The Enhancement of CO Oxidation Performance and Stability in SO2 and H2S Environment on Pd-Au/FeOX/Al2O3 Catalysts

Carbon monoxide (CO) is a colourless, odourless, and toxic gas. Long-term exposure to high concentrations of CO causes poisoning and even death; therefore, CO removal is particularly important. Current research has focused on the efficient and rapid removal of CO via low-temperature (ambient) catalytic oxidation. Gold nanoparticles are widely used catalysts for the high-efficiency removal of high concentrations of CO at ambient temperature. However, easy poisoning and inactivation due to the presence of SO2 and H2S affect its activity and practical application. In this study, a bimetallic catalyst, Pd-Au/FeOx/Al2O3, with a Au:Pd ratio of 2:1 (wt%) was formed by adding Pd nanoparticles to a highly active Au/FeOx/Al2O3 catalyst. Its analysis and characterisation proved that it has improved catalytic activity for CO oxidation and excellent stability. A total conversion of 2500 ppm of CO at −30 °C was achieved. Furthermore, at ambient temperature and a volume space velocity of 13,000 h−1, 20,000 ppm CO was fully converted and maintained for 132 min. Density functional theory (DFT) calculations and in situ FTIR analysis revealed that Pd-Au/FeOx/Al2O3 exhibited stronger resistance to SO2 and H2S adsorption than the Au/FeOx/Al2O3 catalyst. This study provides a reference for the practical application of a CO catalyst with high performance and high environmental stability.


Introduction
Carbon monoxide produced from the inadequate combustion of fossil fuels in the chemical industry, with inefficient fire bursts and explosions in a sealed environment, poses significant safety risks to human life. Carbon monoxide poisoning is the most common cause of gas poisoning in the population, and there is no effective antidote. The pathophysiology of carbon monoxide poisoning includes reducing overall oxygen transmission and inhibiting mitochondrial respiration [1]. Therefore, the removal of CO via efficient catalysts is essential to ensure human health and safety.
Gold nanoparticles have been extensively used in catalytic oxidation reactions because of their excellent catalytic properties [2][3][4][5][6][7]. Supported nano-gold catalysts exhibit improved performance in the catalytic oxidation of CO at low (constant) temperatures, achieving complete conversion at high concentrations and room temperature [8,9]. Their performance is affected mainly by the size of the gold nanoparticles and their interaction with the support [10]. The preparation methods for supported nano-gold catalysts include impregnation [11], deposition-precipitation [12], coprecipitation [13], chemical vapour precipitation [14], and liquid-phase reduction [15]. Among them, the deposition-precipitation method is widely used because it is convenient to operate and can prepare particles with small sizes and uniform distributions [10]. FeO x particles are a mixture of Fe 2 O 3 and Fe 3 O 4 , which are supported on Al 2 O 3 . These are beneficial for the good dispersion of nano-gold particles. The oxygen vacancy on the supported nano-gold catalyst is the activation centre of oxygen molecules [16], and support selection plays a significant role in improving activity. α-Fe 2 O 3 has abundant oxygen vacancies that can activate oxygen during the reaction process and is an excellent carrier for gold nanoparticles [17][18][19]. Guoyan Ma et al. [20] prepared γ-Al 2 O 3 -supported Cu, and the oxidation activity of the catalyst with Fe as an additive was significantly higher than that of pure Cu because the addition of Fe formed more oxygen vacancies and promoted the activity of the catalyst. Chunlei Wang et al. [21] reported the selective deposition of FeO x coatings onto SiO 2 -supported Ir nanoparticles using atomic layer deposition (ALD) technology, which allows precise customisation of IrFeO interfaces to optimize the catalytic performance of PROX reactions. Compared with the uncoated Ir/SiO 2 samples, the FeO x -coated Ir/SiO 2 samples showed significantly enhanced activity.
Studies have also shown that the formation of various gold compounds during urea-assisted deposition-precipitation increases the loading capacity of gold nanoparticles [22,23]. However, Au agglomerates were formed during the calcination process when chloroauric acid was used to prepare supported nano-gold catalysts using the depositionprecipitation method due to residual chlorine. Washing the catalyst with ammonia can effectively prevent gold sintering by removing residual chlorine [16]. Therefore, a supported nano-gold catalyst was prepared using the deposition-precipitation method with urea (CO(NH 2 ) 2 ) (DPU) and washed with ammonia. The resulting supported nano-gold catalyst had a small particle size and uniform distribution.
Long-term SO 2 exposure reduces the activity of supported nano-gold catalysts, severely limiting their application in treating pollutants in industrial flue gas [8]. The catalytic activity of supported nano-gold catalysts for CO oxidation significantly decreases in the presence of SO 2 [24], possibly because of the increased adsorption strength between Au and CO after SO 2 treatment, which inhibits the movement of adsorbed CO on Au particles to the gold-carrier interface to form CO 2 [25]. The poor anti-sulphide ability of nano-gold catalysts is a crucial limiting factor in their application. Supported nano-Pd catalysts also exhibit strong CO catalytic oxidation performance [26,27], but they suffer from low activity at low temperatures [28]. Pd-Au bimetallic catalysts can be efficiently used in various catalytic reactions [29][30][31][32]. A theoretical calculation study reported by Jin Zhang et al. [33] showed that adding gold inhibits Pd/TiO 2 poisoning in CO oxidation reactions by forming the "Golden Crown" Pd-Au structure. However, experimental studies investigating the effects of SO 2 and H 2 S on the performance of supported Pd-Au catalysts have not been reported. Using in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) and temperature programmed desorption (TPD), Wilburn et al. [34] showed that the adsorption of SO 2 by Pd-Pt alloys depends on the Pd:Pt molar ratio and that the effect of SO 2 and H 2 S on catalyst performance can be reduced by regulating the metal proportion in bimetallic catalysts.
The Al 2 O 3 used in this study was spherical alumina. As an excellent carrier, it is the skeleton material of the Au/FeO x /Al 2 O 3 catalyst, which is beneficial for the industrial application of catalysts and for preserving and filling in the catalyst reaction bed. Therefore, in this study, a Pd/FeO x /Al 2 O 3 catalyst was prepared via ultrasonic-assisted impregnation using γ-Al 2 O 3 as the first carrier and ferric oxide as the second carrier. Gold nanoparticles were prepared using a deposition-precipitation method. Pd-Au/FeO x /Al 2 O 3 catalysts with a mass ratio of 1:2 (Pd:Au) were obtained. The loading capacity of Au nanoparticles was 2 wt%. The physicochemical properties of the Au/FeO x /Al 2 O 3 and Pd-Au/FeO x /Al 2 O 3 catalysts were compared using XRD, TEM, ICP, BET, XPS, and CO-TPR to evaluate catalytic CO oxidation activity and stability. The mechanism of catalytic CO oxidation and its adsorption on SO 2 and H 2 S were calculated and analysed using density functional theory (DFT) studies and in situ FTIR, revealing the reasons for the more substantial anti-SO 2 and anti-H 2 S effects of Pd-Au/FeO x /Al 2 O 3 .

FeO x /Al 2 O 3 Preparation
First, alumina was pre-treated with activation: γ-Al 2 O 3 (size 1-3 mm) was placed in a crucible, heated in a muffle furnace at 600 • C and a rate of 10 • C/min, and kept at that temperature for 2 h. The FeO x /Al 2 O 3 support was obtained by preparing a 0.4 mol/L solution of Fe(NO 3 ) 3 (AR) and supporting it on γ-Al 2 O 3 using equal volume impregnation, followed by drying in an oven at 120 • C for 12 h. This support was then placed in a muffle furnace at 500 • C, with a heating rate of 10 • C/min, and maintained at this temperature for 2 h. This process was repeated for a 0.2 mol/L Fe(NO 3 ) 3 solution.

Preparation of Au/FeO x /Al 2 O 3 Catalysts
The FeO x /Al 2 O 3 carrier was placed in a conical bottle with an equal weight of urea as the precipitator, followed by a solution of chloroauric acid with 2 wt% Au in 100 mL of deionised (DI) water. The resulting solution was stirred and heated to 80 • C. The heating was stopped when the pH reached 8-8.5. After 4 h, the catalyst precursor was washed with a large amount of deionised water and a small amount of ammonia water until no precipitation was observed when testing with a 0.5 mol/L silver nitrate solution. The washed catalyst was oven dried for 12 h at 120 • C. The dried samples were roasted in a tubular furnace in an oxygen atmosphere for 2 h at 300 • C and in a hydrogen atmosphere for 2 h at 300 • C to obtain a Au/FeO x /Al 2 O 3 catalyst with 2 wt% Au. BET analysis of Au/FeO x /Al 2 O 3 catalysts is shown in Figure S3a.

Preparation of Pd-Au/FeO x /Al 2 O 3 Catalysts
After characterization analysis of Pd-Au/FeO x /A 2 O 3 catalysts with 5 types of Pd: Au mass ratios (Table S1, Figures S1 and S2). Pd-Au/FeO x /Al 2 O 3 catalysts were prepared, as shown in Figure 1.

Test Methods
The catalyst structure was characterised using X-ray diffractometry (XRD, Panalytical, X' Pert Pro MPD, Almelo, The Netherlands) at a scanning angle of 10-80° and a scanning rate of 2°/min, The wavelength was 1.5418 nm, the voltage was 40 kV, and the current was 40 mA. Transmission electron microscopy (TEM, JEM-F200, JEOL, Tokyo, Japan) at 200 kV was used to characterise the surface morphology and elemental distributions of the samples. An inductively coupled plasma optical emission spectrometer (ICP-OES, 5110, Agilent, Santa Clara, CA, USA) was used to determine the metal content of the catalysts. X-ray photoelectron spectroscopy (XPS, K-Alpha, Thermo Scientific, Waltham, MA, USA) was used to analyse the valence changes of the elements on the catalyst surface. The C1s binding energy was used to correct 284.8 eV. A temperature-programmed chemisorption instrument (CO-TPR, BELCat II, Microtrac, Osaka, Japan) was used to evaluate cata- (1) Preparation of Pd/FeO x /Al 2 O 3 catalysts Ten grams of FeO x /Al 2 O 3 carriers was placed in a conical flask with 0.1 g of Pd(NO 3 ) 2 to obtain a final solution volume of 50 mL. After standing for 30 min and undergoing ultrasonic-assisted impregnation for 30 min, the solution was placed in a water bath shaker at 25 • C and 135 rpm for 24 h. After the reaction was completed, the product was dried for 12 h at 120 • C. The dried samples were roasted in a tube furnace at 300 • C in an oxygen atmosphere for 2 h and at 300 • C in a hydrogen atmosphere for 30 min to obtain Pd/FeO x /Al 2 O 3 catalysts. After step (1), The Pd/FeO x /Al 2 O 3 catalysts were placed in a conical bottle with an equal weight of urea added as the precipitator, followed by a solution of chloroauric acid (2 wt% Au in 100 mL of deionised water). The solution was then stirred and heated to 80 • C. The heating was stopped when the pH reached 8-8.5. After 4 h, the obtained Pd-Au/FeO x /Al 2 O 3 catalyst precursor was washed with a large amount of deionised water and a small amount of ammonia water until no precipitation was produced when testing with a 0.5 mol/L silver nitrate solution. The washed catalyst was oven dried for 12 h at 120 • C. The dried samples were roasted in a tubular furnace in an oxygen atmosphere for 2 h at 300 • C and in a hydrogen atmosphere for 2 h at 300 • C to obtain a Pd-Au/FeO x /Al 2 O 3 catalyst with a Pd:Au load-mass ratio of 1:2. BET analysis of Pd-Au/FeO x /Al 2 O 3 catalysts is shown in Figure S3b.

Test Methods
The catalyst structure was characterised using X-ray diffractometry (XRD, Panalytical, X' Pert Pro MPD, Almelo, The Netherlands) at a scanning angle of 10-80 • and a scanning rate of 2 • /min, The wavelength was 1.5418 nm, the voltage was 40 kV, and the current was 40 mA. Transmission electron microscopy (TEM, JEM-F200, JEOL, Tokyo, Japan) at 200 kV was used to characterise the surface morphology and elemental distributions of the samples. An inductively coupled plasma optical emission spectrometer (ICP-OES, 5110, Agilent, Santa Clara, CA, USA) was used to determine the metal content of the catalysts. X-ray photoelectron spectroscopy (XPS, K-Alpha, Thermo Scientific, Waltham, MA, USA) was used to analyse the valence changes of the elements on the catalyst surface. The C1s binding energy was used to correct 284.8 eV. A temperature-programmed chemisorption instrument (CO-TPR, BELCat II, Microtrac, Osaka, Japan) was used to evaluate catalyst performance in CO oxidation; the samples were heated at 10 • C/min from room temperature to −300 • C for drying pretreatment. The air flow (50 mL/min) was purged for 1 h and then cooled to 50 • C. The samples were desorbed at 10 • C/min in 10% CO/He air flow and then desorbed at 900 • C, and the reduction gas was detected using TCD. In situ FTIR (80 V Brucker, Billerica, MA, USA) was used to detect and analyse the adsorption processes of the materials on the catalyst surface using the diffuse reflection integrating sphere model.

SO 2 and H 2 S Pretreatment
For SO 2 and H 2 S pretreatment, air was used as the balance gas before catalytic CO oxidation. The catalyst was treated with 2 ppm SO 2 (100 mL/min for 1 h) and 2 ppm H 2 S (100 mL/min for 1 h) at 25 • C under atmospheric pressure. Air was used as the balance gas when evaluating the effect of temperature on the CO conversion rate; the CO concentration was 2500 ppm, and the flow rate was 50 mL/min. CO stability at room temperature (25 • C) was tested with a CO concentration of 20,000 ppm and a volume space velocity of 13,000 h −1 .

Evaluation of Catalytic Activity
The catalytic oxidation of CO to CO 2 was performed using a small fixed-bed continuousflow reactor with mass flowmeter control of the raw gas flow and a U-shaped quartz tube with an inner diameter of 1 cm as the reaction tube. A supported nano-gold catalyst (0.5 g) was weighed into the reaction tube, and the catalyst bed was fixed using quartz cotton at both ends. The raw gas consisted of 0.25% CO, 21.1% N 2 , and 78.5% O 2 . The flow rate was 50 mL/min, and the reaction temperature was −30-60 • C. Gas chromatography using an HP6890 gas chromatograph with high-purity hydrogen as the carrier gas, a hydrogen ion flame detector (FID), and a thermal conductivity detector (TCD) containing a reformer was used to detect the concentration of CO in the tail gas. The CO conversion rate (X CO , %) was calculated using the following equation: where C in and C out represent the initial CO concentration and the CO concentration in the tail gas , respectively.

Evaluation of the Stability of the Catalysts
Catalyst stability was also evaluated using a small fixed-bed continuous-flow reaction device, as shown in Figure 2. Here, the CO standard gas was diluted with a specific flow of compressed air to control the rate of CO generation. A U-shaped hard quartz tube with an inner diameter of 1 cm was used as the reaction tube. The height of the catalyst bed in the reaction tube was 8 cm, the bed density was 0.64 g/cm 3 , the volume space velocity was 13,000 h −1 , the reaction temperature was 25 • C, and the concentration was 20,000 ppm. The concentrations of CO and CO 2 in the tail gas were determined using gas chromatography (HP6890).
reformer was used to detect the concentration of CO in the tail gas. The CO conversion rate (XCO, %) was calculated using the following equation: where Cin and Cout represent the initial CO concentration and the CO concentration in the tail gas, respectively.

Evaluation of the Stability of the Catalysts
Catalyst stability was also evaluated using a small fixed-bed continuous-flow reaction device, as shown in Figure 2. Here, the CO standard gas was diluted with a specific flow of compressed air to control the rate of CO generation. A U-shaped hard quartz tube with an inner diameter of 1 cm was used as the reaction tube. The height of the catalyst bed in the reaction tube was 8 cm, the bed density was 0.64 g/cm 3 , the volume space velocity was 13,000 h −1 , the reaction temperature was 25 °C, and the concentration was 20,000 ppm. The concentrations of CO and CO2 in the tail gas were determined using gas chromatography (HP6890).

DFT Calculations
All DFT calculations were performed using the Vienna ab initio simulation (VASP5.4.4) code [35]. The exchange correlation was simulated with the PBE functional, and the ion-electron interactions were described using the projector augmented wave (PAW) method [36,37] Van der Waals (vdW) interactions were included using the empirical DFT-D3 method [38]. The Fe2O3 (110) surface supported the Au13 and Au9Pd4 nanoclusters. The atoms in the upper two layers of the surface were allowed to move freely, whereas the bottom two layers were fixed to simulate the surface of the structure. The Monkhorst-Pack-grid-mesh-based Brillouin zone k-points were set to 2 × 2 × 1 for all periodic structures, with a cut-off energy of 450 eV. The convergence criteria were 0.01 eV A −1 and 10 -5 eV in force and energy, respectively.
The free energy for species adsorption (ΔG) was calculated using the following equation: where ΔE, ΔEZPE, and ΔS represent the changes in electronic energy, zero-point energy, and entropy, respectively, caused by intermediate adsorption. ΔH0 → T represents the change in enthalpy when heated from 0 to T K.

DFT Calculations
All DFT calculations were performed using the Vienna ab initio simulation (VASP5.4.4) code [35]. The exchange correlation was simulated with the PBE functional, and the ion-electron interactions were described using the projector augmented wave (PAW) method [36,37] Van der Waals (vdW) interactions were included using the empirical DFT-D3 method [38]. The Fe 2 O 3 (110) surface supported the Au 13 and Au 9 Pd 4 nanoclusters. The atoms in the upper two layers of the surface were allowed to move freely, whereas the bottom two layers were fixed to simulate the surface of the structure. The Monkhorst-Packgrid-mesh-based Brillouin zone k-points were set to 2 × 2 × 1 for all periodic structures, with a cut-off energy of 450 eV. The convergence criteria were 0.01 eV A −1 and 10 −5 eV in force and energy, respectively. The free energy for species adsorption (∆G) was calculated using the following equation: where ∆E, ∆E ZPE , and ∆S represent the changes in electronic energy, zero-point energy, and entropy, respectively, caused by intermediate adsorption. ∆H 0→T represents the change in enthalpy when heated from 0 to T K.  [39,40]. However, AuNPs were not detected because they were highly dispersed on the FeO x /Al 2 O 3 support. In the Pd-Au/FeO x /Al 2 O 3 catalyst, both the characteristic peaks of γ-Al 2 O 3 and α-Fe 2 O 3 and the characteristic peaks of gold nanoparticles (44.6 • , 52.0 • , and 76.7 • [41]) were detected, possibly because of the effect of Pd on the dispersion of gold nanoparticles. As the XRD pattern of this bimetallic catalyst did not show the characteristic peak of Pd, we confirmed its loading onto the FeO x /Al 2 O 3 carrier through HRTEM characterisation and mapping of the energy spectrum. Moreover, after catalysis, the structures of these three catalysts remained unchanged.  Figure 3 shows the XRD patterns of Pd/FeOx/Al2O3, Au/FeOx/Al2O3 and Pd-Au/FeOx/Al2O3 catalysts. The Au/FeOx/Al2O3 and Pd-Au/FeOx/Al2O3 catalysts exhibited characteristic peaks of γ-Al2O3 and α-Fe2O3 [39,40]. However, AuNPs were not detected because they were highly dispersed on the FeOx/Al2O3 support. In the Pd-Au/FeOx/Al2O3 catalyst, both the characteristic peaks of γ-Al2O3 and α-Fe2O3 and the characteristic peaks of gold nanoparticles (44.6°, 52.0°, and 76.7° [41]) were detected, possibly because of the effect of Pd on the dispersion of gold nanoparticles. As the XRD pattern of this bimetallic catalyst did not show the characteristic peak of Pd, we confirmed its loading onto the FeOx/Al2O3 carrier through HRTEM characterisation and mapping of the energy spectrum. Moreover, after catalysis, the structures of these three catalysts remained unchanged.  Figure 4 shows HRTEM images and elemental distribution diagrams of the prepared Au/FeOx/Al2O3, Pd/FeOx/Al2O3, and Pd-Au/FeOx/Al2O3 catalysts. Au was uniformly dispersed in Au/FeOx/Al2O3 (Figure 4a,b). The particle size distribution of the Au nanoparticles was measured using Nano Measurer software, revealing an average particle size of 3.5 nm (Figure 4d). Pd nanoparticles were uniformly dispersed in Au/FeOx/Al2O3 (Figure 4e,f). Pd nanoparticles in the Pd-Au/FeOx/Al2O3 catalyst had a larger particle size ( Figure  4I,k), with smaller nanometre Au particles on the Pd particles. XEDS atlas and S element mapping were performed for the sulphurated Au/FeOx/Al2O3 and Pd-Au/FeOx/Al2O3 catalysts under similar conditions ( Figure 5), revealing an S adsorption of 7.8 and 33.66 wt% onto Pd-Au/FeOx/Al2O3 and Au/FeOx/Al2O3 catalysts, respectively. This confirms that SO2 and H2S are easily adsorbed onto the Au/FeOx/Al2O3 catalyst.  (Figure 4a,b). The particle size distribution of the Au nanoparticles was measured using Nano Measurer software, revealing an average particle size

XPS Analysis
XPS spectra were used to analyse the Au4f and Pd3d valence states of Au/FeOx and Pd-Au/FeOx/Al2O3 after treatment with SO2 and H2S (Figures 6 and 7). Figure 6 the Au4f binding spectra of both catalysts. The peaks of Au/FeOx/Al2O3 at 87.3 eV an eV were assigned to Au4f 5/2 and Au4f 7/2, respectively, both of which correspon Au 0 (Figure 6a) [42,43]. After vulcanisation, the peaks corresponding to Au4f 5/2 an 7/2 increased in energy, appearing at 87.5 eV and 83.7 eV, respectively. This may be by the reaction of SO2 and H2S with gold nanoparticles to form Au δ+ . Fo Au/FeOx/Al2O3, the Au4f 5/2 and Au4f 7/2 peaks remained unchanged with SO2 an addition (Figure 6b), indicating that the addition of Pd weakened the effect of SO2 a on the Au nanoparticles.    (Figures 6 and 7). Figure 6 shows the Au4f binding spectra of both catalysts. The peaks of Au/FeO x /Al 2 O 3 at 87.3 eV and 83.5 eV were assigned to Au4f 5/2 and Au4f 7/2, respectively, both of which corresponded to Au 0 (Figure 6a) [42,43]. After vulcanisation, the peaks corresponding to Au4f 5/2 and Au4f 7/2 increased in energy, appearing at 87.5 eV and 83.7 eV, respectively. This may be caused by the reaction of SO 2 and H 2 S with gold nanoparticles to form Au δ+ . For Pd-Au/FeO x /Al 2 O 3 , the Au4f 5/2 and Au4f 7/2 peaks remained unchanged with SO 2 and H 2 S addition (Figure 6b), indicating that the addition of Pd weakened the effect of SO 2 and H 2 S on the Au nanoparticles.

XPS Analysis
XPS spectra were used to analyse the Au4f and Pd3d valence states of Au/FeOx/Al2O3 and Pd-Au/FeOx/Al2O3 after treatment with SO2 and H2S (Figures 6 and 7). Figure 6 shows the Au4f binding spectra of both catalysts. The peaks of Au/FeOx/Al2O3 at 87.3 eV and 83.5 eV were assigned to Au4f 5/2 and Au4f 7/2, respectively, both of which corresponded to Au 0 (Figure 6a) [42,43]. After vulcanisation, the peaks corresponding to Au4f 5/2 and Au4f 7/2 increased in energy, appearing at 87.5 eV and 83.7 eV, respectively. This may be caused by the reaction of SO2 and H2S with gold nanoparticles to form Au δ+ . For Pd-Au/FeOx/Al2O3, the Au4f 5/2 and Au4f 7/2 peaks remained unchanged with SO2 and H2S addition (Figure 6b), indicating that the addition of Pd weakened the effect of SO2 and H2S on the Au nanoparticles.   and H2S increased after SO2 and H2S treatment because the SO2 and H2S adsorbed to Pd to form Pd-S [48]. By calculating the content of Pd in different states before and after SO2 and H2S treatment (Table 1), we determined that Pd-Au and PdOx/Pd decreased after SO2 and H2S treatment, whereas the Pd 0 content increased. We hypothesised that this was due to PdOx reduction to Pd 0 in the presence of SO2 and H2S. The Pd in Pd-Au exhibited stronger adsorption with SO2 and H2S [49,50].   Figure 8 shows the CO conversion for the catalytic oxidation of CO at different temperatures using Au/FeOx/Al2O3 and Pd-Au/FeOx/Al2O3 as catalysts before and after SO2 and H2S pretreatment. Compared with Pd/FeOx/Al2O3 catalysts, the activities of Pd-Au/FeOx/Al2O3 and Au/FeOx/Al2O3 catalysts were significantly improved. Compared with Au/Al2O3 and Pd-Au/Al2O3 catalysts, after the addition of FeOx, the activities of Pd-Au/FeOx/Al2O3 and Au/FeOx/Al2O3 catalysts were significantly improved. The Pd-Au/FeOx/Al2O3 and Au/FeOx/Al2O3 catalysts without SO2 and H2S pretreatment had excellent CO conversion rates, achieving complete conversion at a CO concentration of 2500 ppm in the −30-60 °C temperature range. After SO2 and H2S pretreatment, the Pd-Au/FeOx/Al2O3 catalyst conversion rate for a CO concentration of 2500 ppm was 87.2% at −30 °C, and the conversion rate was 100% for temperatures above 25 °C. However, the conversion rate of Au/FeOx/Al2O3 after SO2 and H2S pretreatment was only 23-26% at all tested temperatures, indicating that the Pd-Au/FeOx/Al2O3 catalyst was affected less by SO2 and H2S, matching the observed low SO2 and H2S adsorption measured by XEDS.   [46]. The peaks in the 332-347 eV range were attributed to Pd3d 5/2, with peaks at 336.7, 334.7, and 332.4 eV, corresponding to PdO x /Pd, Pd 0 , and Pd of the interacting Pd-Au species [44,47]. The binding energy of Pd-S interacting with SO 2 and H 2 S increased after SO 2 and H 2 S treatment because the SO 2 and H 2 S adsorbed to Pd to form Pd-S [48]. By calculating the content of Pd in different states before and after SO 2 and H 2 S treatment (Table 1), we determined that Pd-Au and PdO x /Pd decreased after SO 2 and H 2 S treatment, whereas the Pd 0 content increased. We hypothesised that this was due to PdO x reduction to Pd 0 in the presence of SO 2 and H 2 S. The Pd in Pd-Au exhibited stronger adsorption with SO 2 and H 2 S [49,50].  Figure 9 shows the catalytic oxidation stability test curves of the Au/FeOx/Al2O3 and Pd-Au/FeOx/Al2O3 catalysts for CO at 25 °C before and after SO2 and H2S treatment. The conversion rate of the Au/FeOx/Al2O3 catalyst decreased from 97.9% to 52.4% after 726 min, which was slower than that of the Au/FeOx/Al2O3 catalyst pretreated with SO2 and H2S (91.9% to 37% after 550 min). The conversion rate of Pd-Au/FeOx/Al2O3 was 100% in the first 132 min, the TON of it was 205, and it declined slowly to 97.9% after 748 min. Similarly, the conversion rate of SO2and H2S-pretreated catalyst decreased from 98.1% to 96.5% after 726 min. Thus, Pd-Au/FeOx/Al2O3 maintained strong stability after SO2 and H2S treatment, with negligible conversion reduction (approximately 2%) after 726 min. It is deduced that this is due to the adsorption of SO2 and H2S on the catalyst surface, which inhibits the adsorption and reaction of CO.   Figure 9 shows the catalytic oxidation stability test curves of the Au/FeO x /Al 2 O 3 and Pd-Au/FeO x /Al 2 O 3 catalysts for CO at 25 • C before and after SO 2 and H 2 S treatment. The conversion rate of the Au/FeO x /Al 2 O 3 catalyst decreased from 97.9% to 52.4% after 726 min, which was slower than that of the Au/FeO x /Al 2 O 3 catalyst pretreated with SO 2 and H 2 S (91.9% to 37% after 550 min). The conversion rate of Pd-Au/FeO x /Al 2 O 3 was 100% in the first 132 min, the TON of it was 205, and it declined slowly to 97.9% after 748 min. Similarly, the conversion rate of SO 2 -and H 2 S-pretreated catalyst decreased from 98.1% to 96.5% after 726 min. Thus, Pd-Au/FeO x /Al 2 O 3 maintained strong stability after SO 2 and H 2 S treatment, with negligible conversion reduction (approximately 2%) after 726 min. It is deduced that this is due to the adsorption of SO 2 and H 2 S on the catalyst surface, which inhibits the adsorption and reaction of CO.  Figure 9 shows the catalytic oxidation stability test curves of the Au/FeOx/Al2O3 and Pd-Au/FeOx/Al2O3 catalysts for CO at 25 °C before and after SO2 and H2S treatment. The conversion rate of the Au/FeOx/Al2O3 catalyst decreased from 97.9% to 52.4% after 726 min, which was slower than that of the Au/FeOx/Al2O3 catalyst pretreated with SO2 and H2S (91.9% to 37% after 550 min). The conversion rate of Pd-Au/FeOx/Al2O3 was 100% in the first 132 min, the TON of it was 205, and it declined slowly to 97.9% after 748 min. Similarly, the conversion rate of SO2and H2S-pretreated catalyst decreased from 98.1% to 96.5% after 726 min. Thus, Pd-Au/FeOx/Al2O3 maintained strong stability after SO2 and H2S treatment, with negligible conversion reduction (approximately 2%) after 726 min. It is deduced that this is due to the adsorption of SO2 and H2S on the catalyst surface, which inhibits the adsorption and reaction of CO.   Figure 10 shows the in situ FTIR spectra of Al 2 O 3 , FeO x /Al 2 O 3 , Au/FeO x /Al 2 O 3 , and Pd-Au/FeO x /Al 2 O 3 with a 2% CO mass fraction. Both the Pd-Au/FeO x /Al 2 O 3 and Au/FeO x /Al 2 O 3 catalysts exhibited peaks corresponding to -OH (3600-3700 cm −1 ), -COOH (1300-1700 cm −1 ), CO (~1900 cm −1 , 2000-2200 cm −1 ), and CO 2 (2300-2400 cm −1 ) vibrations [51,52]. -OH represents water adsorbed on the catalyst surface and reacting with the catalyst surface active oxygen species. CO ad reacted with -OH ad to produce -COOH, which adsorbed on the catalyst surface [53]. The peak intensities of -OH and -COOH on Pd-Au/FeO x /Al 2 O 3 were similar to those of Au/FeO x /Al 2 O 3 . However, for the Pd-Au/FeO x /Al 2 O 3 catalyst, CO was adsorbed on Pd 0 , Pd δ+ , Au 0 , and Au δ+ . The peak at 2077 cm −1 corresponded to CO adsorption on Pd 0 [52]. The peak of adsorbed CO on the Pd-Au/FeO x /Al 2 O 3 catalyst (2183 cm −1 ) had a higher frequency than that of Au δ+ on the Au/FeO x /Al 2 O 3 catalyst (2190 cm −1 ), possibly because of overlapping peaks [51,52,[54][55][56]. Bridge adsorption (CO B ) of CO occurs at 1957 cm −1 ; that is, CO is adsorbed on the metal centre of Au and Pd [53,57]. Based on the literature and in situ FTIR measurements, we concluded the reaction path of CO adsorbed at the active centre of the catalyst (Au 0 , Au δ+ , Pd 0 , and Pd δ+ ), reacting with adsorbed water to produce CO 2 and -COOH, while -COOH partially accumulated on the catalyst's surface and partially reacted with oxygen on the catalyst's surface to produce CO 2 and release gradually [53]. Au/FeOx/Al2O3 catalysts exhibited peaks corresponding to -OH (3600-3700 cm −1 ), -COOH (1300-1700 cm −1 ), CO (~1900 cm −1 , 2000-2200 cm −1 ), and CO2 (2300-2400 cm −1 ) vibrations [51,52]. -OH represents water adsorbed on the catalyst surface and reacting with the catalyst surface active oxygen species. COad reacted with -OHad to produce -COOH, which adsorbed on the catalyst surface [53]. The peak intensities of -OH and -COOH on Pd-Au/FeOx/Al2O3 were similar to those of Au/FeOx/Al2O3. However, for the Pd-Au/FeOx/Al2O3 catalyst, CO was adsorbed on Pd 0 , Pd δ+ , Au 0 , and Au δ+ . The peak at 2077 cm −1 corresponded to CO adsorption on Pd 0 [52]. The peak of adsorbed CO on the Pd-Au/FeOx/Al2O3 catalyst (2183 cm −1 ) had a higher frequency than that of Au δ+ on the Au/FeOx/Al2O3 catalyst (2190 cm −1 ), possibly because of overlapping peaks [51,52,[54][55][56].

Mechanistic Analysis of CO Oxidation
Bridge adsorption (COB) of CO occurs at 1957 cm −1 ; that is, CO is adsorbed on the metal centre of Au and Pd [53,57]. Based on the literature and in situ FTIR measurements, we concluded the reaction path of CO adsorbed at the active centre of the catalyst (Au 0 , Au δ+ , Pd 0 , and Pd δ+ ), reacting with adsorbed water to produce CO2 and -COOH, while -COOH partially accumulated on the catalyst's surface and partially reacted with oxygen on the catalyst's surface to produce CO2 and release gradually [53].  Figure 11 shows an in situ FTIR diagram of Au/FeOx/Al2O3 and Pd-Au/FeOx/Al2O3 after treatment with 2 ppm SO2 for 1 h and 2 ppm H2S for 1 h at ambient temperature (25 °C). The surfaces of the carrier and catalyst formed -OH (3400-3800 cm −1 ), -COOH (1400-1600 cm −1 ), H2O (1600 cm −1 ), and S-O (~1000 cm −1 ) bonds [58][59][60]. Since there was the presence of air in the reaction gas, the negative peak in the figure was CO2 in the air. During SO2 adsorption, S-O bonds were observed on the surface of both the carrier and catalyst, classified as SO4 2− [58]. The SO4 2− peak of Pd-Au/FeOx/Al2O3 was smaller than that of Au/FeOx/Al2O3, indicating a more difficult SO2 adsorption to form SO4 2− groups on the surface. After H2S treatment, the intensity of the peak corresponding to S-O vibration in Au/FeOx/Al2O3 increased significantly. In contrast, there was no significant variation in Pd-Au/FeOx/Al2O3, indicating that this catalyst had improved anti-SO2 and anti-H2S adsorption ability. However, the effect of H2S on the catalyst was not clear because the test method in this paper cannot be analysed for the time being. After sulphide treatment, the surface of the catalyst and carrier adsorbed -COOH, which was generated by the reaction of the CO component in the mixture with adsorbed water [53,57].  Figure 11 shows an in situ FTIR diagram of Au/FeO x /Al 2 O 3 and Pd-Au/FeO x /Al 2 O 3 after treatment with 2 ppm SO 2 for 1 h and 2 ppm H 2 S for 1 h at ambient temperature (25 • C). The surfaces of the carrier and catalyst formed -OH (3400-3800 cm −1 ), -COOH (1400-1600 cm −1 ), H 2 O (1600 cm −1 ), and S-O (~1000 cm −1 ) bonds [58][59][60]. Since there was the presence of air in the reaction gas, the negative peak in the figure was CO 2 in the air. During SO 2 adsorption, S-O bonds were observed on the surface of both the carrier and catalyst, classified as SO 4 2− [58]. The SO 4 2− peak of Pd-Au/FeO x /Al 2 O 3 was smaller than that of Au/FeO x /Al 2 O 3 , indicating a more difficult SO 2 adsorption to form SO 4 2− groups on the surface. After H 2 S treatment, the intensity of the peak corresponding to S-O vibration in Au/FeO x /Al 2 O 3 increased significantly. In contrast, there was no significant variation in Pd-Au/FeO x /Al 2 O 3 , indicating that this catalyst had improved anti-SO 2 and anti-H 2 S adsorption ability. However, the effect of H 2 S on the catalyst was not clear because the test method in this paper cannot be analysed for the time being. After sulphide treatment, the surface of the catalyst and carrier adsorbed -COOH, which was generated by the reaction of the CO component in the mixture with adsorbed water [53,57]. After SO2 and H2S pretreatment, air was used as the equilibration gas. A mixture with 20,000 ppm of CO was injected, and the material was characterised by in situ FTIR ( Figure  12). The vibration peak of CO increased after CO was injected into the support Au/FeOx/Al2O3, and Pd-Au/FeOx/Al2O3. In contrast, the vibration peak of CO2 (2349 cm −1 ) only increased for the Pd-Au/FeOx/Al2O3 catalyst. On the Pd-Au/FeOx/Al2O3 catalyst bridging CO was adsorbed on the active sites of Au and Pd at 1956 cm −1 [53], and the vibration peak at 2071 cm −1 corresponded to CO adsorbed on Pd 0 . The vibration peak a 2115 cm −1 corresponded to CO adsorbed on Au 0 . The vibration peak at 2173 cm −1 corresponded to Au δ+ and Pd δ+ [52,57]. Thus, a large amount of CO was adsorbed on the bimetallic active sites of the Pd-Au/FeOx/Al2O3 catalyst, reacting to produce CO2. In contrast no CO2 production was observed on the Au/FeOx/Al2O3 catalyst, possibly because the sul phate species adsorbed and accumulated on the Au/FeOx/Al2O3 catalyst inhibited the re action of CO with the active centre of the catalyst. Additionally, the peak corresponding to the vibration of -COOH adsorbed on Pd-Au/FeOx/Al2O3 was derived from the reaction of CO adsorbed on the catalyst surface with reactive oxygen species and absorbed water [53].  After SO 2 and H 2 S pretreatment, air was used as the equilibration gas. A mixture with 20,000 ppm of CO was injected, and the material was characterised by in situ FTIR ( Figure 12). The vibration peak of CO increased after CO was injected into the support, Au/FeO x /Al 2 O 3 , and Pd-Au/FeO x /Al 2 O 3 . In contrast, the vibration peak of CO 2 (2349 cm −1 ) only increased for the Pd-Au/FeO x /Al 2 O 3 catalyst. On the Pd-Au/FeO x /Al 2 O 3 catalyst, bridging CO was adsorbed on the active sites of Au and Pd at 1956 cm −1 [53], and the vibration peak at 2071 cm −1 corresponded to CO adsorbed on Pd 0 . The vibration peak at 2115 cm −1 corresponded to CO adsorbed on Au 0 . The vibration peak at 2173 cm −1 corresponded to Au δ+ and Pd δ+ [52,57]. Thus, a large amount of CO was adsorbed on the bimetallic active sites of the Pd-Au/FeO x /Al 2 O 3 catalyst, reacting to produce CO 2 . In contrast, no CO 2 production was observed on the Au/FeO x /Al 2 O 3 catalyst, possibly because the sulphate species adsorbed and accumulated on the Au/FeO x /Al 2 O 3 catalyst inhibited the reaction of CO with the active centre of the catalyst. Additionally, the peak corresponding to the vibration of -COOH adsorbed on Pd-Au/FeO x /Al 2 O 3 was derived from the reaction of CO adsorbed on the catalyst surface with reactive oxygen species and absorbed water [53]. After SO2 and H2S pretreatment, air was used as the equilibration gas. A mixture with 20,000 ppm of CO was injected, and the material was characterised by in situ FTIR ( Figure  12). The vibration peak of CO increased after CO was injected into the support, Au/FeOx/Al2O3, and Pd-Au/FeOx/Al2O3. In contrast, the vibration peak of CO2 (2349 cm −1 ) only increased for the Pd-Au/FeOx/Al2O3 catalyst. On the Pd-Au/FeOx/Al2O3 catalyst, bridging CO was adsorbed on the active sites of Au and Pd at 1956 cm −1 [53], and the vibration peak at 2071 cm −1 corresponded to CO adsorbed on Pd 0 . The vibration peak at 2115 cm −1 corresponded to CO adsorbed on Au 0 . The vibration peak at 2173 cm −1 corresponded to Au δ+ and Pd δ+ [52,57]. Thus, a large amount of CO was adsorbed on the bimetallic active sites of the Pd-Au/FeOx/Al2O3 catalyst, reacting to produce CO2. In contrast, no CO2 production was observed on the Au/FeOx/Al2O3 catalyst, possibly because the sulphate species adsorbed and accumulated on the Au/FeOx/Al2O3 catalyst inhibited the reaction of CO with the active centre of the catalyst. Additionally, the peak corresponding to the vibration of -COOH adsorbed on Pd-Au/FeOx/Al2O3 was derived from the reaction of CO adsorbed on the catalyst surface with reactive oxygen species and absorbed water [53].

DFT Calculations
Maumau T.R. et al. [61] compared the electro-oxidation activity of Pd/C, Au/C, Pd (Au/C), and Pd-Au/C catalysts to alcohols and found that Pd-Au/C catalysts were more stable and more toxic tolerant. Comparing the reaction kinetics of Pd-Au/SiO 2 and Pd/SiO 2 catalysts, Han Y.F. et al. [62] showed that the decrease in reactant adsorption on the catalyst and the enhancement of surface adsorbed oxygen/adsorbed oxygen mobility were the main reasons for the enhancement of Pd-Au/SiO 2 catalyst activity. Gao feng et al. [63] compared Pd-Au particles supported on different carriers and applied reaction kinetics analysis to find that the Pd-Au-supported catalyst was more likely to oxidize and lose activity than the Pd-Au-supported catalyst, and Pd-Au was more likely to desorb CO, which was also the main reason for the enhanced activity of the Pd-Au-supported catalyst. As shown in Figure 13, the reaction free energy curve calculated using DFT shows that *CO and *O 2 adsorption on Au/FeO x /Al 2 O 3 was very weak, with an energy change of 1.19 eV. Pd-Au/FeO x /Al 2 O 3 had a lower free energy of 0.07 eV. Regarding the transition state, the energy barriers corresponding to *CO to *CO 2 oxidation by Au/FeO x /Al 2 O 3 and Pd-Au/FeO x /Al 2 O 3 catalysts were 1.12 eV and 0.65 eV, respectively, indicating that Pd-Au/FeO x /Al 2 O 3 is thermodynamically more likely to oxidise CO. Maumau T.R. et al. [61] compared the electro-oxidation activity of Pd/C, Au/C, Pd (Au/C), and Pd-Au/C catalysts to alcohols and found that Pd-Au/C catalysts were more stable and more toxic tolerant. Comparing the reaction kinetics of Pd-Au/SiO2 and Pd/SiO2 catalysts, Han Y.F. et al. [62] showed that the decrease in reactant adsorption on the catalyst and the enhancement of surface adsorbed oxygen/adsorbed oxygen mobility were the main reasons for the enhancement of Pd-Au/SiO2 catalyst activity. Gao feng et al. [63] compared Pd-Au particles supported on different carriers and applied reaction kinetics analysis to find that the Pd-Au-supported catalyst was more likely to oxidize and lose activity than the Pd-Au-supported catalyst, and Pd-Au was more likely to desorb CO, which was also the main reason for the enhanced activity of the Pd-Au-supported catalyst. As shown in Figure 13, the reaction free energy curve calculated using DFT shows that *CO and *O2 adsorption on Au/FeOx/Al2O3 was very weak, with an energy change of 1.19 eV. Pd-Au/FeOx/Al2O3 had a lower free energy of 0.07 eV. Regarding the transition state, the energy barriers corresponding to *CO to *CO2 oxidation by Au/FeOx/Al2O3 and Pd-Au/FeOx/Al2O3 catalysts were 1.12 eV and 0.65 eV, respectively, indicating that Pd-Au/FeOx/Al2O3 is thermodynamically more likely to oxidise CO. Finally, the antivulcanisation of Au/FeOx/Al2O3 and Pd-Au/FeOx/Al2O3 was investigated. The adsorption energy diagrams of H2S and SO2 were obtained from the difference in SO2 and H2S adsorption energies ( Figure 14). Higher adsorption energy corresponded to more difficult SO2 and H2S adsorption. The adsorption energy of Pd-Au/FeOx/Al2O3 was positive for both H2S and SO2, whereas that of Au/FeOx/Al2O3 for H2S was negative. The adsorption energy of Au/FeOx/Al2O3 for SO2 was also significantly lower than that of Pd-Au/FeOx/Al2O3. Thus, Pd-Au/FeOx/Al2O3 has a weak adsorption capacity for SO2 and H2S and a more substantial anti-SO2 and anti-H2S effect, indicative of its long-term stability in a sulphur-containing environment. Finally, the antivulcanisation of Au/FeO x /Al 2 O 3 and Pd-Au/FeO x /Al 2 O 3 was investigated. The adsorption energy diagrams of H 2 S and SO 2 were obtained from the difference in SO 2 and H 2 S adsorption energies ( Figure 14). Higher adsorption energy corresponded to more difficult SO 2 and H 2 S adsorption. The adsorption energy of Pd-Au/FeO x /Al 2 O 3 was positive for both H 2 S and SO 2 , whereas that of Au/FeO x /Al 2 O 3 for H 2 S was negative. The adsorption energy of Au/FeO x /Al 2 O 3 for SO 2 was also significantly lower than that of Pd-Au/FeO x /Al 2 O 3 . Thus, Pd-Au/FeO x /Al 2 O 3 has a weak adsorption capacity for SO 2 and H 2 S and a more substantial anti-SO 2 and anti-H 2 S effect, indicative of its long-term stability in a sulphur-containing environment.

Conclusions
In this study, a Pd-Au/FeOx/Al2O3 catalyst with an Au:Pd ratio of 2:1 (wt%) was prepared using ultrasonic-assisted impregnation and urea-assisted deposition-precipitation methods, and its performance was compared with that of Au/FeOx/Al2O3. XRD and TEM analyses showed that the prepared Au and Pd nanoparticles were evenly dispersed on the

Conclusions
In this study, a Pd-Au/FeO x /Al 2 O 3 catalyst with an Au:Pd ratio of 2:1 (wt%) was prepared using ultrasonic-assisted impregnation and urea-assisted deposition-precipitation methods, and its performance was compared with that of Au/FeO x /Al 2 O 3 . XRD and TEM analyses showed that the prepared Au and Pd nanoparticles were evenly dispersed on the carrier. Mapping and XEDS analyses showed that the Pd-Au/FeO x /Al 2 O 3 catalyst adsorbed less SO 2 and H 2 S after SO 2 and H 2 S pretreatment and had stronger SO 2 and H 2 S resistance than Au/FeO x /Al 2 O 3 . The Pd-Au/FeO x /Al 2 O 3 -catalysed CO conversion was measured before and after SO 2 and H 2 S pretreatment, revealing that the Pd-Au/FeO x /Al 2 O 3 catalyst could fully convert 2500 ppm of CO at 25 • C (room temperature) after SO 2 and H 2 S pretreatment. The conversion rate remained at 100% at higher temperatures. Stability tests of the two catalysts before and after SO 2 and H 2 S pretreatment showed that the conversion rate of Pd-Au/FeO x /Al 2 O 3 with 20,000 ppm of CO was 98.1% at 25 • C (ambient temperature) after SO 2 and H 2 S pretreatment, with a less than 2% decrease after 726 min. In situ FTIR analysis of the Pd-Au/FeO x /Al 2 O 3 catalyst showed that CO can be adsorbed on Pd and Au as well as bridging the two metals. The roles of the two catalysts in catalytic CO oxidation were studied using DFT calculations. The results showed that the conversion to CO 2 on Pd-Au/FeO x /Al 2 O 3 required less energetic CO and O 2 species. By comparing the adsorption energies of SO 2 and H 2 S on the two catalysts, we concluded that the Pd-Au/FeO x /Al 2 O 3 catalyst was less susceptible to the activity-reducing SO 2 and H 2 S effect. Thus, the addition of Pd to Au/FeO x /Al 2 O 3 catalysts resulted in more robust catalytic CO oxidation activity and improved anti-SO 2 and anti-H 2 S stability. Pd-Au/FeO x /Al 2 O 3 has the potential for wide use in the treatment of industrial flue gases.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ma16103755/s1, Table S1: Au and Pd mass loading of catalysts; Figure S1: CO conversion of Pd-Au/FeO x /Al 2 O 3 catalysts with different mass loading ratios; Figure