Synergistic effect of two action sites on a nitrogen-doped carbon catalyst towards acetylene hydrochlorination

Bolin Wang ab, Yuxue Yue a, Xiangxue Pang a, Wenrui Zhu a, Zhi Chen a, Shujuan Shao a, Ting Wang a, Zhiyan Pan b, Xiaonian Li *a and Jia Zhao *a
aIndustrial Catalysis Institute of Zhejiang University of Technology, State Key Laboratory Breeding Bas e of Green Chemistry-Synthesis Technology, Hangzhou, 310014, P. R. China. E-mail: xnli@zjut.edu.cn; jiazhao@zjut.edu.cn
bCollege of Environment, Zhejiang University of Technology, Hangzhou 310014, P. R. China

Received 30th July 2020 , Accepted 7th September 2020

First published on 8th September 2020


Whether the reaction pathway is steady or dynamic over the whole life cycle of a catalyst process can facilitate our understanding of its catalytic behavior. Herein, the dynamic reaction pathways of nitrogen-doped carbon catalysts are investigated in acetylene hydrochlorination. When triggered, the reaction follows the LangmuirHinshelwood mechanism with pyrrolic N and pyridinic N as dual active sites. However, pyridinic N is deactivated first, due to the strong adsorption of hydrogen chloride, causing the reaction to further run with pyrrolic N as the single active site and follow the EleyRideal mechanism. This work provides a new promising way to study the catalytic behavior of nitrogen-doped carbon catalysts.


The study of the catalytic reaction mechanism is an eternal research focus in heterogeneous and homogeneous catalysis.1–4 Currently, carbon catalysis, especially hetero-atom doped carbon-based catalysis, is widely utilized for various reactions of industrial importance, such as dehydrogenation, selective oxidation, oxygen reduction, and hydrochlorination.5–8 Nitrogen-doped carbon catalysts are regarded as one of the most promising substitutes for traditional mercury-based catalysts in producing vinyl chloride monomer (VCM) via acetylene hydrochlorination.9 However, their active sites and the corresponding mechanism are still disputable. Pyrrolic N, pyridinic N and other N configurations are considered as the active sites for this reaction (Table S1, Section 4 of the ESI),10–24 while almost all isolated nitrogen sites have varying catalytic activities. For example, Bao et al.10 demonstrated that carbon atoms bonded with pyrrolic N atoms should be the active sites since the catalytic activity increases monotonously with access to pyrrolic N sites. Dai et al.18 reported that pyridinic N was the active site based on their theoretical simulation. Likewise, Pérez-Ramírez et al.24 concluded that both pyrrolic N and pyridinic N can function as active sites. This is the underlying cause for this controversy, and the difficulty to conclude that the deactivation of one active site means the complete deactivation of the whole catalyst.

Herein, the active sites and the dynamic mechanism of nitrogen-doped carbon catalysts over their whole life cycle are investigated with acetylene hydrochlorination as a probe reaction to explore the self-regulation of reaction pathways on a working N-doped carbon catalyst. Ionic liquid (IL) derived nitrogen-doped carbon (NCT) materials were synthesized via carbonization of a binary IL mixture composed of [EMIM][N(CN)2] and [EMIM][NTf2] at distinct temperatures T (Fig. 1a and Section 1 of the ESI, for more details). Different nitrogen configurations were generated during this carbonization process (Fig. 1b and c). Four types of pyridinic N (N6), pyrrolic N (N5), graphitic N (N3) and pyridinic N+O (N0) configurations with distinct contents were determined in NC600 and NC700 (Fig. 1c). Generally, low temperature pyrolysis results in a high proportion of pyrrolic- and pyridinic-N, because the degree of graphitization of the carbon increases with increasing sintering temperature.22–24 When the thermal treatment temperature continues to increase to 900 °C, all the pyrrolic N was decomposed, resulting in three types of pyridinic N, graphitic N and pyridinic N+O configurations in NC900. In NC1000, both pyrrolic N and pyridinic N were decomposed, leaving only two types of graphitic N and pyridinic N+O configurations. This dynamic evolution process of nitrogen configurations with increasing carbonization temperature has been widely reported in N-doped carbon materials.22–24 Sun et al. reported graphene oxide with hydrogen peroxide at 180 °C to generate pyridinic-N content in N-doped graphene.23 Our previous work indicated that the formation of pyridinic N can be inhibited via the doping of sulfur atoms into the carbon matrix. Sulfur species occupied the site where pyridinic N forms.22 However, the synthesis of specific nitrogen configurations remains a serious challenge. The chemical compositions and textural properties of NCT are shown in Tables S2–S4 (Section 4 of the ESI).


image file: d0cp04043f-f1.tif
Fig. 1 (a) Synthesis process and (b) N1s XPS spectra of NCT. (c) Specific N contents estimated from XPS and (d) reaction rates of NCT. Reaction conditions: 180 °C, GHSV(C2H2) = 90 h−1 and V(HCl)/V(C2H2) = 1.2 and P = 1 bar.

The catalytic performance was evaluated in the kinetic region where the internal and external mass transfer limitations were excluded by following the Weisz-Prater and Mears criterion25,26 (Section 2 of the ESI, for more details). The reaction rates were estimated as 4.96, 5.94, 2.39 and 0.13 molC2H2 h−1 kg−1Cat for NC600, NC700, NC900 and NC1000, respectively (Fig. 1d). Comparing the reaction rates of the catalysts developed in this work and other reported metal-free catalysts (Table S5, ESI), the NCT series, especially NC700, exhibited competitive catalytic activity. It should be noted that during the heat treatment from 600 °C to 1000 °C, the N content decreased rapidly while the catalytic activity did not show a consistent trend (Fig. S1, Section 4 of theESI), indicating that specific N species govern the catalytic activity, which was consistent with the literature reports.10–24

The lowest catalytic activity of NC1000 suggested that the two types of graphitic N and pyridinic N+O in NC1000 were not as active as the other two pyrrolic N and pyridinic N configurations in NC600–900 (abbreviations for NC600, NC700 and NC900). Comparing the change trend of N content (Fig. 1c) and reaction rate (Fig. 1d), the catalytic reaction rate showed an increasing trend to the content of pyrrolic N and pyridinic N rather than graphitic N and pyridinic N+O, which further confirmed this speculation. To verify this speculation, we further studied the adsorption behavior of the substrates on the evaluated active sites by theoretical calculations. The adsorption energies of the reactants over the above four N and graphitic C configurations are shown in Fig. S2 (ESI). The adsorption results implied that both reactants (C2H2 + HCl) were more likely to be activated at pyrrolic N and pyridinic N (Fig. S2, ESI). The adsorption of reactants on the active sites affected the catalytic performance.27,28 The low catalytic activity of graphitic N and pyridinic N+O has been reported in previous results.9,10 In view of the ultra-low catalytic activity of NC1000, the active sites graphitic N and pyridinic N+O on the surface of NC1000 are not to be further studied in this work. The catalytic behavior of the graphitic N and pyridinic N+O configurations were preliminarily studied in Section 3 of the ESI. Since the reactants are adsorbed on the N site in our calculations, the catalytic effect of C atoms in the synthesized N-doped carbon materials may be relatively weak (Fig. S2, ESI), although the catalytic activity of C atoms has been reported in some studies.29,30 Therefore, pyrrolic N and pyridinic N configurations were studied in detail in the following studies in terms of their high activity contribution in NC600–900.

The activity difference between NC600/700 (NC600 and NC700) and NC900 suggested that the reaction rate can be facilitated by the co-existence of pyrrolic N and pyridinic N in comparison to the lack of pyrrolic N in NC900, despite the highest surface area of NC900 (Fig. S3 and Table S4, ESI). Pyrrolic- and pyridinic N could not be converted to other types of N during the reaction process (Fig. S4, ESI). The original difference was explored by sequentially pre-treating the catalyst with reactants (Fig. S5, ESI). The VCM signal can be quickly detected from NC600/700, whether in designed Experiments A or B (Fig. S5, ESI), suggesting the potential existence of three distinct pathways (pathway 1–3 in Table S6, ESI) in NC600/700 during the reaction process. For NC900, the VCM signal can be only detected in Experiment B (Fig. S5, ESI), implying that the reaction was triggered via pathway 2. Comparing the pathways between NC600/700 and NC900 (does not have pyrrolic N), pyrrolic N may act as the adsorption site of C2H2 in pathway 1. Therefore, the adsorption site of HCl can be assumed to be pyridinic N in pathway 2, with higher basicity than pyrrolic N,31 leading to more favourable adsorption of acidic HCl at pyridinic N. This hypothesis was tested by the TPD-MS characterization. For NC600/700, C2H2 and HCl showed similar desorption area when the sample was pre-treated separately with individual or mixed reaction gas (Fig. S6, ESI), indicating the non-competitive adsorption of C2H2 and HCl. A drastically reduced desorption of C2H2 and a slight change of HCl in NC900 further confirmed that the missing pyrrolic N might be the main adsorption site for C2H2. This hypothesis can be further verified by density functional theory (DFT) calculations. In Fig. S7 (ESI), pyrrolic N and pyridinic N were shown to be located at the edges and inner cavities in the graphene layer.32 The adsorption energy results indicated that pyrrolic N and pyridinic N were the preferred adsorption sites of C2H2 and HCl, respectively (Fig. S7, ESI).

It should be noted that when the sample was pre-treated with mixed reaction gas, the desorption temperature of C2H2 on NC600/700 was decreased by 19–26 °C (Fig. S6, ESI), in contrast to a decrease of 11–13 °C in the HCl desorption temperature, implying that HCl adsorbed on pyridinic N might have an additional role in activating the adsorbed C2H2 on pyrrolic N. The activation between the two N sites cannot be reflected in pathway 1 and 2, which were dominated by the Eley–Rideal (E–R) mechanism. Therefore, pathway 3 (Table S6, ESI) may be the dominant reaction pathway in NC600/700. Note that this activation can be effectively activated if there is a strong interaction between pyrrolic N and pyridinic N. Therefore, dual-[pyrrolic- + pyridinic N] sites32 (Fig. S8 site a and site b, ESI) are considered. Combined with the calculation data from Fig. S7 (ESI), the strong adsorption of C2H2 and HCl can be weakened by the existence of pyrrolic N and pyridinic N in the dual N sites (Fig. S8, ESI), which further validates the synergistic effect between the two pyrrolic N and pyridinic N sites during the reaction process.

Subsequently, we calculated the reaction pathway over the dual-[pyrrolic- + pyridinic N] sites. The reaction can be triggered easily from pyrrolic N adsorbed C2H2 and pyridinic N adsorbed HCl (Fig. 2a, steps 1 and 2). Then VCM was generated (step 3) and desorbed (step 4) from the N site to complete a catalytic cycle. The reaction orders of reactants (C2H2 and HCl) were calculated to be ∼0.5 for NC600/700 (Fig. 3), suggesting the dependence of the catalytic activity on the surface coverage of both reactants,24 which shows the characteristics of the Langmuir–Hinshelwood (L–H) mechanism. The above TPD, DFT and kinetic results revealed the dominance of pathway 3 in NC600/700, following the classical L–H mechanism (Table S6, ESI). Interestingly, the reaction order of NC900 was ∼1 for both C2H2 and HCl. Since no pyrrolic N was present on this catalyst surface, the reaction rate of NC900 may be limited by the surface reactions: gaseous C2H2 reacted with the adsorbed HCl to produce VCM, showing a characteristic of a reaction order of 1 for both reactants.


image file: d0cp04043f-f2.tif
Fig. 2 Reaction pathways at (a) the dual-[pyrrolic- + pyridinic N] site, and (b) the dual-[pyrrolic- + pyridinic N] site, the pyridinic N was strongly adsorbed by HCl in this dual N site. The blue, red, grey, green, black and white balls represent the N atom in pyrrolic N, the N atom in pyridinic N, the C atom in the carrier, the Cl atom, the C atom in C2H2 and the H atom, respectively.

image file: d0cp04043f-f3.tif
Fig. 3 Reaction orders of C2H2 and HCl over (a) NC600, (b) NC700 and (c) NC900 catalysts. The partial reaction order of both reactants is calculated from the slope (s) of the fitting lines. Reaction conditions: T = 180 °C, FT = 50 cm−3 min−1, Wcat = 0.05 g, and P = 1 bar. The concentrations of C2H2 and HCl were in the range of 10–25% balanced in Ar at a fixed flow of 50 cm−3 min−1. Each point was obtained in a single date point to eliminate the influence of catalyst deactivation.

The long-term evaluation results of NC600/700 and NC900 catalysts showed the deactivation of the catalysts (Fig. 4). The deactivation rate was 0.259 h−1 for NC900 with complete deactivation at 23 h. However, the deactivation curves of NC600/700 exhibited two distinct characteristics: the nearly monotonous shaped curves of stage 1 and stage 2, implying that the deactivation of NC600/700 was probably caused by a single factor at each stage. The surface area of NC600/700 decreased slightly at stage 1 and dramatically at stage 2 (Fig. S9, ESI), indicating that the blockage of the pore channels by carbon-like species was limited at stage 1 but very severe at stage 2. The carbon mass balance results confirmed this conclusion (Table S3, ESI). ∼1.3 wt% (stage 1) and ∼13.1 wt% (stage 2) carbon-like species were generated on the catalyst surface. The difference between the calculated carbon mass balance value and the ideal 100% implied the deposition contents of carbon species on the catalytic surface in this work. According to the adsorption behavior of C2H2 on pyrrolic N, the carbon-like species in stage 2 might be derived from the spontaneous polymerization of C2H2. This hypothesis can be well supported by the dramatically increased C/N ratios in stage 2 over stage 1 (Fig. S10, ESI) in NC600/700.


image file: d0cp04043f-f4.tif
Fig. 4 Long-term evaluations of NC600, NC700 and NC900 catalysts. Reaction conditions: 180 °C, GHSV = 90 h−1, V(HCl)/V(C2H2) = 1.2 and P = 1 bar.

The deactivation rate was ca.∼3-fold higher in stage 1 than in stage 2, implicating the presence of two distinct deactivation mechanisms in the whole catalytic process. We have calculated the adsorption energies of C2H2 and HCl on dual-[pyrrolic- + pyridinic N] sites in Fig. S8 (ESI). The adsorption energy of HCl on pyridinic N was much lower than the adsorption energy of C2H2 on pyrrolic N, suggesting that HCl adsorbed on pyridinic N was more stable than C2H2. This result, in turn, makes it difficult for HCl to desorb from pyridinic N, which can be confirmed by the sharp increase in Cl/N ratio and the slight change in C/N ratio in stage 1 (Fig. S10, ESI). To our knowledge, pyridine hydrochloride is a chemical intermediate for paper, medicine, and plastics, which can be easily synthesized via the reaction between pyridine and HCl.33 Consequently, in the reaction of acidic HCl with basic pyridinic N31 to form a stable pyridine hydrochloride-like salt, the pyridinic N site was occupied and unable to continue to activate HCl, resulting in catalyst deactivation in stage 1. Consistent with NC600/700 in stage 1, NC900 exhibited a similar Cl/N and C/N trend (Fig. S10, ESI). The spent NC600/700 and NC900 showed similar characteristic peaks to those of pure pyridine hydrochloride (Fig. S11, ESI), suggesting the strong adsorption of HCl on pyridinic N. Based on the above results, the deactivation of NC600/700 catalysts could be divided into two stages: deactivation of pyridinic N to form pyridine hydrochloride-like salts (stage 1), and blockage of the pyrrolic N site by carbon-like deposition (stage 2). This deactivation sequence can be supported by the above TPD-MS results. When HCl was adsorbed on pyridinic N, the desorption temperature of C2H2 adsorbed on pyrrolic N was decreased by 19–26 °C (Fig. S6, ESI), indicating the inhibition of carbon-like deposition at pyrrolic N in stage 1.

Since the NC600/700 catalysts feature two distinct deactivation stages, when pyridinic N is deactivated first, the other active site pyrrolic N can still function normally. Specifically, for NC600/700 catalysts, the reaction follows the classical L–H mechanism with dual-[pyrrolic- + pyridinic N] sites as active sites before initiation of deactivation (pathway 3), but when the catalyst begins to deactivate at pyridinic N, the reaction pathway gradually turns to E–R mechanism with pyrrolic N as the single active site (pathway 1). This conclusion can also be validated by the reaction pathway calculated by DFT. As shown in Fig. 2b, HCl was strongly adsorbed at the pyridinic N and cannot be timely desorbed (step 1), leading to the occupation of a pyridinic N active site. The remaining pyrrolic N site can still activate C2H2 and then react with gaseous HCl (step 2) to produce VCM (steps 3 and 4). In contrast, NC900 can only run once due to the absence of pyrrolic on it (Fig. 4, bottom). The above results indicated that the catalytic behavior can be regulated via the dual-N sites. At present, the optimization of the catalytic performance through reasonable regulation of the mixed-active sites on the catalyst surface is attracting widespread attention.34,35

In summary, we have revealed a dynamic process of reaction pathway evolution over the life cycle of N-doped carbon catalysts in acetylene hydrochlorination. The dynamic transition from L–H to E–R mechanism was proposed and proved during the reaction process. Dual-N sites are superior to single N sites in terms of catalytic activity and stability. Further research on dual-N site catalysts will focus on how to inhibit the deactivation of the pyridinic N site. Despite limitations in the absence of in situ characterization, this study has shed light on the potential self-regulation of reaction pathways over a variety of N site coexisting catalysts during a dynamic reaction process. Going beyond acetylene hydrochlorination, the dynamic mechanistic pathways proposed herein can provide new insights for understanding N-doped carbon materials in target reactions and further guide the design of catalysts.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Financial support from the National Natural Science Foundation of China (NSFC: grant No. 21606199 and 21476207) and the Science and Technology Department of Zhejiang Province (LGG20B060004) are gratefully acknowledged.

Notes and references

  1. G. Malta, S. A. Kondrat, S. J. Freakley, C. J. Davies, L. Lu, S. Dawson, A. Thetford, E. K. Gibson, D. J. Morgan, W. Jones, C. J. Kiely and G. J. Hutchings, Science, 2017, 355, 1399–1403 CrossRef CAS.
  2. G. Malta, S. A. Kondrat, S. J. Freakley, C. J. Davies, S. Dawson, X. Liu, L. Lu, K. Dymkowski, F. Fernandez-Alonso, S. Mukhopadhyay, E. K. Gibson, P. P. Wells, S. F. Parker, C. J. Kiely and G. J. Hutchings, ACS Catal., 2018, 8, 8493–8505 CrossRef CAS.
  3. P. Johnston, N. Carthey and G. J. Hutchings, J. Am. Chem. Soc., 2015, 137, 14548–14557 CrossRef CAS.
  4. J. Zhao, B. Wang, Y. Yue, S. Di, Y. Zhai, H. He, G. Sheng, H. Lai, Y. Zhu, L. Guo and X. Li, J. Catal., 2018, 365, 153–162 CrossRef CAS.
  5. L. Qin, L. Wang, C. Wang, X. Yang and B. Lv, Mol. Catal., 2019, 462, 61–68 CrossRef CAS.
  6. V. Y. Evtushok, A. N. Suboch, O. Y. Podyacheva, O. A. Stonkus, V. I. Zaikovskii, Y. A. Chesalov, L. S. Kibis and O. A. Kholdeeva, ACS Catal., 2018, 8, 1297–1307 CrossRef CAS.
  7. L. Han, Y. Sun, S. Li, C. Cheng, C. E. Halbig, P. Feicht, J. L. Hübner, P. Strasser and S. Eigler, ACS Catal., 2019, 9, 1283–1288 CrossRef CAS.
  8. J. Zhao, B. Wang, X. Xu, Y. Yu, S. Di, H. Xu, Y. Zhai, H. He, L. Guo, Z. Pan and X. Li, J. Catal., 2017, 350, 149–158 CrossRef CAS.
  9. J. Zhao, B. Wang, Y. Yue, G. Sheng, H. Lai, S. Wang, L. Yu, Q. Zhang, F. Feng, Z. Hu and X. Li, J. Catal., 2019, 373, 240–249 CrossRef CAS.
  10. X. Y. Li, X. L. Pan, L. Yu, P. J. Ren, X. Wu, L. T. Sun, F. Jiao and X. H. Bao, Nat. Commun., 2014, 5, 3688–3694 CrossRef CAS.
  11. X. Y. Li, P. Li, X. L. Pan, H. Ma and X. H. Bao, Appl. Catal., B, 2017, 210, 116–120 CrossRef CAS.
  12. C. Zhang, L. Kang, B. M. Zhu and B. Dai, RSC Adv., 2015, 5, 7461–7468 RSC.
  13. S. Chao, F. Zou, F. Wan, X. Dong, Y. Wang, Y. Wang, Q. Guan, G. Wang and W. Li, Sci. Rep., 2017, 7, 39789 CrossRef CAS.
  14. X. Li, J. Zhang, Y. Han, M. Zhu, S. Shang and W. Li, J. Mater. Sci., 2018, 53, 4913–4926 CrossRef CAS.
  15. J. Wang, F. Zhao, C. Zhang, L. Kang and M. Zhu, Appl. Catal., A, 2018, 549, 68–75 CrossRef CAS.
  16. X. Dong, S. Chao, F. Wan, Q. Guan, G. Wang and W. Li, J. Catal., 2018, 359, 161–170 CrossRef CAS.
  17. W. Zhao, M. Zhu and B. Dai, Catalysts, 2016, 6, 193 CrossRef.
  18. X. Li, Y. Wang, L. Kang, M. Zhu and B. Dai, J. Catal., 2014, 311, 288–294 CrossRef CAS.
  19. P. Li, H. B. Li, X. L. Pan, K. Tie, T. T. Cui, M. Z. Ding and X. H. Bao, ACS Catal., 2017, 7, 8572–8577 CrossRef CAS.
  20. X. Li, J. Zhang and W. Li, J. Ind. Eng. Chem., 2016, 44, 146–154 CrossRef CAS.
  21. B. Dai, K. Chen, Y. Wang, L. Kang and M. Zhu, ACS Catal., 2015, 5, 2541–2547 CrossRef CAS.
  22. B. Wang, J. Zhao, Y. Yue, G. Sheng, H. Lai, J. Rui, H. He, Z. Hu, F. Feng, Q. Zhang, L. Guo and X. Li, ChemCatChem, 2019, 11, 1002–1009 CAS.
  23. J. Sun, L. Wang, R. Song and S. Yanga, Nanotechnol., 2016, 27, 055404 CrossRef.
  24. R. Lin, S. K. Kaiser, R. Hauert and J. Pérez-Ramírez, ACS Catal., 2018, 8, 1114–1121 CrossRef CAS.
  25. H. S. Fogler, Elements of Chemical Reaction Engineering, Prentice Hall Professional Technical Reference, 2006 Search PubMed.
  26. R. Perry and D. Green, Perry's Chemical Engineers’ Handbook, McGraw-Hill Education, 8th edn, 2008 Search PubMed.
  27. S. Ali, Y. Qiu, Z. Lian, S. Olanrele, G. Lan, Y. Li, D. Su and B. Li, Appl. Surf. Sci., 2019, 478, 574–580 CrossRef CAS.
  28. X. Qiao, Z. Zhou, X. Liu, C. Zhao, Q. Guan and W. Li, Catal. Sci. Technol., 2019, 9, 3753–3762 RSC.
  29. G. Lan, Y. Qiu, J. Fan, X. Wang, H. Tang, W. Han, H. Liu, S. Song and Y. Li, Chem. Commun., 2019, 55, 1430–1433 RSC.
  30. Y. Qiu, S. Ali, G. Lan, H. Tong, J. Fan, H. Liu, B. Li, W. Han, H. Tang, H. Liu and Y. Li, Carbon, 2019, 146, 406–412 CrossRef CAS.
  31. B. Li, X. Sun and D. Su, Phys. Chem. Chem. Phys., 2015, 17, 6691–6694 RSC.
  32. S. Kaiser, R. Lin, S. Mitchell, E. Fako, F. Krumeich, R. Hauert, O. Safonva, V. Kondratenko, E. Kondratenok, S. Collins, P. Midgley, N. Lopez and J. Pérez-Ramírez, Chem. Sci., 2019, 10, 359–369 RSC.
  33. F. Mongin, O. Mongin and F. Trécourt, Tetrahedron Lett., 1996, 37, 6695–6698 CrossRef CAS.
  34. X. Zhang, Z. Yang, Z. Lu and W. Wang, Carbon, 2018, 130, 112–119 CrossRef CAS.
  35. J. Liu, H. Liu, H. Chen, X. Du, B. Zhang, Z. Hong, S. Sun and W. Wang, Adv. Sci., 2020, 7, 1901614 CrossRef CAS.

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/d0cp04043f

This journal is © the Owner Societies 2020