Broadening the Action Spectrum of TiO2-Based Photocatalysts to Visible Region by Substituting Platinum with Copper

In this study, TiO2-based photocatalysts modified with Pt and Cu/CuOx were synthesized and studied in the photocatalytic reduction of CO2. The morphology and chemical states of synthesized photocatalysts were studied using UV-Vis diffuse reflectance spectroscopy, high-resolution transmission electron microscopy, and X-ray photoelectron spectroscopy. A series of light-emitting diodes (LEDs) with maximum intensity in the range of 365–450 nm was used to determine the action spectrum of photocatalysts. It is shown for, the first time, that the pre-calcination of TiO2 at 700 °C and the use of Cu/CuOx instead of Pt allow one to design a highly efficient photocatalyst for CO2 transformation shifting the working range to the visible light (425 nm). Cu/CuOx/TiO2 (calcined at 700 °C) shows a rate of CH4 formation of 1.2 ± 0.1 µmol h−1 g−1 and an overall CO2 reduction rate of 11 ± 1 µmol h−1 g−1 (at 425 nm).


Introduction
Today, humankind faces the problem of an increase in the emission of greenhouse gases, and their utilization is increasingly becoming the actual goal. Additionally, special attention is paid to utilizing the major component of greenhouse gases-CO 2 [1,2]. A promising approach to utilizing CO 2 is its photocatalytic conversion to more reactive substances, such as CO and CH 4 . This technology allows reducing the amount of CO 2 in the atmosphere [3]. Moreover, the products of this process are organic compounds, such as methane, methanol, ethanol, and formaldehyde, which could be then converted into valuable organic components or fuels [4]. It is worth noting that such organic compounds could be synthesized using renewable sources: light, CO 2 , and water [5].
The thermodynamic stability and kinetic inertia of CO 2 limit its use in the chemical industry. The development of technologies that allow the conversion of CO 2 into useful organic compounds under "soft" conditions will increase the appeal of this process and facilitate its large-scale implementation. Exposure to visible or UV light irradiation is known to remove thermodynamic limitations on the reductive conversion of carbon dioxide in the presence of water into organic compounds, which makes the process of photocatalytic CO 2 reduction one of the most promising technologies, but the large-scale implementation of this process is hampered by the lack of effective photocatalysts.
Metal oxides, especially TiO 2 , are widely used as photocatalysts because of their stability and their ability to generate the electron-hole pairs under light irradiation and at a low cost [6,7]. Modification with metals allows one to enhance the activity of titania due to the separation of photogenerated charge carriers (electron-hole pairs) [8,9]. Currently, platinum is the most used metal as a cocatalyst, as it has a high working function of the electron (5.6 eV) and, therefore, promotes effective charge separation [10,11]. Despite the high activity of materials including Pt as cocatalyst, there is a necessity to develop an effective photocatalyst with non-noble metal from a practical point of view. One of the promising cocatalysts is Cu and its oxides, both Cu 2 O and CuO have narrow bandgaps and are able to absorb visible light [12][13][14][15][16]. It should be noted that calcination of TiO 2 at high temperature leads to a shift in adsorption edge to a longer wavelength [17]. Such thermal activation should lead to enhanced photocatalytic activity under visible light irradiation. The advantage of this method is its simplicity compared with other methods for shifting the TiO 2 absorption edge, such as reconstructing the TiO 2 surface, embedding quantum dots, creating defects, or doping TiO 2 lattice [18][19][20][21][22].
At the same time, although research on this process has been ongoing for the past 20 years, there is still no consensus on which photocatalysts are the most active, how the reduction of CO 2 occurs, and what affects the distribution of reduction products. Most studies are aimed at the synthesis of active material for photocatalytic reduction of CO 2 [3], while the identification of regularities between the nature of the products formed, as well as the rate of CO 2 reduction and the state of the catalyst, facilitate the formulation of a scientific basis for creating an efficient photocatalytic system. In the case of using Cu as a cocatalyst, special attention should be paid to studying the photocatalyst surface due to the chemical versatility of Cu and the possible formation of copper oxides [23].
This research focused on a detailed study of the products of CO 2 reduction that depend on the nature of surface TiO 2 -based photocatalyst using Cu and Pt as cocatalysts.
To explain the mechanism of the photocatalyst operation, the reduction of CO 2 was studied under irradiation of light-emitting diodes with different wavelengths in the range from 365 to 450 nm, and the catalysts were characterized by transmission electron microscopy, X-ray photoelectron spectroscopy, and UV-Vis spectroscopy. In the current manuscript, the influence of Pt and Cu/CuO x , the thermal activation of TiO 2 , and the light source wavelength on the product distribution and photocatalyst activity in CO 2 reduction are defined under the same conditions, for the first time. The article reveals the characteristics of the photocatalytic system that contribute to the predominant formation of CH 4 during photocatalytic CO 2 reduction. A simple approach involving heat treatment of TiO 2 and Cu/CuO x deposition is proposed to synthesize visible light active photocatalyst.

Photocatalyst Characterization
The synthesized catalysts were characterized by UV-Vis diffuse reflectance spectroscopy, high-resolution transmission electron microscopy (HR TEM), inductively coupled plasma atomic emission spectroscopy, and X-ray photoelectron spectroscopy (XPS).
The diffuse reflectance UV-Vis spectra were measured using a UV-2501 PC spectrophotometer with an ISR-240A diffuse reflectance unit (Shimadzu, Kyoto, Japan).
The structure and microstructure of the photocatalysts were studied with HR TEM using a ThemisZ electron microscope (Thermo Fisher Scientific, Waltham, MA, USA) operated at an accelerating voltage of 200 kV. The microscope is equipped with a corrector of spherical aberrations, which provided a maximum lattice resolution of 0.06 nm, and a SuperX energy-dispersive spectrometer (Thermo Fisher Scientific). Images were recorded using a Ceta 16 CCD sensor (Thermo Fisher Scientific). For electron microscopy studies, Nanomaterials 2022, 12, 1584 3 of 15 samples were deposited on perforated carbon substrates attached to aluminum grids using an ultrasonic dispersant.
Inductively coupled plasma atomic emission spectroscopy (ICP-AES) was used to perform elemental analysis of the synthesized photocatalysts. ICP-AES was carried out on an Optima 4300 5DV spectrometer (PerkinElmer Inc., Waltham, MA, USA).
The relative surface concentration of elements and their chemical states were investigated using XPS. The study was carried out using an X-ray photoelectron spectrometer (SPECS Surface Nano Analysis GmbH, Berlin, Germany), equipped with an XR-50 X-ray source and PHOIBOS-150 hemispherical electron energy analyzer. The Cu2p, Pt4f, Ti2p, and O1s XPS core-level spectra were recorded under ultrahigh vacuum conditions. The core-level spectra were obtained using the non-monochromatic Al Kα radiation (hυ = 1486.61 eV). The charge correction was performed by setting the Ti2p 3/2 peak at 459.0 eV, which corresponded to titanium in the Ti 4+ state of the TiO 2 lattice. Relative concentrations of elements were determined from the integrated intensities of the core-level spectra using their cross-sections according to Scofield [26]. For detailed analysis, the spectra were fitted into several peaks after background subtraction using the Shirley method. The fitting procedure was performed using the CasaXPS software [27]. The line shape of the peaks was approximated by the multiplication of the Gaussian and Lorentzian functions.

Photocatalytic Activity Measurements
The reduction of CO 2 was carried out in a batch reactor (70 mL) with a quartz window (16 cm 2 ) under light irradiation ( Figure 1). A suspension consisting of a catalyst (30 mg) and water (700 µL) was deposited on a glass support (8 cm 2 ) and dried at 50 • C in an air atmosphere. The resulting thickness of catalyst coverage is about 100 µm. After the deposition, the catalyst was irradiated by UV light (380 nm) for 30 min to remove organic contaminations from the photocatalyst surface and then installed in the reactor containing 1 mL of deionized water. Pure water without any additional sacrificial agents was used in all photocatalytic experiments. After that, the reactor was purged with CO 2 (>99.94% purity) for 1 h. A number of light emission diodes (LEDs) with wavelength and surface power density of 365 (10 mW/cm 2 ), 380 (17 mW/cm 2 ), 400 (60 mW/cm 2 ), 425 (56 mW/cm 2 ), and 450 (42 mW/cm 2 ) nm was used as a light source. The surface power density was measured at the sample position. SuperX energy-dispersive spectrometer (Thermo Fisher Scientific). Images were recorded using a Ceta 16 CCD sensor (Thermo Fisher Scientific). For electron microscopy studies, samples were deposited on perforated carbon substrates attached to aluminum grids using an ultrasonic dispersant. Inductively coupled plasma atomic emission spectroscopy (ICP-AES) was used to perform elemental analysis of the synthesized photocatalysts. ICP-AES was carried out on an Optima 4300 5DV spectrometer (PerkinElmer Inc., Waltham, MA, USA).
The relative surface concentration of elements and their chemical states were investigated using XPS. The study was carried out using an X-ray photoelectron spectrometer (SPECS Surface Nano Analysis GmbH, Berlin, Germany), equipped with an XR-50 X-ray source and PHOIBOS-150 hemispherical electron energy analyzer. The Cu2p, Pt4f, Ti2p, and O1s XPS core-level spectra were recorded under ultrahigh vacuum conditions. The core-level spectra were obtained using the non-monochromatic Al Kα radiation (hυ = 1486.61 eV). The charge correction was performed by setting the Ti2p3/2 peak at 459.0 eV, which corresponded to titanium in the Ti 4+ state of the TiO2 lattice. Relative concentrations of elements were determined from the integrated intensities of the core-level spectra using their cross-sections according to Scofield [26]. For detailed analysis, the spectra were fitted into several peaks after background subtraction using the Shirley method. The fitting procedure was performed using the CasaXPS software [27]. The line shape of the peaks was approximated by the multiplication of the Gaussian and Lorentzian functions.

Photocatalytic Activity Measurements
The reduction of CO2 was carried out in a batch reactor (70 mL) with a quartz window (16 cm 2 ) under light irradiation ( Figure 1). A suspension consisting of a catalyst (30 mg) and water (700 µL) was deposited on a glass support (8 cm 2 ) and dried at 50 °C in an air atmosphere. The resulting thickness of catalyst coverage is about 100 µm. After the deposition, the catalyst was irradiated by UV light (380 nm) for 30 min to remove organic contaminations from the photocatalyst surface and then installed in the reactor containing 1 mL of deionized water. Pure water without any additional sacrificial agents was used in all photocatalytic experiments. After that, the reactor was purged with CO2 (>99.94% purity) for 1 h. A number of light emission diodes (LEDs) with wavelength and surface power density of 365 (10 mW/cm 2 ), 380 (17 mW/cm 2 ), 400 (60 mW/cm 2 ), 425 (56 mW/cm 2 ), and 450 (42 mW/cm 2 ) nm was used as a light source. The surface power density was measured at the sample position.  A kinetic experiment was carried out for 24 h. Gas probe was taken using a gas syringe (500 µL) and analyzed with a gas chromatograph "GH-1000" (Chromos, Moscow, Russia) equipped with the flame ionization detector and thermal conductivity detector to identify the products of CO 2 reduction and H 2 , respectively. The total rate of photocatalytic CO 2 reduction was calculated according to the following equation: where n(CH 4 ) and n(CO) are the amounts of CH 4 and CO (µmol), 2 and 8 are the coefficients for electron balance, t is the time of reaction (h), and m is the weight of photocatalyst (g). The selectivity of CH 4 formation was defined as The apparent quantum efficiency (AQE) was calculated according to the following equation: where N phot is the calculated photon flux equal to 5.8 mmol h −1 for LED-425.

Photocatalyst Characterization
The optical properties of the synthesized photocatalysts were studied by UV-Vis diffuse reflectance spectroscopy (Figure 2a). The non-modified titania has a single, steep absorption edge around 380 nm, which is typical for Evonik (Degussa) P25. After calcination, the TiO 2 700 sample demonstrates a redshift, and the absorption edge is about 400 nm. Evonik P25 is known to consist of two TiO 2 phases with different bandgap energy: anatase (3.2 eV) and rutile (3.0 eV) [28,29]. The observed effect indicates a structural change in TiO 2 during calcination and an increase in rutile content for TiO 2 700. Calculated bandgap energies using Tauc plots are 3.14 eV for unmodified TiO 2 and 3.02 eV for TiO 2 700 ( Figure 2b). Earlier, it has been shown that the calcination of TiO 2 at 700 • C changes anatase to rutile ratio from 85:15 to 22:78 [24], which is in agreement with the UV-Vis spectroscopy study of the synthesized catalysts. Nanomaterials 2022, 12, x FOR PEER REVIEW 4 of A kinetic experiment was carried out for 24 h. Gas probe was taken using a gas s ringe (500 µL) and analyzed with a gas chromatograph "GH-1000" (Chromos, Mosco Russia) equipped with the flame ionization detector and thermal conductivity detector identify the products of CO2 reduction and H2, respectively.
The total rate of photocatalytic CO2 reduction was calculated according to the follow ing equation: where n(CH4) and n(CO) are the amounts of CH4 and CO (µmol), 2 and 8 are the coef cients for electron balance, t is the time of reaction (h), and m is the weight of photocataly (g). The selectivity of CH4 formation was defined as The apparent quantum efficiency (AQE) was calculated according to the followin equation: where Nphot is the calculated photon flux equal to 5.8 mmol h −1 for LED-425.

Photocatalyst Characterization
The optical properties of the synthesized photocatalysts were studied by UV-Vis d fuse reflectance spectroscopy ( Figure 2a). The non-modified titania has a single, steep a sorption edge around 380 nm, which is typical for Evonik (Degussa) P25. After calcinatio the TiO2 700 sample demonstrates a redshift, and the absorption edge is about 400 nm Evonik P25 is known to consist of two TiO2 phases with different bandgap energy: anata (3.2 eV) and rutile (3.0 eV) [28,29]. The observed effect indicates a structural change TiO2 during calcination and an increase in rutile content for TiO2 700. Calculated bandga energies using Tauc plots are 3.14 eV for unmodified TiO2 and 3.02 eV for TiO2 700 (Figu 2b). Earlier, it has been shown that the calcination of TiO2 at 700 °C changes anatase rutile ratio from 85:15 to 22:78 [24], which is in agreement with the UV-Vis spectroscop study of the synthesized catalysts.  The actual metal content established by ICP-AES was around 0.7% for platinized samples and 4.5% for Cu-containing catalysts.
The deposition of platinum and copper promotes the adsorption of visible light for both TiO 2 and TiO 2 700. Platinum was shown to possess a localized surface plasmon resonance in the visible region [30]. In the case of Cu-containing samples, the increased absorption is caused by the intrinsic adsorption of copper oxides due to the narrow bandgap width of CuO (1.7 eV) and Cu 2 O (2.1 eV) [31,32]. The difference in the reflectance between Cu/TiO 2 and Cu/TiO 2 700 photocatalysts may be due to the different content of CuO x and Cu since oxides enhance the light absorption of the samples, whereas metal Cu does not [33]. Adsorption beyond 600 nm is related to the bandgap transitions of CuO [12]. We speculated that the content of CuO in Cu/TiO 2 is higher than in Cu/TiO 2 700, which was further confirmed by HR TEM and XPS.

Kinetic Experiments
The activity of the synthesized photocatalysts was studied in the reduction of CO 2 using a series of LEDs with an intensity maximum at about 365, 380, 400, 425, and 450 nm as light sources to define the spectrum of action of photocatalysts.
Since H 2 O was added to the studied system, the following processes could proceed on the photocatalyst surface [8,34]: The products of reactions were analyzed using gas chromatography: CO and CH 4 were the main products, and H 2 was detected in negligible amount-less than 0.05 µmol after 24 h of reaction for the most active catalyst, i.e., Cu/TiO 2 .
The rates of CO, CH 4 formation, and total CO 2 reduction are presented in Figure 3. One can see that the Pt deposition changes the distribution of products and greatly increases the amount of CH 4 . The Pt/TiO 2 and Pt/TiO 2 700 samples predominantly reduce CO 2 to CH 4 , and only a little amount of CO is detected, except for the LED-450 nm. When using Cu-containing samples, the dependence is more complex. Under UV irradiation, the formation rates of CO are higher than those of CH 4 . Nevertheless, the amounts of CH 4 produced over Cu/TiO 2 and Cu/TiO 2 700 are similar to those for Pt/TiO 2 . At the same time, Cu/TiO 2 and Cu/TiO 2 700 yield CO at significantly higher rates than Pt/TiO 2 , resulting in a higher overall CO 2 reduction efficiency. The activity of these photocatalysts is close to that of Pt/TiO 2 700 under LED-365 nm, which demonstrates that the highest rate of overall CO 2 reduction under UV light is equal to 2.
According to UV-Vis spectra, all photocatalysts exhibit strong light absorption at a wavelength of 365 nm. However, with an increase in the wavelength to 380 nm, the absorption of the samples obtained without preliminary calcination of TiO 2 decreases compared with the calcined samples. Pt/TiO 2 700 demonstrates higher activity under LED-380 nm and LED-365 nm, compared with Pt/TiO 2 . Therefore, the difference in activity between Pt/TiO 2 and Pt/TiO 2 700 is not caused only by different light absorption. The most likely reason for the improved activity under UV light is the increased lifetime of photogenerated charge carriers-electrons and holes-for Pt/TiO 2 700 due to the efficient transfer of photogenerated electrons from TiO 2 to metal particles. To clarify the reason for the increased activity of the Pt/TiO 2 700 sample under irradiation with LED-365 nm, compared with Pt/TiO 2 , the photocatalysts were characterized using HR TEM and XPS.
The HR TEM micrographs of the Pt/TiO 2 and Pt/TiO 2 700 are presented in Figure 4. For Pt/TiO 2 , one can identify well-crystallized TiO 2 nanoparticles with a size of 10-25 nm (Figure 4a), whereas in the case of Pt/TiO 2 700 catalyst (Figure 4b), the average size of crystallized TiO 2 nanoparticles increases to about 35-45 nm. The effect of thermal treatment has been previously studied and shown to lead to a decrease in the specific surface area of the TiO 2 support, an increase in crystallite size, and the part of rutile [24]. It should be noted that, for both supports, the platinum nanoparticles have a narrow particle size distribution, with mean particle sizes of 1.9 and 2.2 nm for Pt/TiO 2 and Pt/TiO 2 700, respectively. The difference between these catalysts is the aggregates of Pt nanoparticles observed in the case of Pt/TiO 2 . For Pt/TiO 2 700, Pt nanoparticles are well distributed on the TiO 2 surface and do not form agglomerate structures.  Designation: "Pt" and "Cu" means Pt/TiO2 and Cu/TiO2, respectively; "Pt-700" and "Cu-700" means Pt/TiO2 700 and Cu/TiO2 700, respectively.
The HR TEM micrographs of the Pt/TiO2 and Pt/TiO2 700 are presented in Figure 4. For Pt/TiO2, one can identify well-crystallized TiO2 nanoparticles with a size of 10-25 nm (Figure 4a), whereas in the case of Pt/TiO2 700 catalyst (Figure 4b), the average size of crystallized TiO2 nanoparticles increases to about 35-45 nm. The effect of thermal treatment has been previously studied and shown to lead to a decrease in the specific surface area of the TiO2 support, an increase in crystallite size, and the part of rutile [24]. It should be noted that, for both supports, the platinum nanoparticles have a narrow particle size distribution, with mean particle sizes of 1.9 and 2.2 nm for Pt/TiO2 and Pt/TiO2 700, respectively. The difference between these catalysts is the aggregates of Pt nanoparticles observed in the case of Pt/TiO2. For Pt/TiO2 700, Pt nanoparticles are well distributed on the TiO2 surface and do not form agglomerate structures.  The XPS study of the Pt/TiO2 and Pt/TiO2 700 samples revealed that the [Pt]/[Ti] ra in the Pt/TiO2 700 photocatalyst increases by a factor of two, compared with Pt/TiO2 (Ta 1). This effect is caused by a decrease in the surface area of TiO2, as shown in our previo research [24]. Since the thermal treatment of TiO2 was carried out in the same manner, assumed a similar change in surface area. The binding energy of the Pt4f7/2 peak is 70.9 for both catalysts, which corresponds to platinum in the metallic state ( Figure S1a). Ad The XPS study of the Pt/TiO 2 and Pt/TiO 2 700 samples revealed that the [Pt]/[Ti] ratio in the Pt/TiO 2 700 photocatalyst increases by a factor of two, compared with Pt/TiO 2 (Table 1). This effect is caused by a decrease in the surface area of TiO 2 , as shown in our previous research [24]. Since the thermal treatment of TiO 2 was carried out in the same manner, we assumed a similar change in surface area. The binding energy of the Pt4f 7/2 peak is 70.9 eV for both catalysts, which corresponds to platinum in the metallic state ( Figure S1a). Additionally, these samples were studied using XPS after a photocatalytic reaction under LED-400 nm, and no change in the state of Pt was observed ( Figure S1b). The [O]/[Ti] ratio after calcination of the initial TiO 2 sample and after carrying out the photocatalytic reaction almost does not change. This small increase could be related to the formation of cation vacancies in a negligible amount [35]. Thus, the higher activity of Pt/TiO 2 700 is due to the high surface concentration of platinum nanoparticles and enhanced absorption properties of calcinated TiO 2 . Under light irradiation with a maximum intensity at 400 nm, both Cu-containing photocatalysts demonstrate higher CH 4 formation rates than CO (Figure 5a). Moreover, their activities are much higher than that of both platinized samples and achieve 37 ± 4 and 31 ± 3 µmol h −1 g −1 for Cu/TiO 2 and Cu/TiO 2 700, respectively (Figure 5b). When using LED-425 nm, the activity of photocatalysts based on calcinated TiO 2 (Pt/TiO 2 700 and Cu/TiO 2 700) exceeds the activity of photocatalysts based on uncalcined TiO 2 (Figure 5c,d). Additionally, Cu/TiO 2 produces more CO than CH 4 , while Cu/TiO 2 700 is still more active toward CH 4 formation. Finally, the use of LED-450 nm leads to the predominant formation of CO for all studied catalysts, but the efficiency of this process is low (Figure 5e,f). Nevertheless, calcined photocatalysts are still more active in overall CO 2 reduction, with a reaction rate of about 1 µmol h −1 g −1 .
The calcinated samples have a shifted adsorption edge, compared with unmodified ones and, therefore, enhanced absorption and improved photocatalytic activity under visible light (425 and 450 nm). Since the dependence under LED-400 nm is more complex, the Cu-containing samples were studied using XPS and HR TEM in the high-angle annular dark-field scanning mode (HAADF STEM). The HAADF STEM allows one to make micrographs of samples with high contrast, with the bright area corresponding to the atoms with higher Z.
The HAADF STEM micrographs of the Cu/TiO 2 and Cu/TiO 2 700 are presented in Figure 6. For Cu/TiO 2 , one can see that copper nanoparticles have a wide particle size distribution in the range of 1.5-5 nm, whereas, in the case of Cu/TiO 2 700, a narrow particle size distribution in the range of 1-2 nm is found. It is worth noting that, in the case of Cu/TiO 2 700, the agglomerates of Cu nanoparticles are not observed. Moreover, in the case of Cu/TiO 2 , the EDX mapping reveals the presence of massive Cu particles of 200-400 nm ( Figure S2a). Thus, the use of calcined TiO 2 as support allows the synthesis of a photocatalyst with a fine size distribution of Cu nanoparticles.  According to the results of UV-Vis spectroscopy, Cu/TiO 2 and Cu/TiO 2 700 absorb visible light to varying degrees, but light absorption in this region indicates the presence of copper oxides in both samples. This assumption was clearly confirmed by the XPS study (Table 2). According to the XPS analysis, the fraction of Cu 2+ cations is about 78% and 52% for Cu/TiO 2 and Cu/TiO 2 700, respectively ( Figure S3a). However, it should be stressed that the identification of Cu 1+ and Cu 0 states by analyzing the Cu2p spectrum is difficult, as the Cu2p 3/2 binding energies of Cu 1+ and Cu 0 states are similar [36]. Nevertheless, the Auger parameter α, which is equal to the sum of the Cu2p 3/2 peak binding energy and the Cu LMM peak kinetic energy, can be used for the identification of Cu 1+ and Cu 0 states. Unfortunately, the surface concentration of copper cations did not allow us to measure the Cu LMM peak kinetic energy and identify the cation distribution. It should also be noted that the [Cu]/[Ti] surface ratios are slightly different for Cu/TiO 2 and Cu/TiO 2 700. The observed deviation is explained by the surface sensitivity of XPS, which is affected by different size distributions of copper nanoparticles, the presence of large Cu/CuO x particles (in the case of Cu/TiO 2 ), and a decrease in the surface area of TiO 2 . Although the presence of CuO x is beneficial for light absorption in the visible region, due to the agglomeration of copper nanoparticles and the presence of large Cu/CuO x particles in the sample Cu/TiO 2 ( Figure S2a Nanomaterials 2022, 12, x FOR PEER REVIEW 9 The calcinated samples have a shifted adsorption edge, compared with unmod ones and, therefore, enhanced absorption and improved photocatalytic activity unde ible light (425 and 450 nm). Since the dependence under LED-400 nm is more com the Cu-containing samples were studied using XPS and HR TEM in the high-angle a lar dark-field scanning mode (HAADF STEM). The HAADF STEM allows one to m micrographs of samples with high contrast, with the bright area corresponding to th oms with higher Z.
The HAADF STEM micrographs of the Cu/TiO2 and Cu/TiO2 700 are presente Figure 6. For Cu/TiO2, one can see that copper nanoparticles have a wide particle distribution in the range of 1.5-5 nm, whereas, in the case of Cu/TiO2 700, a narrow par size distribution in the range of 1-2 nm is found. It is worth noting that, in the ca Cu/TiO2 700, the agglomerates of Cu nanoparticles are not observed. Moreover, in the of Cu/TiO2, the EDX mapping reveals the presence of massive Cu particles of 200-40 ( Figure S2a). Thus, the use of calcined TiO2 as support allows the synthesis of a photo lyst with a fine size distribution of Cu nanoparticles. According to the results of UV-Vis spectroscopy, Cu/TiO2 and Cu/TiO2 700 ab visible light to varying degrees, but light absorption in this region indicates the pres of copper oxides in both samples. This assumption was clearly confirmed by the study (Table 2). According to the XPS analysis, the fraction of Cu 2+ cations is about and 52% for Cu/TiO2 and Cu/TiO2 700, respectively ( Figure S3a). However, it shoul stressed that the identification of Cu 1+ and Cu 0 states by analyzing the Cu2p spectru difficult, as the Cu2p3/2 binding energies of Cu 1+ and Cu 0 states are similar [36]. Neve less, the Auger parameter α, which is equal to the sum of the Cu2p3/2 peak binding en and the Cu LMM peak kinetic energy, can be used for the identification of Cu 1+ and states. Unfortunately, the surface concentration of copper cations did not allow u measure the Cu LMM peak kinetic energy and identify the cation distribution. It sh also be noted that the [Cu]/[Ti] surface ratios are slightly different for Cu/TiO2 Cu/TiO2 700. The observed deviation is explained by the surface sensitivity of XPS, w is affected by different size distributions of copper nanoparticles, the presence of Cu/CuOx particles (in the case of Cu/TiO2), and a decrease in the surface area of T Although the presence of CuOx is beneficial for light absorption in the visible region to the agglomeration of copper nanoparticles and the presence of large Cu/CuOx part in the sample Cu/TiO2 (Figure S2a), Cu/TiO2 cannot work as efficiently as the Cu/TiO photocatalyst. The study of catalysts after irradiation by LED-400 nm and LED-450  Based on the kinetic results and the photocatalyst characterization, we can speculate that the product distribution of the CO 2 reduction process depends on the nature of the cocatalyst and the optical absorption edge of the photocatalyst. Provided that there is enough energy of photons for the generation of a relatively high number of electrons, electrons can be collected in a cocatalyst and participate in the multi-electron process of CO 2 reduction to CH 4 (Equation (3)) [37]. The photon energy required could be reduced by depositing a cocatalyst, which can absorb light with a longer wavelength, and/or by changing the adsorption edge of TiO 2 . The combination of these two approaches makes it possible to obtain a photocatalyst exhibiting high activity with respect to methane under irradiation with 425 nm wavelength. The formation rate of CH 4 over the Cu/TiO 2 700 sample achieves 1.2 ± 0.1 µmol h −1 g −1 and an overall CO 2 reduction rate of 11 ± 1 µmol h −1 g −1 (for a LED-425 nm). A possible reason for the low CH 4 formation rate over Cu/TiO 2 and Cu/TiO 2 700 when using LEDs with a wavelength of 365 and 380 nm is the different electron surface densities in the photocatalyst. Since TiO 2 has strong adsorption in the UV region, it can effectively generate electron-holes pairs under LED-365 and LED-380. It has been shown that CO 2 reduction on the TiO 2 surface predominantly led to CO formation [38]. Under visible light irradiation, copper oxides absorb light to a much greater extent than TiO 2 , therefore, the main generation of charge carriers occurs in cocatalyst particles (Cu/CuO x ), and the most of the electrons are located in the metal, which leads to an increase in the rate of methane formation [39].
The photocatalytic activity of Cu/TiO 2 700 at 425 nm is caused by several factors. Firstly, the absorption edge of the sample is around 400 nm, but since the light absorption of TiO 2 700 decreases gradually from 330 to 440 nm, the photocatalyst is still able to absorb light irradiation beyond 400 nm, although to a lesser extent. Additionally, since LEDs have non-monochromatic radiation, LED-425 emits light with a wavelength of both more and less than 425 nm, as shown in Figure 1a. Finally, the presence of CuO x greatly promotes the adsorption of visible light, as stated in the Introduction section. We believe that the observed behavior of Cu/TiO 2 700 will promote the use of Cu-TiO 2 -based catalysts for the conversion of CO 2 under visible light.
Additionally, it should be noted that the fraction of Cu 2+ and Cu 0 /Cu 1+ cations varies after the photocatalytic reaction for both the Cu/TiO 2 and Cu/TiO 2 700 catalysts ( Table 2). This behavior indicates that copper cations play key roles in the reduction of CO 2 and could influence product distribution. The photogenerated electrons could reduce the copper oxides according to the following equations [40]: Cu 2+ + e − = Cu + , E 0 (Cu 2+ /Cu 1+ ) = 0.16 V Cu 2+ + 2e − = Cu 0 , E 0 (Cu 2+ /Cu 0 ) = 0.34 V Cu + + e − = Cu 0 , E 0 (Cu 1+ /Cu 0 ) = 0.52 V while the reduced states could be oxidized by the free oxygen anion radicals generated due to the reaction of O 2 with e − (Equation (10)) or by singlet oxygen, which is the product of a reaction between O 2 − and h + (Equation (11)) [41].
Therefore, ternary heterostructures, including Cu, CuO x , and TiO 2 , can be formed during the photocatalytic reaction. Both CuO x and TiO 2 compounds adsorb light irradiation and generate electron-hole pairs. CuO (p-type) and TiO 2 (n-type) form a p-n heterojunction, and the electron transfer from TiO 2 to CuO suppresses the charge recombination [42][43][44]. The charge separation between TiO 2 and Cu 2 O is realized according to the Z-scheme, according to which electrons from the conduction band (CB) of TiO 2 recombine with holes from the valence band of Cu 2 O [45]. The CB of Cu 2 O is −1.55 eV [46], which is much lower than the potential of CO 2 reduction to CO and CH 4 in the presence of water (Equations (2) and (3)). Therefore, the remaining electrons in Cu 2 O can easily react with CO 2 to produce CO or CH 4 . At the same time, holes in TiO 2 react with water vapor. The working function of Cu is higher than that of TiO 2 and CuO x ; therefore, the presence of metallic Cu promotes electron transfer to metal particles from TiO 2 , as well as from CuO x , thus improving the charge separation and the lifetime of eand h + [47][48][49][50]. Depending on the position of copper and which contacts are present in the catalyst (Cu-TiO 2 , Cu-CuO x , or both of them), various mechanisms for the transfer of photogenerated charge carriers can be realized. This is a difficult task and requires another detailed study, for example, in situ/operando X-ray absorption spectroscopy (XAS) study of Cu K-edge during the photoreduction of CO 2 . It should be noted that XAS is more sensitive to the oxidation states of copper cations than XPS and allows the Cu 2+ , Cu 1+ , and Cu 0 states to be easily distinguished [51].

Stability Test
To study the catalytic stability of the most active catalyst (Cu/TiO 2 700), stability tests were carried out. The activity of Cu/TiO 2 700 catalyst was tested during five 24 h runs under irradiation of LED-400 nm (Figure 7). Between the runs, light irradiation was turned off, and the reactor was purged with CO 2 for 1 h.
during the photocatalytic reaction. Both CuOx and TiO2 compounds adsorb light irradiation and generate electron-hole pairs. CuO (p-type) and TiO2 (n-type) form a p-n heterojunction, and the electron transfer from TiO2 to CuO suppresses the charge recombination [42][43][44]. The charge separation between TiO2 and Cu2O is realized according to the Zscheme, according to which electrons from the conduction band (CB) of TiO2 recombine with holes from the valence band of Cu2O [45]. The CB of Cu2O is −1.55 eV [46], which is much lower than the potential of CO2 reduction to CO and CH4 in the presence of water (Equations (2) and (3)). Therefore, the remaining electrons in Cu2O can easily react with CO2 to produce CO or CH4. At the same time, holes in TiO2 react with water vapor. The working function of Cu is higher than that of TiO2 and CuOx; therefore, the presence of metallic Cu promotes electron transfer to metal particles from TiO2, as well as from CuOx, thus improving the charge separation and the lifetime of eand h + [47][48][49][50]. Depending on the position of copper and which contacts are present in the catalyst (Cu-TiO2, Cu-CuOx, or both of them), various mechanisms for the transfer of photogenerated charge carriers can be realized. This is a difficult task and requires another detailed study, for example, in situ/operando X-ray absorption spectroscopy (XAS) study of Cu K-edge during the photoreduction of CO2. It should be noted that XAS is more sensitive to the oxidation states of copper cations than XPS and allows the Cu 2+ , Cu 1+ , and Cu 0 states to be easily distinguished [51].

Stability Test
To study the catalytic stability of the most active catalyst (Cu/TiO2 700), stability tests were carried out. The activity of Cu/TiO2 700 catalyst was tested during five 24 h runs under irradiation of LED-400 nm (Figure 7). Between the runs, light irradiation was turned off, and the reactor was purged with CO2 for 1 h.  The CO formation rate slightly decreases during the second and third runs and then begins to grow; the rate of CH 4 formation stays constant throughout the first four runs and then decreases. In general, the photocatalyst exhibits high stability during 120 h of CO 2 reduction. The total amount of CH 4 is 400 µmol g −1 and, for CO, is 70 µmol g −1 after 120 h of irradiation. The observed decrease in photocatalytic activity could be due to the decrease in the number of active centers. It was mentioned above that eight electrons are necessary to convert CO 2 to CH 4 , whereas for CO production, two electrons are needed. Thus, the decrease in the number of heterojunctions could lead to a shift in the rate of CO 2 reduction to CO. It is likely that a decrease in the number of active centers is due to the increased size of copper nanoparticles. It is known that copper oxides could undergo photocorrosion caused by photoelectron reduction [52]. It has been previously observed that the reduction of transition metal oxides to the metallic phase could lead to the migration of metallic nanoparticles over the surface and the formation of metallic agglomerates [53,54]. The formation of agglomerates could lead to a decrease in the number of active sites and, thus, a decrease in photocatalytic activity. We believe that the calcination of TiO 2 leads to the formation of sites where the copper nanoparticle could be fixed, so as not to allow them to migrate over the surface and form the agglomerates. Indeed, as it was mentioned above, the EDX mapping shows that, in the case of Cu/TiO 2 catalyst, the agglomerates of copper nanoparticles are found, as well as large copper particles of 100-200 nm ( Figure S2a). At the same time, the EDX analysis does not reveal any copper particles of 10-100 nm in the case of Cu/TiO 2 700 catalysts (Figure 8a). The stabilization of nanoparticles, which prevents the aggregation of Cu/CuO x nanoparticles on the support, is the key to increasing the long-term stability of photocatalysts. Additionally, the EDX analysis of Cu/TiO 2 catalyst exposed under LED-450 nm for 24 h reveals an increase in the number of copper particles with sizes of 100-200 nm ( Figure S2b,c). To investigate its long-term stability, the Cu/TiO 2 700 catalyst was studied using EDX after the fifth run of CO 2 reduction under irradiation of LED-400 nm (Figure 8b). One can see two regions enriched with Cu. The size of these agglomerates is about 20 nm, and that is much lower than Cu particles observed after 24 h of irradiation of Cu/TiO 2 ( Figure S2c). Additionally, the distribution of oxygen ( Figure S4a,b) in Cu/TiO 2 700 after five runs indicates the absence of an area enriched with O at the place of Cu atom agglomeration. It can be concluded that this region, enriched in Cu atoms, mainly consists of metallic copper. Such metallic particles can be formed by the migration of metallic nanoparticles due to the reduction of copper oxides, as mentioned above. Therefore, the decrease in CH 4 production clearly observed on the fifth run could be caused by the agglomeration of Cu/CuO x particles, thus decreasing contact between cocatalyst and TiO 2 particles and reducing the lifetime of electrons. Additionally, the reduction of CuO x to Cu decreases the visible light absorption by photocatalyst. The results reveal that the calcination of TiO2 at 700 °C facilitates the deposition of well-dispersed cocatalyst particles and prevents Cu/CuOx aggregation during even 100 h of the photocatalytic reaction. This effect promotes the CH4 formation rate rather than the formation of CO. Additionally, the presence of copper oxides provides increased adsorption of visible light by the photocatalyst, which is vital for carrying out the multielectron process of CO2 reduction to CH4 upon irradiation in the visible region.

Conclusions
In this study, the influence of a cocatalyst and thermal activation of TiO2-based photocatalysts on the CO2 reduction process was investigated. Platinized photocatalysts were shown to be more active in relation to the CH4 formation under UV light irradiation, while Cu-containing ones demonstrated superior activity in CH4 production upon irradiation with visible light. The maximum overall CO2 reduction rate under UV irradiation was 2.8 µmol h −1 g −1 , which was achieved over TiO2 calcinated at 700 °C and modified with Pt. The Cu/TiO2 showed an activity rate of 2.6 µmol h −1 g −1 . The highest CO2 reduction rate was 37 µmol h −1 g −1 , achieved over Cu/TiO2 under light irradiation with a maximum intensity of 400 nm. At 425 nm, the Cu/TiO2 700 demonstrated a maximum CO2 reduction rate of 11 µmol h −1 g −1 . Both rates are high, compared with the published data on CO2 reduction over The results reveal that the calcination of TiO 2 at 700 • C facilitates the deposition of well-dispersed cocatalyst particles and prevents Cu/CuO x aggregation during even 100 h of the photocatalytic reaction. This effect promotes the CH 4 formation rate rather than the formation of CO. Additionally, the presence of copper oxides provides increased adsorption of visible light by the photocatalyst, which is vital for carrying out the multielectron process of CO 2 reduction to CH 4 upon irradiation in the visible region.

Conclusions
In this study, the influence of a cocatalyst and thermal activation of TiO 2 -based photocatalysts on the CO 2 reduction process was investigated. Platinized photocatalysts were shown to be more active in relation to the CH 4 formation under UV light irradiation, while Cu-containing ones demonstrated superior activity in CH 4 production upon irradiation with visible light. The maximum overall CO 2 reduction rate under UV irradiation was 2.8 µmol h −1 g −1 , which was achieved over TiO 2 calcinated at 700 • C and modified with Pt. The Cu/TiO 2 showed an activity rate of 2.6 µmol h −1 g −1 . The highest CO 2 reduction rate was 37 µmol h −1 g −1 , achieved over Cu/TiO 2 under light irradiation with a maximum intensity of 400 nm. At 425 nm, the Cu/TiO 2 700 demonstrated a maximum CO 2 reduction rate of 11 µmol h −1 g −1 . Both rates are high, compared with the published data on CO 2 reduction over different photocatalysts [3,55,56]. Under purely visible light irradiation (LED-425 nm), AQE and S(CH 4 ) for Cu/TiO 2 700 photocatalyst were achieved by 0.23% and 90%, respectively. Therefore, the replacement of platinum with copper is an effective method for the synthesis of a visible active photocatalyst for CO 2 reduction, provided that calcined TiO 2 (700 • C) is used.
The stability of Cu/TiO 2 700 was tested during 120 h of reaction under LED-400 nm. It was shown that the cocatalyst particles began to aggregate during experiments, but this process proceeded much more slowly on the surface of calcined TiO 2 . Platinum was in the metallic state on the surface of both TiO 2 and TiO 2 700 samples. Copper was present in various states, and the Cu/TiO 2 700 sample contained less Cu 2+ than Cu/TiO 2 , but this sample had a uniform distribution of cocatalyst, which is beneficial for the CH 4 production under irradiation of visible light. This photocatalyst demonstrated fairly good stability over 120 h of CO 2 reduction. A slight decrease in the CH 4 formation rate is probably associated with the beginning enlargement of Cu/CuO x particles and the reduction of copper oxides to metal particles, which reduces light absorption in the visible region.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/nano12091584/s1, Figure S1: Pt4f core-level spectra of the fresh (a) Pt-TiO 2 photocatalysts and after irradiation (b) under LED-400 nm. The spectra are normalized to the integral intensity of the corresponding Ti2p core-level spectra. Figure S2: EDX mapping (in HAADF STEM mode) of the fresh Cu/TiO 2 (a) catalyst and after irradiation (b,c) for 24 h under LED-450 nm. Figure S3: Cu2p core-level spectra of the fresh (a) Cu-TiO 2 photocatalysts and after irradiation (b) under LED-400 nm. Figure S4: EDX mapping (in HAADF STEM mode) of the Cu/TiO 2 700 catalyst after irradiation for 120 h under LED-400 nm (Cu (a), O (b)).
Author Contributions: Investigation, data curation, visualization, writing-original draft preparation, funding acquisition, project administration, A.A.S.; investigation, data curation, visualization, writing-original draft preparation, A.Y.K.; investigation, data curation, E.Y.G.; writing-original draft preparation, supervision, conceptualization, project administration, E.A.K. All authors have read and agreed to the published version of the manuscript.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.