Mesoporous organosilica nanotubes containing a chelating ligand in their walls

Nanotubes composed of various materials, such as carbon,^1,2 silica,^3–5 and titania,⁶ have attracted much attention as catalyst supports due to their high surface areas, easy access to active sites in the tubes, and confinement effects inside the cavity. As an example, Chen et al. reported that Fe nanoparticles confined inside carbon nanotubes showed significant increases in both activity and selectivity for the Fischer-Tropsch synthesis in comparison to Fe nanoparticles deposited on the external surfaces of carbon nanotubes or Fe nanoparticles deposited on activated carbon.⁷ Kamegawa et al. reported higher catalytic activity for Ti-containing silica nanotubes during the epoxidation of alkanes and unsaturated fatty acid methyl esters, compared to Ti-containing mesoporous silica and zeolite (TS-1), attributed to the efficient transport of reactants to the active sites in the tubes.⁸ These results suggest the significant potential of nanotubes as catalyst supports.

Recently, organosilica nanotubes have been prepared from bridged organosilane precursors [(R^′O)₃Si–R–Si(OR^′)₃, R = organic group, and R^′ = Me,Et] using a single-micelle-templating approach. These materials have an advantage in which various organic functionalities can be incorporated into the tube walls,^9,10 although the organosilica nanotubes reported to date have contained only simple organic compounds such as ethane or benzene, which limit the functionality of the walls. The 2,2′-bipyridine molecule is a typical organic chelating ligand used during the formation of a variety of metal complexes showing specific catalytic functions, mainly in homogeneous solutions.^11–13 Thus, the synthesis of nanotubes containing 2,2′-bipyridine in their walls is an important step towards forming a catalyst support that enables the use of unique metal complexes as heterogeneous catalysts. Conventionally, 2,2′-bipyridine has been grafted on porous solids such as silica gel or mesoporous silica, serving as an organic linker for the fixation of metal complexes.^14,15 However, in many cases, the activity of the metal complexes catalysts grafted on porous solids is greatly decreased due to unfavorable interactions between the metal complex and the solid surface and/or limited diffusion of molecules in the narrowed pores. Recently, however, ordered mesoporous organosilicas^16,17 and metal-organic-frameworks (MOFs)^18–21 containing 2,2′-bipyridine in their frameworks have been reported.

Herein, we report the synthesis of organosilica nanotubes containing bipyridine ligands in their walls (denoted as BPy-NT) and the direct formation of an iridium complex on the tube walls (denoted as Ir-BPy-NT). The bipyridine-containing organosilica nanotubes were prepared by the co-condensation of 2,2′-bipyridine- and benzene-bridged organosilane precursors 1¹⁶ and 2 (Scheme 1), since the synthesis from a 100% bipyridine-based precursor proved too difficult. Unfortunately, the co-condensation of the two organosilanes sometimes resulted in segregated condensation into different domains, and so we carefully studied the distribution of 2,2′-bipyridine groups in the tube walls by scanning transmission electron microscopy (STEM) observations. The iridium complex [IrCp*Cl(bpy), Cp*: η⁵-pentamethylcyclopentadienyl] was directly formed on the walls of the BPy-NT without any linker, using the framework bipyridine as the chelating ligand (Scheme 1). The resulting unique heterogeneous iridium complex exhibited significant catalysis for water oxidation to generate oxygen, using Ce⁴⁺ as an oxidant.

In a typical synthesis, 0.55 g of triblockcopolymer [(EO)₂₀(PO)₇₀(EO)₂₀, P123] and 1.75 g of KCl were dissolved in 180 ml of an aqueous 2M HCl solution at 38 °C. After the copolymer was fully dissolved, 3.15 mmol of 2 was added with vigorous stirring. After stirring for 6 h at 38 °C, 0.35 mmol of 1 dissolved in 2.8 ml ethanol was added dropwise. The molar ratio of benzene to bipyridine precursors in the resulting mixture was 9. The mixture was stirred at 38 °C for a further 24 h and then heated to 100 °C under static conditions for an additional 24 h. The solid product was recovered by filtration and was then air dried at room temperature overnight. Finally, the surfactant was extracted by refluxing 1.0 g of the as-synthesized material in 200 ml of ethanol containing 1.5 g concentrated HCl for 24 h. The obtained sample was deprotonated by treatment with an ammonia solution (28 wt.%) to yield BPy-NT. In addition, organosilica nanotubes containing only benzene groups were also prepared, using 3.5 mmol of the benzene-bridged precursor 2, as a reference material. The synthetic conditions applied to this material were the same as used during the synthesis of BPy-NT, and the resulting sample was denoted B-NT. The Ir-BPy-NT was prepared by adding BPy-NT (50 mg) to a solution of [Cp*IrCl(μ-Cl)]₂ (5 mg) in 30 ml anhydrous ethanol under an argon atmosphere. After the suspension was stirred under refluxing conditions for 24 h, the solid phase was removed by filtration and washed with DMF (N,N-dimethylformamide) and distilled water to remove unreacted [Cp*IrCl(μ-Cl)]₂.

Figure 1 presents scanning electron microscopy (SEM) and bright-field scanning transmission electron microscopy (BF-STEM) images of the BPy-NT acquired by using a HITACHI S-5500 at 30 kV. These clearly show the formation of nanotubes with outer diameters less than 20 nm and lengths of several hundred nanometers. The nitrogen adsorption-desorption isotherm of the BPy-NT corresponded to type IV, which is typical for mesoporous materials (Figure 2(a)), while the pore size distribution plot shows a narrow peak centered at 7.8 nm (Fig. 2(b)). The Brunauer-Emmett-Teller (BET) surface area and pore volume of the BPy-NT as calculated from the adsorption branch of the nitrogen adsorption isotherm were 961 m² g⁻¹ and 0.9 cm³ g⁻¹ (at P/P₀ = 0.80), respectively (Table I). The textural properties of the B-NT are almost the same as those of the BPy-NT (Figure 2 and Table I). The pore diameter and BET surface area were found to be slightly decreased following the formation of the iridium complex on the BPy-NT (Ir-BPy-NT), as shown in Figure 2 and Table I.

The presence of 2,2′-bipyridine in the BPy-NT was confirmed by UV-vis absorption spectra. Figure 3(a) shows the UV-vis spectrum of the final BPy-NT product together with that of the same product prior to treatment with NH₃. The pre-treated BPy-NT exhibits two main absorption peaks at approximately 270 and 315 nm, while the peak at 315 nm is seen to shift to 290 nm after the NH₃ treatment. The bipyridine groups of the BPy-NT were protonated immediately after their preparation since they were synthesized under acidic conditions and then treated with an acidic solution to remove the surfactant. Thus, the peak shift following the NH₃ treatment is due to the deprotonation of the bipyridine groups.²² A mixture of the benzene- and bipyridine-bridged precursors (benzene/bipyridine molar ratio of 9) generated a similar UV-vis spectrum to that of the BPy-NT, indicating that both benzene and bipyridine groups were present in the BPy-NT sample (Figure 3(b)). In the precursor mixture spectrum, the intensity of the bipyridine peak was greater than that of the benzene peak even though the benzene/bipyridine molar ratio was 9, due to the very different molar extinction coefficients of benzene and bipyridine (200 and 14,000 l mol⁻¹ cm⁻¹, respectively, at 300 nm). CHN elemental analysis of the BPy-NT showed that its bipyridine content was 0.36 mmol g⁻¹ and its benzene/bipyridine molar ratio was 14, meaning that this ratio was significantly larger than the ratio in the starting solution. The ¹³C cross-polarization (CP) magic-angle-spinning (MAS) nuclear magnetic resonance (NMR) spectrum of the BPy-NT showed very small peaks at 120-160 ppm due to the bipyridine groups, in addition to a strong peak at 133 ppm due to the benzene groups (Figure 4(a)). This spectrum, therefore, demonstrates the presence of a small amount of bipyridine groups along with a larger quantity of benzene groups in the walls, which is in good agreement with the results of CHN analysis. The signals at 59 and 17 ppm can be assigned to ethoxy group carbons reacted with silanol groups during the ethanol (EtOH) treatment for removal of the surfactant. The ²⁹Si MAS NMR spectrum of the BPy-NT shows three resonance peaks at −62, −71, and −80 ppm (Figure 4(b)), assigned to silicon species T¹ [SiC(OH)₂(OSi)], T² [SiC(OH)(OSi)₂], and T³ [SiC(OSi)₃], respectively. The absence of SiO₄ species such as Q³ [Si(OH)(OSi)₃] and Q⁴ [Si(OSi)₄] in the range of −90 to −120 ppm indicates that no carbon-silicon bond cleavage occurred during the synthesis.

The above results demonstrate only the existence of benzene and bipyridine groups in the sample but do not show the homogeneous distribution of both in the tube walls. In order to study the distribution of bipyridine groups in the walls, high-angle annular dark-field (HAADF)-STEM images of the Ir-BPy-NT were obtained by using a JEOL JEM-2100F at 200 kV (Figure 5). Two regions are clearly seen in these images; broad, curved lines (a few nanometers in width) with a dark contrast and small bright spots less than ca. 0.5 nm in size. The former regions correspond to projections of the nanotube walls along the incident electron beam direction, while the latter spots indicate that heavier atoms or sub-nano clusters consisting of heavy atoms were present, which in this case would be Ir. Since the distribution of the bright spots is not concentrated on the outer/inner surfaces of the nanotube walls but is rather uniform throughout the walls, one can conclude that the Ir atoms/clusters were uniformly distributed along the nanotube walls. The formation of the IrCp*Cl(bpy) complex was also confirmed by UV-vis spectroscopy. The UV-vis spectrum of the Ir-BPy-NT shows a new band appearing at approximately 360 nm (Figure 3(c)). This can be assigned to the metal-to-ligand charge transfer (MLCT) band of the IrCp*Cl(BPy-NT) since the exact same band was observed in a model compound (IrCp*Cl(bpy-Si), bpy-Si: Me₃Si-bpy-SiMe₃). The absorption band at 290 nm due to the bipyridine groups in the BPy-NT is also shifted to 303 nm after the treatment with the Ir precursor, indicating that the bipyridine ligands in the tube walls acted as ligands for the Ir complex. The formation of the Ir complex using the bipyridine groups in the walls was also confirmed by ¹³C CP MAS NMR spectra. Following treatment with the Ir precursor, some of the signals due to the bipyridine groups are seen to have shifted (Figure 4(a)), and a similar shift is observed in the ¹³C NMR peaks of the bipyridine molecule after the formation of the IrCp*Cl(bpy) complex (Figure 4(c)). Cp* signals are also evident in the spectrum of the Ir complex, at 9 and 92 ppm. EDX analysis indicated that the Ir content of the Ir-BPy-NT was 0.32 mmol g⁻¹, suggesting that more than 90% of the bipyridine ligands in the BPy-NT formed a IrCp*Cl(bpy) complex. These results strongly suggest the homogeneous distribution of bipyridine groups in the tube walls.

The Ir-BPy-NT was subsequently used as a heterogeneous catalyst for a water oxidation reaction using Ce⁴⁺ (cerium(IV) ammonium nitrate) as the oxidant. Catalytic water oxidation using homogeneous metal complexes such as IrCp*Cl(bpy) has recently attracted significant attention because of the growing importance of photochemical energy storage with regard to fuel production.^23–25 The heterogenization of such metal complexes is crucial in developing advanced future practical applications of this technology. Figure 6(a) shows the time-dependent oxygen evolution curves obtained with varying amounts of the Ir-BPy-NT (7.5, 10, and 15 μM Ir) and the homogeneous IrCp*Cl(bpy) catalyst (10 μM Ir). In each case, the amount of O₂ gas produced in the head space of the reaction vessel was analyzed by a gas chromatography system equipped with a thermal conductivity detector and an active carbon (mesh 60/80, 3 m long) column. From these trials, the Ir-BPy-NT was found to effectively catalyze the water oxidation reaction to generate oxygen. In each case, the oxygen evolution curve is seen to gradually level off as the Ce⁴⁺ oxidant is consumed with a final O₂ yield of about 90% (corresponding to the limitation imposed by the amount of Ce⁴⁺ used). The initial turnover frequency (TOF) obtained with the Ir-BPy-NT was 3.1 min⁻¹ based on 1/4 O₂, as calculated by multiplying the slope (0.77 min⁻¹) of the initial oxygen evolution rate (over the first 5 min) against the concentration of the Ir-BPy-NT by 4 because the formation of a single O₂ molecule needs four Ce⁴⁺ (Figure 6(b)). This is only 18% of the TOF value (17.2 min⁻¹) obtained for the homogeneous catalyst; however, this TOF is also approximately eight times larger than that (0.4 min⁻¹) obtained using the bipyridine-based IrCp* complex fixed in a MOF with a pore diameter of 1 nm and evaluated under similar conditions (3 mM Ce⁴⁺ and 10 μM Ir).²⁰ The increased catalytic activity of the BPy-NT may be due to the easier access of Ce⁴⁺ to the active sites in the nanotubes, as a result of the larger pore diameters and shorter channel lengths associated with the tubes. Furthermore, the Ir-BPy-NT was readily recovered from the reaction mixture by centrifugation and could be used at least three times without any evident loss of activity (Figure 6(c)). The nanotube structure was maintained even after the third use, indicating that BPy-NT is sufficiently stable for application to the water oxidation reaction (Figure 7).

In summary, organosilica nanotubes containing bipyridine ligands in their walls were successfully obtained. SEM observations and nitrogen adsorption isotherms clearly showed that one-dimensional nanotube structures with pore diameters of 7.8 nm were obtained, while UV-vis and STEM measurements demonstrated the homogeneous distribution of bipyridine groups in the tube walls. Water oxidation catalysis by the immobilized Ir complex showed the significant potential of BPy-NT as a solid support for metal complex catalysts.

This work was supported in part by a Grant-in-Aid for Scientific Research on Innovative Areas ‘Artificial Photosynthesis’ (No. 2406) from the Japan Society for the Promotion of Science (JSPS).