Thermal CO Oxidation and Photocatalytic CO₂ Reduction over Bare and M-Al₂O₃ (M = Co, Ni, Cu, Rh, Pd, Ag, Ir, Pt, and Au) Cotton-Like Nanosheets

Abstract

Aluminum oxide (Al₂O₃) has abundantly been used as a catalyst, and its catalytic activity has been tailored by loading transition metals. Herein, γ-Al₂O₃ nanosheets were prepared by the solvothermal method, and transition metals (M = Co, Ni, Cu, Rh, Pd, Ag, Ir, Pt, and Au) were loaded onto the nanosheets. Big data sets of thermal CO oxidation and photocatalytic CO₂ reduction activities were fully examined for the transition metal-loaded Al₂O₃ nanosheets. Their physicochemical properties were examined by scanning electron microscopy, high-resolution transmission electron microscopy, X-ray diffraction crystallography, and X-ray photoelectron spectroscopy. It was found that Rh, Pd, Ir, and Pt-loading showed a great enhancement in CO oxidation activity while other metals negated the activity of bare Al₂O₃ nanosheets. Rh-Al₂O₃ showed the lowest CO oxidation onset temperature of 172 °C, 201 °C lower than that of bare γ-Al₂O₃. CO₂ reduction experiments were also performed to show that CO, CH₃OH, and CH₄ were common products. Ag-Al₂O₃ nanosheets showed the highest performances with yields of 237.3 ppm for CO, 36.3 ppm for CH₃OH, and 30.9 ppm for CH₄, 2.2×, 1.2×, and 1.6× enhancements, respectively, compared with those for bare Al₂O₃. Hydrogen production was found to be maximized to 20.7 ppm during CO₂ reduction for Rh-loaded Al₂O₃. The present unique pre-screening test results provided very useful information for the selection of transition metals on Al₂O₃-based energy and environmental catalysts.

1. Introduction

Aluminum oxide (Al₂O₃) has extensively been used as a heterogeneous catalyst in diverse catalytic reactions of CO oxidation [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19], CO₂ reduction [20,21,22,23], CO₂ methanation/hydrogenation [20,24,25,26,27], and preferential oxidation of CO [28,29]. The efforts to increase the catalytic activity of the metal oxide have been devoted to the modification of the metal surface by loading of transition metals in groups of 9 (Co, Rh, and Ir), 10 (Ni, Pd, and Cu), and 11 (Cu, Ag, and Au). The morphology of a metal oxide support has also been a key factor for the enhancement of the catalytic activity [30]. It was reported that the role of an overlayer metal becomes different when the support metal oxide is different [2]. The relative catalytic activities of overlayer metals also become different when the catalytic application areas are different for the application to CO oxidation using transition metal-loaded Al₂O₃ catalysts.

Chen et al. prepared Pt nanoparticles (NPs) on Al₂O₃ and observed 100% CO conversion at −20 °C [1]. For the extremely high activity compared with a commercial Pt/Al₂O₃, based on the experimental and the density functional theory (DFT) calculations, they proposed that CO was initially adsorbed on Pt(OH) kink sites and reacted with OH to release gaseous CO₂. Afterward, OH was regenerated by activation of O₂ on terrace sites. Lou and Liu studied CO oxidation of single Pt atoms dispersed on Fe₂O₃ (highly reducible), ZnO (reducible), and γ-Al₂O₃ (irreducible) supports, and observed that the catalytic activity was in the order of Pt/γ-Al₂O₃ < Pt/ZnO < Pt/Fe₂O₃ [2], where the highly reducible support showed the highest catalytic activity. Chen et al. tested Pt/Al₂O₃ for preferential oxidation (PROX) of CO in H₂ [29]. They concluded that CO conversion and CO₂ selectivity reached up to 100% in a wide range of −30 °C to 120 °C. The high performance was attributed to a combination of Pt(OH) and metallic Pt on the Al₂O₃ support. Therefore, the adsorption of CO and the activation of O₂ were optimally tuned to maximize the performance. For monodispersed single Pt atoms on θ-Al₂O₃, Moses-DeBusk et al. found that the CO oxidation did not follow a conventional Langmuir-Hinshelwood mechanism [11]. The Pt atom was first oxygenated, and then CO was bound to form a carbonate (CO₃), which dissociated to generate gaseous CO₂ [11]. Yang et al. employed the DFT calculation to investigate the relative CO oxidation for single-atom catalysts of Ni/γ-Al₂O₃ and Pd/γ-Al₂O₃ [7]. They reported that Ni showed an unexpectedly higher CO oxidation activity than the Pd. Ananth et al. synthesized Ag₂O/γ-Al₂O₃ and (Ag₂O + RuO₂)/γ-Al₂O₃ catalysts and tested the CO oxidation performances to show that the catalytic activity was increased by the addition of RuO₂ [6]. Han et al. reported a high CO oxidation activity at 30 °C for NiO (≤1 nm) on mesoporous Al₂O₃ prepared using atomic layer deposition [16]. The deactivation was found to be lowered with increasing the pre-annealing temperature.

For the application of Al₂O₃ to CO₂ reduction, Zhao et al. synthesized Au/Al₂O₃/TiO₂ nanocomposites, where the atomic-layer Al₂O₃ was sandwiched between the two layers [21]. They tested the photocatalytic CO₂ reduction activity and observed CO (major) and CH₄ (minor) as products. It was concluded that the charge transfer and surface charge recombination were highly influenced by Al₂O₃ interlayer thickness. Therefore, the maximum photocatalytic activity (37 μmol/g of CO and 2 μmol/g of CH₄) was obtained by achieving optimum Al₂O₃ thickness (5 Å). Kwak et al. performed a temperature-programmed CO₂ reduction with H₂ on Ru/Al₂O₃ catalysts and observed CO and CH₄ formation yields with activation energies of 82 kJ/mol and 62 kJ/mol, respectively [20]. It was found that CO formation selectivity was increased with increasing Ru metal dispersion but decreased with increasing Ru clustering and concluded that CO was not an intermediate species for CH₄ formation. Chein and Wang tested CO₂ methanation activities using Ni/Al₂O₃, Ru/Al₂O₃, and Ru-Ni/Al₂O₃ catalysts [27] and found that the hybrid bimetallic Ru-Ni showed higher performance than the monometallic catalysts.

Although numerous detailed in-depth studies have been performed using transition metal-loaded Al₂O₃ catalysts, no systematic comparison studies have been reported among diverse (M = Co, Ni, Cu, Rh, Pd, Ag, Ir, Pt, and Au) transition metal-loaded Al₂O₃ catalysts prepared by the same synthesis method. Motivated by this, we synthesized transition metal-loaded Al₂O₃ nanosheets and evaluated thermal CO oxidation activity as well as photocatalytic CO₂ reduction activity. Consequently, the roles of overlayer transition metals were comparatively investigated in two totally different application reactions. Thereby, the present pre-screening test results provided useful information on the quick-selection of catalysts for thermal CO oxidation and photocatalytic CO₂ reduction.

2. Materials and Methods

2.1. Catalysts Synthesis Procedures

For the synthesis of Al precursor nanosheets, 1 mmol of aluminum nitrate nonahydrate (Al(NO₃)₃ 9H₂O, 98%, Sigma-Aldrich, St. Louis, MO, USA), 0.008 g of polyethylene glycol (PEG, M_n = 4000, Sigma-Aldrich, St. Louis, MO, USA), and 20 μL of oleic acid (≥99%, Sigma-Aldrich, St. Louis, MO, USA) were fully dissolved by magnetic stirring in a mixed solvent of 10 mL of deionized water and 15 mL of ethanol (99.9%, Samchun Chem., Gyounggi, Korea) for 20 min. After that, the solution was transferred into a Teflon-lined stainless-steel autoclave reactor, which was then tightly capped for sealing. The tightly capped reactor was placed in an oven setting at 200 °C for 12 h. After the thermal reaction, the reactor was naturally cooled to room temperature, and the finally obtained white precipitates were collected by washing with deionized water and ethanol repeatedly by centrifugation at 3600 rpm. The collected wet powder was fully dried in an oven setting at 80 °C for 24 h. To obtain Al₂O₃ nanosheets, the dried powder sample was thermally annealed at 600 °C for 2 h.

For transition metal (M = Co, Ni, Cu, Rh, Pd, Ag, Ir, Pt, and Au) loadings, 50 mg of Al₂O₃ nanosheets were fully dispersed in 20 mL ethanol, followed by adding 2 mol% of metal ions. The chemicals for metal ions were cobalt(II) nitrate hexahydrate (≥98%, Sigma-Aldrich, St. Louis, MO, USA), nickel(II) nitrate hexahydrate (≥98.5%, Sigma-Aldrich, St. Louis, MO, USA), copper(II) nitrate trihydrate (99%, Daejung, Gyounggi, Korea), rhodium(III) chloride hydrate (≥99.9%, Sigma-Aldrich, St. Louis, MO, USA), palladium(II) chloride (99.8%, Sigma-Aldrich, St. Louis, MO, USA), silver nitrate solution (0.1 N, Samchun Pure Chem., Gyounggi, Korea), platinum (III) chloride (≥99.9%, Sigma-Aldrich, St. Louis, MO, USA), iridium (III) chloride hydrate (≥99.9%, Sigma-Aldrich, St. Louis, MO, USA), and gold(III) chloride trihydrate (≥99.9%, Sigma-Aldrich, St. Louis, MO, USA). After complete mixing, the solvent was slowly evaporated by gentle heating (50 °C) while stirring. The dried M-loaded Al₂O₃ nanosheets were, again, thermally annealed at 600 °C for 2 h.

2.2. Sample Characterization

The surface morphologies of the Al-precursor, Al₂O₃ nanosheets, and M-loaded Al₂O₃ nanosheets were examined using a scanning electron microscope (SEM, Hitachi S-4800, Hitach Ltd., Tokyo, Japan) at conditions of 10 kV and 10 mA. X-ray crystallographic diffraction patterns were recorded using a PANalytical X’Pert Pro MPD diffractometer (PANalytical, Almelo, Netherland) with Cu Kα radiation (40 kV and 30 mA). Transmission electron microscopic (TEM) and high-resolution TEM (HRTEM) were obtained for bare Al₂O₃ nanosheets and the selected Ni- and Rh-loaded Al₂O₃ nanosheets using an FEI Tecnai G2 F20 TEM (Hillsboro, OR, USA) operated at 300.0 kV. X-ray photoelectron spectra were taken using a Thermo-VG Scientific K-alpha⁺ spectrometer (Thermo VG Scientific, Waltham, MA, USA) with a hemispherical energy analyzer. Attenuated total reflection Fourier-transform infrared spectroscopy (FT-IR) was employed using a Nicolet iS 10 FT-IR spectrometer (Thermo Scientific Korea, Seoul, Korea). The Brunauer-Emmett-Teller (BET) surface areas were measured using a ChemBET TPR/TPD analyzer (Quantachrome Instruments Corp., Boynton Beach, FL, USA) equipped with a thermal conductivity detector.

2.3. Thermal CO Oxidation and Photocatalytic CO₂ Reduction Experiments

For thermal CO oxidation reactions, 20 mg of a catalyst was initially loaded into a U-shape quartz tube. After that, the tube was positioned in a temperature-programmed furnace. The temperature heating rate was 20 °C/min, and the flowing gas was CO(1.0%)/O₂(2.5%)/N₂ at a flow rate of 40 mL/min. The gas products from the outlet of the tube were monitored using a quadrupole mass spectrometer (RGA200, Stanford Research Systems, Sunnyvale, CA, USA). After the first run to a maximum temperature of 500 °C, the sample cell was naturally cooled to a room temperature of 25 °C. After that, the second run was performed at room temperature.

For photocatalytic CO₂ reduction experiments, 3 mg of a catalyst was fully dispersed on a quartz disc (an area of 15.9 cm²) and placed in a stainless-steel reactor (volume ~40 mL) with additional deionized water (20 μL) beside the disc. After that, the reactor was tightly closed with a quartz window (0.3 cm thick and 4.5 cm diameter) on top. Afterward, pure (99.999%) CO₂ gas was fully flushed and filled with the gas. For the photocatalytic CO₂ reduction test, the reactor with the quartz window was placed under UVC (200–280 nm) lamps (a power density of 5.94 mW/cm²) for 12 h. After the UVC irradiation time, 0.5 mL of gas was taken and injected into a YL 6500 gas chromatograph (GC, Young In Chromass Co., Ltd., Seoul, Korea). For the analysis of the CO, CH₃OH, CH₄, and H₂ products, the GC system was equipped with HP-Plot Q-PT column (Agilent Technologies, Inc., Santa Clara, CA, USA), 40/60 Carboxen-1000 column (Sigma-Aldrich, St. Louis, MO, USA), a Ni catalyst methanizer assembly, a thermal conductivity detector, and a flame ionization detector.

3. Results and Discussion

Figure 1(a,a1) show the SEM image of the as-prepared Al-precursor with a morphology of cotton-like nanostructures. Figure 1b shows the sample after thermal annealing at 600 °C, abbreviated as bare Al₂O₃. It appears that the morphology showed no significant change, but the nanosheets became somewhat compacted. The Brunauer-Emmett-Teller (BET) surface area of bare Al₂O₃ was measured to be 154.4 m²/g. The corresponding transmission electron microscope (TEM) image clearly showed the morphology of nanosheets. For the high-resolution TEM image of bare Al₂O₃, clear lattice fringes were seen, and the lattice spacing was estimated to be 0.197 nm. This was well-matched to the (002) crystal plane of cubic phase gamma-Al₂O₃. This was further discussed in detail below. The structure projection of the (002) and (022) planes for Al₂O₃ are shown in Figure 1(b1) for visual understanding.

Figure 2 shows the SEM and TEM images of selected Ni- and Rh-loaded Al₂O₃ nanosheets. The SEM images of other M-loaded Al₂O₃ nanosheets are provided in the Supporting Information, Figure S1. The SEM image (Figure 2a) of Rh-Al₂O₃ showed small nanoparticles embedded on the nanosheets. The nanoparticles appeared as a result of Rh particle formation. The color of burlywood was clearly different from the white color for bare Al₂O₃. On the other hand, the SEM image (Figure 2b) of Ni-Al₂O₃ nanosheets showed only cotton-like nanosheets. It was difficult to discriminate Ni species from the bare Al₂O₃ support. However, the color clearly changed from white to pale blue upon Ni-loading. The photos and optical microscope images of the M-loaded Al₂O₃ nanosheets are provided in the Supporting Information, Figures S2 and S3, respectively. Although the SEM images showed no clear metal embedment, the color change was a clear indication of metal-loading on the Al₂O₃ support. The metal-loading was also confirmed by the X-ray photoelectron spectroscopy (XPS) data, discussed below.

The TEM images (Figure 2(a1)) of Rh-Al₂O₃ clearly showed NPs (with a size of ~20 nm) embedded onto the nanosheets. For the HRTEM image (Figure 2(b2)) of an Rh-NP, clear lattice fringes were observed, and the distances were estimated to be 0.263 nm and 0.254 nm. These distances matched well with the (114) and (200) crystal planes of orthorhombic (Pbca) Rh₂O₃ (ICSD ref. 98-000-9206), respectively. This indicated that Rh was embedded not in the metallic form but rather in the oxide form. This was further confirmed by the XPS data below. The fast Fourier transform (FFT) pattern of the HRTEM image reflected the crystallinity of the Rh oxide. The TEM images (Figure 2(b1)) of Ni-Al₂O₃ nanosheets showed only nanosheet morphology, consistent with the corresponding SEM image (Figure 2b).

For the HRTEM image in Figure 2(b2), the lattice fringes with distances of 0.227 nm and 0.196 nm matched well with the (111) and (002) crystal planes of the cubic phase γ-Al₂O₃. The lattices showed poor crystallinity compared with those of bare Al₂O₃, seen in Figure 1(b3). Interestingly, some areas (dotted circles) showed very poor crystallinity, and these appeared like amorphous particles. This is likely an indication of Ni embedment on the Al₂O₃ nanosheets. Very similarly for Co-Al₂O₃ nanosheets, although particles were not clearly seen in the TEM image (Supporting Information, Figure S4), the corresponding HRTEM image showed the areas with very poor crystallinity. The areas appeared like Co-embedment in the Al₂O₃ support.

The BET surface areas of Ni-Al₂O₃ and Rh-Al₂O₃ nanosheets were measured to be 151.2 m²/g and 153.8 m²/g, respectively. The surface areas were very similar to that of bare Al₂O₃. This indicated that the surface area was not significantly impacted by the metal-loading.

Figure 3 displays the X-ray diffraction patterns of bare Al₂O₃ and M-loaded Al₂O₃ nanosheets. For the XRD patterns of bare Al₂O₃, two distinctive peaks were observed at 2θ = 45.9° and 67.0°. These two peaks could be assigned to the (002) and (022) planes of cubic phase (Fm-3m) γ-Al₂O₃, (ICSD ref. 98-003-0267), respectively. The XRD result was in good consistency with the HRTEM result of the bare Al₂O₃ nanosheet. For the XRD profiles of M-loaded Al₂O₃ nanosheets, two peaks were commonly observed, as expected. Interestingly, Co, Ni, Cu, Rh, and Ag-loaded samples showed no significant extra peaks in the corresponding XRD profiles. These results indicated that the metals were loaded with an amorphous oxide state (discussed below in XPS) or embedded very uniformly without forming good crystal phases. In addition, because the metal amount was only 2 mol%, the XRD patterns could not be clearly observed when the phase was an amorphous oxide form. As seen in the HRTEM images of Ni-Al₂O₃ and Co-Al₂O₃ nanosheets discussed above (Figure 2(b2) and Figure S4, respectively), the particle-like areas showed very poor crystallinity. On the other hand, Pd, Ir, Pt, and Au-loaded samples showed new peaks in the corresponding XRD profiles.

For the XRD patterns of Pd-Al₂O₃ nanosheets, several peaks at 2θ = 33.8°, 42.0°, 54.7°, 60.1°, and 71.5° showed good matches with the (011), (110), (112), (013), and (121) crystal planes of tetragonal (p 42/mmc) PdO (ICSD ref. 98-002-9281), respectively [13]. For Ir-Al₂O₃ nanosheets, several strong XRD peaks were observed at 2θ = 27.9°, 34.6°, 39.9°, 53.9°, 57.9°, 58.3°, 66.0°, 69.0°, and 73.0°, with good matches with the (110), (011), (020), (121), (220), (002), (130), (112), and (031) crystal planes of tetragonal (p 42/mnm) IrO₂ (ICSD ref. 98-008-4577), respectively. For Pt-Al₂O₃ nanosheets, three major peaks were observed at 2θ = 39.8°, 46.2°, and 67.5°, assigned to the (111), (002,) and (022) crystal planes of the cubic (Fm-3m) crystal phase of metallic Pt (ICSD ref. 98-007-6153), respectively. For Au-Al₂O₃ nanosheets, three strong peaks were observed at 2θ = 38.1°, 44.3°, and 64.5°, assigned to the (111), (002), and (022) crystal planes of the cubic (Fm-3m) crystal phase of Au (ICSD ref. 98-061-1624), respectively.

XPS was employed to confirm the loading of the transition metals and examine the oxidation states. Figure 4 shows Co 2p, Ni 2p, Cu 2p, Rh 3d, Pd 3d, Ag 3d, Ir 4d, Pt 4d, and Au 4d of Co-Al₂O₃, Ni-Al₂O₃, Cu-Al₂O₃, Rh-Al₂O₃, Pd-Al₂O₃, Ag-Al₂O₃, Ir-Al₂O₃, Pt-Al₂O₃, and Au-Al₂O₃ nanosheets, respectively. XPS valence band spectra (Figure 4, right panel) are also displayed for the corresponding samples. The survey, Al 2p, O 1s, and C 1s profiles are provided in the Supporting Information, Figure S5. All the binding energies (BEs) were referenced to the C 1s XPS peak at 284.8 eV. The survey spectra commonly showed the elements of Al, O, and C (surface impurities), as expected. The XPS peaks of the loaded transition metals were very weakly observed.

For Co-Al₂O₃ nanosheets, Co 2p_1/2 and Co 2p_3/2 XPS peaks were observed at binding energies (BEs) of 797.4 eV and 781.6 eV, respectively, with a spin-orbit splitting of 15.8 eV. This could be attributed to Co²⁺ of CoO and Co(OH)₂ [31,32]. The corresponding satellite peaks for Co²⁺ were clearly observed around 803 eV and 786 eV. For Ni-Al₂O₃ nanosheets, Ni 2p_1/2 and Ni 2p_3/2 XPS peaks were observed at binding energies (BEs) of 873.4 eV and 856.1 eV, respectively, with a spin-orbit splitting of 17.3 eV. This could be attributed to Ni²⁺ of NiO and Ni(OH)₂ [15,31,32]. The corresponding satellite peaks for Ni²⁺ were clearly observed around 880 eV and 862 eV. For Cu-Al₂O₃ nanosheets, Cu 2p_1/2 and Cu 2p_3/2 XPS peaks were observed at binding energies (BEs) of 952.4 eV and 932.7 eV, respectively, with a spin-orbit splitting of 19.7 eV. This could be attributed to Cu²⁺ of CuO and Cu(OH)₂ [32,33]. The corresponding satellite peak for Cu²⁺ was clearly observed around 942 eV. For Co, Ni, and Cu, no metallic XPS peaks were observed. On the basis of XRD, HRTEM, and XPS data, Co, Ni, and Cu appeared to be embedded as an amorphous oxide form.

For Rh-Al₂O₃ nanosheets, Rh 3d_3/2 and Rh 3d_5/2 XPS peaks were observed at BEs of 314.3 eV and 309.7 eV, respectively, with a spin-orbit splitting of 4.6 eV. The XPS BEs were attributed to an oxidation state of Rh³⁺ [32,34]. As discussed above, the lattice distances in the HRTEM image confirmed orthorhombic Rh₂O₃. An additional weak shoulder peak was seen around 308 eV for Rh 3d_5/2 peak. This could be due to metallic Rh [32,34]. On the basis of the XPS and HRTEM data, Rh-species appeared to be consistent with Rh@Rh₂O₃ core-shell type structure.

For Pd-Al₂O₃ nanosheets, Pd 3d_3/2 and Pd 3d_5/2 XPS peaks were observed at BEs of 341.7 eV and 336.2 eV, respectively, with a spin-orbit splitting of 5.5 eV. The XPS peaks were attributed to an oxidation state of Pd²⁺ [32,34,35]. There was a good coincidence between the oxidation state of the XPS and the XRD profiles of tetragonal PdO. A weak shoulder of the Pd 3d_5/2 peak was seen around 335.5 eV, plausibly due to metallic Pd [32]. For Ag-Al₂O₃ nanosheets, Ag 3d_3/2 and Ag 3d_5/2 XPS peaks were observed at BEs of 374.5 eV and 368.6 eV, respectively. This was attributed to metallic Ag [6,32,36]. The shoulder XPS peak at 367.6 eV for Ag 3d_5/2 was plausibly due to AgO [36]. On the basis of the XPS profile for each M-Al₂O₃ sample, it could be concluded that the transition metal was loaded on the Al₂O₃ support.

For Ir-Al₂O₃ nanosheets, Ir 4d_3/2 and Ir 4d_5/2 XPS peaks were observed at BEs of 313.6 eV and 297.7 eV, respectively. These peaks were assigned to the Ir⁴⁺ oxidation state [37], in good coincidence with the XRD profiles of tetragonal IrO₂, shown above. A weak shoulder of Ir 4d_5/2 peak was seen around 295 eV, plausibly due to metallic Ir [37]. For Pt-Al₂O₃ nanosheets, Pt 4d_3/2 and Pt 4d_5/2 XPS peaks were observed at BEs of 332.9 eV and 314.9 eV, respectively, with a spin-orbit splitting of 18.0 eV. The XPS peaks were attributed to metallic Pt [34,35], which was well-fitting with the XRD result of metallic Pt. For Au-Al₂O₃ nanosheets, the Au 4f_7/2 and Au 4f_5/2 XPS peaks were observed at BEs of 87.3 eV and 83.6 eV, respectively, with a spin-orbit splitting of 3.7 eV. The XPS peaks were attributed metallic Au [32]. This result was in good agreement with the XRD profiles of metallic Au, shown above.

For the Al 2p XPS profiles (Supporting Information, Figure S5), a broad peak was commonly observed around 74.1 eV, attributed to Al of the Al₂O₃ support [32,38]. An additional peak at 75.0 eV was observed and attributed to the Al of surface Al-OH species [32,38]. For the O 1s XPS profiles (Supporting Information, Figure S5), a broad peak was commonly observed around 530.9 eV due to lattice O of Al₂O₃ support. A broad shoulder at 532.5 eV was attributed to oxygen defects and surface OH/H₂O species [39].

The valence band (VB) spectra are shown in Figure 4 to further examine electronic structures. For the VB of bare Al₂O₃ nanosheets, two broad features were seen around 9 eV and 6 eV, attributed bonding 2pσ (mixed with Al 3s, Al 3p, and Al 3d) and antibonding 2pπ of the oxygen [40]. For VB spectra of M-Al₂O₃ nanosheets, the density of states (DOS) was observed to be closer to the Fermi level. Especially, Rh, Pd, Ir, and Pt showed more clearly new features near 2 eV below the Fermi level, attributed to the Rh 4d, Pd 4d, Ir 5d, and Pt 5d, respectively. This could be related to the higher CO oxidation activities for these metals, discussed below. However, the DOS profiles showed no explicit relationship with the photocatalytic CO₂ reduction activity. The detailed roles of the overlayer elements could be understood with the aid of density functional theory.

Temperature-programmed CO oxidation profiles (Supporting Information, Figure S6) were obtained to examine thermal CO oxidation catalytic activities for bare Al₂O₃ and M-Al₂O₃ nanosheets. To evaluate the catalytic activities of the catalysts, Figure 5a,b display the CO oxidation onset temperatures for the first and the second runs, respectively.

Table 1. CO oxidation onset temperatures (T_{M-Al2O3,onset}) in the first and second runs. Differences in CO oxidation onset temperatures (T_M-Al2O3,2nd − T_M-Al2O3,1st) between the first and second runs.

Group #9	First Run	Second Run	Diff.	Group #10	First Run	Second Run	Diff.	Group #11	First Run	Second Run	Diff.
Co	390	385	−5	Ni	490	480	−10	Cu	360	389	29
Rh	135	172	37	Pd	201	207	6	Ag	340	397	57
Ir	195	217	22	Pt	197	206	9	Au	375	415	40

summarizes the onset temperatures (T_{M-Al2O3,onset}) and the temperature difference (T_M-Al2O3,2nd − T_M-Al2O3,1st) between the first and the second runs. The onset temperatures of Ir-, Pt-, Pd-, and Rh-loaded Al₂O₃ nanosheets were observed to be much lower than those of Au-, Ag-, Cu-, Co-, and Ni-loped Al₂O₃ nanosheets. The group 11 (Au, Ag, and Cu) and the period 4 (Co, Ni, and Cu) elements showed much poor catalytic activity on the Al₂O₃ support. Additionally, the onset temperatures of Au and Ag-loaded Al₂O₃ nanosheets were unexpectedly even higher than expected [6,10,15]. In other words, the Au- and Ag-loaded Al₂O₃ nanosheets showed poorer CO oxidation activity. In the first run, the Rh-Al₂O₃ nanosheets showed the lowest onset of 135 °C, while the Ni-Al₂O₃ nanosheets showed the highest onset of 490 °C. The temperature difference between the two samples was estimated to be 335 °C. In the second run, the Rh-Al₂O₃ nanosheets also showed the lowest onset of 172 °C while the Ni-Al₂O₃ nanosheets showed the highest onset of 480 °C. The temperature difference was estimated to be 308 °C. Pd, Ir, and Pt showed the CO oxidation onsets at 207 °C, 217 °C, and 216 °C, respectively, in the second run. For highly dispersed (or single atom state) 0.2 wt % Pt on mesoporous Al₂O₃ support, Zhang et al. reported CO oxidation onset at ~200 °C, which was in good coincidence with the present result [5]. These results clearly indicated that the CO oxidation activity was highly influenced by the nature of overlayer metal species.

Figure 5c shows the CO oxidation profiles for the first and the second runs of the selected samples (bare Al₂O₃, Ni-Al₂O₃, and Rh-Al₂O₃ catalysts). As seen in the Figure 5, the CO oxidation onset of Rh-Al₂O₃ occurred much earlier than that of bare Al₂O₃. The onsets of Rh-Al₂O₃ in the first and the second runs were observed to be 251 °C and 201 °C lower than those of bare Al₂O₃, respectively. However, the onset temperatures became much higher upon loading Ni.

To examine the difference in catalytic activity between the first and the second runs, Figure 5d plots the temperature differences (T_M-Al2O3,2nd − T_M-Al2O3,1st) in the CO oxidation onsets between the first and the second runs. In the first run, the CO oxidation reactions were performed with the as-prepared samples. In the second run, the CO oxidation reactions were performed with samples, which were already participated in the first run. Therefore, the surface states (or the catalytic-active sites) were expected to be different for the samples in the first and the second runs. The values (T_M-Al2O3,2nd − T_M-Al2O3,1st) are summarized in Table 1. The positive value (Figure 5d) indicated that the CO oxidation started at a higher temperature in the second run. In other words, the CO oxidation catalytic activity became lower in the second run.

For Co- and Ni-Al₂O₃ nanosheets, the onset temperatures in the second run were observed to be slightly lower than those in the first run. However, the other samples commonly showed higher onset temperatures in the second run, compared with the first run. This indicated that, for the latter, the catalytic activity became somewhat lower after the first run. The lower catalytic activity appeared to be mainly due to a change in crystallinity and lower catalytic-active sites.

To evaluate the roles of the transition metals in catalytic activities, compared with bare Al₂O₃, Figure 5e,f show the relative CO oxidation onsets (T_Al2O3 − T_M-Al2O3), compared with those of the first and the second runs of the bare Al₂O₃, respectively. The values are summarized in

Table 2. Differences in CO oxidation onset temperatures (T_Al2O3,onset − T_{M-Al2O3,onset}) in the first and second runs, compared with that of bare Al₂O₃ nanosheets. The CO oxidation onset temperatures of bare Al₂O₃ were 386 °C and 373 °C for the first and the second runs, respectively.

Group #9	First Run	Second Run	Group #10	First Run	Second Run	Group #11	First Run	Second Run
Co	4	12	Ni	104	107	Cu	−26	16
Rh	−251	−201	Pd	−185	−166	Ag	−46	24
Ir	−191	−156	Pt	−189	−167	Au	−11	42

. In the first runs, the T_Al2O3,1st − T_M-Al2O3,1st values of Co and Ni showed positive, and others showed negative values. In the second runs, Co, Ni, Cu, Ag, and Au showed positive, and others showed negative values. On the basis of Figure 5e,f, the catalytic activity became poorer upon loading Co and Ni, compared with bare Al₂O₃. Unexpectedly, the Au, Ag, and Cu (group 11) showed somewhat higher activities in the first run but showed poorer catalytic activity in the second run, compared with the bare Al₂O₃ nanosheet. The Rh, Pd, Ir, and Pt showed much higher (with lowering of onset temperatures between 156 °C and 261 °C) CO oxidation activity in the first and second runs. Conclusively, the CO oxidation activity showed the order of Ni < Co < Au < Cu < Ag < Pd < Pt < Ir < Rh in the first run, and Ni < Au < Ag < Cu < Co < Ir < Pt ≈ Pd < Rh in the second run.

For CO oxidation, a simplified mechanism is described below;

M-Al₂O₃ + CO (g) → CO(ad)-M-Al₂O₃      adsorption of CO

(1)

M-Al₂O₃ + 1/2O₂ (g) → O(ad)-M-Al₂O₃   dissociative adsorption of O₂

(2)

CO (g) + O(ad)-M-Al₂O₃ → M-Al₂O₃ + CO₂ (g)

(3)

CO (g) + HO-M-Al₂O₃ → M-Al₂O₃-H + CO₂ (g)

(4)

CO (g) + O₂(ad)-M-Al₂O₃ → CO₃(ad)-M-Al₂O₃   carbonate formation

(5)

CO₃(ad)-M-Al₂O₃ → O(ad)-M-Al₂O₃ + CO₂ (g)

(6)

The CO oxidation mechanism was explained by the Langmuir-Hinshelwood mechanism [12,13] and the non-Langmuir-Hinshelwood mechanism [4,11], depending on the overlayer transition metals. In reaction (1), CO was adsorbed on metal site, and in reaction (2), oxygen was dissociatively adsorbed on the surface. In reaction (3), gaseous CO and surface O reacted to release CO₂ [12,13]. When moisture was present in the reaction, the surface OH group was plausibly formed and CO might also react with the surface metal hydroxide to form the CO₂ in reaction (4) [15]. On the basis of the FT-IR spectra (Supporting Information, Figure S7), surface OH groups were observed in the as-prepared samples. Therefore, reaction (4) was likely involved in the first run of CO₂ formation. If H was not desorbed as H₂O, the surface H was recycled as shown in reaction (4). Otherwise, the H was removed from the surface as gaseous H₂O, and, thus, the reaction (4) was diminished in the second run. Reaction (5) was also reported for surfaces such as Pt/Al₂O₃ [4,11]. In reaction (5), CO was adsorbed on oxygenated metal atoms to initially form carbonate. Then, the carbonate dissociated to generate CO₂ in reaction (6).

Photocatalytic CO₂ reduction products were examined for bare Al₂O₃ and M-Al₂O₃ nanosheets and are displayed in Figure 6 [39,41,42]. Major CO₂ reduction products were observed to be carbon monoxide (CO), methanol (CH₃OH), and methane (CH₄) with an order: CH₄ < CH₃OH < CO. CO was the most dominantly produced species. CH₃OH showed a higher production amount compared with CH₄. Hydrogen (H₂) was additionally observed as a photocatalytic water splitting product during CO₂ reduction. Figure 6a plots all of the product amounts (μmol/mol = ppm) for bare Al₂O₃ and M-Al₂O₃ nanosheets. As a quick glance, Ag-Al₂O₃ nanosheets showed the highest amounts of CO₂ reduction products: 237.3 ppm for CO, 36.3 ppm for CH₃OH, and 30.9 ppm for CH₄, and Rh-Al₂O₃ nanosheets showed the highest H₂ production (20.7 ppm). For the bare Al₂O₃ nanosheets in Figure 6b, CO, CH₃OH, and CH₄ were observed to be 107.5 ppm, 29.6 ppm, and 19.5 ppm, respectively. No H₂ was detected. CO reduction yields (μmol/mol) in different groups of 9, 10, and 11, and with different units (μmol/g), are provided in the Supporting Information, Figures S8 and S9, respectively.

For bare Al₂O₃, the selectivities for CO, CH₃OH, and CH₄ were estimated to be 68.6%, 18.9%, and 12.5%, respectively. Upon Co- and Cu-loading, CH₄ and H₂ showed meaningful (>25%) enhancements. However, the amounts of CO and CH₃OH showed no critical change. CO, CH₃OH, and CH₄ productions were enhanced by 28%, 17%, and 24% upon Ni-loading. CO was increased by 2.2× upon loading Ag in Figure 6c. CH₃OH and CH₄ were also increased by 1.23× and 1.58×, respectively, upon loading Ag. Rh and Pd-loadings had a smaller effect on the CO production relative to the bare support. CH₃OH and CH₄ productions were not meaningfully enhanced by Rh- and Pd-loadings. Instead, interestingly the H₂ production was commonly observed in these metal-loadings. For Ir, Pt, and Au elements in period 6, CO productions were all decreased by metal-loadings. CH₃OH productions were somewhat increased by 19% and 16% upon loading of Pt and Au, respectively. The CH₄ production was only increased upon loading Pt relative to the bare substrate.

For H₂ production, Ag, Pd, and Rh (in period 5) metals commonly showed H₂ productions with amounts of 2.1 ppm, 3.0 ppm, and 20.7 ppm, respectively. For the metals of Co, Ni, and Cu (in period 4), the H₂ production amounts were observed to be 1.9 ppm, 0 ppm, and 3.0 ppm, respectively. That is, Ni showed no H₂ production. The metals of Au, Pt, and Ir in period 6 commonly showed no H₂ production at all. The Rh-Al₂O₃ nanosheets predominantly showed the highest H₂ production with an amount of 20.7 ppm.

The photocatalytic CO₂ reduction mechanism is generally written as xCO₂ + yH⁺ + ze⁻ → C_aH_bO_c products + dH₂O [41,42]. Electrons (e⁻) and holes (h⁺) were generated under UVC irradiation in reaction (7). H⁺ ion was generated via the reactions in (8)–(11). The generation of electrons was an important factor for the multielectron processes. The mechanisms for the productions of CO (in reaction (12)), CH₃OH (in reaction (13)), and CH₄ (in reaction (14)) are written as below and shown in Figure 6 [38,39].

Al oxides + UVC → Al oxides (e⁻ + h⁺)

(7)

H₂O → H⁺ + OH

(8)

OH⁻ + h⁺ → •OH

(9)

•OH + H₂O + 3h⁺ → O₂ + 3H⁺

(10)

H⁺ + e⁻ → 1/2H₂

(11)

CO₂ + 2H⁺ + 2e⁻ → CO + H₂O, −0.530 V vs. standard hydrogen electrode (SHE)

(12)

CO₂ + 6H⁺ + 6e⁻ → CH₃OH + H₂O, −0.380 V vs. SHE

(13)

CO₂ + 8H⁺ + 8e⁻ → CH₄ + 2H₂O, −0.240 V vs. SHE

(14)

These reaction channels were closely spaced in free energy change, and, thus, the hydrogen production channel (H⁺ + e⁻ → 1/2H₂, −0.42 V vs. SHE) occurred competitively. In the mechanism, CO₂ was initially adsorbed to form COOH. The COOH was then attacked by H⁺ and e⁻ to generate gaseous CO. The CO production channel was only enhanced by loading Ag or Ni on Al₂O₃ support. CH₃OH production was likely formed when surface CO_ad underwent step-wise hydrogenation. This production was enhanced by loading Ni, Rh, Ag, Pt, or Au on Al₂O₃ support. CH₄ production was formed via C–O bond scission of hydrogenated ≡C‒OH and new C‒H bond formation. This production was somewhat enhanced by loading Co, Ni, Cu, Ag, or Pt. The present pre-screening tests need further investigations to understand the detailed roles of the overlayer elements, with the aid of density functional theory.

4. Conclusions

In summary, γ-Al₂O₃ nanosheets were prepared by the solvothermal method followed by thermal calcination at 600 °C for 2 h. Transition metals (M = Co, Ni, Cu, Rh, Pd, Ag, Ir, Pt, and Au) were loaded on Al₂O₃ nanosheet supports, and their thermal CO oxidation and photocatalytic CO₂ reduction activities were fully tested.

The thermal CO oxidation activity showed the order of Ni < Co < Au < Cu < Ag << Pd < Pt < Ir < Rh in the first run, and Ni < Au < Ag < Cu < Co << Ir < Pt ≈ Pd < Rh in the second run. The Au, Ag, Co, Ni, and Cu elements reduced the catalytic activity on the Al₂O₃ support. CO oxidation activity was greatly enhanced by the loading of Ir, Pt, Pd, and Rh elements. Rh-Al₂O₃ nanosheets showed the highest CO oxidation activity with onset temperatures of 135 °C and 172 °C for the first and the second runs, respectively.

Photocatalytic CO₂ reduction experiments were also performed to show that CO, CH₃OH, and CH₄ were common products with an order of CH₄ (14.6–30.9 ppm range) < CH₃OH (23.0–36.3 ppm range) << CO (76.5–237.3 ppm range). The highest performance was achieved after Ag-loadings with yields of 237.3 ppm for CO, 36.3 ppm for CH₃OH, and 30.9 ppm for CH₄, corresponding to 2.2×, 1.2×, and 1.6× enhancements, respectively, compared with those for the bare Al₂O₃. CO production was substantially decreased by the loading of Pd and Pt. Hydrogen production was enhanced by Rh-loadings with a yield of 20.7 ppm. Conclusively, Rh-Al₂O₃ and Ag-Al₂O₃ showed the best thermal CO oxidation and photocatalytic CO₂ reduction performances, respectively, among Co, Ni, Cu, Rh, Pd, Ag, Ir, Pt, and Au element loadings.

The present pre-screening test results could be a useful quick guide for the selection of overlayer transition metals in the groups of 9, 10, and 11 when Al₂O₃ is used as a support catalyst material. It also enriched the understanding of the role of an overlayer transition metal.