Comparison of CO2 Reduction Performance with NH3 and H2O between Cu/TiO2 and Pd/TiO2

The aim of this study is to clarify the effect of doped metal type on CO2 reduction characteristics of TiO2 with NH3 and H2O. Cu and Pd have been selected as dopants for TiO2. In addition, the impact of molar ratio of CO2 to reductants NH3 and H2O has been investigated. A TiO2 photocatalyst was prepared by a sol-gel and dip-coating process, and then doped with Cu or Pd fine particles by using the pulse arc plasma gun method. The prepared Cu/TiO2 film and Pd/TiO2 film were characterized by SEM, EPMA, TEM, STEM, EDX, EDS and EELS. This study also has 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 the CO2 reduction performance with Cu/TiO2 under the illumination condition of Xe lamp with UV light is the highest when the molar ratio of CO2/NH3/H2O = 1:1:1 while that without UV light is the highest when the molar ratio of CO2/NH3/H2O = 1:0.5:0.5. It is revealed that the CO2 reduction performance of Pd/TiO2 is the highest for the molar ratio of CO2/NH3/H2O = 1:1:1 no matter the used Xe lamp was with or without UV light. The molar quantity of CO per unit weight of photocatalyst for Cu/TiO2 produced under the illumination condition of Xe lamp with UV light was 10.2 μmol/g, while that for Pd/TiO2 was 5.5 μmol/g. Meanwhile, the molar quantity of CO per unit weight of photocatalyst for Cu/TiO2 produced under the illumination condition of Xe lamp without UV light was 2.5 μmol/g, while that for Pd/TiO2 was 3.5 μmol/g. This study has concluded that Cu/TiO2 is superior to Pd/TiO2 from the viewpoint of the molar quantity of CO per unit weight of photocatalyst as well as the quantum efficiency.


Introduction
Because of large concerns around the world, the global warming problem is a hot area of R&D. Each country has set a goal to reduce the amount of CO 2 emissions. In Japan, the prime minister has declared the intent to reduce the effective CO 2 emissions to zero by 2050. However, the global mean concentration of CO 2 in atmosphere had increased up to 410 ppmV in September 2019, which was 25 ppmV increase from the value in 2009 [1]. Therefore, development of technologies which can reduce the amount of CO 2 in the atmosphere the is urgently required.
Solar conversion of CO 2 to fuel seems a promising procedure to solve the global warming problem for sustainable development of society. Solar energy, is the form of direct solar irradiation, is widely available and it is imperious to utilize it for solar fuel production [2]. One pathway to realize solar conversion of CO 2 is photochemical reactions. According to a literature survey by the authors, photocatalysts can convert CO 2 into fuel species such as CO, CH 4 , CH 3 OH, etc. [3][4][5]. TiO 2 is a popular photocatalyst used for CO 2 reduction since it is convenient to obtain, inexpensive, and has strong resistance to chemicals and corrosion [6]. Pure TiO 2 can only function under UV light illumination which represent only 4% of the energy available in solar radiation [4]. CO 2 reduction performance is thus greatly improved if the TiO 2 or modified TiO 2 can function under visible light illumination.
Some studies have reported on the development of various modified TiO 2 forms [3][4][5]. The modifications included loading TiO 2 with Au, Ag, Pd, Pt, Rh, Ir or bimetals (e.g., Ag-Au and Au-Pt) [7]. A hierarchical pore network and morphology to prepare the biotemplated TiO 2 catalyst [8], heteroleptic iridium complex supported on graphite carbon nitride [9], TiO 2 synthesis using superficial fluid technology [10] and N-doped reduced graphene oxide promoted nano TiO 2 [11] were attempted modifications to prepare TiO 2 to respond to visible light. According to a review paper on the surface modification of TiO 2 to enhance its CO 2 reduction performance [12], there are many approaches for the surface modification of TiO 2 such as impurity doping, metal deposition, alkali modification, heterojunction construction and carbon-based material loading. As an example of impurity doping, the CH 3 OH production of Cu/TiO 2 increased with an increase in the amount of Cu doping, while the over-doping of Cu would lead to high defect density in the TiO 2 , resulting in degradation of the CO 2 reduction performance [12]. Therefore, there is the optimum ratio of dopant to enhance the CO 2 reduction performance of photocatalyst. As to an example of heterojunction construction, it was reported that the photocatalytic CO 2 reduction performance over Cu/TiO 2 hollow nanoparticles was much better than that of pure TiO 2 and Cu 2 O [12]. On the other hand, another recent review paper reported that not only Cu 2 O but also Cu 2 O/TiO 2 hybrid photocatalyst exhibited a higher CO 2 reduction performance compared to pure TiO 2 [13]. Cu 2 O which has a small band gap energy (2.4 V) can help the absorption efficiency in the visible range of the solar spectrum. In addition, Cu 2 O can promote the CO 2 reduction performance by concomitantly increasing the electron-hole separation efficiency. It was also reported that Cu 2 O/TiO 2 proceeded the photocatalytic reaction, resulting in the increase in the selective formation of CO under the illumination condition of light whose wave length was over 305 nm.
Though various metals have been used for doping, the reductant which is a partner for CO 2 reduction is also important. According to the literatures survey by the authros, H 2 O or H 2 were normally used as the reductants for CO 2 reduction [2,7]. The reaction scheme to reduce CO 2 with H 2 O can be summarized as shown below according to the previous studies [14][15][16]: Photocatalytic reaction Oxidation Reduction The reaction scheme to reduce CO 2 with H 2 can be summarized as follows [17]: Photocatalytic reaction Oxidation Reduction In the reduction process, the same number of H + and e − are necessary. Since the doping metal emits the electron which is contributed to prevent the recombination of h + and e − [17], the number of H + should be arranged. Therefore, the combination of doped metal type and reductants is important.
Though various metals have been used for doping, Cu and Pd are favorite candidates [2]. Cu can improve TiO 2 photoactivity and selectivity in the CO 2 photocatalytic application [2]. Cu can extend the absorption band to 400-800 nm [18,19] which covers the whole visible light range. It was reported that Cu/TiO 2 was superior to pure TiO 2 . Cu/Cu + fabricated Ti 3+ /TiO 2 can produce 8 µmol/g of CH 4 which is 2.6 times more than in the case of Ti 3+ /TiO 2 [20]. Cu/TiO 2 prepared by a facile solvothermal method had yields of CO and CH 4 up to 4.48 µmol/g and 5.34 µmol/g, which are 10 times higher than those of TiO 2 [21]. It was reported that the synthesized Cu 2 O/TiO 2 showed a performance of 3.5 µmol/g of CO production while that of TiO 2 was 0.1 µmol/g [18]. These results [18,20,21] were achieved by CO 2 reduction with H 2 O under visible light illumination conditions. On the other hand, Pd can also extend the absorption band to 400 -800 nm [22,23], which covers the whole visible light range. Pd/TiO 2 exhibited a higher reduction performance to produce hydrocarbons and H 2 compared to pure TiO 2 [22][23][24]. This is due to the work function of Pd, which reflects its electron donating or accepting ability. In addition, it is thought that Pd loaded on TiO 2 functions to increase the efficiency of photogenerated electrons for the formation of reductive products. Pd/TiO 2 nanowire produced 50.4 µmol/g of CO and 26.7 µmol/g of CH 4 which were an improvement by 54% and 7%, respectively, compared to those of TiO 2 nanowire [25]. The other study reported that the production of Pd/TiO 2 was 3.50 µmol/g which was 2.5 times as large as that of pure TiO 2 [26]. Pd/TiO 2 prepared by a photochemical deposition method exhibited 0.28 µmol/g of CH 4 which was 14 times as large as that of pure TiO 2 (Degussa P-25) [27]. These results [25][26][27] were obtained from CO 2 reduction with H 2 O under the visible light illumination conditions.
Though there are some reports on CO 2 reduction with H 2 O or H 2 [3,27], the effect of NH 3 having 3H + , which is superior to H 2 O and H 2 , on CO 2 reduction performance of photocatalyst is not investigated yet with the exception of the previous studies conducted by Nishiura et al. using Fe [28] or Cu [29]. In addition, other doped metals have not been investigated yet from the viewpoint of comparison of several metal ion types. When the combination of CO 2 /NH 3 /H 2 O is considered, the ion number of dopants is important to match the number of electrons emitted from the dopant with H + as shown in the reaction scheme. The same number of electrons and H + is necessary to produce fuel. The reaction scheme to reduce CO 2 with NH 3 can be summarized as shown below [15,30]: Photocatalytic reaction Oxidation Reduction It is thought that the total amount of electron which is needed for photochemical reaction is large due to the combination of two H + suppliers such as NH 3 and H 2 O, according to the reaction scheme. Since Pd has a high reduction performance [23,24,31] which can assist the progress of reduction reaction in CO 2 reduction with NH 3 and H 2 O, this study selected Pd as a dopant as well as Cu.
The purpose of this study was to clarify the effect of doped metal type on the CO 2 reduction characteristics of TiO 2 with NH 3 and H 2 O. The CO 2 reduction performance with NH 3 and H 2 O using Cu/TiO 2 or Pd/TiO 2 coated on netlike glass fiber as photocatalyst has been investigated under the illumination conditions of a Xe lamp with or without UV light. In the study, the ratio of CO 2 /NH 3 /H 2 O has been set at 1:1:1, 1:0.5:1, 1:1:0.5, 1:0.5:0.5, 3:2:3, 3:8:12, respectively, to determine the optimum molar ratio of CO 2 /NH 3 /H 2 O with Cu/TiO 2 or Pd/TiO 2 as photocatalyst. According to the reaction scheme to reduce CO 2 with H 2 O or NH 3 , shown above, the theoretical molar ratio of CO 2 /H 2 O to produce CO or CH 4 should be 1:1 or 1:4, respectively, while that of CO 2 /NH 3 to produce CO or CH 4 should be 3:2, 3:8, respectively. Therefore, this study assumes that the molar ratio of CO 2 /NH 3 /H 2 O = 3:2:3 and 3:8:12 are theoretical molar ratios to produce CO and CH 4 , respectively.

Preparation of Cu/TiO 2 and Pd/TiO 2 Photocatalyst
TiO 2 film was prepared by sol-gel and dip-coating process [29]. [(CH 3 ) 2 CHO] 4 Ti (purity of 95 wt%, Nacalai Tesque Co., Kyoto, Japan) of 0.3 mol, anhydrous C 2 H 5 OH (purity of 99.5 wt%, Nacalai Tesque Co., Kyoto, Japan) of 2.4 mol, distilled water of 0.3 mol, and HCl (purity of 35 wt%, Nacalai Tesque Co., Kyoto, Japan) of 0.07 mol were mixed for preparing TiO 2 sol solution. This study coats TiO 2 film on netlike glass fiber (SILIGLASS U, Nihonmuki Co., Kyoto, Japan) by a sol-gel and dip-coating process. Glass fiber having diameter of about 10 µm weaved as a net is collected to give a diameter of approximately 1 mm. The pore diameter of the glass fiber and the specific surface area are approximately 1 nm and 400 m 2 /g, respectively from the specifications of the netlike glass fiber. The netlike glass fiber is composed of SiO 2 (96 wt%). The opening space of the net glass is approximately 2 mm × 2 mm. Since the netlike glass fiber has porous characteristics, the netlike glass fiber can capture TiO 2 film easily during the sol-gel and dip-coating process. Additionally, we can expect that CO 2 is more easily absorbed by the prepared photocatalyst due to the porous characteristics of the netlike glass fiber. This study cut the netlike glass fiber into disc forms having a diameter and thickness of 50 mm and 1 mm, respectively. The netlike glass disc was dipped into a TiO 2 sol solution at the speed of 1.5 mm/s and pulled up it at the fixed speed of 0.22 mm/s. After that, the net was dried out and fired under a controlled firing temperature (FT) and firing duration time (FD) to fasten TiO 2 film on the base material. This study set FT and FD at 623 K and 180 s, respectively.
After the coating of TiO 2 , this study loaded Cu or Pd on the TiO 2 coated netlike glass fiber by a pulse arc plasma gun method [29] emitting nanosized Cu or Pd particles uniformly under an applied high voltage potential difference. The pulse number can control the quantity of metal loaded on TiO 2 . This study set the pulse number at 100. This study applied an ARL-300pulse arc plasma gun device (ULVAC, Inc., Chigasaki, Japan) with a Cu or Pd electrode whose diameter was 10 mm for Cu or Pd loading, respectively. After the netlike glass fiber coated with TiO 2 was set in the evacuated vessel of the pulse arc plasma gun device, the Cu or Pd electrode emitted nanosized Cu or Pd particles by applying a voltage potential difference of 200 V. The pulse arc plasma gun can evaporate Cu or Pd electrodes into fine particle form over the target in a concentric area whose diameter is 100 mm under the condition that the distance between Cu or Pd electrode and the target is set to be 160 mm. Due to the distance between Cu or Pd electrode and TiO 2 film of 150 mm, these conditions can uniformly spread Cu or Pd particles over the TiO 2 film.
These measuring instruments use electrons to characterize materials, meaning that the samples should conduct electricity. Because the netlike glass disc used for base material to coat Cu/TiO 2 or Pd/TiO 2 film can't conduct electricity, a carbon vapor was depositied by a dedicated device (JEE-420, produced by JEOL Ltd., Tokyo, Japan) on the netlike glass discs before characterization. The thickness of the carbon deposited on samples is controlled to be approximately 20-30 nm. The electrode emits the electrons to the sample by setting the acceleration voltage of 15 kV and the current at 3.0 × 10 −8 A in order to analyze the external structure of samples by SEM. After the X-ray characteristics are analyzed by EPAM, the concentration of chemical elements is clarified referring to the relationship between the characteristic X-ray energy and the atomic number. SEM and EPMA have a spatial resolution of 10 mm. The EPMA analysis can help clarify the structure of the prepared photocatalysts as well as to measure the quantity of loaded metal within TiO 2 film on the netlike glass disc as base material. 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 TEM and STEM. The size, thickness and structure of loaded Cu and Pu were evaluated by TEM and STEM, respectively. The X-ray characteristics of the sample is detected by EDX and EDS at the same time, so the concentration distribution of chemical element in the thickness direction of the samples is known. The size, thickness and structure of loaded Cu and Pd were evaluated by TEM and STEM, respectively. The characterization of X-ray is detected by EDX and EDS at the same time, resulting in that the concentration distribution of chemical elements in the thickness direction of the samples is analyzed. EELS is used to detect elements as well as to determine the oxidation states of transition metals. The EELS characterization was determined by a JEM-ARM200F system equipped with GIF Quantum having 2048 ch. The dispersion of 0.5 eV/ch for the full width ad half maximum of the zero loss peak was measured in this study. , a sharp 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., Hamamatsu, Japan), mass flow controller and CO 2 gas cylinder [29]. The reactor size to charge CO 2 is 1.25 × 10 −4 m 3 . The light of Xe lamp which is positioned on the top of the stainless tube illuminates Cu/TiO 2 film or Pd/TiO 2 coated on the netlike glass disc through the sharp cut filter and the quartz glass disc that are located on the top of the stainless tube. Xe lamp has the wavelength of light ranged from 185 nm to 2000 nm. The sharp cut filter can get rid of UV from the Xe lamp, resulting that the wavelength of light illuminating to Cu/TiO 2 film or Pd/TiO 2 film ranges from 401 nm to 2000 nm with the filter. In this study, the average light intensity of Xe lamp without and with the sharp cut filter is 58.7 mW/cm 2 and 47.1 mW/cm 2 , respectively. After filling CO 2 gas of 99.995 vol% purity in the reactor which was pre-evacuated by a vacuum pump for 15 min, the valves positioned at the inlet and the outlet of reactor were closed in the CO 2 reduction experiment with NH 3 + H 2 O. After that, we confirmed that the pressure and gas temperature in the reactor at 0.1 MPa and 298 K, respectively. Then, we injected NH 3 aqueous solution (NH 3 ; 50 vol%), which was changed depending on the planed molar ratio, into the reactor via gas sampling tap, and turned on Xe lamp at a time. Due to the heat of infrared light components illuminated from Xe lamp, the injected NH 3 aqueous solution vaporized completely in the reactor. The temperature in the reactor reached at 343 K within an hour and it was maintained at approximately 343 K during the experiment. We changed the molar ratio of CO 2 /NH 3 /H 2 O at 1:1:1, 1:0.5:1, 1:1:0.5, 1:0.5:0.5, 3:2:3, 3:8:12, respectively. The reacted gas in the reactor was extracted by gas syringe via gas sampling tap and it was analyzed by FID gas chromatography (GC353B, GL Science, Tokyo, Japan) and a methanizer (MT221, GL Science, Tokyo, Japan). The FID gas chromatograph and methanizer have a minimum resolution of 1 ppmV. Figures 2 and 3 show SEM and EPMA images of Cu/TiO 2 and Pd/TiO 2 film coated on netlike glass disc, respectively. Black and white SEM images at 1500 times magnification were obtained in this study, which were also used for EPAM analysis. As to the EPMA image, the concentrations of each element in observation area are displayed by diverse colors. Light colors, e.g., white, pink, and red are used to display a large amount of an element. On the other hand, dark colors like black and blue are used to display a small amount of element. According to Figures 2 and 3, it is observed that TiO 2 film having teeth-like shape coated on the netlike glass fiber is formed irrespective of pulse number. Since the thermal conductivity of Ti and SiO 2 at 600 K are 19.4 W/(m·K) and 1.82 W/(m·K), respectively [33], the temperature distribution of TiO 2 solution adhered on the net like glass disc was not even during the firing process. Thermal expansion and shrinkage around netlike glass fibers occurred, resulting in the formation of thermal cracks within the TiO 2 film. Therefore, it is believed that TiO 2 film on netlike glass fiber has a teeth-like form. In addition, it was found that nanosized Cu and Pd particles were loaded on TiO 2 film uniformly. The observation area which is the center of netlike glass disc having the diameter of 300 µm was analyzed by EPMA to measure the amount of loaded Cu or Pd within the TiO 2 film. The ratio of Cu or Pd to Ti is calculated by averaging the data detected in this area. The weight percentage of element Cu within Cu/TiO 2 film was 1.62 wt%, while the weight percentage of element Pd within Pd/TiO 2 film was 1.64 wt%. The weight percentages of loaded Cu and Pd were almost the same, indicating that pulse arc plasma gun method could control the amount of metal doped on TiO 2 irrespective of metal type. On the other hand, total weights of Cu/TiO 2 and Pd/TiO 2 which were measured by an electron balance and averaged among 10 samples are 0.05 g and 0.07 g, respectively.  Cu particles are loaded on the upside of TiO 2 film, it is not confirmed that a Cu layer is formed [32].   Figure 6 shows STEM and EDS results of Pd/TiO 2 film coated on the netlike glass disc. A 250,000 times magnification STEM image was used in the EDS. It is observed from the STEM image that Pd is coated on the TiO 2 film, which is confirmed by EDS images, too. It is also observed that the layers of Pd and Ti are separated. It is seen that the thickness of the Pd coated is approximately 60 nm. The observation area is small compared to EPMA images shown in Figure 3, suggesting that nano-sized Pd particles are loaded on TiO 2 dispersedly [34]. Figure 7 shows the EELS spectra of Cu in Cu/TiO 2 film. According to this figure, peaks at around 932 eV and 952 eV can be observed. Compared to a report investigating the spectral peaks of Cu 2 O and CuO [35], the EELS spectra of Cu 2 O matches Figure 7. Therefore, Cu in Cu/TiO 2 prepared in this study exists as Cu + ion in Cu 2 O. It was reported that Cu + was more active than Cu 2+ [36]. Consequently, it is expected that Cu + plays a role in enhancing the CO 2 reduction performance. Figure 8 shows EELS spectra of Pd in Pd/TiO 2 film which displays peaks at around 540 eV. Comparing the spectra peaks of Pd nanowire with that of Pd metal and PdO [28], it is seen that the EELS spectra of Pd metal matches that in Figure 8. Therefore, it can be thought that Pd in Pd/TiO 2 prepared in this study exists as Pd metal. Since the photore-Molecules 2021, 26, 2904 9 of 18 duction performance of Pd/TiO 2 was higher than that of PdO/TiO 2 [31,37], it is confirmed that the desirable Pd/TiO 2 without oxidation was prepared in this study.  Figure 6 shows STEM and EDS results of Pd/TiO2 film coated on the netlike glass disc. A 250,000 times magnification STEM image was used in the EDS. It is observed from the STEM image that Pd is coated on the TiO2 film, which is confirmed by EDS images, too. It is also observed that the layers of Pd and Ti are separated. It is seen that the thickness of the Pd coated is approximately 60 nm. The observation area is small compared to EPMA images shown in Figure 3, suggesting that nano-sized Pd particles are loaded on TiO2 dispersedly [34].

Phys. Astron. Int. J.
7 shows the EELS spectra of Cu in Cu/TiO2 film. According to this figure, peaks at around 932 eV and 952 eV can be observed. Compared to a report investigating the spectral peaks of Cu2O and CuO [35], the EELS spectra of Cu2O matches Figure 7. Therefore, Cu in Cu/TiO2 prepared in this study exists as Cu + ion in Cu2O. It was reported that Cu + was more active than Cu 2+ [36]. Consequently, it is expected that Cu + plays a role in enhancing the CO2 reduction performance.

CO 2 Reduction Characteristics of Cu/TiO 2
Tables 1 and 2 list the changes in molar quantity of CO per unit weight of photocatalyst for Cu/TiO 2 film coated on netlike glass disc with the time under the condition of Xe lamp illumination with and without UV light, respectively. In these tables, the impact of molar ratio of CO 2 , NH 3 and H 2 O is also evaluated. In addition, fuels other than CO were not detected in this study. Before this experiment, a blank test under the condition of CO 2 /NH 3 /H 2 O or CO 2 /H 2 O without Xe lamp illumination had been carried out as a reference, resulting that no fuel was detected as expected. As to the reproducibility of experiments, this study shows the data from averaging three experiments. Tables 1 and 2 also list the maximum value of molar quantity of CO per unit weight of photocatalyst which is written in bold font.    It can be seen from Table 1 that the CO 2 reduction performance for the molar ratio of CO 2 /NH 3 /H 2 O = 1:1:1 is the highest where the molar quantity of CO per unit weight of photocatalyst is 10.2 µmol/g at 6 h. According to the reaction scheme of CO 2 reduction with H 2 O or NH 3 shown above, the theoretical molar ratio of CO 2 /H 2 O to produce CO or CH 4 is 1:1 or 1:4, respectively. In addition, the theoretical molar ratio of CO 2 /NH 3 to produce CO or CH 4 is 3:2, 3:8, respectively. Based on these theoretical molar ratios, the molar ratio of CO 2 /NH 3 /H 2 O = 3:2:3 should be the theoretical molar ratio to produce CO. However, it is revealed that the molar ratio of CO 2 /NH 3 /H 2 O = 1:1:1, which exhibits the highest performance of CO production as shown in Table 1, is different from the theoretical molar ratio assumed. Since the ionized Cu doped with TiO 2 can provide free electrons to be used for the reduction reaction process [38,39], the theoretically required quantity of the reductant NH 3 and H 2 O is reduced from the values according to the theoretical reaction scheme with TiO 2 i.e., CO 2 /NH 3 /H 2 O = 3:3:3 to 3:2:3. It is also observed from Table 1 that the produced CO decreases after reaching a maximum value. It is believed that the decrease in the produced CO was caused by the reoxidation reaction with CO and O 2 [40], and not caused by the deactivation of the photocatalyst.
As to the impact of NH 3 on CO 2 reduction characteristics using Cu/TiO 2 , the authors' previous study [30] had drawn the following conclusions: Comparing the concentration change of CO along the time under the Xe lamp with UV light for the molar ratio of CO 2 /H 2 O = 1:1 to that for the molar ratios of CO 2 /NH 3 /H 2 O = 1:1:1, 1:0.5:1, 3:2:3, it is observed that the concentration of formed CO for the molar ratio of CO 2 /H 2 O = 1:1 shows the peak soon after the start of illumination of Xe lamp and decreases gradually. It is also observed that the concentration of formed CO for the molar ratios of CO 2 /NH 3 /H 2 O = 1:1:1, 1:0.5:1, 3:2:3 are larger than that for the molar ratio of CO 2 /H 2 O = 1:1. In addition, the decrease of formed CO is small after the concentration of formed CO performs the highest value compared to the molar ratio of CO 2 /H 2 O = 1:1. Therefore, it is revealed that the combination of NH 3 and H 2 O, that is, the existence of NH 3 is effective for the promotion of the CO 2 reduction performance of prepared photocatalyst. On the other hand, comparing the concentration change of CO along the time under the Xe lamp with UV light for the molar ratio of CO 2 /H 2 O = 1:0.5 to that for the molar ratio of CO 2 /NH 3 /H 2 O = 1:1:0.5, 1:0.5:0.5, it is observed that the concentration of formed CO for the molar ratio of CO 2 /H 2 O = 1:0.5 shows the peak soon after the start of illumination of Xe lamp and decreases gradually, which displays the same tendency as the result for the molar ratio of CO 2 /H 2 O = 1:1. It is also observed that the concentration of CO for the molar ratios of CO 2 /NH 3 /H 2 O = 1:1:0.5 and 1:0.5:0.5 are larger than that for the molar ratio of CO 2 /H 2 O = 1:0.5. In addition, the concentration of formed CO keeps some value approximately without rapid decrease before 24 h for CO 2 /NH 3 /H 2 O conditions compared to the molar ratio of CO 2 /H 2 O = 1:0.5. According to the reaction scheme to reduce CO 2 with NH 3 as shown above, the more reaction step is needed to produce CO since NH 3 should be converted into H 2 at first. Consequently, it is believed that the time to produce CO is longer compared to the molar ratio of CO 2 /H 2 O = 1:0.5. Moreover, comparing the concentration change of formed CO along the time under the Xe lamp with UV light for the molar ratio of CO 2 /H 2 O = 3:12 to that for the molar ratio of CO 2 /NH 3 /H 2 O = 3:8:12, it is observed that the concentration of formed CO for the molar ratio of CO 2 /H 2 O = 3:12 shows the peak soon after the illumination of Xe lamp and decreases gradually. In addition, the concentration of formed CO restarts to increase gradually and decrease again. This trend is different from the other CO 2 /H 2 O condition. The ratio of H 2 O is larger in this condition compared to the others, which indicates larger reductants provided for reduction reaction. Therefore, it is thought to keep CO production even though the oxidization reaction with CO and O 2 starts, which is the reason for the decrease in concentration of CO. Furthermore, it is also observed that the concentration of formed CO for the molar ratio of CO 2 /NH 3 /H 2 O = 3:8:12 is larger than that for the molar ratio of CO 2 /H 2 O = 3:12. Consequently, it is revealed that the combination of NH 3 and H 2 O, that is, the existence of NH 3 , is effective for promotion of the CO 2 reduction performance of prepared photocatalyst for all conditions of CO 2 /NH 3 /H 2 O. When comparing the highest quantity produced CO in the case of CO 2 /NH 3 /H 2 O = 3:8:12 to that in the case of CO 2 /H 2 O = 3:12, it is confirmed that the highest produced CO in the case of molar ratio of CO 2 /NH 3 /H 2 O = 3:8:12 is approximately three times as large as that in the case of molar ratio of CO 2 /H 2 O = 3:12. Consequently, it is clear that NH 3 could promote CO 2 reduction performance of Cu/TiO 2 .
It can be seen from Table 2 that the CO 2 reduction performance for the molar ratio of CO 2 /NH 3 /H 2 O = 1:0.5:0.5 is the highest where the molar quantity of CO per unit weight of photocatalyst is 2.5 µmol/g. In addition, Table 2 also reveals that the amount of total reductants required is smaller than that in the case with UV light shown in Table 1. When the Xe lamp is illuminated without UV light, the light intensity and wavelength range of light are smaller and narrower respectively, compared to the condition with UV light as described above. According to the reaction scheme of CO 2 reduction with H 2 O or NH 3 that an electron is produced by the photochemical reaction which is influenced by the light illumination condition. Additionally, H + whose amount is the same as that of electron is needed to produce CO. Since the number produced electrons might be smaller due to the less light input without UV, it is believed that the numbers of required H + are small. Therefore, the CO 2 reduction performance for the molar ratio of CO 2 /NH 3 /H 2 O = 1:0.5:0.5 was the highest, while total reductants required were smaller than that in the case with UV light. Tables 3 and 4 show the trends in molar quantity of CO per unit weight of photocatalyst (Pd/TiO 2 ) under the condition of Xe lamp illumination with and without UV light, respectively. Tables 3 and 4 also list the maximum value of molar quantity of CO per unit weight of photocatalyst which is written in bold font. It is seen from Table 3 that the highest CO 2 reduction performance reached at the illumination time of Xe lamp with UV light of 12 h irrespective of molar ratio of CO 2 /NH 3 /H 2 O. Though it is observed from Table 3 that the molar quantity of CO per unit weight of photocatalyst for the molar ratio of CO 2 /NH 3 /H 2 O = 3:2:3 has the highest value at the illumination time of 12 h, the molar quantity of CO per unit weight of photocatalyst decreased rapidly after 12 h. In these figures, the impact of molar ratio of CO 2 , NH 3 and H 2 O is also presented. Additionally, the other fuels except for CO were not detected in this study. Before the experiment, a blank test without Xe lamp illumination had been carried out as a reference, resulting that no fuel was detected as expected. As to the reproducibility of experiments, this study shows the data from three averaged experiments. According to Table 3, with the UV illumination, the CO 2 reduction performance for the molar ratio of CO 2 /NH 3 /H 2 O = 1:1:1 is the highest where the molar quantity of CO per unit weight of photocatalyst is 5.5 µmol/g. Though this study assumes that the molar ratio of CO 2 /NH 3 /H 2 O = 3:2:3 is the theoretical molar ratios to produce CO, it is revealed that the optimum molar ratio of CO 2 /NH 3 /H 2 O is 1:1:1. Additionally, the molar ratio exhibiting the highest CO 2 reduction performance for Pd/TiO 2 is the same as that for Cu/TiO 2 . As described above, since the ionized Cu doped with TiO 2 can provide free electron to be used for the reduction reaction process [39], the theoretically required quantity of the reductants of NH 3 and H 2 O is reduced from the values according to the theoretical reaction scheme i.e., CO 2 /NH 3 /H 2 O = 3:3:3 to 3:2:3. According to the EELS spectra analysis as shown above, it is believed that Pd in Pd/TiO 2 prepared in this study was in the form of Pd metal, which is the different from Cu + ion in Cu/TiO 2 prepared in this study.
Moreover, it can be seen from Table 4 that, without UV light illumination, the CO 2 reduction performance for the molar ratio of CO 2 /NH 3 /H 2 O = 1:1:1 is the highest where the molar quantity of CO per unit weight of photocatalyst is 3.5 µmol/g. This optimum molar ratio (1:1:1) without UV light illumination is the same as the optimum molar ratio with UV light illumination. However, it is different from the optimum molar ratio obtained with Cu/TiO 2 photocatalyst. Pd acts as an electron-transfer mediator, rapidly transferring the photoexcited electrons from the conduction band of Pd/TiO 2 and the photoexcited electrons are transported to the surface of Pd/TiO 2 [24]. Although Cu can also act an electron-transfer mediator, Pd might conduct the higher performance compared to Cu. As a result, Pd/TiO 2 exhibits better CO 2 reduction performance than Cu/TiO 2 under the visible light illumination condition [24]. The amount of light absorbed by Pd might be enough even the smaller light input under the illumination condition of Xe lamp without UV light. As a result, the required amount of needed H + for CO 2 reduction without UV light is not smaller than the case with UV light, which is different from the case of Cu/TiO 2 without UV light. Consequently, the CO 2 reduction performance with Pd/TiO 2 at the molar ratio of CO 2 /NH 3 /H 2 O = 1:1:1 is the highest for Pd/TiO 2 even the illumination condition of Xe lamp without UV light.

The Quantum Efficiency Evaluation
Quantum efficiency is a well-known criterion used to indicate the photocatalytic activity and efficiency [41]. The quantum efficiency is generally calculated by the following equations [34,42]: where η is the quantum efficiency [%], N input is the number of protons absorbed by photocatalyst [-], N output is the photon number used in photocatalytic reaction [-] Figures 9 and 10 show the comparison of quantum efficiencies among different molar ratios of CO 2 /NH 3 /H 2 O for Cu/TiO 2 under the condition of Xe lamp illumination with and without UV light, respectively. It is revealed from Figures 9 and 10 that the highest quantum efficiency under the condition of Xe lamp illumination with and without UV light is obtained for the molar ratio of CO 2 /NH 3 /H 2 O = 1:1:1 and 1:0.5:0.5, respectively, which agrees with the results shown in Tables 1 and 2. Comparing the quantum efficiencies shown in Figures 9 and 10, the highest quantum efficiency of 1.96 × 10 −4 is obtained when the Xe lamp with UV light is illuminated. If illumination time, t for the molar ratio of CO 2 /NH 3 /H 2 O = 1:1:1 with UV light is 6 h when the highest molar quantity of CO per unit weight of photocatalyst is obtained, the highest quantum efficiency for Cu/TiO 2 is 3.14 × 10 −3 . Figures 11 and 12 show the comparison of quantum efficiencies among different molar ratios of CO 2 /NH 3 /H 2 O for Pd/TiO 2 under the condition of Xe lamp illumination with and without UV light, respectively. Figures 11 and 12 reveal that the highest quantum efficiency under the condition of Xe lamp illumination with and without UV light is obtained when the molar ratio of CO 2 /NH 3 /H 2 O is 1:1:1, which agrees the results shown in Tables 3 and 4. Comparing the quantum efficiencies shown in Figures 11 and 12, the highest quantum efficiency of 4.20×10 −4 is obtained with UV light. If t is set at 96 h which is the same time as the case of Cu/TiO 2 , the highest quantum efficiency for Pd/TiO 2 under the condition of Xe lamp illumination with UV light is 0.53 × 10 −4 .