Catalytic Oxidative Decomposition of Dimethyl Methyl Phosphonate over CuO/CeO 2 Catalysts Prepared Using a Secondary Alkaline Hydrothermal Method

: Bimetallic synergism plays an important role in lattice-doped catalysts. Therefore, lattice-doped bimetallic CuO/CeO 2 catalysts were prepared by secondary alkaline hydrothermal reaction. During this process, the CeO 2 nanomaterials were partially dissolved and recrystallized; thus, Cu ions were doped into the CeO 2 lattice. The physical and chemical properties of CeO 2 , CuO/CeO 2 , and CuO were investigated. H 2 temperature-programmed reduction characterization showed that the oxidation activity of CuO/CeO 2 was signiﬁcantly improved. X-ray photoelectron spectroscopy results showed that electron transfer occurred between Ce and Cu in the CuO/CeO 2 catalyst. Additionally, Raman characterization conﬁrmed the strong interaction between Cu and Ce. After CuO was loaded, the thermal catalytic decomposition performance of the catalyst was signiﬁcantly improved with respect to the sarin simulant dimethyl methyl phosphonate (DMMP); with an increase in the Cu/Ce ratio, the performance ﬁrst strengthened and then weakened. Additionally, the reaction tail gas and catalyst surface products were analyzed using mass spectrometry and ion chromatography, and the changes in the surface products during the thermal catalytic decomposition of DMMP were characterized at different temperatures using in situ diffuse reﬂectance infrared Fourier transform spectroscopy. Finally, the catalytic reaction pathways of DMMP on CeO 2 , CuO/CeO 2 , and CuO were inferred. The study results not only demonstrate an effective catalyst for the removal of nerve agent but also a feasible preparation method for lattice-doped bimetallic catalysts in the ﬁeld of environmental protection. Ce 3+ content decreased after loading CuO. It has been speculated that a reversible reaction of Ce 3+ + Cu 2+ ↔ Ce 4+ + Cu + and 2Ce 3+ + Cu 2+ ↔ 2Ce 4+ + Cu 0 occurs on the surface of the CuO/CeO 2 catalyst [42–44].


Introduction
The threat posed by chemical warfare agents (CWAs) in terrorist incidents, wars, and conflicts has not yet been addressed. Research on the degradation of CWAs is crucial for human health and national security [1]. Traditional and currently used protection methods involve adsorption through carbon-based materials. However, the adsorption capacity of carbon-based materials is limited, and weak adsorption can lead to secondary pollution [2,3]. Recent studies have shown that the catalytic oxidation and decomposition of CWAs by metal oxides is an effective method, with the advantage of destroying the molecular structure of CWAs and generating low-toxicity or non-toxic products [4][5][6][7][8]. From the perspective of CWAs treatment, protection time, which specifically refers to the time required for complete conversion of CWAs, is an important parameter for evaluating catalyst performance [9]. Sarin, a typical nerve CWA, has been repeatedly used in military conflicts and terrorist attacks, causing massive casualties of troops and innocent civilians [2]. Therefore, it is significant to study the catalytic degradation of sarin. Sarin is an organic phosphate ester. The chemical structure of sarin is shown in Figure 1a. Its poisoning mechanism involves the combination of organophosphate ester with the acetylcholinesterase (AchE) to form a stable phosphoryl cholinesterase, which inhibits the activity of the enzyme and paralyzes the nervous system [1,10]. The poisoning mechanism of sarin is similar to that of organophosphorus pesticides. Dimethyl methyl phosphonate (DMMP) is often used as a simulant for organophosphorus pesticides [11]. Owing to its similar molecular structure (Figure 1b), low toxicity, and similar poisoning mechanism, DMMP is also frequently used as a simulant for sarin [12][13][14].
Sarin, a typical nerve CWA, has been repeatedly used in rorist attacks, causing massive casualties of troops and innocen is significant to study the catalytic degradation of sarin. Sarin ester. The chemical structure of sarin is shown in Figure 1a. Its volves the combination of organophosphate ester with the acet form a stable phosphoryl cholinesterase, which inhibits the a paralyzes the nervous system [1,10]. The poisoning mechanism of organophosphorus pesticides. Dimethyl methyl phosphonat a simulant for organophosphorus pesticides [11]. Owing to its s (Figure 1b), low toxicity, and similar poisoning mechanism, DM as a simulant for sarin [12][13][14]. Metal oxides, such as Al2O3 [15], MgO [16], Y2O3 [7], TiO2 [1 CeO2 [21,22], and others, have a good P−OCH3 bond-breaking a for DMMP decomposition. The decomposition products of DM oxides include methanol, dimethyl ether, and methylphosphon ucts on variable metal oxides can be further oxidized to CO, CO son for this is that the redox couple of the metal ions in the re energy reaction path for the oxidation reaction.
Variable-valence CeO2 has an excellent oxygen storage/re reversibly changed between Ce 3+ and Ce 4+ , accompanied by th tion of oxygen vacancies, making it widely applicable in het Chen et al. reported that DMMP completely decomposed on heating, forming products of methanol, formaldehyde, CO, an the interaction of DMMP with CeO2 on different surfaces using [22]. Loading transition metal oxides onto CeO2 nanomaterials further improve the catalytic activity of the catalysts. CuO/CeO studied two-component, non-noble metal catalysts, and the s CuO and CeO2 promotes electron transfer between Cu and Ce tion in CeO2. Presently, CuO/CeO2 catalysts are commonly use Metal oxides, such as Al 2 O 3 [15], MgO [16], Y 2 O 3 [7], TiO 2 [17], Fe 2 O 3 [18], CuO [19,20], CeO 2 [21,22], and others, have a good P-OCH 3 bond-breaking ability and are widely used for DMMP decomposition. The decomposition products of DMMP on monovalent metal oxides include methanol, dimethyl ether, and methylphosphonic acid, whereas the products on variable metal oxides can be further oxidized to CO, CO 2 , and PO X . The main reason for this is that the redox couple of the metal ions in the reaction can provide a low-energy reaction path for the oxidation reaction.
Variable-valence CeO 2 has an excellent oxygen storage/release capacity and can be reversibly changed between Ce 3+ and Ce 4+ , accompanied by the generation and elimination of oxygen vacancies, making it widely applicable in heterogeneous catalysis [23]. Chen et al. reported that DMMP completely decomposed on ceria films using vacuum heating, forming products of methanol, formaldehyde, CO, and H 2 [21]. Li et al. studied the interaction of DMMP with CeO 2 on different surfaces using density functional theory [22]. Loading transition metal oxides onto CeO 2 nanomaterials is an effective method to further improve the catalytic activity of the catalysts. CuO/CeO 2 is one of the most widely studied twocomponent, non-noble metal catalysts, and the strong interaction between CuO and CeO 2 promotes electron transfer between Cu and Ce, as well as oxygen migration in CeO 2 . Presently, CuO/CeO 2 catalysts are commonly used in CO oxidation and hydrogenation reactions, VOC degradation, water gas shift reactions, and CO 2 hydrogenation reactions [24][25][26][27][28]. However, CuO/CeO 2 catalysts have not yet been studied for the catalytic oxidation of DMMP. Therefore, the performance and reaction mechanism of the CuO/CeO 2 catalyst for the thermal catalytic decomposition of DMMP were comprehensively investigated in this study.
Traditional preparation methods for CuO/CeO 2 include the one-time hydrothermal, impregnation, and deposition precipitation methods [29][30][31][32]. George et al. prepared a CuO/CeO 2 catalyst using ammonia solution and acid solution post-synthesis modification methods and reported significantly improved catalytic CO activity [33,34]. Zhou et al. prepared highly dispersed CuO/CeO 2 catalysts using a secondary solvothermal method (ethanol as the solvent) to mix a CeO 2 support with Cu(NO 3 ) 2 dissolved in an ethanol solution, then dried and calcined it after the reaction [35]. Inspired by the above method, solid CeO 2 was mixed with a Cu(NO 3 ) 2 solution, and a given concentration of NaOH solution was added for the secondary alkaline hydrothermal method. The addition of NaOH resulted in the dissolution and recrystallization of CeO 2 during the reaction. During this process, Cu was doped into the CeO 2 lattice, which increased the interaction between Cu and Ce, promoted electron transfer between Cu and Ce, and formed a Cu-V-Ce (V is donated as oxygen vacancy) structure. Thus, the performance of the CuO/CeO 2 catalyst further improved.
The traditional impregnation and deposition precipitation methods have low utilization rates of raw materials, and when the loading amount increased, serious agglomeration occurs. In our study, high-loading and uniform-distribution CuO were prepared using the secondary alkaline hydrothermal method. This method is characterized by a high utilization rate of raw materials. According to the ICP results obtained in the present study, the theoretical loading of CuO is comparable to actual loading.
In this study, three types of CeO 2 , CuO/CeO 2 , and CuO catalysts were prepared. First, a CeO 2 support was prepared using a hydrothermal method. Secondly, CuO/CeO 2 catalysts with varying Cu/Ce ratios were prepared using the CeO 2 support mixed with Cu(NO 3 ) 2 solution and NaOH solution by the alkaline hydrothermal method. In order to study the changes in physicochemical properties and catalytic performance after CeO 2 -supported CuO, CuO was also prepared by a hydrothermal method for comparison. The structure was characterized using X-ray diffraction (XRD) and transmission electron microscopy (TEM). The surface elemental distribution, redox properties, and surface intermediate products were analyzed using X-ray photoelectron spectroscopy (XPS), H 2 -temperature-programmed reduction (H 2 -TPR), O 2 -temperature-programmed desorption (O 2 -TPD), and in situ diffuse reflectance infrared Fourier transform spectroscopy (in situ DRIFTS). The reaction path was deduced for the thermal decomposition of DMMP on the catalyst. In this study, CuO/CeO 2 catalysts with high loading and uniform distribution were synthesized, and the initial morphology of CuO was not changed. The dissolution and recrystallization of CeO 2 in the alkali-thermal process promoted the doping of Cu into the CeO 2 lattice, and bimetallic synergism enhanced the catalytic performance. The results of this study provide a reference for the design and optimization of CWA decomposition catalysts.

Characteristics of CuO/CeO 2 Catalysts
The as-prepared materials were characterized using XRD to identify crystal phases ( Figure 2). The Figure 2 shows that the CeO 2 and CuO diffraction patterns match the characteristic peaks of the fluorite structure of CeO 2 (JCPDS 34-0394) and the monoclinic structure of CuO (JCPDS 80-1916), respectively. The intensity of the CuO diffraction peaks gradually increased with increased Cu/Ce ratio, and some CeO 2 diffraction peaks disappeared. When the Cu/Ce ratio was 10%, weak diffraction peaks appeared at 36.49 • and 38.68 • , corresponding to the (−111) and (111) crystal planes of the CuO diffraction peaks, respectively. This phenomenon may have been caused by the low loading of CuO. When the Cu/Ce ratio was 50%, the intensities of the diffraction peaks of the (−111) and (111) crystal planes increased, and new diffraction peaks appeared at 48.6 • and 61.6 • , corresponding to the (−202) and (−113) crystal planes of CuO, respectively. The crystalline size of the catalyst was calculated using the Scherrer equation, as shown in Table 1. Owing to the smaller Cu diffraction peaks in 10%Cu/Ce and 20%Cu/Ce, the crystalline size of CuO could not be calculated. Table 1 shows that the grain sizes of CeO 2 and CuO prepared using the secondary alkaline method did not change significantly. The actual ratio of Cu/Ce determined by ICP-OES was higher than the theoretically calculated value (Table 1), indicating that CeO2 was partially dissolved in the secondary alkaline hydrothermal environment. The specific surface area (SBET) of the samples was determined using the N2 adsorption-desorption method. As shown in Table 1, the SBET of the CuO/CeO2 catalyst decreased, mainly due to the formation of sheet-morphology CuO (Figure 3f). The SBET of sheet-morphology CuO is significantly lower than that of CeO2 (Table 1). Therefore, the SBET of the CuO/CeO2 catalyst decreased with increased in CuO doping. Moreover, the SBET of 50% Cu/Ce and 80% Cu/Ce catalysts decrease more significantly compared with that of 10% Cu/Ce and 20%Cu/Ce catalysts as a result of the appearance of a large amount of sheet-morphology CuO (Figure 3d,e) on CeO2 nanorods.   The actual ratio of Cu/Ce determined by ICP-OES was higher than the theoretically calculated value (Table 1), indicating that CeO 2 was partially dissolved in the secondary alkaline hydrothermal environment. The specific surface area (S BET ) of the samples was determined using the N 2 adsorption-desorption method. As shown in Table 1, the S BET of the CuO/CeO 2 catalyst decreased, mainly due to the formation of sheet-morphology CuO (Figure 3f). The S BET of sheet-morphology CuO is significantly lower than that of CeO 2 (Table 1). Therefore, the S BET of the CuO/CeO 2 catalyst decreased with increased in CuO doping. Moreover, the S BET of 50% Cu/Ce and 80% Cu/Ce catalysts decrease more significantly compared with that of 10% Cu/Ce and 20%Cu/Ce catalysts as a result of the appearance of a large amount of sheet-morphology CuO (Figure 3d,e) on CeO 2 nanorods. Figure 3 shows a TEM image of the CuO/CeO 2 catalyst. Figure 3a,f shows that the morphologies of pure CeO 2 and CuO were nanorods and nanoplates, respectively. Figure 3b-e shows TEM images of the CuO/CeO 2 catalysts with Cu/Ce ratios of 10, 20, 50, and 80%, respectively. The morphology of the CeO 2 nanorods did not change after loading with CuO, and when the Cu/Ce ratio was 50 and 80%, CuO nanoplates began to appear (Figure 3d,e). Figure 4 shows EDX images of the Cu/Ce catalysts with varying loading ratios, demonstrating that Cu was evenly distributed on the CeO nanorods (Figure 4a the CuO/CeO2 catalyst decreased, mainly due to the formation of sheet-morphology CuO (Figure 3f). The SBET of sheet-morphology CuO is significantly lower than that of CeO (Table 1). Therefore, the SBET of the CuO/CeO2 catalyst decreased with increased in CuO doping. Moreover, the SBET of 50% Cu/Ce and 80% Cu/Ce catalysts decrease more signifi cantly compared with that of 10% Cu/Ce and 20%Cu/Ce catalysts as a result of the ap pearance of a large amount of sheet-morphology CuO (Figure 3d,e) on CeO2 nanorods.     Figure 3 shows a TEM image of the CuO/CeO2 catalyst. Figure 3a,f shows that th morphologies of pure CeO2 and CuO were nanorods and nanoplates, respectively. Figur 3b-e shows TEM images of the CuO/CeO2 catalysts with Cu/Ce ratios of 10, 20, 50, and 80%, respectively. The morphology of the CeO2 nanorods did not change after loading with CuO, and when the Cu/Ce ratio was 50 and 80%, CuO nanoplates began to appea (Figure 3d,e). Figure 4 shows EDX images of the Cu/Ce catalysts with varying loading ratios, demonstrating that Cu was evenly distributed on the CeO nanorods (Figure 4a Figure 5 shows HRTEM images of the catalysts. The exposed (111) and (200) crysta planes of the CeO2 nanorods correspond to the lattice fringes at 0.31 and 0.28 nm, respec tively ( Figure 5a); the exposed (020) crystal planes of the CuO nanoplates correspond to  Figure 5 shows HRTEM images of the catalysts. The exposed (111) and (200) crystal planes of the CeO 2 nanorods correspond to the lattice fringes at 0.31 and 0.28 nm, respectively ( Figure 5a); the exposed (020) crystal planes of the CuO nanoplates correspond to the lattice fringes at 0.17 nm (Figure 5f). Figure 4c-e shows that a lattice spacing of 0.29 nm appeared with increasing CuO load due to the change in lattice spacing caused by the doping of Cu atoms into the CeO 2 lattice.  The oxidation activity of the catalyst material was characterized using H2−TPR (Fig  ure 6a and Table 2). There are usually two reduction peaks of CeO2: a peak at approxi mately 675 °C corresponding to the surface oxygen reduction peak of CeO2 and a peak a approximately 835 °C corresponding to the bulk oxygen reduction peak of CeO2 [36]. A broad peak appeared in CuO at 170-500 °C, which may be due to the closeness of the surface oxygen and bulk oxygen reduction peaks of CuO, such that the two peaks canno be completely separated. CuO/CeO2 has three main peaks at 170-202 °C, 250 °C, and 275 °C, which are denoted by α, β, and γ, respectively. The α peak corresponds to CuOx, which interacts strongly with CeO2; the β peak corresponds to CuOx, which interacts weakly with CeO2; and the γ peak is attributed to bulk CuO [37]. The H2−TPR diagram ( Figure 6) clearly shows that after CuO was loaded onto CeO2, the reduction temperature of the CuO/CeO catalyst was significantly lower than that of the CeO2, indicating that the loaded catalys had more active oxygen. H2 consumption amounts of as-prepared catalyst were calculated using H2−TPR ( Table 2). As shown in Table 2, the theoretical and the calculated value o H2 consumption increased with an increase in the Cu/Ce loading ratio. The calculated hy drogen consumption amount of 10%Cu/Ce catalyst exceeded theoretical values; however with an increase in the Cu/Ce ratio, the theoretical calculated value exceeded the calcu lated value, implying a interaction between CeO2 and CuO species [38]. Evidently, the addition of CuO improved the CeO2 oxidation activity. The oxidation activity of the catalyst material was characterized using H 2 -TPR ( Figure 6a and Table 2). There are usually two reduction peaks of CeO 2 : a peak at approximately 675 • C corresponding to the surface oxygen reduction peak of CeO 2 and a peak at approximately 835 • C corresponding to the bulk oxygen reduction peak of CeO 2 [36]. A broad peak appeared in CuO at 170-500 • C, which may be due to the closeness of the surface oxygen and bulk oxygen reduction peaks of CuO, such that the two peaks cannot be completely separated. CuO/CeO 2 has three main peaks at 170-202 • C, 250 • C, and 275 • C, which are denoted by α, β, and γ, respectively. The α peak corresponds to CuO x , which interacts strongly with CeO 2 ; the β peak corresponds to CuO x , which interacts weakly with CeO 2 ; and the γ peak is attributed to bulk CuO [37]. The H 2 -TPR diagram ( Figure 6) clearly shows that after CuO was loaded onto CeO 2 , the reduction temperature of the CuO/CeO 2 catalyst was significantly lower than that of the CeO 2 , indicating that the loaded catalyst had more active oxygen. H 2 consumption amounts of as-prepared catalyst were calculated using H 2 -TPR ( Table 2). As shown in Table 2, the theoretical and the calculated value of H 2 consumption increased with an increase in the Cu/Ce loading ratio. The calculated hydrogen consumption amount of 10%Cu/Ce catalyst exceeded theoretical values; however, with an increase in the Cu/Ce ratio, the theoretical calculated value exceeded the calculated value, implying a interaction between CeO 2 and CuO species [38]. Evidently, the addition of CuO improved the CeO 2 oxidation activity. O2−TPD tests were performed on the CuO/CeO2 catalysts, CeO2, and CuO gate the desorption behavior of oxygen ( Figure 6b and Table 3). Two types of o in the catalyst, namely surface-adsorbed oxygen and lattice oxygen [39]. The °C is related to chemisorbed oxygen around the oxygen vacancies, and the pea is related to oxygen escaping from the lattice. As shown in Table 3, with an inc Cu/Ce ratio, the amount of oxygen adsorbed at low temperatures decreased, tice oxygen at high temperatures increased.    Figure 6b and Table 3). Two types of oxygen exist in the catalyst, namely surface-adsorbed oxygen and lattice oxygen [39]. The peak at 180 • C is related to chemisorbed oxygen around the oxygen vacancies, and the peak at 640 • C is related to oxygen escaping from the lattice. As shown in Table 3, with an increase in the Cu/Ce ratio, the amount of oxygen adsorbed at low temperatures decreased, and the lattice oxygen at high temperatures increased. XPS characterizations were used to obtain the composition information and determine the chemical state on the surface of CeO 2 , CuO, and CuO/CeO 2 . Figure 7 shows the Ce 3d XPS spectra of the CeO 2 and CuO/CeO 2 catalysts, which are deconvoluted into ten peaks [36,40,41]. The Ce 3d spectra are composed of two-group spin orbitals of the overlapping peaks, labeled as U 0 -U for 3d 3/2 and V 0 -V for 3d 5/2 . The binding energy peaks of V (882.14 eV), V (888.51 eV), V (898.05 eV), U (900.63 eV), U (907.32 eV), and U (916.45 eV) are related to Ce 4+ , whereas the binding energy peaks of V 0 (880.2 eV), V (884.7 eV), U 0 (898.65 eV), and U (903.09 eV) are associated with Ce 3+ . The Ce 3+ content was calculated using the peak area ratio of Ce 3+ /(Ce 3+ + Ce 4+ ), and the results are listed in Table 4. The Ce 3+ content decreased after loading CuO. It has been speculated that a reversible reaction of Ce 3+ + Cu 2+ ↔ Ce 4+ + Cu + and 2Ce 3+ + Cu 2+ ↔ 2Ce 4+ + Cu 0 occurs on the surface of the CuO/CeO 2 catalyst [42][43][44]. XPS characterizations were used to obtain the composition information and deter mine the chemical state on the surface of CeO2, CuO, and CuO/CeO2. Figure 7 shows the Ce 3d XPS spectra of the CeO2 and CuO/CeO2 catalysts, which are deconvoluted into ten peaks [36,40,41]. The Ce 3d spectra are composed of two-group spin orbitals of the over lapping peaks, labeled as U0-U``` for 3d3/2 and V0-V``` for 3d5/2. The binding energy peaks of V (882.14 eV), V`` (888.51 eV), V``` (898.05 eV), U (900.63 eV), U`` (907.32 eV), and U`` (916.45 eV) are related to Ce 4+ , whereas the binding energy peaks of V0 (880.2 eV), V` (884.7 eV), U0 (898.65 eV), and U` (903.09 eV) are associated with Ce 3+ . The Ce 3+ content was cal culated using the peak area ratio of Ce 3+ /(Ce 3+ + Ce 4+ ), and the results are listed in Table 4 The Ce 3+ content decreased after loading CuO. It has been speculated that a reversible re action of Ce 3+ + Cu 2+ ↔ Ce 4+ + Cu + and 2Ce 3+ + Cu 2+ ↔ 2Ce 4+ + Cu 0 occurs on the surface o the CuO/CeO2 catalyst [42][43][44].  Figure 8 shows the Cu 2p spectra and Cu LMM Auger spectra of the CuO/CeO2 cat alysts, where two characteristic peaks are visible, namely Cu 2p3/2 and Cu 2p1/2, along with a broad satellite peak in the Cu 2p spectra. The peaks centered at 934 and 954 eV corre spond to Cu 2p3/2 and Cu 2p1/2, respectively, and the broad satellite peak is in the range o   Figure 8 shows the Cu 2p spectra and Cu LMM Auger spectra of the CuO/CeO 2 catalysts, where two characteristic peaks are visible, namely Cu 2p 3/2 and Cu 2p 1/2 , along with a broad satellite peak in the Cu 2p spectra. The peaks centered at 934 and 954 eV correspond to Cu 2p 3/2 and Cu 2p 1/2 , respectively, and the broad satellite peak is in the range of 937-947 eV. The peaks located at 934 eV and the corresponding satellite peaks are associated with Cu 2+ , and the peaks located at 932-933 eV are associated with Cu + or/and Cu 0 [38,45]. The two reduced states (Cu + and Cu 0 ) are difficult to distinguish in the Cu 2p spectra [46][47][48][49]. In order to distinguish Cu + and Cu 0 , Auger Cu LMM spectra were used. The broad feature in the kinetic energy spectra consists of three peaks at 917.5 eV, 913.2-916.0 eV, and 918.5 eV-919 eV, which are assigned to Cu 2+ , Cu + , and Cu 0 species, respectively [45,[50][51][52]. Although the Cu LMM Auger spectra show that both Cu 0 and Cu + occur simultaneously in the prepared catalysts, the presence of Cu 0 species possible due to the photoreduction in the XPS test [37,49]. Table 4 presents the calculated Cu 0 and Cu + to Cu 2+ ratios; it is clear that (Cu 0 + Cu + )/Cu 2+ is related to the change in Ce 3+ content (∆Ce 3+ ), which further indicates that reversible reactions of Ce 3+ + Cu 2+ ↔ Ce 4+ + Cu + and 2Ce 3+ + Cu 2+ ↔ 2Ce 4+ + Cu 0 occurred on the surface of the CuO/CeO 2 catalyst. The (Cu 0 + Cu + )/Cu 2+ of 20%Cu/Ce, 50%Cu/Ce, and 80%Cu/Ce were nearly the same and correspond to a minimal change in ∆Ce 3+ . Figure 9 shows that the O1s spectra of the CeO 2 , CuO, and CuO/CeO 2 catalysts were deconvoluted into two distinct peaks [36]. The low-binding-energy peak located at 528.5-529.5 eV (denoted as Oα) was attributed to surface lattice oxygen. The higher-bindingenergy peak located at 530-531.5 eV (denoted as Oβ) was attributed to oxygen vacancies, surface-adsorbed O 2− /O − , surface-adsorbed O 2 , hydroxyl groups, and carbonates [39,53]. The Oα and Oβ contents were calculated based on the areas of the deconvoluted peaks (Table 4). With an increase in the Cu/Ce ratio, the Oβ content decreased, whereas the Oα content increased; this trend is consistent with the O 2 -TPD results.
Visible Raman spectroscopy provides additional information about the lattice vibrations and internal defects of catalysts. Figure 10 shows that a strong band was detected at 450-460 cm −1 , which was ascribed to the Raman-active vibrational F 2g mode of CeO 2 with a fluorite-type structure. Another broad peak located at approximately 598 cm −1 was also detected (Figure 10 inset), which corresponds to the defect-induced (D) mode associated with O vacancies caused by the presence of Ce 3+ [36]. The F 2g mode of CeO 2 shifted from 463 to 445 cm −1 for the CuO/CeO 2 catalysts, which is generally attributed to the shrinkage and distortion of the CeO 2 lattice caused by the electronic redistribution of the Cu-V-Ce structure. The ratio of the two peak intensities between the defect-induced (D) mode and the F 2g mode (I D /I F2g ) represents the relative concentration of O vacancies (Table 4) [54,55]. The I D /I F2g first increased and then decreased with an increase in the Cu/Ce ratio, and 50%Cu/Ce exhibited the highest O vacancy concentration of 0.083. photoreduction in the XPS test [37,49]. Table 4 presents the calculated Cu 0 and Cu + to ratios; it is clear that (Cu 0 +Cu + )/Cu 2+ is related to the change in Ce 3+ content (ΔCe 3+ ), w further indicates that reversible reactions of Ce 3+ + Cu 2+ ↔ Ce 4+ + Cu + and 2Ce 3+ + Cu 2Ce 4+ + Cu 0 occurred on the surface of the CuO/CeO2 catalyst. The (Cu 0 +Cu + )/Cu 20%Cu/Ce, 50%Cu/Ce, and 80%Cu/Ce were nearly the same and correspond to a min change in ΔCe 3+ .  Figure 9 shows that the O1s spectra of the CeO2, CuO, and CuO/CeO2 catalysts w deconvoluted into two distinct peaks [36]. The low-binding-energy peak located at 52 529.5 eV (denoted as Oα) was attributed to surface lattice oxygen. The higher-bind energy peak located at 530-531.5 eV (denoted as Oβ) was attributed to oxygen vacan surface-adsorbed O 2− /O − , surface-adsorbed O2, hydroxyl groups, and carbonates [39 The Oα and Oβ contents were calculated based on the areas of the deconvoluted p (Table 4). With an increase in the Cu/Ce ratio, the Oβ content decreased, whereas the content increased; this trend is consistent with the O2−TPD results. Visible Raman spectroscopy provides additional information about the lattice vibra tions and internal defects of catalysts. Figure 10 shows that a strong band was detected a 450-460 cm −1 , which was ascribed to the Raman-active vibrational F2g mode of CeO2 with a fluorite-type structure. Another broad peak located at approximately 598 cm −1 was also detected (Figure 10 inset), which corresponds to the defect-induced (D) mode associated with O vacancies caused by the presence of Ce 3+ [36]. The F2g mode of CeO2 shifted from with O vacancies caused by the presence of Ce 3+ [36]. The F2g mode of CeO2 shifted from 463 to 445 cm −1 for the CuO/CeO2 catalysts, which is generally attributed to the shrinkage and distortion of the CeO2 lattice caused by the electronic redistribution of the Cu−V−Ce structure. The ratio of the two peak intensities between the defect-induced (D) mode and the F2g mode (ID/IF2g) represents the relative concentration of O vacancies (Table 4) [54,55]. The ID/IF2g first increased and then decreased with an increase in the Cu/Ce ratio, and 50%Cu/Ce exhibited the highest O vacancy concentration of 0.083. Figure 10. Raman spectra of the CeO2-supported CuO catalysts. Figure 10. Raman spectra of the CeO 2 -supported CuO catalysts.

Catalytic Performance Testing
Protection time is an important parameter for the evaluation of catalyst performance [9]. Figure 11 presents the protection times of the six catalysts against DMMP, as shown in Table 3. The protection times of the catalysts were ranked in the following order: 50%Cu/Ce (322 min) > 20%Cu/Ce (266 min) > 10%Cu/Ce (252 min) > 80%Cu/Ce (140 min) > CeO 2 (140 min) > CuO (56 min). The protection time of the CuO/CeO 2 catalyst was significantly improved after loading with CuO. The protection time first increased and then decreased with an increasing Cu/Ce ratio, and 50%Cu/Ce exhibited the longest protection time. The variation trend of protection time is the same as that of oxygen vacancy concentration in Raman, indicating that the importance of oxygen vacancies in the catalytic DMMP reaction. A comparison of protection time on various reported catalysts is summarized in Table 5.

Catalytic Performance Testing
Protection time is an important parameter for the evaluation of catalyst performance [9]. Figure 11 presents the protection times of the six catalysts against DMMP, as shown in Table 3. The protection times of the catalysts were ranked in the following order 50%Cu/Ce (322 min) > 20%Cu/Ce (266 min) > 10%Cu/Ce (252 min) > 80%Cu/Ce (140 min) > CeO2 (140 min) > CuO (56 min). The protection time of the CuO/CeO2 catalyst was significantly improved after loading with CuO. The protection time first increased and then decreased with an increasing Cu/Ce ratio, and 50%Cu/Ce exhibited the longest protection time. The variation trend of protection time is the same as that of oxygen vacancy concentration in Raman, indicating that the importance of oxygen vacancies in the catalytic DMMP reaction. A comparison of protection time on various reported catalysts is summarized in Table 5.

Catalyst
Reaction Condition Protection Time Reference

Reaction Mechanism for the Thermocatalytic Decomposition of DMMP
50%Cu/Ce catalyst with the longest protection time, CeO 2 , and CuO were selected to investigate the DMMP thermocatalytic decomposition mechanism. After the protectiontime test, XPS analyses were performed on the deactivated catalyst; the results are shown in Figure 12. A P2p peak was obvious on the spent catalysts, indicating that the deactivated catalyst surface was covered with P species. The O1s spectrum of the deactivated catalyst was fitted to four peaks, namely surface lattice oxygen (Oα), P-O, adsorbed oxygen (O Ads ), and adsorbed water/hydroxyl oxygen [19]. The appearance of the P-O peak indicates the formation of P-O byproducts on the catalyst surface. Oα contents before and after the catalytic decomposition of DMMP were calculated (Table 6). It is obvious that the proportion of Oα in the deactivated catalyst was significantly lower that of the fresh catalyst, which indicates that lattice oxygen participates in the reaction and may play an important role in the catalytic oxidative decomposition of DMMP. Figure 12c shows the Cu2p spectra of CuO before and after the reaction. The Cu2p 3/2 peak changed from 534 to 535.5 eV, with a broad satellite peak at 935-945 eV. The Cu2p satellite peak of CuO in the deactivated catalyst was similar to that of Cu2p 3/2 when comparing the Cu2p of copper carbonate and copper phosphate, and the peak position was consistent with that of Cu2p 3/2 . Therefore, copper phosphate and copper carbonate products were speculated to exist on the surface of the deactivated catalyst.  The phosphorus-containing byproducts on the surface of the deactivated catal were qualitatively analyzed using ion chromatography, which showed the presenc PO4 3− and methyl phosphonate species, as shown in Table 7. DMMP decomposi amounts and the corresponding P element amounts are presented in Table 7. Mass sp trometry was used to qualitatively analyze the tail gas produced by the catalytic dec position of DMMP on CeO2, 50%Cu/Ce, and CuO to determine its composition. Figur shows the mass spectra of the tail gas generated by CeO2, 50%Cu/Ce, and CuO, dem strating that methanol, CO2, H2O, and H2 were detected in the reaction tail gas. Addit ally, except for the fragment mass-to-charge ratio of methanol and CO2 (28), the mass charge ratio of CO was 28. Therefore, gas chromatography was used to determine presence of CO in the generated exhaust gas. CO was detected in the reaction tail ga CeO2 and CuO; however, no CO was detected for 50%Cu/Ce, indicating that the oxida activity of the 50%Cu/Ce catalyst was higher than that of CeO2 and CuO.  The phosphorus-containing byproducts on the surface of the deactivated catalysts were qualitatively analyzed using ion chromatography, which showed the presence of PO 4 3− and methyl phosphonate species, as shown in Table 7. DMMP decomposition amounts and the corresponding P element amounts are presented in Table 7. Mass spectrometry was used to qualitatively analyze the tail gas produced by the catalytic decomposition of DMMP on CeO 2 , 50%Cu/Ce, and CuO to determine its composition. Figure 13 shows the mass spectra of the tail gas generated by CeO 2 , 50%Cu/Ce, and CuO, demonstrating that methanol, CO 2 , H 2 O, and H 2 were detected in the reaction tail gas. Additionally, except for the fragment mass-to-charge ratio of methanol and CO 2 (28), the mass-to-charge ratio of CO was 28. Therefore, gas chromatography was used to determine the presence of CO in the generated exhaust gas. CO was detected in the reaction tail gas of CeO 2 and CuO; however, no CO was detected for 50%Cu/Ce, indicating that the oxidative activity of the 50%Cu/Ce catalyst was higher than that of CeO 2 and CuO.    The in situ DRIFTS of DMMP on CeO 2 , 50%Cu/Ce, and CuO at varying temperatures were determined to ascertain the species change of DMMP on the surface of the three catalysts. As shown in Figure 1, the molecular structure of DMMP contains P-CH 3 , O-CH 3 , P=O, P-O, P-C, C-O, and C-H bonds. The vibration peaks of chemical bonds in DMMP are mainly distributed in two regions. C-H bond-stretching vibrations of P-CH 3 and O-CH 3 are located in the high-frequency region from approximately 2800 cm −1 to 3200 cm −1 , whereas stretching vibrations of P=O, C-O, and P-O, as well as deformation vibrations and rocking vibrations of PCH 3 and OCH 3 are located in the low-frequency region from 800 cm −1 to 1500 cm −1 . The corresponding functional groups of the DMMP molecules were consistent with those reported in the literature [58,59], as shown in Table 8. The corresponding functional groups of surface species on CeO 2 , 50%Cu/Ce, and CuO are shown in Table 9. The characteristic peak located at 3500-3800 cm −1 was attributed to the stretching vibration of the surface hydroxyl groups [60,61]. As the temperature increased, hydroxyl vibration peaks of the negative-going OH bands were observed. The depletion of the surface hydroxyl groups may have originated from the formation of H bonds between the DMMP molecules and surface hydroxyl groups of CeO 2 , CuO, and Cu/Ce or from the surface hydroxyl groups participating in the DMMP decomposition reaction [59,62] (Figures 14-16). Figure 14a shows the in situ DRIFTS of CeO 2 in the high-frequency region.  in Table 9. The characteristic peak located at 3500-3800 cm −1 was attributed to the stretching vibration of the surface hydroxyl groups [60,61]. As the temperature increased, hydroxyl vibration peaks of the negative-going OH bands were observed. The depletion of the surface hydroxyl groups may have originated from the formation of H bonds between the DMMP molecules and surface hydroxyl groups of CeO2, CuO, and Cu/Ce or from the surface hydroxyl groups participating in the DMMP decomposition reaction [59,62] (Figures 14-16). Figure 14a shows the in situ DRIFTS of CeO2 in the high-frequency region.  The characteristic infrared peaks of ν a (P-CH 3 ), ν a (O-CH 3 ), ν s (P-CH 3 ), and ν s (O-CH 3 ) in the DMMP molecules gradually weakened with increasing temperature and disappeared at 400 • C. New characteristic peaks simultaneously appeared at 2931 cm −1 and 2846 cm −1 , and the peak intensity weakened with increasing temperature. Therefore, we speculated that these two peaks correspond to the ν(C-H) vibrational modes of the intermediates generated during DMMP decomposition. Morris et al. reported that cleavage of P-OCH 3 led to the formation of methoxy groups, which combined with Ti ions to form surface methoxy groups. The vibrational peak of the surface methoxy shifts to a lower frequency than that of DMMP [63]. A similar phenomenon was observed in the in situ DRIFTS of CeO 2 . Therefore, the peak at 2803 cm −1 was attributed to ν(Ce-O-CH 3 ). Figure 14a shows a partially enlarged view of the 2600-2900 cm −1 band in the high-frequency region, which clearly demonstrates that the peak disappeared when the temperature reached 400 • C. This indicates that the methoxy group in Ce-O-CH 3 is an unstable intermediate product that can be completely decomposed by the CeO 2 catalyst at 400 • C. The vibrational peak at 2722 cm −1 shown in Figure 14a was assigned to the 2δ(C-H) vibrational mode of the formate product [61]. The characteristic infrared peaks of DMMP molecules in the midlow frequency range of CeO 2 in situ DRIFTS (Figure 14b [61]. Additionally, the vibrational mode ν(CO 3 ) of the carbonate appeared at 1357 cm −1 [60]. of the δa(O-CH3), ρ‖(O-CH3), νa(C-O), and νs(C-O) vibration modes in DMMP indicated that the methoxy group completely decomposed. Although the characteristic peaks of ρ‖(P-CH3) and δs(P-CH3) in DMMP were still visible, these characteristic peaks were speculated to be characteristic peaks of methyl-containing products. At 400 °C, the peak intensities of the two bands at 1290-1150 and 1100-1000 cm -1 significantly increased, indicating that a large number of phosphorus and carbon oxygen products were produced on the surface of CuO at 400 °C [21,65].  Figure 16 shows the in situ DRIFTS spectra of 50%Cu/Ce at varying temperatures. In the in situ DRIFTS (Figure 16a) in the high-frequency region of 50%Cu/Ce, the stretching vibration mode νa(P-CH3) of DMMP molecules shifted from 2996 to 2993 cm -1 ; νs(P-CH3) shifted from 2927 to 2925 cm -1 , and the peak did not decay with increasing temperature. Therefore, we speculated that the vibration peaks of ν(P-CH3) at 2993 cm -1 and 2925 cm -1 Figure 15. In situ DRIFTS spectra of CuO at varying temperatures in (a) the high-frequency region and (b) in the mid-low frequency region. Figure 15 shows the in situ DRIFTS spectra of CuO at varying temperatures. CuO exhibited a multimodal overlap at 3100-2800 cm −1 in the high-frequency region of the in situ DRIFTS. Moreover, the in situ DRIFTS spectral peak signal of CuO in the high-frequency region was relatively weak, and it was impossible to directly distinguish the peak attribution in this region. The ν(C-H) characteristic peak of the intermediate product appeared in this region with reference to the in situ infrared spectrum of CeO 2 (Figure 15a). Therefore, we speculated that the CuO peaks in this region might be the ν(C-H) characteristic peaks of the undecomposed DMMP itself and the ν(C-H) characteristic peaks of the partially decomposed DMMP products. This is similar to the in situ DRIFTS spectrum of CeO 2 ; a ν(Cu-O-CH 3 ) vibrational peak was also generated on the CuO surface at 2820 cm −1 . The in situ DRIFTS (Figure 15b) in the mid-low frequency region clearly shows that the ν(P=O) vibration mode in DMMP disappeared, and a new bridging P species ν s (O-P-O) appeared at 1055 cm −1 , implying an interaction between the phosphoryl oxide and catalyst surface [21,64]. The disappearance of the ν(P-O) vibration mode in DMMP indicated that the methoxy group was separated from the DMMP molecules, and the disappearance of the δ a (O-CH 3 ), ρ (O-CH 3 ), ν a (C-O), and ν s (C-O) vibration modes in DMMP indicated that the methoxy group completely decomposed. Although the characteristic peaks of ρ (P-CH 3 ) and δ s (P-CH 3 ) in DMMP were still visible, these characteristic peaks were speculated to be characteristic peaks of methyl-containing products. At 400 • C, the peak intensities of the two bands at 1290-1150 and 1100-1000 cm −1 significantly increased, indicating that a large number of phosphorus and carbon oxygen products were produced on the surface of CuO at 400 • C [21,65]. cm is the ν(C-H) vibrational mode of the intermediate product generated during the decomposition of DMMP. Similar to CeO2 and CuO, the ν(Ce/Cu-O-CH3) vibrational mode of the unstable intermediate appeared near 2805 cm -1 in the 50%Cu/Ce in situ DRIFTS and disappeared at 400 °C. The spectral characteristic peaks in the low-frequency region of the in situ DRIFTS (Figure 16b) of 50%Cu/Ce at 400 °C were similar to those of CuO, i.e., the methoxy group in DMMP completely decomposed. A methyl-containing product peak was present at 1307 cm -1 . Large quantities of phosphorus and carbon oxygen products were produced in the 1250-900 cm -1 band, and the magnified image of the 1350-850 cm -1 band shows that the 1307 cm -1 and 900-893 cm -1 bands are characteristic peaks of methyl-containing products. Figure 16. In situ DRIFTS spectra of 50%Cu/Ce at varying temperatures in (a) the high-frequency region and (b) in the mid-low frequency region, the inset is a magnified image of the tagged area.   Figure 16 shows the in situ DRIFTS spectra of 50%Cu/Ce at varying temperatures. In the in situ DRIFTS (Figure 16a) in the high-frequency region of 50%Cu/Ce, the stretching vibration mode ν a (P-CH 3 ) of DMMP molecules shifted from 2996 to 2993 cm −1 ; ν s (P-CH 3 ) shifted from 2927 to 2925 cm −1 , and the peak did not decay with increasing temperature. Therefore, we speculated that the vibration peaks of ν(P-CH 3 ) at 2993 cm −1 and 2925 cm −1 were the DMMP decomposition products. The stretching vibration mode ν a (O-CH 3 ) in DMMP shifted from 2958 to 2949 cm −1 , and the stretching vibration of ν a (O-CH 3 ) weakened with increasing temperature, disappearing at 400 • C. The vibrational mode of ν s (O-CH 3 ) in DMMP was similar to that of CeO 2 ; therefore, we speculated that the peak at 2849 cm −1 is the ν(C-H) vibrational mode of the intermediate product generated during the decomposition of DMMP. Similar to CeO 2 and CuO, the ν(Ce/Cu-O-CH 3 ) vibrational mode of the unstable intermediate appeared near 2805 cm −1 in the 50%Cu/Ce in situ DRIFTS and disappeared at 400 • C. The spectral characteristic peaks in the low-frequency region of the in situ DRIFTS (Figure 16b) of 50%Cu/Ce at 400 • C were similar to those of CuO, i.e., the methoxy group in DMMP completely decomposed. A methyl-containing product peak was present at 1307 cm −1 . Large quantities of phosphorus and carbon oxygen products were produced in the 1250-900 cm −1 band, and the magnified image of the 1350-850 cm −1 band shows that the 1307 cm −1 and 900-893 cm −1 bands are characteristic peaks of methyl-containing products.
The reaction pathways for the catalytic decomposition of DMMP on CeO 2 , CuO, and CuO/CeO 2 were speculated using the combined ion chromatography, mass spectrometry, and in situ DRIFTS data (Scheme 1). Scheme 1(I) shows the presumed reaction pathway for the decomposition of DMMP on the surface of CeO 2 . First, the electron-rich O in DMMP combines with Ce or surface hydroxyl groups on the surface of CeO 2 . The surface hydroxyl groups nucleophilically attack P, breaking the P-OCH 3 bond, and combine the methoxy group with H to generate gaseous methanol. Simultaneously, methyl phosphates with bridged O-P-O structures formed on the surface [4,21,59] The surface lattice O in CeO 2 nucleophilically attacks the methyl group in methyl phosphate, forming a PO x species on the CeO 2 surface, and the methyl group is further dehydrogenated to CO, H 2 , CO 2 , and H 2 O (Scheme 1a). According to Wu et al., methanol can induce a series of side reactions on the CeO 2 surface [61]. First, in Scheme 1(II) methanol reacts with the CeO 2 surface lattice O to form surface methoxy Ce-O-CH 3 . The methoxy groups continue to react with the surface lattice O or hydroxyl groups, fully dehydrogenating to form CO, H 2 , CO 2 , and H 2 O (Scheme 1b) and partially dehydrogenating to form formate (Scheme 1c) [66], which reacts with the surface lattice O and completely dehydrogenates to form carbonate (Scheme 1d) [67]. Based on the in situ DRIFTS data, we speculated that the carbonates may contain one or more species, including polydentate, bidentate, bridged, and monodentate carbonates, due to the broadband nature of their vibrational modes. The catalytic decomposition pathways of DMMP on CuO and CuO/CeO 2 are essentially the same as those on CeO 2 (Scheme 1(I)). The difference is that there is no formate in the decomposition products of the CuO and CuO/CeO 2 catalytic decomposition of DMMP (Scheme 1c), and no CO is generated in the decomposition products of the CuO/CeO 2 catalytic decomposition of DMMP. This indicates that the synergistic effect of the bimetals enhances the selectivity for the fully catalytic oxidation product of DMMP, CO 2 .
Catalysts 2022, 12, x FOR PEER REVIEW 2 same as those on CeO2 (Scheme 1(Ⅰ)). The difference is that there is no formate in t composition products of the CuO and CuO/CeO2 catalytic decomposition of D (Scheme 1c), and no CO is generated in the decomposition products of the CuO/Ce alytic decomposition of DMMP. This indicates that the synergistic effect of the bi enhances the selectivity for the fully catalytic oxidation product of DMMP, CO2. Various mechanisms have been proposed over ceria-based catalysts in the lite [68][69][70], which can be generally broken down into three main categories: the Eley− mechanism [71], the Langmuir−Hinshelwood mechanism [72], and Mars−van−Krevelen mechanism [73,74]. In catalytic reactions at higher temperatur mechanism of CeO2-based materials is mainly the Mars−van−Krevelen mechanism Mars−van−Krevelen mechanism consists of a redox cycle. First, reactant molecules Various mechanisms have been proposed over ceria-based catalysts in the literature [68][69][70], which can be generally broken down into three main categories: the Eley-Rideal mechanism [71], the Langmuir-Hinshelwood mechanism [72], and the Mars-van-Krevelen mechanism [73,74]. In catalytic reactions at higher temperatures, the mechanism of CeO 2 -based materials is mainly the Mars-van-Krevelen mechanism. The Mars-van-Krevelen mechanism consists of a redox cycle. First, reactant molecules interact with the lattice oxygen on the catalyst surface, and oxygen vacancies are formed on the catalyst surface. Then, the oxygen vacancies are refilled by gaseous oxygen or oxygen atoms from the bulk. Kinetic experimentation is needed in future studies to verify the thermal catalytic DMMP decomposition mechanism of as-prepared materials reported in the present study.

Synthesis of CeO 2 Catalyst
Ce(NO 3 ) 3 ·6H 2 O (99.9%) and NaOH (99.9%) were obtained from Macklin, Shanghai, China. First, 2.6 g Ce(NO 3 ) 3 ·6H 2 O was dissolved in deionized water (10 mL). Then, 21.6 g NaOH was dissolved in 45 mL of deionized water. The Ce(NO 3 ) 3 solution was added dropwise to the NaOH solution and continuously stirred at room temperature for 30 min. This mixture was hydrothermally treated in 100 mL Teflon-lined autoclaves for 24 h at 100 • C. After naturally cooling to room temperature, the products were centrifuged and washed several times with deionized water and ethanol. The precipitate was dried at 100 • C for 8 h and used for the synthesis of CuO/CeO 2 catalysts. Then, the precipitate was calcined at 400 • C for 4 h and used for characterization, catalytic performance testing, and reaction mechanism analysis.

CuO/CeO 2 Catalyst Preparation
A secondary alkaline hydrothermal approach was adopted to load CuO on CeO 2 nanorod supports to synthesize CuO/CeO 2 catalysts. Specifically, 3 g of the precipitates was fully dispersed in 10 mL of deionized water, including the calculated amount of Cu(NO 3 ) 2 ·3H 2 O, and slurries were obtained with 30 min of continuous stirring. Then, 10.8 g of NaOH was dissolved in 45 mL of deionized water. The NaOH solution was dropped into the slurry and stirred for 30 min. The mixture slurries were hydrothermally treated in a 100 mL Teflon-lined autoclave for 24 h at 100 • C. After naturally cooling to room temperature, the products were centrifuged and washed several times with deionized water and ethanol, dried at 100 • C for 8 h, and calcined at 400 • C for 4 h in air. Four types of CuO/CeO 2 catalysts were prepared using 0.87, 1.74, 4.36, and 6.97 g of Cu(NO 3 ) 2 ·3H 2 O and labeled as 10%Cu/Ce, 20%Cu/Ce, 50%Cu/Ce, and 80%Cu/Ce, respectively.

Synthesis of CuO Catalyst
CuO was prepared using a hydrothermal method. First, 8.7 g of Cu(NO 3 ) 2 ·3H 2 O was fully dispersed in deionized water (10 mL). Then, 10.8 g of NaOH was dissolved in 45 mL of deionized water. The Cu(NO 3 ) 2 solution was added dropwise to the NaOH solution and continuously stirred at room temperature for 30 min. This mixture was hydrothermally treated in 100 mL Teflon-lined autoclaves for 24 h at 100 • C. After naturally cooling to room temperature, the products were centrifuged and washed several times with deionized water and ethanol. The precipitates were dried at 100 • C for 8 h and calcined at 400 • C for 4 h in air.

Characterization of CuO/CeO 2 Catalysts
XRD analyses were performed on a 9 kW SmartLab (Rigaku, Saitama, Japan) with Cu Kα radiation (λ = 1.5418 Å). The Cu and Ce contents were analyzed using inductively coupled plasma optical emission spectrometry (ICP-OES; Agilent 725 ES, Santa Clara, CA, USA). N 2 adsorption-desorption measurements were performed on a Nova 4200e Quantachrome surface area analyzer (Quantachrome Instruments, Boynton Beach, FL, USA) at 77 K and calculated using the Brunauer-Emmett-Teller (BET) model. TEM images and energy-dispersive X-ray spectroscopy (EDX) images of the samples were obtained using a TECNAI F20 electron microscope. H 2 -TPR was performed using an AutoChem II 2920 instrument (Micromeritics, Norcross, GA, USA). The samples (60 mg) were heated to 300 • C at 50 mL/min of He for 3 h and cooled to 50 • C. Temperature-programmed reduction was performed by heating the sample at 10 • C/min from 50 to 900 • C in an H 2 /Ar mixture (10% H 2 , 90% Ar, 50 mL/min). O 2 -TPD analyses were performed on an AutoChem II 2920 (Micromeritics, Norcross, GA, USA) as follows. The sample (60 mg) was first treated with He (50 mL/min) at 300 • C for 3 h to remove adsorbed impurities. After cooling to 50 • C, the sample was exposed to O 2 /Ar (10% O 2 , 90% Ar, 30 mL/min) for 1 h. Then, the sample was purged with He (30 mL/min) for 1 h. Finally, the sample was heated to 800 • C at 10 • C/min in He (30 mL/min). XPS measurements were conducted using a Thermo Fisher ESCALAB 250Xi (Waltham, MA, USA) system with Al-Kα radiation. The binding energy was calibrated using the C 1s peak (284.8 eV) as the internal standard. The XPS data were analyzed using Avantage software v5.976 (Thermo Fisher). All peaks of the corrected spectra were fitted with a Gaussian-Lorentzian shape function to fit the data. Raman spectroscopy was performed on a HORIBA LabRAM HR Evolution (Longjumeau, France) with 532 nm laser excitation. In situ DRIFTS spectra were acquired during the reaction process using a Nicolet iS50 instrument (Thermo Fisher Scientific, Waltham, MA, USA) to monitor the DMMP thermal catalytic decomposition reaction. During the reaction, 64 scans were recorded at an 8 cm −1 resolution (to minimize collection time) to form the spectra shown.

Thermal Catalytic Decomposition Performance Tests
The performances of the catalysts were evaluated using a custom-built catalytic evaluation system ( Figure 17). Compressed air at 100 mL/min was used as the carrier gas. The compressed air flowed through a bubbler filled with DMMP at 30 • C to create a DMMP vapor stream. To avoid DMMP vapor condensation, the temperature of the gas line was maintained at 100 • C. Under these conditions, the inlet DMMP concentration was 8.46 g/m 3 . A gas chromatograph (Agilent 6890N, chromatographic column DB-1701, Santa Clara, CA, USA) equipped with a flame ionization detector was used for online analysis. A blank DMMP decomposition experiment was conducted in an empty reactor to verify the thermal stability of DMMP at 400 • C. Then, 0.46 g of CuO/CeO 2 catalysts (20-40 mesh) was loaded in a 4 mm reaction tube with a catalytic temperature of 400 • C and a gas hourly space velocity (GHSV) of 20,000 h −1 . The catalytic decomposition performance was expressed in terms of the protection time (Protection time is defined as the duration of DMMP conversion at 100%). The DMMP conversion rate is defined by Equation (1).
where C out is the content of DMMP in the reaction tail gas, and C in is the initial content of DMMP created by the bubbler. reaction. During the reaction, 64 scans were recorded at an 8 cm −1 resolution (to minimize collection time) to form the spectra shown.

Thermal Catalytic Decomposition Performance Tests
The performances of the catalysts were evaluated using a custom-built catalytic evaluation system ( Figure 17). Compressed air at 100 mL/min was used as the carrier gas. The compressed air flowed through a bubbler filled with DMMP at 30 °C to create a DMMP vapor stream. To avoid DMMP vapor condensation, the temperature of the gas line was maintained at 100 °C. Under these conditions, the inlet DMMP concentration was 8.46 g/m 3 . A gas chromatograph (Agilent 6890N, chromatographic column DB-1701, California, USA) equipped with a flame ionization detector was used for online analysis. A blank DMMP decomposition experiment was conducted in an empty reactor to verify the thermal stability of DMMP at 400 °C. Then, 0.46 g of CuO/CeO2 catalysts (20-40 mesh) was loaded in a 4 mm reaction tube with a catalytic temperature of 400 °C and a gas hourly space velocity (GHSV) of 20,000 h −1 . The catalytic decomposition performance was expressed in terms of the protection time (Protection time is defined as the duration of DMMP conversion at 100%). The DMMP conversion rate is defined by Equation (1).
where Cout is the content of DMMP in the reaction tail gas, and Cin is the initial content of DMMP created by the bubbler.

Characterization of Products in the Thermal Catalytic Decomposition Reaction
Thermal catalytic decomposition microreactions were performed in a CATLAB microreactor (HIDEN ANALYTICAL, Warrington, UK). First 0.05 g CuO/CeO2 was loaded in the reaction tube and dried for 2 h at 300 °C under He (100 mL/min). After the sample Figure 17. Schematic diagram of the catalytic evaluation system. MFC denotes mass flow meter, and the reaction tube size is 4 mm.

Characterization of Products in the Thermal Catalytic Decomposition Reaction
Thermal catalytic decomposition microreactions were performed in a CATLAB microreactor (Hiden Analytical, Warrington, UK). First 0.05 g CuO/CeO 2 was loaded in the reaction tube and dried for 2 h at 300 • C under He (100 mL/min). After the sample naturally cooled to room temperature, it was heated to 400 • C at 10 • C/min. A given concentration of DMMP vapor was introduced to the reaction tube using an 80% Ar and 20% O 2 mixture bubbled in a flask containing DMMP at 10 • C with a constant flux of 50 mL/min. The reaction tail gas was monitored using an HPR-20 mass spectrometer (Hiden Analytical, Warrington, UK). Ion chromatography was performed on an ICS-5000 instrument (Thermo Scientific, Waltham, MA, USA). The spent catalyst was dissolved in aqua regia and hydrofluoric acid, with a 15 mM KOH buffer used as eluent.

Conclusions
CeO 2 nanorods were used as catalyst supports to prepare CuO/CeO 2 catalysts with varying loading ratios using a secondary hydrothermal method. The addition of NaOH during the secondary hydrothermal process partially dissolved CeO 2 . During the recrystallization process, Cu was doped into the CeO 2 lattice, resulting in more defects in the CeO 2 lattice and the formation of a Cu-V-Ce structure, which facilitated oxygen transfer. This preparation method provides a new and feasible method for lattice-doped bimetallic catalysts. Furthermore, the performance of CeO 2 , CuO/CeO 2 , and CuO in the thermal catalytic decomposition of DMMP was investigated. The load of CuO considerably improved the protection time as compared to the CeO 2 and CuO catalysts. Among the Cu/Ce catalysts with varying ratios, the protection time first increased and then decreased with an increasing loading ratio, and the 50%Cu/Ce catalyst had the longest protection time. A strong interaction was generated between CeO 2 and CuO through electron transfer, and the bimetallic synergy between Cu and Ce enhanced the catalytic decomposition performance of the CeO 2 . The catalytic reaction pathways of CeO 2 , CuO, and CuO/CeO 2 were similar, and the gas-phase products were CH 3 OH, CO 2 , H 2 O, and H 2 . However, CO was still present in the gas-phase products of CeO 2 and CuO, whereas the gas phase of CuO/CeO 2 was CO 2 , indicating that the CuO/CeO 2 bimetallic catalyst had higher selectivity for CO 2 . Carbonate and formate byproducts were formed on the CeO 2 catalyst surface, whereas only carbonates were formed on the surfaces of the CuO and CuO/CeO 2 catalysts. The accumulation of P-O and C-O products on the surface of the three catalysts resulted in their eventual deactivation. This study preliminarily revealed the mechanism of the CuO/CeO 2 catalyst for catalysis of the decomposition of DMMP, which can inform the design of catalysts for the efficient degradation of chemical poisons in the future.
Author Contributions: W.K. and S.Z. designed the experiments, made the main contribution to the experimental works, and managed the experimental and writing process as co-first authors; X.W. and Q.H. assisted in accomplishing a part of the experimental works and characterizations; S.Z., P.Y. and Y.D. provided the concept of this research, participated in the guidance of this work, and provided advice on the synthesis of CeO 2 supports and catalyst characterization as the corresponding authors. Y.Y. participated in the guidance of this work and provided advice on the synthesis of CeO 2 supports and catalyst characterization as the corresponding authors. All authors have read and agreed to the published version of the manuscript.