3.1 Catalyst synthesis and characterization
Phosphorus is easily doped into the carbon framework during the calcination process after impregnation. The XPS and EDS results listed in the Table S1 and Fig. 1a show that there is a certain amount of phosphorus in this batch of catalysts, indicating that phosphorus is doped in the carbon framework.
In addition, nitrogen adsorption and desorption isotherms are used to measure P-doped activated carbons. The Table S2 lists the specific surface area and pore structure parameters of P-doped activated carbons calcined at different temperatures. The BET surface areas of Cu/PC200, Cu/PC400, Cu/PC600, and Cu/PC800 are 203, 420, 910, 1005 m2g− 1, respectively. Compared with the specific surface area of activated carbon of 1204 m2g− 1, both the specific surface area and pore volume of activated carbon after phosphorus doping treatment are smaller. The addition of phosphorus element may fill and block part of the pore of the carrier and occupy some available space. The larger specific surface area and pore volume of the catalysts calcined at 600℃ and 800℃ may be caused by the thermal decomposition of phosphorus ligand at high temperature and the reduction of blocked pores. The pore size of phosphorus-doped activated carbon is similar, and it is also relatively close to activated carbon. Previous reports generally agree that a higher specific surface area can expose more active sites to promote the transfer of the substrate, which is conducive to improving the activity. The results of this study are also the same. The carbon carrier calcined at 800°C has the largest specific surface area, and the corresponding acetylene conversion rate is also the highest among several catalysts. As shown in the Fig. 1b, the N2 adsorption-desorption isotherm of PC800 with the best effect is a typical IV type curve with obvious hysteresis loop characteristics of mesoporous structure, indicating that the PC800 carrier has a multi-stage pore structure and other catalysts also have hysteresis loop characteristics. The Fig. S1a,b,c shows that several other catalysts also have hysteresis loop characteristics.
Figure 1c shows the XRD patterns of each fresh Cu-based catalyst. The amorphous diffraction peaks of the carbon carrier at 25° and 43° correspond to the plane of (002) and (101), respectively37. In addition, no other discernible diffraction peaks are detected in the Cu-based catalyst, which means that the copper particle size is below the detection limit of the XRD instrument, or the copper species on the activated carbon support is in an amorphous form38–39. It can be speculated that the copper species can be well dispersed on the P-doped carbon support, and the presence of phosphorus can enhance the dispersibility of copper active sites.
The Fig. S2 shows the morphology of each phosphorus-doped catalyst, showing a classic two-dimensional (2D) carbon nanosheet structure, with obvious wrinkles which may be caused by phosphorus doping. The size of phosphorus atoms is larger than carbon atoms, resulting in local geometric distortion in the carbon skeleton. Meanwhile, there are almost no copper nanoparticles in these fresh catalysts, indicating that the introduction of phosphorus atoms can make the copper species well dispersed on the support. The presence of phosphorus atoms on the carrier may provide new active sites, and the coordination structure formed with copper species may be the main active sites. These results are consistent with the previous XRD spectra analysis.
In addition, further analysis of HAADF-STEM image (Fig. 1d and Fig. S3a,c,e) reveals the existence of highly dispersed isolated copper species, and no copper nanoparticles are found. It’s confirmed that the single center copper species supported on carbon support is the active center of acetylene hydrochlorination reaction, indicating that the active component of the catalyst is composed of atom dispersed copper40–42. The element mapping of the catalyst Cu/PC800 (Fig. 1e) reveals that C, P and Cu elements are uniformly distributed on the surface of the catalyst, which is also consistent with the previous XPS results, verifying the successful doping of phosphorus in the carbon support, as well as the other catalysts (Fig. S3b,d,f).
Figure2.(a) The conversion of acetylene over P-doped Cu-based catalysts. Reaction conditions: temperature=150℃, GHSV(C2H2)=90 h−1, V(HCl)/V(C2H2)=1.2/1, (b)Comparison of acetylene conversions for Cu/PC200, Cu/PC400, Cu/PC600 and Cu/PC800 catalysts and their respective treated carbons,(c) Kinetic studies of Cu/AC and Cu/PC800 catalyst: apparent activation energy, kJ mol−1, (d) GHSV plotted against the TOF for some copper-based catalysts reported in literature and Cu/PC800 catalyst with better catalytic performance in this article.
3.2 Catalytic performance of Cu-based catalysts
The catalyst shown in the Fig. 2a has the same copper load and phosphorus doping amount in the preparation process. Under the test conditions of T = 150℃, GHSV(C2H2) = 90 h− 1 and V(HCl): V(C2H2) = 1.2, the initial conversion of acetylene is significantly different due to the different calcination temperatures of phosphorous doped carbon carriers. The initial conversion of acetylene increases with the increase of calcination temperature. These catalysts don’t deactivate within 10 h, and Cu/PC800 shows a better catalytic performance with the highest conversion reaching 83.1%. As we can see, the Fig. S4 clearly shows that the VCM selectivity of all catalysts has reached more than 99%. Obviously, all the P-doped Cu-based catalysts in the Fig. 2b show a higher initial conversion than pure Cu/AC (the initial conversion is 35.46%). However, the activity of several phosphorus-doped carbon supports without the active component copper is very low, indicating that the enhanced activity of the copper-based catalyst is due to the interaction and synergistic effect between the active copper species and the phosphorus-doped activated carbon, rather than simply the sum of the parts.
In addition, through experiments at different temperatures, the Arrhenius equation is used to plot and the experimental activation energy (Ea) is obtained through linear fitting (Fig. 2c). At this time, the internal and external diffusion of the reaction have been eliminated, and the reaction is under kinetic control. The apparent activation energy of the Cu/PC800 catalyst with the best catalytic performance in the figure is calculated to be 23 kJ mol− 1. The activation energy of the classic Cu-based catalyst Cu/AC is 34 kJ mol− 1, and the calculated result of the P-doped Cu-based catalyst is lower than this value, indicating that acetylene is more likely to react with hydrogen chloride on the P-doped Cu-based catalyst than the classic copper catalyst. In order to demonstrate the excellent performance of P-doped Cu-based catalysts, the Cu/PC800 catalyst is compared with some copper-based catalysts published in literature (note that the experimental conditions are not necessarily the same). The turnover frequency (TOF) at the beginning of the experiment is plotted versus GHSV, and is shown in Fig. 2d. Various copper catalysts from the literature3,7,43−51 is used for this comparison, and it is obvious that the Cu/PC800 catalyst is one of the better catalysts that can provide higher yields of vinyl chloride at relatively high space velocities (colored areas in the figure) compared to some of the catalysts reported.
3.3 Identification of the catalytic active sites
The Fig. 3a shows the XRD pattern of the used sample, which is similar to the fresh catalyst. The two main diffraction peaks at approximately 25° and 43° correspond to the (002) and (101) crystal planes of carbon, respectively52–56. Except for the two diffraction peaks, no other characteristic peaks were found. The active copper species on the phosphorus-doped carbon support was always dispersed very well during the reaction.
As shown in Fig. 3b, Raman spectroscopy shows that all samples have two characteristic peaks near 1350 and 1590 cm− 1, which are attributed to the absorption peaks of the D band and G band in the carbon material respectively. The G band is generated by the vibration of sp2 hybridized graphite-type carbon atoms, indicating the degree of graphitization of carbon materials, while the D band is usually caused by sp3 hybridized carbon atoms and structural defects, indicating the disordered structure and defect. The doping of phosphorus atoms destroys the hexagonal symmetry of the graphene plane57, which increases the number of defect sites in the activated carbon framework. The higher the temperature, the lower the regularity and order of the sample.
XPS spectroscopy can be used to analyze the chemical state of copper on the catalyst surface, and the relative content of different copper species can be calculated according to the relative deconvolution peak area. The Fig. 3d shows that in each P-doped Cu-based catalyst, the main peak with a binding energy of about 934.5 eV belongs to Cu2+, and the peak with a binding energy of about 932.3 eV belongs to Cu+ and Cu0 39,49. The XAES spectrum can be used to distinguish Cu+ from Cu0. In the XAES spectrum shown in the Fig. 3e, the peak of Cu+ can be observed at 916.6eV, and the peak of Cu0 can be observed at about 918.6eV7,30. The Table S3 and Fig. S5a lists the binding energy positions and the relative content of different copper species Cu2+, Cu+ and Cu0. It can be found that Cu2+ and Cu+ are both active components for P-doped Cu-based catalysts, but with the increase of calcination temperature, the proportion of Cu2+ increases significantly, and the amount of Cu+ and Cu0 decreases, combined with the result of the acetylene conversion (Fig. 4a-c), the catalytic activity is the best when calcination temperature is 800 ℃, so the presence of Cu2+ is more conducive to the improvement of catalyst activity.
The P 2p spectrum shown in the Fig. 3f can be deconvolved into three peaks to determine the relative content and species of phosphorus. The peaks with binding energies around 135.1eV, 134.2eV and 133.0eV correspond to three different phosphorus species, P = O, P-O and P-C respectively53–56, which is also confirmed by FT-IR (Fig. 3c). The P-O bond represents all phosphorus-containing functional groups related to oxygen. The P-C bond indicates that the phosphorus atom is indeed successfully incorporated into the carbon lattice. The binding energy and the relative amount of each phosphorus species in the p-doped Cu-based catalysts are listed in the Table S4. Although the P-O bond occupies the highest proportion in each catalyst, only the relative contents of the P-C bond increase significantly with the increase of roasting temperature, while the relative contents of the P = O bond and the P-O bond decrease. Combined with the catalytic activity of Cu-based catalysts (Fig. 4d-f), the P-C bond has the largest proportion (41.6%) in the catalysts calcined at 800℃, indicating that the higher the content of P-C bond, the more phosphorus atoms entering the carbon skeleton, the more favorable the catalyst to obtain higher catalytic activity.
In order to further analyze the chemical bond configuration of phosphorus in the catalyst, the O 1s peak is deconvolved into two components. As shown in the Fig. S5b, the peak at about 531.4 eV of all catalysts is attributed to non-bridging oxygen (P = O) 50,58, and the peak at about 532.6eV of binding energy is attributed to P-O-Cu bond and accounts for a large proportion54, which is consistent with the result of P 2p spectrum. It further indicates that P-O-Cu bond is included in the oxygen related phosphorus-containing functional group represented by P-O bond, and the coordination structure of Cu and O in P-O bond is relatively stable. In addition, combined with the results of the Table S5 and the acetylene conversion (Fig. S6), as the calcination temperature increases, the relative content of P = O bond decreases slightly, and the relative content of P-O-Cu bond increases slightly, but the change range of both is very small, indicating that the variation of calcination temperature has little effect on the proportion of P = O bond and P-O-Cu bond and they have no significant positive effect on the activity of the catalyst, which can also be verified by the P 2p spectra. It should be noted that although the oxygen related phosphorus-containing functional groups represented by the P-O bond will decrease with the increase of the calcination temperature, the change of P-O-Cu bond contained in it is negligible. In general, the distribution of phosphorus species on the surface of the support can be adjusted by changing the roasting temperature. For the P-doped Cu-based catalyst in this study, the increase of the calcination temperature only leads to a significant increase in the relative content of P-C bond, which has a positive impact on the catalytic activity. Combined with the results of P 2p spectra, it can be inferred that Cu2+ and P-C bond can play a positive role in the hydrochlorination of acetylene, and the coordination structure formed by their interaction is the main active site of the Cu-based catalyst.
It has been reported that experimentally, many phosphorus doping methods, including high-temperature firing in an inert environment, are often accompanied by oxygen doping, so that different types of phosphorus and oxygen-containing functional groups can be formed59–60. Referring to the XPS results, P-O bonds, P = O bonds and P-C bonds may constitute various phosphorus and oxygen-containing groups and the bonding configuration of phosphorus entering the carbon matrix (Fig. 5c).
The TPR curves of each catalyst are shown in the Fig. 5a. Two main reduction peaks can be detected in all catalysts. The first hydrogen consumption peak appears in the range of 230°C to 330°C, and the second peak appears in the temperature range of 470°C to 570°C, these two reduction peaks are attributed to the reduction of Cu2+ species to Cu+ species and the change of Cu+ species to metallic copper, respectively 61. The reduction peaks of Cu2+ and Cu+ move to higher temperatures with the increase of the calcination temperature of the phosphorus-doped carbon support. The temperatures of the two reduction peaks of Cu/PC800 have increased to different degrees compared with other catalysts. It shows that compared with other catalysts calcined at other temperatures, there is a stronger interaction between copper and phosphorus-doped carbon support in Cu/PC800 catalyst, which effectively improves the anti-reduction ability of Cu2+ and Cu+ species. In addition, the reduction peak area of Cu2+ in the Cu/PC800 catalyst is significantly larger than other catalysts, indicating that the coordination structure formed by Cu and P atoms stabilizes the high-valent copper, and to a certain extent delays the reduction of the oxidation state Cu2+ species during the preparation process. This is the same as the result of XPS spectra. The relatively excellent catalytic performance of Cu/PC800 catalyst comes from the interaction between the Cu2+ species with high valence and the P-C bond.
TPD characterization can analyze the adsorption of the catalyst to the reactants, but the phosphorus ligand will be thermally decomposed at a higher temperature. In order to avoid interference to the acetylene signal from the sample desorption, TPD-MS experiment is used to study the adsorption of acetylene by the active site of the catalyst. The results are shown in Fig. 5b. Each catalyst has two desorption peaks, indicating that there are two active sites capable of adsorbing substrates, which may be Cu2+ active centers and Cu+ active centers, respectively. It has been reported that in the Cu(Ⅱ)/AC catalyst, the adsorption energy of acetylene on the copper center is less than that of the Cu(Ⅰ)/AC catalyst, and the smaller adsorption energy is usually desorbed first49. Therefore, the desorption peak at lower temperature and the desorption peak at higher temperature are likely to be attributed to the active center of Cu2+ and Cu+ respectively. The Cu2+ active species is the main active center. By comparing the different catalysts, it can be found that with the increase of calcination temperature, the peak area and desorption temperature of the desorption peak related to Cu2+ increase, especially for the catalyst with the calcination temperature of 800℃. The peak area of desorption peak is widely considered to represent the adsorption capacity. The larger the area is, the more active sites exist and the more acetylene is adsorbed on the corresponding active sites. The desorption temperature represents the strength of adsorption, and a high temperature indicates a stronger adsorption capacity for acetylene. Therefore, more Cu2+ active sites are conducive to the improvement of the catalytic performance of our series of phosphorus-doped copper-based catalysts, and the strong adsorption of related active sites on acetylene is beneficial to improve the activity and stability of the catalyst. Some results of TPD-MS are the same as those of XPS and TPR above.
To understand the activity of Cu2+ active sites loaded on phosphorus-doped carbon support, considering five phosphorus species (C3P, P = O, (OH)2P=O, P(OH)2, (OH)P = O) mentioned in Fig. 4d-f, we carry out a series of theoretical studies aimed at exploring the underlying reaction mechanism. When encountering computational modelling for the active center, we have taken our computing power and time efficiency into full consideration, so we adopt a simple monolayer carbonic cluster saturated by hydrogen atoms (C13H9) in Fig. S7a to simulate carbon support12. The overall calculated results are shown in Fig. 6 and Fig. S7.
The adsorption energy of C2H2 on five different phosphorus species active sites are − 31.77, -42.27, -85.07, -21.26, -16.54kJ/mol respectively (Fig. S7b-f). It is obvious that P(OH)2 and (OH)P = O copper active sites show weak interaction with C2H2. The P = O active site shows better adsorption to C2H2, but O of P = O bond may interact strongly with the Cl of HCl, which may cause complex reaction mechanism. Here we do not discuss the P = O species. The (OH)2P=O would decompose at our reaction temperature. Then only the C3P specie was left.
The optimized structure of Cu2+ species adsorbed on the C3P carbon support exist strong interaction between P and Cl, which may be real reason for the activity of our P-doped Cu-based catalyst. After C2H2 and HCl adsorbed on the active site, the C ≡ C bond length and H-Cl bond length is 1.2254Å and 1.3354Å respectively (Fig. 6a, d), which is 1.2012Å and 1.2895Å in the gas phase in reference. As is shown in the Fig. 6b-c, the molecular orbital of C2H2-complex displays interaction between C2H2 and Cu2+ active sites, and there is an orbital overlap between the Cu atom d orbital and the acetylene π* orbital, while the molecular orbital of HCl-complex is different between HCl and active sites. However, the adsorption energy of C2H2 and HCl on Cu2+ species are calculated at -31.77 and − 34.13 kJ/mol. With cooperation of the nearly equal adsorption energy, it renders that the Cu2+ species after loaded to the P-doped carbon support can active both C2H2 and HCl of the same class. Subsequently, the adsorption of C2H2 and HCl at the active site may cause acetylene hydrochlorination reaction following the L-H mechanism shown in Fig. 6e. C2H2 and HCl co-adsorbed species (C2H2* and HCl*) reacted with each other to produce vinyl chloride.