Titanium dioxide (TiO₂) has been widely used as a model catalyst material in various areas such as photo and thermal catalytic applications for energy and environment [1-6]. To satisfy increasing demand, considerable effort has been devoted to the designing of new catalysts with higher catalytic performance [1,7-9]. Among various strategies, guest transition metal doping/loading of a host metal oxide has been extensively employed [1,7,10-13]. For recycling energy production using CO₂ reduction, Ola and Maroto-Vale loaded Cr, V, and Co onto honeycomb monolithic TiO₂ structures, tested vapor-phase CO₂ photoreduction, and observed CH₄ and acetaldehyde production rates of 4.87 and 11.13 μmol g⁻¹ h⁻¹, respectively, for 0.5 wt% V–TiO₂ monoliths [9]. Low et al. prepared a TiO₂/Ti3C2 composite and tested photocatalytic CO₂ reduction to achieve a CH₄ production rate of 0.22 μmol h⁻¹ under a 300 W Xe arc lamp. Xu et al. reported 1D/2D TiO₂/MoS₂ hybrid nanostructures prepared using the hydrothermal transformation method. They reported that CH₄ and CH₃OH yields reached up to 2.86 and 2.55 μmol g⁻¹ h⁻¹, respectively, for photocatalytic CO₂ reduction over 10 % MoS₂/TiO₂ [12]. Bare TiO₂ fibers were reported to show only CH₃OH as a reduction product with a yield of 0.72 μmol g⁻¹ h⁻¹. Jung et al. reported that when mesoporous TiO₂ and few-layered MoS₂ were decorated on graphene, the CO production yield by CO₂ photocatalytic reduction reached 93.2 μmol g⁻¹ h⁻¹, which is 14.5 times that of bare TiO₂ [13]. Pham and Lee used Cu/V co-doped TiO₂ on polyurethane as a CO₂ photoreduction catalyst and reported CH₄ and CO production rates of 933 and 588 μmol g⁻¹ h⁻¹ under visible-light irradiation (0.05 W/cm²) [8].

For thermal CO oxidation tests [7,14,15], Fang et al. prepared CuO/TiO₂ catalysts using coprecipitation and impregnation methods and reported a CO oxidation onset at approximately 50 °C [14]. The high catalytic activity was attributed to the interfacial bond formation of hybrid Ti–O–Cu and the consequent stabilization of Cu+ species. DeSario et al. prepared low-valent (0, 1+) Cu nanoparticles (2–3 nm) on TiO₂ aerogels and reported CO oxidation onset below 100 °C with a high conversion rate, which was primarily determined by the Cu^0/1+: Cu²⁺ ratio [16].

To acquire new information on the relationship between thermal CO oxidation and photocatalytic CO₂ reduction, we introduced Cu/Ni-bimetallic catalysts loaded onto TiO₂ nanoparticles (NPs) and performed both CO oxidation and CO₂ reduction experiments. The present thermal CO oxidation and CO₂ reduction demonstration tests could be very useful methodologies for developing Ti oxide-based thermal and photocatalysts.

Anatase titanium (IV) oxide nanopowder (99.7 %, < 25 nm particle size, Sigma-Aldrich) was used as received as a support material. For the preparation of the Cu/Ni-loaded TiO₂ NPs, appropriate amounts of Ni²⁺ (nickel (II) nitrate hexahydrate, 98 %, Samchun, Korea) and Cu²⁺ (copper (II) nitrate trihydrate, 99 %, Daejung, Korea) solutions were added to a TiO₂ dispersed solution. The mixed solution was fully dried before characterization. A Hitachi S-4800 scanning electron microscope (SEM) was used to examine the morphology of the prepared samples (1, 2, and 5 mol% Cu/Ni-loaded TiO₂ NPs). A Hitachi H-7600 transmission electron microscope (TEM) at 100 kV was also used to examine the morphology. The crystal phases of the powder samples were examined using a PANalytical X’Pert Pro MPD X-ray diffractometer (XRD) under Cu Kα radiation (40 kV and 40 mA). Using a Thermo-VG Scientific K-alpha⁺ spectrometer with a monochromatic Al Ka X-ray source and a hemispherical energy analyzer, the surface chemical states were examined by conducting a survey of Cu 2p, Ni 2p, Ti 2p, and O 1s X-ray photoelectron spectra. UV-visible absorption and Raman spectra of the powder samples were obtained using a SCINCO NeoSys-2000 UV–Vis spectrophotometer and a Bruker Senterra Raman spectrometer (a laser excitation wavelength of 532 nm), respectively. Thermal CO oxidation experiments in a flow reactor were conducted using two different detectors of a CO₂ sensor and an SRS RGA200 quadrupole mass spectrometer. A powder sample of 20 mg was mounted in a U-type quartz tube. The flow rate of CO (1.0 %)/O₂ (2.5 %)/N₂ was fixed at 40 mL/min. A heating rate of 20 K/min was fixed for the detection of CO₂ (CO oxidation product) using a mass spectrometer. To use the CO₂ sensor, the sample temperature was increased to the desired temperature and stabilized for 30 min before measuring the CO₂ concentration using the CO₂ sensor. Photocatalytic CO₂ reduction experiments were performed in a stainless-steel reactor with a quartz window on top. The powder sample was pasted on a glass slide (3 cm × 3 cm) and placed in the reactor with 40 μL of deionized water. Afterwards, the reactor was flushed and filled with pure CO₂ (99.999 %) gas. The reactor was then placed under four 15 W UVC (200–280 nm) lamps for 12 h. The CO₂ reduction gas products (e.g., CO, CH₄, and CH₃OH) were analyzed using a YL 6500 gas chromatograph (Young In Chromass Co., Ltd.) equipped with a flame ionization detector and a thermal conductivity detector. A gas volume of 0.5 mL was injected into two different columns of 40/60 Carboxen-1000 and HP-PlotQ-PT for GC analysis.

Figure 1 shows the XRD patterns of bare TiO₂ NPs and 5 mol% Cu/Ni-loaded TiO₂ NPs. For the bare TiO₂ powder sample, all the XRD peaks were attributed to the tetragonal crystal phase (JCPDS 1-084-1285). A dominant peak at 2θ = 25.3° was assigned to the (101) plane. The (004), (200), (105), (211), (204), and (220) planes were assigned to the corresponding peaks [17]. For the 5 mol% Cu/Ni-loaded TiO₂ powder sample, new (but weak) XRD peaks clearly appeared at 2θ = 37.3° and 43.3°. These peaks were attributed to the (111) and (200) planes of the cubic crystal phase (Fm-3m) of NiO (JCPDS 1-078-0429) or Ni_xCu_1-xO (JCPDS 1-078-0645). The 2θ peak at approximately 43° was also possibly due to the (111) plane of metallic Cu. This is discussed below.

Figure 1. XRD patterns of bare TiO₂ NPs and 5 mol% Cu/Ni-loaded TiO₂ NPs.

Figure 2 shows the SEM and TEM images of bare TiO₂ NPs and 5 mol% Cu/Ni-loaded TiO₂ NPs. The SEM image of the bare TiO₂ powder shows large cluster aggregates of small particles. The SEM image of the Cu/Ni-loaded (5 mol%) TiO₂ sample appeared to be bigger cluster aggregates of particles of various sizes. TEM images show particle size distribution of approximately 20 nm.

Figure 2. SEM (top) and TEM (bottom) images of the bare TiO₂ and Cu/Ni-loaded (5 mol%) TiO₂ samples.

Figure 3 displays the UV-visible absorption and Raman spectra of the bare TiO₂ and Cu/Ni-loaded TiO₂ powder samples. The absorption edge appeared at approximately 390 nm, corresponding to a band gap of 3.2 eV for TiO₂ [17]. Upon introducing Cu and Ni, the absorption edge increased to a longer wavelength in the visible region. The absorption in the visible region increased with increasing amounts of Cu and Ni. Four major Raman active peaks were found at 144, 394, 514, and 637 cm⁻¹, and became weaker and broader as the number of metal ions increased. The former two peaks (144 and 394 cm⁻¹) were attributed to the E_g and B_1g O–Ti–O bending modes, respectively, and shifted to higher wavenumbers with increasing guest metal ions. The latter two were attributed to A_1g and E_g Ti–O stretching modes, respectively, and shifted to lower wavenumbers with increasing metal ions [10,15]. This indicates that the bending and stretching modes are oppositely affected by the loaded metal ions.

Figure 3. Reflectance UV-Vis absorption (left) and Raman spectra (right) of bare TiO₂ and Cu/Ni-loaded (1,2, and 5 mol%) TiO₂ samples.

Figure 4 displays CO oxidation profiles along with the reaction temperature obtained using two different methods: 1) CO₂ gas sensor detection and 2) temperature-programmed reaction mass spectrometry [7,11,18]. For the CO₂ gas sensor detection method, the sample temperature was increased to the desired temperature and stabilized for 30 min. Subsequently, the CO₂ concentration reading on the sensor was recorded. For the first run of 1 and 2 mol% Cu/Ni-loaded TiO₂ samples, the CO oxidation started to appear at approximately 300 °C. The 5 mol% loaded sample showed CO oxidation onset just above 200 °C. For the second run, all the samples showed CO oxidation onset at approximately 200 °C. However, the slope of the CO₂ concentration change was different and showed the order of 2 mol% << 1 mol% < 5 mol%. The concentration for the 5 mol% sample exhibited the sharpest changes at different temperatures. For bare TiO₂ NPs, no CO oxidation activity was observed below 500 °C, which is in good agreement with the literature [14]. For the CO oxidation profiles obtained using temperature-programmed reaction spectrometry, the sample heating rate was fixed at 20 K/min. For the first run, CO oxidation onsets were observed at 240, 300, and 340 °C for 5, 2, and 1 mol% samples, respectively. The 5 mol% sample showed the highest CO oxidation activity. For the second run, the differences in the onset temperatures became smaller. CO oxidation onsets were observed at 200, 240, and 250 °C for 5, 2, and 1 mol% samples, respectively. The CO oxidation activity increased after the first run, possibly due to an increase in crystallinity and metal alloy formation [7]. Based on the CO oxidation profiles using two different methods, it was concluded that the heating rate and stabilization were important factors affecting the catalytic activity.

Figure 4. 1st (left panel) and 2nd run (right panel) CO oxidation reaction profiles measured using a sensor (top) and a mass spectrometry (bottom).

Photocatalytic CO₂ reduction experiments (Fig. 5) were performed, and three different reduction products of carbon monoxide (CO), methane (CH₄), and methanol (CH₃OH) were obtained. The total CO production yields after UVC irradiation for 13 h were observed to be 0.4, 14.7, 10.2, and 9.1 μmol for bare TiO₂ NPs, CuNi (1 mol%)/TiO₂, CuNi (2 mol%)/TiO₂, and CuNi (5 mol%)/TiO₂, respectively. Compared with that of bare TiO₂ NPs, the CO yield saw an increase by a factor of 35. The total CH₄ production yields for 13 h were observed to be 0.6, 3.2, 2.0, and 1.9 μmol for bare TiO₂ NPs, CuNi (1 mol%)/TiO₂, CuNi (2 mol%)/TiO₂, and CuNi (5 mol%)/TiO₂, respectively. The total CH₃OH production yields were observed to be 1.0, 6.0, 4.9, and 5.1 μ mol for bare TiO₂ NPs, CuNi (1 mol%)/TiO₂, CuNi (2 mol%)/TiO₂, and CuNi (5 mol%)/TiO₂, respectively. The total reduction yields showed the order of CH₄ < CH₃OH < CO. In addition, the photocatalytic CO₂ reduction activity showed the order of bare TiO₂ NPs << CuNi (5 mol%)/TiO₂ < CuNi (2 mol%)/TiO₂ < CuNi (1 mol%)/TiO₂. Overall, the CuNi (1 mol%)/TiO₂ sample showed the highest CO₂ reduction activity and the CO, CH₄, and CH₃OH production rates were observed to be 1.13, 0.25, and 0.46 μmol h^-1, respectively. These values were higher than those reported in the literature [19]. For the mechanism of CO₂ reduction, photogenerated electrons (e⁻) and holes (h⁺) are created, which participate in the formation of H⁺, consequently forming CO, CH₄, and CH₃OH, as described below.

Figure 5. Total CO₂ reduction (CO, CH₄, and CH₃OH) yields (μmol) for of bare TiO₂ and Cu/Ni-loaded (1,2, and 5 mol%) TiO₂ samples.

A survey XPS scan (not shown here) was performed to examine the surface chemical species. Two strong Ti 2p and O 1s peaks were observed because of the host TiO₂ matrix. Weaker Cu 2p, Ni 2p, and C 1s peaks were observed and were attributed to guest metal (Cu and Ni) ions and surface impurities, respectively. Figure 6 displays high-resolution Cu 2p, Ni 2p, Ti 2p, and O 1s XPS spectra. For the Cu 2p XPS peaks, the dominant Cu 2p_3/2 and Cu 2p_1/2 peaks were found at binding energies (BEs) of 932.7 and 952.4 eV, with a spin-orbit splitting energy of 19.7 eV. These peaks can be assigned to metallic Cu and/or Cu₂O [20]. A negligible satellite peak at approximately 942.0 eV indicated that Cu was not in the +2 oxidation state [15]. To distinguish Cu from Cu₂O, Cu LMM Auger lines were examined in the XPS survey. Two broad Cu LMM Auger peaks were observed at BEs of 570 (minor) and 565 eV (major). This indicates that Cu is mainly metallic with a minor Cu₂O state. For Ni 2p XPS peaks, two Ni 2p_1/2 and Ni 2p_3/2 peaks were observed at BEs of 856.1 and 873.8 eV, respectively, with a spin-orbit splitting energy of 17.7 eV. These peaks could be attributed to the surface NiO and/or Ni(OH)₂ species. Broad and strong satellite peaks at 862 eV indicate the +2 oxidation state of Ni. For the Ti 2p XPS peaks, two Ti 2p_3/2 and Ti 2p_1/2 XPS peaks were observed at 459.0 and 464.6 eV, with a spin-orbit splitting of 5.6 eV. This is attributed to the 4+ oxidation state of TiO₂ [12,15]. Two O 1s XPS peaks were observed at 530.2 and 532.0 eV, mainly due to lattice oxygen (O^2-) of TiO₂ and other surface oxygen species (e.g., OH and H₂O) species, respectively [12,15].

Figure 6. Cu 2p, Ni 2p, Ti 2p, and O 1s XPS spectra for 5 mol% CuNi-loaded TiO₂ sample.

Cu and Ni co-loaded TiO₂ NPs were prepared and tested for both thermal CO oxidation and photocatalytic CO₂ reduction performance. The CuNi (1 mol%)/TiO₂ sample showed the highest CO₂ reduction activity, and the total CO, CH₄, and CH₃OH production yields for UVC irradiation after 13 h were 14.7, 3.2, and 6.1 μmol, respectively. Bare TiO₂ NPs exhibited CO₂ reduction yields below 1 μ mol. For thermal CO oxidation, the CuNi (5 mol%)/TiO₂ sample showed the highest CO oxidation activity with a CO oxidation onset at 200 °C. This demonstration test could provide guidance for developing thermal CO oxidation and photocatalytic CO₂ reduction catalysts.

This research was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MEST) (2016R1D1A3B04930123).