Impact of Pd Loading on CO2 Reduction Performance over Pd/TiO2 with H2 and H2O

This study investigated the impact of molar ratio of CO2 to reductants H2O and H2, as well as Pd loading weight on CO2 reduction performance with Pd/TiO2 as the photocatalyst. The Pd/TiO2 film photocatalyst is prepared by the sol-gel and dip-coating process to prepare TiO2 film and the pulse arc plasma method is used to dope Pd on TiO2 film. The prepared Pd/TiO2 film was characterized by SEM, EPMA, STEM, EDS, and EELS. This study also investigated the performance of CO2 reduction under the illumination condition of Xe lamp with or without ultraviolet (UV) light. As a result, it is revealed that when the molar ratio of CO2/H2/H2O is set at 1:0.5:0.5, the best CO2 reduction performance has been obtained under the illumination condition of Xe lamp with and without UV light. In addition, it is found that the optimum Pd loading weight is 3.90 wt%. The maximum molar quantities of CO and CH4 produced per unit weight of photocatalyst are 30.3 μmol/g and 22.1 μmol/g, respectively, for the molar ratio of CO2/H2/H2O = 1:0.5:0.5 under the condition of Xe lamp illumination with UV light. With UV light, C2H4 and C2H6, as well as CO and CH4 are also produced by the Pd/TiO2 film photocatalyst prepared in this study.


Introduction
The Paris Agreement adopted in 2015 set the goal that the increase in average temperature in the world from the industrial revolution by 2030 should be kept less than 2 K. However, the global mean concentration of CO 2 in the atmosphere has increased up to 410 ppmV in December 2019, which increased by 25 ppmV since 2009 [1]. Therefore, it is requested to develop a new CO 2 reduction/utilization technology in order to reduce the amount of CO 2 in the atmosphere.
Reducing or converting CO 2 into fuel by photocatalyst became a hot R&D area. TiO 2 is commonly used as a photocatalyst for CO 2 reduction since it is convenient, inexpensive, and has strong durability for chemicals and corrosion [2]. TiO 2 is a popular photocatalyst that can reduce CO 2 into CO, CH 4 , CH 3 OH, and H 2 etc. with ultraviolet (UV) light [3][4][5].
Since pure TiO 2 can only be activated under UV light illumination, it is not very effective under sunlight illumination as UV light accounts for only approximately 4% in the solar spectrum. In addition, the rate of electron/hole pair recombination is faster than the rate of chemical interaction between the absorbents during redox reactions when using pure TiO 2 [6].
Many attempts have been reported to improve the performance of the TiO 2 [3]. Doping precious metals such as Pt [7], Ag [8], Au [9], Cu [10,11], using composite materials formed by GaP and TiO 2 [12], the authors [24]. In addition, the metal loading weight with TiO 2 is important to improve the CO 2 reduction performance [25,26]. However, there was no study so far to qualify the improvement.
To promote the CO 2 reduction performance, the optimum reductant providing the proton (H + ) should be clarified. According to the previous studies [27][28][29][30], the reaction mechanism to reduce CO 2 with H 2 O can be summarized as shown below: <Photocatalytic reaction> TiO 2 + hν → h + + e − (1) <Reduction> The reaction mechanism to reduce CO 2 with H 2 can be summarized as shown below [30,31]. <Photocatalytic reaction> TiO 2 + hν → h + + e − <Oxidization> Molecules 2020, 25 Though a few studies using pure TiO 2 under CO 2 /H 2 /H 2 O condition were reported [32][33][34], the effect of ratio of CO 2 , H 2 , and H 2 O, as well as the effect of Pd loading on CO 2 reduction characteristics was not investigated previously.
The purpose of this study is to clarify the effect of molar ratio of CO 2 to reductants of H 2 and H 2 O on CO 2 reduction characteristics with Pd/TiO 2 . Additionally, the present study also aims to clarify the optimum combination of reductants, as well as Pd loading weight with TiO 2 .
The present study employed TiO 2 films coated on netlike glass fibers (SILIGLASS U, Nihonmuki Co., Tokyo, Japan) by the sol-gel and dip-coating process. The glass fiber whose diameter is about 10 µm is weaved as a net, resulting in the diameter of collected fiber of approximately 1 mm. As to the specification of each fiber, the porous diameter is approximately 1 nm and the specific surface area is approximately 400 m 2 /g. The composition of netlike glass fiber is SiO 2 of 96 wt%. The aperture area is approximately 2 mm × 2 mm. Due to the porous structure of the netlike glass fiber, the TiO 2 film can be captured on netlike glass fiber easily in the step of preparation by sol-gel and dip-coating procedure. Additionally, it was believed that CO 2 would be more easily absorbed by the prepared photocatalyst since the porous fiber has a large surface area [35,36].
After the coating of TiO 2 , nanosized Pd particles were loaded on TiO 2 by the pulse arc plasma method applying high voltage. The pulse number can be controlled by the quantity of Pd loaded. The Pd loading weight on TiO 2 was measured by Electron Probe Micro Analyzer (EPMA).
In this paper, the characterization of Pd/TiO 2 was conducted by Scanning Electron Microscope (SEM), EPMA, Scanning Transmission Electron Microscope (STEM), Energy Dispersive X-ray Spectrometer (EDS), and Electron Energy Loss Spectrum (EELS) analysis before the CO 2 reduction experiment. The performances of CO 2 reduction with H 2 and H 2 O under the condition of illuminating Xe lamp including or excluding UV light were investigated in this paper. The combination of CO 2 /H 2 /H 2 O was changed for 1:0.5:0.5, 1:0.5:1, 1:1:0.5, 1:1:1, and 1:2:2 based on molar ratio to clarify the optimum combination of CO 2 /H 2 /H 2 O for CO 2 reduction with Pd/TiO 2 . If the amount of H 2 is larger than that of H 2 O, it is thought that the effect of H 2 O on the photocatalytic reaction is higher. On the other hand, if the amount of H 2 O is larger than that of H 2 , it is thought that the effect of H 2 O on the photocatalytic reaction is higher. This study investigated the effect of H 2 or H 2 O on the CO 2 reduction performance of Pd/TiO 2 under the condition of CO 2 /H 2 /H 2 O for the first time, so the originality of this study could be justified. In addition, the effect of Pd loading weight with TiO 2 on CO 2 reduction performance was also investigated in this study. Figure 2 show SEM images of TiO 2 film and Pd/TiO 2 film coated on netlike glass disc, respectively. The SEM images were taken at 1500 times magnification. In these figures, the red circles indicate TiO 2 according to EPMA results. Figures 3 and 4 show EPMA results of TiO 2 and Pd/TiO 2 film coated on netlike glass disc, respectively. The data with the weight percentage of Pd to Pd/TiO 2 film of 4.97 wt% are shown in Figure 4 as an example. In these figures, the different colors indicate the concentrations of each element in the observation area. For example, light colors such as white, pink, and red mean the quantity of element is small.

Characterization Analysis of Pd/TiO 2 Film
According to these figures, it is clear that TiO 2 film was coated on netlike glass fiber. In addition, it is observed that the crack is formed on the TiO 2 film. Since the thermal conductivity is different between Ti and SiO 2 , which are 19.4 W/(m K) and 1.82 W/(m K), respectively at 600 K [37], the temperature distribution of TiO 2 solution adhered on the netlike glass disc was not uniform during the firing process, as a result, cracks were formed on the TiO 2 film by the thermal expansion and shrinkage around netlike glass fiber. As to the crystal structure of TiO 2 , it is thought to be anatase since the firing temperature was set at 623 K in this study. A previous study [38] found the crystal structure of prepared TiO 2 was anatase if the firing temperature was from 673 K to 873 K, while it would be rutile if the firing temperature was 973 K. The uniform loading of nanosized Pd particles on TiO 2 was observed according to Figure 3.
The observation area of diameter of 300 µm is analyzed by EPMA to evaluate the quantity of loaded Pd within TiO 2 film. Twenty observation points obtained from several samples were used to determine the weight percentages of Pd and Ti in this study. As a result, the weight percentages of Pd to Pd/TiO 2 film prepared by changing pulse number in this study are 0.49 wt%, 3.90 wt%, and 4.97 wt%, while the weight percentage of Ti are 99.51 wt%, 96.10 wt%, and 95.03 wt%, respectively.
Molecules 2020, 25, x 4 of 18 the firing process, as a result, cracks were formed on the TiO2 film by the thermal expansion and shrinkage around netlike glass fiber. As to the crystal structure of TiO2, it is thought to be anatase since the firing temperature was set at 623 K in this study. A previous study [38] found the crystal structure of prepared TiO2 was anatase if the firing temperature was from 673 K to 873 K, while it would be rutile if the firing temperature was 973 K. The uniform loading of nanosized Pd particles on TiO2 was observed according to Figure 3. The observation area of diameter of 300 μm is analyzed by EPMA to evaluate the quantity of loaded Pd within TiO2 film. Twenty observation points obtained from several samples were used to determine the weight percentages of Pd and Ti in this study. As a result, the weight percentages of Pd to Pd/TiO2 film prepared by changing pulse number in this study are 0.49 wt%, 3.90 wt%, and 4.97 wt%, while the weight percentage of Ti are 99.51 wt%, 96.10 wt%, and 95.03 wt%, respectively.    the firing process, as a result, cracks were formed on the TiO2 film by the thermal expansion and shrinkage around netlike glass fiber. As to the crystal structure of TiO2, it is thought to be anatase since the firing temperature was set at 623 K in this study. A previous study [38] found the crystal structure of prepared TiO2 was anatase if the firing temperature was from 673 K to 873 K, while it would be rutile if the firing temperature was 973 K. The uniform loading of nanosized Pd particles on TiO2 was observed according to Figure 3. The observation area of diameter of 300 μm is analyzed by EPMA to evaluate the quantity of loaded Pd within TiO2 film. Twenty observation points obtained from several samples were used to determine the weight percentages of Pd and Ti in this study. As a result, the weight percentages of Pd to Pd/TiO2 film prepared by changing pulse number in this study are 0.49 wt%, 3.90 wt%, and 4.97 wt%, while the weight percentage of Ti are 99.51 wt%, 96.10 wt%, and 95.03 wt%, respectively.     Figure 5 shows STEM and EDS results of Pd/TiO 2 film coated on netlike glass disc. 250,000 times magnification STEM image was used for the EDS analysis. It is observed that Pd is coated on TiO 2 film according to STEM image, which is confirmed from EDS image. It is also observed that the layout of Pd and Ti are separated. The thickness of the Pd coated is approximately 60 nm. Nanosized Pd particles are loaded on TiO 2 dispersedly.    shows STEM and EDS results of Pd/TiO2 film coated on netlike glass disc. 250,000 times magnification STEM image was used for the EDS analysis. It is observed that Pd is coated on TiO2 film according to STEM image, which is confirmed from EDS image. It is also observed that the layout of Pd and Ti are separated. The thickness of the Pd coated is approximately 60 nm. Nanosized Pd particles are loaded on TiO2 dispersedly.    Figure 5 shows STEM and EDS results of Pd/TiO2 film coated on netlike glass disc. 250,000 times magnification STEM image was used for the EDS analysis. It is observed that Pd is coated on TiO2 film according to STEM image, which is confirmed from EDS image. It is also observed that the layout of Pd and Ti are separated. The thickness of the Pd coated is approximately 60 nm. Nanosized Pd particles are loaded on TiO2 dispersedly.  Figure 6 shows EELS spectra of Pd in Pd/TiO 2 film which peaks at around 540 eV. Comparing the spectra peaks of Pd nanowire with that of Pd metal and PdO in [39], the EELS spectra of Pd metal matches that in Figure 6. Therefore, it is believed that the Pd in Pd/TiO 2 prepared in this study exists as Pd metal. Since the photoreduction performance of Pd/TiO 2 was higher than that of PdO/TiO 2 [40,41], the desirable Pd/TiO 2 without oxidization was proved to be prepared in this study.  Figure 6 shows EELS spectra of Pd in Pd/TiO2 film which peaks at around 540 eV. Comparing the spectra peaks of Pd nanowire with that of Pd metal and PdO in [39], the EELS spectra of Pd metal matches that in Figure 6. Therefore, it is believed that the Pd in Pd/TiO2 prepared in this study exists as Pd metal. Since the photoreduction performance of Pd/TiO2 was higher than that of PdO/TiO2 [40,41], the desirable Pd/TiO2 without oxidization was proved to be prepared in this study.   Figure 6 shows EELS spectra of Pd in Pd/TiO2 film which peaks at around 540 eV. Comparing the spectra peaks of Pd nanowire with that of Pd metal and PdO in [39], the EELS spectra of Pd metal matches that in Figure 6. Therefore, it is believed that the Pd in Pd/TiO2 prepared in this study exists as Pd metal. Since the photoreduction performance of Pd/TiO2 was higher than that of PdO/TiO2 [40,41], the desirable Pd/TiO2 without oxidization was proved to be prepared in this study.   Figures 7,8,9,and 10 show the change in concentration of formed CO, CH4, C2H4, and C2H6 with Pd/TiO2 film coated on netlike glass disc with the time under the condition of Xe lamp illumination with UV light, respectively. In these figures, the impact of molar ratio of CO2, H2, and H2O, as well as Pd loading weight are also presented. Before this experiment, a blank test without Xe lamp illumination had been carried out as a reference, resulting that no fuel was detected as expected. Table 1 Figures 7,8,9,and 10 show the change in concentration of formed CO, CH4, C2H4, and C2H6 with Pd/TiO2 film coated on netlike glass disc with the time under the condition of Xe lamp illumination with UV light, respectively. In these figures, the impact of molar ratio of CO2, H2, and H2O, as well as Pd loading weight are also presented. Before this experiment, a blank test without Xe lamp illumination had been carried out as a reference, resulting that no fuel was detected as expected. Table 1 Tables 1-4, the CO 2 reduction performance to produce CO, CH 4 , C 2 H 4 , and C 2 H 6 is the highest at the molar ratio of CO 2 /H 2 /H 2 O = 1:0.5:0.5. Since the reaction scheme of CO 2 /H 2 /H 2 O has not been fully understood, Equations (1)-(15) are used to explain the results. Equations (1)- (15) show that the theoretical molar ratio of CO 2 with H 2 O or H 2 to produce CO is 1:1. On the other hand, the theoretical molar ratio of CO 2 with H 2 O or H 2 to produce CH 4 is 1:4. In addition, CH 4 , C 2 H 4 , and C 2 H 6 are produced in the series after CO is produced. For example, producing CH 4 needs four times H + and electrons as many as producing CO needs. The other fuels such as C 2 H 4 and C 2 H 6 need more H + and electrons compared to producing CH 4 . Since Pd has a high reduction performance [21,22,40], it is thought that the optimum molar ratio of CO 2 /total reductants to produce CH 4 , C 2 H 4 , and C 2 H 6 is smaller than the theoretical molar ratio required. Moreover, since the molar ratio of H 2 is the same as that of H 2 O under the molar ratio of CO 2 /H 2 /H 2 O = 1:0.5:0.5 condition, the effect of H 2 or H 2 O is not higher than that of the other to obtain the optimum molar ratio of CO 2 /H 2 /H 2 O over Pd/TiO 2 photocatalyst. However, according to Tables 1-4, the CO 2 reduction performance for the condition that the molar ratio of H 2 O is larger than that of H 2 is better, resulting in that the effect of H 2 O is bigger than that of H 2 to promote the CO 2 reduction performance over Pd/TiO 2 totally in this study.

Performance
In addition, it is known from Figures 7-10 and Tables 1-4 that the maximum concentration of produced fuel is obtained when Pd loading weight is 3.90 wt% irrespective of fuel type. One might think that the CO 2 reduction performance is promoted with increasing Pd loading weight. However, it is believed that too much Pd loading causes covering the surface of TiO 2 film [42,43], resulting in that CO 2 and reductants cannot attain the surface of TiO 2 film sufficiently. Consequently, it is clear that there is an optimum Pd loading weight to promote CO 2 reduction performance with H 2 and H 2 O.
Tables 5-8 list the maximum molar quantities of CO, CH 4 , C 2 H 4 , and C 2 H 6 per unit weight of photocatalyst under the condition of Xe lamp illumination with UV light, respectively. The quantities of Pd/TiO 2 coated on netlike glass disc for Pd loading weight of 0.44 wt%, 3.90 wt%, and 4.97 wt% are 0.05 g, 0.05 g, and 0.09 g, respectively. These quantities of Pd/TiO 2 coated on netlike glass disc were measured by an electric balance comparing the weights of several samples before and after preparing Pd/TiO 2 film on netlike glass fiber. The photocatalytic activity evaluation using molar quantities of product per weight of photocatalyst was adopted as in the recent photocatalyst studies [44][45][46][47].
According to Tables 5-8, the maximum molar quantities of CO, CH 4 , C 2 H 4 , and C 2 H 6 per unit weight of photocatalyst are obtained for the molar ratio of CO 2 /H 2 /H 2 O = 1:0.5:0.5. In addition, it is known that the maximum molar quantity of fuel per unit weight of photocatalyst is obtained for Pd loading weight of 3.90 wt% irrespective of fuel type. It is thought that these results agree with the results shown in Figures 11-14.     Figure 11 shows the change in concentration of formed CO with the Pd/TiO 2 film with the time under the condition of Xe lamp illumination without UV light. In this figure, the impact of molar ratio of CO 2 , H 2 , and H 2 O, as well as Pd loading weight is also presented. Before this experiment, a blank test without Xe lamp illumination had been carried out as a reference, resulting in that no fuel was detected as expected. Table 9 lists the maximum concentration of formed CO under the condition shown in Figure 11.   According to Figure 11 and Table 9, the CO2 reduction performance to produce CO is the highest at the molar ratio of CO2/H2/H2O = 1:0.5:0.5 and the maximum concentration of produced fuel is obtained for Pd loading weight of 3.90 wt%. These results are the same as that in the case of illuminating Xe lamp with UV light. The reason why these results are obtained is thought to be the same as explained above in the case of illuminating Xe lamp with UV light. It is found from Figure  11 that the concentration of formed CO is smaller than that under the condition of Xe lamp with UV light. There were no other fuels such as CH4, C2H4, and C2H6 detected under the condition of Xe lamp illumination without UV light. It is thought that the responsiveness of visible light with Pd/TiO2 prepared in this study was too low.   According to Figure 11 and Table 9, the CO 2 reduction performance to produce CO is the highest at the molar ratio of CO 2 /H 2 /H 2 O = 1:0.5:0.5 and the maximum concentration of produced fuel is obtained for Pd loading weight of 3.90 wt%. These results are the same as that in the case of illuminating Xe lamp with UV light. The reason why these results are obtained is thought to be the same as explained above in the case of illuminating Xe lamp with UV light. It is found from Figure 11 that the concentration of formed CO is smaller than that under the condition of Xe lamp with UV light. There were no other fuels such as CH 4 , C 2 H 4 , and C 2 H 6 detected under the condition of Xe lamp illumination without UV light. It is thought that the responsiveness of visible light with Pd/TiO 2 prepared in this study was too low. Table 10 shows the maximum molar quantity of CO per unit weight of photocatalyst under the condition of Xe lamp illumination without UV light. The maximum molar quantity of CO per unit weight of photocatalyst is obtained for the molar ratio of CO 2 /H 2 /H 2 O = 1:0.5:0.5 at Pd loading weight of 3.90 wt%. This result is the same as that in the case of illuminating Xe lamp with UV light. In this study, the maximum molar quantity of CH 4 per unit weight of photocatalyst is 22.1 µmol/g for the molar ratio of CO 2 /H 2 /H 2 O = 1:0.5:0.5 at Pd loading weight of 3.90 wt% under the condition of Xe lamp illumination with UV light. This maximum value is obtained after 6 h of illumination. According to the previous studies reported, the molar quantities of CH 4 per unit weight of photocatalyst in the case of CO 2 /H 2 O with Pd/TiO 2 were 25 µmol/g, 4.8 µmol/g, and 1.9 µmol/g [21,22,40]. These molar quantities of CH 4 per unit weight of photocatalyst were obtained after 8 [21], 6 [22], and 24 [40] h of illumination, respectively. Another study [23] reported that the molar quantity of CH 4 per unit weight of photocatalyst in the case of CO 2 /H 2 with Pd/TiO 2 was 356 µmol/g which was obtained after 3 h of illumination.
In this study, the maximum molar quantity of CO per unit weight of photocatalyst is 30.3 µmol/g for the molar ratio of CO 2 /H 2 /H 2 O = 1:0.5:0.5 at Pd loading weight of 3.90 wt% under the condition of Xe lamp illumination with UV light. This maximum value is obtained after illumination time of Xe lamp of 6 h. The previous studies reported that the molar quantities of CO per unit weight of photocatalyst in the case of CO 2 /H 2 O with Pd/TiO 2 were 0.12 µmol/g and 0.13 µmol/g [22,39], while the study reported that the molar quantity of CO per unit weight of photocatalyst in the case of CO 2 /H 2 with Pd/TiO 2 was 45 µmol/g [23]. These molar quantities of CO per unit weight of photocatalyst were obtained after illumination of 6 [22], 5 [39], and 3 [23] h, respectively.
Compared to the other studies, CO 2 reduction performance in terms of producing CH 4 or CO per unit weight of photocatalyst obtained in this study does not necessarily imply that the photocatalyst was prepared. Additionally, the best time to obtain the highest molar quantity of produced fuel per unit weight of photocatalsyt is almost the same as the previous studies. However, in terms of producing the other fuels such as C 2 H 4 and C 2 H 6 , which are difficult to produce through CO 2 reduction and were not reported in the other studies, are confirmed in this study. According to the previous study [21], Pd/TiO 2 could produce hydrocarbon such as C 2 H 6 more effectively compared to the other photocatalysts. The CO 2 molecules activated at Pd sites react with H + and the electrons to produce the intermediate Pd-C=O. Meanwhile, a small amount of CO is generated by C=O desorption, but Pd-C=O further interacts with the dissociated H to form a Pd-C species. Finally, the carbon species generated continue to react with the H species at Pd sites to produce CH 4 . During the CH 4 formation process, some intermediates (such as ·CH, ·CH 2 , and ·CH 3 ) are produced, and C 2 H 6 is obtained when two ·CH 3 species interact with each other. Since C 2 H 4 and C 2 H 6 have high heating values, producing these fuels have a profound significance in CO 2 utilization. Therefore, it can be said that this study has realized the photocatalyst having high CO 2 reduction performance.
Though it is thought that the doped Pd can provide the free electron not only to prevent the recombination of electron and hole produced but also to improve light absorption effect, it is necessary to improve the CO 2 reduction performance further. This study suggests that different metals should be doped on TiO 2 to promote the CO 2 reduction further. The co-doped TiO 2 such as PbS-Cu/TiO 2 , Cu-Fe/TiO 2 , Cu-Ce/TiO 2 , Cu-Mn/TiO 2 , and Cu-CdS/TiO 2 were reported to promote the CO 2 reduction performance of TiO 2 with H 2 O [4,48]. Then, the promotion of CO 2 reduction performance by different metal doping is expected when the combination of CO 2 /H 2 /H 2 O is considered. For example, Fe which can absorb the shorter wavelength light than Pd can [48] should be co-used since the amount of light absorbed by the photocatalyst can be increased and an effective utilization of wide range light can be realized by the combination of Fe and Pd.

Preparation of Pd/TiO 2 Photocatalyst
The TiO 2 film used in this study was prepared using the sol-gel and dip-coating procedure [24,49,50]. At first, [(CH 3 ) 2 CHO] 4 Ti (95 wt% purification, produced by Nacalai Tesque Co., Kyoto, Japan) of 0.3 mol, anhydrous C 2 H 5 OH (99.5 wt% purification, produced by Nacalai Tesque Co.) of 2.4 mol, distilled water of 0.3 mol, and HCl (35 wt% purification, produced by Nacalai Tesque Co.) of 0.07 mol were mixed to make the TiO 2 sol solution. As the basis to coat TiO 2 film, the sheet of netlike glass fiber was cut into a disc shape whose diameter and thickness were 50 mm and 1 mm, respectively. The disc shaped netlike glass fiber was then immersed into the TiO 2 sol solution at a speed of 1.5 mm/s and lifted at 0.22 mm/s. The disc was dried and heated at the controlling firing temperature (FT) and the firing duration time (FD) of 623 K and 180 s, respectively. After the TiO 2 film was coated on netlike glass disc, the pulse arc plasma method was selected to load Pd on the TiO 2 film. The pulse arc plasma gun device (ARL-300, produced by ULVAC, Inc., Suzuka, Japan) with Pd electrode having a diameter of 10 mm was used in this study. The quantity of loaded Pd was controlled by pulse number. In this study, the pulse number was varied from 100 to 500, and Pd loading weight with TiO 2 was measured by EPMA, for each pulse number. It is confirmed that the Pd/TiO 2 film prepared in this way could not be removed from the netlike glass fiber by rubbing. Figure 12 shows the photos of netlike glass disc before and after coating of Pd/TiO 2 . Since the sheet of netlike glass disc does not have a scouring structure inside it, the TiO 2 film is coated on the surface of netlike glass fiber and Pd can be deposited on TiO 2 film by pulse arc plasma method.
Molecules 2020, 25, x 13 of 18 0.07 mol were mixed to make the TiO2 sol solution. As the basis to coat TiO2 film, the sheet of netlike glass fiber was cut into a disc shape whose diameter and thickness were 50 mm and 1 mm, respectively. The disc shaped netlike glass fiber was then immersed into the TiO2 sol solution at a speed of 1.5 mm/s and lifted at 0.22 mm/s. The disc was dried and heated at the controlling firing temperature (FT) and the firing duration time (FD) of 623 K and 180 s, respectively. After the TiO2 film was coated on netlike glass disc, the pulse arc plasma method was selected to load Pd on the TiO2 film. The pulse arc plasma gun device (ARL-300, produced by ULVAC, Inc., Suzuka, Japan) with Pd electrode having a diameter of 10 mm was used in this study. The quantity of loaded Pd was controlled by pulse number. In this study, the pulse number was varied from 100 to 500, and Pd loading weight with TiO2 was measured by EPMA, for each pulse number. It is confirmed that the Pd/TiO2 film prepared in this way could not be removed from the netlike glass fiber by rubbing. Figure 12 shows the photos of netlike glass disc before and after coating of Pd/TiO2. Since the sheet of netlike glass disc does not have a scouring structure inside it, the TiO2 film is coated on the surface of netlike glass fiber and Pd can be deposited on TiO2 film by pulse arc plasma method.

Characterization of Pd/TiO2 Film
The structural and crystal characteristics of Pd/TiO2 film prepared were evaluated by using SEM (JXA-8530F, produced by JEOL Ltd., Tokyo, Japan), EPMA (JXA-8530F, produced by JEOL Ltd., Tokyo, Japan), and EELS (JEM-ARM2007 Cold, produced by JEOL Ltd., Tokyo, Japan). In order to analyze the sample by these equipments, carbon was coated on Pd/TiO2 whose thickness was approximately 15 nm by the dedicated device (JEC-1600, produced by JEOL Ltd.) before analysis. This carbon coating was conducted for analysis, while the CO2 reduction experiment was carried out without carbon coating. The carbon coating was not conducted for the right photo in Figure 1.
The electron was emitted on the sample by the electron probe applying the acceleration voltage of 15 kV and the current at 3.0 × 10 −8 A to analyze the surface structure of the sample by SEM. Simultaneously, EPMA detects the characteristic X-ray. The space resolutions for SEM and EPMA were set at 10 μm. The state of prepared photocatalyst, as well as the quantity of doped metal within TiO2 film could be known by EPMA analysis.
The electron probe emits electrons to the sample at the acceleration voltage of 200 kV, when the inner structure of the sample is analyzed by STEM. The size, thickness, and structure of loaded Pd were evaluated. The X-ray characteristics of the sample is detected by EDS at the same time. Therefore, the concentration distribution of chemical elements toward thickness direction of the sample is known. In the present paper, the concentration distribution of Ti, Pd, and Si were analyzed.
EELS is used to detect elements, as well as to determine oxidation states of transition metals. The EELS characterization was determined by JEM-ARM200F equipped with GIF Quantum having 2048 ch. The dispersion of 0.5 eV/ch for the full width at half maximum of the zero loss peak was achieved in the study.

Characterization of Pd/TiO 2 Film
The structural and crystal characteristics of Pd/TiO 2 film prepared were evaluated by using SEM (JXA-8530F, produced by JEOL Ltd., Tokyo, Japan), EPMA (JXA-8530F, produced by JEOL Ltd., Tokyo, Japan), and EELS (JEM-ARM2007 Cold, produced by JEOL Ltd., Tokyo, Japan). In order to analyze the sample by these equipments, carbon was coated on Pd/TiO 2 whose thickness was approximately 15 nm by the dedicated device (JEC-1600, produced by JEOL Ltd.) before analysis. This carbon coating was conducted for analysis, while the CO 2 reduction experiment was carried out without carbon coating. The carbon coating was not conducted for the right photo in Figure 1.
The electron was emitted on the sample by the electron probe applying the acceleration voltage of 15 kV and the current at 3.0 × 10 −8 A to analyze the surface structure of the sample by SEM. Simultaneously, EPMA detects the characteristic X-ray. The space resolutions for SEM and EPMA were set at 10 µm. The state of prepared photocatalyst, as well as the quantity of doped metal within TiO 2 film could be known by EPMA analysis.
The electron probe emits electrons to the sample at the acceleration voltage of 200 kV, when the inner structure of the sample is analyzed by STEM. The size, thickness, and structure of loaded Pd were evaluated. The X-ray characteristics of the sample is detected by EDS at the same time. Therefore, the concentration distribution of chemical elements toward thickness direction of the sample is known. In the present paper, the concentration distribution of Ti, Pd, and Si were analyzed.
EELS is used to detect elements, as well as to determine oxidation states of transition metals. The EELS characterization was determined by JEM-ARM200F equipped with GIF Quantum having 2048 ch. The dispersion of 0.5 eV/ch for the full width at half maximum of the zero loss peak was achieved in the study. Figure 13 shows the experimental setup of the reactor composed of a stainless tube (height of 100 mm and inside diameter of 50 mm), Pd/TiO 2 film coated on netlike glass disc (diameter of 50 mm and thickness of 1 mm) located on the teflon cylinder (height of 50 mm and diameter of 50 mm), a quartz glass disc (diameter of 84 mm and thickness of 10 mm), an edge cut filter cutting off the light whose wavelength is below 400 nm (SCF-49.5C-42L, produced by SIGMA KOKI CO. LTD., Tokyo, Japan), a 150 W Xe lamp (L2175, produced by Hamamatsu Photonics K. K.), mass flow controller, gas cylinder of CO 2 and H 2 .

CO 2 Reduction Experiment
The reactor volume available for CO 2 is 1.25 × 10 −4 m 3 . The light of Xe lamp which is located outside the stainless tube illuminates Pd/TiO 2 film coated on the netlike glass disc through the edge cut filter and the quartz glass disc that are at the top of the stainless tube. The wavelength of light illuminating by Xe lamp is distributed from 185 nm to 2000 nm. Since an edge cut filter can remove UV components of the light from the Xe lamp, the wavelength from Xe lamp is distributed from 401 to 2000 nm with the filter. Figure 14 shows the spectra data on light intensity of Xe lamp without the edge filter according to the catalog of Xe lamp company. Figure 15 shows the performance of the edge cut filter to cut off the wavelength of light whose wavelength is below 400 nm. The average light intensities of Xe lamp without and with the edge cut filter are 65.0 W/cm 2 and 40.5 W/cm 2 , respectively. CO 2 gas and H 2 gas whose purity were 99.995 vol% and 99.99999 vol%, respectively were controlled by mass flow controller and mixed in the buffer chamber before the experiment. The mixing ratio of CO 2 and H 2 was checked and confirmed by TCD gas chromatograph (Micro GC CP4900, produced by GL Science, Tokyo, Japan) before being introduced into the reactor. The distilled water was then injected into the reactor via gas sampling tap and when Xe lamp was turned on. The water was injected and vaporized by the heat of Xe lamp completely. The molar ratio of CO 2 /H 2 /H 2 O was set at 1:0.5:0.5, 1:0.5:1, 1:1:0.5, 1:1:1, 1:2:2. The temperature in reactor rose up to 343 K within 1 h and was kept at about 343 K during the entire experiment. Figure 13 shows the experimental setup of the reactor composed of a stainless tube (height of 100 mm and inside diameter of 50 mm), Pd/TiO2 film coated on netlike glass disc (diameter of 50 mm and thickness of 1 mm) located on the teflon cylinder (height of 50 mm and diameter of 50 mm), a quartz glass disc (diameter of 84 mm and thickness of 10 mm), an edge cut filter cutting off the light whose wavelength is below 400 nm (SCF-49.5C-42L, produced by SIGMA KOKI CO. LTD., Tokyo, Japan), a 150 W Xe lamp (L2175, produced by Hamamatsu Photonics K. K.), mass flow controller, gas cylinder of CO2 and H2.

CO2 Reduction Experiment
The reactor volume available for CO2 is 1.25×10 −4 m 3 . The light of Xe lamp which is located outside the stainless tube illuminates Pd/TiO2 film coated on the netlike glass disc through the edge cut filter and the quartz glass disc that are at the top of the stainless tube. The wavelength of light illuminating by Xe lamp is distributed from 185 nm to 2000 nm. Since an edge cut filter can remove UV components of the light from the Xe lamp, the wavelength from Xe lamp is distributed from 401 to 2000 nm with the filter. Figure 14 shows the spectra data on light intensity of Xe lamp without the edge filter according to the catalog of Xe lamp company. Figure 15 shows the performance of the edge cut filter to cut off the wavelength of light whose wavelength is below 400 nm. The average light intensities of Xe lamp without and with the edge cut filter are 65.0 W/cm 2 and 40.5 W/cm 2 , respectively.
CO2 gas and H2 gas whose purity were 99.995 vol% and 99.99999 vol%, respectively were controlled by mass flow controller and mixed in the buffer chamber before the experiment. The mixing ratio of CO2 and H2 was checked and confirmed by TCD gas chromatograph (Micro GC CP4900, produced by GL Science, Tokyo, Japan) before being introduced into the reactor. The distilled water was then injected into the reactor via gas sampling tap and when Xe lamp was turned on. The water was injected and vaporized by the heat of Xe lamp completely. The molar ratio of CO2/H2/H2O was set at 1:0.5:0.5, 1:0.5:1, 1:1:0.5, 1:1:1, 1:2:2. The temperature in reactor rose up to 343 K within 1 h and was kept at about 343 K during the entire experiment.
In the CO2 reduction experiment with UV light, samples of the gas in the reactor were taken every 6 h, while in the CO2 reduction experiment without UV light samples were taken every 24 h due to the difference of reaction speed of prepared photocatalyst under these two conditions. The gas samples were analyzed using FID gas chromatograph (GC353B, produced by GL Science) and methanizer (MT221, produced by GL Science). FID gas chromatograph and metanizer can be analyzed in the minimum range of 1 ppmV.   In the CO 2 reduction experiment with UV light, samples of the gas in the reactor were taken every 6 h, while in the CO 2 reduction experiment without UV light samples were taken every 24 h due to the difference of reaction speed of prepared photocatalyst under these two conditions. The gas samples were analyzed using FID gas chromatograph (GC353B, produced by GL Science) and methanizer (MT221, produced by GL Science). FID gas chromatograph and metanizer can be analyzed in the minimum range of 1 ppmV.

Conclusions
The following conclusions could be drawn from this study: • The nanosized Pd particles could be loaded on TiO2 uniformly by the pulse arc plasma method. Pd in Pd/TiO2 prepared by this method exists in the form of Pd metal.
• The highest CO2 reduction performance to produce CO, CH4, C2H4, and C2H6 was obtained at the molar ratio of CO2/H2/H2O = 1:0.5:0.5 with Xe lamp illumination with or without UV light. It is revealed that the molar ratio of CO2/total reductants = 1:1 is the optimum to produce fuels.
• The maximum molar quantity of fuel per unit weight of photocatalyst is obtained at Pd loading weight of 3.90 wt% irrespective of fuel type. In this study, the maximum molar quantities of CO and CH4 per unit weight of photocatalyst were 30.3 μmol/g and 22.1 μmol/g, respectively, for the molar ratio of CO2/H2/H2O = 1:0.5:0.5 at Pd loading weight of 3.90 wt% under the condition of Xe lamp illumination with UV light.
• The Pd/TiO2 photocatalyst prepared in this study could produce C2H4 and C2H6, as well as CO and CH4, therefore, it can be said that the photocatalyst prepared in this study has realized to have the higher CO2 reduction performance.

Conclusions
The following conclusions could be drawn from this study: • The nanosized Pd particles could be loaded on TiO2 uniformly by the pulse arc plasma method. Pd in Pd/TiO2 prepared by this method exists in the form of Pd metal.
• The highest CO2 reduction performance to produce CO, CH4, C2H4, and C2H6 was obtained at the molar ratio of CO2/H2/H2O = 1:0.5:0.5 with Xe lamp illumination with or without UV light. It is revealed that the molar ratio of CO2/total reductants = 1:1 is the optimum to produce fuels.
• The maximum molar quantity of fuel per unit weight of photocatalyst is obtained at Pd loading weight of 3.90 wt% irrespective of fuel type. In this study, the maximum molar quantities of CO and CH4 per unit weight of photocatalyst were 30.3 μmol/g and 22.1 μmol/g, respectively, for the molar ratio of CO2/H2/H2O = 1:0.5:0.5 at Pd loading weight of 3.90 wt% under the condition of Xe lamp illumination with UV light.
• The Pd/TiO2 photocatalyst prepared in this study could produce C2H4 and C2H6, as well as CO and CH4, therefore, it can be said that the photocatalyst prepared in this study has realized to have the higher CO2 reduction performance.

Conclusions
The following conclusions could be drawn from this study: • The nanosized Pd particles could be loaded on TiO 2 uniformly by the pulse arc plasma method. Pd in Pd/TiO 2 prepared by this method exists in the form of Pd metal. • The highest CO 2 reduction performance to produce CO, CH 4 , C 2 H 4 , and C 2 H 6 was obtained at the molar ratio of CO 2 /H 2 /H 2 O = 1:0.5:0.5 with Xe lamp illumination with or without UV light. It is revealed that the molar ratio of CO 2 /total reductants = 1:1 is the optimum to produce fuels.

•
The maximum molar quantity of fuel per unit weight of photocatalyst is obtained at Pd loading weight of 3.90 wt% irrespective of fuel type. In this study, the maximum molar quantities of CO and CH 4 per unit weight of photocatalyst were 30.3 µmol/g and 22.1 µmol/g, respectively, for the molar ratio of CO 2 /H 2 /H 2 O = 1:0.5:0.5 at Pd loading weight of 3.90 wt% under the condition of Xe lamp illumination with UV light.

•
The Pd/TiO 2 photocatalyst prepared in this study could produce C 2 H 4 and C 2 H 6 , as well as CO and CH 4 , therefore, it can be said that the photocatalyst prepared in this study has realized to have the higher CO 2 reduction performance.