Self-assembly of [PNb_xW_12−xO₄₀]ⁿ⁻ Keggin anions – a simple way to mixed Nb–W polyoxometalates

Abstract

This paper summarizes preparations of mixed phosphoniobotungstates [PNb_xW_12−xO₄₀]^(3+x)− (x = 1–3) in two different ways: (i) incorporation of Nb into pre-existing lacunary phosphotungstates and (ii) self-assembly reactions. A niobium oxalate complex as the source of niobium is used. The reaction products were monitored by ³¹P and ¹⁸³W NMR in solution as well as by ³¹P MAS NMR techniques for precipitated samples. Single-substituted species [PNbW₁₁O₄₀]⁴⁻ forms rapidly and selectively from lacunary [PW₁₁O₃₉]⁷⁻ with one equivalent of [NbO(C₂O₄)₂(H₂O)₂]⁻. With the increasing amount of the niobium oxalate the highest number of Nb atoms was found to be three per Keggin anion, but the disubstituted [PNb₂W₁₀O₄₀]⁵⁻ was always the preferable product whatever the tungsten reagent was. The reaction products with x = 2 and 3 are present as mixtures of all possible isomers, but the ratio between the individual isomers depends on the synthetic pathway.

---

Introduction

Of the polyoxoniobate family mixed-addendum Nb/W POMs, explored by Klemperer,¹ Hill,^2–6 and Finke^7–9 groups, Lindqvist-type [Nb_xW_6−xO₁₉]^(2+x)−,^10–12 Keggin-type [XW₉Nb₃O₄₀]ⁿ⁻ (X = Si^IV, Ge^IV, n = 7; X = P^V, As^V, n = 6)^2,3,13–15 and Dawson-type [P₂W₁₂(NbO₂)₆O₅₆]¹²⁻,¹⁶ [P₂W₁₅Nb₃O₆₂]⁹⁻,^7,17 and [P₂W₁₇(NbO₂)O₆₁]^7− 6 merit particular attention because of their unique characteristics. Due to the presence of Nb, these POMs possess higher nucleophilicity than the parent tungstates.^15,18,19 Such complexes demonstrate photocatalytic activity in oxidation⁵ and reduction²⁰ processes, and can be used as precursors for hydrodesulfurization/isomerization catalysts¹² or as active catalysts in the preparation of linearly-extended polyalkylenepolyamines.²¹

The intrinsic high negative charge density of [Nb₆O₁₉]⁸⁻ and other polyniobates can be reduced by the substitution of one or more Nb^V addenda sites with W^VI, rendering these POMs less basic and more stable in neutral or acidic media. This goal can be achieved in various ways. Dabbabi and Boyer reported that Na₂WO₄ and K₈[Nb₆O₁₉] in aqueous H₂O₂ yield the mixed metal complex [NbW₅O₁₉]³⁻.²² This compound is stable at pH 1.5–5 and gradually transforms into [cis-Nb₂W₄O₁₉]⁴⁻, [fac-Nb₃W₃O₁₉]⁵⁻, [cis-Nb₄W₂O₁₉]⁶⁻, [Nb₅WO₁₉]⁷⁻, and [Nb₆O₁₉]⁸⁻ with increasing alkalinity. The isomeric purity of [cis-Nb₂W₄O₁₉]⁴⁻, [fac-Nb₃W₃O₁₉]⁵⁻, and [cis-Nb₄W₂O₁₉]⁶⁻ was claimed from infrared studies²³ and further confirmed, for [cis-Nb₂W₄O₁₉]⁴⁻, by electrochemical measurements²⁴ as well as by ¹⁷O and ¹⁸³W NMR spectroscopy.^10,25 The other traditional way of making the Nb/W mixed POMs consists of reacting peroxoniobates prepared in situ by solubilizing [Nb₆O₁₉]⁸⁻ in H₂O₂ with lacunary polyoxotungstates or sodium tungstate. This methodology was extended to produce first Ta/W POMs.¹⁸ The peroxoniobates have good stability in neutral to acidic media owing to the presence of the strongly coordinating η²-peroxide ligands, which makes them compatible with lacunary POMs, unlike simple oxoniobates, which can be handled only in alkaline solutions where polytungstates are unlikely to survive.²⁶ However, the use of peroxoniobates still presents several drawbacks, such as: (i) the excess of H₂O₂ destroys polytungstate precursors, and careful titration of peroxide with sodium sulfite is needed to avoid this;^27,28 (ii) to destroy the niobium peroxo complexes prolonged heating is needed which is not always desirable; (iii) the control of the stoichiometric ratio Nb : W is complicated by some Nb ending up as insoluble Nb₂O₅. To avoid these difficulties, we have set out to develop an alternative approach based on the use of a niobium oxalate complex, (NH₄)[NbO(C₂O₄)₂(H₂O)₂]·3H₂O, as a Nb precursor. The oxalate complex offers the advantages of being a well-defined stoichiometric solid, easily soluble in water and stable enough in solution even at a pH range between 1 and 3.²⁹ We have found the only example of the use of niobium oxalate, namely, for the preparation of mixed Nb–W catalysts for decomposition of aromatics.³⁰ In this work we used niobium oxalate for incorporation of Nb into lacunary Keggin phosphotungstates, and for self-assembly reactions with phosphate and tungstate, leading to mixed [PNb_xW_12−xO₄₀]ⁿ⁻ anions (x = 1–3).

Results and discussion

V and Nb-substituted heteropolytungstates

Vanadium substituted Keggin-type heteropolyanions [XV_xM_12−xO₄₀]ⁿ⁻ have been investigated as catalysts for selective oxidation of acids, aldehydes and hydrocarbons, oxidative dehydrogenation of alkanes etc.^31–35 Vanadium for tungsten substitution in a Keggin anion does not affect its overall structure. Since X-ray diffraction analysis does not show any preferential position of vanadium in the Keggin structure, one can only obtain information about the cation site occupancy by analytical determination of the M : V ratio.^34,36,37 Extra vanadium can also enter the Keggin-type anions by capping one or two rectangular faces of the Keggin polyhedra with {VO} caps.^37,38

The niobium substituted Keggins have been studied on a more modest scale. A few of them, [XW₉Nb₃O₄₀]ⁿ⁻ (X = Si^IV, Ge^IV, n = 7; X = P^V, As^V, n = 6),^9,39,40 [XW₉O₃₇(NbO₂)₃]ⁿ⁻,^4,5 [γ-SiW₁₀O₃₈(NbO₂)₂]⁶⁻,⁴¹ and [XNbW₁₁O₄₀],^5–42 have been obtained tested in alkene epoxidation, but their catalytic activity was found to be much lower than that of the vanadium analogues. However, the terminal oxide ligand at Nb seems to be more labile than that at W in substitution reactions, which offers some unexpected possibilities of making POMs more “soft”, in the HSAB theory sense. Thus Sécheresse et al. have shown that [PW₁₁NbO₄₀]⁴⁻, obtained by addition of a NbO³⁺ group from the controlled hydrolysis of NbCl₅ into the monovacant α-[PW₁₁O₃₉]⁷⁻, selectively exchanges its terminal oxo group in the reaction with p-methoxyphenylthionophosphine sulfide (R₂P₂S₄), with the formation of α-[PW₁₁NbSO₃₉]⁴⁻, with a sulfide replacing oxide.⁴³ There remain also some general problems to be answered, such as the maximum number of Nb atoms entering a given Keggin structure, isomer formation and distribution etc.

Synthesis of P-Nb_xW_12−x Keggin anions from lacunary type POM

Reactions between (NH₄)[NbO(C₂O₄)₂(H₂O)₂]·3H₂O (Nb-Ox) and the trilacunary POMs, [A-PW₉O₃₄]⁹⁻, γ-[PW₁₀O₃₆]⁷⁻, and [PW₁₁O₃₉]⁷⁻, were carried out in water by addition of a calculated amount of Nb-Ox to a solution of the POM, and heating the final solution for 30 min at 80 °C at pH 3. Cs salts for ³¹P MAS NMR were precipitated with CsCl. The white precipitates were collected by filtration, extensively rinsed with water and dried in vacuo before the NMR experiments.

The ³¹P MAS NMR spectra of the three products, together with the reference compound H₃PW₁₂O₄₀·nH₂O, are shown in Fig. 1. The multiple peaks in the range −14.9 to −15.1 ppm for H₃PW₁₂O₄₀·nH₂O result from different local environments of [PW₁₂O₄₀]³⁻ due to the non-uniform degree of hydration in various domains of the bulk sample. These chemical shifts agree with the published values for the solid state (−15 ppm)^44,45 and in solution (−14.4 ppm).⁴⁶ The spectra of the solids from the reactions of Nb-Ox with [PW₁₁O₃₉]⁷⁻ and [PW₁₀O₃₆]⁷⁻ (Fig. 1b and c) are almost the same and exhibit three resonances at −15.0 (3%), −13.6 (85%), and −12.7 ppm (12%). In the spectrum of the product from the reaction with [PW₉O₃₄]⁹⁻ (Fig. 1d) only two resonances are observed at −13.6 (5%) and −12.4 ppm (95%). Consequently, these two resonances at −13.6 and −12.4 ppm correspond to (Cs,K)₄[PNbW₁₁O₄₀] (−13.6 ppm) and (Cs,Na)₅[PNb₂W₁₀O₄₀] (−12.4 ppm). Note the systematic shift in the ³¹P resonance (Δδ ∼ 1.3 ppm) for [PNb_xW_12−xO₄₀]^(3+x)− with increasing x from 0 to 2. According to these results, with Nb-Ox stoichiometric (or quantitative) introduction of Nb into the Keggin phosphotungstate structure can be achieved only with [PW₁₁O₃₉]⁷⁻, while with [PW₁₀O₃₆]⁷⁻ and [PW₉O₃₄]⁹⁻ simultaneous incorporation of one and two Nb instead of strictly two and strictly three Nb atoms, respectively, expected, takes place. This means that partial decomposition and rearrangement of these bi- and trilacunary POMs occurs during the reaction. To gain further insight into these reactions, we followed them with conventional ³¹P NMR techniques in solution.

Fig. 1 ³¹P MAS NMR spectra of (a) H₃PW₁₂O₄₀·nH₂O, Cs salt of reaction products of Nb-Ox and (b) [PW₁₁O₃₉]⁷⁻, (c) [PW₁₀O₃₆]⁷⁻, and (d) [A-PW₉O₃₄]⁹⁻.

Fig. 2 shows the ³¹P NMR spectra of the reaction mixtures from Nb-Ox and K₇[PW₁₁O₃₉] (Fig. 2a) and Na₉[A-PW₉O₃₄] (Fig. 2b). It was difficult to perform the same experiment with Cs₇[PW₁₀O₃₆] due to its low solubility leading to a suspension rather than to a clear solution. The spectrum of the reaction mixture with K₇[PW₁₁O₃₉] (Fig. 2a) shows the dominant resonance line at −13.2 ppm assigned to [PNbW₁₁O₄₀]⁴⁻ and an unidentified minor signal at −12.4 ppm. The latter was already present in the spectrum of the K₇[PW₁₁O₃₉] precursor as an impurity and could correspond to [PW₁₀O₃₆]⁷⁻.⁴⁷¹⁸³W NMR of the reaction solution with K₇[PW₁₁O₃₉] (Fig. 3) confirms the formation of [PNbW₁₁O₄₀]⁴⁻ as the main product, since the six resonances at −77.1, −94.7, −99.5, −104.0, −105.5, and −112.0 ppm with relative intensities 2 : 2 : 2 : 1 : 2 : 2 are compatible with the C_s symmetry group of the monosubstituted Keggin structure. These NMR data agree with the published data for [PNbW₁₁O₄₀]⁴⁻ POM.^43,48,49

Fig. 2 ³¹P NMR spectra of the reaction mixture of (a) Nb-Ox and [PW₁₁O₃₉]⁷⁻, and (b) Nb-Ox and [A-PW₉O₃₄]⁹⁻.

Fig. 3 ¹⁸³W NMR spectrum of the reaction mixture of Nb-Ox and [PW₁₁O₃₉]⁷⁻.

However, reaction with Na₉[A-PW₉O₃₄] produces a complicated spectrum (Fig. 2b). The signal of [PNbW₁₁O₄₀]⁴⁻ and two sets of resonances for the isomers of [PNb₂W₁₀O₄₀]⁵⁻ (−11.7 to −12.0 ppm) and [PNb₃W₉O₄₀]⁶⁻ (−10.1 to −11.4 ppm) are clearly visible, together with the signal from free phosphate (0.6 ppm). The signal distribution corresponds to a mixture consisting of 10% of [PNbW₁₁O₄₀]⁴⁻, 58% of [PNb₂W₁₀O₄₀]⁵⁻, 12% of [PNb₃W₉O₄₀]⁶⁻, and 20% of free H_nPO₄⁽³⁻ⁿ⁾⁻. These observations match the ³¹P MAS NMR data for the Cs salts, indicating that: (i) [PNb₃W₉O₄₀]⁶⁻ is less favored than [PNb₂W₁₀O₄₀]⁵⁻, and (ii) structural decomposition/rearrangement takes place in the presence of excess of the Nb-Ox reagent. The starting [A-PW₉O₃₄]⁹⁻ is entirely consumed, judging from the absence of its ³¹P NMR signature at −5 ppm.⁵⁰

Self-assembly of P-Nb_xW_12−x Keggin anions

Self-assembly of P-Nb_xW_12−x Keggin anions in the range of chemical composition H₃PO₄ : Nb-Ox : Na₂WO₄ = 1 : n : 12−n, n = 1–6, was monitored by ³¹P NMR (Fig. 4). The spectra revealed only the formation of [PW₁₁O₃₉]⁷⁻ (−10.2 ppm), [PNbW₁₁O₄₀]⁴⁻ (−13.2 ppm), and the isomers of [PNb₂W₁₀O₄₀]⁵⁻ (−11.6 to −12.0 ppm) and [PNb₃W₉O₄₀]⁶⁻ (−10.3 to −11.3 ppm), as well as free phosphate (0.7 ppm). The species distribution as a function of number of Nb equivalents is shown in Fig. 5. With addition of one equivalent of Nb all phosphate is consumed, and the prominent species is the monosubstituted Keggin [PNbW₁₁O₄₀]⁴⁻ together with some [PNb₂W₁₀O₄₀]⁵⁻ and the lacunary POM [PW₁₁O₃₉]⁷⁻ as minor products. Starting from two equivalents of Nb, the disubstituted [PNb₂W₁₀O₄₀]⁵⁻ becomes the dominant product. Simultaneously the share of unreacted phosphate increases with the increase in the niobium amount. The self-assembly effectively stops at [PNb₂W₁₀O₄₀]⁵⁻ even if some [PNb₃W₉O₄₀]⁶⁻ is also detectable. These results support previous observations and explain why di- and trilacunary phosphotungstates produced only mixtures of mono- and disubstituted (Cs,K)₄[PNbW₁₁O₄₀] and (Cs,Na)₅[PNb₂W₁₀O₄₀].

Fig. 4 ³¹P NMR spectra of reaction mixtures of H₃PO₄ : Nb-Ox : Na₂WO₄ at the molar ratios (a) 1 : 1 : 11, (b) 1 : 2 : 10, (c) 1 : 3 : 9, (b) 1 : 4 : 8, (c) 1 : 5 : 7, (b) 1 : 6 : 6.

Fig. 5 ³¹P NMR species distribution as a function of the number of Nb eq. in the reaction mixtures of H₃PO₄ : Nb-Ox : Na₂WO₄ at the molar ratios 1 : n : 12−n, where n = 0–6.

The selectivity of the formation of the monosubstituted compound from the lacunary K₇[PW₁₁O₃₉] is superior to that in the one-pot self-assembly. The yield of [PNbW₁₁O₄₀]⁴⁻ is 85% (based on P) with [PW₁₁O₃₉]⁷⁻, and drops to modest 38% in the one-pot synthesis. In order to check whether the precipitation with Cs⁺ alters the nature and distribution of the products, we precipitated the product of the one-pot reaction for the molar ratio H₃PO₄ : Nb-Ox : Na₂WO₄ = 1 : 4 : 8 with Cs⁺ and redissolved it in water. Fig. 6 shows the ³¹P NMR spectra before and after the precipitation–dissolution procedure. They show no significant changes after precipitation of the products as Cs⁺ salts. Even the isomer distribution is retained, indicating that the thermodynamic equilibrium is rapidly reached during the synthesis.

Fig. 6 ³¹P NMR spectra of (a) aqueous reaction mixtures of H₃PO₄ : Nb-Ox : Na₂WO₄ at a molar ratio of 1 : 4 : 8 and (b) after precipitation with Cs⁺ and redissolution in H₂O/D₂O.

Isomer distribution analysis

As already mentioned, two sets of resonances were observed, corresponding to the isomers of [PNb₂W₁₀O₄₀]⁵⁻ and [PNb₃W₉O₄₀]⁶⁻. The numbers of isomers and their degeneracy for the series [PNb_xW_12−xO₄₀]^(3+x)− have been computed with a homemade routine running in the Matlab software package (ESI†) and the results are summarized in Table 1. These calculations are based on statistical distribution resulting from random W/Nb site occupancies. There are five different isomers with a statistical distribution of 1 : 2 : 2 : 2 : 4 for [PNb₂W₁₀O₄₀]⁵⁻, and 13 isomers with multiplicity of 1 : 1 : 2 : 3 : 3 : 3 : 6 : 6 : 6 : 6 : 6 : 6 : 6 are possible for [PNb₃W₉O₄₀]⁶⁻. We indeed observed five ³¹P resonances within the range −11.5 to −12.0 ppm, assigned to [PNb₂W₁₀O₄₀]⁵⁻ species, and ca. 15 signals in the range −10.0 to −11.5 ppm for [PNb₃W₉O₄₀]⁶⁻ (Fig. 4). Fig. 7 presents a zoom on the ³¹P NMR spectra in these ranges, as observed in the mixtures of H₃PO₄ : Nb-Ox : Na₂WO₄ at molar ratios 1 : x : 12−x, x = 2–6, as well as when Nb-Ox reacted with [A-PW₉O₃₄]⁹⁻, for comparison. It is possible to observe evolution in the relative signal intensity of the resonances of [PNb₂W₁₀O₄₀]⁵⁻, labeled “a” through “e”, when the number of Nb equivalents increases (Fig. 8). There are three groups of signals: (i) the most intense, 30–40% (signal c), (ii) the less intense, 2–5% (signal e), and (iii) the three remaining evolving around the value of 20% of the total intensity. This may indicate that the statistical distribution is nearly reached in all self-assembly processes regardless of the W/Nb stoichiometry used (theoretical ratio is 9% : 3 × 18% : 36%, represented by dotted lines in Fig. 8). Interestingly enough, the distribution of the isomers in the product from Nb-Ox and the lacunary POM (Fig. 7a) is significantly different: a: 43%; b: 9%; c: 41%; d: 6%; e: 1%. This clearly indicates regioselective substitution occurring in this case.

Table 1 Number of isomers and their distribution and degeneracy in the mixed Keggin [PNb_xW_12−xO₄₀]^(3+x)−, 0 ≤ x ≤ 6

x	Number of isomers	Isomers distribution (degeneracy)
0	1	1(1)
1	1	1(1)
2	5	1(1) : 3 × 1(2) : 1(4)
3	13	2 × 1(1) : 1(2) : 3 × 1(3) : 7 × 1(6)
4	26	1(2) : 1(3) : 8 × 1(4) : 16 × 1(8)
5	38	10 × 1(5) : 28 × 1(10)
6	47	2 × 1(2) : 2 × 1(3) : 2 × 1(4) : 8 × 1(6) : 33 × 1(12)

---

Fig. 7 Expanded ³¹P NMR spectra in the ranges of resonances of [PNb₂W₁₀O₄₀]⁵⁻ (−11 to −12 ppm) and [PNb₃W₉O₄₀]⁶⁻ (−10 to −11 ppm) observed in (a) reaction between Nb-Ox and Na₉[A-PW₉O₃₄], and in H₃PO₄ : Nb-Ox : Na₂WO₄ mixtures at the molar ratios (b) 1 : 2 : 10, (c) 1 : 3 : 9, (d) 1 : 4 : 8, (e) 1 : 5 : 7, (f) 1 : 6 : 6.

Fig. 8 ³¹P NMR species distribution as a function of Nb equivalents introduced in the mixtures of H₃PO₄ : Nb-Ox : Na₂WO₄ at the molar ratios 1 : n : 12−n, where n = 0–6.

From these distributions we can propose an assignment of the signals (Fig. 9). The most intense signal (“c”) should correspond to the most statistically abundant isomer, while the less intense signal (“e”) – to the least abundant isomer. The three remaining isomers occur at the same frequency, and one of them (signal “a”) was particularly prominent in the regioselective reaction with [PW₉O₃₄]⁹⁻. Such a degree of selectivity implies that a significant part of the starting lacunary POMs does not undergo any decomposition during the reaction. In this case the probability to find two Nb atoms in the same trimeric unit filling the vacancy of the PW₉ unit becomes high, and this species should therefore correspond to signal “a”.

Fig. 9 The five isomers in the mixed Keggin [PNb₂W₁₀O₄₀]⁵⁻ and possible ³¹P NMR assignments: signals a–e in Fig. 7.

Conclusions

In this research we have developed an alternative synthetic way to niobium-substituted Keggin phosphotungstate with (NH₄)[NbO(C₂O₄)₂(H₂O)₂]·3H₂O. In the reaction of [PW₁₁O₃₉]⁷⁻ with this niobium oxalate selective formation of [PNbW₁₁O₄₀]⁴⁻ is achieved. Reactions with bi- and tri-lacunary phosphotungstate or self-assembly reactions of the niobium oxalate complex, sodium tungstate and phosphoric acid give mixtures of niobium substituted Keggins, further complicated with the presence of all possible niobium positional isomers in the di- and tri-substituted products, even though the formation of [PNb₂W₁₀O₄₀]⁵⁻ is always preferable. Increasing the niobium-to-tungsten ratio above 1 : 3 does not lead to the uptake of extra Nb into the Keggin backbone.

Application of this strategy to the synthesis of other mixed Nb–W polyoxometalates is in progress.

Experimental section

General information

(NH₄)[NbO(C₂O₄)₂(H₂O)₂]·3H₂O (Nb-Ox) was prepared starting from K₇H[Nb₆O₁₉]·13H₂O.⁵¹ The purity of the compound was confirmed with powder X-ray diffraction by comparison of the XRD patterns with CCDC AMOXNB. The lacunary POMs Na₉[A-PW₉O₃₄]·7H₂O, Cs₇[PW₁₀O₃₆]·H₂O and K₇[PW₁₁O₃₉]·17H₂O were synthesized following the published protocols.⁵² Fourier transformed infrared (FT-IR) spectra were recorded on a 6700 FT-IR Nicolet spectrophotometer. The spectra were recorded on undiluted samples and ATR correction was applied. EDX analysis was carried out on a HITACHI TM 3000 Tabletop Microscope. TGA experiments were performed using a Seiko TG/DTA 320 thermogravimetric balance (5 °C min⁻¹, under O₂).

Reactions of lacunary POMs with Nb-Ox

(a) 0.46 g (1.1 mmol) of solid Nb-Ox was added to a clear solution of 1.00 g (0.4 mmol) of Na₉[A-PW₉O₃₄]·7H₂O in 20 mL of distilled water. The pH value dropped down to pH 3.1 upon mixing. After complete dissolution of Nb-Ox, the mixture was kept at 80 °C for 30 min. Reaction products were then studied with ³¹P NMR in solution. For the solid-state NMR study, the sample was prepared by addition of 1.00 g (6 mmol) of CsCl. The crude product was collected by filtration, washed with water and dried in vacuo.

(b) 0.47 g (1.2 mmol) of solid Nb-Ox was added to a suspension of 2.00 g (0.6 mmol) of Cs₇[PW₁₀O₃₆]·H₂O in 20 mL of distilled water. The pH value was adjusted to pH 3 with conc. HCl. After complete dissolution of Nb-Ox, the mixture was kept at 80 °C for 30 min. The crude product (1.80 g) was collected by filtration, washed with water and dried in vacuo.

(c) 0.12 g (0.3 mmol) of solid Nb-Ox was added to a clear solution of 2.00 g (0.6 mmol) of K₇[PW₁₁O₃₉]·17H₂O in 20 mL of distilled water. The pH value was adjusted to pH 3 with conc. HCl. After complete dissolution of Nb-Ox, the mixture was kept at 80 °C for 30 min. The reaction products were then studied with ³¹P and ¹⁸³W NMR in solution. For the solid-state NMR study, the sample was prepared by precipitation with 2.00 g (0.012 mol) of CsCl. The crude product (1.90 g) was collected by filtration, washed with water and dried in vacuo. Cesium salt IR (diamond, cm⁻¹): 1088 (sh), 1073 (s), 983 (vs), 882 (s), 781 (vs), 596 (m), 518 (m), 377 (m), 333 (m). TGA loss 6.2%. EDX calc. for Cs₄[PNbW₁₁O₄₀]·12H₂O Cs, P, Nb, W (%): 15.04, 0.88, 2.63, 57.23; found Cs, P, Nb, W (%) 14.82, 0.81, 2.45, 57.02. Potassium salt is more soluble and can be isolated by addition of 3 g of KCl to the final solution, it gives 0.8 g of the crude product. IR (diamond, cm⁻¹): 1088 (sh), 1074 (s), 977 (vs), 882 (s), 781 (vs), 596 (m), 518 (m), 372 (m), 330 (m). TGA loss 4.4%. EDX calc. for K₄[PNbW₁₁O₄₀]·7H₂O K, P, Nb, W (%): 5.10, 1.01, 3.03, 65.90; found K, P, Nb, W (%) 5.09, 0.87, 3.12, 65.66.

Self-assembly reactions

All syntheses were done by dissolving a fixed amount of Na₂WO₄2H₂O in 20 mL of distilled water, followed by addition of the appropriate amounts of Nb-Ox and H₃PO₄ (85%). When the amount of niobium oxalate was small the pH was adjusted to 3 with 1M HCl. The reaction mixtures were thermostated for 30 min at 80 °C and then analyzed with ³¹P NMR. Depending on the required W : Nb : P molar ratio, the quantities are as follows:

W : Nb :P 11 : 1 : 1; 2.00 g (6 mmol) Na₂WO₄·2H₂O, 0.22 g of Nb-Ox and 44 μL of H₃PO₄;

W : Nb : P 10 : 2 : 1; 2.00 g (6 mmol) Na₂WO₄·2H₂O, 0.48 g of Nb-Ox (10 : 2) and 44 μL of H₃PO₄;

W : Nb : P 9 : 3 : 1; 2.00 g (6 mmol) Na₂WO₄·2H₂O, 0.79 g of Nb-Ox (9 : 3) and 44 μL of H₃PO₄;

W : Nb : P 8 : 4 : 1; 2.00 g (6 mmol) Na₂WO₄·2H₂O, 1.20 g of Nb-Ox (8 : 4) and 44 μL of H₃PO₄;

W : Nb : P 7 : 5 : 1; 2.00g (6 mmol) Na₂WO₄·2H₂O, 1.70 g of Nb-Ox (7 : 5) and 44 μL of H₃PO₄;

W : Nb : P 6 : 6 : 1; 2.00 g (6 mmol) Na₂WO₄·2H₂O, 2.40 g of Nb-Ox (6 : 6) and 44 μL of H₃PO₄.

NMR experiments

The ³¹P MAS NMR experiments were performed on a Bruker Avance 500 spectrometer equipped with a 11.7 T wide-bore magnet. The ³¹P resonance frequency was 202.5 MHz and the experiments were performed on a 3.2 mm MAS probe with a spinning rate of 10 kHz. ³¹P CP/MAS NMR spectra were recorded with a contact time of 10 ms and a recycle delay of 15 s. Quantitative ³¹P MAS NMR spectra were also recorded using high-power proton decoupling. A total of 16 scans were accumulated with a π/2 pulse width of 1.96 μs and a 475 s recycle delay. Chemical shifts were referenced to 85% H₃PO₄ at 0 ppm.

The solution ³¹P NMR (162 MHz) spectra in D₂O were recorded in 5 mm outer diameter tubes on a Bruker 400 MHz spectrometer. Quantitative measurements were performed by accumulating 256 scans of π/2 pulses with a relaxation delay of 200 s. The NMR spectra were referenced to an external standard of 85% H₃PO₄ in H₂O. The ¹⁸³W NMR (20.8 MHz) spectra were recorded in 10 mm outer diameter tubes, on a Bruker Avance 500 spectrometer equipped with a Bruker low-frequency tunable probe. These spectra were referenced to an external standard (saturated Na₂WO₄–D₂O solution) by the substitution method. Chemical shifts were reported on the δ scale with negative scale up-field from Na₂WO₄ (δ = 0). The ¹⁸³W NMR scale was calibrated relative to the signal of D₂O/H₂O solution of H₄SiW₁₂O₄₀ at −103.8 ppm with respect to the saturated Na₂WO₄ solution.

Acknowledgements

This work was supported by RFBR (grant number 15-33-20651), LIA FRANCE-RUSSIA No. 1144 CLUSPOM and COST STSM program (COST-STSM-ECOST-STSM-CM1203-101015-062506).

Notes and references

1. V. W. Day, W. G. Klemperer and C. Schwartz, J. Am. Chem. Soc., 1987, 109, 6030.
2. G. S. Kim, H. D. Zeng, D. Van Derveer and C. L. Hill, Angew. Chem., Int. Ed., 1999, 38, 3205.
3. M. K. Harrup, G. S. Kim, H. D. Zeng, R. P. Johnson, D. Van Derveer and C. L. Hill, Inorg. Chem., 1998, 37, 5550.
4. G. S. Kim, D. A. Judd, C. L. Hill and R. F. Schinazi, J. Med. Chem., 1994, 37, 816.
5. M. K. Harrup and C. L. Hill, J. Mol. Catal. A: Chem., 1996, 106, 57.
6. D. A. Judd, J. H. Nettles, N. Nevins, J. P. Snyder, D. C. Liotta, J. Tang, J. Ermolieff, R. F. Schinazi and C. L. Hill, J. Am. Chem. Soc., 2001, 123, 886.
7. A. Trovarelli and R. G. Finke, Inorg. Chem., 1993, 32, 6034.
8. R. G. Finke, K. Nomiya, C. A. Green and M. W. Droege, Inorg. Synth., 1992, 29, 239.
9. R. G. Finke and M. W. Droege, J. Am. Chem. Soc., 1984, 106, 7274.
10. F. Bannani, H. Driss, R. Thouvenot and M. Debbabi, J. Chem. Crystallogr., 2007, 37, 37.
11. T. M. Anderson, M. A. Rodriguez, T. A. Stewart, J. N. Bixler, W. Q. Xu, J. B. Parise and M. Nyman, Eur. J. Inorg. Chem., 2008, 3286.
12. K. Bouadjadja-Rohan, C. Lancelot, M. Fournier, A. Bonduelle-Skrzypczak, A. Hugon, O. Mentre and C. Lamonier, Eur. J. Inorg. Chem., 2015, 2067.
13. B. Yue, L. B. Tang, C. G. Rui, H. Z. Liu, S. L. Jin, S. S. Zhu, J. Liu and M. Q. Chen, Chem. Res. Chin. Univ., 1997, 13, 195.
14. A. Stein, M. Fendorf, T. P. Jarvie, K. T. Mueller, A. J. Benesi and T. E. Mallouk, Chem. Mater., 1995, 7, 304.
15. E. V. Radkov and R. H. Beer, Inorg. Chim. Acta, 2000, 297, 191.
16. Y. H. Ren, Y. C. Hu, Y. C. Shan, Z. P. Kong, M. Gu, B. Yue and H. Y. He, Inorg. Chem. Commun., 2014, 40, 108.
17. C. C. Li, S. X. Liu, S. J. Li, Y. Yang, H. Y. Jin and F. J. Ma, Eur. J. Inorg. Chem., 2012, 3229–3234.
18. P. I. Molina, D. J. Sures, P. Miro, L. N. Zakharov and M. Nyman, Dalton Trans., 2015, 44, 15813.
19. J. M. Maestre, J. P. Sarasa, C. Bo and J. M. Poblet, Inorg. Chem., 1998, 37, 3071.
20. D. D. Zhang, C. Zhang, P. T. Ma, B. S. Bassil, R. Al-Oweini, U. Kortz, J. P. Wang and J. Y. Niu, Inorg. Chem. Front., 2015, 2, 254.
21. R. G. Bowmann, D. C. Molzahn and G. E. Hartwell, US Pat., US5030740 A, 1991.
22. M. Dabbabi and M. Boyer, J. Inorg. Nucl. Chem., 1976, 38, 1011.
23. D. J. Sures, P. I. Molina, P. Miro, L. N. Zakharov and M. Nyman, New J. Chem., 2016, 40, 928.
24. M. Dabbabi, M. Boyer, J. Pierrelaunay and Y. J. Jeannin, J. Electroanal. Chem., 1977, 76, 153.
25. W. Clegg, M. R. J. Elsegood, R. J. Errington and J. Havelock, J. Chem. Soc., Dalton Trans., 1996, 681.
26. M. Nyman, Dalton Trans., 2011, 40, 8049.
27. M. W. Droege and R. G. Finke, J. Mol. Catal., 1991, 69, 323.
28. N. Mizuno, H. Weiner and R. G. Finke, J. Mol. Catal. A: Chem., 1996, 114, 15.
29. J. M. Jehng and I. E. Wachs, J. Raman Spectrosc., 1991, 22, 83.
30. M. Hino, M. Kurashige, H. Matsuhashi and K. Arata, Appl. Catal., A, 2006, 310, 190.
31. S. M. Wang, L. Liu, Z. Y. Huang and Z. B. Han, RSC Adv., 2016, 6, 38782.
32. X. B. Dong, D. J. Wang, K. B. Li, Y. Z. Zhen, H. M. Hu and G. L. Xue, Mater. Res. Bull., 2014, 57, 210.
33. J. Arichi, M. Eternot and B. Louis, Catal. Today, 2008, 138, 117.
34. J. E. Molinari, L. Nakka, T. Kim and I. E. Wachs, ACS Catal., 2011, 1, 1536.
35. B. C. Ma, Z. X. Zhang, W. F. Song, X. L. Xue, Y. Z. Yu, Z. S. Zhao and Y. Ding, J. Mol. Catal. A: Chem., 2013, 368, 152.
36. O. W. Howarth, L. Pettersson and I. Andersson, J. Chem. Soc., Dalton Trans., 1991, 1799.
37. Y. Xu, J. Q. Xu, C. X. Guo, G. Y. Yang, R. Z. Wang and H. R. Sun, Chem. J. Chin. Univ., 1999, 20, 14.
38. Q. Lan, Z. M. Zhang, Y. G. Li, Y. Lu and E. B. Wang, Dalton Trans., 2014, 43, 16265.
39. X. Lopez, C. Bo and J. M. Poblet, J. Am. Chem. Soc., 2002, 124, 12574.
40. K. Nomiya, C. Nozaki, A. Kano, T. Taguchi and K. Ohsawa, J. Organomet. Chem., 1997, 533, 153.
41. N. Satake, T. Hirano, K. Kamata, K. Suzuki, K. Yamaguchi and N. Mizuno, Chem. Lett., 2015, 44, 899.
42. D. R. Park, S. Park, Y. Bang and I. K. Song, Appl. Catal., A, 2010, 373, 201.
43. E. Cadot, V. Béreau and F. Sécheresse, Inorg. Chim. Acta, 1995, 239, 39.
44. E. V. Ramos-Fernandez, C. Pieters, B. Van der Linden, J. Juan-Alcaniz, P. Serra-Crespo, M. Verhoeven, H. Niemantsverdriet, J. Gascon and F. Kapteijn, J. Catal., 2012, 289, 42.
45. W. Salomon, C. Roch-Marchal, P. Mialane, P. Rouschmeyer, C. Serre, M. Haouas, F. Taulelle, S. Yang, L. Ruhlmann and A. Dolbecq, Chem. Commun., 2015, 51, 2972.
46. K. Uehara, T. Oishi, T. Hirose and N. Mizuno, Inorg. Chem., 2013, 52, 11200.
47. W. H. Knoth and R. L. Harlow, J. Am. Chem. Soc., 1981, 103, 1865.
48. E. Radkov and R. H. Beer, Polyhedron, 1995, 14, 2139.
49. E. Radkov, Y. J. Lu and R. H. Beer, Inorg. Chem., 1996, 35, 551.
50. Z. G. Sun, Q. Liu and J. F. Liu, Transition Met. Chem., 2000, 25, 374.
51. M. Jurić, J. Popović, A. Šantić, K. Molčanov, N. Brničević and P. Planinić, Inorg. Chem., 2013, 52, 1832.
52. P. J. Domaille, Inorg. Synth., 1992, 27, 96.

---

Footnote

† Electronic supplementary information (ESI) available: Matlab code (Topo_MH) to compute isomers in the series [PNbxW12−xO40](3+x)−. See DOI: 10.1039/c6nj02637k

---