Mesoporous RuO₂/TiO₂ composites prepared by cyclodextrin-assisted colloidal self-assembly: towards efficient catalysts for the hydrogenation of methyl oleate

Abstract

Mesoporous RuO₂/TiO₂ composites were prepared using a template-directed colloidal self-assembly approach combined with a cyclodextrin (CD)-assisted aqueous impregnation method. The supramolecular assemblies formed between the randomly methylated β-cyclodextrin (RaMeβ-CD) and the block copolymer P123 acted as a template for the formation of a highly porous TiO₂ network over which uniform dispersion of ruthenium nanoparticles was achieved. By combining dynamic light scattering, X-ray diffraction, N₂-adsorption, temperature-programmed reduction, field-emission scanning electron microscopy and high-resolution transmission electron microscopy, we show that CD-based assemblies provide a versatile and easily accessible toolbox with different functionalities for generating metal-supported catalysts with controlled pore architecture and uniform metal distribution. The performance of these supported catalysts was evaluated in the liquid phase hydrogenation of methyl oleate (MO, C18:1) to methyl stearate (MS, C18:0). Control of ruthenium dispersion into the large pores of RaMeβ-CD-P123-templated TiO₂ material enhanced catalyst activity and selectivity for the hydrogenation of the internal C=C bond and permitted catalyst separation and reuse without loss of activity. Our findings highlight the pivotal role played by the CD-based assemblies on the performance of supported ruthenium catalysts.

---

1. Introduction

Creating self-assembled nanostructured materials with well-defined architectures and enhanced catalytic performance is of central importance for applications in the field of heterogeneous catalysis.^1–6 Template-directed colloidal self-assembly, based on the spontaneous organisation of colloidal nanoparticles around an organic template by non-covalent interactions, provides a simple and versatile strategy for building robust materials with hierarchical order at multiple length scales.^7–12 Metal nanoparticles can be uniformly dispersed within these materials providing nanostructured composites with tunable properties. The importance of metal–oxide interfaces has long been recognised and different possibilities have been proposed for improving the catalyst performance including the porosity of the support, the size and distribution of metal nanoparticles, the metal–support interactions,^13–17 as well as the surface characteristics of nanoparticles through chemical modification.¹⁸ For this purpose, the design of nanostructured support materials with high specific surface area and tailored porosity is of crucial importance for enhancing the catalyst effectiveness.^19,20

Recently, our group has explored the possibility of using cyclodextrin (CD)-based supramolecular assemblies as template or structure directing agents for the elaboration of inorganic materials with a large variety of morphologies, porosities and crystal phase compositions.^21–25 A particular feature of CDs is their ability to form, in association with polymers, supramolecular assemblies with many different kinds of architectures such as for instance, spherical or ellipsoidal micelles, vesicles, elongated polypseudorotaxanes and three-dimensional hydrogels.^26–29 Interestingly, silica oligomers²³ or colloidal nanoparticles prepared by sol–gel process^21,22,25 may easily infiltrate within these assemblies and replicate their original structure. Moreover, CDs have also the unusual property to form adducts with a large number of metal salts enabling their uniform dispersion over solid supports.^30–35 We aim to take advantage of the experience that our group has acquired on the elaboration of nanostructured titanium dioxide (TiO₂) in order to use this material as support for uniform dispersion of ruthenium nanoparticles. Indeed, besides its interest as efficient photocatalyst for fundamentals and industrial applications,^36–39 titania has also proved to be an important support in heterogeneous catalysis due to its high surface area, chemical and thermal stability, as well as ability to strongly interact with various metal nanoparticles.^40–44 Moreover, ruthenium has been identified as a promising catalyst for applications in the hydrogenation of biomass-derived compounds such as levulinic acid,⁴⁵ furanic molecules,^46,47 arabinoic acid,⁴⁸ xylose,⁴⁴ and fatty esters^49–51 under mild conditions.

Herein, the TiO₂ support is synthesized using the recently developed RaMeβ-CD-P123 directed colloidal self-assembly approach^24,25 while the native β-CD is employed as dispersing agent to improve dispersion of metal nanoparticles over the solid support. Since the efficiency of the final catalyst depends on various factors including the characteristics of the support,⁵² the size and composition of individual metal nanoparticles^53,54 as well as the metal–support interactions,^1,13,14 control of the preparation process is essential. For this purpose, we have undertaken a detailed investigation of the effect of both the supramolecular template (RaMeβ-CD-P123 assemblies) and the dispersing agent (native β-CD) on the characteristics of the heterogeneous Ru/TiO₂ catalysts. Finally, the efficiency of these latter was evaluated in the liquid phase hydrogenation of methyl oleate (MO, C18:1) to methyl stearate (MS, C18:0) under mild conditions (50 °C, 40 bar H₂).^55–57

2. Experimental

2.1 Chemicals

Randomly methylated β-cyclodextrin (denoted RaMeβ-CD, average degree of molar substitution (DS) 1.8 and average molar weight (M_w) 1310 g mol⁻¹) was provided by Wacker Chemie GmbH while the native β-cyclodextrin (denoted β-CD, average M_w 1135 g mol⁻¹) was a gift from Roquette Frères. PEO-PPO-PEO triblock copolymer [PEO poly(ethylene oxide) and PPO poly(propylene oxide)], denoted Pluronic P123 (average composition PEO₂₀PPO₇₀PEO₂₀, M_w 5800 g mol⁻¹), titanium isopropoxyde, Ti(OⁱPr)₄ (M_w 284.3 g mol⁻¹, d 0.96 g cm⁻³), nitric acid (HNO₃, 68%), and methyl oleate (M_w 296.49 g mol⁻¹) were purchased from Sigma Aldrich. Ruthenium(III) nitrosyl nitrate (Ru(NO)(NO₃)₃, 1.5% Ru, M_w 318.10 g mol⁻¹) was procured from STREM Chemicals. All chemicals were used as received without further purification.

2.2 Synthesis of TiO₂ materials and RuO₂/TiO₂ composites

Titanium dioxide nanoparticles were synthesised according to a previously reported sol–gel method.⁵⁸ Briefly, 30 mL (0.1 mol) of Ti(OⁱPr)₄ was first dissolved in 27 mL (0.35 mol) of isopropanol. Then, 160 mL of hot distilled water was added rapidly at 85 °C under vigorous stirring (hydrolysis ratio h = [H₂O]/[Ti] = 88). After 15 min, 1.3 mL of nitric acid ([HNO₃]/[Ti] = 0.2) was added dropwise to peptise the viscous precipitate and the mixture was maintained under reflux at 85 °C for 16 h. The final product was a stable translucent suspension of TiO₂ nanoparticles crystallised in anatase (70%) and brookite (30%). The concentration of titanium in the sol was 0.5 mol L⁻¹, as determined by weight loss on ignition at 1000 °C for 2 h.⁵⁹ Subsequently, Pluronic P123 (5 wt%, P123/Ti molar ratio = 0.017) and RaMeβ-CD (50 mg mL⁻¹, RaMeβ-CD/Ti molar ratio = 0.076) were added successively to 100 mL of the above titania sol. The mixture was stirred for 3 h, and then allowed to equilibrate at room temperature for 24 h. A xerogel was recovered after drying the sol at 60 °C for 72 h after which time it was calcined in air at 500 °C for 2 h using a heating ramp of 5 °C min⁻¹. In a second step, 800 mg of the calcined titania materials, prepared without or with the supramolecular template (RaMeβ-CD/P123 molar ratio = 4.5), were impregnated with an aqueous solution of ruthenium(III) nitrosyl nitrate prepared without or with β-CD (total volume 6 mL, 0.0098 mol L⁻¹ β-CD, β-CD/Ru molar ratio = 0.09). The suspension was maintained under stirring at 75 °C until the water was completely evaporated,⁴⁴ then dried at 120 °C (24 h) and calcined at 400 °C (4 h) under air flow. The obtained catalysts were identified according to the following notation: RuxTiO₂-sg or RuxTiO₂-ns where x indicates the weight percentage of ruthenium in the TiO₂ support, while sg and ns refer to conventional sol–gel and nanostructured TiO₂ respectively. If the native β-cyclodextrin is used as dispersing agent, β-CD is also added in the sample notation. Thus, Ru2.5β-CDTiO₂-ns indicates a catalyst prepared with 2.5% Ru supported on a nanostructured TiO₂ material with assistance of β-CD, while Ru2.5TiO₂-sg indicates a catalyst prepared also with 2.5% Ru, but supported on a conventional sol–gel TiO₂ support without assistance of β-CD. In some experiments, two commercial TiO₂ were also used as supports, i.e. pure anatase and a mixture of anatase (80%) and rutile (20%) (P25).

2.3 Characterisation methods

Dynamic light scattering (DLS) measurements were performed at 25 °C with a Malvern Zeta Nanosizer instrument equipped with a 4 mW He–Ne laser operating at 633 nm and using a backscattering detection system (scattering angle θ = 173°). Powder X-ray diffraction data were collected on a Siemens D5000 X-ray diffractometer in a Bragg–Brentano configuration with a Cu Kα radiation source. Scans were run over the angular domains 5° < 2θ < 80° with a step size of 0.02° and a step time of 2 s. Crystalline phases were identified by comparing the experimental diffraction patterns to Joint Committee on Powder Diffraction Standards (JCPDS) files. The RuO₂ crystallite sizes were estimated from the line broadening of the (110) diffraction peak using the Scherrer equation,⁶⁰ i.e. D = Kλ/(β cos θ), where K is the shape factor (taken as 0.9 in this study considering that the particles are spherical), λ is the X-ray radiation wavelength (1.54056 Å for Cu Kα), β is the full width at half-maximum (fwhm) and θ is the Bragg angle. Nitrogen adsorption–desorption isotherms were collected at −196 °C using an adsorption analyser Micromeritics Tristar 3020. Prior to analysis, 200–400 mg samples were outgassed at 320 °C overnight to remove the species adsorbed on the surface. From N₂ adsorption isotherms, specific surface areas were determined by the BET method⁶¹ and pore size distributions were calculated using the NLDFT (nonlocal density functional theory) model⁶² assuming a cylindrical pore structure. The relative errors were estimated to be the following: S_BET, 5%; pore volume (pv) (DFT), 5%; pore size (ps) (DFT), 20%. Temperature programmed reduction (TPR) measurements were performed using a Micromeritics AutoChem 2920 chemisorption analyser equipped with a thermal conductivity detector (TCD) for monitoring the H₂ consumption. Prior to analysis, 30–40 mg samples were placed in a U-tube quartz reactor and outgassed under argon flow at 120 °C for 2 h. After cooling to 25 °C, the sample was reheated at 500 °C with a heating rate of 10 °C min⁻¹ and the RuO₂ reduction was performed in a 5% H₂/Ar (v/v) flow with a flow rate of 10 mL min⁻¹. Field emission scanning electron microscopy (FE-SEM) observations were performed to examine the morphology of the support materials using a FEG Hitachi S-4700 field-emission microscope operating at 5 kV. Before imaging, samples were covered with a thin layer of carbon to reduce the accumulation of charges at high magnification. High resolution transmission electron microscopy (HR-TEM) observations were performed on a Tecnai microscope operating at an accelerating voltage of 200 kV. The RuO₂/TiO₂ powder was deposited directly on the surface of a carbon coated copper grid.

2.4 Catalytic tests

The performance of the heterogeneous catalysts was evaluated in the liquid phase hydrogenation of methyl oleate (CH₃(CH₂)₇CH=CH(CH₂)₇CO₂CH₃, C18:1) to methyl stearate (CH₃(CH₂)₁₆CO₂CH₃, C18:0).^56,57 Before catalytic testing, the RuO₂/TiO₂ composite was introduced in a U-tube quartz reactor and reduced at 400 °C for 4 h under H₂ atmosphere with a heating rate of 2 °C min⁻¹ and a flow rate of 30 mL min⁻¹. The catalytic tests were carried out using a 50 mL cylindrical stainless steel autoclave reactor equipped with a magnetic drive stirring and a thermostatic bath for the temperature control. In a typical experiment, 80.8 mg supported catalyst (0.02 mmol Ru) was introduced to a solution containing 593 mg methyl oleate (MO/Ru molar ratio = 100) dissolved in 10 mL of heptane. The mixture was purged three times with nitrogen at room temperature and then, pressurized with H₂ (40 bar). The reactions were run at a constant temperature of 50 °C for 2 h using a stirring speed of 750 rpm. Finally, the reaction products were analysed using a Varian 3900 gas chromatograph (GC) equipped with a CP-Sil-5B column (15 m × 0.25 mm × 0.25 μm) and a flame ionisation detector (FID). The injector and detector temperatures were 250 °C. Conversion and selectivity were quantified based on the relative GC-areas referred to an external standard (dodecane) calibrated to the corresponding pure compounds. Each catalytic run was performed in duplicate and the values reported are the averages between two runs, with an uncertainty equal or less than 5%. For the recycling tests, the catalyst was recovered after each run by filtration, washed with heptane three times, dried at 70 °C for 2 h and reduced under H₂ flow at 400 °C for another 4 h.

3. Results and discussion

3.1 Preparation and characterisation of RuO₂/TiO₂ composites

The synthetic procedure employed for the preparation of the mesoporous RuO₂/TiO₂ composites is schematically illustrated in Fig. 1. First, the support material consisting of mesoporous TiO₂ was prepared using the RaMeβ-CD-P123 assemblies as template and the sol–gel synthesized TiO₂ nanoparticles (H₂O/Ti molar ratio = 88) as building blocks. After removal of the supramolecular template by calcination at 500 °C, a nanostructured and highly porous material was recovered. Next, the porous support was impregnated with a solution of ruthenium(III) nitrosyl nitrate and native β-CD (Ru/β-CD molar ratio 11), followed by drying at 120 °C and calcination at 400 °C under air-flow. In this synthetic procedure, the RaMeβ-CD acts as micelle expander^21,22 providing an enlargement of the pore diameter of the TiO₂ support, while the native β-CD behaves as dispersing agent³⁰ ensuring a uniform repartition of ruthenium nanoparticles over the TiO₂ support. Finally, the performance of these supported catalysts was evaluated in the liquid phase hydrogenation of methyl oleate (MO, C18:1) to methyl stearate (MS, C18:0) under mild conditions (50 °C, 40 bar H₂).

Fig. 1 Schematic illustration of the template-directed synthesis of mesoporous Ru/TiO₂ catalysts where RaMeβ-CD-P123 assemblies act as supramolecular template, the native β-CD as dispersing agent and TiO₂ colloids as building blocks for the construction of a nanostructured framework.

We first examined the impact of RaMeβ-CD-P123 assemblies on the structural and textural characteristics of the support material. The sol–gel TiO₂ material (TiO₂-sg, no template used) (Fig. 2A) presents very low pore volume and surface area (0.03 cm³ g⁻¹ and 21 m² g⁻¹ respectively, Table 1) due to the small size of the crystallites in the titania sol and their high level of aggregation which facilitates the contact between neighbouring grains, thus resulting in a very rapid conversion to rutile (XRD Fig. 2B) and a drastic decrease of the pore volume.²⁵ The isotherm and pore size distribution of nanostructured RaMeβ-CD-P123-templated TiO₂ (TiO₂-ns) (Fig. 2A) indicate that the supramolecular assemblies remarkably improve the textural characteristics of the support material. Indeed, a dramatic increase in the pore size (from 5.3 to 16.3 nm) and pore volume (from 0.03 to 0.36 cm³ g⁻¹) is noticed together with a strong increase in the specific surface area (from 21 to 80 m² g⁻¹) (Table 1). Moreover, from the XRD patterns shown in Fig. 2B, it can be seen that the RaMeβ-CD-based assemblies make the titania material more stable against the anatase-to-rutile and brookite-to-rutile phase transformation, as evidenced by the strong decrease in the intensity of the (110) peak at 2θ = 27.4° assigned to the rutile phase (R) (JCPDS card no. 00-034-0180).²⁵ Finally, from the representative FE-SEM images, it is clear that the RaMeβ-CD-based assemblies also play a key role in restructuring the titania framework. Thus, while the TiO₂-sg material is made up predominantly of shapeless and aggregated particles (Fig. 2C), the RaMeβ-CD-P123-templated material (TiO₂-ns) is mainly comprised of fine and rounded grains randomly packed into a highly porous framework (Fig. 2D).

Fig. 2 N₂-adsorption isotherms and corresponding pore size distributions (inset) (A) and XRD patterns (B) of the sol–gel titania materials prepared without template (a) and with RaMeβ-CD-P123 assemblies (b). FE-SEM images of TiO₂-sg (C) and TiO₂-ns (D) showing the effect of the supramolecular template on the porosity of the resulting material. Samples were calcined at 500 °C.

Table 1 Characteristics of the TiO₂ materials and supported RuO₂/TiO₂ catalysts after thermal treatment at 400 °C

Sample	SBET^a (m^2 g^−1)	Vcum^b (cm^3 g^−1)	PS^c (nm)	RuO2 size^d (nm)
a Specific surface area determined in the relative pressure range 0.1–0.25.b Cumulative pore volume.c Average pore size resulting from NLDFT calculations.d RuO2 crystallite size determined using the Scherrer equation.
TiO2 supports
TiO2-sg	21 ± 1	0.030 ± 0.002	5.3 ± 1.1	—
TiO2-ns	80 ± 5	0.360 ± 0.018	16.3 ± 3.3	—
RuO2-TiO2 catalysts
Ru2.5TiO2-sg	8 ± 0	0.015 ± 0.001	6.2 ± 1.2	15.8 ± 1.6
Ru2.5β-CDTiO2-sg	16 ± 1	0.045 ± 0.002	5.9 ± 1.2	11.5 ± 1.2
Ru2.5TiO2-ns	51 ± 3	0.265 ± 0.012	16.5 ± 3.3	9.8 ± 1.0
Ru2.5β-CDTiO2-ns	75 ± 4	0.314 ± 0.016	16.8 ± 3.3	9.2 ± 0.9

---

Then, the interactions between Ru(III) nitrosyl nitrate and native β-CD were investigated using dynamic light scattering (DLS) combined with X-ray diffraction (XRD) measurements. Fig. 3A displays the DLS plots of native β-CD (0.016 M) before (a) and after equimolar addition of ruthenium salt (Ru/β-CD molar ratio = 1) (b). As expected, β-CD presents a bimodal size distribution plot with two populations centred at 70 and 380 nm associated with the formation of aggregates. Interestingly, after addition of ruthenium nitrosyl nitrate, these aggregates undergo partial disintegration, as evidenced by the shift of the hydrodynamic radius to 22 and 130 nm respectively. Further evidence for such interaction was provided from XRD measurements performed on β-CD and Ru/β-CD assemblies dried at 60 °C (Fig. 3B). Thus, the numerous diffraction lines characteristic of the cage-type crystalline microstructure of β-CD disappear upon addition of ruthenium nitrosyl nitrate indicating disruption of the β-CD microstructure. In analogy with the experiment described above, DLS measurements were also performed on the β-CD/TiO₂ suspensions (Fig. 3C). From the size distribution plots of the TiO₂ sols (0.5 M) prepared without and with the native β-CD (β-CD/Ti molar ratio = 0.032), the decrease in the apparent hydrodynamic radius from 110 to 85 nm suggests rupture of TiO₂ aggregates due to the adsorption of the cyclic molecule onto the TiO₂ surface.²⁴ Moreover, the disappearance of the most intense XRD reflections (at 9.0° and 12.5°) of the corresponding TiO₂ and β-CD/TiO₂ xerogels dried at 60 °C confirms the breakdown of the cage-type structure of β-CD after interaction with titania surface (Fig. 3D).

Fig. 3 Hydrodynamic radius distributions (R_H) of the scattered intensity (A) for the native β-CD solution (0.016 M) prepared without or with Ru(III) nitrosyl nitrate (β-CD/Ru molar ratio = 1). XRD patterns of the corresponding xerogels dried at 60 °C (B). R_H distributions of the native β-CD solution (0.016 M), TiO₂ sol (0.5 M), and β-CD/TiO₂ mixtures (β-CD/Ti molar ratio = 0.032) (C). XRD patterns of the corresponding xerogels dried at 60 °C (D).

Taken together, our results provide evidence of the fact that β-CD presents a strong affinity for both the mesoporous TiO₂ support and the Ru(III) nitrosyl nitrate. Such behaviour is likely to alter the degree of dispersion of the ruthenium nanoparticles over the support and, by consequence, affect the performance of the final catalyst.

On the bases of the results obtained above, we focused the continuation of our study on the preparation and characterization of a series of Ru/TiO₂ catalysts using both conventional sol–gel TiO₂ (TiO₂-sg) and nanostructured RaMeβ-CD-P123-templated TiO₂ (TiO₂-ns) supports. The ruthenium nitrosyl nitrate impregnating solution was prepared without or with assistance of the native β-CD (Ru/β-CD = 11).

XRD patterns of the different catalysts are presented in Fig. 4. Similarly to the undoped TiO₂-ns support (Fig. 2B), Ru2.5TiO₂-ns and Ru2.5β-CDTiO₂-ns catalysts contain only anatase and brookite implying that no anatase-to-rutile or brookite-to-rutile phase transformation occurs during the second calcination at 400 °C. On the other hand, three distinct crystalline phases (anatase, brookite and rutile) are clearly observed for the Ru2.5TiO₂-sg and Ru2.5β-CDTiO₂-sg catalysts. Furthermore, for all catalysts, in addition to the peaks of the support material, other reflexion lines are clearly identified at 2θ = 27.8° and 34.9° which could be indexed to the (110) and (101) planes respectively of the tetragonal RuO₂ (JCPDS 00-043-1027). Note that for the TiO₂-sg supported catalysts, the (110) and (101) planes of RuO₂ and rutile-TiO₂ overlap due to the lattice matching between these two phases both of which adopt a tetragonal structure. It is also worthy to mention that both the support and the dispersing agent affect the size of RuO₂ crystallites. Thus, for the Ru2.5 and Ru2.5β-CD catalysts deposited over TiO₂-ns, the crystallite sizes determined from the line broadening of the (110) diffraction peak are 9.8 nm and 9.2 nm respectively, while the average sizes obtained for the catalysts deposited over TiO₂-sg are beyond 10 nm (Table 1).

Fig. 4 XRD patterns of RuO₂/TiO₂ composites (2.5 wt% Ru) supported over conventional sol–gel TiO₂ (TiO₂-sg) and nanostructured RaMeβ-CD-P123-templated TiO₂ (TiO₂-ns). Both composites were prepared without or with assistance of native β-CD as dispersing agent. R in the figure denotes rutile-TiO₂.

To further investigate the impact of RuO₂ nanoparticles on the textural characteristics of the catalysts, N₂-adsorption analyses were carried out. From the adsorption isotherms (Fig. 5A) and corresponding pore size distributions (Fig. 5B), it can be noticed that, whatever the type of the support (TiO₂-sg or TiO₂-ns), the textural characteristic are only slightly modified after deposition of RuO₂ nanoparticles, indicating that no pore blocking occurs. For instance, the surface area, pore volume and pore size of the Ru2.5β-CDTiO₂-ns catalyst (75 m² g⁻¹, 0.31 cm³ g⁻¹ and 16.8 nm respectively) are very close to those of the undoped TiO₂-ns support (80 m² g⁻¹, 0.36 cm³ g⁻¹ and 16.3 nm respectively) (Table 1). In a somewhat different way, the Ru2.5β-CDTiO₂-sg catalyst presents a higher pore volume compared to the TiO₂-sg support (0.045 cm³ g⁻¹ vs. 0.030 cm³ g⁻¹) as well as a second population of mesopores centred at 14.8 nm. This is an indication that the native β-CD has also a beneficial effect in restructuring the titania framework during the impregnation step due to the ability of this cyclic molecule to adsorb onto the surface of the oxide nanoparticles, as mentioned earlier.

Fig. 5 N₂-adsorption isotherms (A) and corresponding pore size distribution plots (B) of RuO₂/TiO₂ composites (2.5 wt% Ru) supported over conventional sol–gel TiO₂ (TiO₂-sg) and nanostructured RaMeβ-CD-P123-templated TiO₂ (TiO₂-ns) prepared without or with assistance of native β-CD.

To probe the reducibility of mesoporous RuO₂/TiO₂ composites, temperature programmed reduction (TPR) measurements were also carried out. The TPR profiles of the different supported catalysts are shown in Fig. 6.

Fig. 6 H₂-TPR profiles of RuO₂/TiO₂ composites (2.5 wt% Ru) supported over conventional sol–gel TiO₂ (TiO₂-sg) and nanostructured RaMeβ-CD-P123-templated TiO₂ (TiO₂-ns) without or with assistance of β-CD. The TPR profile of non-doped TiO₂-ns support is added for comparison.

As expected, the non-doped TiO₂ material shows a very weak signal and no reduction peak indicating that the support alone can hardly react with H₂. In contrast, the Ru2.5TiO₂-sg catalyst displays a very broad peak centred at ∼320 °C assigned to the reduction of Ru⁴⁺ to metallic Ru⁰.⁶³ Such a high reduction temperature may result from the formation of RuO₂ aggregates which interact strongly with the rutile phase (major polymorph in the TiO₂-sg support) and which are difficult to reduce. On the other hand, the Ru2.5β-CDTiO₂-sg catalyst displays two detached peaks at ∼320 °C and ∼210 °C suggesting that, at least, two different ruthenium species coexist in this sample. Thus, the higher temperature may be attributed to the reduction of RuO₂ aggregates, in strong interaction with the TiO₂-sg support, which exist also in the Ru2.5TiO₂-sg catalyst, while the lower one may be assigned to the reduction of small and more easily reducible RuO₂ species in weak interaction with the support and formed only in the presence of β-CD. This result suggests that, by locating at the metal–support interface, the cyclodextrin should reduce the contact points between TiO₂ and RuO₂ particles, thus preventing the incorporation of these latter into the titania support during the calcination process. Note that the catalysts supported over TiO₂-ns globally display lower reduction temperatures, i.e. in the range of 180–230 °C, compared to those supported over TiO₂-sg materials (210–320 °C). This is an indication that the RaMeβ-CD-P123-templated support promotes the formation of small and readily reducible RuO₂ nanoparticles. Moreover, as RuO₂ has different lattice parameters from TiO₂-anatase and TiO₂-brookite (the major polymorphs in the TiO₂-ns support), the metal–support interactions in these catalysts should be weaker compared to those between RuO₂ and TiO₂-sg support. Taken together, our results provide evidence of the fact that small nano-sized and well-dispersed RuO₂ particles could be obtained using nanostructured TiO₂ prepared by template-directed colloidal self-assembly as support and native β-CD as dispersing agent.

Further evidence for the key role of RaMeβ-CD-based assemblies on the dispersion of RuO₂ nanoparticles was provided by high resolution transmission electron microscopy (HR-TEM) combined with high angle annular dark-field scanning transmission electron microscopy (HAADF-STEM). Observations were performed on two selected catalysts, i.e. Ru2.5β-CDTiO₂-ns and Ru2.5β-CDTiO₂-sg.

From the TEM image shown in Fig. 7a, it can be noticed that the surface of the RaMeβ-CD-P123-templated TiO₂ support is decorated with RuO₂ nanoparticles which can be directly visualised by the bright spots resulting from the difference in Z-contrast between the two oxides (Fig. 7b). Moreover, EDS elemental mapping on different regions of the TEM micrograph (Fig. 7c–f) allows to further differentiate the two phases, i.e. TiO₂ (white regions) and RuO₂ nanoparticles (dark regions).

Fig. 7 TEM image (a) and HAADF-STEM images (b–d) of the Ru2.5β-CDTiO₂-ns catalyst showing uniform dispersion of RuO₂ nanoparticles over the mesoporous RaMeβ-CD-P123-templated TiO₂ material; EDS spectra with the electron beam focused first on a white region (TiO₂) (e), and then in a dark region (RuO₂) (f) of the catalyst images. HR-TEM micrographs of the Ru2.5β-CDTiO₂-ns catalyst indicating uniform dispersion of TiO₂ nanoparticles crystallised in anatase and brookite (g and h), as well as RuO₂ nanoparticles (5–10 nm) (i) TEM and HR-TEM images of the Ru2.5β-CDTiO₂-sg catalyst composed mostly of a non-porous support material showing TiO₂ and RuO₂ nanoparticle aggregation (j–n).

TiO₂ nanoparticles in this sample are identified as a mixture of tetragonal anatase (∼15 nm) and orthorhombic brookite (∼10 nm) (Fig. 7g and h). Interestingly, mesopores of 15–20 nm diameter, resulting from the self-assembly of these nanoparticles around the supramolecular template, can also be clearly visualised in this micrograph (pink circles), in agreement with N₂-adsorption analyses. On the other hand, RuO₂ nanoparticles are more or less spherical and approximately 10–12 nm in size (Fig. 7i), in agreement with XRD results. Conversely, the Ru2.5β-CDTiO₂-sg catalyst shows mainly a dense and less porous network (Fig. 7j) made-up of anatase (∼35%), brookite (∼26%) and rutile (∼39%) and over which RuO₂ particles tend to form elongated heterogeneous structures (Fig. 7k–n). Such unusual epitaxial growth of RuO₂ nanoparticles over rutile-TiO₂ was also observed by Xiang et al.⁶⁴ in a previous study and was suggested to result from the structural matching between these two phases which actually share the same type of lattice symmetry (tetragonal) and have very similar lattice parameters (a = 4.59 Å, c = 2.96 Å for rutile-TiO₂ and a = 4.49 Å, c = 3.10 Å for RuO₂).

3.2 Catalytic performance

Methyl esters of vegetable oils, derived from polyunsaturated fatty acids, can be transformed through partial or complete hydrogenation into saturated fatty acids with improved physical properties (e.g. high melting point and oxidative stability).^55,65 Methyl stearate (MS, C18:0) is an important saturated oil obtained from catalytic hydrogenation of unsaturated methyl oleate (MO, C18:1) and widely used as a feedstock for catalytic hydrogenolysis into its corresponding saturated fatty alcohol (stearyl alcohol, C18:0).^56,66 Ru-based catalysts have been shown to be effective for low temperature hydrogenation of carboxylic acid esters, especially for selective hydrogenation of the olefinic C=C bond without affecting the carbonyl C=O bond. In the present study, we evaluated the performance of the different supported Ru/TiO₂ catalysts in the liquid phase hydrogenation of methyl oleate to methyl stearate.

Prior to catalytic tests, RuO₂/TiO₂ composites were pre-treated under hydrogen flow at 400 °C for 4 h to reduce the RuO₂ in metallic Ru. TEM analyses confirmed that Ru(0)/TiO₂-sg was made-up predominantly of aggregated particles dispersed in a dense and less porous network (TiO₂-sg), while Ru(0)/TiO₂-ns catalyst was mainly comprised of small and well-dispersed nanoparticles dispersed in a porous and nanostructured network (TiO₂-ns) (Fig. S1, ESI†). This is in agreement with what was observed on non-reduced RuO₂/TiO₂-sg and RuO₂/TiO₂-ns composites (Fig. 7).

The results obtained in catalysis after 2 h of reaction (40 bar H₂, 50 °C) with the four catalysts, prepared using either conventional sol–gel or nanostructured TiO₂ supports, with the assistance or not of the native β-CD (Ru/β-CD = 11), are shown in Fig. 8. The results obtained with the non-doped supports (TiO₂-sg and TiO₂-ns) as well as with the Ru2.5 catalyst supported over commercial TiO₂ (Ru2.5TiO₂-A and Ru2.5TiO₂-P25) are added for comparison. Depending on the preparation method, clear differences can be observed in both the catalytic activity (Fig. 8a) and selectivity (Fig. 8b). The control TiO₂-sg and TiO₂-ns supports show almost no activity in the hydrogenation of methyl oleate. On the other hand, all supported Ru/TiO₂ catalysts are selective in the hydrogenation of the olefinic C=C bond producing the fully saturated ester, i.e. methyl stearate (C18:0) as the main product and the methyl elaïdate (C18:1, trans-9) as the only byproduct. Moreover, it is worth noting that the catalytic performance strongly differs depending on the type of the support (TiO₂-sg or TiO₂-ns) and the impregnation method (with or without assistance of native β-CD). Thus, when Ru is deposited over nanostructured RaMeβ-CD-P123-templated TiO₂, the catalytic activity increases sharply from 36% (Ru2.5/TiO₂-sg) to 82% (Ru2.5/TiO₂-ns). Interestingly, further improvement in the conversion efficiency is noticed when the native β-CD is used as dispersing agent. Thus, the conversion rate increases by ∼14% (from 36% to 50%) for the Ru2.5 catalyst deposited over TiO₂-sg and by ∼12% (from 82% to 94%) for the same catalyst deposited over TiO₂-ns support. Note that among the four supported catalysts prepared using colloidal self-assembly approach, Ru2.5β-CDTiO₂-ns is the fastest one, achieving almost full MO conversion (∼94%) in 2 h. Moreover, it is interesting to notice that the activity of this catalyst is almost 3-fold higher than that of ruthenium deposited over commercial anatase-TiO₂ (Ru2.5TiO₂-A) and P25-TiO₂ (Ru2.5TiO₂-P25).

Fig. 8 Comparison of the catalytic activity (a) and selectivity (b) of different supported catalysts in the hydrogenation of methyl oleate (C18:1) to methyl stearate (C18:0). Reaction conditions: Ru (0.02 mmol), methyl oleate (MO/Ru molar ratio = 100), H₂ (40 bar), solvent (heptane, 10 mL), stirring rate (750 rpm), temperature (50 °C), reaction time (2 h).

Similarly to the catalytic activity, clear differences can be observed also in the selectivity of the supported catalysts. Thus, the increase in the methyl oleate conversion when moving from Ru2.5TiO₂-sg to Ru2.5TiO₂-ns leads also to a dramatic increase in the selectivity for the methyl stearate (from 47% to 80%). However, although a beneficial effect of β-CD is noticed for the Ru2.5TiO₂-sg catalyst (MS selectivity increases from 47% to 67%), no clear difference exists between the Ru2.5TiO₂-ns catalysts prepared without or with assistance of β-CD (MS selectivity is the same ∼80%).

Over the investigated mesoporous Ru/TiO₂ catalysts, the catalytic performance agrees well with the textural characteristics of the support and the degree of dispersion of the ruthenium species. Indeed, the highest performance of the Ru2.5β-CDTiO₂-ns catalyst may result from a combined effect of improved textural characteristics of the support and uniform dispersion of ruthenium nanoparticles. Thus, the high surface area of the RaMeβ-CD-P123-templated TiO₂ material (80 m² g⁻¹) should provide a high level of dispersion of ruthenium nanoparticles, facilitating the contact of methyl oleate with the catalyst surface during the hydrogenation reaction. On the other hand, the high pore volume (0.36 cm³ g⁻¹) should allow for more reactant molecules to be adsorbed to the internal surface of the pores, thus improving the diffusion of reactants and products toward the active phase during the catalytic process. Finally, the use of native β-CD as dispersing agent may allow for reducing the interparticle aggregation thanks to the ability of this cyclodextrin to interact with both RuNO(NO₃)₃ and TiO₂ nanoparticles.^30,32,33 Overall, our results reveal a dual role played by the CD-based assemblies: an enhancement of the porosity and surface area of the TiO₂ material and an improvement of the dispersion of Ru nanoparticles over the support, parameters that appear to be essential for enhancing the efficacy of the catalyst in the liquid phase hydrogenation of methyl oleate.

Moreover, as leaching and deactivation are common phenomena encountered in any liquid phase hydrogenation reaction, the ease of recovery and reusability of the catalyst are important features that need to be considered. For this purpose, we have investigated the recyclability of the most efficient catalyst, i.e. Ru2.5β-CDTiO₂-ns, in three successive runs. Interestingly, no loss of activity and selectivity was observed during the consecutive cycles, indicating that the supported catalyst exhibits high stability under the employed reaction conditions (Fig. 9).

Fig. 9 Reusability tests of the Ru2.5β-CDTiO₂-ns catalyst. Reaction conditions: methyl oleate/ruthenium molar ratio (MO/Ru = 100), H₂ (40 bar), solvent (heptane, 10 mL), stirring rate (750 rpm), temperature (50 °C), reaction time (2 h).

We expect that the high surface area and large pore volume generated by the template-directed self-assembly of titania colloids should allow for a good dispersion of nanosized ruthenium particles and, therefore, effectively prevent aggregation while improving the chemical stability of the catalyst. Moreover, the thermal treatment at 400 °C should allow for a stable fixation of ruthenium species over the TiO₂ material, preventing leaching from the matrix to the reaction media during the catalytic process. Hence, the high efficacy of this catalyst combined with ease of recovery and reuse without loss of conversion and selectivity makes the template-directed colloidal self-assembly a simple and versatile approach for the synthesis of supported catalysts with high performance in liquid phase hydrogenation of methyl oleate.

4. Conclusions

In conclusion, we have demonstrated that the approach combining the template-directed colloidal self-assembly with a cyclodextrin-assisted impregnation method allows to prepare highly active and selective Ru/TiO₂ catalysts for the hydrogenation of methyl oleate (MO, C18:1) to methyl stearate (MS, C18:0) under mild conditions (50 °C, 40 bar H₂). The supramolecular assemblies formed between RaMeβ-CD and block copolymer P123 promoted the formation of a nanostructured and highly porous TiO₂ support material over which uniform dispersion of ruthenium species was achieved, without or with the assistance of the native β-cyclodextrin. The high activity of these catalysts was attributed to the ability of the supramolecular CD-based assemblies to generate TiO₂ materials with improved structural and textural characteristics and interact in a specific way with both the TiO₂ support and Ru species, thus improving the metal–support interactions.

Overall, our results revealed that the CD-based assemblies provide a versatile and easily accessible toolbox with different functionalities for generating metal-supported catalysts with controlled pore architecture and uniform metal distribution. The approach combining the use of two different cyclodextrins with different functionalities is versatile and may be generalised to other monometallic or bimetallic catalysts supported over simple or mixed oxides prepared by sol–gel process.

Acknowledgements

The Chevreul Institute (FR 2638), the European Regional Development Fund (ERDF), the Conseil Regional du Nord-Pas de Calais, the CNRS and the Ministère de l'Enseignement Supérieur et de la Recherche are acknowledged for supporting and funding this work. We thank Laurence Burylo (UCCS, University of Lille) for technical assistance in XRD measurements, Antonio Da Costa (UCCS, University of Artois) and Nora Djelal (UCCS, University of Lille) for their help with FE-SEM analyses, as well as Loïc Brunet of the BICeL-Campus Lille 1 Facility for access to instruments and technical advices. We are indebted to the Research Federation FRABio (Univ. Lille, CNRS, FR 3688, FRABio, Biochimie Structurale et Fonctionnelle des Assemblages Biomoléculaires) for providing the scientific and technical environment conducive to achieving this work. We kindly acknowledge Roquette Frères for generous gift of the native β-CD.

References

1. M. Cargnello, V. V. T. Doan-Nguyen, T. R. Gordon, R. E. Diaz, E. A. Stach, R. J. Gorte, P. Fornasiero and C. B. Murray, Science, 2013, 341, 771–773.
2. F. A. Westerhaus, R. V. Jagadeesh, G. Wienhofer, M. M. Pohl, J. Radnik, A. E. Surkus, J. Rabeah, K. Junge, H. Junge, M. Nielsen, A. Bruckner and M. Beller, Nat. Chem., 2013, 5, 537–543.
3. S. Pradhan, J. K. Bartley, D. Bethell, A. F. Carley, M. Conte, S. Golunski, M. P. House, R. L. Jenkins, R. Lloyd and G. J. Hutchings, Nat. Chem., 2012, 4, 134–139.
4. L. B. Sun, X. Q. Liu and H. C. Zhou, Chem. Soc. Rev., 2015, 44, 5092–5147.
5. Y. Y. Li, S. L. Yu, X. F. Wu, J. L. Xiao, W. Y. Shen, Z. R. Dong and J. X. Gao, J. Am. Chem. Soc., 2014, 136, 4031–4039.
6. V. Molinari, C. Giordano, M. Antonietti and D. Esposito, J. Am. Chem. Soc., 2014, 136, 1758–1761.
7. C. Schmuck, Nat. Nanotechnol., 2011, 6, 136–137.
8. Y. Hwang, K. Lee and Y. Kwon, Chem. Commun., 2001, 1738–1739.
9. M. Wong, E. Jeng and J. Ying, Nano Lett., 2001, 1, 637–642.
10. H. Zhang, E. W. Edwards, D. Wang and H. Möhwald, Phys. Chem. Chem. Phys., 2006, 8, 3288–3299.
11. G. Kaune, M. Memesa, R. Meier, M. A. Ruderer, A. Diethert, S. V. Roth, M. D'Acunzi, J. S. Gutmann and P. Mueller-Buschbaum, ACS Appl. Mater. Interfaces, 2009, 1, 2862–2869.
12. J. H. Fendler, Chem. Mater., 1996, 8, 1616–1624.
13. K. Mudiyanselage, S. D. Senanayake, L. Feria, S. Kundu, A. E. Baber, J. Graciani, A. B. Vidal, S. Agnoli, J. Evans, R. Chang, S. Axnanda, Z. Liu, J. F. Sanz, P. Liu, J. A. Rodriguez and D. J. Stacchiola, Angew. Chem., Int. Ed. Engl., 2013, 52, 5101–5105.
14. C. Wen, Y. Zhu, Y. Ye, S. Zhang, F. Cheng, Y. Liu, P. Wang and F. Tao, ACS Nano, 2012, 6, 9305–9313.
15. A. Caballero, J. P. Holgado, V. M. Gonzalez-dela Cruz, S. E. Habas, T. Herranz and M. Salmeron, Chem. Commun., 2010, 46, 1097–1099.
16. C. Mattevi, C. Tobias Wirth, S. Hofmann, R. Blume, M. Cantoro, C. Ducati, C. Cepek, A. Knop-Gericke, S. Milne, C. Castellarin-Cudia, S. Dolafi, A. Goldoni, R. Schloegl and J. Robertson, J. Phys. Chem. C, 2008, 112, 12207–12213.
17. A. Bruix, J. A. Rodriguez, P. J. Ramírez, S. D. Senanayake, J. Evans, J. B. Park, D. Stacchiola, P. Liu, J. Hrbek and F. Illas, J. Am. Chem. Soc., 2012, 134, 8968–8974.
18. A. K. Boal, F. Ilhan, J. E. DeRouchey, T. Thurn-Albrecht, T. P. Russell and V. M. Rotello, Nature, 2000, 404, 746–748.
19. A. T. Bell, Science, 2003, 299, 1688–1691.
20. P. Alphonse, R. Bleta, D. Lavergne, J. Barrault, A. Maciuca, J. M. Tatibouët and C. Rossignol, PCT Pat., WO 2015/055672 A1, 2015.
21. R. Bleta, C. Machut, B. Léger, E. Monflier and A. Ponchel, Macromolecules, 2013, 46, 5672–5683.
22. R. Bleta, C. Machut, B. Léger, E. Monflier and A. Ponchel, J. Inclusion Phenom. Macrocyclic Chem., 2014, 80, 323–335.
23. R. Bleta, S. Menuel, B. Léger, A. Da Costa, E. Monflier and A. Ponchel, RSC Adv., 2014, 4, 8200–8208.
24. R. Bleta, A. Lannoy, C. Machut, E. Monflier and A. Ponchel, Langmuir, 2014, 30, 11812–11822.
25. A. Lannoy, R. Bleta, C. Machut, E. Monflier and A. Ponchel, RSC Adv., 2014, 4, 40061–40070.
26. A. Harada, Acc. Chem. Res., 2001, 34, 456–464.
27. J. Araki and K. Ito, Soft Matter, 2007, 3, 1456–1473.
28. J. Li and X. J. Loh, Adv. Drug Delivery Rev., 2008, 60, 1000–1017.
29. F. Yuen and K. C. Tam, Soft Matter, 2010, 6, 4613–4630.
30. S. Tilloy, H. Bricout, S. Menuel, F. Hapiot and E. Monflier, Curr. Organocatal., 2016, 3, 24–31.
31. A. Jean-Marie, A. Griboval-Constant, A. Y. Khodakov, E. Monflier and F. Diehl, Chem. Commun., 2011, 47, 10767–10769.
32. L. Bai, F. Wyrwalski, J.-F. Lamonier, A. Y. Khodakov, E. Monflier and A. Ponchel, Appl. Catal., B, 2013, 138–139, 381–390.
33. H. Liu, Y. Li, H. Wu, W. Yang and D. He, Fuel, 2014, 136, 19–24.
34. H. Liu, Y. Li, H. Wu, W. Yang and D. He, Chin. J. Catal., 2014, 35, 1520–1528.
35. H. Liu, Y. Li, H. Wu, J. Liu and D. He, Chin. J. Catal., 2015, 36, 283–289.
36. D. Chen and R. A. Caruso, Adv. Funct. Mater., 2013, 23, 1356–1374.
37. L. Gomathi Devi and R. Kavitha, Appl. Catal., B, 2013, 140–141, 559–587.
38. S. U. M. Khan, M. Al-Shahry and W. B. Ingler, Science, 2002, 297, 2243–2245.
39. N. M. Nursam, X. Wang and R. A. Caruso, J. Mater. Chem. A, 2015, 3, 24557–24567.
40. T. Omotoso, S. Boonyasuwat and S. P. Crossley, Green Chem., 2014, 16, 645–652.
41. S. J. Tauster, S. C. Fung, R. T. K. Baker and J. A. Horsley, Science, 1981, 211, 1121–1125.
42. T. S. Kim, J. D. Stiehl, C. T. Reeves, R. J. Meyer and C. B. Mullins, J. Am. Chem. Soc., 2003, 125, 2018–2019.
43. F. Wang, S. Zhang, C. Li, J. Liu, S. He, Y. Zhao, H. Yan, M. Wei, D. G. Evans and X. Duan, RSC Adv., 2014, 4, 10834–10840.
44. C. Hernandez-Mejia, E. S. Gnanakumar, A. Olivos-Suarez, J. Gascon, H. F. Greer, W. Zhou, G. Rothenberg and N. Raveendran Shiju, Catal. Sci. Technol., 2016, 6, 577–582.
45. W. R. H. Wright and R. Palkovits, ChemSusChem, 2012, 5, 1657–1667.
46. R. Herbois, S. Noël, B. Léger, S. Tilloy, S. Menuel, A. Addad, B. Martel, A. Ponchel and E. Monflier, Green Chem., 2015, 17, 2444–2454.
47. O. F. Aldosari, S. Iqbal, P. J. Miedziak, G. L. Brett, D. R. Jones, X. Liu, J. K. Edwards, D. J. Morgan, D. K. Knight and G. J. Hutchings, Catal. Sci. Technol., 2016, 6, 234–242.
48. L. Fabre, P. Gallezot and A. Perrard, J. Catal., 2002, 208, 247–254.
49. Y. Pouilloux, A. Piccirilli and J. Barrault, J. Mol. Catal. A: Chem., 1996, 108, 161–166.
50. Y. Pouilloux, F. Autin, C. Guimon and J. Barrault, J. Catal., 1998, 176, 215–224.
51. N. Gokulakrishnan, G. Peru, S. Rio, J. F. Blach, B. Léger, D. Grosso, E. Monflier and A. Ponchel, J. Mater. Chem. A, 2014, 2, 6641–6648.
52. N. J. Divins, I. Angurell, C. Escudero, V. Pérez-Dieste and J. Llorca, Science, 2014, 346, 620–623.
53. C. T. Campbell, S. C. Parker and D. E. Starr, Science, 2002, 298, 811–814.
54. A. Cao and G. Veser, Nat. Mater., 2010, 9, 75–81.
55. M. Besson, P. Gallezot and C. Pinel, Chem. Rev., 2014, 114, 1827–1870.
56. C. Vasiliou, A. Bouriazos, A. Tsichla and G. Papadogianakis, Appl. Catal., B, 2014, 158–159, 373–381.
57. C. Hu, D. Creaser, S. Siahrostami, H. Grönbeck, H. Ojagh and M. Skoglundh, Catal. Sci. Technol., 2014, 4, 2427–2444.
58. R. Bleta, P. Alphonse and L. Lorenzato, J. Phys. Chem. C, 2010, 114, 2039–2048.
59. P. Alphonse and M. Courty, J. Colloid Interface Sci., 2005, 290, 208–219.
60. P. Scherrer, Nachr. Ges. Wiss. Göttingen Math.-Phys. Kl., 1918, 2, 98–100.
61. S. Brunauer, P. H. Emmett and E. Teller, J. Am. Chem. Soc., 1938, 60, 309–319.
62. R. Evans, U. M. B. Marconi and P. Tarzona, J. Chem. Soc., Faraday Trans. 2, 1986, 82, 1763–1787.
63. E. V. Ramos-Fernandez, J. Silvestre-Albero, A. Sepulveda-Escribano and F. Rodriguez-Reinoso, Appl. Catal., A, 2010, 374, 221–227.
64. G. Xiang, X. Shi, Y. Wu, J. Zhuang and X. Wang, Sci. Rep., 2012, 2(801), 1–6.
65. A. Philippaerts, P. A. Jacobs and B. F. Sels, Angew. Chem., Int. Ed., 2013, 52, 5220–5226.
66. J. Pritchard, G. A. Filonenko, R. van Putten, E. J. M. Hensen and E. A. Pidko, Chem. Soc. Rev., 2015, 44, 3808–3833.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra27161d

---