Enhancing the Ethynylation Performance of CuO-Bi 2 O 3 Nanocatalysts by Tuning Cu-Bi Interactions and Phase Structures

: Catalytic systems consisting of copper oxide and bismuth oxide are commonly employed for the industrial production of 1,4-butynediol (BD) through ethynylation. However, few studies have investigated the inﬂuence mechanism of Bi for these Cu-based catalysts. Herein, a series of nanostructured CuO-Bi 2 O 3 catalysts were prepared by co-precipitation followed by calcination at different temperatures. The obtained catalysts were applied to the ethynylation reaction. The textural and crystal properties of the catalysts, their reduction behavior, and the interactions between copper and bismuth species, were found to strongly depend on temperature. When calcined at 600 ◦ C, strong interactions between Cu and Bi in the CuO phase facilitated the formation of highly dispersed active cuprous sites and stabilized the Cu + valency, resulting in the highest BD yield. Bi 2 O 3 was completely absent when calcined at 700 ◦ C, having been converted into the spinel CuBi 2 O 4 phase. Spinel Cu 2+ was released gradually to form active Cu + species over eight catalytic cycles, which continuously replenished the decreasing activity resulting from the formation of metallic Cu and enhanced catalytic stability. Moreover, the positive correlation between the in-situ-formed surface Cu + ions and BD yield suggests that the amount of Cu + ions is the key factor for ethynylation of formaldehyde to BD on the as prepared CuO-Bi 2 O 3 catalysts. Based on these results and the literature, we propose an ethynylation reaction mechanism for CuO-based catalysts and provide a simple design strategy for highly efﬁcient catalytic CuO-Bi 2 O 3 systems, which has considerable potential for industrial applications. simultaneously activates C 2 H 2 and HCHO. Hence, the present observations and the literature allow us to propose a plausible ethynylation mechanism for CuO-based catalysis. Our work highlights a simple method for the rational design of highly active and stable CuO-based catalysts that simultaneously solves the problem of over-reduction and the aggregation of copper species and maintains sustainable catalysis during the phase transformation of copper oxides, thereby demonstrating important academic and industrial value.


Structure, Texture, and Morphology
To demonstrate the structural evolution, XRD patterns of fresh catalysts are displayed in Figure 1, along with that of pure CuO-600 for comparison. Here "CB" refers to the copper-bismuth catalyst and the number refers to the calcination temperature. Peaks observed at 32.6 • , 35 . In addition, a small peak at 27.7 • , which is attributed to Bi 2 O 3 (121) diffraction (JCPDS41-1449), is also observed in the case of CB-500 [29]. Calcination at 600 • C resulted in dramatic increases in the intensities of the CuO and Bi 2 O 3 XRD peaks. In addition, the weak peaks observed at 28.1 • and 33.2 • are assigned to CuBi 2 O 4 (211) and (102) diffractions (JCPDS48-1886), which is indicative of the formation of the spinel CuBi 2 O 4 phase mainly due to solid-solid interactions between CuO and Bi 2 O 3 [30,31]. The diffraction peaks corresponding to Bi 2 O 3 were entirely absent in the XRD pattern of the material calcined at 700 • C, while the CuBi 2 O 4 diffraction peaks were noticeably more intense. To demonstrate the structural evolution, XRD patterns of fresh catalysts are displayed in Figure  1, along with that of pure CuO-600 for comparison. Here "CB" refers to the copper-bismuth catalyst and the number refers to the calcination temperature. Peaks observed at 32.6°, 35.6°, 38.7°, 48.8°, 53.5°, 58.3°, 61.5°, 66.2°, and 68.1° in the XRD pattern of CuO are assigned to CuO (110), (002), (111), (−202), (020), (202), (−113), (−311), (220), and (004) diffractions, respectively (JCPDS45-0937). In addition, a small peak at 27.7°, which is attributed to Bi2O3 (121) diffraction (JCPDS41-1449), is also observed in the case of CB-500 [29]. Calcination at 600 °C resulted in dramatic increases in the intensities of the CuO and Bi2O3 XRD peaks. In addition, the weak peaks observed at 28.1° and 33.2° are assigned to CuBi2O4 (211) and (102) diffractions (JCPDS48-1886), which is indicative of the formation of the spinel CuBi2O4 phase mainly due to solid-solid interactions between CuO and Bi2O3 [30,31]. The diffraction peaks corresponding to Bi2O3 were entirely absent in the XRD pattern of the material calcined at 700 °C, while the CuBi2O4 diffraction peaks were noticeably more intense.

Characterizing Fresh Catalysts
The CuO crystallite sizes, calculated from the line broadening of the most intense XRD reflections using the Scherrer equation, are listed in Table 1. As the calcination temperature was increased from 500 to 600 °C, the CuO-particle size sharply increased from 17.2 to 33.0 nm. However, elevating the temperature to 700 °C resulted in only a further slight increase in the CuO particle size, to 39.4 nm, indicating that the gradual formation of CuBi2O4 inhibits the further growth of CuO crystals to some extent.   The CuO crystallite sizes, calculated from the line broadening of the most intense XRD reflections using the Scherrer equation, are listed in Table 1. As the calcination temperature was increased from 500 to 600 • C, the CuO-particle size sharply increased from 17.2 to 33.0 nm. However, elevating the temperature to 700 • C resulted in only a further slight increase in the CuO particle size, to 39.4 nm, indicating that the gradual formation of CuBi 2 O 4 inhibits the further growth of CuO crystals to some extent. Taking the CuO(111) diffraction peak at about 38.9 • as an example, this peak in the patterns of the CB-500 and CB-600 samples is shifted to lower diffraction angles compared to that of pure CuO-600, to 2θ values of about 38.8 • and 38.7 • , respectively. These observations indicate that Bi 3+ ions thermally diffuse into the CuO crystal lattice during calcination at 500 • C and 600 • C. Based on their ionic radii (Bi 3+ 0.96 Å, Cu 2+ 0.72 Å), we infer that the Bi 3+ ions doped into the CuO crystal structure in two ways, namely through substitutional and interstitial doping. Both ways expand the cell volume of the CuO crystal [32], which suggests that Cu and Bi interact. The XRD patterns reveal that the CuO in CB-600 exhibits the maximum cell volume, which demonstrates the existence of stronger Cu-Bi interactions in the CuO of the CB-600 catalyst compared to the others. Interestingly, the CuO(111) peak shifted to the same position as that observed for pure CuO in the case of CB-700, suggestive of a lack of interaction between the Cu and Bi in the CuO phase in this catalyst. Figure 2 shows Raman spectra of fresh CuO-Bi 2 O 3 catalysts and pure CuO-600. Three peaks-at about 293.3 cm −1 (A g mode), 345.4 cm −1 (B g mode), and 626.4 cm −1 (B g mode)-in pure CuO-600 are attributed to the vibrations of the oxygen atoms in the lattice (Cu-O and Cu-O-Cu bonds) [33,34]. Compared with pure CuO, the Raman peaks of CB-500 and CB-600 were broader and shifted to lower frequencies; the main peak at 293.3 cm −1 observed for pure CuO shifted to 293.0 cm −1 for CB-500, and to 284.5 cm −1 for CB-600. The red-shifting and broadening of the peaks in the Raman spectra are mainly ascribable to defects in CuO originating from interactions between Cu and Bi [35,36]. Combining this information with the XRD results, we ascertain that Bi 3+ ions diffuse into the CuO lattice to distort the CuO structure. Since a charge difference exists between Bi 3+ and Cu 2+ , the introduction of Bi generates oxygen vacancies in the CuO lattice to balance the charge [35,36]. Compared to CB-500, the CuO peaks in the spectrum of CB-600 were observed at lower frequencies, indicating the existence of stronger interactions in the CB-600 catalyst. No obvious Raman features corresponding to Bi 2 O 3 were observed in the CB-500 and CB-600 spectra, possibly due to its lower content and crystallinity. In addition, four distinct vibrational peaks-at 130. 9 Taking the CuO(111) diffraction peak at about 38.9° as an example, this peak in the patterns of the CB-500 and CB-600 samples is shifted to lower diffraction angles compared to that of pure CuO-600, to 2θ values of about 38.8° and 38.7°, respectively. These observations indicate that Bi 3+ ions thermally diffuse into the CuO crystal lattice during calcination at 500 °C and 600 °C. Based on their ionic radii (Bi 3+ 0.96 Å, Cu 2+ 0.72 Å), we infer that the Bi 3+ ions doped into the CuO crystal structure in two ways, namely through substitutional and interstitial doping. Both ways expand the cell volume of the CuO crystal [32], which suggests that Cu and Bi interact. The XRD patterns reveal that the CuO in CB-600 exhibits the maximum cell volume, which demonstrates the existence of stronger Cu-Bi interactions in the CuO of the CB-600 catalyst compared to the others. Interestingly, the CuO(111) peak shifted to the same position as that observed for pure CuO in the case of CB-700, suggestive of a lack of interaction between the Cu and Bi in the CuO phase in this catalyst.  [33,34]. Compared with pure CuO, the Raman peaks of CB-500 and CB-600 were broader and shifted to lower frequencies; the main peak at 293.3 cm −1 observed for pure CuO shifted to 293.0 cm −1 for CB-500, and to 284.5 cm −1 for CB-600. The red-shifting and broadening of the peaks in the Raman spectra are mainly ascribable to defects in CuO originating from interactions between Cu and Bi [35,36]. Combining this information with the XRD results, we ascertain that Bi 3+ ions diffuse into the CuO lattice to distort the CuO structure. Since a charge difference exists between Bi 3+ and Cu 2+ , the introduction of Bi generates oxygen vacancies in the CuO lattice to balance the charge [35,36]. Compared to CB-500, the CuO peaks in the spectrum of CB-600 were observed at lower frequencies, indicating the existence of stronger interactions in the CB-600 catalyst. No obvious Raman features corresponding to Bi2O3 were observed in the CB-500 and CB-600 spectra, possibly due to its lower content and crystallinity. In addition, four distinct vibrational peaks-at 130.9, 261.8, 402.8, and 586.6 cm −1 -were observed for the CB-700 sample, which corresponds to the presence of CuBi2O4 [30,31].  Table 1 presents the BET specific surface areas of the fresh catalysts. CB-500 exhibits the largest surface area (23.2 m 2 g −1 ); however, a significantly reduced surface area (8.7 m 2 g −1 ) was obtained for  Table 1 presents the BET specific surface areas of the fresh catalysts. CB-500 exhibits the largest surface area (23.2 m 2 g −1 ); however, a significantly reduced surface area (8.7 m 2 g −1 ) was obtained for the sample calcined at 700 • C. This decrease in surface area is normally associated with the growth of larger particles [27].
The morphologies and microstructures of the fresh catalysts were investigated by transmission electron microscopy (TEM). As shown in Figure 3, the catalysts have nearly spherical or irregular polyhedral morphologies. Elevated calcination temperatures lead to larger catalyst particles that are randomly distributed, which is ascribable to metal oxide sintering at high temperatures. of larger particles [27].
The morphologies and microstructures of the fresh catalysts were investigated by transmission electron microscopy (TEM). As shown in Figure 3, the catalysts have nearly spherical or irregular polyhedral morphologies. Elevated calcination temperatures lead to larger catalyst particles that are randomly distributed, which is ascribable to metal oxide sintering at high temperatures.

X-ray Photoelectron Spectroscopy
The fresh CuO-Bi2O3 catalysts were subjected to XPS in order to identify the surface chemical states of their copper species; the Cu2p XPS spectra are displayed in Figure 4. All catalysts displayed two peaks, one centered at about 933.1 eV and the other at about 955.9 eV, which are attributed to Cu2p3/2 and Cu2p1/2, respectively. The main Cu2p3/2 peak at 933.1 eV and its satellite at about 940-945 eV are characteristic of Cu 2+ species [37,38]. Major Cu2p3/2 peaks are observed at around 933.1 eV in the spectra of CB-500 and CB-700; however, this peak was observed at 933.9 eV in the spectrum of CB-600. The positive binding-energy (BE) shifts suggest charge transfer from the metal ions to other components, which is indicative of strong interactions between Cu and Bi [39,40]. In addition, the shoulder peak at 936.0 was observed to gradually increase with increasing calcination temperature, which is attributed to spinel Cu 2+ species [26].

Reduction Behavior
To investigate the interactions between the different components, we employed H2-TPR experiments. As shown in Figure 5, the reduction profiles can be divided into α (around 220-450 °C),

X-ray Photoelectron Spectroscopy
The fresh CuO-Bi 2 O 3 catalysts were subjected to XPS in order to identify the surface chemical states of their copper species; the Cu2p XPS spectra are displayed in Figure 4. All catalysts displayed two peaks, one centered at about 933.1 eV and the other at about 955.9 eV, which are attributed to Cu2p 3/2 and Cu2p 1/2 , respectively. The main Cu2p 3/2 peak at 933.1 eV and its satellite at about 940-945 eV are characteristic of Cu 2+ species [37,38]. Major Cu2p 3/2 peaks are observed at around 933.1 eV in the spectra of CB-500 and CB-700; however, this peak was observed at 933.9 eV in the spectrum of CB-600. The positive binding-energy (BE) shifts suggest charge transfer from the metal ions to other components, which is indicative of strong interactions between Cu and Bi [39,40]. In addition, the shoulder peak at 936.0 was observed to gradually increase with increasing calcination temperature, which is attributed to spinel Cu 2+ species [26].
The morphologies and microstructures of the fresh catalysts were investigated by transmission electron microscopy (TEM). As shown in Figure 3, the catalysts have nearly spherical or irregular polyhedral morphologies. Elevated calcination temperatures lead to larger catalyst particles that are randomly distributed, which is ascribable to metal oxide sintering at high temperatures.

X-ray Photoelectron Spectroscopy
The fresh CuO-Bi2O3 catalysts were subjected to XPS in order to identify the surface chemical states of their copper species; the Cu2p XPS spectra are displayed in Figure 4. All catalysts displayed two peaks, one centered at about 933.1 eV and the other at about 955.9 eV, which are attributed to Cu2p3/2 and Cu2p1/2, respectively. The main Cu2p3/2 peak at 933.1 eV and its satellite at about 940-945 eV are characteristic of Cu 2+ species [37,38]. Major Cu2p3/2 peaks are observed at around 933.1 eV in the spectra of CB-500 and CB-700; however, this peak was observed at 933.9 eV in the spectrum of CB-600. The positive binding-energy (BE) shifts suggest charge transfer from the metal ions to other components, which is indicative of strong interactions between Cu and Bi [39,40]. In addition, the shoulder peak at 936.0 was observed to gradually increase with increasing calcination temperature, which is attributed to spinel Cu 2+ species [26].

Reduction Behavior
To investigate the interactions between the different components, we employed H2-TPR experiments. As shown in Figure 5, the reduction profiles can be divided into α (around 220-450 °C),

Reduction Behavior
To investigate the interactions between the different components, we employed H 2 -TPR experiments. As shown in Figure 5, the reduction profiles can be divided into α (around 220-450 • C), β (around 450-550 • C), and γ (around 550-700 • C) peaks. The α peak is attributed to the reduction of CuO. The stepwise reduction of CuO, i.e., Cu 2+ to Cu + (low temperature) and Cu + to Cu 0 (high temperature) [41,42], can be discounted because of the fast H 2-flow rate and deviations from the ideal pattern (the area of the low-temperature reduction peak is ideally equal to that of the high-temperature peak). The α peak was observed to shift toward higher temperatures with increasing calcination temperature. In the present experiment, both increases in CuO-particle size and interactions between copper and bismuth species improve the reduction temperature [43,44]. To identify the influence of Cu-Bi interactions, we also examined the H 2 -TPR of pure CuO samples ( Figure 6) that were calcined at different temperatures to produce CuO particles of equivalent sizes to those of the corresponding CuO-Bi 2 O 3 . Symmetrical reduction peaks were observed for all pure CuO samples, with reduction temperatures of 257, 326, and 390 • C. By comparing the shapes and temperatures of the reduction peaks for the CB-500 and CB-600 catalysts, we ascribe the first peak-centered at about 266 • C or 339 • C-to the reduction of isolated bulk CuO and/or weak interactions between the CuO and Bi species [39], and the second peak-centered at about 334 • C or 385 • C-to the reduction of CuO that strongly interacts with Bi 2 O 3 [43,44]. Clearly, the peak area corresponding to strong CuO-Bi 2 O 3 interactions is larger in the CB-600 profile than the CB-500 profile, suggesting that stronger interaction exist within CB-600. The reduction profile of CuO-700 is very close to that of the CB-700 catalyst. Moreover, the XRD and Raman results reveal that large CuO particles devoid of interactions are observed in the CB-700 sample. Hence, the single symmetrical peak observed at 392 • C in the profile of the CB-700 catalyst is attributable to the reduction of large CuO particles. These observations suggest that calcination temperature affects the interactions between Cu and Bi and, as a consequence, the reduction behavior of the Cu species. This also illustrates that the reducibility of copper species is hindered by proper interactions between the Cu and Bi species, which might protect the copper species against further reduction by HCHO in the ethynylation reaction. temperature) [41,42], can be discounted because of the fast H2-flow rate and deviations from the ideal pattern (the area of the low-temperature reduction peak is ideally equal to that of the hightemperature peak). The α peak was observed to shift toward higher temperatures with increasing calcination temperature. In the present experiment, both increases in CuO-particle size and interactions between copper and bismuth species improve the reduction temperature [43,44]. To identify the influence of Cu-Bi interactions, we also examined the H2-TPR of pure CuO samples ( Figure 6) that were calcined at different temperatures to produce CuO particles of equivalent sizes to those of the corresponding CuO-Bi2O3. Symmetrical reduction peaks were observed for all pure CuO samples, with reduction temperatures of 257, 326, and 390 °C. By comparing the shapes and temperatures of the reduction peaks for the CB-500 and CB-600 catalysts, we ascribe the first peakcentered at about 266 °C or 339 °C-to the reduction of isolated bulk CuO and/or weak interactions between the CuO and Bi species [39], and the second peak-centered at about 334 °C or 385 °C-to the reduction of CuO that strongly interacts with Bi2O3 [43,44]. Clearly, the peak area corresponding to strong CuO-Bi2O3 interactions is larger in the CB-600 profile than the CB-500 profile, suggesting that stronger interaction exist within CB-600. The reduction profile of CuO-700 is very close to that of the CB-700 catalyst. Moreover, the XRD and Raman results reveal that large CuO particles devoid of interactions are observed in the CB-700 sample. Hence, the single symmetrical peak observed at 392 °C in the profile of the CB-700 catalyst is attributable to the reduction of large CuO particles. These observations suggest that calcination temperature affects the interactions between Cu and Bi and, as a consequence, the reduction behavior of the Cu species. This also illustrates that the reducibility of copper species is hindered by proper interactions between the Cu and Bi species, which might protect the copper species against further reduction by HCHO in the ethynylation reaction. These reduction properties may be related to the coordination structures and electron charge densities (ECDs) of the copper species [45,46]. The partial incorporation of Bi 3+ into the CuO lattice (the Cu 2+ ions are at centers of inversion symmetry in a single four-fold 4c site [47]) results in structural distortion, based on the above XRD and Raman results; hence, the Cu 2+ -coordination structures will also be distorted [32,33]. Moreover, the XPS results reveal that copper species that interact with the bismuth species exhibite higher BEs, indicating that the copper species have low outer-shell electron densities. This result may explain why the copper species that interact with bismuth species are more difficult to reduce than bulk CuO under identical conditions. The above two points may be related to the intrinsic nature of Cu-Bi interactions.
In the higher-temperature region (>450 °C), the β and γ peaks mark the reductions of Bi2O3 and These reduction properties may be related to the coordination structures and electron charge densities (ECDs) of the copper species [45,46]. The partial incorporation of Bi 3+ into the CuO lattice (the Cu 2+ ions are at centers of inversion symmetry in a single four-fold 4c site [47]) results in structural distortion, based on the above XRD and Raman results; hence, the Cu 2+ -coordination structures will also be distorted [32,33]. Moreover, the XPS results reveal that copper species that interact with the bismuth species exhibite higher BEs, indicating that the copper species have low outer-shell electron densities. This result may explain why the copper species that interact with bismuth species are more difficult to reduce than bulk CuO under identical conditions. The above two points may be related to the intrinsic nature of Cu-Bi interactions.
In the higher-temperature region (>450 • C), the β and γ peaks mark the reductions of Bi 2 O 3 and CuBi 2 O 4 , respectively [29][30][31]. Compared to that of the CB-500 sample, the β-reduction-peak area of CB-600 was lower and a γ peak at 578 • C emerged due to the partial transformation of Bi 2 O 3 into CuBi 2 O 4 . The γ peak was even larger and shifted to a slightly higher temperature (608 • C) in the profile of CB-700, while the β peak was absent, suggesting that Bi 2 O 3 had completely transformed into CuBi 2 O 4 , with larger particles, which is consistent with the XRD results. Clearly, the Cu 2+ species in the spinel matrix are less reducible than those in the non-spinel CuO phase. Previous authors have also observed similar phenomena in Cu-Al spinel catalysts [26,48].
In comparison to CB-500 and CB-700, CB-600 contains more copper species that interact with bismuth, which may facilitate the effective conversion of CuO into active cuprous species, thereby enhancing its catalytic performance. It should be noted that CB-700 contains more spinel copper species that seem to play crucial roles in the catalytic stability, as discussed below.

Structure and Texture
Cuprous species (cuprous acetylides) are considered to be the active species in the ethynylation reaction, which are formed in situ from CuO during the reaction [1]. These species undergo phase transformations including reduction from Cu 2+ to Cu + followed by coordination to acetylene. Hence, we employed XRD, BET, and CO-IR techniques to trace the changes in the structures, textures, and surface valencies of the used catalysts. Herein, we characterize the CuO-Bi 2 O 3 catalysts after 15 h of evaluation.
The XRD patterns of the used samples shown in Figure 7 exhibit distinct diffraction-peak differences. No diffraction peaks related to CuO are observed in the pattern of CB-600. Meanwhile, weak broad peaks at about 28.1 • , 33.0 • , and 46.9 • emerged, which may belong to amorphous active cuprous species, which indicates that these active cuprous species form during the complete transformation of CuO and the less-crystalline CuBi 2 O 4 during ethynylation, which is beneficial for increased activity [16,17]. With the exception of the weak broad peaks, a diffraction peak of metallic Cu (JCPDS04-0836) was detected at a 2θ value of 43.3 • for CB-500, indicating that Cu 2+ ions were partially reduced to metallic Cu by HCHO in the ethynylation reaction [16,17]. Based on the H 2 -TPR profiles of the fresh catalysts, the appearance of metallic Cu can be ascribed to noninteracting copper species, which are easily reduced in the CB-500 catalyst. In contrast, some of the CuO and CuBi 2 O 4 XRD peaks remained in the pattern of the CB-700 sample, suggesting that this catalyst was not completely transformed into active cuprous species because of the weak reducibility of the copper species in the large CuO particles and spinel matrices.
BET data for the used catalysts are listed in Table 1. Compared to the corresponding fresh catalysts, the surface areas of the used catalysts were all lower to various extents. This observation is possibly ascribable to the collapse of porous channels and shrinkage of the frameworks of copper oxide species during phase transformation [14,15]. The used CB-600 catalyst shows a minimal decrease in S BET , suggesting that strong Cu-Bi interactions effectively hinder the migration and aggregation of catalyst particles, and retain the original framework of the fresh catalysts to an extent. Moreover, the used CB-600 also exhibits the maximal S BET , which is beneficial for exposing active sites.
partially reduced to metallic Cu by HCHO in the ethynylation reaction [16,17]. Based on the H2-TPR profiles of the fresh catalysts, the appearance of metallic Cu can be ascribed to noninteracting copper species, which are easily reduced in the CB-500 catalyst. In contrast, some of the CuO and CuBi2O4 XRD peaks remained in the pattern of the CB-700 sample, suggesting that this catalyst was not completely transformed into active cuprous species because of the weak reducibility of the copper species in the large CuO particles and spinel matrices. BET data for the used catalysts are listed in Table 1. Compared to the corresponding fresh catalysts, the surface areas of the used catalysts were all lower to various extents. This observation is possibly ascribable to the collapse of porous channels and shrinkage of the frameworks of copper oxide species during phase transformation [14,15]. The used CB-600 catalyst shows a minimal decrease in SBET, suggesting that strong Cu-Bi interactions effectively hinder the migration and aggregation of catalyst particles, and retain the original framework of the fresh catalysts to an extent. Moreover, the used CB-600 also exhibits the maximal SBET, which is beneficial for exposing active sites.

CO-IR Spectroscopy
To further determine the valencies and chemical environments of the surface copper species, CO-IR spectra were recorded and are shown in Figure 8. As detailed previously, the linear adsorption of CO on Cu + species gives rise to peaks with vibrational frequencies at about 2160-2110 cm −1 [49][50][51][52][53][54]. However, Cu 0 −CO and Cu 2+ −CO species are usually unstable and do not exist at room temperature, especially following the evacuation process. Thus, based on our experiment conditions and literature data, we attribute the strong peak at 2110 cm −1 to linearly adsorbed CO on Cu + . In addition, the intensity of this peak followed the order: CB-600 > CB-500 > CB-700, which corresponds to the ordering of the BET specific surface areas of the used catalysts. The results of the integrated areas of the Cu 1+ −CO band are listed in Table 1, which reveals that the used CB-600 catalyst has the highest number of exposed active Cu + sites, which is due to strong interactions between copper and bismuth species. On the basis of the XRD and Raman results from the calcined catalysts, we conclude that Bi 3+ ions diffuse into the CuO lattice and distort the CuO structure. Doping with bismuth species may play a role in diluting the copper species. During phase evolution, the strong interactions between the Cu and Bi species hinder the migration and aggregation of copper species to retain the framework of the fresh catalysts to some extent. The XPS and H2-TPR results confirm that the outer shell electron configurations and the reduction behavior of Cu 2+ are affected by bismuth-species doping regulated

CO-IR Spectroscopy
To further determine the valencies and chemical environments of the surface copper species, CO-IR spectra were recorded and are shown in Figure 8. As detailed previously, the linear adsorption of CO on Cu + species gives rise to peaks with vibrational frequencies at about 2160-2110 cm −1 [49][50][51][52][53][54]. However, Cu 0 −CO and Cu 2+ −CO species are usually unstable and do not exist at room temperature, especially following the evacuation process. Thus, based on our experiment conditions and literature data, we attribute the strong peak at 2110 cm −1 to linearly adsorbed CO on Cu + . In addition, the intensity of this peak followed the order: CB-600 > CB-500 > CB-700, which corresponds to the ordering of the BET specific surface areas of the used catalysts. The results of the integrated areas of the Cu 1+ −CO band are listed in Table 1, which reveals that the used CB-600 catalyst has the highest number of exposed active Cu + sites, which is due to strong interactions between copper and bismuth species. On the basis of the XRD and Raman results from the calcined catalysts, we conclude that Bi 3+ ions diffuse into the CuO lattice and distort the CuO structure. Doping with bismuth species may play a role in diluting the copper species. During phase evolution, the strong interactions between the Cu and Bi species hinder the migration and aggregation of copper species to retain the framework of the fresh catalysts to some extent. The XPS and H 2 -TPR results confirm that the outer shell electron configurations and the reduction behavior of Cu 2+ are affected by bismuth-species doping regulated by calcination temperature. The XRD results for the used catalysts further confirm that the over-reduction of Cu 2+ to metallic Cu can be effectively restrained and that the Cu + valency is stabilized in the used CB-600 catalyst. Similar observations have been reported by Di et al. and Wen et al. [44,55], who noted that the reduction behavior and the dispersion of copper species can be regulated by altering interactions between copper species and promoters or supports. by calcination temperature. The XRD results for the used catalysts further confirm that the overreduction of Cu 2+ to metallic Cu can be effectively restrained and that the Cu + valency is stabilized in the used CB-600 catalyst. Similar observations have been reported by Di et al. and Wen et al. [44,55], who noted that the reduction behavior and the dispersion of copper species can be regulated by altering interactions between copper species and promoters or supports.

Catalytic Performance
The main reaction involved in the ethynylation of formaldehyde can be written as follows: Firstly, we investigated the influence of reaction temperature on the catalytic activities of the CuO-Bi2O3 and CuO catalysts in a simulated slurry reactor at 90 °C and under atmospheric pressure. The main product from the ethynylation reaction is BD; by-products, such as propargyl alcohol, formic acid, and polyformaldehyde are not shown because of their low selectivities. The yield of BD over 15 h with different reaction temperatures are plotted in Figure 9a. The reaction temperature was found to have a great influence on the catalysts performance. The yield of BD could be greatly improved with the temperature elevated. When the reaction temperature was 90 °C or 100 °C, all the catalysts displayed relatively high activity. Figure 9b displays the ethynylation activities of the CuO-Bi2O3 and CuO catalysts as functions of time over 15 h at 90 °C. The CuO-based catalysts gradually became catalytic active after an induction period of around 2 h; this induction period possibly correlates with phase evolution in the catalysts. It is widely believed that cuprous acetylide is the actual active phase involved in the ethynylation reaction, rather than CuO. With extended reaction times, all catalysts showed increased activities. However, the BD yield first increased and then decreased with increasing calcination temperature. As shown in Table 1, the BD yield over CB-600 was 92.4% after 15 h. In contrast, the BD yields over CB-500 and CB-700 were 84% and 74.6% under the same conditions, respectively. Figure 10 reveals that the integrated areas of the CO-IR peaks correlate with reaction yield, which indicates that catalytic activity is proportional to the Cu + surface area and that surface cuprous species are responsible for the formation of BD.

Catalytic Performance
The main reaction involved in the ethynylation of formaldehyde can be written as follows: HC≡CH + 2HCHO → OH-CH 2 C≡CCH 2 -OH Firstly, we investigated the influence of reaction temperature on the catalytic activities of the CuO-Bi 2 O 3 and CuO catalysts in a simulated slurry reactor at 90 • C and under atmospheric pressure. The main product from the ethynylation reaction is BD; by-products, such as propargyl alcohol, formic acid, and polyformaldehyde are not shown because of their low selectivities. The yield of BD over 15 h with different reaction temperatures are plotted in Figure 9a. The reaction temperature was found to have a great influence on the catalysts performance. The yield of BD could be greatly improved with the temperature elevated. When the reaction temperature was 90 • C or 100 • C, all the catalysts displayed relatively high activity. Figure 9b displays the ethynylation activities of the CuO-Bi 2 O 3 and CuO catalysts as functions of time over 15 h at 90 • C. The CuO-based catalysts gradually became catalytic active after an induction period of around 2 h; this induction period possibly correlates with phase evolution in the catalysts. It is widely believed that cuprous acetylide is the actual active phase involved in the ethynylation reaction, rather than CuO. With extended reaction times, all catalysts showed increased activities. However, the BD yield first increased and then decreased with increasing calcination temperature. As shown in Table 1, the BD yield over CB-600 was 92.4% after 15 h. In contrast, the BD yields over CB-500 and CB-700 were 84% and 74.6% under the same conditions, respectively. Figure 10 reveals that the integrated areas of the CO-IR peaks correlate with reaction yield, which indicates that catalytic activity is proportional to the Cu + surface area and that surface cuprous species are responsible for the formation of BD.
The stabilities of the CuO-Bi 2 O 3 catalysts were examined by filtering off the catalysts after 15 h of reaction time and reusing them in another seven cycles under the reaction conditions of 90 • C and atmospheric pressure (Figure 11). The yields of BD after recycling are listed in Table 2. During the eight runs, marginal decreases in BD yield were observed for CB-600 (92.5% to 62.2%), while larger decreases were observed for CB-500 (86% to 32.8%). However, CB-700 showed optimal catalytic stability (72.5% to 65.1%), despite its lower activity. The formation of metallic Cu may be the predominant factor responsible for catalyst deactivation in this work. Hence, the CuO-Bi 2 O 3 catalysts were examined by XRD after 1, 4, and 8 cycles in order to investigate their phase evolution ( Figure 12). The CB-600 and CB-500 catalysts exhibited the formation and growth of metallic Cu that positively correlates with the observed decreases in catalyst stabilities. However, metallic Cu particles were observed to grow on the CB-700 catalyst, while a decline in the CuBi 2 O 4 phase was observed after the eight runs. Hence, active cuprous species gradually formed in situ from the spinel phase, which replenished the decline in activity resulting from a decrease in Cu + species. According to the H 2 -TPR results, spinel Cu 2+ species are more difficult to reduce than non-spinel Cu 2+ species under the same conditions; hence, the formation of active cuprous species from the CuBi 2 O 4 phase will be somewhat slower than that from the CuO phase. From the above discussion, we conclude that the excellent catalytic stability of CB-700 is due to synergism between the CuO and CuBi 2 O 4 phases. During ethynylation, the CB-700 catalyst is first activated by the formation of active cuprous species from non-spinel Cu 2+ , followed by the formation of active cuprous species through the gradual release of Cu 2+ from the spinel structure, as displayed in Figure 13. The stabilities of the CuO-Bi2O3 catalysts were examined by filtering off the catalysts after 15 h of reaction time and reusing them in another seven cycles under the reaction conditions of 90 °C and atmospheric pressure (Figure 11). The yields of BD after recycling are listed in Table 2. During the  The stabilities of the CuO-Bi2O3 catalysts were examined by filtering off the catalysts after 15 h of reaction time and reusing them in another seven cycles under the reaction conditions of 90 °C and atmospheric pressure (Figure 11). The yields of BD after recycling are listed in Table 2. During the eight runs, marginal decreases in BD yield were observed for CB-600 (92.5% to 62.2%), while larger    Based on the above analysis, we conclude that the temperature at which the CuO-Bi 2 O 3 catalyst is calcined greatly influences its phase structure, texture, and the interactions between the metal oxide precursors, as well as the transformations of active cuprous species and the corresponding catalytic performance. There is a range of calcination temperatures at which the catalyst exhibited the best activity. The excellent catalytic performance of Cu/Bi-based catalysts is ascribable to strong interactions between Cu and Bi. On the one hand, these strong interactions stabilize the Cu + valency and inhibit the formation of metallic Cu, which facilitates the efficient transformation of CuO into active cuprous species. On the other hand, this strong interaction retains the catalyst framework and hinders the migration and aggregation of catalyst particles, resulting in more active sites exposed on the catalyst surface. In addition, the gradual release of copper species from spinel structures formed at high temperatures sustainably replenishes catalytic activity [26,56], resulting in improved catalytic stability. Catalysts 2019, 9, x FOR PEER REVIEW 12 of 19

Reaction Mechanism
Given our present observations, phase transformation appears to be important during the reaction, in which the catalysts undergo reduction and carbonization by HCHO and C 2 H 2 ; this phase change can be expressed as: CuO → active cuprous species (Cu 2 C 2 ) → Cu. The first phase transformation results in induction behavior that activates the catalyst. The dispersion and valence stability of Cu + species in the activated catalyst are responsible for catalytic performance. The second step deactivates the CuO-Bi 2 O 3 catalyst. The active cuprous species formed in situ are sufficiently nucleophilic to react with formaldehyde. Hübner et al. [57] also observed net electron transfer from Cu + ions in zeolites that act as electron donors to acetylene by IR spectroscopy and by quantum-chemical calculations, which demonstrated that the acetylene molecules adsorbed on the Cu + ions are nucleophilic. In contrast, it has been suggested that Cu + acts at electrophilic or Lewis acidic sites, thereby improving the reactivity of the C=O bond [39,58,59]. The carbonyl group is coordinated to Cu + through a lone pair of electrons on oxygen, leading to a more-electrophilic carbonyl carbon atom [59]. This assumption was confirmed by the hydrogenation of C=O bonds on copper-based catalysts [34]. According to the above discussion and experimental observations, we have reason to believe that the carbonyl and alkynyl are simultaneously activated by Cu + and, consequently, propose a plausible mechanism to account for the CuO-catalyzed ethynylation ( Figure 14). This mechanism indirectly supports the study of Chu et al. who reported that the adsorption and activation of HCHO and C 2 H 2 molecules on catalyst surfaces are critical to the intrinsic reaction rates [60].

Preparation of the Catalysts
All chemical reagents, namely Cu(NO3)2, Bi(NO3)3, PEG-400, NaOH, were of analytical grade and were used as received. The CuO-Bi2O3 catalyst was prepared by the coprecipitation method. In a typical synthesis procedure, urea (100 g) and PEG-400 (100 g) were dissolved in 200 mL of a solution of Cu(NO3)2 (0.5 M) and Bi(NO3) 3 Figure 14. Plausible reaction mechanism for catalytic reactions involving CuO-Bi 2 O 3 .

Preparation of the Catalysts
All chemical reagents, namely Cu(NO 3 ) 2 , Bi(NO 3 ) 3 , PEG-400, NaOH, were of analytical grade and were used as received. The CuO-Bi 2 O 3 catalyst was prepared by the coprecipitation method. In a typical synthesis procedure, urea (100 g) and PEG-400 (100 g) were dissolved in 200 mL of a solution of Cu(NO 3 ) 2 (0.5 M) and Bi(NO 3 ) 3 (0.1 M) at room temperature. A solution of NaOH (200 mL, 2 M) was then added dropwise to the above solution at 80 • C with stirring. Following reaction, the precipitate was separated by centrifugation and washed three times (successively) with H 2 O and EtOH. The final product was dried at 80 • C for 12 h and calcined at 500, 600, or 700 • C for 3 h. The powders prepared in this manner are referred to as "CB-T", where T indicates the calcination temperature. In addition, the used catalysts were washed, filtered, and dried under vacuum at 60 • C for 48 h. Pure CuO was prepared in the same way as the CuO-Bi 2 O 3 samples. As summarized in Table 3, pure CuO was calcined at different temperatures to obtain samples with essentially equivalent CuO particle sizes to those of CB-500, CB-600, and CB-700. For ease of reference, these samples are referred to as "CuO-500", "CuO-600", and "CuO-700", respectively.

Characterization
XRD patterns were collected on a Bruker D8 Advance spectrometer (Bruker, Billerica, MA, USA) with a Cu Kα target (λ = 0.15418 nm) operating at 40 kV and 40 mA, at a scanning rate of 2.4 • /min and over the 10 • -80 • 2θ range. N 2 -Physisorption analyses were performed using a Micromeritics ASAP-2020 apparatus (Norcross, GA, USA). The fresh and used samples were degassed at 150 • C for 5 h and 60 • C for 24 h respectively prior to adsorption testing. Specific surface areas were obtained using the multi-point BET procedure.
H 2 -TPR experiments were performed using a Micromeritics AutoChemII 2920 instrument (Norcross, GA, USA). A 30 mg sample was loaded into the instrument. A 5 vol% mixture of H 2 in N 2 (30 mL/min) was introduced and the temperature was raised from room temperature to 700 • C at a rate of 10 • C/min. The consumption of H 2 was determined using a thermal conductivity detector, and the end gas was analyzed using an on-line quadrupole mass spectrometer (HIDEN HPR-20 QIC) (Beijing Henven Scientific Instrument Co., Ltd., Beijing, China) with a minimum dwell time of 3 ms.
Raman spectra were recorded with a LabRAM HR Evolution instrument (HORIBA, Tokyo, Japan) equipped with a CCD detector at room temperature. A 532 nm Ventus laser operating at 10 MW was used as the excitation source.
CO-IR spectra were recorded on a Bruker Tensor 27 Fourier-transform infrared spectrometer (Bruker, Billerica, MA, USA) equipped with a highly sensitive MCT detector cooled by liquid N 2 . The used samples were evacuated at 60 • C for 3 h.
XPS experiments were performed in an ultra-high vacuum chamber (THERMO Scientific K-Alpha spectrometer) (Thermo Fisher Scientific, Waltham, MA, USA) at a base pressure of −4 × 10 −10 Torr. The XPS spectrometer was equipped with a high-resolution hemispherical electron analyzer (279.4 mm diameter with 25 meV resolution) and a Mg Kα (hν = 1253.6 eV) X-ray excitation source. All spectra were referenced to adventitious C (1s) at a BE of 284.5 eV.
TEM experiments were performed using a JEOL JEM-2100 transmission electron microscope operating at 200 kV (JEOL, Tokyo, Japan). TEM samples were dispersed by sonication in ethanol followed by deposition of the suspension onto a standard Cu grid covered with a porous carbon film.

Catalyst Test
A 0.5 g sample of the catalyst and 10 mL of aqueous formaldehyde (30 vol%) were sequentially added to a 100-mL round bottom three-port flask equipped with a thermometer and a condenser tube. The flask was purged with N 2 and the system was heated to 90 • C in an oil bath with stirring, after which C 2 H 2 gas was introduced for 15 h. The mixture was cooled to room temperature following the reaction. The solid catalyst was removed by centrifugation and quantitatively analyzed on an Agilent 7890A gas chromatograph fitted with a DB-5 capillary column (0.32 mm × 50 m) and an FID detector. 1,4-Butanediol was used as the internal standard. The unconverted formaldehyde in solution was determined by iodimetry.

Conclusions
In the present study, a series of nanostructured CuO-Bi 2 O 3 catalysts was prepared using simple coprecipitation methods followed by calcination at different temperatures, after which they were applied to the ethynylation of formaldehyde to produce BD. We reveal that the calcination temperature greatly influences the crystal phase structures of the CuO-Bi 2 O 3 catalysts and interactions between the components. The CB-600 catalysts demonstrated strong interactions between CuO and Bi 2 O 3 , which improves the stability of the Cu + valence state and maintains a higher degree of dispersion of the active cuprous species during the transformation from copper oxide to cuprous acetylide, resulting in enhanced catalytic activity. The spinel CuBi 2 O 4 phase formed at high temperatures (≥600 • C) gradually releases copper species to form active cuprous species over multiple cycles, which maintains catalytic stability. Furthermore, we found a linear relationship between catalytic activity and the Cu + area, which indicates that Cu + is the actual active species that simultaneously activates C 2 H 2 and HCHO. Hence, the present observations and the literature allow us to propose a plausible ethynylation mechanism for CuO-based catalysis. Our work highlights a simple method for the rational design of highly active and stable CuO-based catalysts that simultaneously solves the problem of over-reduction and the aggregation of copper species and maintains sustainable catalysis during the phase transformation of copper oxides, thereby demonstrating important academic and industrial value.