Manganese-phosphomolybdate molecular catalysts for the electron transfer reaction of ferricyanide to ferrocyanide

Abstract

Three manganese-containing phosphomolybdate hybrids (H₂bpp)₅[Na(Hbpp)]₆H₁₀{Mn[Mo₆O₁₂(OH)₃(HPO₄)₄]₂}₄·14H₂O (1), [Na₄(H₂bpp)₂Mn(H₂O)₇]{Mn[Mo₆O₁₂(OH)₃(HPO₄)₃(PO₄)]₂}·2H₂O (2), and Na(H₂O)₂(Hbpp)₃[Na₂(bpp)(H₂O)][Mn₂(H₂O)₅]{Mn[Mo₁₂O₂₄(OH)₆(HPO₄)₆(H₂PO₄)(PO₄)]}·(HPO₄)·H₂O (3) (bpp = 1,3-bi(4-pyridyl)-propane) have been constructed and characterized. The inorganic moieties of the three hybrids consist of ‘hourglass-shaped’ anionic clusters, composed of two reduced polymolybdenum phosphate units [P₄^VMo₆^VO₂₈(OH)₃]₉ {P₄Mo₆} bridged by one manganese ion. Preliminary experiments show that these hybrids, as a unique class of molecular catalyst, are highly active for promoting the inorganic electron transfer (redox) reaction of ferricyanide to ferrocyanide by thiosulphate with high rate constants under mild conditions. These catalysts maintain their structural identity both in solution and solid state and can be easily separated from the reaction solution for the next catalytic cycle.

---

1. Introduction

In recent years, researchers have shown great interest in metal nanoparticles (NPs) because of their special nature and their applications in many fields such as catalysis, photonics, electronics, biomedicine and biosensors.^1–8 The major disadvantages of using metal NPs as catalysts are their aggregation during the catalytic process and their separation after the reaction ends (which also causes loss). These problems cause severe restrictions in using costly metal NPs as catalysts for large scale applications. So the search for new materials-based catalysts with structural stability, as well as easy separation and good reusability for several cycles, is ongoing. Recently, metal NPs embedded in mesoporous hosts in the form of powder/bulk have shown some improvements,^9,10 however, the separation and reuse of the catalysts was still tough. The pursuit of more environmentally friendly and economical catalytic systems is still a significant, but challenging, goal.

Polyoxometalates (POMs) are an important type of inorganic building block with special and stable nanoscale structures which have been used to construct novel hybrid materials with specific properties. As an important polyanion in the POM family, phosphomolybdate units {P₄Mo₆X₃₁} (X = O or OH) (abbreviated as {P₄Mo₆}) have attracted a great deal of researchers’ attention due to their unique reduced structural features.^11–16 It is well known that a significant property of POMs is their ability to reversibly accept several electrons and therefore to act as electron reservoirs. The unique reduced structure of {P₄Mo₆}, with the d¹ electronic configuration of Mo^V, can be expected to be useful in such electron transfer reaction catalysts. Our recent research has focused on POM-based supramolecular assemblies by utilizing the noncovalent weak forces that occur between the surface oxygen atoms of POMs and organic molecules.^17–22 A feature of these hybrids is their good structural stability both in water and common organic solvents, which is convenient for their applications as heterogeneous catalysts. With these considerations in mind, three functional analogs²³ have been isolated. Their formulae are: (H₂bpp)₅[Na(Hbpp)]₆H₁₀{Mn[Mo₆O₁₂(OH)₃(HPO₄)₄]₂}₄·14H₂O (1), [Na₄(H₂bpp)₂Mn(H₂O)₇]{Mn[Mo₆O₁₂(OH)₃(HPO₄)₃(PO₄)]₂}·2H₂O (2), and Na(H₂O)₂(Hbpp)₃[Na₂(bpp)(H₂O)][Mn₂(H₂O)₅]{Mn[Mo₁₂O₂₄(OH)₆(HPO₄)₆(H₂PO₄)(PO₄)]}·(HPO₄)·H₂O (3) (bpp = 1,3-bi(4-pyridyl)-propane). Experimental results indicate that the hybrids 1–3 are highly active molecular catalysts for promoting the inorganic electron transfer (redox) reaction of ferricyanide to ferrocyanide by thiosulphate. Since this class of hybrids is insoluble in the reaction medium, the catalysts can be easily filtered after the reaction is complete. The stability and reusability of these hybrids is also investigated.

2. Experimental section

2.1 Materials and measurements

All the reagents were commercially purchased and used without further purification. Hydrothermal synthesis was carried out using a 20 mL Teflon-lined autoclave under autogenous pressure. The reaction vessels were filled to approximately 70% of their volume capacity. Elemental analyses of C, H, N were carried on a Perkin-Elmer 2400 CHN elemental analyzer. FTIR spectra were recorded using KBr pellets with a FTIR-8900 IR spectrometer in the range of 400–4000 cm⁻¹. TG analysis were performed on a Perkin-Elmer Pyris Diamond TG/DTA instrument in flowing N₂ with a heating rate of 10 °C min⁻¹. Powder X-ray diffractions (XRD) were determined using a Bruker AXS D8 Advance diffractometer. UV spectra were measured using a U3010 UV-visible spectrophotometer (Shimadzu).

2.2 Preparation of hybrids 1–3

A mixture of MnCl₂·4H₂O (0.08 g, 0.40 mmol), bpp (0.03 g, 0.15 mmol), Na₂MoO₄·2H₂O (0.24 g, 3.31 mmol), H₃PO₄ (0.5 mL, 7.50 mmol), and H₂O (8 mL, 0.72 mmol) was stirred for 30 min at room temperature, the pH was adjusted to approximately 1.2 with NaOH, then the solution was transferred to a Teflon-lined reactor and heated at 160 °C for 5 days. The reactor was cooled to room temperature at a rate of 8 °C h⁻¹. Pale brown crystals of 1 were obtained, which were washed with distilled water and air-dried to give a yield of 52% based on Mo. Anal. calcd. for C₁₄₃H₂₆₄Mn₄Mo₄₈N₂₂Na₆O₂₆₂P₃₂ (%): C, 27.5; H, 2.96; N, 4.92; found: C, 27.7; H, 2.84; N, 4.97.

Hybrids 2 and 3 were prepared through the similar procedure to that of hybrid 1, except that the pH values were different. The mixture was adjusted with NaOH solution to ca. pH = 2.5 for 2, and pH = 4.5 for 3. Deep brown–red crystals of 2 (yield: 42% based on Mo) and red block crystals of 3 (yield: 48% based on Mo) were obtained. Elemental anal. calcd. for 2 C₂₆H₆₂Mn₂Mo₁₂-N₄Na₄O₇₁P₈ (%): C, 9.24; H, 1.64; N, 1.69; found: C, 9.88; H, 1.65; N, 1.78, and for 3 C₅₂H₉₂Mn₃Mo₁₂N₈Na₃O₇₅P₉ (%): C, 16.54; H, 1.87; N, 3.16; found: C, 16.98; H, 1.90; N, 3.05.

2.3 X-ray crystallographic study

Single crystal data were collected on a Smart Apex CCD diffractometer at 296(2) K with Mo Kα monochromatic radiation (λ = 0.71073 Å) at room temperature. The structures of 1–3 were solved by direct methods and refined by the full-matrix least-squares methods on F² using the SHELXTL crystallographic software package.²⁴ Anisotropic thermal parameters were used to refine all non-hydrogen atoms. The positions of hydrogen atoms attached to carbon atoms were fixed at the ideal positions. Hydrogen atoms attached to nitrogen or oxygen atoms were located from the difference Fourier maps and were constrained during the subsequent refinement calculation. A summary of the crystallographic data and structural determination parameters for 1–3 are provided in ESI Table 1.†

2.4 Catalytic experiments

The catalytic reactions were carried out in a Pyrex reactor of 80 mL capacity. A 20 mg sample of the hybrid 1 (or 2 or 3) was dispersed into aqueous solution. Here, experiments consisting of K₃[Fe(CN)₆] (1.0 × 10⁻³ M) and Na₂S₂O₃ (1.3 × 10⁻¹ M) solutions were performed. The system was magnetically stirred at 50 °C. At 15, 30, 60, 90, 150, and 210 min, a 3 mL aliquot was sampled and then centrifuged to remove any particles of catalyst. The evolution in the absorption spectra due to the reduction of [Fe(CN)₆]³⁻ to [Fe(CN)₆]⁴⁻ was recorded using UV-visible spectrophotometry. Spectra of blank experiments in the absence of any catalyst were also recorded. To test the reusability of the hybrid, the used catalysts were separated from the solution by centrifuge, then washed and dried, ready for use in the next cycle. The ionic equation for the redox reaction is shown in eqn (1).

2S₂O₃²⁻ + 2[Fe(CN)₆]³⁻ ↔ S₄O₆²⁻ + 2[Fe(CN)₆]⁴⁻ (1)

The conversion of Fe(III) is calculated from ([C₀] − [C_i])/C₀ × 100%, where [C₀] is the initial concentration and [C_i] is the concentration at each time point.

3. Results and discussion

3.1 Structural analysis

Single crystal X-ray diffraction analyses reveal that the basic moiety of the hybrids 1–3 is a sandwich-type anionic unit {P₄Mo₆X₃₁}ⁿ⁻ (abbr. {P₄Mo₆}, X = O or OH). As is observed usually, the unit is made up of six {MoO₆} octahedra, four {PO₄} tetrahedra and one {MnO₆} octahedron which links by corner- and edge-sharing modes. According to valence bond calculations,²⁵ all Mo atoms are in the reduced +5 oxidation state (5.140–5.301). P and Mn atoms are in +5 and +2 oxidation states (Table S2–S4†), respectively.

The basic composition of the hybrid 1 consists of four {Mn[P₄Mo₆]₂} clusters, six [Na(Hbpp)]²⁺, five protonated [H₂(bpp)]²⁺ cations, and fourteen lattice water molecules. In the {Mn[P₄Mo₆]₂} cluster, the central Mn atom adopts an octahedral coordination mode which is achieved by binding to six bridging oxygen atoms between Mo–Mo single bonds of the {Mn[P₄Mo₆]₂} cluster (Fig. 1a). There are two kinds of discrete polyanionic {Mn(P₄Mo₆)₂} units in 1, labeled as Mn(13)- and Mn(14)-containing clusters (Fig. 1b). The Mn–O bond lengths are 2.173–2.259 Å. Na⁺ cations are attached to the surface O atoms of the anions via electrostatic forces, forming a 2D inorganic sheet-like structure (Fig. 1c). The cationic [Na(Hbpp)]²⁺ units, protonated bpp, PO₄³⁻ anions, and solvent water molecules reside in the interspace between two adjacent supramolecular layers, stabilised by intermolecular interactions.

Fig. 1 (a) The {Mn(P₄Mo₆)₂} unit in 1; (b) the simplification of six edge-sharing {MoO₆} octahedra; (c) a polyhedral view showing the 2D inorganic sheet linked through O–Na–OH₂–Na–O; (d) the 3D supramolecular framework showing the alternative arrangements of the inorganic and organic layers through bpp-Na–O and C–H⋯O linkages.

The hybrid 2 consists of a chain-like anionic unit [Na₄Mn(H₂O)₇]{Mn[Mo₆O₁₂(OH)₃(HPO₄)₃(PO₄)]₂}¹⁰⁻, sodium and protonated [H₂(bpp)]²⁺ cations, and water molecules. Mn(2) is located at the center of the sandwich-type unit {Mn[P₄Mo₆]₂} while Mn(1), in the form of {Mn(OH₂)₂} fragments, bridges {Mn[P₄Mo₆]₂} clusters to form a 1D inorganic extending structure (Fig. 2a). Worthy of mention is the existence of Na⁺ ions, which further stabilize and extend the inorganic network. As shown in Fig. 2b, four distorted {NaO₆} octahedra surround one Mn(1) octahedron in an edge-sharing mode to form a {Na₄MnO₂} unit. In 2, it can be seen that the adjacent parallel inorganic chains construct a 2D inorganic sheet, which are perpendicular to other chains from adjacent sheets (see Fig. 2c). The protonated bpp and water molecules fill the space formed by the stacking of these inorganic chains.

Fig. 2 (a) The 1D structure of {Mn[Mn(P₄Mo₆)₂]₂}_∞ in 2; (b) the 2D inorganic sheet formed by Na–O interactions; (c) the 3D supramolecular framework of 2.

Similarly, hybrid 3 is composed of {Mn[Mo₁₂O₂₄(OH)₆(HPO₄)₆(H₂PO₄)(PO₄)]}⁸⁻, sodium cations, [Na₂(bpp)]²⁺, [H(bpp)]⁺, isolated [HPO₄]²⁻ ions and lattice water molecules. The difference is that, here, the clusters of {Mn[P₄Mo₆]₂} are linked by double bridge linkages of {Mn(OH₂)₂}₂ (Mn(1) and Mn(1i)). As shown in Fig. 3, all the adjacent inorganic 1D chains are parallel to each other to form 2D inorganic sheets by Na–O linkages. Furthermore, the adjacent 2D sheets are also parallel to each other in the same extending directions and can form a 3D network of 3 through O–Na-bpp-Na–O and hydrogen bonding linkages.

Fig. 3 (a) 1D chain-like structure of {Mn₂[(P₄Mo₆)₂]}_∞ in 3; (b) 2D inorganic sheet; (c) 3D supramolecular framework of 3. The isolated [HPO₄]²⁻ ions, water molecules and hydrogen atoms are omitted for clarity.

3.2 FT-IR spectra, TG analyses, XRD and UV spectra

IR spectra of 1–3 exhibit similar characteristic peaks (see Fig. S1(a–c)†). The characteristic bands in the range of 606–745 cm⁻¹ are attributed to ν(Mo–O–Mo) vibrations. Strong peaks at 956–966 cm⁻¹ are due to ν(Mo=O) vibrations. The ν(P–O) vibration ranges from 1033–1178 cm⁻¹. Strong peaks at 1504–1622 cm⁻¹ are assigned to the C–C and C–N stretches of organic amines. In addition, broad bands at 3247–3440 cm⁻¹ are due to ν(O–H) and ν(N–H). The thermal stabilities were investigated under a N₂ atmosphere from 20 to 800 °C. The results show that the main structures of 1–3 are stable up to ca. 300 °C (Fig. S2†). The simulated and experimental XRD patterns of 1–3 in the angular region 5–55° are shown in Fig. S3.† One can see that the diffraction peaks of the simulated and experimental patterns of 1–3 match in the key positions indicating the phase purity of three compounds. The differences in intensity might be due to the preferred orientation of the powder samples. The UV spectra of the hybrids 1–3 in methanol were measured and are provided in the ESI (Fig. S4).† One absorption peak at 259 nm in the UV region is assigned to the charge-transfer absorption band from terminal and bridging oxygen atoms to Mo metal atoms.

3.3 Catalytic experiments

The highly reduced nature and structure of the hybrids 1–3 motivates us to further study their catalytic performance for electron transfer reactions. The hybrids 1–3 were tested for their heterogeneous catalytic properties for the inorganic redox reaction between Fe(CN)₆³⁻ and S₂O₃²⁻ in aqueous solution at room temperature (25 °C) and at 50 °C. For this study, the optical spectra of the respective solutions in the presence of catalysts were monitored over time by UV-visible spectrometry. The depletion of the 420 nm peak was used to study the rate of the catalyzed reaction, and the results are presented in Fig. 4. It shows the successive UV-vis spectra of the solutions recorded during the redox reaction between Fe(CN)₆³⁻ and S₂O₃²⁻, with a change in color from yellow to light green (see Fig. S5†). As shown in Fig. 4a–c, a depletion in the Fe(CN)₆³⁻ peaks was observed as the reaction proceeded in the presence of the hybrids 1–3. For comparison, a similar test in the absence of any of the hybrids 1–3 was also performed, and only a very small change in absorption values (Fig. 4d) was observed over a period of 180 min at 50 °C. The results clearly show the catalytic activity of the hybrids 1–3 for the reduction of [Fe(CN)₆]³⁻ to [Fe(CN)₆]⁴⁻ ions. It may be noted here that the reaction proceeds through a clear isosbestic point at 286 nm, indicating that this redox reaction proceeds smoothly without forming multiple products. As shown in Fig. 5, the apparent rate constant, calculated from the slope of the plot of −ln A₄₂₀ against time, is found to be 1.5 × 10⁻³ min⁻¹ (2.5 × 10⁻⁵ s⁻¹) at 25 °C. But at 50 °C the apparent rate constant is calculated from the slope to be 8.5 × 10⁻³ min⁻¹ (1.41 × 10⁻⁴ s⁻¹), which is five times higher than at 25 °C. These data are comparable with that of the reduction of [Fe(CN)₆]³⁻ in the presence of Au NPs doped onto mesoporous boehmite films at the same temperatures (25 °C, 3.12 × 10⁻³ min⁻¹; 50 °C, 8.43 × 10⁻³ min⁻¹).²⁶

Fig. 4 Successive UV-visible absorption spectra of the aqueous solutions of the catalytic reduction reaction of [Fe(CN)₆]³⁻ and S₂O₃²⁻ in the presence of 1(a), 2(b), and 3(c). (d) is the blank experiment without any catalyst at 50 °C.

Fig. 5 Pseudo first order plots of −ln A (absorbance intensity at 420 nm) versus time for the Fe(III) reaction at 25 °C and 50 °C.

To experimentally verify that the observed reaction is a truly catalytic process, a blank experiment was designed and performed. The reaction does not occur in the absence of catalyst at these temperatures. There are two possibilities for the catalytic reduction reaction: (i) the reduced POM cluster serves as a reductant; (ii) the reduced POM cluster serves as a catalyst for the reaction of K₃[Fe(CN)₆] and Na₂S₂O₃. In order to investigate whether or not surface catalysis takes place, solutions initially containing 1 mM K₃[Fe(CN)₆] were monitored as a function of time in the presence and absence of Na₂S₂O₃ and the hybrid 1. Fig. 6 reports the results of these experiments performed in aqueous solution. Curve a consists of two stages at 50 °C: the concentration of hexacyanoferrate(III) does not change in only the presence of the hybrid 1 under similar conditions over nearly 350 min, indicating that no hexacyanoferrate(III) is adsorbed onto the surface of 1. When the reductant Na₂S₂O₃ is added, the rapid decrease in [Fe^III]_aq over time means that Fe^III is quickly transformed to Fe^II. These results suggests that the hybrid 1 itself cannot adsorb and cannot reduce K₃[Fe(CN)₆], but can only catalyze the reduction reaction of hexacyanoferrate(III) with Na₂S₂O₃. Curve b indicates that there is a little surface adsorption of Fe^III onto crystals at 25 °C, and that the maximum adsorption amount is ca. 13.2%. The adsorption process of Fe^III onto the hybrid 1 is slow and equilibrium is obtained within 250 min. Curves c and d represent experiments in which both Na₂S₂O₃ and the hybrid 1 are present at 25 °C and 50 °C, respectively. Combining Fig. 5 and 6, it is evident that the effect of temperature enhancing the catalytic ability is significant under similar conditions. This can be ascribed to the increased solvability of the hybrids at a higher temperature, leading to a change from a heterogeneous to a homogeneous process.

Fig. 6 Reduction of 1 mM Fe(III) by the hybrid 1 as a function of time in: (a) Na₂S₂O₃-free solution for 350 min and then adding Na₂S₂O₃, at 50 °C; (b) Na₂S₂O₃-free solution for 350 min at 25 °C; (c) and (d) the presence of both Na₂S₂O₃ and 1 at 25 °C and 50 °C, respectively.

Curves a and b indicate that it is the crystal surface that serves as the catalytic active sites rather than the POM cluster itself acting as a reducer. Here, the hybrid 1 is used as an example to superficially study the kinetics. The overall catalytic redox reaction may be expressed using eqn (2)–(6). In the derivation, the elementary steps are assumed as below (to simplify the expression, [A] represents [[Fe(CN)₆]³⁻], [B] represents [Na₂S₂O₃], [*] represents the concentration of vacant active chemisorption sites on the solid catalysts surface, [A*], [B*], [C*] and [P*] are those of the reaction intermediates, and [P] represents the product [Fe(CN)₆]⁴⁻ ion).²⁷

The hybrids 2 and 3 were also tested for their heterogeneous catalytic properties for this system (Fig. 7). Additionally, a Keggin-type anion-containing hybrid (H₂bpp)₂[PMo₉^VIMo₃^VO₄₀] with mixed valence Mo was conducted as a contrast experiment to check its activity.²⁸ It was observed that this hybrid was ineffective for the reduction reaction (see Fig. S6†). The simple phosphomolybdic acid H₃PMo₁₂O₄₀ does not show catalytic activity for this system, but is soluble in water and quickly reacts with the reducing agent Na₂S₂O₃ to form the secondary pollutant heteropoly blue (Fig. S7†). Compared to the blank experiment, hybrids 1, 2 and 3 show high catalytic activity due to the recombination of the inorganic anion {P₄Mo₆} and organic bpp. It could be concluded that the {Mn[P₄Mo₆]₂} cluster in 1–3 plays a critical role in the reduction reaction of Fe^III, and that the extended structural topologies of the Mn coordination environment are also related to the catalytic performance. Because the reduction step is an electron transfer process, the reported work also confirmed that the electron transfer of the catalytic reaction occurred at the noble metal NP surface. A highly reduced polyanion cluster {Mn[P₄Mo₆^V]} might be more conducive to electron transfer in a much faster manner.

Fig. 7 Plots of conversion of Fe(III) vs. time at 50 °C.

A POM-based crystalline tubular material IFMC-100 was employed as a matrix/microreactor to prepare a Au-anchored material (Au@IFMC-100) by a solid–liquid redox reaction. The as-prepared Au@IFMC-100 microtubes exhibit enhanced catalytic performance for the reduction of K₃Fe(CN)₆ with NaBH₄ in aqueous solution.²⁹ However, thiosulphate salts are a National Sanitation Foundation (NSF, a WHO collaborator) approved water treatment chemical. The development of a catalytic material that could utilize thiosulphate as a reductant for the electron transfer reaction would have significant environmental benefits.³⁰ During our course of catalytic experiments, the POM-based supramolecular frameworks maintained higher catalytic activity and structural stability. Importantly, the catalyst could be easily separated after the reaction and reused several times. The catalytic reaction was performed multiple times using the same sample. It is found that the conversion rate of Fe³⁺ becomes lower with more repeated cycles (see Table 1), which may be ascribed to the loss of some crystals with multiple cycles. In addition, a possible reason is that the surface of the crystal sample adsorbs a lot of poorly soluble Fe²⁺, which impedes the contact of the catalyst and Fe³⁺ and reduces the catalytic activity. IR spectra of the fresh and used catalysts were measured to check the structural stability of the hybrid 1 (Fig. S8†). The characteristic peaks illustrate that the skeleton of the parent {P₄Mo₆} polyanion still remains intact after the catalytic reaction. Furthermore, X-ray photoelectron spectra were measured to monitor the oxidation state of Mo in the hybrid 1 before and after catalysis (see Fig. S9†). The two curves for the Mo atoms gave two peaks at 231.4 and 234.7 eV, attributable to Mo⁵⁺3d_5/2 and Mo⁵⁺3d_3/2, respectively. The results support the fact that the Mo^V ions are stable in their reduced oxidation state (+5) during the catalytic process. The advantages of using such POM hybrid catalysts are therefore manyfold: (i) the catalysts can be easily separated after the reaction is finished, and be reused after washing with water; (ii) the spectra of the solution during the course of the catalytic reaction can be monitored without any overlapping absorption originating from the catalyst; (iii) these hybrid systems are cheap and easily obtained, and in particular, they can be structurally designed to optimize their properties and applications in environmentally friendly catalytic reactions.

Table 1 Catalytic electron transfer (redox) of ferricyanide to ferrocyanide by the hybrids 1–3 for different cycles

	Conversion (%)
	Fresh	Cycle 1	Cycle 2	Cycle 3
Hybrid 1	63.5	60.7	53.4	40.0
Hybrid 2	77.5	71.2	64.5	56.3
Hybrid 3	81.4	64.6	50.8	46.1

---

4. Conclusions

In this paper, three reduced molybdophosphate hybrids based on {P₄Mo₆} building blocks have been constructed and were tested for their catalytic electron transfer performance of ferricyanide to ferrocyanide. The results indicated that this unique class of hybrids, as heterogeneous molecular catalysts, has good performance for the electron transfer reaction of hexacyanoferrate(III). The high catalytic performance of the title hybrids may make them excellent candidates for the catalysis of materials undergoing electron transfer reactions. Compared with noble metal Au/Pt/Pb NPs, the synthesis of 1–3 is cheaper and more straightforward. Importantly, the catalysts can be easily separated and reused several times. The current work might provide a new idea for exploring POM-based materials in catalytic chemistry. Further work on these hybrids and reaction systems in this field is underway.

Acknowledgements

This work was financially supported by the Natural Science Foundation of China (21341003), the Hebei Natural Science Foundation of China (no. B2015205116), and the Key Research Fund of Hebei Normal University (no. L2011Z06).

Notes and references

1. I. J. Buerge and S. J. Hug, Environ. Sci. Technol., 1998, 32, 2092.
2. P. Watts and C. Wiles, Chem. Commun., 2007, 443.
3. B. Deng and A. T. Stone, Environ. Sci. Technol., 1996, 30, 2484.
4. K. Bhowmik, A. Mukherjee, M. K. Mishra and G. De, Langmuir, 2014, 30, 3209.
5. W. Huang, J. N. Kuhn, C. K. Tsung, Y. Zhang, S. E. Habas, P. Yang and G. A. Somorjai, Nano Lett., 2008, 8, 2027.
6. B. Lim, M. Jiang, P. H. C. Camargo, E. C. Cho, J. Tao, X. Lu, Y. Zhu and Y. Xia, Science, 2009, 324, 1302.
7. H. L. Jiang, B. Liu, T. Akita, M. Haruta, H. Sakurai and Q. Xu, J. Am. Chem. Soc., 2009, 131, 11302.
8. X. Huang, P. K. Jain, I. H. El-Sayed and M. A. El-Sayed, Nanomedicine, 2007, 2, 681.
9. J. Sun and X. Bao, Chem.–Eur. J., 2008, 14, 7478.
10. M. Boutros, A. Denicourt-Nowicki, R. A. Alain, L. Gengembre, P. Beaunier, A. Gedeon and F. Launay, Chem. Commun., 2008, 2920.
11. M. T. Pope, Heteropoly, Isopoly Oxometalates, Springer-Verlag, Berlin, 1983.
12. D. L. Long, E. Burkholder and L. Cronin, Chem. Soc. Rev., 2007, 36, 105.
13. C. Streb, D. L. Long and L. Cronin, CrystEngComm, 2006, 8, 629.
14. K. Yu, W. L. Chen, B. B. Zhou, Y. G. Li, Y. Yu, Z. H. Su, S. Gao and Y. Chen, CrystEngComm, 2011, 13, 3417.
15. X. Zhang, J. Q. Xu and J. H. Yu, J. Solid State Chem., 2007, 180, 1949.
16. X. Xu, W. W. Ju, D. W. Yan, N. G. Jian and Y. Xu, J. Coord. Chem., 2013, 66, 2669.
17. Z. G. Han, Y. G. Gao, X. L. Zhai, J. Peng, A. X. Tian, Y. L. Zhao and C. W. Hu, Cryst. Growth Des., 2009, 9, 1225.
18. X. L. Xue, X. F. Zhao, D. S. Zhang, Z. G. Han, H. T. Yu and X. L. Zhai, RSC Adv., 2014, 4, 63670.
19. Z. G. Han, X. Q. Chang, J. S. Yan, K. N. Gong, C. Zhao and X. L. Zhai, Inorg. Chem., 2014, 53, 670.
20. J. S. Yan, X. F. Zhao, J. Huang, K. N. Gong, Z. G. Han and X. L. Zhai, J. Solid State Chem., 2014, 211, 200.
21. J. S. Yan, K. N. Gong, X. L. Xue, X. L. He, C. Zhao, Z. G. Han and H. T. Yu, Eur. J. Inorg. Chem., 2014, 5969.
22. Z. G. Han, Y. N. Wang, X. J. Song, X. L. Zhai and C. W. Hu, Eur. J. Inorg. Chem., 2011, 3082.
23. K. N. Gong, W. J. Wang, J. S. Yan and Z. G. Han, J. Mater. Chem. A, 2015, 3, 6019.
24. G. M. Sheldrick, SHELXTL-97, Programs for Crystal Structure Refinement, University of Göttingen, Germany, 1997.
25. D. Jana, A. Dandapat and G. De, Langmuir, 2010, 26, 12177.
26. I. D. Brown and D. Altermatt, Acta Crystallogr., Sect. B: Struct. Sci., 1985, 41, 244.
27. C. X. Yang, J. H. Meldona, B. Lee and H. Yi, Catal. Today, 2014, 233, 108.
28. Crystal data of (H₂bpp)₃[PMo₉^VIMo₃^VO₄₀]: C₃₉H₄₅Mo₁₂-N₆O₄₂P, Mr = 2452.06, monoclinic, space group Cc, a = 23.74 (Å), b = 12.68 (Å), c = 22.25 (Å); α = 90°, β = 105.81°, γ = 90°. V = 6408.7 (ų), Z = 4, goodness-of-fit on F² 1.043, final R indices [I > 2σ(I)], R₁ = 0.0256, wR₂ = 0.0553.
29. D. Y. Du, J. S. Qin, T. T. Wang, S. L. Li, Z. M. Su, K. Z. Shao, Y. Q. Lan, X. L. Wang and E. B. Wang, Chem. Sci., 2012, 3, 705.
30. A. Pearson and P. O’Mullane, ChemPlusChem, 2013, 78, 1343.

---

Footnote

† Electronic supplementary information (ESI) available: Additional IR spectra, TG curve, PXRD and UV spectra of reaction solutions. CCDC 1015040–1015042. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c5ra06018d

---