Analysis of thermal decomposition of catalyst precursors
Figure 1 shows the TG-DSC curves of the different catalyst precursors after drying at 120 oC, from which it can be seen that both TG and DSC curves are very similar for the two catalyst precursors. Specifically, the TG curve has two weight loss steps, the first one at < 200 oC, corresponding to the removal of water adsorbed on the catalyst surface, and the second one with a significant weight loss accompanied by a large exothermic peak at 200–400 oC, corresponding to the decomposition of the precursor hydroxide. From the above experimental results, it is clear that the precursors of both catalysts have decomposed completely at > 400 oC, so in this paper, 450 oC was selected as calcination temperature for the precursors to prepare the CuO-CeO2 catalysts.
Characterization of catalysts
Structural and textural properties
Figure 2 shows the XRD spectra of the CuO-CeO2 catalysts prepared from different cerium salts. It can be seen from the figure that strong characteristic diffraction peaks of CeO2 are present at 2θ = 28.5 o, 33.0 o, 47.4 o, and 56.3 o for both samples (Jin et al. 2022). In addition, characteristic diffraction peaks of CuO are also present at 2θ = 35.5 o and 38.7 o (Jin et al. 2022). The results obtained in combination with those of the thermal analyses (Fig. 1) indicate that the precursors of the catalysts have decomposed to oxides under calcination at 450 oC.
The average particle sizes of CuO and CeO2 were calculated based on the diffraction peaks of the two crystalline planes of CuO and the diffraction peaks of the crystalline plans of CeO2(111) using the Scherrer formula, and the results are listed in Table 1. As can be seen from Table 1, the CeO2 crystallite sizes of the two catalysts are very close to each other, but the particle size of CuO of the CC-N catalyst is significantly larger than that of CC-NH.
Table 1
Structural and textural properties of the different CuO-CeO2 catalysts
Sample | Cu/(Cu + Ce)a /mol% | DCeO2b /nm | DCuOc /nm | Vpd /(cm3/g) | Dpd /(Å) | SBET /m2/g) | Lattice constantb /nm |
CC-N | 18.9 | 9.35 | 19.0 | 0.12 | 76.5 | 61 | 0.5382 |
CC-NH | 18.0 | 9.98 | 7.4 | 0.08 | 78.9 | 38 | 0.5472 |
a Copper content obtained from ICP-AES tests. |
b Crystal size and lattice constant of CeO2 were obtained from CeO2(111) plane. |
c Crystal size of CuO is based on CuO(002) and CuO(111) planes. |
d Vp (pore volume) and Dp(pore diameter) were measured from N2 adsorption. |
In addition, the lattice constants of CeO2 calculated by the Bragg equation based on CeO2(111) plane are also listed in Table 1. As can be seen, the lattice constant of CeO2 of CC-NH is larger than that of CC-N. Generally, two factors can affect the lattice constants of CeO2. On the one hand, the radius of Cu2+(0.072 nm) is smaller than that of Ce4+(0.097 nm), which leads to lattice contraction when Cu2+ replaces part of Ce4+ into the lattice (Zheng et al. 2016). On the other hand, the formation of oxygen vacancies due to the substitution of Ce4+ by Cu2+ or the formation of Ce3+(0.103 nm) will have the opposite effect on the lattice of CeO2 (Hossain et al. 2018). These two opposing effects together determine the lattice constants of CeO2.
The Raman spectra of CuO-CeO2 catalysts prepared from different cerium precursors are shown in Fig. 3. It can be seen from the figure that both samples have a strong peak at ~ 450 cm− 1 ascribed to the F2g characteristic peak of cubic CeO2 (Jin et al. 2022). In addition, there is a very weak peak at ~ 600 cm− 1 corresponding to the characteristic peak of oxygen vacancy due to lattice distortion of CeO2 (Jin et al. 2022), which is generally believed to be generated due to the entrance of Cu2+ into the CeO2 lattice, leading to its lattice contraction. Comparing the two catalysts prepared from different cerium precursors, there is little difference in their F2g characteristic peaks, indicating that the valence state of the cerium ion does not have much influence on its F2g characteristic peak of CeO2. The ratio of the two peak areas at 600 and 450 cm− 1 (A600/A450) is commonly used to calculate the relative concentration of oxygen vacancies in catalysts (Liu et al. 2020; Jin et al. 2022), but in this paper, it is difficult to use this method to accurately determine the relative concentration of oxygen vacancies in catalysts because the peak at 600 cm− 1 is too weak and too broad (Jin et al. 2022).
The textural properties of CuO-CeO2 catalysts prepared with different cerium precursors are also presented in Table 1. As can be seen, the specific surface area (SBET) and pore volume (Vp) of CC-N are significantly larger than those of CC-NH, but their pore sizes (Dp) are similar.
The actual Cu content in the catalyst was measured by the ICP-AES, and the results are collected in Table 1. Clearly, the actual Cu content of CC-N and CC-NH samples is close to the preseted value (20%).
Surface Characterization (XPS)
The CuO-CeO2 catalysts prepared with different cerium precursors were characterized by XPS to observe the state of metal oxides on the surface and the distribution of surface components. The obtained results are presented in Fig. 4 and Table 2. As can be seen from Table 2, the Cu/Ce and Ce/O atomic ratios on the surface of the CC-NH catalyst are slightly higher than those of the CC-N catalyst (0.66 vs. 0.63, 0.29 vs. 0.25, respectively), indicating that the Cu and Ce contents on the surface of CC-NH catalyst are relatively higher. In addition, compared with the results of ICP-AES analysis (Table 1), it can be seen that the Cu content on the surface of the two catalysts is significantly higher than that of the bulk composition, indicating that Cu species are enriched on catalyst surface because the surface energy of Cu is lower than that of Ce (Hu et al. 2010).
Table 2
XPS data measured for the different CuO-CeO2 catalysts
Sample | Atomic ratio | OA a/% | Ce3 + b/% | Cu+ c/ % |
Cu/Ce | Ce/O |
CC-N | 0.63 | 0.25 | 24.8 | 15.5 | 21.0 |
CC-NH | 0.66 | 0.29 | 22.7 | 14.4 | 18.2 |
a Area ratio of OA peaks to the entire O 1s peaks. |
b Area ratio of peaks attributed to Ce3+ to all Ce 3d peaks. |
c Area ratio of peaks attributed to Cu+ to all Cu 2p peaks. |
According to the Cu 2p spectra of different catalysts (Fig. 4(A)), both catalysts have a strong main peak at 933.7 eV, and a satellite peak appears at 941.2-941.8 eV, indicating that Cu species on the surface of catalysts mainly exist as Cu2+ (Liu et al. 2020; Jin et al. 2022). In addition, a shoulder peak representing Cu+ or Cu0 species at 932.0 eV was also observed in the spectra of the two catalysts (Liu et al. 2020; Jin et al. 2022). At this time, it is necessary to further distinguish with the help of the Cu LMM spectra (Fig. 4(B)). It can be seen from Fig. 4(B) that there is a peak representing the presence of Cu+ at 914.5 eV for both catalysts. In order to explain the reduction of Cu species on the catalyst more clearly, the obtained Cu LMM spectra were peak-separated, and the relative content of Cu+ in different catalysts was expressed as Cu+(%) by the ratio of the area belonging to the Cu+ peak to the total area of all Cu species. According to the calculated data (Table 2), the relative content of Cu+ on CC-NH is less than that of CC-N, which corresponds to the relative content of Ce3+ on the two catalysts obtained later. Under the influence of CuO-CeO2 interaction, the Cu species on the surface of the catalyst will partially reduce and generate a small amount of Cu+, which has been reported to have a better ability to adsorb and activate CO (Liu et al. 2020; Jin et al. 2022).
Figure 4(C) shows the Ce 3d spectra of the different catalysts, where it can be seen that both samples contain eight peaks attributed to four pairs of double spins, with Ce4+ 3d3/2 peaks labeled as u, u'', u''' and Ce4+ 3d5/2 peaks labeled as v, v'', v''', while u' and v' are used to denote the two electron arrangement configurations of the Ce3+ species (Fu et al. 2015). This observation shows that both Ce4+ and Ce3+ species are present on the surface of both catalysts. The ratio of the sum of u' and v' peak areas to the total area of all peaks is usually used to estimate the relative content of Ce3+ species in the catalysts (Ce3+/(Ce4++Ce3+)) (Sun et al. 2015b), and the Ce3+(%) data for both catalysts are listed in Table 2. From Table 2, it can be found that the Ce3+ species are present on both catalysts and the relative content of Ce3+ on CC-NH is less than that on CC-N. According to the literature (Guo and Zhou 2016), the relative content of Ce3+ in the catalyst represents the strength of the Cu-Ce interaction in the catalyst and the concentration of the oxygen deficiency, and the higher the relative content of Ce3+, the higher the concentration of the oxygen deficiency. Additionally, it is well known that the presence of Ce3+ can promote the electron transfer process Ce3+ + Cu2+→ Ce4+ + Cu+, i.e., more Cu+ is generated (Lu et al. 2021; Wang et al. 2022), which is consistent with the previous results obtained by Cu LMM.
It can be seen from the O 1s spectra of both catalysts in Fig. 4(D) that there are three peaks in the spectra of both catalysts at 528.9-529.2 eV, 529.5-529.9 eV, and 531.2-531.4eV. They belong to lattice oxygen (OL), surface adsorbed oxygen (OA), and hydroxyl oxygen (OOH) (Zhao et al. 2019; Jin et al. 2022). The ratio of the area of the OA peak to the total area of the three peaks, designated as OA(%), represents the relative concentration of adsorbed oxygen in catalysts. Therefore, it can be seen from the results in Table 3 that the relative concentration of adsorbed oxygen on the surface of CC-N is greater than that of CC-NH. This result indicates that the concentration of oxygen deficiency on the surface of CC-N is greater than that of CC-NH, which is consistent with the oxygen deficiency concentration inferred from the Ce3+ content above. In addition, the binding energy of lattice oxygen of CC-N is significantly greater than that of CC-NH, indicating that its lattice oxygen has better mobility and higher activity (Zhao et al. 2019).
The above results show that different cerium precursors lead to differences in the contents of Cu+, Ce3+, and oxygen vacancies on the surface of CuO-CeO2 catalysts, and consequently resulting in different catalytic performance in CO oxidation at low-temperatures.
CO - IR analysis
In order to further study the existence status of Cu species in catalysts, in-situ DRIFTs characterization of the adsorbed CO on both catalysts was carried out. As shown in Fig. 5, there is a strong absorption peak at 2103 cm− 1 on the DRIFT spectra of both catalysts, which can be ascribed to the stretching vibration peak attributable to the linear adsorption of CO on Cu+ (Cu+-CO) (Dong et al. 2016), and the intensity of this peak for CC-NH catalyst is weaker than that for CC-N catalyst. This is consistent with the relative content of Cu+ obtained by the above XPS analysis (Table 2).
It can also be seen from Fig. 5 that DRIFT spectra of the two samples in the range of 1290 to 1598 cm− 1 have many peaks belonging to formate and carbonate adsorbed on the CeO2 surface. According to the literature (Sun et al. 2015a; Zhang et al. 2018b), the peak around 1598 cm− 1 belongs to carbonate or formate adsorbed on CeO2, and the peak around 1470 cm− 1 belongs to polydentate or monodentate carbonates, and the peaks around 1386 cm− 1 and 1293 cm− 1 belong to bidentate carbonate. After adding CO at 30 oC, obvious stretching vibration peaks of formate and carbonate appeared on the surface of both samples, and these peaks were mainly caused by the adsorption of CO in the gas phase or CO2 generated by the reaction between CO and the sample on the surface of CeO2. Moreover, the stretching vibration peak of carbonate species of CC-N is stronger than that of CC-NH, indicating that CO adsorbed on its surface is more likely to react with oxygen species on the surface of the catalyst (Sun et al. 2015b).
Therefore, it can be seen from the above analysis results that Cu+ species formed under the promotion of CuO-CeO2 interaction have better CO adsorption capacity than Cu2+, and CC-N has greater CO adsorption capacity than CC-NH.
Reducibility of catalyst (H2-TPR)
As shown in Fig. 6, three hydrogen consumption peaks appear on the H2-TPR curves of both catalysts: the low-temperature α peak corresponds to the reduction of finely dispersed CuO with strong interaction with CeO2; the medium temperature β peak is attributed to the reduction of CuO entering into the CeO2 lattice, and the high-temperature γ peak is attributed to the reduction of bulk CuO with weak or no interaction with CeO2 (Zheng et al. 2016). The temperature and area of each peak of the catalysts are summarized in Table 3.
Table 3
Quantitative data of the H2-TPR tests over the different CuO-CeO2 catalysts
Catalyst | Temperature of peaks / oC | Area of peaks / (a.u.) and relative intensities* / % |
Tα | Tβ | Tγ | Aα | Aβ | Aγ |
CC-N | 165 | 192 | 217 | 126(27) | 195 (42) | 146 (31) |
CC-NH | 169 | 195 | 205 | 91 (20) | 219(49) | 136 (31) |
* It is calculated according to the proportion of each peak in the whole reduction peak. |
It can be observed from Table 3 that the Tα and Tβ of the CC-N catalyst are slightly lower than those of the CC-NH catalyst, suggesting that CC-N has stronger reducibility. However, the Tγ of the CC-N catalyst is higher than that of the CC-NH catalyst. Based on the fact that the larger the grain size of CuO crystallites is, the higher the reduction temperature is (Zheng et al. 2016), it can be judged that the size of CuO crystallites in the bulk phase of the CC-N catalyst is larger than that of the CC-NH catalyst, which is consistent with the previous XRD characterization results (Table 1). In addition, it can be seen from Table 3 that the fraction of Aγ in the two catalysts is the same (both of 31%), indicating that the relative content of CuO in the bulk phase is the same. However, the fraction of Aα in CC-N was higher than that in CC-NH (27 vs 20%), but the fraction of Aβ was lower than that in CC-NH (42 vs 49%). These observations suggest that the relative content of highly dispersed CuO with strong interaction with CeO2 in CC-N was greater than that in CC-NH, and the relative amount of CuO entering into the lattice of CeO2 is less than that of CC-NH. These results clearly indicate the distinction in the distribution of Cu species in different CuO-CeO2 catalysts prepared with cerium nitrate and ammonium cerium nitrate as precursors.
Oxygen adsorption analysis (O2-TPD)
To further investigate the defects on the catalyst surface and the mobility of adsorbed oxygen, O2-TPD characterization was performed. In general, the oxygen species adsorbed on the catalyst surface undergo the following transformation process with increasing electron content: O2(ad) → O2−(ad) → O−(ad) → O2−(ad/lattice). O2(ad) refers to physically adsorbed oxygen, which is usually purged out by helium flow before the desorption temperature rises. The adsorbed oxygen species O2−(ad) and O−(ad) have weak bonding on the catalyst surface and are therefore easily desorbed. O2−(ad/lattice) is surface or bulk phase lattice oxygen and is difficult to be desorbed. According to the literature (He et al. 2022), the desorption peaks below 350 oC generally originate from the surface adsorbed O2−(ad) and/or O−(ad), which are surface oxygen species associated with surface defects of the catalyst, and the peaks above 350 oC from the desorption of surface lattice oxygen and bulk phase lattice oxygen. Therefore, it can be seen from Fig. 7 that the adsorbed oxygen species of CC-N are more active than those of CC-NH.
Reaction Performance Of Catalysts
Activity of catalysts
Figure 8 presents the activity of CuO-CeO2 catalysts prepared by different cerium precursors for CO oxidation at low temperatures. Obviously, CO conversion of both samples gradually increased with the elevation in reaction temperature. In addition, the CO conversion of CC-N is evidently higher than that of CC-NH at all temperature investigated here. Specifically, T30, T50, and T90 (corresponding to the temperature at which the CO conversion equals to 30, 50 and 90%, respectively) of the catalysts are collected in Table 4. Clearly, the T30, T50, and T90 values of CC-N are noticeably lower than those of CC-NH, indicating that CC-N prepared by cerium nitrate shows better catalytic activity for CO oxidation at low temperatures.
Table 4
T30, T50 and T90 of various catalysts for CO oxidation
Sample | T30 / oC | T50 / oC | T90 / oC |
CC-N | 59 | 68 | 86 |
CC-NH | 76 | 84 | 100 |
Kinetic test of catalysts
The apparent activation energies (Ea) of the CC-N and CC-NH catalysts were estimated on the base of CO oxidation results at low temperatures (< 55oC) by using an Arrhenius plot (Fig. 9), which are 61.3 and 63.8 kJ/mol, respectively. The Ea data are consistent with the activity test results of both catalysts (Fig. 8), since lower Ea corresponds to higher activity of the catalyst.
Correlation Between Physicochemical Properties And Catalytic Activity
Based on the results of characterization and activity testing, the correlation between physicochemical properties and catalytic activities of the CuO-CeO2 catalysts prepared by different cerium precursors can be obtained as follows:
(1) Generally speaking, the larger specific surface area and pore volume of catalysts can provide more active sites for adsorption and activation of reactants, which is conducive to the improvement of catalytic activity (Zhao et al. 2018; Xie et al. 2020). N2 adsorption/desorption characterization results (Table 1) indicated that the specific surface area and pore volume of CC-N were significantly larger than those of CC-NH, so the higher activity of CC-N (Fig. 8 and Table 4) could be attributed to its larger specific surface area and pore volume to some extent.
(2) It is well accepted that the activity of CuO-CeO2 catalyst for CO catalytic oxidation is closely related to the CuO species on its surface. It is generally believed that the highly dispersed CuO is the active species, while the bulk CuO contributes little to the activity or even has negative effects (Yang et al. 2014; Zheng et al. 2016). According to the TPR characterization results (Table 3), although the relative amount of bulk CuO in the CC-N and CC-NH is the same, the amount of highly dispersed CuO in CC-N is significantly higher than that in CC-NH. Therefore, the existence of more highly dispersed CuO species on the surface of CC-N is an important factor for its higher activity.
(3) It is well known that the adsorption of CO as one of the reactant molecules on the Cu+ species of CuO-CeO2 catalyst is an important step in CO oxidation. Therefore, more Cu+ on the surface of the catalyst will be conducive to the adsorption of CO, which is beneficial to the improvement of catalytic activity (Sun et al. 2015b; Zhang et al. 2018b). Therefore, the higher activity of CC-N can be attributed to the presence of more amount of Cu+ species on its surface, which can be confirmed by Cu LMM (Fig. 4(B) and Table 2) and in-situ DRIFTs analysis results (Fig. 5).
(4) According to the mechanism of CO oxidation reaction on CuO-CeO2 catalyst, that is, CO adsorbed by Cu+ reacts with oxygen species on the catalyst surface to generate CO2 (Lykaki et al. 2018; Zhang et al. 2020), it can be inferred that the CO oxidation activity of CuO-CeO2 catalyst may be related to the number and reactivity of oxygen species on the catalyst surface. Therefore, the higher activity of CC-N can be attributed to the larger number and stronger reactivity of oxygen species on its surface, which were evidenced by the results obtained from XPS analysis (Fig. 4) and O2-TPD analysis (Fig. 7), respectively.
Anti-toxicity Of Catalysts
As we know, water or CO2 resistance of a catalyst is very important indexes for evaluating its CO oxidation performance, especially at low temperatures. Therefore, the water or CO2 resistance of the CuO-CeO2 catalysts prepared by different cerium precursors was investigated here.
As shown in Fig. 10, the activity of the two catalysts under moisture conditions with different contents of water vapor (0.6vol% and 4.2vol%) is almost the same as that under a dry atmosphere, indicating that both catalysts have preeminent water resistance.
As shown in Fig. 11, the CC-N catalyst achieved 99% CO conversion at 110 oC with 10 vol% CO2 in the reaction gas, which is 10 oC higher than the T99 for the reaction gas without CO2 addition; while the T99 of the CC-NH catalyst increased by 30 oC with the addition of CO2 to the reaction gas. These results clearly indicate that the CC-N catalyst has better resistance to CO2 than CC-NH. It is generally believed that the reduction in the activity of CuO-CeO2 catalysts caused by CO2 is mainly due to competitive adsorption at the active site and the formation of carbonates that inhibit oxygen mobility. Cecilia et al. (2015) suggested that the highly dispersed CuO species in close contact with CeO2 are the active sites for CO oxidation reaction but weakly react with CO2, so the higher the content of the highly dispersed CuO species the better the resistance to CO2 poisoning. In addition, He et al. (2021) claimed that Cu+ sites preferentially adsorb CO compared to CO2, so more amount of Cu+ in the catalyst is favorable for enhancing its resistance to CO2. In this work, the amount of finely dispersed CuO, which strongly interacts with CeO2, is significantly larger in CC-N than in CC-NH, as evidenced by the characterization results of TPR (Table 3). In addition, the amount of Cu+ in the CC-N catalyst is larger than in CC-NH, as shown by the characterization results of XPS (Table 2) and CO-IR (Fig. 5). Furthermore, the results of N2 physical adsorption/desorption characterization (Table 1) indicate that the specific surface area and pore volume of CC-N are significantly larger than those of CC-NH, and the larger specific surface area and pore volume of the catalyst not only provide more active sites for adsorption and activation of reactant molecules but also can accommodate more deposition of the resulting carbonate. Therefore, we suggest that the better resistance of CC-N to CO2 poisoning can be attributed to its larger specific surface area and pore volume, and the presence of more amounts of highly dispersed CuO species in close contact with CeO2 and Cu+ species.