A nickel-phyllosilicate core–echinus catalyst via a green and base additive free hydrothermal approach for hydrogenation reactions

Bing Ma , Huimei Cui and Chen Zhao *
Shanghai Key Laboratory of Green Chemistry and Chemical Processes, School of Chemistry and Molecular Engineering, East China Normal University, Shanghai 200062, China. E-mail: czhao@chem.ecnu.edu.cn

Received 4th August 2017 , Accepted 30th August 2017

First published on 31st August 2017


We report a new hydrothermal and basic-additive free process for synthesizing a core(single-crystalline HBEA zeolite)–echinus(nickel phyllosilicate) catalyst, which exhibits excellent reactivity and stability for hydrogenation reactions. Desilication and dealumination processes generate substantial SiO32− ions and exposed Si–OH groups to form nickel phyllosilicate on the external and internal surfaces of zeolite.


Hydrogenation, as a class of important catalytic reactions, has emerged as a versatile organic tool for the synthesis of fuels and chemicals.1 Nickel-based catalysts have been frequently used in the conversion of biomass2,3 and methane dry reforming reactions,4–6 due to their high catalytic activities and insensitivity to catalyst poisons.

However, the nickel nanoparticles have a tendency to undergo fast agglomeration through migration and coalescence of particles under working conditions at high temperatures and pressures, which is an impediment to stable performance.7 Strategies to incorporate metals into different supports comprise impregnation of metal ions,8 precipitation of metal salts in the presence of basic agents,9 and exchange of ions with acidic sites.10 However, these approaches are not generally applicable, because of the weak metal–support interactions,8 the complex and time-consuming procedures involved,9 or the limited exchange capability with acidic sites.10 Recently, hydrothermally synthesized catalysts with well-defined morphologies have attracted great attention,11,12 due to the strong metal–support interaction and structural confinement of metal nanoparticles in shells or pores. The addition of basic precipitation agents is usually used to partially dissolve the carriers for the hydrothermal synthesis. However, this procedure suffers from the inconvenience of needing large quantities of water to remove the basic reagents, leading to a more tedious and time-consuming procedure, in addition to forming greater amounts of waste water.

Hierarchical porous zeolites are promising candidates for the encapsulation of nanoparticles for confinement catalysis.9 In this contribution, we report an environmentally friendly hydrothermal approach without the addition of any basic additives for synthesizing a core(single-crystalline HBEA zeolite)–echinus(nickel phyllosilicate) precursor, using nickel acetate (a weak acid weak base salt) to desilicate and dealuminate the HBEA carrier. We chose the hydrogenation of stearic acid and ethylbenzene as model reactions to evaluate the effect of confinement of mesopores and shells as well as the metal–support interaction (MSI) on the catalytic performance in terms of activity and stability.

The choice of the metal salt precursor is key for achieving a successful hydrothermal synthesis when a basic precipitation agent is not used. Thus three nickel salts (nickel acetate, nickel chloride, and nickel nitrate) were screened as the metal precursor sources for the synthesis, and a hierarchical single-crystalline HBEA zeolite (Si/Al molar ratio: 12.7) was adopted as the support. Prior to hydrothermal treatment, the pH values of the mixture were 7.2 (Ni(OAc)2), 6.2 (Ni(NO3)2) and 5.8 (NiCl2) (Fig. 1a). After 5 hours of hydrothermal treatment, the highest nickel loading attained was at 39 wt% with nickel acetate as the metal precursor, while loadings of only 1.8 wt% and 1.7 wt% were attained with nickel chloride and nickel nitrate, respectively (Table S1, ESI). The rate for the loading of Ni onto the support was highest with Ni(OAc)2 (0.125 gNi gzeolite−1 h−1), and was much higher than those with NiCl2 and Ni(NO3)2 within 3 h (0.001 and 0.0013 gNi gzeolite−1 h−1) (Fig. S1a, ESI).


image file: c7cc06116a-f1.tif
Fig. 1 (a) The nickel loading and the relative crystallinity of the Ni catalysts prepared with different nickel salts after 5 h, (b) 27Al NMR spectra, (c) 1H NMR spectra, and (d) FTIR spectra of HBEA zeolites after reactions with three nickel salts after hydrothermal treatment.

In order to understand the detailed mechanism taking place between HBEA and the three Ni salts, samples were selected after hydrothermal synthesis after 2 and 5 h for detailed characterization. The 27Al MAS NMR (Fig. 1b) showed that from 2 to 5 h, Al contents in the frameworks (FAl) with Ni(OAc)2 decreased from 8.2% to 5.8%, while those in FAl–Ni(NO3)2 and FAl–NiCl2 increased from 9.5% to 12.9% and from 9.5% to 11.2%, respectively, under identical conditions (Table S1, ESI). Accordingly, as shown from the 1H NMR spectra (Fig. 1c), the content of SiOHAl groups (as determined by integrating the bridging hydroxyl proton at 4.58 ppm) decreased significantly from 71.6% to 23.1% in the HBEA–Ni(OAc)2 sample from 2 to 5 h (Table S1, ESI); in HBEA–Ni(NO3)2 or HBEA–NiCl2 the same signal in the 1H spectrum either increased or was unchanged (Table S1, ESI). This demonstrated that within the initial 2 h of hydrothermal treatment, desilication and dealumination both occurred on HBEA with the three Ni salts, which caused decreased contents in FAl and SiOHAl; but after prolonged periods of 2–5 hours, Al may be re-incorporated in the framework of HBEA in Ni(NO3)2 and NiCl2 solutions via recrystallization, with a concomitant increase in the FAl and SiOHAl content.

Accordingly, there was a marked decrease in the relative crystallinity of HBEA–Ni(OAc)2 from 100% to 55.5% (Fig. S1b, ESI), while, in contrast, the crystallinity of HBEA with the other two Ni salts underwent an initial decrease followed by an increase (from 83.9% to 90.9% with Ni(NO3)2, and from 93.2% to 93.4% with NiCl2, as shown in Table S1, ESI). These results indicate a more significant desilication process with HBEA–Ni(OAc)2 under the selected hydrothermal conditions, leading to a decrease in the crystallinity of the as-formed sample. An increase in the crystallinity with Ni(NO3)2 or NiCl2 may result from the recrystallization process where Al is re-incorporated into the framework of HBEA.

The ICP-AES measurement showed that no Al3+ was detected with HBEA–Ni(OAc)2 during treatment after 2–5 h; however, substantial amounts of Al3+ were detected in HBEA–Ni(NO3)2 and HBEA–NiCl2 (1.632–2.601 ppm, Table S1, ESI). This suggests that Si–O–Al undergoes partial cleavage and that the Al species are still anchored onto the framework of HBEA when treated with Ni(OAc)2, while Al is removed in the form of Al(OH)3 with the two other Ni salts. With respect to the Ni2+ concentrations in the solution, in Ni(OAc)2 they gradually decreased up to 5 h as a result of Ni phyllosilicate formation (Table S1, ESI). However, the concentrations of Ni2+ and Si4+ in the Ni(NO3)2 or NiCl2 solutions increased from 2 to 5 h (Table S1, ESI), suggesting that the formed Ni phyllosilicate may be dissolved under the strongly acidic conditions. The detected framework Si/Al ratios (as analyzed by 29Si-NMR measurements in Fig. S2, ESI) are in agreement with these results, showing that they were slightly higher than the bulky Si/Al ratios (as analyzed by ICP-AES), when treating with Ni(NO3)2 or NiCl2 solution (Table S1, ESI). This probably results from the recrystallization of Si or Al into the framework of HBEA.

The DRIFT spectra showed that the Ni(OAc)2 treated sample showed a remarkable decrease in the 3745 cm−1 peak (Fig. 1d), probably due to the efficient hydrolysis of the terminal Si–OH groups with this salt (Fig. 2a and b).13 The peak at 3670 cm−1 is assigned to the hydroxyl groups bound to aluminum atoms (Al–OH) connected to the zeolite framework through only one or two bonds (red circle in Fig. 2a).14,15 The presence of strong acids (HCl or HNO3) can completely hydrolyze these Al–O bonds to Al(OH)3 (Fig. 2c and d), and thus the intensity of the band at 3670 cm−1 accordingly decreases in the HBEA–NiCl2 and HBEA–Ni(NO3)2 samples. In the same manner, the hydrolysis of the aluminum bound hydroxyl groups directly bound to the zeolite framework led to a remarkable decrease of the Si(OH)Al band (positioned at 3610 cm−1)16 (blue circle in Fig. 2a).


image file: c7cc06116a-f2.tif
Fig. 2 (a and b) The mechanism of the hydrothermal process with HBEA in Ni(OAc)2 solution. (a, c and d) The mechanism of the hydrothermal process with HBEA in NiCl2 or Ni(NO3)2 solutions.

The data obtained (Table S2, ESI) thus far allowed us to propose a plausible mechanism for the hydrothermal synthesis with HBEA–Ni(OAc)2 (see Scheme 1). First, hydroxyl ions produced from the hydrolysis of nickel acetate in turn facilitate the cleavage of Si–O–Si bonds, leading to the formation of silicate ions (Si content in the filtrate increases from 116.0 to 180.3 ppm, Table S2, ESI), and combine with nickel ions to produce nickel phyllosilicate with a rate of 0.08 gNi gzeolite−1 h−1 in 0–2 h on the external surface (fast step, shown on the left side of Scheme 1). During the formation of Ni3Si2O5(OH)4, OH ions would be consumed, thus increasing the concentration of protons and leading to a decrease in the pH from 2 h to the end of the process (Table S2, ESI). It has been reported that the tetrahedral Si–O–Al framework in HBEA is negatively charged,17 and thus, the subsequent slow step is the acid catalyzed hydrolysis of Al–O bonds from the Si–O–Al groups (shown on the right side of Scheme 1). The slow hydrolysis rate was confirmed by the lowered increase rates for Ni loading (0.02 gNi gzeolite−1 h−1 in 5–7 h, as shown in Table S2, ESI), and is probably due to the reduced exposure of the silanol framework which is covered with an additional network of aluminum (the blue circles shown on the right side of Scheme 1).


image file: c7cc06116a-s1.tif
Scheme 1 The entire NixSiyOz/HBEA formation process with nickel acetate precursor and HBEA support within the hydrothermal process.

Having ascertained the optimal metal salt precursor, different types of zeolite (HBEA (12.7), HZSM-5 (20.5), HMOR (7.5), and HUSY (20.9)) were tested with the same conditions for the hydrothermal process. The highest nickel loadings were achieved with the Ni/HBEA and Ni/HUSY samples (Table S3, ESI), which reached 16 wt% and 25 wt%, respectively. These loadings were between three and four times higher than those in the Ni/HZSM-5 and Ni/HMOR catalysts (5 wt% and 6 wt%, respectively). The relative crystallinities of the various Ni/zeolite catalysts as determined by the XRD patterns were as follows: HUSY (66.2%) < HBEA (74.2%) < HZSM-5 (78.8%) < HMOR (82.8%) (Fig. S3a, ESI). Hence, it appears that there is an inverse relationship between the crystallinity of the sample and the Ni loadings of the catalysts (Fig. 3a).


image file: c7cc06116a-f3.tif
Fig. 3 (a) The nickel loadings and the relative crystallinity of the Ni catalysts supported on various zeolite supports, (b) Si concentration in filtrates as determined by ICP and the Si–OH defect density for hydrothermally treated Ni/zeolite samples with nickel acetate solution.

The greater Ni loadings that are achieved with HUSY and HBEA zeolites can be explained in terms of the differences in the framework density (FD)18 and silicon hydroxyl density.19 Hence, zeolites that exhibit greater stability to hydrothermal treatment would be expected to have higher FDs. The FDs of the structural unit of HUSY, HBEA, HMOR, and HZSM-5 are 12.7, 15.1, 17.2, and 17.9 T/1000 A3, respectively. Thus, the FD sequence (HUSY < HBEA < HMOR < HZSM-5) follows the trend of decreasing Ni content and increasing crystallinity of the pre-catalysts after hydrothermal synthesis (Fig. 3a).

In addition, the stability of zeolites in high-temperature water mainly depends on the density of the silicon hydroxyl defects, and higher densities generally lead to zeolites with a greater tendency to collapse.19 According to the DRIFT spectra of the four zeolites (HZSM-5, HMOR, HUSY, and HBEA) in Fig. S3b (ESI), a strong peak at about 3738 cm−1 (assigned to silanol groups) was present in the HUSY and HBEA samples, but not on the HMOR and HZSM-5 zeolites. The density of silicon hydroxyl defects (determined by the intensity of the peak at 3738 cm−1 of the IR spectrum in Fig. S3b ESI, unit: mmol g−1) decreased (HUSY (5.7) > HBEA (5.2) > HMOR (1.07) > HZSM-5 (0.042)) with increasing crystallinity, with the concentrations of dissolved silicon in the solution (measured by ICP) perfectly mirroring the concentration of defects attained (Fig. 3b and Table S4, ESI).

A specific advantage of our developed hydrothermal method is that the aqueous filtrate solution remaining after the hydrothermal process can be reused for other synthetic runs, thus avoiding production of excessive waste water during the production of the pre-catalyst. A depiction of the process for the synthesis of the catalyst after consecutive runs is shown in Fig. 4. The compositions of the filtrate and the nickel loadings of the catalyst after three synthetic runs are summarized in Table S5 (ESI). The increase in the Ni content is expected since each step accumulates an increasingly larger amount of nickel when the filtrate from the previous synthetic run is used, which is added to a fresh batch of nickel acetate. Since the pH gradually decreases from 5.20 to 4.52 after four hydrothermal synthetic runs, it is suggested that WAWB salt precursors can efficiently incorporate Ni onto the zeolitic supports after repeated runs within the observed pH range. The catalysts prepared with the recycled filtrate still had high Ni loadings (around 17 wt%) and well-preserved crystalline structures, as verified by ICP and XRD (Fig. S4, ESI).


image file: c7cc06116a-f4.tif
Fig. 4 The process for re-using the filtrate solvent during the hydrothermal synthesis.

The Ni/HBEA–Ni(OAc)2 catalyst synthesized via the hydrothermal method was extensively characterized by XRD, IR, N2 sorption, SEM, TEM and TEM-EDX measurements (Fig. S5 and Table S6, ESI). The nickel particles were uniformly dispersed, and the mean particle size was 5.9 ± 1.2 nm (Fig. S5e, ESI) as determined by the TEM image, while the Ni dispersion as analyzed by CO chemisorption was determined to be 3.5%. The TEM-EDX experiment (Fig. S5f, ESI) determined that there was a large amount of Ni and a little amount of Si dispersed on the external spherical surface, while Al was evenly dispersed on the support, showing a typical core(single-crystalline HBEA zeolite)–echinus(nickel phyllosilicate) structure. Next, two model reactions (hydrodeoxygenation of stearic acid and hydrogenation of ethylbenzene) were used to test the catalytic activity and stability of the hydrothermally (HT)-prepared reduced Ni/HBEA catalyst, compared to those obtained via conventional impregnation (IW) and deposition–precipitation (DP) methods. The detailed characterization of the three Ni/HBEA catalysts is shown in Fig S6–S8 and Table S7 (ESI). Fig. 5a shows the product yield as a function of time for the hydrodeoxygenation of stearic acid for the three reduced Ni/HBEA catalysts at 260 °C. The highest yields of C17 and C18 alkanes were achieved with HT-Ni/HBEA (66%), while with DP-Ni/HBEA and IW-Ni/HBEA only 47% and 27% alkane yields were respectively attained with the rates of 11.2, 7.6, and 5.2 g g−1 h−1. Moreover, the produced high yield of C18 alkanes evidenced a major route of the hydrogenation–dehydration path (Table S8, ESI).


image file: c7cc06116a-f5.tif
Fig. 5 Comparison of kinetics from (a) hydrodeoxygenation of stearic acid to alkanes, (b) hydrogenation of ethylbenzene to ethylcyclohexane, recycling tests of (c) stearic acid hydrodeoxygenation, and (d) ethylbenzene hydrogenation over three Ni/HBEA samples. General conditions: 80 mL n-dodecane, 4 MPa H2, stirring at 600 rpm; (a) 5.0 g stearic acid, 0.2 g Ni/HBEA, 260 °C; (b) 5.0 g ethylbenzene, 0.1 g Ni/HBEA, 200 °C.

In the hydrogenation of ethylbenzene at 200 °C, an ethylcyclohexane yield of nearly 60% was reached in the initial 20 min with HT-Ni/HBEA (Fig. 5b), but the yields were only 38% and 0.7% for DP-Ni/HBEA and IW-Ni/HBEA, respectively (Table S9, ESI), which correspond to the rates of 120, 56, and 1.0 g g−1 h−1, respectively. Hence, it is evident that HT-Ni/HBEA is a highly active catalyst for hydrogenation reactions. The stability tests showed that the catalyst still maintained a high activity after the four runs for hydrodeoxygenation and hydrogenation reactions, and Ni leaching was not observed (Fig. 5c and d). The high activity and stability on the hydrothermally synthesized Ni/HBEA catalyst with highly dispersed Ni nanoparticle (NPs) on the zeolite support are probably due to the good confinement of Ni NPs in mesopores and shells as well as strong metal–support interaction (MSI), while the NPs un-confinement and weaker MSI may lead to lower activities on the other two Ni/HBEA samples.

The hydrothermal synthesis process can be scaled up by at least 25 times, still obtaining a Ni/HBEA precatalyst with excellent properties (Ni loading of 38%, and a crystalline structure as demonstrated by XRD, Fig. S9, ESI). Importantly, this hydrothermal synthesis is versatile for introducing Co and Cu transition metals to zeolitic materials, as shown in Fig. S10 (ESI).

In summary, we have developed a new hydrothermal process that does not require the addition of any basic additives for synthesizing a nickel phyllosilicate precursor. The as-formed reduced core(single-crystalline HBEA zeolite)–echinus(nickel phyllosilicate) catalyst exhibits excellent reactivity for the hydrogenation of stearic acid and ethylbenzene and is highly durable. During the formation of the catalyst, desilication combined with dealumination processes on HBEA by the aqueous nickel acetate solution generates large amounts of SiO32− ions and exposed Si–OH groups, which combine with metal ions to form large amounts of nickel phyllosilicate on the external and internal surfaces of zeolite pores. The developed synthetic method represents a more environmentally friendly and facile approach compared to conventional approaches, and possesses versatility for introducing Ni, Co, and Cu metals confined into silicate/alumina supports with variable loadings. In the process, the waste water is re-used and the procedure is amenable to scale-up; therefore, the developed method has potential for synthesizing efficient supported metal catalysts with structural confinement and strong metal–support interaction used in high-temperature reactions.

This research was supported by the National Key Research and Development Program of China (Grant No. 2016YFB0701100), the Recruitment Program of Global Young Experts in China, and the National Natural Science Foundation of China (Grant No. 21573075). Bing Ma is grateful for the outstanding doctoral dissertation cultivation plan of action (PY2015031).

Conflicts of interest

There are no conflicts to declare.

Notes and references

  1. J. Pritchard, G. A. Filonenko, R. Putten, E. J. M. Hensen and E. A. Pidko, Chem. Soc. Rev., 2015, 44, 3808 RSC.
  2. J. Löfstedt, C. Dahlstrand, A. Orebom, G. Meuzelaar, S. Sawadjoon, M. V. Galkin, P. Agback, M. Wimby, E. Corresa, Y. Mathieu, L. Sauvanaud, S. Eriksson, A. Corma and J. S. M. Samec, ChemSusChem, 2016, 9, 1392 CrossRef PubMed.
  3. J. Kong, M. He, J. A. Lercher and C. Zhao, Chem. Commun., 2015, 51, 17580 RSC.
  4. C. Zhang, W. Zhu, S. Li, G. Wu, X. Ma, X. Wang and J. Gong, Chem. Commun., 2013, 49, 9383 RSC.
  5. M. V. Sivaiah, S. Petit, J. Barrault, C. Batiot-Dupeyrat and S. Valange, Catal. Today, 2010, 157, 397 CrossRef CAS.
  6. B. Chen, Z. Chao, H. He, C. Huang, Y. Liu, W. Yi, X. Wei and J. An, Dalton Trans., 2016, 45, 2720 RSC.
  7. W. Song, C. Zhao and J. A. Lercher, Chem. – Eur. J., 2013, 19, 9833 CrossRef CAS PubMed.
  8. D. Pakhare and J. Spivey, Chem. Soc. Rev., 2014, 43, 7813 RSC.
  9. B. Ma, J. Hu, Y. Wang and C. Zhao, Green Chem., 2015, 17, 4610 RSC.
  10. D. J. Ostgard, L. Kustov, K. R. Poeppelmeier and W. M. H. Sachtler, J. Catal., 1992, 133, 342 CrossRef CAS.
  11. S. Zhang, W. Xu, M. Zeng, J. Li, J. Li, J. Xu and X. Wang, J. Mater. Chem. A, 2013, 1, 11691 CAS.
  12. B. Ma, H. Cui, D. Wang, P. Wu and C. Zhao, Nanoscale, 2017, 9, 5986 RSC.
  13. A. Vimont, F. Thibault-Starzyk and J. C. Lavalley, J. Phys. Chem. B, 2000, 104, 286 CrossRef CAS.
  14. M. Maache, A. Janin and J. C. Lavalley, Zeolites, 1993, 13, 419 CrossRef CAS.
  15. C. Yang and Q. Xu, Zeolites, 1997, 19, 404 CrossRef CAS.
  16. Y. Oumi, R. Mizuno, K. Azuma, S. Nawata, T. Fukushima, T. Uozumi and T. Sano, Microporous Mesoporous Mater., 2001, 49, 103 CrossRef CAS.
  17. R. M. Ravenelle, F. Schüler, A. D’Amico, N. Danilina, J. A. Bokhoven, J. A. Lercher, C. W. Jones and C. Sivers, J. Phys. Chem. C, 2010, 114, 405 Search PubMed.
  18. W. Lutz, H. Toufar and R. Kurzhals, Adsorption, 2005, 11, 405 CrossRef CAS.
  19. L. Zhang, K. Chen, B. Chen, J. L. White and D. E. Resasco, J. Am. Chem. Soc., 2015, 137, 11810 CrossRef CAS PubMed.

Footnote

Electronic supplementary information (ESI) available: Experimental details and data. See DOI: 10.1039/c7cc06116a

This journal is © The Royal Society of Chemistry 2017