Comment

A polyoxometalate is an aggregate of oxoanions of group V and VI elements such as Mo, W and V. The introduction of peroxide groups to a polyoxometalate leads to the formation of a peroxidopolyoxometalate. The chemistry of peroxidopolyoxometalates has not been widely explored thus far, probably due to the instability of peroxide groups in solution or the solid state. In an attempt to synthesize organic-inorganic hybrid peroxidopolyoxometalates, we obtained a formate complex of a dimeric peroxidomolybdate, namely

-(formato-²O:O')--oxido-bis[oxidobis(peroxido-²O,O')molybdate(IV)], as the potassium, (I) (Fig. 1), rubidium, (II) (Fig. 2), and ammonium, (III) (Fig. 3), salts.

In all three salts, the structure of the complex anion consists of two oxidodiperoxidomolybdate units bridged by a formate ligand in a ₂-manner and by a ₂-oxide ligand. The fundamental structure is similar to the acetate complex [Mo₂(CH₃COO)O₃(O₂)₄]^3- (Hou et al., 2003). Each Mo atom bears two peroxide groups in its equatorial plane to form a pentagonal-bipyramidal geometry. A terminal and a formate O atom are located at the axial positions of the decahedron, and two decahedra are linked by a singly bridging oxide O atom. This dimeric peroxidomolybdate has roughly the same structure as [Mo₂O₃(O₂)₄(H₂O)₂]^2- (Stomberg, 1968; Le Carpentier et al., 1972; Djordjevic et al., 1989), but there are no water molecules and the Mo decahedra are rotated around the bridging O atom to make the coordination of formate possible at their vacant positions. The formate group is a bipodal O:O'-bridging ligand. The Mo-O(formate) bond lengths are long in all three salts due to the trans influence of the terminal O atom (Tables 1-3).

The unit-cell parameters of these three salts are very similar. However, (III) refined in a centrosymmetric space group (Pbcm), whereas (I) and (II) are essentially isostructural in the noncentrosymmetric space group Pca2₁. It was not impossible to apply the Pca2₁ space group to (III), but the refinement did not converge and gave a higher R value and goodness-of-fit. As a result, there is a mirror plane passing through atoms O6, C1 and H1 and perpendicular to the O7-Mo1-Mo1ⁱ-O7ⁱ plane in the anion of ammonium salt (III) [symmetry code: (i) x, y, -z + ], and the macrocycle in the anion is planar, in contrast with the other two salts where the macrocycle is twisted. The (formate)O-Mo-Mo-O(formate) torsion angle is 17.32 (6)° in (I) and 14.81 (11)° in (II). In contrast, atoms O6 and C1 of (III) are displaced by 0.208 (2) and 0.066 (2) Å, respectively, from the least-squares O7-Mo1-Mo1ⁱ-O7ⁱ plane. The reason for this conformational and crystallographic difference is uncertain.

All three crystal structures show quite similar packing modes. As can be seen in Figs. 4 and 5, the arrangements of the anions and cations are quite similar. The anions in (I) and (II) have a slight zigzag arrangement along the c axis, in contrast with the straight arrangement found in (III), reflecting the difference in the space groups. The shortest axes (b for the potassium and rubidium salts, and a for the ammonium salt) lengthen in the order K < NH₄ < Rb, whereas for the other two axes the order is K < Rb < NH₄. The unit-cell volume of (III) is slightly larger than that of (II). Such differences may cause a difference in the cation-anion interactions in these salts. Each K and Rb atom has eight to ten close contacts within 3.2 Å with O atoms of the anion, forming a three-dimensional network in the crystal structure. Similarly, all H atoms in the ammonium cations form hydrogen bonds with O atoms in the anion, resulting in a three-dimensional network. However, the ammonium cations in the general position and on the mirror plane interact with six and five different O atoms, respectively, fewer than the number of O atoms coordinating to the potassium and rubidium cations. Such a difference may also be caused by the differences in lattice parameters and the limited interaction ability due to the tetrahedral orientation of the H atoms in the ammonium cation, whereas the K and Rb cations are spherical.

Powder X-ray diffraction patterns were compared with simulated patterns calculated from the single-crystal results (Brandenburg, 2005) for each of the three salts, and all agreed well. This indicates that the crystals subjected to measurement here were representative of the bulk.

In the peroxidoisomolybdate system, at the high peroxide/Mo ratio used for the preparation of the current compounds, a monomeric diperoxidomolybdate and a dimeric tetraperoxidodimolybdate are important species in the solution (Taube, Hashimoto et al., 2002; Taube, Andersson et al., 2002). The diperoxidomonomolybdate and tetraperoxidodimolybdate anions show very broad ⁹⁵Mo NMR signals at about -262 and -280 p.p.m., respectively, at pH 3, as was employed for their preparation in the present work. The signals of these two species are not well separated and give one broad signal, and the positions of the signals are somewhat ambiguous. A formate-free solution, where nitric acid was used for pH adjustment and the other conditions were similar to those used for the preparation of the present compounds, gave a broad ⁹⁵Mo signal with a shoulder. It can be deconvoluted to -264 and -284 p.p.m., which are in good agreement with those reported for diperoxidomolybdate and tetraperoxidodimolybdate, respectively, by Taube and co-workers (Taube, Hashimoto et al., 2002; Taube, Andersson et al., 2002). When a controlled amount of formic acid was added to the solution, a signal at ca -216 p.p.m., probably due to the present formate complex, appeared and grew as the molar ratio of formic acid to the total concentration of molybdate was increased. The signals of the monomeric and dimeric species almost disappeared when 20 times the molar amount of formic acid versus molybdate was added to the solution, although the ratio of formate to molybdate is 0.5 in the complex formed here. The signal at -216 p.p.m. decreased significantly after crystallization of the complex, which also supports the above hypothesis that the signal was due to the present complex. However, these observations indicated that the coordination of formate to the -oxido-bis(oxidodiperoxidomolybdate) unit was rather weak, although the complexation behaviour could not be detected by ¹H and ¹³C NMR due to small shifts of the signals from their coordination-free positions.

Experimental

For the preparation of the potassium salt, (I), sodium molybdate dihydrate (5 g) was dissolved in water (30 ml). The pH was lowered to ca 4 using formic acid, and 30% hydrogen peroxide (5 ml) was added to the solution. The pH was then adjusted to 3 using formic acid. The volume of the solution was increased to 50 ml by the addition of water, and 1 M potassium chloride (50 ml) was added. The concentrations of molybdate and H₂O₂ were ca 0.2 M and ca 0.5 M, respectively, in the final solution. The concentration of formate was not controlled because formic acid was used for the pH adjustment. The mixture was kept at 278 K. Yellow block-shaped crystals of (I) appeared after 1 d (yield 3.03 g, 55.2% based on Mo).

The rubidium salt, (II), and the ammonium salt, (III), were obtained in a similar manner, but using 1 M rubidium chloride and 1 M ammonium chloride, respectively. Yellow block-shaped crystals were obtained after a few days (yield of rubidium salt: 4.93 g, 71.2%; yield of ammonium salt: 1.53 g, 31.7%).

Caution: There is a risk of explosion or combustion by smashing, striking or grinding crystals of these compounds.

Caesium gives crystals of the same anion by a similar preparative method with CsCl as the cation-supplying reagent. However, very poor quality crystals (severely stacked thin plates) obviated single-crystal X-ray data collection, and all attempts to obtain suitable crystals for single-crystal work have failed. The ¹³C CPMAS NMR of the caesium salt gave a single resonance at 168.0 p.p.m. Comparison of its powder X-ray diffraction pattern with those of the other salts indicates that the caesium salt has a completely different crystal structure from the K, Rb and NH₄ salts. Preliminary X-ray measurement of the Cs salt reveals that the most probable space group was P321, with a = 6.731 (2) Å, c = 17.044 (8) Å, R1 [I > 2(I)] = 0.1238 and wR2 (all data) = 0.3617, and the structure shows severe disordering of anions with the Mo atoms on threefold axes (Wyckoff notation c). A powder pattern of the Cs salt simulated from the `single'-crystal result agrees well with that of the measured data.

Refinement of the absolute structure parameter for (I) and (II) using the TWIN/BASF instructions of SHELXL97 (Sheldrick, 2008), gave values of 0.44 (4) and 0.465 (6), respectively, which indicated that both structures were inversion twins.

The formate H atom in each compound was located geometrically and the ammonium H atoms were located from a differential Fourier map. In the final refinements, the positional parameters of the formate H atom in (I) and (II) were fixed, but those of (III) were refined. The U_iso values of the formate H atoms in (I)-(III) were fixed at ca 1.5 times the U_eq value of the attached C atom. The positional and displacement parameters of all the ammonium H atoms were refined isotropically.

For all compounds, data collection: CrystalClear (Rigaku, 2008); cell refinement: CrystalClear; data reduction: CrystalClear; program(s) used to solve structure: SHELXS97 (Sheldrick, 2008); program(s) used to refine structure: SHELXL97 (Sheldrick, 2008); molecular graphics: ORTEP-3 for Windows (Farrugia, 1997); software used to prepare material for publication: SHELXL97.

---

Supplementary data for this paper are available from the IUCr electronic archives (Reference: SF3186 ). Services for accessing these data are described at the back of the journal.

---