Pomegranate-like Core-Shell Ni-NSs@MSNSs as a High Activity, Good Stability, Rapid Magnetic Separation, and Multiple Recyclability Nanocatalyst for DCPD Hydrogenation.

A novel pomegranate-like Ni-NSs@MSNSs nanocatalyst was successfully synthesized via a modified Stöber method, and its application in the hydrogenation of dicyclopentadiene (DCPD) was firstly reported. The Ni-NSs@MSNSs possessed a high specific area (658 m2/g) and mesoporous structure (1.7-3.3 nm). The reaction of hydrogenation of DCPD to endo-tetrahydrodicyclopentadiene (endo-THDCPD) was used to evaluate the catalytic performance of the prepared materials. The distinctive pomegranate-like Ni-NSs@MSNSs core-shell nanocomposite exhibited superior catalytic activity (TOF = 106.0 h-1 and STY = 112.7 g·L-1·h-1) and selectivity (98.9%) than conventional Ni-based catalysts (experimental conditions: Ni/DCPD/cyclohexane = 1/100/1000 (w/w), 150 °C, and 2.5 MPa). Moreover, the Ni-NSs@MSNSs nanocatalyst could be rapidly and conveniently recycled by magnetic separation without appreciable loss. The Ni-NSs@MSNSs also exhibited excellent thermal stability (≥750 °C) and good recycling performance (without an activity and selectivity decrease in four runs). The superior application performance of the Ni-NSs@MSNSs nanocatalyst was mainly owing to its unique pomegranate-like structure and core-shell synergistic confinement effect.

Introduction

In view of the increasingly serious environment and resource problems in recent years, research in the ″green chemistry″ field becomes more important and prosperous than ever. Catalytic hydrogenation reaction is an atom-economic reaction widely used in the synthesis of a variety of compounds, such as alkenes,[1−4] alkynes,[5,6] carboxylic acids,[7] aldehydes,[8,9] ketones,[10] nitro compounds,[2,11] nitriles,[12] aromatic compounds,[13,14] some natural compounds,[15−17] etc. Catalytic hydrogenation of unsaturated compounds to ideal saturated products is a common liquid fuel processing technology to improve quality and expand usage in high-value fields.

Precious metal and non-precious metal are both used as catalysts in heterogeneous hydrogenation reaction. Precious metal-based catalysts have the advantages of high activity, good selectivity, and low activation temperature, but they are expensive, scarce, and easily poisoned and deactivated. Nickel is the most common and widely used non-noble-metal catalyst in the heterogeneous hydrogenation process because of its reasonable catalytic activity, abundant supply, and cheap price.[18−21]

Raney Ni is the most common Ni-based hydrogenation catalyst used in the industry,[22] but it has some disadvantages such as harsh reaction conditions, low catalytic activity and selectivity, difficulty in regeneration, poor mechanical strength, etc. The shortcomings of Raney Ni can be improved by supported Ni-based catalysts on mechanical strength and catalytic activity.[23,24] Unfortunately, the recyclability and stability of these supported Ni-based catalysts are far from ideal. Consequently, the search for novel Ni-based heterogeneous hydrogenation catalysts with high catalytic activity, good stability, and multiple recyclability is eagerly expected.

The core–shell structure is one of the most effective structures to improve the catalytic performance of metal catalysts.[25−27] A meso-/microporous shell can prevent the active core metal from sintering at high temperatures and increase the specific surface area of the catalyst. For instance, Chen et al. reported that a Ni@mSiO2 catalyst with large specific surface area and controllable cavity structure showed good catalytic activity and high selectivity in the hydrogenation of m-dinitrobenzene (m-DNB) to p-phenylenediamine (m-PDA), with a conversion of m-DNB of 100% and selectivity of m-PDA of over 91%.[28] Niu et al. reported that Ni@mSiO2 and Ni-N2H4@mSiO2 core–shell catalysts showed higher activity than the supported Ni/mSiO2 catalyst in the cinnamaldehyde hydrogenation, with over 90% yield of hydrogenated cinnamaldehyde.[29] It is confirmed that the synergistic effect of core metal and protective shell layer plays an important role in increasing catalytic performances. For instance, Zhou et al. reported that FeNi@Ni nanocomposites showed better catalytic performance than the FeNi nanorods and Ni nanoparticles in reduced hydrogenation of p-nitrophenol to p-aminophenol due to the interfacial synergistic effect.[30]

Nano-metal-catalysts possess excellent catalytic properties superior to conventional metal catalysts due to their large specific surface area and high surface activity. Unfortunately, nano-metal-catalysts are easily oxidized and agglomerated in air and are difficult to separate and recycle, seriously reducing their catalytic performance and bringing obstacles to their industrial applications. Therefore, improving the stability and recyclability of nano-metal-catalysts is particularly important. Magnetic separation is an effective way to improve separation characteristics of nano-metal-catalysts. In recent years, magnetic nanocomposites have been widely used in the field of catalysis, such as photocatalytic degradation of organic dyes,[31] catalytic hydrogenation,[32,33] condensation reactions,[34,35] etc., for their convenient separation. A core–shell structure can protect nanometal particles from oxidation and reduce agglomeration of nanoparticles. To our knowledge, great efforts have been made to integrate active cores with silica shell to develop a core–shell catalyst with extremely high catalytic activity, good stability, and excellent recyclability for various applications because of the chemical inertness, controlled porosity, and excellent thermal stability of SiO2.[36−38] The most common method of encapsulating a SiO2 shell around core NPs is the well-known Stöber method.[39−42] Consequently, the development of a new type of magnetic nano-metal-core@SiO2-shell composite catalyst with high catalytic activity, good stability, easy magnetic separation, and recycling is of great significance.

endo-Tetrahydrodicyclopentadiene (endo-THDCPD) is usually used as a high-energy-density solid fuel but also as an important intermediate for the synthesis of high-value-added products, such as adamantane and exo-tetrahydrodicyclopentadiene (exo-THDCPD/JP-10), as illustrated in Scheme a. In general, endo-THDCPD is synthesized by the hydrogenation of dicyclopentadiene (DCPD), which is mainly separated from the C5 byproducts of the naphtha pyrolysis process. Furthermore, the two double bonds in the DCPD molecule have different hydrogenation activities, and the hydrogenation process is a cascade reaction, resulting in the hydrogenation intermediate mainly being 9,10-dihydrodicyclopentadiene (9,10-DHDCPD).[22] However, it must be emphasized that DCPD can be easily decomposed into a series of C5 byproducts at higher temperatures (Scheme b),[43] which is the main reason for the significant decrease in the selectivity of endo-THDCPD during the hydrogenation of DCPD.

Scheme 1

Main Reaction Route (a) and Side Reaction Route (b) during the Hydrogenation of DCPD

In the present work, a Ni nanospheres (NSs) encapsulated in meso- and microporous silica nanospheres (MSNSs) core–shell catalyst (Ni-NSs@MSNSs) was synthesized by means of a modified Stöber method. The structure and physicochemical properties of the Ni-NSs@MSNSs nanocatalyst were systematically investigated. The catalytic performance of this core–shell catalyst on DCPD hydrogenation was also evaluated. The well-designed pomegranate-like Ni-based core–shell catalyst showed high catalytic activity, good stability on DCPD hydrogenation reaction, and convenient magnetic recycle characteristics. The excellent physical, chemical, and catalytic properties of Ni-NSs@MSNSs were mainly attributed to the high dispersion of the Ni cores, unique pomegranate-like core–shell structure effect, and synergistic confinement effect deriving from the strong interaction between the Ni cores and the MSNSs shell.

Results and Discussion

Synthesis of NiO-NSs@CTAB/SiO2 Core–Shell Precursors

CTAB was the key to the formation of the core–shell structure. The CTAB molecule was the structure-directing template for forming the three-dimensional alkoxide-based silica precursors.[44] The CTAB molecule strongly located on individual NiO particle surface to form a hydrophilic/hydrophobic amphiphilic interface, inducing silicon source orientedly distributed in the outside of NiO-NSs. And the double electric layer formed by the CTAB surfactant protected NiO-NSs from accumulation and settlement. On the other hand, the concentration of CTAB was the key to the morphology of the resulting core–shell nanocomposites (Figure S1). At low concentrations (Figure S1a,b), CTAB could not effectively disperse NiO NPs. This was due to the fact that the CTAB molecule located at NiO-NSs was scarce and repulsive electrostatic force was less than the van der Waals force between the particles, resulting in aggregation and further growth.[45] With the further increase of CTAB concentration, the CTAB molecule located at NiO-NSs increased; thus, the particles gradually changed from agglomerated to monodisperse (Figure S1c,d). When the CTAB concentration exceeded the optimal concentration (Figure S1e,f), the excess CTAB molecule aggregated and nucleated in the solution, leading to the formation of a single SiO2 nanosphere. When the pH of the solution was adjusted to 10–11, condensation reaction preferentially occurred at the interface between CTAB and silicate to form a core–shell structure as the concentration of silicate at the interface was higher than that in the liquid phase.[46] As the polymerization of the oligomeric silicate species continued, the CTAB–silicate underwent a structural transition from spherical to cylindrical micelles, leading to the formation of wormhole-like pores perpendicular to the surface of the NiO-NSs.[47] During the growth of these pores, CTAB acted as a ″bridge″ for the coupling of NiO-NSs and silica shell.

Structural Characteristics

There were two obvious weight-loss steps in the TGA figure for NiO-NSs@CTAB/SiO2 precursors (Figure a). The first weight-loss stage (17.4%) at 200–350 °C corresponded to the decomposition of CTAB, and the second weight-loss stage (13.9%) at 400–550 °C was derived from the removal of partial silanol condensation water remaining in the silica gel. No weight loss was observed at about 550 °C, indicating the complete removal of CTAB and partial O–Si–OH groups. This could be further verified by FT-IR spectroscopy (Figure b). For NiO-NSs, the absorption peak at 433 cm–1 was ascribed to the stretching vibration of the Ni–O bond. SiO2-NSs had five absorption peaks at 1082, 462, 807, 1640, and 3443 cm–1.[48] In particular, the stretching vibration of the Si–O bond appearing at 971 cm–1 was mainly caused by the transformation of silica from the original amorphous structure to an ordered hexagonal arrangement.[49,50] Furthermore, for surfactant CTAB, there were three obvious absorption peaks at 2922, 2854, and 1479 cm–1, respectively.[45] After calcination, CTAB was removed from NiO-NSs@MSNSs core–shell precursors, while the SiO2 and NiO structure was maintained in NiO-NSs@MSNSs core–shell precursors.

Figure 1

TGA curves (a) and FT-IR spectra (b) of catalyst precursors.

The reducibility of the oxide species and their interaction with SiO2 support were studied by TPR. Figure a illustrated the TPR profiles of the NiO-NSs, NiO/MSNSs-15, NiO/MSNSs-40, and NiO-NSs@MSNSs. The TPR curve of NiO-NSs had a hydrogen consumption peak centered at 460 °C. For NiO/MSNSs-15, the H2 consumption peak was significantly broadened (300–630 °C). However, the TPR profile of NiO/MSNSs-40 presented only one distinct H2 consumption peak at about 340 °C, much lower than that of NiO-NSs, which was mainly assigned to NiO NPs have very weak interaction with the MSNSs support.[51] NiO-NSs@MSNSs showed broader and bigger H2 consumption peaks between 250 and 550 °C.[37,42,52] The peak at low temperature was attributed to the reduction of loosely attached NiO crystals with a rather low dispersion. A higher reduction temperature demonstrated a stronger metal–support interaction (MSI) and a bigger particle size. As shown in Figure b, the profile of NiO-NSs@MSNSs could be divided into three hydrogen consumption peaks by fitting, indicating that there were many types of NiO in the silica shell. The broad H2 consumption peaks in the range of 200–500 °C were assigned to the reduction of NiO species.[53] Consequently, the broad peaks around 335 and 410 °C were attributed to the reduction of NiO to Ni0 species with weak interaction with the silica shell, while the small H2 consumption peak around 514 °C was possibly due to the strong interaction between the contact part of NiO cores and silica shell,[54] which can be validated by the FT-IR results (Figure S3B). Moreover, the peak area ratio of peak 1/peak 2/peak 3 was 1.77/4.78/1, revealing that the content of NiO with strong interaction with the silica shell was low. This type of NiO species was difficult to be reduced at 450 °C in H2/Ar.

Figure 2

TPR profiles of catalyst precursors (a) and deconvolution of the reduction curve of the NiO-NSs@MSNSs (b).

The N2 adsorption–desorption isotherms and pore size distributions of the oxide precursors and catalysts were shown in Figure . As revealed in Figure a, the NiO-NSs showed typical type-IV isotherms with a hysteresis loop and the capillary condensation occurred at values around P/P0 = 0.5, indicating the presence of a mesopore structure. Obviously, the prepared mesoporous structure of NiO-NSs had a relatively broad pore size distribution and the average pore diameter was 8.6 nm (Figure b, inset), demonstrating the existence of a large fraction of wide slitlike pores derived from inter- and intraparticle voids that were mainly ascribed to the aggregation of nanoparticles.[55,56] The pore structure of Ni-NSs was collapsed because of sintering during the H2 reduction (Figure c,d). In addition, all supported samples and MSNSs support displayed increasing N2 adsorption at P/P0 = 0.9–1.0 with a relatively small type-H1 hysteresis loop, indicating the existence of a mesopore structure.[57] Moreover, all these supported samples and MSNSs support possessed a steep absorption stage when P/P0 < 0.01 and exhibited type-I adsorption isotherms, demonstrating a micropore structure.[58] These assertions above could be further verified via the pore diameter distribution plots in Figure b,d. The core–shell samples not only had the same pore structure characteristics as MSNSs but also manifested the pore structure characteristics of the inner cores. Particularly, it was emphasized that the reduced Ni-NSs@MSNSs exhibited one noticeable condensation hysteresis loop in the relative pressure (P/P0) range of 0.4–1.0 (Figure c). This was mainly due to the high reduction temperature improving the dispersion of Ni NPs, resulting in abundant mesopores between the Ni particles, which was more conducive to the diffusion and transfer of organic species. Besides, the Ni-NSs@MSNSs widened the pore size of MSNSs (about 2.5 nm) (Figure d), which perfectly confirmed the conclusion of FT-IR (Figure b).

Figure 3

N2 adsorption–desorption isotherms (a, c) and corresponding pore size distributions (b, d) of the prepared nanocomposites.

As shown in Table , the specific surface area and pore volume of NiO-NSs were 93 m2·g–1 and 0.20 cm3·g–1, respectively. On the contrary, the specific surface area (3 m2·g–1) and pore volume (0.01 cm3·g–1) of the freshly reduced Ni-NSs were all severely reduced due to the serious agglomeration and sintering of Ni NPs during the reduction process. All core–shell and supported samples showed structural characteristics similar to the MSNSs, such as high specific surface area (650–766 m2·g–1), relatively large pore volume (0.41–0.56 cm3·g–1), and narrow pore diameter (2.5–3.3 nm). Like the other core–shell catalysts reported in previous studies,[29,58,59] the Ni-NSs@MSNSs possessed a large surface area (658 m2·g–1) and uniform pore (2.0–3.9 nm) provided by the SiO2 shell. Although the surface area of Ni-NSs@MSNSs and Ni/MSNSs-x was mimetic, the pore volume (0.53 cm3·g–1) and mean pore size (3.3 nm) of Ni-NSs@MSNSs were much larger than those of Ni/MSNSs-x, which would facilitate the diffusion and transfer of organic species. The coating of the silica shell provided a large surface area for the Ni-NSs@MSNSs catalysts, promoting the dispersion of nickel particles and reducing the sintering of Ni during the reduction process.

Table 1

Symbols and Several Characteristics of Samples (Specific Pore Volume (Vp), Specific Surface Area (SBET), and Mean Pore Diameter (MPD))

entry	samples	SBET (m2/g)a	Vp (cm3/g)b	MPD (nm)c
1	NiO-NSs	93	0.20	8.6
2	Ni-NSs	3	0.01	13.3
3	NiO-NSs@MSNSs	766	0.56	2.9
4	Ni-NSs@MSNSs	658	0.53	3.3
5	NiO/MSNSs-15	757	0.49	2.6
6	Ni/MSNSs-15	717	0.46	2.6
7	NiO/MSNSs-40	650	0.41	2.5
8	Ni/MSNSs-40	657	0.41	2.5
9	MSNSs	1233	0.77	2.5

Determined by N2 adsorption using the Brunauer–Emmett–Teller (BET) equation.

Barrett-Joyner–Halenda (BJH) desorption pore volume.

BJH desorption average pore size.

Figure a illustrated the XRD patterns of the oxide precursors and MSNSs support. For all samples, the peak at 2θ = 22° was attributed to the amorphous silica phase (JCPDS 29-0085),[40] and peaks at 2θ equal to 37.25, 43.28, 62.88, 75.41, and 79.41° revealed the presence of the face-centered cubic nickel oxide phase (JCPDS 47-1049), demonstrating the existence of NiO and SiO2 in core–shell and supported samples. Figure b illustrated the XRD patterns of fresh Ni catalysts. Three diffraction peaks at 44.51, 51.85, and 76.37° corresponding to the metallic Ni (JCPDS 04-0850) were shown in the patterns of Ni-NSs, Ni/MSNSs-40, and Ni-NSs@MSNSs. Specifically, the peaks of Ni/MSNSs-15 were broad and low, almost invisible, which were due to the lower Ni content and strong interaction between Ni NPs and MSNSs support, resulting in the well dispersion of Ni NPs. The Ni-NSs, Ni-NSs@MSNSs, and Ni/MSNSs-40 showed a higher peak intensity with a higher Ni content. Ni-NSs@MSNSs exhibited a lower peak intensity and broader peak width than Ni-NSs and Ni/MSNSs-40, which meant smaller Ni particles. The Ni crystallite size on Ni-NSs, Ni/MSNSs-40, and Ni-NSs@MSNSs was 33.7, 20.6, and 13.0 nm, respectively (calculated by the Scherrer equation, Table S1). It was demonstrated that the coating of silica shell could inhibit the agglomeration and improve the dispersion of Ni NPs.

Figure 4

XRD patterns of the (a) calcined catalyst precursors and (b) freshly reduced catalysts.

The size and morphological structure of synthesized materials were investigated by SEM and TEM. As shown in Figure a, NiO-NSs possessed a uniform spherical structure, the average particle size was around 150 nm, and no significant aggregation was observed. It could be clearly seen that the 150 nm NiO-NSs secondary particles were composed of a large number of ∼10 nm primary particles, which were confirmed by the XRD results. Nevertheless, the Ni-NSs obtained by reduction at 450 °C (Figure b) showed noticeable sintering and particle agglomeration, which were consistent with the results of XRD and N2 adsorption–desorption isotherms. For NiO/MSNSs-15 (Figure c) and Ni/MSNSs-15 samples (Figure d), it could be clearly observed that the very small NiO NPs and Ni NPs were well dispersed on the MSNSs support, respectively. In addition, high-resolution TEM (HRTEM) was conducted to analyze the crystal structure of Ni/MSNSs-15. The Ni-NPs lattice fringe spacing was 0.178 nm (Figure d, inset), similar to the d value of the face-centered cubic Ni (200) plane, which was also in agreement with the XRD results. Big and irregularly shaped NiO NPs were dispersed on the surface of NiO/MSNSs-40 (Figure e). As revealed in Figure f, the Ni particles on Ni/MSNSs-40 were spherical and severely agglomerated, revealing that sintering occurred during the H2/Ar reduction process. Ni-NSs@MSNSs (Figure g) had a uniform core–shell structure, with Ni cores and a ∼25 nm thick SiO2 shell. The HRTEM image of Ni-NSs@MSNSs (Figure g, inset) showed abundant microcapsule-like pore structures in the Ni-NSs core, which were caused by the removing of lattice oxygen and shrinking of the Ni crystal particle size during H2/Ar reduction. The crystal planes of the (111) of face-centered cubic nickel (about 0.212 nm) were clearly observed. Based on Figure h, the dark black near-spherical-shaped Ni particles were uniformly dispersed inside the silica shell (gray color, no obvious lattice fringe) without obvious aggregation, and 1.7–3.3 nm pores in the silica shell were clearly observed. Further information about the spatial distribution of Ni, Si, and O elements in the Ni-NSs@MSNSs was provided by HRTEM-EDX mapping (Figure i,j). The HRTEM-EDX measurements further confirmed the existence of the pomegranate-like Ni cores and silica shell structure. From the viewpoint of the nanocatalysts, the silica layers existing between the Ni particles might play a great role in stabilizing the Ni particles and protecting Ni-NSs from agglomeration. On the other hand, these meso- and microporous channels existing in silica layers provided convenience for reactants to quickly access and product molecules to leave the Ni active sites, which might facilitate the catalytic process.

Figure 5

SEM image of NiO-NSs (particle distribution, inset) (a) and TEM images of Ni-NSs (HRTEM, inset) (b), NiO/MSNSs-15 (c), Ni/MSNSs-15 (HRTEM, inset) (d), NiO/MSNSs-40 (e), and Ni/MSNSs-40 (f), respectively. TEM image (HRTEM, inset) (g), HRTEM image (h), HAADF-STEM image (i), and EDX mapping (j) of the pomegranate-like Ni-NSs@MSNSs core–shell nanocatalyst.

XPS (Figure ) was adopted to characterize the valence state and surface nickel species of Ni-NSs@MSNSs and Ni/MSNSs-x samples. The survey XPS spectrum (Figure S2) of the Ni-NSs@MSNSs revealed the existence of Ni, Si, and O elements.[60] In the high-resolution XPS spectra, the Ni 2p3/2 and Ni 2p1/2 peaks at 856.5 and 874.3 eV for Ni-NSs@MSNSs were observed, corresponding to the Ni0 species. There were Ni2+ species with binding energies (BE) of 862.1 and 880.1 eV, respectively. It could be seen that most of the Ni element existed in the form of nickel oxide and only a portion of nickel oxide was reduced to a zero-valent state. This corresponded to the pattern of TPR that only a part of NiO could be reduced to metallic Ni at 450 °C. Compared with Ni/MSNSs-40, the Ni species in Ni-NSs@MSNSs exhibited a chemical shift toward a higher binding energy, demonstrating the formation of a stronger interaction between the Ni cores and the meso- and microporous silica shell,[61,62] in agreement with the TPR results (Figure ). The ratios of Ni0/Ni2+ in Ni/MSNSs-15, Ni/MSNSs-40, and Ni-NSs@MSNSs catalysts were 1.71, 1.17, and 0.72, respectively. This demonstrated that the reducibility of NiO in core–shell samples decreased. This conclusion conformed with the result drawn from TPR. It was known that hydrogen molecules were adsorbed and activated on the Ni0 sites. Therefore, the amount of Ni0 on nickel-based catalysts was crucial to its hydrogenation activity. Surprisingly, the Ni-NSs@MSNSs nanocatalysts showed significantly better catalytic activity than the supported catalysts in the subsequent DCPD hydrogenation performance test. This was mainly due to the synergistic confinement effect between the Ni cores and the silica shell. The silica shell provided an extremely large surface area for the Ni catalysts, and the abundant pore structure in the silica shell facilitated the reactants to quickly contact the Ni active sites.

Figure 6

XPS spectra of Ni 2p for the freshly reduced catalysts.

Magnetic Properties

Nanocatalysts had some problems such as poor recovery, large filtration loss, and low centrifugation efficiency. The magnetic separation could achieve rapid recovery and increase the recovery rate of nanocatalysts to a large extent. Magnetic properties were of great importance to the separation and recycle of nanocatalysts. The magnetic properties of nanocatalysts were investigated by room temperature hysteresis loops. As shown in Figure , the typical characteristics of superparamagnetic behavior were observed by undetectable coercivity and remanence. It was apparent that the saturation magnetization (Ms) value of Ni-NSs@MSNSs (10.69 emu/g) was much higher than that of Ni/MSNSs-15 (0.45 emu/g). High saturation magnetic intensity was beneficial to realize high-efficiency magnetic separation. Ni-NSs@MSNSs catalysts could be rapidly (within 40 s) and entirely (no obvious weight loss) separated from the reaction solution by a 4500 G external magnet.

Figure 7

Magnetic hysteresis loops of Ni/MSNSs-15 (a) and Ni-NSs@MSNSs (b) in cyclohexane and the photos of Ni-NSs@MSNSs before and after magnetic separation.

Catalytic Performance

Figure a showed the DCPD conversion curve on Raney Ni, Ni-NSs, Ni/MSNSs-x, and Ni-NSs@MSNSs catalysts. With the increasing of reaction time, the conversion of DCPD gradually increased. For Raney Ni, it took 6 h to reach 85.2% conversion of DCPD. In addition, Ni/MSNSs-15 and Ni/MSNSs-40 got 99.9 and 93.3% conversion of DCPD within 3 h. Interestingly, it took only 20 min to reach over 99.5% conversion of DCPD for Ni-NSs@MSNSs and Ni-NSs catalysts. The conversion speed of DCPD on the core–shell structure Ni-NSs@MSNSs was faster than that of bulk Ni catalysts and supported Ni catalysts. Turnover frequency (TOF) reflected the intrinsic activity of the active sites (exposed surface Ni metal) in the catalysts. The TOF values were 106.0, 69.8, and 46.8 h–1 for Ni-NSs@MSNSs, Ni/MSNSs-15, and Ni/MSNSs-40 nanocatalysts. Figure b illustrated the yield of intermediate 9,10-DHDCPD on different catalysts. The content of 9,10-DHDCPD firstly increased and then decreased. Under the experimental conditions, the conversion abilities of intermediate 9,10-DHDCPD of Raney Ni, Ni/MSNSs-40, and Ni-NSs catalysts were insufficient, so the content of 9,10-DHDCPD was maintained at a high level. As far as Ni/MSNSs-15 and Ni-NSs@MSNSs were concerned, an interesting phenomenon was that the contents of intermediate 9,10-DHDCPD took only 20 min to reach the maximum accumulation and eventually decreased to 0.02%. Since the raw material DCPD, the intermediate 9,10-DHDCPD, and the product endo-THDCPD were transformed through a tandem reaction (Scheme a), the maximum accumulation of the intermediate 9,10-DHDCPD and the time spent to reach the maximum accumulation were all related to the relative rate of the first hydrogenation (NB-bond) and the second hydrogenation (CP-bond). Obviously, Ni/MSNSs-15 and Ni-NSs@MSNSs catalysts significantly increased the relative hydrogenation rate in the second step and the first step, making the intermediate 9,10-DHDCPD reach the maximum accumulation faster and then quickly transform into the product endo-THDCPD. Figure c demonstrated the yield of the product endo-THDCPD. As the reaction proceeded, the yield of the product endo-THDCPD gradually increased for all catalysts. The yields of endo-THDCPD on Ni-NSs, Raney Ni, and Ni/MSNSs-40 catalysts were only 33.3, 48.2, and 56.9%, respectively. Ni/MSNSs-15 and Ni-NSs@MSNSs took only 2 h to reach over 98.5% yields of endo-THDCPD. It was obvious that the coating of silica shell not only enhanced the speed of endo-THDCPD production but also increased the yield of endo-THDCPD. The STY values were 112.7, 31.2, and 3.59 g·L–1·h–1 for Ni-NSs@MSNSs, Ni/MSNSs-15, and Ni/MSNSs-40 nanocatalysts. Figure d described the yield of C5 by-products. For Ni-NSs, Raney Ni, and Ni/MSNSs-40 catalysts, the yields of C5 by-products were all high. The yields of C5 by-products on Ni/MSNSs-15 and Ni-NSs@MSNSs were low (0.3 and 0.6%, respectively). In fact, the C5 by-product generation reaction (Scheme b) and the DCPD hydrogenation reaction were competing reactions. Ni-NSs@MSNSs core–shell catalysts could accelerate the conversion of DCPD hydrogenation due to the synergistic effect, thereby relatively suppressing the generation of C5 by-products.

Figure 8

Conversion of DCPD (a), yield of 9,10-DHDCPD (b), yield of endo-THDCPD (c), and yield of the C5 fraction (d) varying with reaction time. Reaction conditions: weight ratio of Ni/DCPD/cyclohexane = 1/100/1000, temperature = 150 °C, hydrogen pressure = 2.5 MPa, stirring rate = 600 rpm, and reaction time = 6.0 h.

The catalytic performance of catalysts previously reported in the hydrogenation of DCPD is summarized in Table . Compared with other catalysts, our pomegranate-like Ni-NSs@MSNSs core–shell catalysts showed the highest catalytic activity and best selectivity.

Table 2

Comparison of the Catalytic Performance of Several Typical DCPD Hydrogenation Catalysts

entry	samples	temperature (°C)	pressure (MPa)	conversion (%)	selectivity (%)	ref.
1	Pd/Al2O3	90	2.0	85	76.5	(63,64)
2	Pd/C	60	1.0	97		(65)
3	Raney Ni	120	1.5	97.9	3.0	(22)
4	SRNA-4	130	1.5	98	98.5	(22)
5	Ni/HY	170	2.0	99.5	96.8	(23)
6	Pd/HY-LA	70	0.1	100	92.5	(66)
7	Ni/γ-Al2O3	110	5.0	98.3	94.7	(24)
8	NiMo0.2/γ-Al2O3	150	3.5	97	99.7	(67)
9	Ni-NSs@MSNSs	150	2.5	100	98.9	this study

In summary, all of the above phenomena point to the following conclusions: (1) The catalytic performance of DCPD hydrogenation on Ni-based catalysts was greatly affected by the dispersion of nickel particles, and catalysts with smaller Ni particles had better catalytic properties. (2) The Ni-NSs@MSNSs core–shell catalysts showed higher activity and better selectivity on DCPD hydrogenation than bulk and supported Ni catalysts. The higher activity and selectivity of the pomegranate-like Ni-NSs@MSNSs core–shell catalysts were primarily attributed to higher Ni dispersity and the synergistic confinement effect provided by the strong interaction between Ni cores and the silica shell.[68,69]

Thermal Stability

Effect of Annealing Temperature on Core–Shell Nanocomposites

The thermal stability of Ni-NSs@MSNSs core–shell catalysts was studied at 350, 550, and 750 °C for 3 h in Ar. As revealed by XRD in Figure a, no obvious structure change and new species were formed after annealing process, and the crystal structure of Ni cores was relatively stable with only a slight size growth at 750 °C (Table ). TEM showed, after the annealing at 350, 550, and 750 °C, that the core–shell morphologies of Ni-NSs@MSNSs remained intact. Moreover, no obvious sintering and grain aggregation of Ni particles were found (Figure b–d). Therefore, the above results confirmed the excellent thermal stability of the Ni-NSs@MSNSs core–shell catalyst.[70]

Figure 9

XRD patterns (a) and TEM images of pomegranate-like Ni-NSs@MSNSs catalyst at different annealing temperatures in Ar: 350 °C (b), 550 °C (c), and 750 °C (d).

Table 3

Particle Sizes of the Ni-NSs@MSNSs Core–Shell Nanocatalyst at Different Annealing and Reduction Treatment Temperatures

		particle size (nm)a
entry	samples	350 °C	450 °C	550 °C	650 °C	750 °C
1	annealing-SPs	14.8		13.8		17.4
2	reduction treatment-SPs	14.3	13.0	12.1	14.7	12.8

Calculated from the Ni (111) plane by the Scherrer equation.

Effect of Reduction Temperature on Catalytic Activity

The effect of reduction temperatures on Ni-NSs@MSNSs catalysts’ hydrogenation performance was evaluated (Figure ). It was observed that the hydrogenation rate of DCPD on Ni-NSs@MSNSs core–shell catalysts was affected by the reduction temperature. The higher the reduction temperature was, the slower the relative hydrogenation rate was. Catalysts reduced at 350 and 450 °C possessed almost the same excellent catalytic performance. When reduced at a high temperature of 550 °C, Ni-NSs@MSNSs catalysts maintained good catalytic activity and excellent selectivity of DCPD hydrogenation. Surprisingly, when the reduction temperature was raised to 750 °C, Ni-NSs@MSNSs still exhibited remarkable activity and selectivity of DCPD hydrogenation, which were significantly higher than those of Ni-NSs, Raney Ni, and Ni/MSNSs-40 catalysts (reduced at 450 °C). Obviously, it proved that the silica shell could effectively protect Ni particles from agglomeration and improve the thermal stability of the Ni-based catalyst. Moreover, it could be noted that high reduction temperatures had no significant effect on the conversion rate of DCPD and the yield of C5 by-products on Ni-NSs@MSNSs core–shell catalysts but significantly affected the second-stage hydrogenation of DCPD and decreased the production of endo-THDCPD.

Figure 10

Catalytic activity test of the pomegranate-like Ni-NSs@MSNSs nanocatalyst obtained after treatment at different reduction temperatures: (a) conversion of DCPD, (b) yield of 9,10-DHDCPD, (c) yield of endo-THDCPD, and (d) yield of the C5 fraction varying with reaction time. Reaction conditions: weight ratio of Ni/DCPD/cyclohexane = 1/100/1000, temperature = 150 °C, hydrogen pressure = 2.5 MPa, stirring rate = 600 rpm, and reaction time = 6.0 h.

It was generally considered that the Ni active components were susceptible to a high reduction temperature, thereby affecting the catalytic behavior.[71] Surprisingly, even when reduced at 750 °C for 3 h, the Ni-NSs@MSNSs possessed fairly good catalytic hydrogenation properties and maintained the original core–shell structure and morphology with no obvious sintering and aggregation of nickel particles, as shown in Table . The above results confirmed the excellent thermal stability of the Ni-NSs@MSNSs core–shell catalyst.

Cyclic Stability

High activity, excellent selectivity, and good stability were the key features of catalysts.[72,73] Cyclic stability was an important feature for industrial catalysts. Therefore, a preliminary cyclic stability test of Ni/MSNSs-15 and Ni-NSs@MSNSs in DCPD hydrogenation was carried out. The catalysts were separated from the reaction system by an external magnet, washed with cyclohexane several times, and reused in the next run. The conversion of DCPD and selectivity of endo-THDCPD in the recycling experiments were shown in Figure . After four cycles, Ni-NSs@MSNSs catalysts maintained high DCPD conversion activity (99.9%) and good endo-THDCPD selectivity (99%) much higher than that of Ni/MSNSs-15 (58.3%). It was generally known that the hydrogenation reaction of DCPD is a two-step process. And kinetic experiments showed that the reaction activation energy of the first-step hydrogenation was much smaller than that of the second-step hydrogenation.[65,74] Therefore, the hydrogenation in the first step was easier to carry out than that in the second step, resulting in the rapid conversion of DCPD. However, the second-step hydrogenation was relatively difficult to perform thermodynamically and was the control step of the entire hydrogenation reaction, which directly affected the yield and selectivity of the product endo-THDCPD.

Figure 11

Cyclic stability test of Ni/MSNSs-15 supported and pomegranate-like Ni-NSs@MSNSs core–shell catalysts: (a) conversion of DCPD and (b) selectivity of endo-THDCPD.

The textural properties and morphology structures of used catalysts were shown in Figure . As shown in Figure a, the specific surface area and pore volume of used Ni/MSNSs-15 significantly decreased (52 m2/g and 0.08 cm3/g, respectively), and the pore size (6.4 nm) increased significantly. TEM of the used Ni/MSNSs-15 showed the size of Ni particles increased and many Ni NPs fell off from MSNSs (Figure b). The structure of the used Ni/MSNSs-15 catalyst changed significantly, resulting in a decrease in its hydrogenation activity. For the first-step hydrogenation, Ni/MSNSs-15 still exhibited considerable activity, so the change in the conversion of DCPD was not obvious. But for the second-step hydrogenation, the activity decreased significantly, so the ability of the intermediate 9,10-DHDCPD converted to endo-THDCPD was insufficient, resulting in a decrease in product selectivity. Moreover, the decrease in endo-THDCPD selectivity for Ni/MSNSs-15 may be related to the shedding and loss of active components caused by intermittent tank stirring. The specific surface area and pore volume of used Ni-NSs@MSNSs remained almost unchanged. The used Ni-NSs@MSNSs still exhibited characteristics similar to the fresh catalyst without structure destruction and agglomeration of Ni particles (Figure c). More importantly, the good dispersity of Ni active components in used Ni-NSs@MSNSs led to the formation of uniform pores, which accelerated the contact of reactants to and the leaving of products from Ni active sites.[42,75] The excellent stability of Ni-NSs@MSNSs was derived from the synergistic confinement effect between Ni cores and silica shell and the distinctive pomegranate-like structure, which effectively prevented the leaching of Ni active sites.

Figure 12

N2 adsorption–desorption isotherms (pore size distribution, inset) (a). TEM images of Ni/MSNSs-15 (b) and Ni-NSs@MSNSs (c) catalysts after running four cycles of catalytic reduction of DCPD. The scale bars are all 200 nm.

Conclusions

To sum up, a well-defined pomegranate-like Ni-NSs@MSNSs core–shell catalyst was successfully synthesized by a modified Stöber method. The core–shell structure nanocomposite had a 150 nm core composed of a number of dispersed Ni-NSs (about 10 nm) and a 25 nm thickness mesoporous SiO2 outer shell. The Ni-NSs@MSNSs had a high specific area (658 m2/g) and uniform mesoporous structure (1.7–3.3 nm). The catalytic performance of Ni-NSs@MSNSs was evaluated by hydrogenation of DCPD to endo-THDCPD. Under test conditions (Ni/DCPD/cyclohexane = 1/100/1000 (w/w), 150 °C, and 2.5 MPa), the pomegranate-like Ni-NSs@MSNSs core–shell nanocatalyst exhibited superior catalytic activity (TOF = 106.0 h–1 and STY = 112.7 g·L–1·h–1) and selectivity (98.9%) than conventional Ni-based catalysts. Furthermore, the Ni-NSs@MSNSs nanocatalyst could be rapidly and completely magnetically separated from the reaction solution by a 4500 G external magnet in only 40 s with no obvious weight loss. The Ni-NSs@MSNSs nanocatalyst exhibited excellent thermal stability (≥750 °C) and superior recyclability (without an activity and selectivity decrease in four runs), which was owing to their unique pomegranate-like structure and core–shell synergistic effect. As an effective strategy, the present work also provided an approach to embed ultrafine metal NPs into a meso- and microporous silica shell to produce high-performance catalysts with broad application prospects in various catalytic fields.

Experimental Section

Materials

Nickel nitrate (Ni(NO3)2·6H2O, AR), sodium oleate (AR), oleic acid (OA, AR), ammonium hydroxide (AR, 25–28%), tetraethyl orthosilicate (TEOS, AR), and texadecyltrimethylammonium bromide (CTAB, 99%) were purchased from Sinopharm Chemical Reagent Co., Ltd. Absolute ethanol (AR, 99.7%) was obtained from Beijing Tongguang Fine Chemical Co., Ltd. Trichloromethane (CHCl3, AR) and acetone (AR) were purchased from Tianjin Fuyu Fine Chemical Co., Ltd. 1-Octadecene (ODE, 90%) and dicyclopentadiene (DCPD, AR) were obtained from Shanghai Energy Chemical Co., Ltd. Cyclohexane (AR, 99.5%) and n-hexane (AR, 97%) were purchased from Shanghai Macklin Biochemical Technology Co., Ltd. Deionized water was made in the lab. All chemicals were of analytical grade and used as received without any purification. Air, nitrogen, hydrogen, and argon were all purchased from Beijing Yongsheng Gas Technology Co., Ltd.

Catalyst Preparation

Synthesis of Ni-NSs Catalysts

Hydrophobic OA-capped NiO-NSs were synthesized by the thermal decomposition method with modification (Scheme a).[76,77] Sodium oleate (120 mmol) and Ni(NO3)2·6H2O (60 mmol) were dissolved in a mixture solvent composed of 80 mL of absolute ethanol, 60 mL of distilled water, and 140 mL of n-hexane. Then, the resulting solution was heated to 60 °C and reacted for 4.0 h to prepare nickel oleate. The synthesized nickel oleate complex and 2.3 mL of OA were dissolved in 55.6 mL of ODE, heated to 320 °C for 1.0 h, cooled down to room temperature, and washed with n-hexane to produce hydrophobic NiO-NSs. Lastly, the obtained hydrophobic NiO-NSs were dispersed in chloroform and centrifuged.

Scheme 2

Synthetic Procedure and Structure of Oleic Acid Capped NiO Nanospheres (a), NiO-NSs@CTAB/SiO2 Precursors (b), and the Pomegranate-like Ni-NSs@MSNSs Core–Shell Nanocatalyst (c)

The Ni-NSs catalysts were obtained by calcination of OA-capped NiO-NSs in air at 350 °C for 3 h and subsequent reduction in 5% H2/Ar.

Synthesis of Ni-NSs@MSNSs Core–Shell Catalysts

The hydrophilic CTAB-modified NiO-NSs were prepared via a phase-transfer process from prepared OA-capped NiO-NSs and surfactant CTAB as the capping agent.[78] OA-capped NiO-NSs (15 mg in 1.0 mL of CHCl3) were added to an aqueous CTAB solution (20 mL, 54.8 mM) and then evaporated at 70 °C to remove the chloroform and obtain the hydrophilic CTAB-modified NiO-NSs. Subsequently, a series of NiO-NSs@CTAB/SiO2 core–shell precursors were prepared by polymerizing the silica layers around the surface of NiO-NSs via the modified Stöber method (Scheme b).[79] Firstly, CTAB-modified NiO-NSs (30 mg) were added to a mixture of CTAB (0.3 g), distilled water (80 mL), and absolute ethanol (40 mL) and then dispersed by sonication for 0.5 h, and TEOS was added and sonicated for another 1 h. After that, the silica-CTAB layer was formed around the NiO-NSs under basic conditions (pH 10–11) through an electrostatic interaction between the cationic (CTAB) and anionic (silicate) species. Then, the synthesized NiO-NSs@CTAB/SiO2 NPs were washed three times with ethanol and distilled water and dried at 60 °C followed by calcination at 350 °C for 2 h in static air to remove CTAB surfactants. Finally, the Ni-NSs@MSNSs nanocatalysts were obtained by H2 reduction (Scheme c).

Synthesis of Ni/MSNS Supported Catalysts

The SiO2 nanospheres were synthesized by a modified Stöber method[80,81] with TEOS as a silica source. Ni/MSNSs supported catalysts (Ni 15 wt % and Ni 40 wt %) were prepared by an equal volume impregnation method with silica nanospheres and an aqueous solution of Ni(NO3)2·6H2O.[67] The impregnated samples were obtained after rotary evaporation and dried at 110 °C overnight, followed by calcination at 350 °C for 2 h in static air. Then, the NiO/MSNSs precursors were reduced in a horizontal quartz tubular reaction chamber under 5% H2/Ar atmosphere to produce Ni/MSNSs-x (x = 15, 40) catalysts.

Characterizations

The crystal structures of the prepared nanoparticles were characterized using the XRD (Philips PW1700) with Cu Kα radiation (1.54 Å) operated at 40 kV and 30 mA. The scanning speed is 10°/min from 5 to 85°. The crystalline size of the Ni species was calculated according to the Scherrer Equation (eq ), where D is the mean size of Ni species, K is a dimensionless shape factor, λ is the X-ray wavelength (1.542 Å), β is the line broadening at half the maximum intensity (FWHM), and θ is the Bragg angle.

The size and morphological structure of synthesized materials were observed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The crystalline structure and lattice fringes of individual catalyst were observed by high-resolution transmission electron microscopy (HRTEM) in addition to the X-ray energy dispersive spectrometer (EDS) spectra (JEOL JEM-2100F).

The thermogravimetric analysis (TGA) of synthesized nanoparticles was performed using a Mettler Toledo TGA-SDTA851 analyzer at a heating rate of 10 °C/min from 25 to 900 °C in the presence of air.

The functional groups of the synthesized materials were investigated using Fourier transform infrared (FT-IR) spectroscopy (Perkin-Elmer GX 2000 spectrometer) in the range from 4000 to 400 cm–1.

N2 adsorption–desorption isotherms of catalysts were measured by using a BELSORP-max instrument to estimate the porosity, specific surface area, and pore size distribution at 77 K. Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH) methods were applied to determine the specific surface area and the pore size distribution considering the desorption and adsorption branch of the N2 isotherms, respectively.

X-ray photoelectron spectroscopy (XPS) was performed using a ThermoFisher ESCALAB 250XI instrument, and the C1s binding energy of carbon (284.8 eV) was used as an internal standard to correct the binding energies of the other elements.

Temperature-programmed reduction (TPR) was performed on a dynamic adsorption instrument (TP-5076) equipped with a thermal conductivity detector (TCD). In the TPR experiment, the sample (100 mg) was pretreated at 300 °C in Ar for 1 h and then cooled to room temperature. A gas mixture of 5% H2/Ar was used as the reducing agent with a total flow rate of 50 mL/min. The samples were heated to 980 °C (10 °C/min), and the TPR profile was recorded simultaneously.

The hydrogenation reaction mixture was analyzed by a gas chromatograph (GC-2014, SHIMADZU, Japan) equipped with an HP-5MS column (30 m × 0.25 mm × 0.25 μm) and a flame ionization detector (FID). The area normalization method was used to quantitatively analyze the conversion of DCPD and the yield of each product. The conversion rate of DCPD and endo-THDCPD selectivity were evaluated according to eqs and 3, respectively, where C0 is the DCPD dosage before the reaction, C1 is the DCPD content after the reaction, and S1 is the endo-THDCPD content after the reaction. The turnover frequency (TOF) of the hydrogenation of DCPD on the catalytic active sites (Ni) of the catalysts was calculated by dividing the number of DCPD converted per hour by the number of exposed Ni atoms on the surface of the catalysts determined by XPS.

The magnetic properties of the synthesized Ni-based catalysts were tested by VSM (Lake Share 7404) with an external magnetic field ranging from −25 to +25 kOe.

An ultrasonic processor (KQ-250DE) with a maximum power of 650 W was used for the preparation of the series of core–shell nanocomposites.

Catalytic Activity Evaluation

Hydrogenation of DCPD was carried out in a 100 mL stainless steel autoclave (Shanghai Laibei Scientific Instruments Co., Ltd.) equipped with a mechanical agitator, a pressure controller, and a temperature controller. Cyclohexane was selected as the reaction solvent. Prior to the reaction, a certain mass ratio of pre-reduced catalyst, DCPD, and cyclohexane was introduced in the reactor vessel. The reaction conditions were set at 150 °C, 2.5 MPa, and a stirring rate of 600 rpm. A back-pressure valve was used to guarantee constant pressure in the autoclave. Liquid samples of 1.5 mL were taken from the reactor and quantitatively analyzed by a gas chromatograph. The space–time yield (STY) of the catalyst was calculated according to eq , where M is the weight of raw material before the reaction, C is the conversion of DCPD, S is the selectivity of endo-THDCPD after the reaction, V is the volume of the reactor, and t is the reaction time. The used catalysts were separated using an external magnetic field, washed with cyclohexane, and then kept under cyclohexane for recycling.