Ti plates of 1 mm thickness were cut by guillotine to be 5.0 cm × 8.0 cm. To smooth the surface of the plates, they were sanded with 800, 1,000, 1,200, and 1,500 grit sandpapers, respectively. Then, they were cleaned in an ultrasonic bath for 10 min in acetone, ethanol and distilled water, respectively.

The Ti plates were then chemically cleaned in a solution medium containing 4% HF, 31% HNO₃, and 65% water for 30 s (Özcan et al., 2018). Afterward, the Ti plates were again cleaned in an ultrasonic bath in water for 10 min and dried at room temperature to remove all contaminating ions coming from the use of HF and HNO₃.

The two-electrode system in which the anodic oxidation process is applied to form nano-structured TiO₂ samples on the Ti plate surface is shown in Supplementary Figure S1. The solution in which anodic oxidation will be carried out was prepared by dissolving 0.3% by mass of NH₄F in a solution containing 2% water and 98% ethylene glycol by volume (Özcan et al., 2018). Anodic oxidation was carried out in this prepared electrolyte solution by applying a definite constant voltage (50 or 60 V) to the electrodes for a certain time (3 or 5 h) and washed with pure water. These chosen values are similar to those used in the literature (Özcan et al., 2018). Then, the anode was placed in a beaker containing pure water and subjected to an ultrasonic bath, which allowed TiO₂ in nanotube/nanowire structure to break and separate from the plate (Ali and Hannula 2017; Ali et al., 2018; Yang et al., 2019).

A sufficient amount of TiO₂ nanoparticles were obtained by applying this process repeatedly. The crystallinity of the obtained TiO₂ nanoparticles was very low. In order to obtain photocatalysts of different crystallinity, the partitioned TiO₂ photocatalyst was kept in an air atmosphere for 3 h (Protherm, PLF-110/10 model) (temperature rise rate: 3°C/min) in a muffle furnace at different temperatures (100–700°C).

The resulting photocatalysts were named TiO₂-xV-yh-z. Here, x: voltage value applied while preparing photocatalyst by anodic oxidation, y: time (h) during which the voltage was applied, and z: thermal treatment temperature of the anode. For example, the powdered photocatalyst TiO₂-60V-3h-250 was prepared at 60 V for 3 h, separated from the titanium plate by means of an ultrasonic bath, and calcined at 250°C. The photoanode used without thermal treatment for the sake of comparison was named TiO₂-xV-yh-25 because it was prepared at room temperature (ca. 25°C).

The potential value applied in the production of all nano-structured electrodes produced by anodic oxidation was reached by increasing ∼0.2 V per second from a voltage of 0.0 V.

The produced nano-structured TiO₂ samples were characterized by X-ray diffraction (XRD) which was performed with a Bruker D8 Advance diffractometer using the Cu Kα radiation (1.544 Å) and a 2θ scan rate of 1.281°/min, while scanning electron microscope observations (SEM) were performed with an FEI microscope (NanoSEM 650 model, FEI Company). The samples on the stab were covered with a thin gold film before SEM observations. XRD analyses were performed to establish the crystal phase(s) and the primary particle sizes of the photocatalysts, while SEM analyses to determine their morphological properties.

BET specific surface area values of the prepared samples were determined by the multi-point BET method using a Micromeritics (Gemini VII model) apparatus. The nano-structured TiO₂ photocatalysts were degassed at 200°C for 5 h before the measurement. The crystalline and morphologic properties of the samples could be changed only slightly by the used degassing temperature, however we could not perform degassing process at low temperatures because the pores would not be completely free and the measurement would be wrong (i.e., room temperature).

A 250 mL Pyrex beaker was used as a photoreactor for experiments performed under UVA light (Supplementary Figure S2). UVA rays reaching the suspension were provided by four fluorescent lamps (Philips, each lamp is 8 W) emitting mainly at 365 nm. The spectrum of the lamp is shown in Supplementary Figure S3. The distance to the suspension of these lamps was 6.8 cm, and the lamps were placed parallel to each other. The average value of the radiation energy reaching the suspensions, determined using a radiometer (Delta Ohm, DO9721), was 21 W m⁻² in the range 315–400 nm.

The initial substrate concentration for glycerol was 2 or 10 mM while for 3-pyridinemethanol was 0.50 mM. The amount of used photocatalyst was 0.20 g L⁻¹. Before the lamps were turned on, the suspension was stirred in the dark for 30 min to reach thermodynamic and adsorption-desorption equilibrium. Moreover, the suspension was continuously stirred magnetically at 500 rpm during the photocatalytic experiments that were carried out at pH 7 and at room temperature (ca. 25°C). The aqueous suspension was in contact with the atmosphere which provided the necessary oxygen. Samples withdrawn at certain time intervals were filtered through a hydrophilic membrane (0.45 µm, HA, Millipore) before being injected in the high performance liquid chromatography (HPLC) and total organic carbon (TOC) analyzers. All the photocatalytic experiments were carried out twice or three times to check their reproducibility.

Qualitative and quantitative analyses of the samples withdrawn from the experimental system at certain times during the photocatalytic experiments of glycerol and 3-pyridinemethanol were performed with HPLC (Shimadzu Prominence LC-20A model and SPD-M20A photodiode array detector). Two consecutive C-18 columns were used (Phenomenex and Syncronis, respectively), and the column temperature was kept at 30°C. Retention times of analyzed substrates and intermediates and UVA spectra of compounds were compared with known standards (Sigma-Aldrich, purity ≥ 98%). For glycerol and its intermediates, the mobile phase was a 5 mM aqueous solution of sulfuric acid, and the flow rate was 0.3 cm³ min⁻¹. For 3-pyridinemethanol and its intermediates, the mobile phase was 40% methanol and 60% deionized water, and the flow rate was 0.2 cm³ min⁻¹.

Total organic carbon (TOC) analyses were performed with a Shimadzu (TOC-LCPN model) device to determine the amount of mineralized carbon dioxide during the oxidation of the substrates studied. The selectivity towards the main products (%) and the substrate conversion (%) have been calculated according to the formula below: %   Selectivity = ( Amount   of   product   formed , mmol ) ( Amount   of   substrate   reacted , mmol ) x 100 % %   Conversion = ( Amount   of   reacted   substrate , mmol ) ( Initial   amount   of   substrate ,   mmol ) x 100 %

Figures 1–3 show XRD patterns of nanotube/nanowire-structured TiO₂ photocatalysts calcined at different temperatures (up to 700°C). Since the behaviors of all three sample types (TiO₂-60V-5h, TiO₂-60V-3h, and TiO₂-50V-3h) with respect to heat treatment were similar, TiO₂-60V-5h was calcined at several different temperatures, while TiO₂-60V-3h and TiO₂-50V-3h samples were calcined only at 250, 500, and 700°C. The peak values ​​at 2θ = 27.5°, 36.5°, 41.0°, 54.1°, and 56.5° refer to the rutile phase, while those at 2θ = 25.58°, 38.08°, 48.08°, 54.58° belong to the anatase phase (Yurdakal et al., 2012; Fiorenza et al., 2020). Low and broad XRD peaks are indicative of a low crystallinity, consequently, uncalcined TiO₂ samples showed to have a low crystallinity (high amorphous content) and a small amount of anatase phase content (Yurdakal et al., 2012; Çetinkaya and Yurdakal 2021). For all samples, as the calcination temperature increased up to 500°C, the intensity and the sharpness of the anatase phase peak increased also. Consequently, TiO₂ crystallinity increased by increasing the thermal treatment temperature. At 700°C, a sharp peak belonging to the rutile phase was also formed. The XRD peak intensity of the anatase phase decreased at 700°C with respect to the sample calcined at 500°C, because the transformation from anatase phases to rutile occurred.

Supplementary Figure S4 shows the XRD patterns of commercial TiO₂ (Merck and Degussa P25) photocatalysts. These photocatalysts, which are not in nanotube/nanowire structure, are in powder form and have been used to compare the investigated home prepared photocatalysts. Merck TiO₂ photocatalyst is entirely in the anatase phase, while Degussa P25 is in the anatase and rutile phases (ca. 82% A, 18% R).

In Table 1, phases of TiO₂ samples, percentage of each crystal phase, and primary particle size values ​​of each phase calculated from the Scherrer equation are reported. As the calcination temperature of the photocatalysts raised, the primary particle sizes increased due to the sintering of the particles. When the TiO₂-60V-5h samples were examined, although the primary particle size raised slightly up to 300°C, it showed a significant increase at 500°C and especially at 700°C. The primary particle size value was only 10 nm at 300°C, while it increased at 18 and 34 nm at 500 and 700°C, respectively. Both anatase and rutile phases were formed for all samples calcined at 700°C. Anatase and rutile percentages of these samples are different. The rutile phase was dominant in TiO₂-60V-5h-700, while anatase phase was present in other samples (especially TiO₂-60V-3h-700). The primary particle size of the anatase phase for all samples was about 33 nm, while that of the rutile phase 52 nm. In Degussa P25, the most common commercial photocatalyst, the primary particle size values ​​of both phases are close to each other and are approximately 20 nm.

Figure 4 shows the nitrogen adsorption-desorption isotherms of the prepared TiO₂ samples, and the results derived from these isotherms are reported in Table 2. All the prepared TiO₂ photocatalysts showed the type IV isotherm characteristic with an H3-type loop according to the IUPAC classification, indicating the irregular structure of the inner surface of the pores (Sing et al., 1985; Yurdakal et al., 2019; Zhang et al., 2022). The samples prepared without thermal treatment showed values of specific surface area in the range ca. 78–89 m²/g; however, their surface area decreased to ca. 47–54 m²/g after calcination at 500°C. A drastic reduction of the surface area up to 7.4 m²/g occurred after the treatment at 700°C. This finding is due to the sinterization of the nanotubes at a very high temperature (see SEM results). The BET isotherms of the three different samples (TiO₂-60V-5h, TiO₂-60V-3h and TiO₂-50V-3h) showed the same behavior if the thermal treatment temperature was the same, 25, 500, or 700°C (see Figure 4).

BJH Adsorption cumulative volume values (between 1.7 and 300 nm width) of pores of the samples, treated thermally at room temperature and 500°C, are similar. In contrast, that of the treated ones at 700°C decreased significantly. For instance, the pore volume of TiO₂-60V-5h-500 was 0.261 cm³/g, while TiO₂-60V-5h-700 was only 0.0957 cm³/g because of the sinterization of the TiO₂ nanotubes. Indeed, the BET adsorption-desorption isotherm of TiO₂-60V-5h-700 is very different from those of TiO₂-60V-5h-25 and TiO₂-60V-5h-500; it has a tiny hysteresis loop compared with the other two. In addition, BJH adsorption mean pore width values of the samples increased with the heat treatment temperature as there is an opposite trend between pore volume and width. Consequently, as seen in Figures 5, 6, pore size distribution curves were shifted to larger values by increasing the temperature of thermal treatment. For instance, peak maximum of the pore width vs. pore volume curves of TiO₂-60V-3h-25, TiO₂-60V-3h-500, and TiO₂-60V-3h-700 are ca. 13, 22, and 44 nm, respectively (see Figure 6B).

In this section, also some SEM images of nanotube/nanowire-structured TiO₂ samples prepared using the anodic oxidation method on the Ti surface at different potential values and different times are discussed. In order to obtain a sufficient amount of photocatalyst, TiO₂ structures were formed on the Ti surface many times by the anodic oxidation method and removed from the surface by ultrasonic treatment.

Figure 7; Supplementary Figure S5 show SEM images of TiO₂-60V-5h-25 nanoparticles taken at different magnifications. The TiO₂ particles show a heterogeneous distribution of their size from a few microns to 400 μm, but the size of most of them ranges between 50 and 100 µm. Figure 7 reports SEM images taken from the top of the nanomaterial. These pictures show that the structure is porous and in a nanotube structure. However, there are also nanotube structures with closed mouths on the surface. Supplementary Figures S5B–D are cross-sectional views of TiO₂-60V-5h-25 material. From these images, the diameter of the TiO₂ nanotube is between 100 and 150 nm, and according to the picture in Supplementary Figure S5C, the nanotube length is about 8 µm.

In Figure 7 the SEM images of the same material (TiO₂-60V-5h) calcined in the air at 250°C (TiO₂-60V-5h-250) can be observed. Figure 7 is top images of the material, with a clear nanotube structure. Supplementary Figures S6A,B are filled nanotube/nanowire photos. In the picture of Supplementary Figure S6B, it can be noticed that some nanowire structures are separated from each other. The diameters of these nanowires are between about 60 and 90 nm. These SEM images indicate that some nanowire structures formed together with the nanotube structure. The separation of some of the nanowire structures from each other also increases the effective surface area of the material. Supplementary Figure S6C is a cross-sectional view of this material, which shows that the section thickness of the nanostructure is approximately 18 µm.

SEM images of TiO₂-60V-5h-700 are provided in Figure 7. Figure 7; Supplementary Figure S7A are images of the upper part of the material that appears to be in nanotube structures. Due to the high calcination temperature of 700°C, its crystallinity is quite high and some sintering can be observed in Figure 7. Indeed, the BET surface area of this sample was very low with respect to the uncalcined ones because of the partial sintering of the nanostructures; 12.2 vs. 78.3 m²/g. Supplementary Figure S7B is a side sectional view of the same material.

SEM images of the TiO₂-50V-3h-25 nanomaterial are given in Figure 7. Supplementary Figure S8A shows the overall particle distribution. This photocatalyst has smaller particles (up to about 100 µm) than the “TiO₂-60V-3h” materials. Figure 7; Supplementary Figure S8B show the nanotube structures in the cross-sectional view. These photos were taken from the upper part of the material, and nanotube structures with a homogeneous distribution are evident. Supplementary Figure S8D reports an image of the bottom of the material. According to the cross-sectional view, the length of the nanotubes is approximately 9 µm (Supplementary Figure S8E). In some places, the nanotubes were separated from the main structure, individually or combined (Figure 7). The diameter of these nanotubes is approximately 140–160 nm.

SEM photos of TiO₂-60V-3h-25 sample are given in Supplementary Figure S9. Large nanoparticles up to about 180 µm are present (Supplementary Figure S9A). There are also smaller particles of the order of nanometers in size. Supplementary Figure S9B indicates that the material is mostly in nanowire structure. The diameter of the nanostructures is about 100–135 nm (Supplementary Figures S9E,F). Supplementary Figure S9C shows the traces left by the broken nanowires on the surface. According to Supplementary Figure S9D, the layer length of the nanomaterial is approximately 8 µm. Supplementary Figure S10 gives an SEM picture of the same material calcined at 500°C. The nanostructures in Supplementary Figure S10B appear to consist of tiny crystals with a 20–30 nm diameter.

Photocatalytic glycerol oxidation experiments were carried out under UVA irradiation at pH 7 to work under environmentally friendly conditions (Yurdakal et al., 2017). No activity was detected in the presence of light but in the absence of photocatalyst or in the dark in the presence of photocatalyst (Augugliaro et al., 2012). Light, photocatalyst and oxygen were essential for the occurrence of the photocatalytic reaction.

Table 3 shows the results of photocatalytic experiments performed with 10 mM initial concentration of glycerol. The conversion values ​​are low for our prepared samples due to the high initial concentration. From the qualitative and quantitative analyses of the samples during the reaction, glyceraldehyde (GAD), 1,3-dihydroxyacetone (DHA), and formic acid (FA) were determined as the main intermediates (see Scheme 1A). The low carbon balance is due to the formation of undetected aliphatic species. We could not detect the other intermediates, because of their trace amounts. In addition, CO₂ was formed as a mineralization product. In the tables of the photocatalytic experiments conversion, selectivity (vs. the formation of GAD, DHA and FA), TOC decrease (%), and the pH values ​​of the suspension at the end of the experiment after 8 h are reported. TOC analyses were carried out only for some experiments. In general, the pH value decreased due to the formation of FA and other acidic species. While the photocatalysts prepared on the Ti plate surface showed low glycerol conversion (up to 10%), the commercial Degussa P25 photocatalyst gave rise to a 27% conversion that was the highest value among the photocatalysts tested, and consequently the highest TOC decrease was achieved with this photocatalyst. First of all, the experiments were carried out with the TiO₂-60V-5h photocatalyst at different temperatures. TiO₂-60V-5h-500 photocatalyst gave the highest conversion with 9%. Therefore, other photocatalysts were calcined at 500°C (TiO₂-50V-3h-500 and TiO₂-60V-3h-500) and were also used for comparison purposes. Among them, TiO₂-50V-3h-500 showed 10% conversion. This photocatalyst is in the anatase crystalline phase. Photocatalysts with low heat treatment showed to have scarce activities. Since the selectivity values were calculated at different reaction times, there is an inverse relationship between the selectivity values and the conversion values. In fact, as the reaction progresses, the intermediate products formed are more likely to reoxidize several times.

Therefore, the TiO₂-60V-5h-200 photocatalyst with the lowest activity showed the highest GAD and DHA selectivity, i.e., 56% and 17%, respectively. Although the photocatalysts have a porous structure, as they detached from the Ti surface in large pieces (in microns), their effective surface is less than those in powder and showed a lower activity than the commercial ones. Furthermore, the selectivity values towards GAD are higher than those towards DHA. This is due to the fact that to obtain DHA it is necessary to oxidize the carbon in position two of glycerol, while for the formation of GAD it is necessary to oxidize one of the carbons in position 1 or 3. Consequently, the probability of obtaining GAD is greater (Augugliaro et al., 2010).

Since the conversion observed during photocatalytic experiments carried out with an initial concentration of 10 mM glycerol was low, a set of photocatalytic experiments were also performed starting from an initial concentration of 2 mM glycerol (see Table 4). Due to the low initial glycerol concentration, the amount of GAD in some experiments was too low to be analysed, and it was not reported in table. As the calcination temperature of each type of nanomaterial increased, the glycerol conversion values increased, and the highest conversions were obtained with samples calcined at 700°C. The most significant result was found for the TiO₂-50V-3h photocatalyst, with glycerol conversions of 5, 14 and 25% for TiO₂-50V-3h-25, TiO₂-50V-3h-500 and TiO₂-50V-3h-700, respectively. In particular, the TiO₂-50V-3h-700 sample proved to be much more photoactive per surface unit than the corresponding samples treated at lower temperatures. This finding indicates that crystallinity, higher at higher temperatures, is the most important factor for the photocatalytic oxidation of glycerol. Glycerol, a highly hydrophilic molecule, could interact very strongly with the surface of uncalcined TiO₂ samples which are mainly in the amorphous phase (Augugliaro et al., 2008). Furthermore, its oxidation products (i.e., DHA and GAD) could remain on the surface blocking the active sites for a longer time than what happens on hydrophobic surfaces. Conversely, glycerol and its oxidation products could more easily desorb from crystalline and probably less hydroxylated surfaces, which would justify greater activity of the latter (Augugliaro et al., 2010). For this reason the Merck TiO₂ sample was shown to possess the highest activity despite its specific surface area being quite small (approx. 11 m²/g).

Some works were published on the partial photocatalytic oxidation of glycerol (Augugliaro et al., 2010; Imbault et al., 2020; Payormhorm and Idem 2020; Rangarajan et al., 2021). Previously, we performed photocatalytic oxidation of glycerol in aqueous suspensions containing home-prepared TiO₂ photocatalysts (Augugliaro et al., 2010). The initial glycerol concentration (50–180 mM) and the reaction time were higher than in the present work. Moreover, the photocatalysts were bulk anatase, rutile, or anatase–rutile polymorphic phases. In that work, Merck and Degussa P25 TiO₂ photocatalysts showed higher activity/selectivity than the home-prepared ones, like in the present study. The fact that a high crystallinity of the photocatalysts is beneficial for the activity is confirmed by the present investigation. In addition, we also detected the same intermediates. Consequently, the TiO₂ morphology, whether nanotube/nanowire or bulk structure, does not play a significant role in the product nature and distribution.

Payormhorm and Idem (2020) used C-doped TiO₂ prepared by a sol-microwave method for photocatalytic glycerol oxidation under visible light irradiation. Mainly DHA, GAD, and FA products and trace amounts of acetic acid were obtained. Unlike Payormhorm and Idem (2020), in our work, we used UVA irradiation and did not detect acetic acid.

Imbault et al. (2020) investigated photocatalytic oxidation of glycerol to DHA, GAD, glyceric acid, and several other chemicals using TiO₂ photocatalysts under simulated solar irradiation and in acetonitrile as solvent. Reaction rate and selectivity towards the products were higher in acetonitrile compared with an aqueous suspension. For instance, after 6 h of reaction, glycerol conversion and selectivity towards DHA were 96.8% and 17.8% in acetonitrile compared to 36.1% and 14.7% in water, respectively.

Rangarajan et al. (2021) prepared Ag-AgBr/TiO₂ ternary nanocomposites for selective photocatalytic glycerol oxidation into GAD and DHA in a non-aqueous solvent under visible light and aerobic conditions. Although a lower glycerol conversion was obtained in acetonitrile than in water after 4 h, a significantly higher yield of DHA (13%) and GAD (19%) was obtained due to the less significant formation of non-selective ˙OH radicals in the organic solvent.

Oxidation of 3-pyridinemethanol, an aromatic primary alcohol, was also carried out with the same photocatalysts used for glycerol oxidation. 3-Pyridinemethanal and vitamin B₃ were determined as main products. Scheme 1B shows a proposed mechanism for photocatalytic oxidation starting from 3-pyridinemethanol. The first step can be considered the interaction of h⁺ or hydroxyl radicals with formation of 3-pyridinemethanal subsequently oxidized to vitamin B₃. The over-oxidation products of these substrates and their corresponding molecules were mainly aliphatic species and CO₂ (Yurdakal et al., 2020).

Table 5 shows the results of the photocatalytic oxidation tests of 3-pyridinemethanol oxidation under UVA light in the presence of TiO₂-50V-3h photocatalysts obtained at different thermal treatment temperatures. Initial reaction rates, pseudo-first-order rate constants, substrate conversion values ​​for one and 3 h reaction times, selectivity towards 3-pyridinemethanal and vitamin B₃ after 15% and 30% conversion, and pH values ​​of the suspension at the end of the reaction are given in the table. Only 15% conversion was achieved after 3 h with the TiO₂-50V-3h-25 photocatalyst. This photocatalyst appeared to be in the anatase phase, but its crystallinity was low. Although crystalline samples showed higher activity, it can be noted that TiO₂-50V-3h-250 photocatalyst was both the most active and the most selective towards 3-pyridinemethanal and vitamin B₃ formation. In fact, the k-value for the 3-pyridinemethanol oxidation of this photocatalyst was about 3 times that of the TiO₂-50V-3h-25 photocatalyst (0.156 h⁻¹ vs. about 0.052 h⁻¹). In addition, 87% 3-pyridinemethanal and 11% vitamin B₃ selectivities were obtained after 30% conversion.

In this work seems that the temperature of 250°C, at least for the samples prepared as above described, is the optimum choice both for the conversion of the substrates studied and for production of 3-pyridinemethanal and vitamin B₃ (see Table 5). It should be noted that the significantly higher selectivity values towards 3-pyridinemethanal compared to vitamin B₃ may be due to the relatively short reaction times and consequently to the low conversion values. In fact, the first oxidation product deriving from 3-pyridinemethanol is aldehyde (3-pyridinemethanal), while acid vitamin B₃ is subsequently formed. The amount of vitamin B₃, therefore, increased with the reaction time, while the amount of 3-pyridinemethanal decreased. The carbon balance values were quite good, indicating that the CO₂ formation resulting from complete mineralization of the substrate was negligible; consequently, TOC values were not inserted in the following tables. Finally, there was no significant change in the pH values measured at the end of the experiment, and this finding confirms that vitamin B₃ was formed in small quantities.

In Table 5, the results of the photocatalytic oxidation of 3-pyridinemethanol under UVA light with TiO₂-60V-3h photocatalysts obtained at different thermal treatment temperatures are given. TiO₂-60V-3h-250 photocatalyst gave the highest activity (k = 0.144 h⁻¹) with 36% conversion. This photocatalyst, which is thermally treated at 250°C, still has the highest activity and product selectivity values. It showed 85% 3-pyridinemethanal and 11% vitamin B₃ selectivity (96% total) for 30% conversion.

Photocatalytic results under UVA light of 3-pyridinemethanol oxidation with TiO₂-60V-5h photocatalysts calcined at different temperatures are given in Table 5. It can be noted that 250°C was the optimum temperature. The values of the rate constant with respect to the calcination temperature of the photocatalysts can be arranged almost like a Gaussian curve. The optimum calcination temperature depends on the type of substrate. The highly crystalline TiO₂ samples were found to be the most effective both in terms of conversion and selectivity for the oxidation of glycerol (best results in the presence of calcined TiO₂ at 700°C), while a medium/low crystallinity of TiO₂ was required for the oxidation of 3-pyridinemethanol (best results in the presence of calcined TiO₂ at 250°C).

Table 5 also shows the results of two commercial TiO₂ photocatalysts (Degussa P25, Merck) for the sake of comparison. The commercial TiO₂ catalysts were prepared at high temperatures, and their high crystallinity was evidenced by their sharp and intense XRD peaks (Yurdakal et al., 2020). TiO₂-60V-5h-250 showed the highest activity among the home prepared photocatalysts, and in particular the k-value was about 3 times higher than that found for the Merck commercial photocatalyst (0.276 vs. 0.0856 h⁻¹). Conversely, Degussa P25 showed greater photocatalytic activity (k = 0.596 h⁻¹). However, while the total selectivity value of aromatic aldehyde and vitamin B₃ after the 30% substrate conversion was only 50% in the case of Degussa P25, this value was 92% for the TiO₂-60V-5h-250 sample. The photocatalytic oxidation results of 3-pyridinemethanol versus time in the presence of Degussa P25 and TiO₂-60V-5h-250 are shown in Figure 8A. It was evidenced in Figure 8B that the reaction rate fits the pseudo-first-order kinetics.

Although the conversion rate of alcohol was lower in the presence of TiO₂-60V-5h-250 than in Degussa P25, the TiO₂-60V-5h-250 photocatalyst produced more products. This finding indicates that the latter photocatalyst (TiO₂-60V-5h-250) can potentially be used for fast and selective photocatalytic synthetic reactions under environmentally friendly conditions, although a handicap is the laborious preparation process.

Some authors (Alfe et al., 2014; Spasiano et al., 2015a, 2015b, 2016) performed this reaction under acidic conditions (pH 1–4) using commercial TiO₂ and also TiO₂-graphene composite photocatalysts. The reaction took place in the presence of Cu²⁺ in the absence of oxygen, but at pH >4 the reaction almost did not occur (Alfe et al., 2014; Spasiano et al., 2015a, 2015b, 2016). Our group investigated the selective oxidation of pyridinemethanols (o-, m-, p-) under UVA, UV-Vis, and visible light sources in the presence of Pt-loaded TiO₂ photocatalysts at pH 2–12 (Yurdakal et al., 2017). High selectivity towards vitamin B₃ was obtained only in basic conditions. Another recent study on photocatalytic oxidation of 3-pyridinemethanol was carried out by Çetinkaya and Yurdakal (2021). In that case, samples prepared at low temperature showed high product selectivity with respect to Degussa P25. However, in the present work we investigated the thermal treatment temperature in detail and found that the optimum thermal treatment is 250°C. Low crystalline TiO₂ photocatalysts have a hydrophilic surface, but allowed the desorption of the products before their further oxidation up to a complete mineralization, and this is the reason for the high product selectivity (Augugliaro et al., 2008).