Unique solubility of polyoxoniobate salts in methanol: coordination to cations and POM methylation

Abstract

Sodium salts of hybrid organometallic POM complexes consisting of [M₆O₁₉]⁸⁻ (M = Nb, Ta) and half-sandwich fragments {Cp*M}²⁺ (Cp* = η⁵-C₅(CH₃)₅; M = Rh, Ir) and {(C₆H₆)Ru}²⁺ dissolve only in one organic solvent, that is, in methanol. From methanolic solutions, crystals of Na₄[trans-{(C₆H₆)Ru}₂Nb₆O₁₉]·14.125MeOH·2H₂O (1), K₄[trans-{Cp*Rh}₂Nb₆O₁₉]·4MeOH·10H₂O (2) and K₄[trans-{Cp*Ir}₂Nb₆O₁₉]·10MeOH·4H₂O (3) were isolated and characterized by X-ray analysis. Methoxo species [{LM′}₂M₆O_19−n(OCH₃)_n] (n = 1–3) were detected in solution by ESI-MS and NMR, and they account only for about half of the total speciation in the solution. DFT calculations were used to calculate ¹³C NMR chemical shifts in the methoxo complexes and to assess their relative stability. Reasons for the preferred solubility in methanol are discussed.

---

Introduction

The chemistry of polyoxoniobates and tantalates is routinely confined to strongly alkaline aqueous solutions, but a remarkably rich chemistry is displayed even under these restrictions.¹ Non-aqueous chemistry of Nb and Ta polyoxometalates is by far less studied. Hydrated Nb₂O₅ can be solubilized in methanol under solventothermal conditions in the presence of Me₄NOH as a base.² It also reacts with different organic bases in acetonitrile. In this way [Nb₁₀O₂₈]⁶⁻,³ [Nb₂₀O₅₄]⁸⁻,⁴ as well as the recently reported [Ta₁₀O₂₈]^6− 5 were prepared. The main reason of the strong bias in favor of choosing water as the solvent is the high charge of the polyniobates, which requires large numbers of counter cations, typically alkali metal cations. This leads to high lattice energies, which can be counterbalanced only by a solvent with a high dielectric constant and donor/acceptor number, preferably with water. Hydrogen bonds also play an important part in stabilizing the highly charged polyanions, and those are again best provided by water. Can hydrated alkali metal salts of highly charged polyoxometalates be dissolved and handled in organic solvents? One can expect that coordination of a positively charged organometallic moiety such as {Cp*M}²⁺ (M = Rh, Ir) or {(arene)Ru}²⁺ (arene is benzene and its derivatives) to [Nb₆O₁₉]⁸⁻ would reduce the overall negative charge and simultaneously increase the hydrophobicity of the hybrid complex, thus enabling solubility in organic solvents. Indeed, Proust et al. reported that the neutral complex [{(p-cym)Ru}₄Nb₆O₁₉] (p-cym = 1-methyl-4-(1-methylethyl)benzene) was soluble in CH₃OH, and the negative ion mode electrospray mass spectrum of methanolic solutions detected formation of the methoxo complexes [{(p-cym)Ru}₂Nb₆O₁₆(OCH₃)₃]⁻ and [{(p-cym)Ru}₃Nb₆O₁₈(OCH₃)]⁻, arising from de-complexation of one or two {(p-cym)Ru}²⁺ fragments and formation of Nb–OCH₃ bonds.⁶ This solvolitic behavior in methanol is strongly reminiscent of the reactivity of [Nb₂W₄O₁₉]⁴⁻ towards CH₃OH, and of [P₂W₁₅V₃O₆₂]⁹⁻ with CH₃C(CH₂OH)₃.^7,8 In the present study we found that sodium salts of [{(C₆H₆)Ru}₂Nb₆O₁₉]⁴⁻, [{Cp*Rh}₂Nb₆O₁₉]⁴⁻ and [{Cp*Ir}₂Nb₆O₁₉]⁴⁻ were soluble in methanol, but not in ethanol or other organic solvents. Both the solvolysis of the anion and solvation of the alkali metal cation are the driving force in our case: the methoxo species [{LM′}₂M₆O_19−n(OCH₃)_n] (n = 1–3) are detected by ESI-MS and NMR, and they account for about half of the total speciation of the hybrid complexes in methanol. From these solutions, crystals of Na₄[{(C₆H₆)Ru}₂Nb₆O₁₉]·14.125MeOH·2H₂O (1), K₄[trans-{Cp*Rh}₂Nb₆O₁₉]·4MeOH·10H₂O (2) and K₄[trans-{Cp*Ir}₂Nb₆O₁₉]·10MeOH·4H₂O (3) were isolated in high yields, which bear no Nb–OCH₃ groups, but show direct coordination of CH₃OH to Na⁺ instead. These observations are presented and discussed in detail below.

Results and discussion

ESI-MS experiments

The first indications of the formation of methoxo complexes were obtained from ESI-MS experiments. Reactions of [M₆O₁₉]⁸⁻ (M = Nb and Ta) salts with [Cp*RhCl₂]₂ at a 2 : 1 Rh/M₆ ratio in water yielded solids containing [(Cp*Rh)₂M₆O₁₉]⁴⁻ anions.

When their ESI (−) mass spectra were investigated using methanol as the solvent, signals attributed to methoxo species were readily identified on the basis of their m/z value as well as their isotopic distribution. In particular signals from the 2:1 species [{Cp*Rh}₂Nb₆O₁₉H₃]⁻ (m/z 1340), [{Cp*Rh}₂Nb₆O₁₈(OCH₃)H₂]⁻ (m/z 1354), [{Cp*Rh}₂Ta₆O₁₉H₃]⁻ (m/z 1868) and [{Cp*Rh}₂Ta₆O₁₈(OCH₃)H₂]⁻ (m/z 1882) were identified. Table S1† shows the m/z values for the methoxo species. Signals from the 1 : 1 adducts, namely [{Cp*Rh}Nb₆O₁₉H₅]⁻ (m/z 1104), [{Cp*Rh}Nb₆O₁₇(OCH₃)₂H₃]⁻ (m/z 1132), [{Cp*Rh}Nb₆O₁₈(OCH₃)H₄]⁻ (m/z 1118), [{Cp*Rh}Ta₆O₁₈(OCH₃)H₃]²⁻ (m/z 824), and [{Cp*Rh}Ta₆O₁₈(OCH₃)H₄]⁻ (m/z 1646),⁹ were also observed in their respective ESI mass spectra in methanol. We also found that sodium salts of the 2 : 1 and 1 : 1 complexes Na₄[{Cp*Ir}₂Nb₆O₁₉]·22H₂O and Na₆[{Cp*Ir}Ta₆O₁₉]·27H₂O were soluble in CH₃OH, and signals from the corresponding anions as well as from their mono- and di-methylated derivatives [{Cp*Ir}₂Nb₆O₁₈(OCH₃)H]²⁻ (m/z 765) and [{Cp*Ir}₂Nb₆O₁₇(OCH₃)H₂]⁻ (m/z 1532) were detected (see for example Fig. S1 and S2†) in these solutions.¹⁰ Methoxo complexes were also detected in the ESI-MS spectra of [{Cp*Ir}₂TeNb₅O₁₉]³⁻ and [{Cp*Ir}TeNb₅O₁₉]⁵⁻.¹¹ This indicates the tendency of the Lindqvist POMs to form methoxo species upon dissolving in methanol.

Since in all cases formation of methylated hybrid anions was detected by ESI-MS experiments in freshly prepared methanolic solutions, it would appear that they rapidly form in solution in significant amounts, up to a certain equilibrium point between the starting oxo and methoxo species. However, we are aware that the large methanol excess employed for sample preparation for ESI-MS analysis (sample concentration ca. 5 × 10⁻⁵ M in CH₃OH) might exaggerate the propensity of the oxo or hydroxo groups to be replaced by methoxo; in addition, the appearance of the signals due to the methoxo species may be due to secondary processes operative only in the gas phase under experimental conditions. Therefore we tried to isolate methoxo complexes directly from methanolic solutions and demonstrate their formation in solution by X-ray crystallography and NMR techniques, respectively.

X-ray structures

The reaction of sodium hexaniobate with [(C₆H₆)RuCl₂]₂ proceeds slowly upon dissolution of the ruthenium complex in a water solution of hexaniobate, which typically takes 1–2 hours for completion depending on the loads and temperature, and then the yellow solution is heated and stirred for 12 h. The product is precipitated with acetone (slow diffusion gives yellow crystals and quick addition gives a yellow precipitate). Recrystallization from CH₃OH yields large yellow plates of Na₄[{(C₆H₆)Ru}₂Nb₆O₁₉]·14.125MeOH·2H₂O (1), whose structure was determined. The phase purity of the product was confirmed by XRPD.

The crystal structure of 1 consists of trans-[{(C₆H₆)Ru}₂Nb₆O₁₉]⁴⁻ hybrid anions and sodium cations, which aggregate into [(MeOH)₃Na(μ-MeOH)Na(H₂O) (MeOH)₂]²⁺ dimers; hence the formula of 1 is better rendered as [(MeOH)₃Na(μ-MeOH)Na(H₂O)(MeOH)₂]₂[{(C₆H₆)Ru}₂Nb₆O₁₉]·2.125MeOH. The cationic dimers unite the anions into a 3D framework (Fig. 1). This is the first example of breaking the standard layered structural motif common to all previously known hybrid complexes bearing two coordinated C₆H₆ or Cp* fragments.^6,9 The formation of the 3D framework (the building block is represented in Fig. S3†) creates infinite channels running along the [010] and [100] crystallographic directions, which are filled with disordered methanol molecules (2.125 molecules per formula unit from XRD refinement) (Fig. S4†). The methanol molecules that are coordinated to the cations form strong hydrogen bonds with bridging oxo ligands (d(MeOH⋯μ₂-O) is 2.645 and 2.701 Å) of the hybrid POM. Moreover, one solvate water molecule forms a weaker hydrogen bond with μ₂-O (the H₂O⋯O distance is 2.821 Å).

Fig. 1 Infinite channels occupied with disordered methanol molecules (not shown for clarity) in the crystal packing of 1. Lindqvist anions are shown in the blue polyhedra, and sodium cations are colored green.

trans-K₄[{Cp*Rh}₂Nb₆O₁₉]·20H₂O (ref. 9) and trans-K₄[{Cp*Ir}₂Nb₆O₁₉]·22H₂O (ref. 10) also dissolve in methanol, but crystalline samples could not be obtained by straightforward evaporation of the solvent. In one of the crystallization attempts we added a few milligrams of dibenzo-18-crown-6 to the methanolic solutions, which then were left at 2 °C. Yellow single crystal plates were obtained in both cases, but curiously enough, no crown ether molecules entered the structure, the formulae of the products being trans-K₄[{Cp*Rh}₂Nb₆O₁₉]·4MeOH·10H₂O (2) and trans-K₄[trans-{Cp*Ir}₂Nb₆O₁₉]·10MeOH·4H₂O (3). Possibly the crown ether binds excess water, forming associates with strong hydrogen bonds, which remain in solution, but this hydration shifts the equilibrium and facilitates formation of 2 and 3. The two salts are obviously not isostructural, and 2 crystallizes in the P2₁/c space group while 3 adopts the P1̄ group. This difference is due to the different values in the content of methanol and water molecules. However, whatever the differences are, in both structures the hybrid anions trans-[{Cp*M}₂Nb₆O₁₉]⁴⁻ are combined with the cationic part (there are infinite chains of the cations connected with the solvate methanol, water molecules and oxygen ligands of the anions, Fig. S5 and S6†) into the layers spreading along the [011] (M = Rh) or [110] (M = Ir) crystallographic directions (Fig. S7 and S8†). Both in 2 and 3 additional K⁺ cations coordinate free {Nb₃O₃} faces of the anions. In 2 such cations have CN 7 (with three oxo ligands from the anion face, three water molecules and one methanol, Fig. S9†), and in 3 K⁺ has CN 8 (three oxo ligands from the anion face, three water molecules and two methanol molecules, Fig. S10,† left). Methanol molecules, coordinated to the cations, form hydrogen bonds with POM both in 2 and 3. In 2 only the methanol molecules that are located in the positions near to the anion produce hydrogen bonds (d(Nb=O⋯HOCH₃) is 2.761 Å) stronger than those in water molecules (d(Nb=O⋯H₂O) is 2.900 Å) (Fig. S9†). This effect is more significant in the case of 3, which has more CH₃OH molecules in the structure. All methanol molecules are oriented in a way to afford hydrogen bonding with the anion (d(O⋯O) is 2.717–2.748 Å) (Fig. S10,† right). Details of XRD experiments and refinement are summarized in Table S2.†

Qualitative experiments have shown that also Cs⁺ (Ta), K⁺ (Nb), and Na⁺ (Ta, Nb) salts for [{Cp*M}₂M₆O₁₉]⁴⁻ complexes (M = Rh, Ir), and K⁺ (Nb), and Na⁺ (Nb, Ta) salts for [{(C₆H₆)Ru}₂M₆O₁₉]⁶⁻ complexes are soluble in methanol. Attempts to get crystals from these solutions failed. The crystal packing of all starting materials consists of the layers formed by the hybrid anions and solvated cations which stack together only through π–π interactions between the aromatic rings of the organic ligands. Methanol can penetrate between such layers and replace water in the coordination spheres of the cations, while CH₃CN or CH₂Cl₂ cannot do this.

¹³C and ¹H NMR experiments

The compounds 1–3 were isolated in high yields. This means either that the methoxo complexes are present only in solution in a rapidly shifting equilibrium so that they dissociate upon crystallization, or that they are kinetically stable, but only as a minor component of the complicated equilibria. In order to check this hypothesis we studied methanolic solutions of 2 and 3 by NMR. In the ¹³C NMR spectra of 2 and 3 in CD₃OD there appears two signals at 68 and 61 ppm for the Rh complex (Fig. 2), and at 61 and 56 ppm for the Ir complex (Fig. 3), in addition to the signals from the Cp* ligands.

Fig. 2 ¹³C NMR spectrum of d⁴-MeOH solution of 2.

Fig. 3 ¹³C NMR spectrum of 3 in CD₃OD.

Since free methanol gives a ¹³C NMR resonance at 49.5 ppm, these signals can be assigned to coordinated CH₃O⁻. Indeed, Errington et al. reported (Bu₄N)₂[(MeO)NbW₅O₁₈] in CD₃CN δ_CH3O at 65.91 ppm.¹² Similar chemical shifts for various methoxy species were reported elsewhere.¹³ Karaliota et al. even discriminated between terminal Nb–OCH₃ and bridging Nb–(μ-OCH₃)–Nb modes, assuming δ 67 ppm for the bridging and 60–64 ppm for the terminal methoxy ligands for solutions of NbCl₅ in methanol.¹⁴ Thus ¹³C NMR gives evidence for the formation of methoxo complexes in methanolic solutions of 2 and 3. In order to confirm this assignment we carried out DFT calculation of the ¹³C chemical shifts taking methylated isomers of [{Cp*Ir}₂Nb₆O₁₈(OCH₃)]³⁻ as models, which yielded δ_CH3 60.6 ppm and 63.4 ppm for the terminal and bridging methoxide, respectively. These results show that the difference between the chemical shifts of the terminal and bridged methoxo ligands is meaningful and agrees with Karaliota’s assignment. According to ¹³C NMR, we have complexes with both terminal (upfield) and bridging (downfield) CH₃O⁻ ligands in solution. Three isomers are possible, as shown in Fig. 4. The formation of the new species in CH₃OH solutions can be also deduced from new signals from Cp* which appear at 93.70 and 94.23 ppm as broad resonances, together with signals from the starting trans-[{Cp*Rh}₂Nb₆O₁₉]⁴⁻ at 92.91 and 92.68 ppm.

Fig. 4 Atomic numbering scheme used in DFT calculations for trans-[{Cp*Ir}₂Nb₆O₁₈(OCH₃)]³⁻. Isomer I: CH₃ group added to terminal oxygen atoms (any of the six equivalent atoms O1, O24, O13, O17, O22, and O12). Isomer II: CH₃ group added to any of the six equivalent bridging atoms in the “equatorial belt” (O14, O15, O23, O5, O6, and O19). Isomer III: CH₃ group added to any of the six capping oxygen atoms (O4, O7, O16, O18, O20, and O21).

The methyl group proton (¹H NMR) resonances which appear in solutions of NbCl₅ in CH₃OH fall in three series around the 4.5 ppm, 4.2–3.8 ppm, and 3.3 ppm regions, and are due to bridging methoxides, terminal methoxides, and to free methanol, respectively.^{14,15 1}H NMR of 1 in MeOH gives δ = 3.530, 3.520, 3.504, 3.491, and 3.468 ppm for terminal coordination and δ = 4.284 ppm for bridging coordination, but in ¹³C NMR we observed only one signal at 61 ppm. For complex 2 ¹H NMR also shows formation of terminal MeO⁻ species (δ = 3.67 ppm), but peaks from bridging ligands also appear at 4.53–4.55 ppm. The Cp* signals from trans-[{Cp*Rh}₂Nb₆O₁₉]⁴⁻ appear at 1.90 and 1.87 ppm, together with new signals at 1.92 and 1.94 ppm. The latter two signals can be assigned to the signals from Cp* groups in the CH₃O⁻ complexes. The signal intensity at 3.67 ppm is 5% of that of the two signals with δ = 1.90 and 1.87 ppm, hence the corresponding Cp* signal must have a 10-fold intensity, which in turn corresponds to that of the peak with δ = 1.92 ppm. In the same way, the integral intensity of the signals from the bridging CH₃O group (from various isomers at 4.53–4.55 ppm) is only 1%, which correlates with the intensity of the Cp* peak at 1.94 ppm (10%). The total intensity of the Cp* signals from the methoxo complexes is 60% of that of trans-[{Cp*Rh}₂Nb₆O₁₉]⁴⁻ which indicates a high degree of methylation (37.5%), in qualitative agreement with the ESI data. A methanolic solution of 3 demonstrates signals from terminal (3.68 ppm) and bridging (4.35, 4.33, 4.32, 4.30, 4.29, and 4.26 ppm) groups. The signals from Cp* appear at 1.83 and 1.81 ppm (for trans-[{Cp*Ir}₂Nb₆O₁₉]⁴⁻) and 1.85–1.87 ppm (for methylated POMs). The signal at 1.85 ppm can be assigned to the isomers with a terminal CH₃O group (10% of the intensity of the Cp* groups from the starting trans-[{Cp*Ir}₂Nb₆O₁₉]⁴⁻), and that of 1.87 ppm to the isomers with a bridging methoxide (100% of the intensity). This corresponds to the total methylation of ca. 53% of the starting complex. The formation of the new species in CH₃OH solutions can be also deduced from new signals from Cp* which appear at 84.75 and 85.65 ppm as broad resonances, together with the signals from the starting trans-[{Cp*Ir}₂Nb₆O₁₉]⁴⁻ at 83.67 and 83.64 ppm. Observed signals in ¹H NMR for all complexes are summarized in Table 1. Also both for Rh and Ir there is a signal at δ = 29.4 ppm (¹H NMR δ = 3.37 ppm) which appears also in ref. 14, and was assigned by the authors of this work to CH₃Cl. But we assign it as an acetone trace (30 ppm, http://nmrshiftdb.nmr.uni-koeln.de/).

Table 1 ¹H NMR data for 1–3 in CD₃OD

Complex	δ(term)^a (ppm)	δ(bridged)^a (ppm)
a Assigned from ref. 14.
1	3.530, 3.520, 3.504, 3.491, 3.468	4.284
2	3.670	4.002, 3.987, 4.553, 4.537, 4.528
3	3.678, 3.630	4.347, 4.334, 4.307, 4.294, 4.260

---

The conclusion from this part of the work is that the formation of the methoxo complexes does take place in methanolic solutions and contributes to enhanced solubility of Na₄[{(C₆H₆)Ru}₂Nb₆O₁₉] and K₄[{Cp*M}₂Nb₆O₁₉] (M = Rh, Ir) in methanol as a result of specific solvolysis.

DFT calculations

In order to gain more insight into the formation of the methoxo complexes, DFT calculations were carried out (atomic numbering is presented in Fig. 4). Calculation of the charge distribution on the terminal oxoligands of the free Lindqvist anion [Nb₆O₁₉]⁸⁻ shows equal Hirshfeld charges of −0.754 (see Table S3†). Coordination of a {Cp*Ir}²⁺ fragment leads, as expected, to an overall decrease in the negative charge for all terminal oxygen atoms, but the terminal oxo ligands at the free face opposite to {Cp*IrNb₃O₃} retain a greater negative charge (−0.620 vs. −0.565) (Table S3†).

This means that the coordination of {Cp*Ir}²⁺ “polarizes” the Lindqvist structure in such a way that more negatively charged oxygen atoms belong to the face opposite to the {Cp*IrNb₃O₃} group. This perhaps explains preferential formation of the trans-isomers observed for the bicapped [{XM}₂M₆O₁₉]⁴⁻ species.^9–11

Capping of [Nb₆O₁₉]⁸⁻ with half-sandwich moieties also significantly increases the positive charge at the Nb atoms (by almost 0.2 e). This may lead to stabilization of an intermediate with CN 7, where a CH₃OH molecule is coordinated to Nb; in fact there are examples of POMs with seven-coordinated Nb.¹⁶ Following this, we can envision proton transfer to one of the terminal or bridging oxides, followed by substitution of CH₃O⁻ for OH⁻. We have calculated total energies and Hirshfeld charges for various isomers of monomethylated trans-[{Cp*Ir}₂Nb₆O₁₈(OCH₃)]³⁻. The formation of a single possible terminal methoxo complex by methylation of one of the six equivalent terminal Nb=O groups (Fig. 4, isomer I) is slightly favored over the bridging isomer where methylation occurs at either of the six equivalent Nb–O–Nb groups (isomer II). The isomer with capping μ₃-OCH₃ (isomer III) is the least favored (Table S3†).

This is in agreement with Errington’s isolation of (Bu₄N)₂[(MeO)NbW₅O₁₈] with a terminal Nb–OCH₃ group,¹² but does not exclude the presence of methoxo-bridged isomers in equilibrium, as can be seen from the NMR spectra in our case.

To conclude, exclusive solubility of the hybrid organometallic–POM complexe salts with alkali metal cations in methanol is due to the formation of the methoxo complexes (as detected in solution) and solvation of Na⁺ and K⁺ (as is seen in the crystals of the solids which were isolated from such solutions), and hydrophobic interactions with Cp* and C₆H₆ moieties. This, prima facie, does not explain the lack of solubility in ethanol, which is a better donor (being more basic). But CH₃OH strongly differs from C₂H₅OH in having a significantly larger value of the dielectric constant (ε), of 31 vs. 24.

Thus methanol is better suited for dismantling the crystal lattice of these ionic compounds due to a sufficiently high ε value. The packing in the crystals of hydrates such as trans-K₄[{Cp*Rh}₂Nb₆O₁₉]·20H₂O (ref. 9) and trans-K₄[{Cp*Ir}₂Nb₆O₁₉]·22H₂O (ref. 10) consists of the layers formed by hybrid anions and solvated cations which stack together only through π–π interactions between organic ligands. Methanol can penetrate between such layers and replace water in the coordination spheres of the cations.

Experimental

General procedures

The starting Na₇H[Nb₆O₁₉]·15H₂O was synthesized as described in the literature.¹⁷ [(C₆H₆)RuCl₂]₂ was obtained from the reaction of RuCl₃·xH₂O (Krastsvetmet) with 1,3-cyclohexadiene (Sigma Aldrich) by the standard procedure.¹⁸ K₄[trans-{Cp*Rh}₂Nb₆O₁₉]·20H₂O and K₄[trans-{Cp*Ir}₂Nb₆O₁₉]·22H₂O were prepared as described in the literature.^9,10 Other reagents were of commercial quality and used as purchased. Methanol was purified according to standard methodology. IR spectra (4000–400 cm⁻¹) were recorded on an IFS-85 Bruker spectrometer.

Electrospray ionization (ESI) mass spectra

Electrospray ionization (ESI) mass spectra were recorded on a QTOF Premier (quadrupole-T-wave-time-of-flight) instrument. The temperature of the source block was set to 100 °C and the desolvation temperature to 200 °C. A capillary voltage of 3.3 kV was used in the negative scan mode and the cone voltage was set to 10 V to control the extent of fragmentation of the identified species. Mass calibration was performed with a solution of sodium iodide in isopropanol : water (50 : 50) in the mass range from m/z 50 to 3000. Sample solutions ca. 5 × 10⁻⁵ M in water or methanol were infused via a syringe pump directly connected to the ESI source at a flow rate of 10 μl min⁻¹. The observed isotopic pattern of each compound perfectly matched the theoretical isotope pattern calculated from their elemental composition using the MassLynx 4.1 program.

NMR experiments

¹H and ¹³C NMR spectra were recorded on a Bruker Avance III 500 spectrometer with inner references in CD₃OD at different temperatures.

DFT calculations

The experimentally determined geometries of Na₇H[Nb₆O₁₉]·14H₂O (ref. 19) and K₄[trans-{Cp*Ir}₂Nb₆O₁₉]·22H₂O (ref. 10) were used as the initial input for the geometry optimization of the anions. The converged molecular structures changed only slightly from the geometries found in the crystal. The structures with coordinated MeO ligands were constructed in silico. Quantum chemical calculations were carried out with the Amsterdam Density Functional (ADF2013.01.c) program (ADF2013, SCM, Theoretical Chemistry, Vrije Universiteit, Amsterdam, the Netherlands, http://www.scm.com). The calculations were performed with an all-electron Slater type TZ2P basis set. As with the geometry optimization, the VWN (local density approximation) and Becke–Perdew (generalized gradient approximation) functionals were used. Relativistic corrections were introduced by both scalar and spin–orbit relativistic zeroth order regular approximation (ZORA).

A Hirshfeld population analysis was performed and NMR spectra obtained according to the corresponding modules of the ADF program.

X-ray crystallography

Crystallographic data and refinement details are given in Table S2.† The diffraction data were collected on a Bruker Apex Duo (for 1) and Bruker X8 Apex (for 2 and 3) diffractometers with MoKα radiation (λ = 0.71073 Å) by doing φ and ω scans of narrow (0.5°) frames at 100 K. Structures were solved by direct methods and refined by full-matrix least-squares treatment against |F|² in an anisotropic approximation in SHELX 2014/7 with a ShelxLe program.²⁰ Absorption corrections were applied empirically with the SADABS program.²¹ All non-hydrogen atoms of main structural units were refined anisotropically. Positions of all cations in both structures were fully occupied, which excluded models with protonated polyoxoanions. Further details may be obtained from the Cambridge Crystallographic Data Center on quoting the depository numbers CCDC 1423580–1423582.

Na₄trans-[{(C₆H₆)Ru}₂Nb₆O₁₉]·14.125MeOH·2H₂O (1). 0.4 g (0.3 mmol) of Na₇H[Nb₆O₁₉]·15H₂O was dissolved upon heating (80–90 °C) and stirring in 8 mL of distilled water. To the clear solution 0.154 g (0.3 mmol) of crude [(C₆H₆)RuCl₂]₂ was added. The mixture was kept at 80 °C overnight upon stirring producing a clear yellow-orange solution. After cooling, the crude product (0.480 g) was precipitated by addition of 20 mL of acetone and isolated by filtration. Then the product was dissolved in 10 mL of freshly distilled methanol under stirring and heating at 50 °C. If some insoluble material appeared it was filtered off and discarded. Slow evaporation of the clear solution gives nice rhombohedral orange crystals of 1, which were characterized with XRD. Afterwards the filtration crystals were washed with cold methanol (2 × 5 ml) and carefully dried in vacuo. The solid material is very hygroscopic and absorbs water when kept in air. Yield: 80%. TGA: weight loss corresponds to ca. 12 CH₃OH. Analysis calc. for Na₄C₁₂H₁₂Ru₂Nb₆O₁₉(CH₃OH)₂(H₂O)₂C, H (%): 11.91, 1.71; found C, H (%): 11.53, 1.55. IR (KBr, cm⁻¹): 3400(s), 2922(m), 2853(m), 1628(m), 1431(m), 1139(w), 1012(sh), 867(s), 766(sh), 622(s), 609(sh), 530(s), 465(s).

K₄trans-[{Cp*Rh}₂Nb₆O₁₉]·4MeOH·10H₂O (2). 0.1 g (0.05 mmol) of K₄trans-[{Cp*Rh}₂Nb₆O₁₉]·20H₂O was dissolved in 5 mL of freshly distilled methanol upon gentle heating (ca. 50 °C). To the clear solution 30 mg of dibenzo-18-crown-6 was added, and after that the solution was allowed to evaporate slowly in air at 2 °C. Large yellow rhombohedral crystals of 2 were collected overnight. Drying the sample in vacuo gives a sample without methanol molecules. The solid material is very hygroscopic and absorbs water when kept in air. Yield: 80%. Analysis calc. for K₄C₂₀H₅₀Rh₂Nb₆O₂₉ C, H (%): 14.31, 3.04; found C, H (%): 14.02, 2.84. IR (KBr, cm⁻¹): 3455(s), 3358(s), 2922(m), 1651(m), 1459(m), 1387(m), 1190(m), 1085(m), 1023(m), 852(s), 782(s), 656(s), 606(s), 549(s), 506(m), 411(s).

K₄trans-[{Cp*Ir}₂Nb₆O₁₉]·10MeOH·4H₂O (3). Complex 3 was prepared in the same way as 2. Drying the sample in vacuo causes a loss of approximately eight methanol molecules. The solid material is very hygroscopic and absorbs water when kept in air. Yield: 80%. Analysis calc. for K₄C₂₂H₄₆Ir₂Nb₆O₂₅ C, H (%): 14.61, 2.52; found C, H (%): 14.22, 2.13. IR (KBr, cm⁻¹): 3372(s), 2915(m), 2834(m), 1621(m), 1435(m), 1382(m), 1080(w), 1024(m), 875(s), 850(s), 776(s), 670(s), 611(s), 524(s), 500(s), 409(s).

Conclusions

The hybrid complex A₄[{LM′}₂M₆O₁₉] ({LM′} = {Cp*Rh}²⁺, {Cp*Ir}²⁺, {(C₆H₆)Ru}²⁺; M = Nb, Ta; A = Na, K, Cs) is soluble only in a single organic solvent, namely CH₃OH. ESI-MS and NMR data of the methanolic solutions confirm the presence of methoxo complexes [{LM′}₂M₆O_19−n(OCH₃)_n] (n = 1–3). NMR experiments show that about half of the starting [{LM′}₂M₆O₁₉]⁴⁻ is present as methylated derivatives. Crystallization of the solutions of Na₄trans-[{(C₆H₆)Ru}₂Nb₆O₁₉], K₄trans-[{Cp*Rh}₂Nb₆O₁₉]·20H₂O and K₄trans-[{Cp*Ir}₂Nb₆O₁₉]·22H₂O in methanol leads to isolation of Na₄trans-[{(C₆H₆)Ru}₂Nb₆O₁₉]·14.125MeOH·2H₂O (1), K₄trans-[{Cp*Rh}₂Nb₆O₁₉]·4MeOH·10H₂O (2) and K₄trans-[{Cp*Ir}₂Nb₆O₁₉]·10MeOH·4H₂O (3) where direct coordination of CH₃OH to Na⁺ or K⁺ is observed. The crystal structure of 1 is the first example of a 3D framework built from dimeric [(MeOH)₃Na(μ-MeOH)Na(H₂O)(MeOH)₂]²⁺ cations and hybrid anions, which is typically new in the case of all known crystal structures of hybrid complexes. Its stability and sorption properties are under investigation.

DFT calculations were used for comparison of the ¹³C NMR shifts of different positions of methoxo ligands in the [{Cp*Ir}Nb₆O₁₈(OCH₃)]⁵⁻ backbone. Comparison of the energies of [{Cp*Ir}Nb₆O₁₈(μ-OCH₃)]⁵⁻ and [{Cp*Ir}Nb₆O₁₈(OCH₃)]⁵⁻ gives very close equal values suggesting that rapid exchange of the MeO⁻ ligand between different positions can be very fast in the solution.

Acknowledgements

This research was supported by Russian Science Foundation (grant number RScF 14-13-00645). The authors are also grateful to the SCIC of the Universitat Jaume I for providing mass spectrometric facilities.

Notes and references

1. M. Nyman, Dalton Trans., 2011, 40, 8049; H.-L. Wu, Z.-M. Zhang, Y.-G. Li, X.-L. Wang and E.-B. Wang, CrystEngComm, 2015, 17, 6261; P. A. Abramov, M. N. Sokolov, S. Floquet, M. Haouas, F. Taulelle, E. Cadot, E. V. Peresypkina, A. V. Virovets, C. Vicent, N. B. Kompankov, A. A. Zhdanov, O. V. Shuvaeva and V. P. Fedin, Inorg. Chem., 2014, 53, 12791; P. A. Abramov, C. Vicent, N. B. Kompankov, A. L. Gushchin and M. N. Sokolov, Chem. Commun., 2015, 51, 4021.
2. J.-H. Song and W. H. Casey, Chem. Commun., 2015, 51, 12744; J.-H. Song, D. H. Park, D. A. Keszler and W. H. Casey, Chem.–Eur. J., 2015, 21, 6727; J.-H. Song, J. R. Wang and W. H. Casey, Dalton Trans., 2014, 43, 17928; J.-H. Song, C. A. Ohlin, R. L. Johnson, P. Yu and W. H. Casey, Chem.–Eur. J., 2013, 19, 5191; J.-H. Song, C. A. Ohlin, E. C. Larson, P. Yu and W. H. Casey, Eur. J. Inorg. Chem., 2013, 1748.
3. E. J. Graeber and B. Morosin, Acta Crystallogr., 1977, 33, 2137; E. M. Villa, C. A. Ohlin, E. Balogh, T. M. Anderson, M. D. Nyman and W. H. Casey, Angew. Chem., Int. Ed., 2008, 47, 4844.
4. M. Maekawa, Y. Ozawa and A. Yagasaki, Inorg. Chem., 2006, 45, 9608.
5. M. Matsumoto, Y. Ozawa, A. Yagasaki and Y. Zhe, Inorg. Chem., 2013, 52, 7825.
6. D. Laurencin, R. Thouvenot, K. Boubekeur and A. Proust, Dalton Trans., 2007, 1334.
7. V. W. Day, W. G. Klemperer and C. Schwartz, J. Am. Chem. Soc., 1987, 109, 6030.
8. C. Hill and H. Yugi, J. Am. Chem. Soc., 1993, 25, 11823.
9. P. A. Abramov, M. N. Sokolov, A. V. Virovets, S. Floquet, M. Haouas, F. Taulelle, E. Cadot, C. Vicent and V. P. Fedin, Dalton Trans., 2015, 44, 2234.
10. P. A. Abramov, C. Vicent, N. B. Kompankov, A. L. Gushchin and M. N. Sokolov, Eur. J. Inorg. Chem., 2016, 154–160.
11. P. A. Abramov, T. P. Zemerova, N. K. Moroz, N. B. Kompankov and M. N. Sokolov, Inorg. Chem., 2016 DOI:10.1021/acs.inorgchem.5b01806.
12. W. Clegg, M. R. J. Elsegood, R. J. Errington and J. Havelock, J. Chem. Soc., Dalton Trans., 1996, 681.
13. E. Pretsh, T. Clerc, J. Seibl and W. Simon, Tables of Spectral Data for Structure Determination of Organic Compounds, Springer Verlag, Berlin, 2nd edn, 1989; C. Tsiao, D. Corbin and C. Dybowski, J. Am. Chem. Soc., 1990, 112, 7140; D. Crans, R. Felty, H. Chen and H. Ectert, Inorg. Chem., 1994, 33, 2427.
14. A. Karaliota, M. Kamariotaki and D. Hatzipanayioti, Transition Met. Chem., 1997, 22, 411.
15. L. Hubert-Pfalzgraf and J. Riess, Inorg. Chim. Acta, 1980, 41, 111; L. Hubert-Pfalzgraf, M. Postel and J. Riess, in Comprehensive Coordination Chemistry Pergamon, ed. G. Wilkinson, Oxford, 1987, vol. 3, pp. 585–698, and ref. therein; L. Hubert-Pfalzgraf, Polyhedron, 1994, 13, 1181; F. A. Cotton, M. Diebold and W. Roth, Inorg. Chem., 1987, 26, 3319; F. A. Cotton, M. Diebold and W. Roth, Inorg. Chem., 1987, 26, 3323.
16. R. Tsunashima, D. L. Long, H. Miras, D. Gabb, C. P. Pradeep and L. Cronin, Angew. Chem., Int. Ed., 2010, 49, 113 ; in solid state see for example: L. Jahnberg, J. Solid State Chem., 1970, 1, 454; L. Jahnberg, Mater. Res. Bull., 1981, 16, 513; N. Masó, D. I. Woodward, P. A. Thomas, A. Várez and A. R. West, J. Mater. Chem., 2011, 21, 2715; R. Hofmann and R. Gruehn, Z. Anorg. Allg. Chem., 1990, 590, 81; J. Busch and R. Gruehn, Z. Anorg. Allg. Chem., 1994, 620, 1066.
17. C. M. Flynn and G. D. Stucky, Inorg. Chem., 1969, 8, 178.
18. M. A. Bennett, T.-N. Huang, T. W. Matheson and A. K. Smith, Inorg. Synth., 1982, 21, 74.
19. A. Goiffon, E. Philippot and M. Maurin, Rev. Chim. Miner, 1980, 17, 466.
20. C. B. Hübschle, G. M. Sheldrick and B. Dittrich, J. Appl. Crystallogr., 2011, 44, 1281.
21. APEX2 (Version 1.08), SAINT (Version 7.03), and SADABS (Version 2.11), Bruker Advanced X-ray Solutions, Bruker AXS Inc., Madison, Wisconsin, USA, 2004.

---

Footnote

† Electronic supplementary information (ESI) available: ESI-MS data, NMR data, crystal packing information, details of X-ray experiments and refinement. CCDC 1423580–1423582. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c5ra23918d

---