Enhanced catalytic properties of rhodium nanoparticles deposited on chemically modified SiO₂ for hydrogenation of nitrile butadiene rubber

Abstract

An efficient and recyclable catalyst was prepared by depositing rhodium nanoparticles (Rh NPs) on a SiO₂ support with the surface modified by 3-aminopropyl-trimethoxysilane (APTS). The obtained catalyst, Rh(III)/MSiO₂, was extensively characterized by FTIR, SEM, XRD, XPS, BET, and ICP to demonstrate the incorporation of amine groups on the surface of the modified SiO₂. The hydrogenation of acrylonitrile butadiene rubber (NBR) was used to evaluate the catalytic properties of Rh(III)/MSiO₂, and a 98.68% conversion and nearly 100% selectivity to C=C were obtained under optimum conditions (120 °C, 3.0 MPa H₂, 8 h). A comparative study was made between Rh(III)/MSiO₂ and Rh(III)/SiO₂ in the hydrogenation of NBR, and Rh(III)/MSiO₂ showed higher activity and higher stability. The results indicated that the increase of the amount of deposited metal particles and the improvement of the adhesion of metal particles to the support are the reasons for the high performance of Rh(III)/MSiO₂. In addition, the Rh(III)/MSiO₂ could be easily recovered by centrifugation and reused 3 times with only a slight loss in activity and selectivity. Furthermore, this modified method can be extended to other heterogeneous catalyst systems to improve their activity, recyclability and stability.

---

1. Introduction

The catalytic hydrogenation of unsaturated polymers is an effective method to prepare products that exhibit improved mechanical, chemical, and thermal properties over those of the original polymers.^1–5 The hydrogenation of acrylonitrile butadiene rubber (NBR) is a representative example to produce hydrogenated acrylonitrile butadiene rubber (HNBR), a polymer that shows excellent resistance to thermal and oxidative degradation. Due to the increasing demand of HNBR on the market of high performance elastomers, the hydrogenation of NBR has attracted significant attention.

Generally, the hydrogenation of NBR can be realized by heterogeneous or homogeneous processes depending on the nature of the catalyst.^6,7 Homogeneous hydrogenation allows the reaction to be carried out at milder operating conditions and without diffusion problems.^6,8–11 Nevertheless, the removal of expensive catalysts from the final polymer product, i.e., HNBR, is often difficult, and the residual catalysts in the product can cause polymer degradation. Therefore, in order to decrease the production cost, the heterogeneous systems have received much attention as an alternative to the use of conventional homogeneous catalysts in the past few decades and has been applied to the hydrogenation of diene-based polymers. The heterogeneous systems can be defined as modified homogeneous catalysts to “break” the uniformity of the catalyst and products by making them in separate phases to allow their separation.¹² The simplest method is to immobilize homogeneous catalysts on various types of organic and inorganic supports.¹³ The most widely explored heterogeneous catalysts for hydrogenation of NBR are Pd/C, Pd/SiO₂, and Pd/TiO₂, all of which have been developed by Zeon Corporation since the 1980s. With these catalysts, a degree of hydrogenation of above 95% can be achieved.^14,15 However, the Pd nanoparticles have poor adhesion to the bare support due to the absence of chemical bonding, which will lead to the leaching of active Pd particles from the support surface during the reaction. The catalyst Pd/SiO₂ was found to give decreasing degree of hydrogenation from 98.8% to 33.3% after one times recycling.¹⁵ Also, the high viscosity of the polymer solution and the steric hindrance of the polymer chains lead to limited contact between the polymer and the supported catalysts. Further, in recent years, little effort has been directed towards the development of new heterogeneous catalytic systems for hydrogenation of NBR with good activity and recyclability.⁶ Therefore, researches in this area are urgently needed to optimize the heterogeneous NBR hydrogenation process.

The amount and quality of deposited active metal particles depend on the surface properties of the support.^16,17 Because of the absence of chemical bonding, the metal particles usually display poor adhesion to the support, resulting in their easy removal from the support surface during reactions. Nevertheless, surface modification with organic functional groups has been found to be useful for increasing the amounts of the metal particles loaded and improving the adhesion between the metal particles and the support.^18,19 In previous investigations, APTS was one of the most commonly used coupling agents for modifying the SiO₂ surfaces.^20–22 And, modified SiO₂ plays a vital role in the development of materials for various applications, such as adsorption of metal ions, catalysis and drug delivery.^23–26 In addition, this method overcomes the problems of metal leaching in comparison with the ion exchange or adsorption methods used for immobilization on porous supports.²⁰ Therefore, in this study, APTS was chosen as a modifier to enhance the metal loading and improve the properties of the supported catalysts.

To the best of our knowledge, Rh NPs loaded on surface-modified SiO₂ have not been studied for the hydrogenation of NBR. Herein, the purpose of this research was to fabricate a surface modified SiO₂ immobilized Rh NPs catalyst, which acts as a highly active, selective and easily recyclable catalyst for hydrogenation of NBR. Also, a comparative study of the Rh(III)/MSiO₂ and Rh(III)/SiO₂ was also performed. The catalytic activity, selectivity, and recyclability were evaluated to better understand the importance of modification.

2. Experimental

2.1 Materials

SiO₂ with different particle sizes (40 nm, 1 μm, 10 μm, 48 μm, and 150 μm) was purchased from XF NANO, Inc. The rhodium precursor RhCl₃·3H₂O was purchased from Shanghai Jiulin Chemical Co., Ltd., with an Rh content of 39.0 wt%. Nitrile rubber (Nancar1052M30) was obtained from Taiwan NANTEX Chemical Industry Co., Ltd., with a number average and a weight average molecular weight of 108 and 307 kg mol⁻¹, respectively, as measured by gel permeation chromatography with tetrahydrofuran (THF) as the solvent at 35 °C. 3-Aminopropyl-triethoxysilane (APTS) was purchased from Tianjin Guangfu Research Institute of Fine Chemicals. Chlorobenzene (AR), anhydrous ethanol (AR), acetone (AR), and ammonia (25 wt%) were all obtained from Beijing Chemical Factory. Triphenylphosphine and deuterated chloroform were obtained from National Chemical Reagent Co., Ltd. Nitrogen (research grade; N₂ > 99.99%) and hydrogen (research grade; H₂ > 99.99%) were donated by Beijing Beifen Gas Industries Co., Ltd.

2.2 Preparation of modified silica (MSiO₂)

The SiO₂ was modified with APTS in order to covalently attach a –NH₂ termination to its surface according to published procedures.²⁷ SiO₂ (5.0 g, dried under vacuum at 100 °C for 6 h before use) was first dispersed into 200 mL of deionized water under stirring at room temperature for 15 min to form slurry. APTS (2 mL) was added into the slurry slowly, and heated to 40 °C for 1 h. Then, 8 mL of ammonia (25 wt%) was added dropwise to promote the reaction, and the slurry was further stirred at 40 °C for another 2 h. After reaction, the mixture was sonicated for 15 min, recovered by filtration, and washed with copious amounts of fresh deionized water in order to remove the loosely bound APTS. Finally, the MSiO₂ was dried at 80 °C overnight.

2.3 Preparation of supported catalysts

The MSiO₂ (2.5 g) was impregnated in 50 mL of deionized water and dispersed for 30 min in an ultrasonic bath. 5 mL of RhCl₃·3H₂O solution (10 g L⁻¹) was added to the suspension, following by stirring vigorously for 10 h. Because the amine group is a strong electron donor having a strong capability for chelating to transition metal ions, Rh³⁺ can react with –NH₂ to form a complex and be bound to the SiO₂ surface. After impregnation, the slightly yellow powder was recovered by filtration, washed with deionized water for several times, and dried under vacuum at 80 °C for 12 h. The powder sample was designated as Rh(III)/MSiO₂ 1. The Rh(III)/MSiO₂ 1 was later reduced by NaBH₄ as follows: Rh(III)/MSiO₂ 1 (2.5 g) was dispersed into deionized water (200 mL) under stirring for 30 min. Then, a freshly prepared NaBH₄ solution (1.0 g of NaBH₄ in 30 mL of deionized water) was added dropwise into the mixture, stirring for an additional 3 h. The final solids were recovered by filtration, washed repeatedly with deionized water, dried under vacuum, and stored in a glass bottle for further use. The sample was designated as Rh(0)/MSiO₂ 2. The preparation processes of Rh(III)/MSiO₂ 1 and Rh(0)/MSiO₂ 2 are illustrated in Scheme 1.

Scheme 1 Preparation of supported catalysts.

For comparison, under the same conditions, the SiO₂ without modification was impregnated with an aqueous solution of RhCl₃·3H₂O to form Rh(III)/SiO₂ 3, which was then reduced with NaBH₄ to form Rh(0)/SiO₂ 4. The surface features of Rh(III)/SiO₂ 3 and Rh(III)/MSiO₂ 1 are shown in the Fig. S1†.

2.4 Catalytic activity for hydrogenation of NBR

The hydrogenation reaction was carried out in a 500 mL high pressure reactor. The NBR chlorobenzene solution (3 wt%) and catalyst were placed into the reactor. Before being heated to the reaction temperature, the reactor was degassed with N₂ for 3 times to remove air. Then the reactor was heated to the desired temperature, and the reaction was initiated by flushing H₂ to 3.0 MPa and adjusted the agitation rate to 480 rpm. The hydrogen pressure and reaction temperature were kept constant throughout the reaction. The conversion of the double bonds, also called the degree of hydrogenation (HD), was calculated as:

HD = (1 − [C=C]_t/[C=C]₀) × 100% (1)

where [C=C]_t is the concentration of C=C bonds at reaction time t and [C=C]₀ is the initial double bond concentration. The hydrogenated NBR sample was analyzed by ¹HNMR.²⁸

2.5 Catalyst recovery

After each hydrogenation reaction, catalyst was removed from the NBR glue by centrifugation (8000 rpm min⁻¹ for 30 min).

Then the catalyst washed with acetone for 3–4 times followed by decantation and dried under vacuum at 60 °C for subsequent reaction.

2.6 Catalyst reactivation

A known amount of recovered catalyst and 100 mL of chlorobenzene were transferred into a flask, heated to 60 °C in an oil bath, stirring vigorously for 8 h. Then, the catalyst was separated from the chlorobenzene and dispersed in acetone overnight in an Erlenmeyer flask. The final recovered catalyst was obtained by centrifugation, fully washed with acetone and ethanol, and dried under vacuum at 60 °C for use in the next hydrogenation reaction.

2.7 Characterization

The physical and chemical properties of the prepared samples were measured by several methods. FTIR spectra were obtained by a Bruker Tensor 27 spectrophotometer in the wave-number range of 400–4000 cm⁻¹. The morphology and particle size of the samples were examined by scanning electron microscopy (SEM, Hitachi S4800 Dual Stage Electron Microscope) at 10 kV. The crystal phases of the samples were determined by X-ray diffraction (XRD, Bruker D8 Advance) at a scanning rate of 0.02° s⁻¹ in the 2θ range of 5° to 90°. The specific surface area, pore size distribution, pore volume, and adsorption–desorption isotherms of the samples were measured by N₂ physical adsorption at 77 K (Micrometrics ASAP-2020M adsorption apparatus). X-ray photoelectron spectroscopy (XPS) was performed on a Thermo Fisher Scientific ESCALAB 250 system with monochromatized Al KR radiation at 15 kV. A thin layer of the sample was dispersed on the surface of a carbon tape, evacuated in a load lock, and then transferred into the analysis chamber for measurement. The Rh concentration in the samples was determined by using Inductive Coupled Plasma (ICP) Optical Emission Spectroscopy (ICPS-7500). For ICP analyses, the samples were digested in 10% (v/v) nitric acid and 40 wt% hydrogen fluoride (HF) at 120 °C for 30 min.

3. Results and discussion

3.1 Characterization of supported catalysts

The changes in chemical structure of SiO₂ modified with APTS were investigated by FTIR, as shown in Fig. 1. The intensities of the peaks of APTS are weak due to the low concentration. The peak at 471 cm⁻¹ attributed to the bending vibration of Si–O–Si; the peaks at 800 cm⁻¹ and 1103 cm⁻¹ represent the asymmetric and symmetric stretching vibrations of Si–O; and the peaks at 1103 cm⁻¹ and 3441 cm⁻¹ are the characteristic peaks of the hydroxyl groups on the SiO₂ surface. For MSiO₂, however, in addition to the absorption peaks in SiO₂, the peak at 3441 cm⁻¹ broadens, having a clear shoulder structure around 3378 cm⁻¹, which is the characteristic peak of N–H stretching vibration, and the peaks for the bending vibrations of the –CH₂ and –CH₃ structures at 2924 cm⁻¹ and 2854 cm⁻¹ become more visible, but the band at around 1632 cm⁻¹ appears slightly less intense, indicating successful incorporation of APTS onto the SiO₂.

Fig. 1 FTIR spectra of SiO₂ and MSiO₂.

Fig. S2† shows that the interface of SiO₂ particles is indistinct, whereas the MSiO₂ particles remain relatively well dispersed. The large numbers of hydroxyls on the surface of SiO₂ generate strong hydrogen bonding, resulting in aggregation of SiO₂ particles. After modification, the APTS reacts with the hydroxyls on the surface of SiO₂, leading to a decreased hydrogen bonding interaction. In addition, the surface tension of SiO₂ decreases after hydrolysis and condensation, and APTS can act as a physical barrier to prevent the SiO₂ particles from re-aggregation. Therefore, the particles of MSiO₂ display a well-dispersion compare with that of SiO₂. Both Rh(III)/SiO₂ 3 or Rh(III)/MSiO₂ 1 consists of small particles of 40–60 nm in diameter, and the latter is slightly better dispersed than the former.

The XRD patterns of RhCl₃, SiO₂, MSiO₂, Rh(III)/SiO₂ 3, and Rh(III)/MSiO₂ 1 are illustrated in Fig. 2. It can be seen that RhCl₃, SiO₂, MSiO₂, Rh(III)/SiO₂ 3 and Rh(III)/MSiO₂ 1 all present the characteristic peak of amorphous SiO₂ at 24.40°, indicating the structure of SiO₂ is well preserved after modification. However, the peaks at 15.51°, 36.75°, 52.75°, and 65.94°, corresponding to the (001), (131), (060), and (004) planes, respectively, of Rh, are not observed in Rh(III)/SiO₂ 3 and Rh(III)/MSiO₂ 1 because of (1) the strong diffraction peak of SiO₂ hiding the weak diffraction peak of Rh element; (2) the small amount of Rh, which is difficult to detect; (3) the poor dispersion of Rh ion on the surface of SiO₂; and (4) the Rh ion infiltrating into the pores of SiO₂.

Fig. 2 XRD-patterns of RhCl₃, SiO₂, MSiO₂, Rh(III)/SiO₂ 3 and Rh(III)/MSiO₂ 1.

N₂ adsorption–desorption isotherms were used to determine the changes in the BET surface area, pore volume, and average pore diameter of SiO₂ and MSiO₂, and the results were shown in Fig. S3†. The calculated surface area, pore volume, and pore diameter are listed in Table S1†. This information is useful for comparing the adsorption properties of SiO₂ and MSiO₂. The obtained isotherms for the SiO₂ and MSiO₂ show reversible type IV behavior.²⁹ After modification, the pores of SiO₂ are filled with APTS, leading to a decrease in average pore diameter. Therefore, the surface area and pore volume of MSiO₂ are smaller than those of SiO₂, indicating successful modification by APTS.

XPS survey was conducted to study the attachment of APTS layer onto the SiO₂. The XPS spectra of SiO₂ and MSiO₂ are shown in Fig. S4†. The main components of SiO₂, such as O and Si, exhibit strong absorption intensity. An absorption peak with low intensity near 400 eV is detected for the MSiO₂ due to the nitrogen atom from APTS. There is a distinct increase in the absorption intensity of the C 1s peak because the propyl chains of APTS are adsorbed onto SiO₂. In the case of SiO₂, the C 1s peak is mainly due to adventitious carbon contamination.

The XPS Si 2p and N 1s spectra of SiO₂ (a and c) and MSiO₂ (b and d), respectively, are seen in Fig. 3. In the Si 2p spectrum of SiO₂, the peak at BE ∼103.72 eV is characteristic of the Si⁴⁺ in SiO₂; a smaller peak at BE ∼105.92 eV indicates the presence of Si–OH species. For the MSiO₂, the characteristic peak of Si⁴⁺ is at a BE of ∼103.72 eV and the Si–OH peak is at 105.69 eV. The important data from the XPS Si 2p spectra are calculated in Tables S2 and S3†. The ratio of the Si⁴⁺ to Si–OH is approximately 97.66/2.34 for SiO₂, but nearly 100/0 for MSiO₂. The reason for the disappearance of the Si–OH peak in MSiO₂ is the formation of Si–O–Si bonds during the modification process.

Fig. 3 X-ray photoelectron spectra of (a and c) SiO₂ and (b and d) MSiO₂.

The presence of APTS on the surface of SiO₂ was further confirmed by the N 1s spectrum of MSiO₂ in Fig. 3d. It shows two peaks attributable to the hydrogen-bonded NH₂ of APTS at about 401.3 eV and the “free” NH₂ terminal groups of APTS, which are not hydrogen bonded, at ∼399.8 eV. An extremely weak signal in the N 1s spectrum of the SiO₂ may be due to adventitious nitrogen contamination. The relatively strong peak at a BE of ∼400 eV in the N 1s spectrum of MSiO₂ indicates that the surface of SiO₂ has been successfully modified. For MSiO₂, the ratio of the hydrogen bonded NH₂ to the “free” NH₂ is approximately 76/24. The important data from the XPS N 1s spectra of MSiO₂ are presented in Table S4†.

The main components of the two catalysts, Rh(III)/SiO₂ 3 and Rh(III)/MSiO₂ 1, such as Si and O, exhibit strong absorption intensity, while Rh shows relatively weak absorption intensity due to its low concentration (Fig. 4). Table 1 reports the surface composition (in atom%) of the Rh(III)/SiO₂ 3 and Rh(III)/MSiO₂ 1 samples as determined by XPS. The Rh concentration increases eightfold after modification. The weak peak in the Rh 3d spectrum of Rh(III)/SiO₂ 3 is attributed to the low rhodium content of 0.16 atom%, which is too low for detection by XPS. The increase in the C/O atomic ratio from 24.33% in Rh(III)/SiO₂ 3–37.08% in Rh(III)/MSiO₂ 1 also confirms that the SiO₂ has been successfully modified by APTS.

Fig. 4 XPS survey spectra (a) Rh(III)/SiO₂ 3 and (b) Rh(III)/MSiO₂ 1; XPS Rh 3d spectra for (c) Rh(III)/SiO₂ 3 and (d) Rh(III)/MSiO₂ 1.

Table 1 Surface compositions (in atom%) of Rh(III)/SiO₂ 3 and Rh(III)/MSiO₂ 1

Samples	C (%)	O (%)	Si (%)	N (%)	Rh (%)	C/O (%)
Rh(III)/SiO2 3	13.61	55.95	30.06	0.21	0.16	24.33
Rh(III)/MSiO2 1	18.92	51.03	27.55	1.88	1.22	37.08

---

3.2 Catalytic performance of rhodium catalysts

The as-prepared rhodium catalysts were used in hydrogenation of NBR to evaluate the catalytic behaviour. Table 2 compares the catalytic performance of the various rhodium catalysts for hydrogenation of NBR under similar conditions.

Table 2 Catalytic performance of various rhodium catalysts for hydrogenation of NBR

Catalysts	Wcat./WNBR	WPPh3/WNBR	HD/%^a
a HD = degree of hydrogenation, Wcat. = weight of rhodium catalyst, WNBR = weight of NBR, and WPPh3 = weight of triphenylphosphine. The reaction is conducted at 3.0 MPa of pressure of H2 and 120 °C for 8 h.
Rh(0)/SiO2 4	0.05	0	<5
Rh(0)/SiO2 4	0.05	0.025	8.46
Rh(0)/MSiO2 2	0.05	0	6.81
Rh(0)/MSiO2 2	0.05	0.025	22.57
Rh(III)/SiO2 3	0.05	0.025	52.44
Rh(III)/MSiO2 1	0.05	0.025	98.68

---

It can be seen that the degree of hydrogenation is very low for Rh(0)/SiO₂ 4 and Rh(0)/MSiO₂ 2. A possible reason is that Rh⁰ has low activity towards NBR. However, under identical reaction conditions, the degree of hydrogenation increases to 8.46% and 22.57%, respectively, for Rh(0)/SiO₂ 4 and Rh(0)/MSiO₂ 2 with triphenylphosphine (PPh₃) as assistant agent. The reason can be attributed to the chemical adsorption of PPh₃ on the surface of the catalyst, promoting in situ growth of the active Rh^δ+ (0 < δ ≤ 1) species,³⁰ which is beneficial to the catalytic hydrogenation of NBR. This observation also suggests that the activity of these catalysts is largely associated with the situ growth of the active species Rh^δ+ (0 < δ ≤ 1), rather than the Rh⁰.

The degree of hydrogenation is 52.44% and 98.68% for Rh(III)/SiO₂ 3 and Rh(III)/MSiO₂ 1, respectively, with PPh₃ as assistant agent. The high degree of hydrogenation is attributed to the in situ growth of the Rh⁺ active species between Rh³⁺ and PPh₃.

The degree of hydrogenation was determined by ¹HNMR. Fig. 5 shows the ¹HNMR spectra of (a) NBR and (b) HNBR. It can be seen that Rh(III)/MSiO₂ 1 shows higher activity than Rh(III)/SiO₂ 3 for hydrogenation of NBR under the same conditions. All the peaks in the olefinic proton region have disappeared in the spectrum for Rh(III)/MSiO₂ 1, which gives a degree of hydrogenation of 98.68%, higher than the degree of hydrogenation of 52.44% for Rh(III)/SiO₂ 3, which remained a few olefinic chains. The characteristic peaks for –NH₂ and –NH– are not observed in the ¹HNMR spectrum of Rh(III)/MSiO₂ 1, indicating that the catalyst is selective in the hydrogenation of NBR, which takes place on the C=C double bonds without reduction of the –CN group.

Fig. 5 ¹H NMR spectra of (a) NBR and (b) HNBR.

To understand the underlying mechanism that lead to the higher activity of Rh(III)/MSiO₂ 1 than that of Rh(III)/SiO₂ 3 for hydrogenation of NBR, we focus on the cause for the different activity between the two catalysts. In a previous research,²⁷ the activity of a catalyst was strongly affected by the amount of metal active centers and their dispersion. In this study, because the –NH₂ groups on the surface of MSiO₂ can significantly affect the adsorption of metal ions, we hypothesized that the utilization of APTS increases the amount of the rhodium loaded and improves the adhesion between the rhodium ions and the MSiO₂. Comparison experiments were carried out to confirm this hypothesis.

First, the MSiO₂ was calcined at 550 °C for 4 h to remove the –NH₂ groups on its surface. Then, the calcined MSiO₂ was impregnated with RhCl₃ solution (see Fig. 6c) to fabricate the supported catalyst Rh(III)/MSiO₂ with calcination. The variations of the degree of hydrogenation for different catalysts are shown in Fig. 7. The catalytic performance of Rh(III)/MSiO₂ with calcination is similar to that of Rh(III)/SiO₂ 3 because of the removal of the –NH₂ by calcination. The results indicates that APTS plays a key role in increasing the amount of rhodium loaded and enhancing the adhesion between the rhodium ions and the MSiO₂ support, causing the higher activity of Rh(III)/MSiO₂ 3 than that of Rh(III)/SiO₂ 1.

Fig. 6 Images of different silica supports, (a) SiO₂, (b) MSiO₂, and (c) MSiO₂ calcined, impregnated with RhCl₃ solution.

Fig. 7 Degree of hydrogenation as function of time for various silica-supported catalysts: Rh(III)/SiO₂, Rh(III)/MSiO₂, and Rh(III)/MSiO₂ with calcination. Reaction conditions: NBR solution (200 mL, 3 wt%), rhodium supported catalyst (0.30 g, 5 wt%), and triphenylphosphine (0.15 g, 0.25 wt%) were heated at 120 °C under H₂ pressure of 3.0 MPa.

Different solvents, such as water, a mixture of water–ethanol (volume ratio = 1 : 1), and ethanol were investigated to optimize the surface coating process for the modification of SiO₂. The catalytic performance of Rh(III)/MSiO₂ 1 prepared by using different solvents is illustrated in Table S5†. All three solvents give a similar degree of hydrogenation, implying that the use of these solvents results in similar amounts and dispersion of surface Rh NPs.

The support has an effect on catalyst selectivity as well as activity.^31,32 Therefore, we also investigated the hydrogenation of NBR, using catalyst Rh(III)/MSiO₂ 1 with different SiO₂ particles sizes: 40 nm, 1 μm, 10 μm, 48 μm, and 150 μm. It shows that the degree of hydrogenation increases with decreasing particle size of SiO₂ because of the increase in surface area (see Table S6†). However, Rh(III)/MSiO₂ 1 with a particle size of 1 μm exhibits a higher degree of hydrogenation than Rh(III)/MSiO₂ 1 with a particle size of 0.04 μm because the large NBR coils cannot access the micro and mesopore structures of the SiO₂, where the active Rh species are located.

Of the catalysts prepared in this study, Rh(III)/MSiO₂ 1 with a particle size of 48 μm exhibited relatively high activity, selectivity, and separation for the hydrogenation of NBR. In the following investigations, we used only the heterogeneous catalyst system of Rh(III)/MSiO₂ 1 with a particle size of 48 μm and PPh₃ as ligand for the hydrogenation of NBR.

The effects of various conditions, such as reaction temperature, time, and pressure, concentration of NBR, and concentration of catalyst, on the hydrogenation of NBR using Rh(III)/MSiO₂ 1 are shown in Fig. S5†. Each factor was studied in turn, all other factors being kept constant. The degree of hydrogenation gradually increases with increasing reaction time and concentration of catalyst. The degree of hydrogenation increases with increasing temperature, reaches a maximum at about 120 °C, and then decreases with further increases in temperature, probably because of catalyst deactivation. Similarly, the degree of hydrogenation reaches a maximum of 98.40% at a concentration of NBR of 3 wt%. The degree of hydrogenation decreases with further increases in the concentration of NBR because the increasing viscosity of the NBR solution limits diffusion. The degree of hydrogenation increases with increasing hydrogen pressure because of the limited solubility of H₂ gas in the solvent.

Therefore, we concluded that the modification of SiO₂ can effectively improve the amount of active rhodium species and enhance the adhesion between the rhodium ions and the MSiO₂ support. The superior catalytic performance of Rh(III)/MSiO₂ 1 is ascribed to its superior physicochemical properties.

3.3 Reusability of Rh(III)/MSiO₂ 1

A significant advantage of heterogeneous catalyst is the ability to easily remove from the reaction mixture and reuse it for subsequent reactions.^33,34 The reusability of Rh(III)/MSiO₂ 1 was examined for the hydrogenation of NBR at 120 °C under high pressure by using 5 wt% of catalyst with a Rh loading of 1.18 wt%. The results of the recycling experiments are shown in the Table 3.

Table 3 Hydrogenation of NBR catalyzed by recycle catalyst of Rh(III)/MSiO₂ 1^a

Recycle	1	2	3	4a
a 1: the fresh catalyst; 2, 3: the used catalyst; 4a: the reactivated catalyst. Reaction conditions: concentration of NBR solution = 3 wt%, concentration of catalyst = 5 wt%, WPPh3/Wcat. = 0.5, PH2 = 3.0 MPa, and T = 120 °C.
HD (%)	98.53	90.35	67.61	81.66

---

In a previous research,¹⁵ a Pd/SiO₂ catalyst lost about 65% of its activity in the hydrogenation of NBR after one continuous cycles, whereas Rh(III)/MSiO₂ 1 lost about 8.18% under the same conditions. The results indicate that the catalytic stability of Rh(III)/MSiO₂ 1 is superior to that of Pd/SiO₂. In the Rh(III)/MSiO₂ 1 system, the SiO₂ was modified by APTS before the conventional impregnation process, and thus the Rh NPs are attached to the SiO₂ surface with chemical bonds, as shown in Scheme 1. However, in the Pd/SiO₂ system, the Rh NPs were attached to the SiO₂ surface with van der Waals forces. Because the binding force of a covalent bond is much stronger than the van der Waals forces, it is harder for the Rh NPs to detach from the SiO₂ surface and higher stability can be obtained.

Also, to understand the deactivation mechanism of the Rh(III)/MSiO₂ 1, we carried out leaching experiments to determine whether the hydrogenation reaction occurs with the Rh species on the support surface or the leached homogeneous Rh species.

Rh(III)/MSiO₂, PPh₃, and NBR solution were added into a 500 mL high pressure reactor, maintaining stirring at 120 °C for 2 h, and the reaction products were centrifuged to obtain a solid catalyst Rh(III)/MSiO₂ (A) and liquid (B) (Scheme 2). Both the solid (A) and the liquid (B) were used for next hydrogenation reaction; the result is shown in the Fig. S6†. When the recycled catalyst (A) was used for the hydrogenation of NBR, HNBR was obtained with a degree of hydrogenation of 87.32%. In contrast, when the liquid (B) was allowed to react under similar conditions, but without a catalyst, the degree of hydrogenation was only 11.86%. The result shows that the leaching of a small amount of Rh species occurred due to the viscous NBR solution, high temperature, and high pressure. Therefore, the leaching of Rh species is one of the reasons giving rise to the decrease in activity. Also, the amount of leaching Rh species in the recovered catalyst can be verified by ICP analysis. Table S7† shows that after two consecutive cycles, the rhodium amount in the recovered catalyst is 0.98 wt%, which is 83% of the initial rhodium content in the fresh catalyst. Also, it should be noted that the rhodium amount decreases from the first to the second run and then remains constant.

Scheme 2 Leaching experiment for Rh(III)/MSiO₂ 1 for hydrogenation of NBR.

SEM was employed to investigate the evolution of the morphology of the recycled Rh(III)/MSiO₂ 1 (Fig. S7†). It can be seen that the Rh NPs re-aggregate and a thin film covers the surface of the SiO₂, possibly because of the high viscosity of the NBR solution, thus resulting in a decrease in activity. In addition, XPS and FITR demonstrate that the chemical structure of reused catalyst Rh(III)/MSiO₂ 1 has not changed compare with the fresh catalyst (Fig. S8 and S9†)

Hence, we conclude that the leaching of Rh active centers is not the only reason for catalyst deactivation. In order to verify this conclusion, another experiment was carried out to restore the activity of the used catalyst. The recovered Rh(III)/MSiO₂ 1 was dispersed in chlorobenzene, and then the system was heated to 60 °C for 8 h to remove the NBR glue on the surface or inside the pore of MSiO₂. In addition, the resulting catalyst was washed with acetone for three times and dried under vacuum at 60 °C. After reactivation, the activity of Rh(III)/MSiO₂ 1 was restored to 81.66% (Table 3, cycle 4a), compare with 67.61% in the third recycle (Table 3, cycle 3).

Therefore, the present results suggest a deactivation mechanism: a small quantity of Rh species (<10 wt%) leach into the reaction solution, the SiO₂ particles re-aggregate, and the highly viscous NBR solution infiltrates the pores of the SiO₂, leading to a large mass transfer resistance.

Some fresh catalyst (15 wt% initial weight) was added along with the recovered catalyst from the first hydrogenation cycle to apply in hydrogenation of NBR, and the results are shown in Table 4. With 15 wt% fresh catalyst added, the activity of the blend consisting of recovered and fresh catalyst was fully restored to 96.42%. After the three times cycle, the degree of hydrogenation for the blend still remains at 96.13%, an indication of the high stability and utility of the catalyst Rh(III)/MSiO₂ 1.

Table 4 Comparison of activity of fresh catalyst and blend of 15 wt% fresh catalyst and recovered catalyst^a

Recycle	1	2	3	4
a Reaction conditions: concentration of NBR solution = 3 wt%, concentration of catalyst = 5 wt%, WPPh3/Wcat. = 0.5, PH2 = 3.0 MPa, and T = 120 °C.
HD (%)	97.53	96.42	95.88	96.13

---

Further, this modification method is not restricted to the SiO₂ support and the Rh species. The hydrogenation of NBR was carried out by using other supports, such as the molecular sieves of 13×, MCM-41 and SBA-15. Fig. S10† shows that before modification, the degree of hydrogenation of the supported catalysts was 67.64%, 48.75%, and 67.42% for 13×, MCM-41, and SBA-15, respectively. After modification, the degree of hydrogenation for the supported catalysts increased to 96.34%, 89.46%, and 98.86%, respectively. These obvious increases show that modification of these supports with APTS enhances the activity of the Rh catalyst concerned. In addition, other congeneric transition metal (Ru³⁺ and Pd²⁺) complexes deposited on the surface of modified SiO₂ also have enhanced catalytic performance (Fig. S11†). For example, the degree of hydrogenation increases from 20.49% to 45.58% after modification for a Ru catalyst in hydrogenation of NBR. These results indicate that this method of surface modification by APTS can be extended to other catalyst systems to improve their activity, reusability, and stability.

4. Conclusions

The surface of SiO₂ was chemically modified with the silane coupling agent APTS, and Rh NPs are deposited on MSiO₂ surface to fabricate an Rh-loaded supported catalyst, Rh(III)/MSiO₂. FTIR, SEM, XRD, XPS, BET, and ICP demonstrated that the Rh NPs were well dispersion on the surface of MSiO₂. In addition, the Rh(III)/MSiO₂ showed a higher catalytic activity and stability in hydrogenation of NBR than Rh(III)/SiO₂. Factors that influence the degree of hydrogenation were also fully investigated and the optimum reaction conditions were as follows: 200 mL of NBR solution (3 wt%), 0.12 g of supported catalyst (2 wt%), 0.06 g of triphenylphosphine (1 wt%), temperature of 140 °C, H₂ pressure of 3.0 MPa, and reaction time of 8 h. A degree of hydrogenation above 98% was obtained. And the Rh(III)/MSiO₂ could be recovered by a simple centrifugation and recycled for three consecutive runs with less loss of activity. After treatment of the recovered catalyst with chlorobenzene and acetone, the catalytic activity of the catalyst was restored from 67.61% to 81.66%. The enhanced interaction between the metal ions and the support, and the increase of the amount of metal species as well as their dispersion could be the reasons for the high activity of Rh(III)/MSiO₂. This modified method may be a promising way to improve the activity and stability of other supported catalysts.

Acknowledgements

The financial supports of the National Natural Science Foundation of China under Grant no. 20774009 are gratefully acknowledged.

Notes and references

1. K. Y. Han, H. R. Zuo, Z. W. Zhu, G. P. Cao, C. Lu and Y. H. Wang, Ind. Eng. Chem. Res., 2013, 52, 17750–17759.
2. S. Maheshwari, M. Tsapatsis and F. S. Bates, Macromolecules, 2007, 40, 6638–6646.
3. H. Wang, Q. Pan and G. L. Rempel, J. Polym. Sci., Part A: Polym. Chem., 2012, 50, 4656–4665.
4. H. Wang, Q. Pan and G. L. Rempel, J. Polym. Sci., Part A: Polym. Chem., 2012, 50, 2098–2110.
5. L. Wei, J. Jiang, Y. Wang and Z. Jin, J. Mol. Catal. A: Chem., 2004, 221, 47–50.
6. H. Wang, L. Yang and G. L. Rempel, Polym. Rev., 2013, 53, 192–239.
7. L. B. Dong, S. Turgman-Cohen, G. W. Roberts and D. J. Kiserow, Ind. Eng. Chem. Res., 2010, 49, 11280–11286.
8. J. S. Parent, N. T. McManus and G. L. Rempel, Ind. Eng. Chem. Res., 1996, 35, 4417–4423.
9. J. S. Parent, N. T. McManus and G. L. Rempel, Ind. Eng. Chem. Res., 1998, 37, 4253–4261.
10. A. Mahittikul, P. Prasassarakich and G. L. Rempel, J. Mol. Catal. A: Chem., 2009, 297, 135–141.
11. D. M. Yue, Z. M. Shen, R. Q. Xu and Y. K. Wei, J. Elastomers Plast., 2002, 34, 225–237.
12. A. E. C. Collis and I. T. Horvath, Catal. Sci. Technol., 2011, 1, 912–919.
13. J. M. Thomas, J. C. Hernandez-Garrido, R. Raja and R. G. Bell, Phys. Chem. Chem. Phys., 2009, 11, 2799–2825.
14. Y. Kubo and K. Ohura, US Patent, 4337329, 1982.
15. Y. Kubo and K. Oura, US Patent, 4452951, 1984.
16. H. Dai, H. Li and F. Wang, Appl. Surf. Sci., 2006, 253, 2474–2480.
17. G. O. Williams, G. M. O'Connor, P. T. Mannion and T. J. Glynn, Appl. Surf. Sci., 2008, 254, 5921–5926.
18. R. Chen, Y. Jiang, W. Xing and W. Jin, Ind. Eng. Chem. Res., 2011, 50, 4405–4411.
19. H. Wu, Z. Liu, X. Wang, B. Zhao, J. Zhang and C. Li, J. Colloid Interface Sci., 2006, 302, 142–148.
20. S. J. J. Titinchi and H. S. Abbo, Catal. Today, 2013, 204, 114–124.
21. I. Slowing, B. G. Trewyn and V. S. Y. Lin, J. Am. Chem. Soc., 2006, 128, 14792–14793.
22. V. Antochshuk, O. Olkhovyk, M. Jaroniec, I.-S. Park and R. Ryoo, Langmuir, 2003, 19, 3031–3034.
23. F. Balas, M. Manzano, P. Horcajada and M. Vallet-Regí, J. Am. Chem. Soc., 2006, 128, 8116–8117.
24. A. M. Liu, K. Hidajat, S. Kawi and D. Y. Zhao, Chem. Commun., 2000, 1145–1146.
25. X. Wang, K. S. K. Lin, J. C. C. Chan and S. Cheng, J. Phys. Chem. B, 2005, 109, 1763–1769.
26. D. J. Macquarrie and D. B. Jackson, Chem. Commun., 1997, 1781–1782.
27. H. Li, H. Jiang, R. Chen, Y. Wang and W. Xing, Ind. Eng. Chem. Res., 2013, 52, 14099–14106.
28. A. J. Marshall, I. R. Jobe, T. Dee and C. Taylor, Rubber Chem. Technol., 1990, 63, 244–255.
29. C. H. Huang, K. P. Chang, H. D. Ou, Y. C. Chiang and C. F. Wang, Microporous Mesoporous Mater., 2011, 141, 102–109.
30. L. Yan, Y. J. Ding, H. J. Zhu, J. M. Xiong, T. Wang, Z. D. Pan and L. W. Lin, J. Mol. Catal. A: Chem., 2005, 234, 1–7.
31. I. Lee, M. A. Albiter, Q. Zhang, J. Ge, Y. Yin and F. Zaera, Phys. Chem. Chem. Phys., 2011, 13, 2449–2456.
32. F. Zaera, Chem. Soc. Rev., 2013, 42, 2746–2762.
33. W. Long, N. A. Brunelli, S. A. Didas, E. W. Ping and C. W. Jones, ACS Catal., 2013, 3, 1700–1708.
34. S. Moussa, A. R. Siamaki, B. F. Gupton and M. S. El-Shall, ACS Catal., 2012, 2, 145–154.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra11711e

---