Synthesis and characterization of a multifunctional nanocatalyst based on a novel type of binary-metal-oxide-coated Fe₃O₄–Au nanoparticle

Abstract

A novel type of binary-metal-oxide-coated Au nanocatalyst, including a mixed oxide layer, a moveable magnetic Fe₃O₄ core and some Au NPs of 2–5 nm, has been synthesized successfully by a facile hydrothermal synthesis method. SEM, TEM, EDX, FTIR, XRD, and TGA were employed to characterize the prepared samples. The results showed the mSiO₂–TiO₂ layer could increase the thermal stability and reactivity of metal nanocatalysts compared to a pure TiO₂ or SiO₂ layer. The reduction of 4-nitrophenol (4-NP) to 4-aminophenol (4-AP) was employed as a model reaction to test catalytic performance in this work. The results showed that the binary-metal-oxide-coated nanocatalyst (550 °C) exhibited significantly enhanced catalytic performance compared with the pure SiO₂ (550 °C) or TiO₂ (550 °C). In particular, the mSiO₂–TiO₂/Au/C/Fe₃O₄ particles calcined at 550 °C showed the highest catalytic activity, compared to the samples calcined at 700 °C, 300 °C and RT. Meanwhile, because of C layer burning, the sample presented a few white spots between the Fe₃O₄ microsphere and the oxide layer, suggesting that the specific surface area was increased by calcination. The sample (550 °C) still has a certain degree of magnetism, suggesting the desired samples could be separated by magnet. Finally, to explain the reduction of 4-nitrophenol (4-NP) to 4-aminophenol (4-AP), a possible reaction mechanism was also proposed.

---

1. Introduction

Nowadays, Au NPs, as a kind of high-activity noble metal catalyst, have been studied extensively for chemical reactions, such as oxidation of CO,¹ oxidation of alcohols,^2,3 hydrogenation of acetophenone⁴ and propane dehydrogenation.^5,6 As is well known, Au NPs are less prone to over-oxidation, self-poisoning, and metal leaching and thus have the highest selectivity. However, some limitations still exist. For example, during the pretreatment and catalytic process, Au nanoparticles tend to migrate and sinter because of moderate-to-high temperature.^7,8 With these limitations, fabricating core–shell nanostructures has been extensively known as an efficient and promising strategy to prevent aggregation of Au NPs. Recently, for preventing aggregating, various metallic oxides, such as TiO₂, CeO₂, ZnO and ZrO₂, have been widely used as supporting materials.^9–14 In heterogeneous catalysis, it has been known that a high-surface-area oxide support can enhance anti-sintering ability and reaction activity of the noble metal nanocatalyst. For example, Zhou et al. synthesized mSiO₂/Pt magnetic nanocatalysts with a TiO₂ or CeO₂ layer successfully. The mesoporous SiO₂ shell prevented the migration and aggregation of Pt NPs during calcination and improved the catalytic activity.¹⁵ Zaera et al. prepared a new Au@TiO₂ catalyst. The use of titania as the shell material proved critical to achieve much higher reactivity than the catalysts based on other oxides at low temperatures.¹⁰ To study the nature of the active gold sites on inert supports, it is important to maximize the number of the atomically dispersed gold sites and fully eliminate the formation of Au NPs.¹⁶ Yang et al. reported that the addition of alkali ions to gold on KLTL-zeolite and mesoporous MCM-41 silica could stabilize mononuclear gold in Au-O(OH)_x-(Na or K) ensembles.¹⁷ According to some previous research results of oxide package, we can easily find that oxide package can have an impact on the reactivity of Au nanocatalyst. Meanwhile, some researchers, to acquire multiple actions, have assembled Au NPs deposited on the inner of mixed oxide layer. Brioude¹⁸ synthesized an eccentric Au@(SiO₂, TiO₂) nanostructure and Zhou¹⁹ took the thermal stability and catalytic properties of this distinctive structure into consideration. The calcined Au@(SiO₂, TiO₂) particles showed the highest catalytic activity due to the easy mass transfer, improved thermal stability and increased synergy effect of Au with TiO₂. However, Brioude and Zhou et al. only studied the distribution, the thermal stability and catalytic properties of Au NPs with 10–50 nm diameters, and did not take Au NPs with the smaller diameter and oxide core into consideration. It is generally accepted that the catalytic activity of Au NPs decreases inevitably with the growth in the size of Au NPs. Furthermore, the smaller size of Au NPs may aggregate easily in the process of equipping mixed oxide layer, which makes their preparation difficult to a certain extent.¹⁹ Thus, in this work, we have prepared the novel mixed oxide/C/Au/MO_x nanoparticles successfully with the Au NPs (ca. 2–5 nm in diameter).

For saving resources, magnetically recoverable noble metal nanoparticles are promising catalysts for chemical reactions.²⁰ Kusumoto et al.²¹ fabricated a functionalized core–shell structure that the metallic Au-ultrathin layer was successfully functionalized on the magnetic nanocubes surface for the fabrication of the core–shell structure (Fe₃O₄@Au) by the borohydrate reduction of HAuCl₄ in water/poly-L-histidine solution. Cai et al.²² reported a facile approach for fabrication of Fe₃O₄@Au nanocomposite particles as a dual mode contrast agent for both magnetic resonance and computed tomography imaging applications. However they both did not test the magnetism of Fe₃O₄ about the magnetically recoverable performance and the chemical reaction activity of Au catalyst. So it is very necessary for using Fe₃O₄ to synthesize a magnetically recoverable noble metal catalyst to save resources.

Herein, based on above hypothesis, we report a novel nanocatalyst of spherical MO_x/C/Au hierarchical nanostructures with Au NPs embedded in the inner surfaces of mixed oxide (mSiO₂, TiO₂). The mixed oxide could serve as excellent active-sites and enhance the catalytic activity due to the strongly interacting with Au NPs. In addition, the synthesized spherical NPs could show a higher diffusion and catalytic rate because of the high aspect ratio using spherical magnetic Fe₃O₄ as support. Furthermore, Fe₃O₄ can be moved by the magnet and could be easily separated. With this goal in mind, in this work, we designed the configuration of Au/C/Fe₃O₄, then, covered with mSiO₂–TiO₂ shell. The C layer, on the one hand, could enhance the dispersity and deposited capacity of Au NPs on support, on the other hand, can increase the specific surface area of catalyst to improve the reactivity. The detailed structure and synthetic procedures are depicted in Fig. 1. In the first step, Fe₃O₄ NPs were prepared via a hydrothermal method. In the second step, well-dispersed Au NPs were deposited on the surface of the spherical NPs via electrostatic interactions and a layer of mSiO₂–TiO₂ was coated on the surface of the as-synthesized Au/C/Fe₃O₄ NPs. At last, the obtained sample was suffered to calcination to get the desired mSiO₂–TiO₂/Fe₃O₄–Au NPs. In this work, we also assembled mSiO₂/Fe₃O₄–Au NPs and TiO₂/Fe₃O₄–Au NPs compared with the desired mSiO₂–TiO₂/Fe₃O₄–Au NPs about the sintering resistance and reactivity. At the end of the experiment, it was shown that the obtained mSiO₂–TiO₂/Fe₃O₄–Au samples exhibited the strong magnetic.

Fig. 1 Schematic illustration for the preparation of a novel mSiO₂–TiO₂/Fe₃O₄–Au catalyst.

2. Experiment

2.1. Materials

Tetrabutyl titanate, iron(III) chloride hexahydrate (FeCl₃·6H₂O), ethylene glycol (EG) were purchased from Tianjin no. 1 Chemical Reagent Factory (Tianjin, China). HAuCl₄·3H₂O (≥99.9%), Ti(OEt)₂(acac)₂ (75 wt% in isopropanol), and 4-nitrophenol (≥99%) were purchased from Aldrich. Tetraethyl orthosilicate (TEOS), sodium citrate, isopropanol, ethanol, ammonia solution (28 wt%), polyvinylpyrrolidone (M.W. ∼ 50 000), and glucose were of analytical grade and all of them were purchased from Sinopharm Chemical Reagent Co., Ltd. Hydroxypropyl cellulose (HPC, M.W. ∼ 100 000) were obtained from Alfa Aesar. Deionized water was used in all experiment. All chemicals were used as received.

2.2. Synthesis

Synthesis of spherical Fe₃O₄ NPs. Fe₃O₄ magnetic microspheres were synthesized by a modified reduction reaction between FeCl₃·6H₂O and ethylene glycol (EG) in a solvothermal system.²³ Briefly, FeCl₃·6H₂O (2.70 g, 5 mmol) was first dissolved in EG (80 mL). After 40 minutes, anhydrous NaAc (7.20 g) and PEG₁₀₀₀₀ (2.00 g) were added to the above solution. The resulting solution was carefully transferred into a Teflon-lined stainless steel autoclave. Then the autoclave was sealed and heated at 200 °C for 10 h, the autoclave was naturally cooled to room temperature. Then the black products were collected with the help of a magnet, followed by washing with a cycle of deionized water and ethanol six times. The products were then dried under vacuum at 60 °C for 12 h.

Synthesis of Au/C/Fe₃O₄. The first step, C/Fe₃O₄ microspheres were synthesized by a solvothermal system. In short, Fe₃O₄ magnetic microspheres (0.2 g), cetyltrimethyl ammonium bromide (CTAB, 0.10 g) and glucose (2.00 g) were first dissolved in mixed solution (80 mL) (ethanol/deionized water = 1 : 1). After being stirred for 1 h, the resulting solution was carefully transferred into a Teflon-lined stainless steel autoclave, sealed and heated at 170 °C for 24 h. Then the black products were collected with the help of a magnet.

The second step, the as-obtained C/Fe₃O₄ magnetic microspheres were dispersed in deionized water (100 mL) at room temperature for 1 h. Then aniline (0.04 mL) was added into the solution and the product was maintained at 25 °C for 24 h after adding an ammonium peroxydisulfate aqueous solution (Wt = 20%). At the protection of trisodium citrate (TSC, 1.2 mL, 40 mg mL⁻¹), the HAuCl₄ (0.55 mL, 10 mg mL⁻¹) was injected into the as obtained solution at room temperature (RT) and continually stirred for 10 min. Then fresh sodium borohydride (9 mL) was immediately injected to the solution under vigorous stirring. The Au/C/Fe₃O₄ colloids with the Au NPs of 2–5 nm were obtained after stirring at RT for 2 h. The product was collected by drying at 25 °C for 12 h.²⁴

The coating of the mixed oxide (mSiO₂, TiO₂). The as-prepared 0.1 g of Au/C/Fe₃O₄ cores was mixed with 50 mL of isopropanol and 1.0 mL of ammonia solution under vigorous stirring. Then, 12 mL of a mixed solution (0.05 mL TEOS + 0.10 mL Ti(OEt)₂(acac)₂) in isopropanol was added drop by drop. The resulting mixture was then vigorously stirred at room temperature for additional 12 h. The mSiO₂–TiO₂/Au/C/Fe₃O₄ composite particles were collected by a magnet and washed with ethanol.

Synthesis of mSiO₂/Au/C/Fe₃O₄ and TiO₂/Au/C/Fe₃O₄ micro-spheres. The procedures for preparation of mSiO₂/Au/C/Fe₃O₄ microspheres were similar to the preparation process of synthesizing mSiO₂–TiO₂/Au/C/Fe₃O₄ microspheres, but except for increasing the amount of TEOS to 0.10 mL and no addition of titanium butoxide (TBOT). The amount of titanium butoxide was added about 8 mL and no TEOS was involved in the preparation of TiO₂/Au/C/Fe₃O₄ microspheres.

2.3. Characterization

Transmission electron microscopy (TEM) experiments were conducted on a JEM-1230 microscope operated at 100 kV. The samples for the TEM measurements were suspended in ethanol and supported onto a Cu grid. Scanning electron microscope (SEM) was performed on a Hitachi S-3400N scanning electron microscope and energy dispersive X-ray spectroscopy (EDX) analysis were conducted on a JEM-1230 microscope operated at 100 kV. The powder X-ray diffraction (XRD) patterns were recorded on a Bruker D8 Advance Diffractometer (Germany) with Cu Ka radiation (γ = 1.5406 Å). FT-IR spectra were measured by a Bruker Fourier spectrometer (Tensor27, German) and thermal gravity analysis curve was measured by a SDT Q600 Simultaneous TGA (TA Instruments-Waters LLC, USA). UV-vis spectra were recorded on a Shimadzu UV 3600 spectrometer.

2.4. Catalytic reduction of p-nitrophenol

To study catalytic properties of the mSiO₂–TiO₂/Au/C/Fe₃O₄ nanoparticles, reduction of 4-NP by NaBH₄ was chosen as a model reaction. In this experiment, the activities of as-prepared mSiO₂–TiO₂/Fe₃O₄–Au nanoparticles were tested by comparison with the mSiO₂/Fe₃O₄–Au and TiO₂/Fe₃O₄–Au samples. Note that 10 mg of the catalyst was dispersed in 10 mL deionized water, and then NaBH₄ aqueous solution (10 mL, 1.2 M) and 4-NP aqueous solution (10 mL, 3.4 mM) were added. The reaction solution was stirred at room temperature, and a UV-vis spectrometer was used to monitor the progress of the reaction at regular intervals. Finally, the activities of the mSiO₂–TiO₂/Fe₃O₄–Au nanoparticles calcined at different temperature were tested by using the same test method.

3. Results and discussion

3.1. Characterization of mSiO₂–TiO₂/Fe₃O₄–Au nanocomposites

The work aims to design and synthesize Au nanocatalyst based on C/Fe₃O₄, then, encapsulated by SiO₂ and TiO₂ mixed oxides shells. The morphology and structure of C/Fe₃O₄ microspheres were first investigated by SEM. As observed in Fig. S1,† after most of Fe₃O₄ spheres were coated by C, Fe₃O₄ still remains spherical and highly disperse, showing the synthesis of C/Fe₃O₄. The main experimental procedures and the observations are summarized in Fig. 2. Fig. 2a shows that the Fe₃O₄ are mostly spherical with overall diameters about 500 nm. The small inset at the top right of Fig. 2a represents a single Fe₃O₄ sphere with the resolution boosted, clearly showing the surfaces of Fe₃O₄ nanospheres are rough, because of the existence of hydroxyl and carbanyl groups on the edge of Fe₃O₄ compound. The TEM image (Fig. 2b) clearly shows that the C layer (ca. 30 nm in thickness) is uniformly layered on the Fe₃O₄ nanospheres. As shown in Fig. 2c, the Au nanoparticles (ca. 2–5 nm in the diametre) are highly dispersed on the surface of C layer.

Fig. 2 TEM images of (a) Fe₃O₄, (b) C/Fe₃O₄, (c) Au/C/Fe₃O₄, (d) TiO₂/Au/C/Fe₃O₄, (e) mSiO₂/Au/C/Fe₃O₄, (f) mSiO₂–TiO₂/Au/C/Fe₃O₄. The insets at the top represent the higher magnification of the corresponding particle's edge. The insets at the bottom are the size distribution histograms of Fe₃O₄ microspheres, the thickness of carbon, gold nanoparticles and the thickness of TiO₂, SiO₂, and mSiO₂–TiO₂ layer, respectively.

Fig. 2d–f displays the TEM images of TiO₂/Au/C/Fe₃O₄ (d), mSiO₂/Au/C/Fe₃O₄ (e) and mSiO₂–TiO₂/Au/C/Fe₃O₄ (f) magnetic microspheres, and the insets at the top of Fig. 2d–f are the higher magnification images. In Fig. 2d, the changes in size and morphology are apparent compared with Au/C/Fe₃O₄. It can be inferred that the thickness of TiO₂ layer reaches nearly 20 nm. The TiO₂ layer consists of a large number of titania nanoparticles, observed in the inset of Fig. 2d.²⁵ In addition, as shown in Fig. 2e, a uniform mesoporous silica layer was deposited onto the surface of Au/C/Fe₃O₄ magnetic microspheres about 20 nm in thickness. The TEM image of mSiO₂–TiO₂/Au/C/Fe₃O₄ is displayed in Fig. 2f. The diameter reaches about 565 nm with a cladding layer (ca. 35 nm in thickness), compared with the C/Fe₃O₄ microspheres. The insets at the bottom are the size distribution histograms of Fe₃O₄ microspheres, the thickness of carbon, gold nanoparticles and the thickness of TiO₂, SiO₂, and mSiO₂–TiO₂ layer, respectively. The size distribution histograms present the size and thickness of microspheres and layers clearly.

For further discussing deeply the load of Au on C/Fe₃O₄, as shown in Fig. 3a–c, the high-resolution TEM image (HRTEM) (Fig. 3a and the inset) represents a lattice fringes measured with a spacing of 0.235 nm and the SAED pattern (Fig. 3b) describes a dim ring corresponding to the (111) plane of face centered cubic golden, which demonstrates the load of Au NPs. Besides, since Fe₃O₄ particles have a grey black appearance similar to Au NPs, the loaded Au NPs can't be distinguished apparently, in spite of the existence of C layer. As shown in Fig. 3c, the small bright spots represent Au nanoparticles and the others are Fe₃O₄ nanospheres. The visible appearance suggests that the load of Au is successful. To confirm that the TiO₂/Au/C/Fe₃O₄, mSiO₂/Au/C/Fe₃O₄ and mSiO₂–TiO₂/Au/C/Fe₃O₄ microspheres were prepared successfully, EDX spectroscopy was applied to analyze the microspheres obtained. After the direct coating procedures, the EDX analysis of TiO₂/Au/C/Fe₃O₄ microspheres (Fig. 4a) shows the existence of Fe, Ti, Au, C and O elements, which further manifests the coating of titania onto the C/Fe₃O₄ microspheres. The EDX analysis of the obtained mSiO₂/Au/C/Fe₃O₄ microspheres (Fig. 4b) reveals the presence of Fe, Si, Au, C and O elements, indicating the formation of silica on the surface of the C/Fe₃O₄ microspheres. Fig. 4c verifies the existence of Fe, Si, Ti, Au, C and O elements, further confirming the successful synthesis of the mSiO₂–TiO₂/Au/C/Fe₃O₄ microspheres. In addition, elemental Cu which exists in all the EDX spectra is derived from the copper mesh. Meanwhile, the chemical composition of the surfaces of the synthesized microspheres was investigated by EDX spectroscopy analysis (Fig. S2†).

Fig. 3 (a) A high resolution TEM image of the loaded Au NPs. (b, c) The SAED pattern obtained from Au/C/Fe₃O₄ sample.

Fig. 4 The EDX spectrum data for the obtained (a) TiO₂/Au/C/Fe₃O₄, (b) mSiO₂/Au/C/Fe₃O₄ and (c) mSiO₂–TiO₂/Au/C/Fe₃O₄ magnetic microspheres.

From the pattern of the mean value of atomic% (Fig. S2†) in the TiO₂/Au/C/Fe₃O₄ and mSiO₂/Au/C/Fe₃O₄ microspheres, the titanium content and the silica content can be observed to be 2.8% and 13.60% in mean atomic value, respectively. The total titanium and silica content in the mSiO₂–TiO₂/Au/C/Fe₃O₄ microspheres was measured to be 12.7%. It is worth mentioning that the weight ratio of TEOS and Ti(OEt)₂(acac)₂ used in the synthesis process is about 1 : 2, but, the atomic ratio of Si/Ti in the mSiO₂–TiO₂/Au/C/Fe₃O₄ microspheres is about 2 : 1. The reason is that the hydrolysis of Ti precursor compared with TEOS is preferable at the early stage of hydrolysis, which leads to a bigger content of TiO₂ on the Au/C/Fe₃O₄ surface and a smaller content in core–shell surface. What is important is that the enriched Ti species on the Au NPs surface will be advantageous to increase the activity of Au catalyst for enhanced synergy interaction between Au and TiO₂.^19,26,27

An FTIR instrument is also employed to characterize the prepared microspheres. Fig. S3a–c† shows a characteristic band for Fe–O stretching vibration at 576 cm⁻¹, suggesting the existence of Fe₃O₄ microspheres. Fig. S3c† shows characteristic absorption peaks at 500–750 cm⁻¹ of titania, further confirming the successful preparation of TiO₂/Au/C/Fe₃O₄ microspheres. Fig. S3a† demonstrates not only the characteristic band for Fe–O stretching vibrations at 576 cm⁻¹, but also a new absorption peak corresponding to the absorption of silica at 463 cm⁻¹, proving the successfully synthesis of mSiO₂/Au/C/Fe₃O₄ microspheres. Fig. S3b† displays the broad absorption peaks which appeared at 450–700 cm⁻¹ can be assigned to Ti–O, Si–O and Fe–O stretching vibrations, providing additional evidence of the successful formation of mSiO₂–TiO₂/Au/C/Fe₃O₄ microspheres.

The X-ray diffraction (XRD) patterns of the synthesized microspheres are shown in Fig. 5. Fig. 5A shows the XRD patterns of the Fe₃O₄ microspheres, and the strong diffraction peaks at 2θ = 30.16, 35.50, 43.22, 57.11 and 62.72 can be indexed to the (220), (311), (400), (511) and (440) reflections, respectively, which clearly verifies the typical cubic structure of Fe₃O₄ (JCPDS no. 19-629).²⁸ The detected peaks at 38.3°, 43.2° are indexed to (111) and (200) planes of Au diffraction (JCPDS file 89-3697, face-centered).²⁹ There are no obvious XRD peaks corresponding to carbon material, indicating that the carbon coatings are amorphous.³⁰ However, the presence of carbon has been proved by the above EDX analysis in our work. It is easy to find that gold mainly exists in Au(111) forms based on the XRD diffraction height of Au(111)and Au(220), corresponding to the above SAED pattern in Fig. 5C. The wide-angle XRD pattern of the mSiO₂/Au/C/Fe₃O₄ sample indicates the prepared SiO₂ is amorphous.²⁸ In contrast, Fig. 5F suggests that TiO₂ coating onto the mSiO₂–TiO₂/Au/C/Fe₃O₄ microspheres is also amorphous. As shown in Fig. 5G, the (101), (004), (105) and (211) reflections of anatase phase TiO₂ (JCPDS no. 21-1272) can be observed at 2θ = 25.28, 37.86, 48.09 and 55.15. This result confirmed the existence of titanium dioxide on the surface of the mSiO₂–TiO₂/Au/C/Fe₃O₄ microspheres, meanwhile, also demonstrating that the TiO₂ layer heated appears crystal structure transformation on the surface of the mSiO₂–TiO₂/Au/C/Fe₃O₄ microspheres.²⁷ The peak at 33.2° corresponding to the (104) crystal planes of a-Fe₂O₃ confirms that a small portion of Fe₃O₄ transfers into the α-Fe₂O₃.¹⁵ According to XRD pattern and the above work, it is clear to demonstrate that mSiO₂–TiO₂/Au/C/Fe₃O₄ microspheres have been prepared successfully.

Fig. 5 X-ray diffraction patterns of the as-synthesized (A) Fe₃O₄, (B) C/Fe₃O₄, (C) Au/C/Fe₃O₄, (D) mSiO₂/Au/C/Fe₃O₄, (E) TiO₂/Au/C/Fe₃O₄, (F) mSiO₂–TiO₂/Au/C/Fe₃O₄ and (G) mSiO₂–TiO₂/Fe₃O₄–Au (mSiO₂–TiO₂/Au/C/Fe₃O₄ particles calcined at 550 °C) microspheres. The diffraction peaks of Fe₃O₄, TiO₂ and Au NPs are marked with F, T and A, respectively.

Investigating the dosage of ammonia is a very meaningful thing, for the accuracy control of the coating of the mixed oxide. As shown in Fig. 6a, the fact is that mixed oxide package grows around Au NPs rather than growing only on one side. Au NPs stuck in the middle of mixed oxide, like a hot dog burger, can get more favorable interaction with TiO₂, promoting the catalytic performance of Au NPs, as shown in the inset in Fig. 6a. Meanwhile, the SiO₂ component of mixed oxide enhances the thermal stability of the core–shell Au NPs catalyst, preferably. Besides, in the process of coating, the addition of ammonia leads to the ionic strength increase of the solution due to the dissociation of ammonia, which leads to the aggregation of gold nanoparticles.²⁶ By controlling experimental conditions such as the coating time, the concentration of gold nanoparticles and alkoxides, the structure and thickness of the coating can be varied in a large range.³¹ As shown in Fig. 6a–c, with the increase of dosage of ammonia, the structure of coating can be changed in appearance and quality. It's not hard to see the structure of coating shown in Fig. 6c is more compact than that shown in Fig. 6b. Meanwhile, the contents of Si and Ti species are also higher, which is good for catalytic activity and thermal ability. Finally, we find the optimum dose of ammonium for coating is 1.0 mL. Hence, the right usage of ammonia is the key to synthesizing the desired sample, simultaneously taking the aggregation of gold nanoparticles and the structure of the coating into account.

Fig. 6 TEM images of mSiO₂–TiO₂/Au/C/Fe₃O₄ microspheres prepared by adding the different dosage of ammonia, (a) 0.55 mL, (b) 1 mL, (c) 2 mL ammonia solution.

The thermal stability of the core–shell particles was tested in this study. Fig. 7 displays the TEM images of the samples of mSiO₂/Au/C/Fe₃O₄, TiO₂/Au/C/Fe₃O₄ and mSiO₂–TiO₂/Au/C/Fe₃O₄ calcined at 550 °C for 4 h, respectively. It is clear that the (Si, Ti)-550 remains virtually unchanged upon calcination, not only on coating structure, but also on the Au diameter size, demonstrating that the mSiO₂–TiO₂/Au/C/Fe₃O₄ is quite thermally stable to prevent Au NPs from sintering. However, the Ti-550 exhibits a significant structure collapse and severe Au aggregation. Yin^32,33 has shown that calcination of TiO₂ shell may lead to an extensive crystallization and grain growth of TiO₂, thus, a significant structural rearrangement of the shell is inevitable. The extensive structure reorganization of the shell may lead to the encaged Au core escaping from TiO₂ shell and ultimately, resulting in coalescence and sintering to about 15 nm on the Ti-550 sample. In contrast, the mixed oxide shell acts as a barrier preventing the crystal growth of TiO₂ during calcination.^33–35 In this work, the Au NPs size of mSiO₂–TiO₂/Fe₃O₄–Au changed from 3.5 to 4.5 nm after calcination at 550 °C (Fig. S4a†). Nevertheless, the Au NPs of traditional “naked” Au/C/Fe₃O₄ sample changed to 20 nm in diameter (Fig. S6†). These experimental evidences have intuitively demonstrated the thermal stability of the oxide shells. Meanwhile, C layer between Fe₃O₄ microspheres and mixed oxide layer will be removed during calcination. After the C layer was removed, in the original C layer position would leave gaps, increasing the specific surface area, and enlarging the contact space between Au nanoparticles and reactants. Not difficult to find from Fig. 7c, the Si-550 also presents a gap between Fe₃O₄ microsphere and SiO₂ layer, suggesting mixed oxide represents the same properties to the pure SiO₂ shell.³⁶ The TGA analysis of the synthesized mSiO₂–TiO₂/Au/C/Fe₃O₄ composites is shown in Fig. 7d. As shown, the weight loss of composites from room temperature to 250 °C should owe to the evaporation of the absorbed water and solvent. With an increase of temperature, a drastic weight loss which may be attributed to the oxidations of carbon to CO₂/CO appears between 250 °C and 370 °C, describing the weight loss of mSiO₂–TiO₂/Au/C/Fe₃O₄ composites is 6%. It is easy to find that the thermogravimetric slope becomes small between 350 °C and 500 °C, because the oxidation of Fe₃O₄ to Fe₂O₃, which occurs when the temperature is higher than 350 °C, results in weight gain.³⁷ After the roasting temperature reaches 750 °C, the quality of composites keeps constant.

Fig. 7 TEM images of (a) mSiO₂/Fe₃O₄–Au (Si-550), (b) TiO₂/Fe₃O₄–Au (Ti-550) and (c) mSiO₂–TiO₂/Fe₃O₄–Au ((Si, Ti)-550) calcined at 550 °C. (d) The TGA images of mSiO₂–TiO₂/Au/C/Fe₃O₄ calcined from 100 °C to 800 °C.

3.2. Catalytic reduction of 4-nitrophenol

Employing the reduction of 4-nitrophenol (4-NP) into 4-aminophenol (4-AP) as a model reaction has been testified to be useful for the analysis of the catalytic activity of noble metal nanocatalyst.^35,38,39 In the reaction, an excess amount of NaBH₄ was added to avoid the influence of the concentration of NaBH₄ on the reduction rate, and the reduction process could be considered as a first-order reaction. As the NaBH₄ was added, the absorption maximum of 4-NP at 317 nm shifts to 400 nm due to the formation of 4-nitro-phenolate ion.⁴⁰ Besides, because of the inertia of 4-NP to NaBH₄, the reduction reaction would not proceed without catalyst. For reducing errors, the dosage of every reagent kept constant in all catalytic runs.

Fig. 8b shows the linear relationships between ln(C_o/C_t) and reaction time in the reaction catalyzed by different samples (C_t is the ordinate values of the absorption peak at 400 nm), and the plots well match the first-order reaction kinetics. Considering the concentration of NaBH₄ is much higher than 4-NP, the reaction rate can be assumed to be independent of the concentration of the reductant. Thus, a pseudo-first-order rate kinetics can be used to evaluate the catalytic rate with regard to 4-NP.¹⁹ The rate constant k_app is calculated from the slope of the linear relationships between ln(C_o/C_t) and reaction time. The apparent rate constants (k_app) of the as-prepared (Si, Ti)-700 sample, (Si, Ti)-300 and (Si, Ti)-RT sample are 0.1368 min⁻¹, 0.4039 min⁻¹ and 0.3177 min⁻¹, respectively. However, in sharp contrast, the sample of (Si, Ti)-550 shows the highest catalytic activity (0.6040 min⁻¹) in this work which is roughly 4.42 times higher than that of the (Si, Ti)-700 sample. The fact reveals that calcination temperature plays a key role in the reaction rate. From Fig. S4b,† the used (Si, Ti)-550 NPs keep its morphology similar to the sample not used (Fig. 7c and S4a†), revealing that the repeated use has little effect on the structure of the (Si, Ti)-550 nanocatalyst. As can been seen from Fig. 8c and S5,† due to the high calcination temperature (700 °C), the Au NPs, breaking away from the interlayer of mSiO₂–TiO₂/Fe₃O₄, begin aggregating and growing, proving why low catalytic activity of (Si, Ti)-700 NPs.⁴¹

Fig. 8 (a) Successive UV-vis absorption spectra of the reduction of 4-NP by NaBH₄ in the presence of (Si, Ti)-550 catalyst. (b) Plots of ln(C_t/C_o) of 4-NP against time; (1) (Si, Ti)-550, (2) (Si, Ti)-300, (3) (Si, Ti)-RT, (4) Si-550, (5) Ti-550 and (6) (Si, Ti)-700. (c) TEM image of (Si, Ti)-700 catalyst. (d) The image of magnetic separation.

It is well known that k_app values are in direct proportion to the content of Au NPs. Therefore, it is necessary to eliminate the influence of Au content. As can be seen from Fig. S2,† the contents of Au in Si-550, Ti-550 and (Si, Ti)-550 are 1.3%, 0.6% and 1.1% respectively. As a result, the k_app of Si-550 is 0.28814 min⁻¹ almost equal to that of Ti-550 (0.22907 min⁻¹). Seemingly, the Si-550 has a higher activity than Ti-550. Fact, Au contents have an effect on the k_app of catalysts. The content of Au in Si-550 was twice in Ti-550, so the Ti-550 has a higher activity than Si-550. Consequently, under the condition of considering the gold content, TiO₂ layer shows excellent co-catalysis effect compared to SiO₂ layer. Meanwhile, the k_app of (Si, Ti)-550 is 0.60398 min⁻¹ almost 2.6 times higher than that of Ti-550 (0.22907 min⁻¹), revealing (Si, Ti)-550 have a higher activity than Ti-550, due to the help of SiO₂ constituents.¹⁵ Consequently, mSiO₂–TiO₂ layer shows excellent co-catalysis effect compared to TiO₂ or SiO₂ layer when the catalyst used in relatively high temperature (for example, 550 °C). Finally, by comparison, Fig. 8d demonstrates that after used five times, the mSiO₂–TiO₂/Fe₃O₄–Au (550 °C) nanospheres still keep excellent magnetic.

A possible catalytic mechanism, to more clearly explain the experimental results, is illustrated vividly in Fig. 9. Generally, when a metal and semiconductor are placed in contact, Fermi level alignment might occur, which results in the charge redistribution and the formation of a depletion layer surrounding the metal.³⁶ As Au (5.1 eV) has a higher work function than TiO₂ (3.2 eV), electrons leave from the TiO₂ into the near Au particles, which results in an electron-enriched region.^19,42,43 Thus, BH₄⁻¹ may be easily absorbed on the surface of TiO₂, where it is oxidized to BO₂⁻¹ by giving away the electron. Furthermore, the enhanced catalytic activity of Au NPs should have close relation to the improvement of the crystallinity of TiO₂, corresponding to the higher reaction activity at appropriately high temperature (550 °C). Meanwhile, we speculate that unreacted BH₄⁻¹ into the reserved gap reacts once, improving the reaction speed and selectivity. Indeed, Cai⁴⁴ has recently reported that the interface of the deposited nanoparticles has a critical role in promoting the activation of reactants by the DFT analysis and calculation; thus, we suggest that the synergistic effect of TiO₂ layer, Au NPs and gaps enhances the reactivity.

Fig. 9 Speculated mechanism of the catalytic reduction of 4-NP with the mSiO₂–TiO₂/Fe₃O₄–Au catalyst.

4. Conclusion

A novel type of binary-metal-oxide-coated nanocomposite catalyst has been synthesized successfully, verified by the above datum. The hybrids have a mixed oxide layer and a moveable magnetic Fe₃O₄ core. Besides, the outer layer of mSiO₂–TiO₂ could increase the thermal stability and reactivity of the nanocatalyst. The increased k_app of (Si, Ti)-550 compared to Ti-550 demonstrated the high thermal stability due to the existence of SiO₂. The increased k_app of (Si, Ti)-550 compared to Si-550 demonstrates the high reactivity due to the interaction between Au and TiO₂. Furthermore, the excellent performance of (Si, Ti)-550, compared to (Si, Ti)-700, (Si, Ti)-300 and (Si, Ti)-RT, reveals that the mixed oxide layer is suit to high temperature but below 700 °C reaction. In addition, the mSiO₂–TiO₂/Fe₃O₄–Au sample presents a few white spots between Fe₃O₄ microsphere and mixed oxide layer, suggesting specific surface area is increased by calcination and obtaining the expected results of the experiment. Finally, the sample of (Si, Ti)-550 kept high performance and excellent magnetic even after five cycles of reusing. Future work will research a novel preparation method of Au NPs.

Acknowledgements

The authors are grateful for the financial supports of the National Natural Science Foundation of China (Grant No. 21376051, 21106017, 21306023), the Natural Science Foundation of Jiangsu Province of China (Grant No. BK20131288), the Fund Project for Transformation of Scientific and Technological Achievements of Jiangsu Province of China (Grant No. BA2014100), the Fundamental Research Funds for the Central Universities (No. 3207045421), Instrumental Analysis Fund of Southeast University and the Priority Academic Program Development of Jiangsu Higher Education Institutions.

References

1. S. Zhang, X. S. Li, B. Zhu, J. L. Liu, X. B. Zhu, A. M. Zhu and B. W. L. Jang, Catal. Today, 2015, 256, 142–147.
2. V. I. Sobolev and K. Y. Koltunov, Kinet. Catal., 2015, 56, 343–346.
3. A. A. Rodriguez, C. T. Williams and J. R. Monnier, Catal. Lett., 2015, 145, 750–756.
4. M. Freifelder, T. Anderson, Y. H. Ng and V. Papendic, J. Pharm. Sci., 1964, 53, 967.
5. W. Sangthong and J. Limtrakul, Abstr. Paper Am. Chem. Soc. Natl. Meet., 2011, 242, 111.
6. H. Dyrbeck and E. A. Blekkan, Stud. Surf. Sci. Catal., 2007, 167, 331–336.
7. R. X. Jin, Y. Yang, Y. F. Li, L. Fang, Y. Xing and S. Y. Song, Chem. Commun., 2014, 50, 5447–5450.
8. R. Guttel, M. Paul, C. Galeano and F. Schuth, J. Catal., 2012, 289, 100–104.
9. L. Wang, H. Ji, S. Wang, L. Kong, X. Jiang and G. Yang, Nanoscale, 2013, 5, 3793–3799.
10. I. Lee, J. B. Joo, Y. Yin and F. Zaera, Angew. Chem., Int. Ed., 2011, 50, 10208–10211.
11. Y. Xu, Y. Zhang, Y. Zhou, Z. Zhang, S. Xiang, X. Sheng, Q. Wang and C. Zhang, RSC Adv., 2015, 5, 34549–34556.
12. R. Liu, D. Jiang, Q. Zhang, R. Li, Q. Duan, J. Qin, X. Shen, J. Hou, J. Zhao, Q. Liang and S. Gao, J. Phys. D: Appl. Phys., 2015, 48, 275104–275111.
13. W. Cui, H. Zhu, M. Jia, W. Ao, Y. Zhang and B. Zhaorigetu, React. Kinet., Mech. Catal., 2013, 109, 551–562.
14. Y. Zhang, S. Xiang, Y. Zhou, Y. Xu, Z. Zhang, X. Sheng, Q. Wang and C. Zhang, RSC Adv., 2015, 5, 48187–48193.
15. C. Zhang, Y. Zhou, Y. Zhang, Q. Wang and Y. Xu, RSC Adv., 2015, 5, 12472–12479.
16. G. M. Veith, A. R. Lupini, S. J. Pennycook and N. J. Dudney, ChemCatChem, 2010, 2, 281–286.
17. Y. Ming, L. Sha, W. Yuan, A. H. Jeffrey, X. Ye, F. A. Lawrence, L. Sungsik, H. Jun, M. Manos and F.-S. Maria, Science, 2014, 346, 1498–1501.
18. D. D. Lekeufack, A. Brioude, A. Mouti, J. G. Alauzun, P. Stadelmann, A. W. Coleman and P. Miele, Chem. Commun., 2010, 46, 4544–4546.
19. Z. Zhang, Y. Zhou, Y. Zhang, X. Sheng, S. Zhou and S. Xiang, RSC Adv., 2014, 4, 40078–40084.
20. C. Zhang, Y. Zhou, Y. Zhang, Z. Zhang, Y. Xu and Q. Wang, RSC Adv., 2015, 5, 64951–64960.
21. M. Abdulla-Al-Mamun, Y. Kusumoto, T. Zannat, Y. Horie and H. Manaka, RSC Adv., 2013, 3, 7816–7827.
22. H. Cai, K. Li, M. Shen, S. Wen, Y. Luo, C. Peng, G. Zhang and X. Shi, J. Mater. Chem., 2012, 22, 15110–15120.
23. H. Deng, X. L. Li, Q. Peng, X. Wang, J. P. Chen and Y. D. Li, Angew. Chem., Int. Ed., 2005, 44, 2782–2785.
24. S. Xiang, Y. Zhou, Y. Zhang, Z. Zhang, X. Sheng, S. Zhou and Z. Yang, Dalton Trans., 2014, 43, 11039–11047.
25. J. B. Joo, Q. Zhang, I. Lee, M. Dahl, F. Zaera and Y. Yin, Adv. Funct. Mater., 2012, 22, 166–174.
26. D. D. Lekeufack, A. Brioude, A. Moui, J. G. Alauzun, P. Stadelmann, A. W. Coleman and P. Miele, Chem. Commun., 2010, 46, 4544–4546.
27. W. Li, Q. Deng, G. Fang, Y. Chen, J. Zhan and S. Wang, J. Mater. Chem. B, 2013, 1, 1947–1961.
28. Q. Yuan, N. Li, W. Geng, Y. Chi and X. Li, Mater. Res. Bull., 2012, 47, 2396–2402.
29. Y. Xu, Y. Zhang, Y. Zhou, S. Xiang, Q. Wang, C. Zhang and X. Sheng, RSC Adv., 2015, 5, 58237–58245.
30. Y. M. Zhou, H. B. Wang, M. Gong, Z. Y. Sun, K. Cheng, X. K. Kong, Z. Guo and Q. W. Chen, Dalton Trans., 2013, 42, 9906–9913.
31. K. L. Kelly, E. Coronado, L. L. Zhao and G. C. Schatz, J. Phys. Chem. B, 2003, 107, 668–677.
32. J. B. Joo, Q. Zhang, I. Lee, M. Dahl, F. Zaera and Y. D. Yin, Adv. Funct. Mater., 2012, 22, 166–174.
33. J. B. Joo, Q. Zhang, M. Dahl, I. Lee, J. Goebl, F. Zaera and Y. D. Yin, Energy Environ. Sci., 2012, 5, 6321–6327.
34. Z. Zhang, Y. Zhou, Y. Zhang, S. Zhou, J. Shi, J. Kong and S. Zhang, Dalton Trans., 2013, 42, 5004–5012.
35. Y. Zhang, Y. Zhou, Z. Zhang, S. Xiang, X. Sheng, S. Zhou and F. Wang, Dalton Trans., 2014, 43, 1360–1367.
36. C. Zhang, Y. Zhou, Y. Zhang, Z. Zhang, Y. Xu and Q. Wang, Adv. Powder Technol., 2015, 284, 387–395.
37. X. Zhang, Z. Hu, X. Xiao, L. Sun, S. Han, D. Chen and X. Liu, New J. Chem., 2015, 39, 3435–3443.
38. S. Xiang, Y. Zhou, Y. Zhang, Z. Zhang, X. Sheng, S. Zhou and Z. Yang, Dalton Trans., 2014, 43, 11039–11047.
39. X. Chen and S. S. Mao, Chem. Rev., 2007, 107, 2891–2959.
40. Q. R. Geng and J. Z. Du, RSC Adv., 2014, 4, 16425–16428.
41. X. Wang, D. P. Liu, S. Y. Song, L. Zeng and Y. Zhang, Dalton Trans., 2012, 41, 7193–7195.
42. R. L. Park, Science, 1988, 241, 1839.
43. J. M. Song, S. Y. Chen, Y. L. Shen, C. H. Tsai, S. W. Feng, H. T. Tung and I. G. Chen, Appl. Surf. Sci., 2013, 285, 450–457.
44. Y. Q. Cai, Z. Q. Bai, S. Chintalapati, Q. F. Zeng and Y. P. Feng, J. Chem. Phys., 2013, 138, 154711.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra27136c

---