 2 ·h, whereas, in th microflow system, it reached 268 µ mol/m 2 h.
The specific surface areas determined from N2 physisorption for the Cu-Au/TiO2 an Fe-Au/TiO2 varied from 251 to 547 g/m 2 , which was predominantly controlled by the typ of precursor used for synthesis ( Table 1). The use of nitrate precursor during the synthesi procedure caused a significant increase (about two-fold) in the specific surface area wit reduced crystallinity (Figure 3a) compared to pure TiO2. Au phase is not a dominant phas for these catalysts; therefore a small peak of Au is observed. Amorphous phases dominat in CuN-Au/TiO2 and FeN-Au/TiO2 samples, such phases do not give peaks in the XRD pattern, and on the background of such phases, even a small peak from nanocrystallin gold is easily detectable. The crystallite size of Au for each catalyst are shown in Table 1   Upon the deposition of catalysts on the walls of the microreactor, the synthesized bimetallic materials revealed better photo reactivity in regard to both BnOH specific conversion rates, an absolute contrast trend with the batch experiments ( Figure 1a). After 3 h of light experiment, the specific conversion rate of Cu-Au/TiO 2 prepared using copper nitrate precursor (CuN-Au/TiO 2 ) in the batch system was 14 µmol/m 2 ·h, whereas, in the microflow system, it reached 268 µmol/m 2 h.
The specific surface areas determined from N 2 physisorption for the Cu-Au/TiO 2 and Fe-Au/TiO 2 varied from 251 to 547 g/m 2 , which was predominantly controlled by the type of precursor used for synthesis ( Table 1). The use of nitrate precursor during the synthesis procedure caused a significant increase (about two-fold) in the specific surface area with reduced crystallinity (Figure 2a) compared to pure TiO 2 . Au phase is not a dominant phase for these catalysts; therefore a small peak of Au is observed. Amorphous phases dominate in CuN-Au/TiO 2 and FeN-Au/TiO 2 samples, such phases do not give peaks in the XRD pattern, and on the background of such phases, even a small peak from nanocrystalline gold is easily detectable. The crystallite size of Au for each catalyst are shown in Table 1.  Upon the deposition of catalysts on the walls of the microreactor, the synthesized bimetallic materials revealed better photo reactivity in regard to both BnOH specific conversion rates, an absolute contrast trend with the batch experiments ( Figure 2a). After 3 h of light experiment, the specific conversion rate of Cu-Au/TiO2 prepared using copper nitrate precursor (CuN-Au/TiO2) in the batch system was 14 µ mol/m 2 ·h, whereas, in the microflow system, it reached 268 µ mol/m 2 h.
The specific surface areas determined from N2 physisorption for the Cu-Au/TiO2 and Fe-Au/TiO2 varied from 251 to 547 g/m 2 , which was predominantly controlled by the type of precursor used for synthesis ( Table 1). The use of nitrate precursor during the synthesis procedure caused a significant increase (about two-fold) in the specific surface area with reduced crystallinity (Figure 3a) compared to pure TiO2. Au phase is not a dominant phase for these catalysts; therefore a small peak of Au is observed. Amorphous phases dominate in CuN-Au/TiO2 and FeN-Au/TiO2 samples, such phases do not give peaks in the XRD pattern, and on the background of such phases, even a small peak from nanocrystalline gold is easily detectable. The crystallite size of Au for each catalyst are shown in Table 1.  A combination of type II and IV adsorption isotherms was noticed for sol-gel-synthesized TiO 2 and the bimetallic catalysts with nitrate precursors (Figure 2b) [24]. The H2-type hysteresis loop of these catalysts suggested a complex pore structure [25]. On the other hand, the bimetallic catalyst with acetate precursor (CuA-Au/TiO 2 ) showed a type-IV adsorption isotherm [24], suggesting the presence of mesopores. The shape of the hysteresis loops was of type H3, which indicates the presence of aggregates of slit-shaped conical pores composed of primary particles, which can give rise to piled-up pores [26][27][28].
The Tauc plots ( Figure S2) revealed that the bandgaps for all the synthesized catalysts were~3.4 eV. In addition, the BJH (Barrett-Joyner-Halenda) method was used to calculate the average pore size distribution, and the values are provided in Table 1.
The average pore size was lower for samples synthesized using nitrate precursors than for samples that were prepared with acetate precursors. This observation can be explained by a partial modification of the TiO 2 surface by nitrate decomposition and evolving gaseous NO x during the synthesis procedure. The photocatalytic activity in the microflow system ( Figure 1b) of CuN-Au/TiO 2 and FeN-Au/TiO 2 accompanied by their larger specific area, as a consequence, may provide more active sites and shorten the bulk diffusion length of charge carriers, thus suppressing bulk recombination [29]. However, the pore volume was much higher for the sample CuN-Au/TiO 2 (0.4 cm 3 /g) than for FeN-Au/TiO 2 (0.2 cm 3 /g), even though nitrogen salts were used in both cases during the preparation. This dependency can be caused by steric hindrance during the growth of the TiO 2 crystallites via Ostwald ripening [30]. The molecular structure of iron (III) nitrate nonahydrate is bigger than copper (II) nitrate trihydrate (computed by PubChem), which consequently causes more steric hindrance.
The synthesized bimetallic catalysts were investigated using transmission electron microscopy (TEM). The composition of the particles was studied using STEM (high-angle annual dark field, HAADF) and EDXS mapping. In the CuA-Au/TiO 2 sample, the TEM analysis revealed that primary nanoparticles were clustered to form more oversized agglomerates during the formation stage itself ( Figure 3). The results of the TEM analysis showed that Ti, O, and Cu are evenly distributed throughout the sample surface (Figure 3d,e).
The application of the HAADF method confirmed the presence of the expected elements, such as Ti, Cu, and Au (Figure 3c-e), and also, from XPS analysis of CuA-Au/TiO 2 , we were able to confirm the presence of 0.1 at% Cu and 0.03 at% Au in the catalyst surface ( Figure S4). Gold and copper nanoparticles were evenly distributed in the tested sample (Figure 3b-d) which is very important in terms of the influence on the photocatalytic activity. Based on TEM analysis (Figure 3), UV-Vis DRS spectra ( Figure S3), and XPS analysis ( Figure S4), it can be deduced that surface incorporation of Cu and Au atoms on the TiO 2 has occurred. On the UV-Vis spectra, the red shift of absorption for CuA-Au/TiO 2 in comparison to bare TiO 2 and no obvious peak from copper nanoparticles (NPs) can be seen. The absence of the peak coming from Cu NPs could be due to the low amount of Cu NPs on the TiO 2 surface. The same results regarding the slight red shift of absorption to 403 nm for copper nanospheres coupled with TiO 2 were obtained by Monga et al. and also other research groups [31][32][33]. The distinct peak derived from Au nanoparticles occurred at 560 nm due to the charge transfer from the metal ion to TiO 2 [32]. The abovementioned absorption in the range of visible irradiation may be due to the localized surface plasmon resonance (LSPR). It has been shown that in the mechanism using LSPR, metal can act as an electron trap and thus inhibit the recombination process of electron-hole pairs [34,35]. The characteristic LSPR band maximum at 555 nm [36] indicates the presence of Au nanoparticles in the CuN-Au/TiO 2 sample, shown on the UV-Vis DRS spectra ( Figure S3).
It can be clearly seen in the TEM pictures ( Figure 3b) that the obtained gold nanoparticles have a spherical shape, and the particle size distribution of Au was in the range of 10-40 nm, with the highest contribution of the fraction from 15 to 25 nm and mean size equal to 22.1 nm (Figure 3a). The nanometric size and shape of the obtained Au particles are good enough to provoke LSPR [31,37]. This proves that the particles in nanoscale with a favorable spherical shape were successfully obtained during the synthesis procedure. Gołąbiewska et al. [38] have shown that the spherical shape of gold nanoparticles was very beneficial for increasing the photocatalytic activity in the range of visible irradiation. They compared the photoactivity of different shapes of Au particles deposited on TiO 2 microspheres and proved that visible light activity decreased in the following order: spheres > rods > stars [38]. It can also be added that in the case of the CuA-Au/TiO 2 sample, the inter-planar spacing for gold was 0.24 nm, which corresponds to the separation between the (111) lattice planes of Au (Figure 3b), which was consistent with the results obtained by XRD analysis (Table 1).  The application of the HAADF method confirmed the presence of the expected elements, such as Ti, Cu, and Au (Figure 4c-e), and also, from XPS analysis of CuA-Au/TiO2, we were able to confirm the presence of 0.1 at% Cu and 0.03 at% Au in the catalyst surface ( Figure S4). Gold and copper nanoparticles were evenly distributed in the tested sample ( Figure 4b-d) which is very important in terms of the influence on the photocatalytic activity. Based on TEM analysis (Figure 4), UV-Vis DRS spectra ( Figure S3), and XPS analysis ( Figure S4), it can be deduced that surface incorporation of Cu and Au atoms on the TiO2 has occurred. On the UV-Vis spectra, the red shift of absorption for CuA-Au/TiO2 in comparison to bare TiO2 and no obvious peak from copper nanoparticles (NPs) can be seen. The absence of the peak coming from Cu NPs could be due to the low amount of Cu NPs on the TiO2 surface. The same results regarding the slight red shift of absorption to 403 nm for copper nanospheres coupled with TiO2 were obtained by Monga et al. and also other research groups [33][34][35]. The distinct peak derived from Au nanoparticles occurred at 560 nm due to the charge transfer from the metal ion to TiO2 [34]. The abovementioned absorption in the range of visible irradiation may be due to the localized surface plasmon resonance (LSPR). It has been shown that in the mechanism using LSPR, metal can act as an electron trap and thus inhibit the recombination process of electron-hole pairs [36,37]. The characteristic LSPR band maximum at 555 nm [38] indicates the presence of Au nanoparticles in the CuN-Au/TiO2 sample, shown on the UV-Vis DRS spectra ( Figure S3).
It can be clearly seen in the TEM pictures ( Figure 4b) that the obtained gold nanoparticles have a spherical shape, and the particle size distribution of Au was in the range of 10-40 nm, with the highest contribution of the fraction from 15 to 25 nm and mean size equal to 22.1 nm (Figure 4a). The nanometric size and shape of the obtained Au particles a b c d e Very similar to the CuA-Au/TiO 2 TEM analysis results were the results obtained for the CuN-Au/TiO 2 sample, where the gold particles have a nanometric size and spherical shape; in addition, they were evenly distributed over the entire sample (Figure 4a-e). The difference in the photocatalytic activity of these samples was probably due to the precursors used (nitrate and acetate). The average size of the Au particles ranges between 10-50 nm, also less than 5% in the range 60-70 nm, with a significant predominance of particles with sizes from 10 to 30 nm, where the mean size was 24.7 nm (Figure 4a) Moreover, the Au particle has planes (111) and (200), which correspond to 0.24 and 0.20 nm, respectively. (Figure 4b).
The TEM micrographs of FeN-Au/TiO 2 shown in Figure 5 gave a bright observation of the sample in which the particles are much bigger than those obtained in the case of the samples containing copper (the mean sizes of the particles were 22.1 nm for CuN-Au/TiO 2 and 24.7 nm for CuA-Au/TiO 2 ). The particle size was found in the range of up to 60 nm and also in the range from 80 to140 nm, with the mean size equal to 60.6 nm. (Figure 5a). It should be emphasized that in the UV-Vis DRS spectrum ( Figure S3), no characteristic peak from the LSPR of gold particles was detected. Similar results were obtained by Duan et al., where despite the presence of gold particles in the sample, confirmed by other analytical methods, the characteristic peak of approx. 550 nm was not observed [39]. Similar results were obtained for the FeN-Au/TiO 2 sample; there were big agglomerates which caused an uneven distribution of Au particles on the TiO 2 ( Figure 5d,e), which could be the reason for the absence of the peak from Au NPs in the UV-Vis DRS spectra. However, a significant shift of absorption towards the visible spectrum (>400nm) for FeN-Au/TiO 2 in comparison with unmodified TiO 2 was observed on UV-Vis DRS ( Figure S3), and the presence of Fe 2p 3/2 on the surface of the catalyst (confirmed by XPS analysis) also indicates the probable surface incorporation of Fe and Au atoms on TiO 2 . The aim of introducing the Fe atoms to the sample was to increase the absorption toward the visible light wavelength and thus increase the photocatalytic activity in the visible range [40]. The distribution of Fe atoms observed on TEM connected with UV-Vis DRS and XPS results can testify to the surface incorporation of Fe atoms [41,42].
very beneficial for increasing the photocatalytic activity in the range of visible irradiation. They compared the photoactivity of different shapes of Au particles deposited on TiO2 microspheres and proved that visible light activity decreased in the following order: spheres > rods > stars [40]. It can also be added that in the case of the CuA-Au/TiO2 sample, the inter-planar spacing for gold was 0.24 nm, which corresponds to the separation between the (111) lattice planes of Au (Figure 4b), which was consistent with the results obtained by XRD analysis (Table 1).
Very similar to the CuA-Au/TiO2 TEM analysis results were the results obtained for the CuN-Au/TiO2 sample, where the gold particles have a nanometric size and spherical shape; in addition, they were evenly distributed over the entire sample (Figure 5a-e). The difference in the photocatalytic activity of these samples was probably due to the precursors used (nitrate and acetate). The average size of the Au particles ranges between 10-50 nm, also less than 5% in the range 60-70 nm, with a significant predominance of particles with sizes from 10 to 30 nm, where the mean size was 24.7 nm (Figure 5a) Moreover, the Au particle has planes (111) and (200), which correspond to 0.24 and 0.20 nm, respectively. (Figure 5b). The TEM micrographs of FeN-Au/TiO2 shown in Figure 6 gave a bright observation of the sample in which the particles are much bigger than those obtained in the case of the samples containing copper (the mean sizes of the particles were 22.1 nm for CuN-Au/TiO2 and 24.7 nm for CuA-Au/TiO2). The particle size was found in the range of up to 60 nm and also in the range from 80 to140 nm, with the mean size equal to 60.6 nm. (Figure 6a). It should be emphasized that in the UV-Vis DRS spectrum ( Figure S3), no characteristic peak from the LSPR of gold particles was detected. Similar results were obtained by Duan et al., where despite the presence of gold particles in the sample, confirmed by other analytical methods, the characteristic peak of approx. 550 nm was not observed [41]. Similar a b c d e In the SEM images of the bimetallic TiO 2 particles, presented in Figure 6, it was observed that the sample prepared using acetate precursor has a more sponge-like surface, with more cavities in the structure, compared to that obtained with nitrate precursor. CuN-Au/TiO 2 and FeN-Au/TiO 2 particles were poorly formed with highly irregular shapes.
The use of sonication improves catalyst deposition in a microreactor aided by enhanced mass deposition, which subsequently improves the overall photocatalytic conversion [43]. We link this activity to the higher availability of the active sites upon the deposition on the wall of the microreactor. On the deposition of these catalysts in the microflow system under the influence of ultrasound, the breakage of the agglomerations into smaller sizes was observed. The big aggregates (300-500 nm) in the case of CuA-Au/TiO 2 disturb the flow of the solution inside the microtube, reducing the activity of the catalyst (Figure 1). The small agglomerations (80-250 nm) and good dispersion in the case of CuN-Au/TiO 2 and FeN-Au/TiO 2 ( Figure 6) can lead to the improved transfer of light and better interaction of the reagent with the surface of the catalyst. The thickness of the deposited catalyst inside the wall of the microreactor was 5-7 µm, measured with an optical microscope image, hence the better photocatalytic activity in the microflow system for this catalyst.
caused an uneven distribution of Au particles on the TiO2 (Figure 6d,e), which could be the reason for the absence of the peak from Au NPs in the UV-Vis DRS spectra. However, a significant shift of absorption towards the visible spectrum (>400nm) for FeN-Au/TiO2 in comparison with unmodified TiO2 was observed on UV-Vis DRS ( Figure S3), and the presence of Fe 2p 3/2 on the surface of the catalyst (confirmed by XPS analysis) also indicates the probable surface incorporation of Fe and Au atoms on TiO2. The aim of introducing the Fe atoms to the sample was to increase the absorption toward the visible light wavelength and thus increase the photocatalytic activity in the visible range [42]. The distribution of Fe atoms observed on TEM connected with UV-Vis DRS and XPS results can testify to the surface incorporation of Fe atoms [43,44]. In the SEM images of the bimetallic TiO2 particles, presented in Figure 7, it was observed that the sample prepared using acetate precursor has a more sponge-like surface, with more cavities in the structure, compared to that obtained with nitrate precursor. CuN-Au/TiO2 and FeN-Au/TiO2 particles were poorly formed with highly irregular shapes. The use of sonication improves catalyst deposition in a microreactor aided by enhanced mass deposition, which subsequently improves the overall photocatalytic conversion [45]. We link this activity to the higher availability of the active sites upon the deposition on the wall of the microreactor. On the deposition of these catalysts in the microflow system under the influence of ultrasound, the breakage of the agglomerations into smaller  The synthesized CuA-Au/TiO 2 catalyst was considered for the oxidation of aromatic alcohols (vanillyl alcohol (VanOH), coniferyl alcohol (ConOH), and cinnamyl alcohol (CinOH) under UV (375 nm) light, as this catalyst showed better activity under UV light for BnOH oxidation (Figure 1a) in the batch system. Before light experiments, dark adsorption studies were performed to determine the adsorption/reactivity, and the duration to reach adsorption equilibrium was 30 min. Strong adsorption of VanOH (37%) and ConOH (30%) was observed on the surface of bimetallic CuA-Au/TiO 2 in the dark, which might be because of the complex formation between these alcohols with TiO 2 .
After 2 h of the light experiment, BnOH and VanOH conversion were 54% and 95%, respectively (Figure 7b). Comparing the structure of the molecules of BnOH and VanOH (Figure 7a), it can be seen that the main difference between these two alcohols was that VanOH has in its structure an OH group directly connected with the aromatic ring and also a methoxy (OCH 3 ) group connected with the next carbon atom of the aromatic ring. In contrast, benzyl alcohol has only an OH group at the end of the alkyl chain, which was not directly connected with the aromatic ring, and no methoxy group. As mentioned above, the OH and OCH 3 groups in the vanillyl alcohol are electron-donating groups which were providing the electrons to the aromatic ring. An excess of electrons causes lower production of aldehyde, which results in low selectivity towards aldehyde in the case of VanOH (25%) in comparison to BnOH (100%) under UV irradiation (Figure 7b) after 2 h of the light experiment. As the selectivity of producing vanillyl aldehyde (VanAld) was strongly reduced, other products of vanillyl alcohol oxidation were achieved ( Figures S5-S7). Furthermore, the lower selectivity can also be ascribed to the strong adsorption (37%) of alcohol on the surface of the catalyst, which can lead to partial oxidation, dimerization, or C-O, C-C coupling products.  Figures S5-S7). Furthermore, the lower selectivity can also be ascribed to the strong adsorption (37%) of alcohol on the surface of the catalyst, which can lead to partial oxidation, dimerization, or C-O, C-C coupling products. Considering ConOH and CinOH alcohols, it can be seen that very high conversions were achieved (95% and 81%, respectively). In the case of ConOH, the selectivity of producing coniferyl aldehyde (ConAld) was very low (5%), with 7% selectivity towards VanAld (Figure 8), with additional products verified by GC-MS analysis (Figures S5-S7). Along with the OCH3 and OH groups, ConOH has a double bond in its structure, which can potentially also provide the electrons to the aromatic ring, which causes further reduction, lowering the selectivity. The side products formed are benzaldehyde from CinOH and vanillyl aldehyde from ConOH (partial oxidation) [46], as well as 3-phenoxy Considering ConOH and CinOH alcohols, it can be seen that very high conversions were achieved (95% and 81%, respectively). In the case of ConOH, the selectivity of producing coniferyl aldehyde (ConAld) was very low (5%), with 7% selectivity towards VanAld (Figure 7), with additional products verified by GC-MS analysis (Figures S5-S7). Along with the OCH 3 and OH groups, ConOH has a double bond in its structure, which can potentially also provide the electrons to the aromatic ring, which causes further reduction, lowering the selectivity. The side products formed are benzaldehyde from CinOH and vanillyl aldehyde from ConOH (partial oxidation) [44], as well as 3-phenoxy benzaldehyde from ConOH (C-O coupling product), which probably was the result of many complex chemical reactions leading to the dimerization of ConOH under the influence of the oxidizing environment (UV radiation) [45]. The conversion of aromatic alcohols and the selectivity to its aldehyde were the main criteria to estimate catalyst performance as the other products were in traces. For CinOH oxidation, the selectivity towards cinnamyl aldehyde (CinAld) was 45%. Furthermore, the catalyst showed highly selective toward BnOH oxidation to BnAld.
After conducting the selective photocatalytic oxidation experiments in UV radiation, the CuA-Au/TiO 2 photocatalyst was selected for experiments also under visible light in the batch system (Figure 8). Under visible irradiation (515 nm), the satisfactory conversion of the alcohols was obtained only for ConOH and VanOH (93% and 52%, respectively), with 8% and 17% selectivity towards their respective aldehydes, with the formation of other side products. Increased photocatalytic conversion in the range of visible radiation for ConOH and VanOH, with a complete lack of activity for BnOH and CinOH, was probably related to the structures of these alcohols. In the case of BnOH and CinOH, there are no groups that could be a direct source of additional electrons for the aromatic ring and no group which can form complexes with TiO 2 . In order to explain this, UV-Vis DRS and FTIR analyses were performed for the aromatic alcohol-adsorbed TiO 2 complexes (Figures 9 and 10) [46,47].
Molecules 2022, 27, x FOR PEER REVIEW 12 other side products. Increased photocatalytic conversion in the range of visible radia for ConOH and VanOH, with a complete lack of activity for BnOH and CinOH, was p ably related to the structures of these alcohols. In the case of BnOH and CinOH, ther no groups that could be a direct source of additional electrons for the aromatic ring no group which can form complexes with TiO2. In order to explain this, UV--Vis DRS FTIR analyses were performed for the aromatic alcohol-adsorbed TiO2 complexes ( Fig  10 and 11) [48,49]. As a result of the adsorption of ConOH and VanOH on the TiO 2 catalyst, a significant change in the color of the catalyst from white to yellow can clearly be seen, which resulted in a shift of absorption toward visible radiation (Figure 9). A plausible reason can be complex formation by the ligand-to-metal charge transfer (LMCT) [48,49]. During this process, electrons are transferred from the highest occupied molecular orbital (HOMO) of substrates/adsorbates to the conduction band of TiO 2 upon visible light irradiation. (Scheme 1) [46]. Consequently, the samples forming the colorful complexes exhibited an excellent conversion of aromatic alcohols under visible light ( Figure 10).  As a result of the adsorption of ConOH and VanOH on the TiO2 catalyst, a significa change in the color of the catalyst from white to yellow can clearly be seen, which result in a shift of absorption toward visible radiation ( Figure 10). A plausible reason can complex formation by the ligand-to-metal charge transfer (LMCT) [50,51]. During th process, electrons are transferred from the highest occupied molecular orbital (HOMO) substrates/adsorbates to the conduction band of TiO2 upon visible light irradiatio (Scheme 1) [48]. Consequently, the samples forming the colorful complexes exhibited excellent conversion of aromatic alcohols under visible light (Figure 9). As a result of the adsorption of ConOH and VanOH on the TiO2 catalyst, a significant change in the color of the catalyst from white to yellow can clearly be seen, which resulted in a shift of absorption toward visible radiation ( Figure 10). A plausible reason can be complex formation by the ligand-to-metal charge transfer (LMCT) [50,51]. During this process, electrons are transferred from the highest occupied molecular orbital (HOMO) of substrates/adsorbates to the conduction band of TiO2 upon visible light irradiation. (Scheme 1) [48]. Consequently, the samples forming the colorful complexes exhibited an excellent conversion of aromatic alcohols under visible light (Figure 9). The formation of complexes between TiO2 and ConOH and TiO2 and VanOH was also supported by the FTIR analysis results (Figure 11). Characteristic IR bands for alcohols have also been observed for alcohol-adsorbed titania samples, especially in TiO2-Va-nOH and TiO2-ConOH ( Figure S8). This provided an indication of complex formation via The formation of complexes between TiO 2 and ConOH and TiO 2 and VanOH was also supported by the FTIR analysis results (Figure 10). Characteristic IR bands for alcohols have also been observed for alcohol-adsorbed titania samples, especially in TiO 2 -VanOH and TiO 2 -ConOH ( Figure S8). This provided an indication of complex formation via the adsorption of alcohol.
It was mentioned above that the adsorption of alcohols (VanOH and ConOH) on the surface of TiO 2 occurred by a dissociative mechanism, which means that OH and O-CH 3 groups interacted between titania and aromatic alcohols. In that case, alcohol and TiO 2 would be linked via the C-O bond. The presence of C-O stretching vibrations was observed at 1000-1200 cm −1 (Figure S8), which can be seen in the case of TiO 2 -VanOH and TiO 2 -ConOH; two bands were observed in this region (1118 and 1158 cm −1 ). The bands in the range of 1000-1050 cm −1 were assigned to the presence of =C-H bonds derived from the aromatic ring ( Figure 10 and Figure S8). The bands around 1460 cm −1 originated from CH 2 vibrations and around 1640 cm −1 from C=C [46]. Similar spectra were observed for bimetallic CuA-Au/TiO 2 as well, confirming the formation of the complex with the above aromatic alcohol ( Figure S8), which confirms that the metals did not take part in this complex formation. It is worth adding that the oxidation of VanOH and ConOH was successful under visible irradiation due to the formation of the abovementioned complex with TiO 2 . To further confirm the role of the OH group in LMCT complex formation, the synthesized TiO 2 was calcined at a high temperature (600 • C) to remove surface OH through condensation ( Figure S9). Interestingly, this catalyst was found inactive, suggesting the OH groups are crucial for LMCT complex formation and the visible light activity of the catalyst.
The experiment with the CuN-Au/TiO 2 -coated microreactor for other aromatic alcohols oxidation was performed under visible light irradiation as (1) CuN-Au/TiO 2 in the microflow system showed better activity ( Figure 1) and (2) the other aromatic alcohols showed better conversion under visible light (Figure 8). Like the batch experiment, BnOH and CinOH did not show any activity in the microflow system under visible light, whereas ConOH and VanOH did. The specific conversion rate of ConOH was higher (198 µmol/m 2 h) compared to VanOH (46 µmol/m 2 h), with 11 % and 10 % selectivity towards their respective aldehydes ( Figure 11). Other side products, such as batch experiments, were also observed. TiO2-ConOH; two bands were observed in this region (1118 and 1158 cm −1 ). The bands in the range of 1000-1050 cm −1 were assigned to the presence of =C-H bonds derived from the aromatic ring (Figures 11 and S8). The bands around 1460 cm −1 originated from CH2 vibrations and around 1640 cm −1 from C=C [48]. Similar spectra were observed for bimetallic CuA-Au/TiO2 as well, confirming the formation of the complex with the above aromatic alcohol ( Figure S8), which confirms that the metals did not take part in this complex formation. It is worth adding that the oxidation of VanOH and ConOH was successful under visible irradiation due to the formation of the abovementioned complex with TiO2.
To further confirm the role of the OH group in LMCT complex formation, the synthesized TiO2 was calcined at a high temperature (600 °C) to remove surface OH through condensation ( Figure S9). Interestingly, this catalyst was found inactive, suggesting the OH groups are crucial for LMCT complex formation and the visible light activity of the catalyst.
The experiment with the CuN-Au/TiO2-coated microreactor for other aromatic alcohols oxidation was performed under visible light irradiation as (1) CuN-Au/TiO2 in the microflow system showed better activity ( Figure 2) and (2) the other aromatic alcohols showed better conversion under visible light (Figure 9). Like the batch experiment, BnOH and CinOH did not show any activity in the microflow system under visible light, whereas ConOH and VanOH did. The specific conversion rate of ConOH was higher (198 μmol/m 2 h) compared to VanOH (46 μmol/m 2 h), with 11 % and 10 % selectivity towards their respective aldehydes ( Figure 12). Other side products, such as batch experiments, were also observed. After deposition, the catalyst's surface was more selective towards ConAld for ConOH oxidation than in the batch system (8%). The elevated selectivity in the batch system compared to the microflow system might be because of the higher availability of the catalyst's active sites upon deposition. Additional side products (as discussed above) were After deposition, the catalyst's surface was more selective towards ConAld for ConOH oxidation than in the batch system (8%). The elevated selectivity in the batch system compared to the microflow system might be because of the higher availability of the catalyst's active sites upon deposition. Additional side products (as discussed above) were confirmed from GC/MS; however, because of the trace amount of some products, it was not easy to calculate the selectivity towards them ( Figure S7). The sol-gel-synthesized CuN-Au/TiO 2 catalyst showed promising activity in the batch system as well as in the microflow system for ConOH oxidation. Furthermore, from the catalyst-deposited microtubes, no leaching was confirmed with the ED-XRF analysis after 60 min of the experiment ( Figure S10). The recycling test with the best performing catalyst (CuA-Au/TiO 2 ) and the one which showed lower activity (TiO 2 ) in the batch system were performed by washing the catalyst with acetonitrile and water, and the activity of the catalyst was retained ( Figure S11).

Catalyst Synthesis
The bimetallic Cu-Au/TiO 2 and Fe-Au/TiO 2 nanoparticles were prepared based on the previously established sol-gel method [50]. The required amount of metal precursors (ratio of atomic percent of Au to Cu or Fe was adjusted to 1:4) were added to 15 mL of isopropanol and stirred (magnetic stirrer) at 600 rpm. Under the rotational conditions, 4 mL of TTIP was added dropwise to the mixture. Subsequently, milli-Q water (5 mL) was added to the solution at a rate of 0.167 mL/min using a syringe infusion pump (Programmable Double Syringe Pump (WPI), NE-4000). After 6 h of aging with mixing at 1000 rpm, the resulting solid samples were obtained by centrifugation, washing with deionized water (2 times) and ethanol (1 time), and drying at 80 • C in an oven for 24 h.

Microreactor Preparation
A visible light transparent perfluoroalkoxy alkane (PFA, 0.8 mm ID, BOLA: S 1811-02) tube was used as a microreactor [50,51]. A 0.5 g/L concentration (previously optimized [50]) of nanoparticles was dispersed in Milli-Q water by ultrasonication for 15 min using an ultrasonic bath (Elma Elmasonic P, 37 kHz, 70% power). Twenty mL of the homogeneous nanoparticle suspension was passed through the cleaned PFA microtube under the influence of ultrasound using a syringe pump. The spiralized fragment (Figure 12b) was the effective length subjected to the ultrasound treatment (Elma Elmasonic P, 37 kHz, 100% power) for 75 min (the flow rate of the suspension was 0.26 mL/min). The tube was placed in the oven for 24 h at 80 • C, and later cleaned by passing Milli-Q water and ethanol, dried with airflow, and then placed in the oven for 1 h at 80 • C. After this procedure, the catalyst-coated PFA tube was used for photocatalytic experiments (Figure 12b).
influence of ultrasound using a syringe pump. The spiralized fragment (Figure 1b) w the effective length subjected to the ultrasound treatment (Elma Elmasonic P, 37 k 100% power) for 75 min (the flow rate of the suspension was 0.26 mL/min). The tube w placed in the oven for 24 h at 80 °C, and later cleaned by passing Milli-Q water and et nol, dried with airflow, and then placed in the oven for 1 h at 80 °C. After this procedu the catalyst-coated PFA tube was used for photocatalytic experiments (Figure 1b).

Catalytic Performance Test
The photocatalytic oxidation of the aromatic alcohols over the synthesized cataly was carried out in batch ( Figure 1a

Catalytic Performance Test
The photocatalytic oxidation of the aromatic alcohols over the synthesized catalysts was carried out in batch ( Figure 12a) and microflow reactors (Figure 12b, Figure S1). UV-LED and Vis-LED systems were used as light sources (375 nm and 515 nm wavelength, respectively). The flow rate was set at 0.167 mL/min (after optimization) to reproduce enough space-time according to the reactor's dimensions, and the whole experiment was carried out for 60 min [50]. Experiments in the batch photocatalytic reactor were performed using 20 mL of the reaction solution and 0.5 g/L of catalyst concentration for 60 min under UV light. Alcohol conversion, the specific conversion rate, and selectivity to each product were calculated according to the following equations: Conversion (%) = ((Converted moles of aromatic alcohols)/(Initial moles of aromatic alcohols)) × 100% Selectivity (%) = ((Produced moles of aromatic aldehydes)/(Converted moles of aromatic alcohols)) × 100% Specific Conversion Rate (µmol/m 2 ·h) = (C Oho − C OHt )/(S C × time) where C OHo = initial concentration of aromatic alcohols, C OHt = concentration of aromatic alcohols after time, t, and S C is the specific surface area (surface area multiplied by the catalyst concentration) of the catalyst taking part in the photocatalytic conversion.

Characterization
The synthesized samples were characterized using diffuse reflectance spectroscopy (DRS, Shimadzu UV-2600i), and the bandgap was calculated from the Tauc plot. Powder X-ray diffraction (XRD) measurements were performed employing the Bragg-Brentano configuration. This type of arrangement was provided using a PANalytical Empyrean diffraction platform, powered at 40 kV × 40 mA and equipped with a vertical goniometer, with theta-theta geometry using Ni-filtered CuKα (λ = 1.5418 Å) radiation. The elemental maps of the samples were also obtained by FEI Nova Nanolab 200 scanning electron microscopy (SEM). The textural properties of TiO 2 were determined by N 2 physisorption using a micromeritics automated system (Micromeritics Instrument Corporation, Norcross, GA, USA) with the Brunauer-Emmet-Teller (BET) and the Barret-Joyner-Halenda (BJH) methods. Before adsorption, the samples were degassed under vacuum (0.1 Pa) for 12 h at 80 • C. The presence of functional groups on the surface of the catalyst and substrate was determined using a Bruker Tensor II Fourier transform (FT) IR spectrometer.
The samples collected from the outlet of the catalyst-deposited PFA capillary were examined using the energy dispersive X-ray fluorescence (EDXRF) spectrometer (Mini-Pal 4, PANalytical Co., Malvern, UK) with a Rh tube and silicon drift detector to check the titania residual and other metals. The spectra were collected in an air atmosphere, without a filter, at a tube voltage of 30 kV. To identify and quantify the alcohols, aldehydes, and acids present, the collected samples were analyzed using high-pressure liquid chromatography (HPLC, Waters) with mass spectrometry using a mobile phase containing a mixture of organic solvents and a 0.05% H 3 PO 4 (5M) aqueous solution (CH 3 CN:CH 3 OH: H 2 O = 20:2.5:77.5 v/v).

Conclusions
In this study, we have made an attempt to promote green chemistry by improving the utilization of lignin-an important waste from the paper and pulp industry, which otherwise would have been disposed of, thus contaminating the environment. To this end, we propose to improve the conversion of the model alcohols by targeting and improving the catalytic activity of TiO 2 doped with bimetals such as Cu-Au, and Fe-Au. In the batch system, as high as 100% benzaldehyde (BnAld) selectivity was obtained for benzyl alcohol (BnOH) conversion of~60% using CuA-Au/TiO 2 photocatalyst. The high conversion could be explained by the high average pore size (6.8 nm) and better crystallinity of the CuA-Au/TiO 2 catalyst, which were confirmed in subsequent XRD and N 2 physisorption analysis.
Our studies further revealed that the LMCT complex formation of TiO 2 with the methoxy (OCH 3 ) and OH groups (directly connected with the aromatic ring) exists in the structure of coniferyl alcohol (ConOH) and vanillyl alcohol (VanOH), which was crucial to activate the catalyst under visible light. This hypothesis has been further corroborated by UV-Vis and FTIR studies. Additionally, the whole process was green/environmentfriendly as the catalyst synthesis process did not include any high-temperature calcination steps, unlike commercial P25 TiO 2 , and the photocatalytic selective oxidation route was additive-free (no additional molecular oxygen). We also developed a catalyst-decorated microreactor using ultrasonic irradiation, which helped to increase the turbulence of the liquid phase and to improve the active surface area of our catalyst via the de-agglomeration and fragmentation of the nanoparticles.
In a broader context, we believe that the presented work demonstrates the potential of an ultrasonic-assisted bimetallic TiO 2 wall-coated microreactor for selective oxidation of lignin-based model compounds using solar energy and this will serve as a conceptual blueprint for further developments.
Author Contributions: S.R.P.: writing-original draft preparation, data analysis, catalyst characterization, conceptualization, investigation, and validation. M.P.-G.: writing-review and editing and data curation. D.Ł.: writing-review and editing and visualization. D.L.: investigation and data processing. J.C.C.: writing-review and editing, resources, conceptualization, supervision, and project administration. All authors have read and agreed to the published version of the manuscript.