Introduction

For transition metals of the groups 3–8, tri- and tetranuclear cluster chalcogenide complexes can be obtained, which can be described as having a triangle (Ti, V, Mo, W, Re, Os) or square (Ln, Ta) of metal atoms, capped by a single chalcogen (Q) atom (μ3 in the triangular and μ4 in the square planar clusters), and bridged over each side by three (triangular clusters) or four (square planar clusters) tilted μ2-dichalogenide ligands, so that M3Q n+7 or M4Q n+9 cluster cores result. These clusters have therefore three different chalcogen sites (one capping, one bridging, residing almost in the M3 (or M4) plane (so-called equatorial position) and another bridging above the same plane, on the opposite side from the capping chalcogen (termed axial chalcogens). Thus each μ2-Q2 ligand is represented as μ2-QeqQax. In most cases these clusters are obtained by self-assembly reactions [1–4]. This method of preparation poses an interesting problem: will this self-assembly be site-differentiating when two different chalcogen enter the core, preferentially taking a specific position each, or clusters with randomly distributed chalcogen atoms in the core would form? In fact for the trinuclear clusters it seems that the smaller chalcogen always occupies the capping position. This has been proven for the Re2O7/OsO4–S–SeCl2 systems which always give the {M3(μ3-S)(μ2-Se2)3} clusters [5], and for the Mo(W)–S–Se–Br2 systems which selectively produce {M3(μ3-S)(μ2-Se2)3} clusters [A. L. Gushchin and M. N. Sokolov, unpublished results]. It turns out now that the preference of S for the capping position holds also for square clusters. Here we describe a new tetranuclear Ta cluster, Ta4S1.5Se7.5I8, which contains the planar Ta4 core bridged by one μ4-S atom, and by four μ2-SeQ (Q ≈ Se0.87S0.13) units.

Experimental Section

General Procedures

High purity Ta and S powders, Se granules and I2 crystals were used. Raman spectra were obtained by means of a Triplimate SPEX spectrometer with a 632.8 nm line of He–Ne laser for excitation. X-ray powder diffraction data were obtained on a DRON-2 powder diffractometer (CuKα radiation).

Preparation of Ta4S1.5Se7.5I8 (1)

Ta powder (0.36 g, 2.0 mmol), S (0.016 g, 0.5 mmol), Se (0.32 g, 4.0 mmol) and small excess of I2 (0.55 g, 2.2 mmol) were loaded in a glass ampoule, which was evacuated, flame sealed and heated at 300 °C (4 days) in a furnace with a small natural temperature gradient. A crop of large single crystals together with a fine powder were obtained. The powder was sifted out leaving the crystals of 1. The yield was 50%. Element ratio: Ta4.0S1.4Se7.7I7.9 (EDAX). Raman (cm−1): 396w (Ta4–μ4-S), 303w, 294w, 207m, 204sh, 187m, 159w, 142s, 140sh, 109s, 94s, 86vs, 73s, 70s, 65s, 59m.

Electrochemistry

As the title complex is not soluble in common solvents the electrochemical behavior was studied by the method of immobilized solid particles [6]. The cyclic volatmmograms were recorded on a 797 VA Computrance setting (Metrohm, Switzerland). A 10 mL-volume three-electrode cell was employed. As main electrode a paraffin-impregnated graphite (PIGE) with solid particles of the complexes under investigation immobilized on it. Immobilization was achieved by rubbing crushed crystals of the complexes into the end surface of PIGE. More details concerning electrode preparation ate to be found in [7, 8]. An Ag/AgCl reference electrode, filled with 3 M KCl, was used, and auxiliary electrode was a Pt wire (6.0343 Metrohm). Background electrolyte was 0.1 M KCl, made by dissolving potassium chloride (Ultrapure) in redistilled water.

X-ray Crystallography

The diffraction data were collected on a Bruker X8APEX CCD diffractometer with MoKα radiation (λ = 0.71073 Å) using φ-scans of narrow (0.5°) frames. The structure was solved by direct methods and refined by full-matrix least-squares method against |F|2 in anisotropic approximation with SHELXTL programs set. Absorption correction was applied empirically with SADABS program (T min/T max = 0.510) [9–11]. The detailed data are collected in Table 1.

Table 1 Crystallographic data and details of diffraction experiment for Ta4S1.5Se7.5I8
Full size table

Results and Discussion

The title compound is obtained by heating the elements in the required stoichiometric ratio at 300 °C for 4 days. It is easily separated as large black single crystals. The exact composition was determined from the X-ray data. Ta4S1.5Se7.5I8 is the first compound obtained in the system Ta–S–Se–I2. The structural analogues of 1, the selenoiodide Ta4Se9I8 (2) and the thiobromide Ta4S9Br8 (3) were recently obtained from the elements [1, 2].

The molecule of Ta4S1.5Se7.5I8 is shown in Fig. 1. The molecular structure is identical to those of 2 and 3. Four Q2 (Q=Se, SeS) ligands are asymmetrically coordinated to the Ta–Ta edges in the μ2–η2:η2 manner. The equatorial chalcogen atoms lie almost in the Ta4 plane, and the axial ones deviate from the plane in the opposite direction. The terminal iodine atoms are coordinated slightly asymmetrically. The interatomic distances and some angles are summarized in Table 2. The coordination polyhedron around Ta can be described as pentagonal bipiramid with μ4-S (100 % sulfur occupancy) and one of the I atoms trans to it in the axial position (I trans ), and the two groups Q2 and another I atom (I cis ) in the equatorial position. Rather long Ta–Ta distances (3.26–3.32 Å) are in agreement with the average Ta oxidation state of +4.5 in this highly electron-deficient cluster. In Ta4Se9I8 and Ta4S9Br8 the distances between the Ta atoms are 3.32–3.39 and 3.30 Å, respectively [1, 2]. The fact that the Ta–Ta distances in 1 and 3 are almost identical shows that the nature of the μ4-Q atom is more decisive for the M–M distance than that of the Q2 bridge. The Ta4–μ4-S bond length (2.43–2.46 Å) is practically the same as found in Ta4S9Br8 (2.46 Å) (Table 2). The Se–Se distances correspond to single bond expected for the Se2 2− formalism.

Fig. 1
figure 1

Molecular structure of Ta4S1.5Se7.5I8 (1) (ellipsoids of 50% probability level)

Full size image
Table 2 Selected bond lengths [Å] for Ta4S1.5Se7.5I8 (1), Ta4Se9I8 (2) and Ta4S9Br8 (3)
Full size table

The intermolecular interactions in the crystals of Ta4S1.5Se7.5I8 and Ta4Se9I8 are identical. In both 1 and 2 four shortened non-valent contacts Seax…I (3.48–3.59 Å) are all directed to the same iodine atom so that the molecules are joined into zigzag chains (Fig. 2).

Fig. 2
figure 2

The chains of molecules in the crystal packing of Ta4S1.5Se7.5I8. The short interatomic Se…I contacts are shown as dashed lines

Full size image

The Raman spectrum of Ta4S1.5Se7.5I8 shows a weak band at 396 cm−1 that comes from Ta–μ4-S vibrations. The position of this band correlates well with the positions of this vibration in Ta4S9Br8 (407 cm−1) [1].

Electrochemistry

The cyclic voltammograms of Ta4S1.5Se7.5I8 (1) mechanically immobilized on the surface of PIGE at scan rate of 20 mV s−1 are shown in Fig. 3. In 0.1 M KCl it exhibits a response of redox couple at E1/2 = (E am  + E cm ) = 0.522 V, which is due to the transitions in the metallic core [Ta4]18+/(18+n). In order to calculate the number of electrons (n) involved in the rate-determining step, a Tafel plot (Fig. 4) was drawn from background-corrected data taken from the rising part of anodic current-potential curves (where there is no concentration polarization). As is seen from Fig. 4 the Tafel plot indicates one-electron process in the rate-limiting step, assuming transfer coefficients of β = 0.5 with error not exceeding 2% (the theory requires 0.120 V/decade slope for one-electron transfer). The Tafel plot was also constructed according to the equation valid for a totally irreversible diffusion process [12], which gives E1/2 = (b/2) log v + Const., where b is the Tafel slope. On the basis of this equation, the slope of Em vs lg v is b/2 = ∂Em/∂lg v. Figure 5 shows cyclic voltammograms of Ta4S1.54Se7.46I8 with different scan rates and plot a maximum potential vs logarithm of scan rate for the anodic current (see inset). The ∂Em/∂lg v is 0.0579 V/decade, which gives b = 2 × 0.0579 = 0.116 V/decade. This b value also points to one-electron transfer to be a rate-limiting step, if we assume transfer coefficient β = 0.5, with error not exceeding 2%.

Fig. 3
figure 3

Cyclic voltammograms of Ta4S1.5Se7.5I8 (1) mechanically immobilized on the surface of PIGE: background electrolyte 0.1 M KCl (dashed line); potential route −0.7 → 0.8 → −0.7 V; scan rate 20 mV s−1

Full size image
Fig. 4
figure 4

Tafel plot obtained from anodic current-potential curves recoded for Ta4S1.5Se7.5I8 (1) as in Fig. 1

Full size image
Fig. 5
figure 5

Cyclic voltammograms of Ta4S1.5Se7.5I8 mechanically immobilized on the surface of PIGE (solid line): background electrolyte 0.1 M KCl; potential route −0.3 → 0.9 → −0.3 V; scan rate 10 (—), 20 (— —) and 50 (····) mV s−1. (Points denote maximum of oxidation current). Inset plot of maximum potential (Em) versus logarithm of scan rate (v) for the oxidation current at the cyclic voltammograms

Full size image

Therefore, the redox-process may be described as:

$$ [{\text{Ta}}_{2}^{5 + } {\text{Ta}}_{2}^{4 + } ]^{18 + } \,\underset{{ + {\text{e}}^{ - } \left( {\text{slowly}} \right)}}{\overset{{ - {\text{e}}^ - \left( {\text{fast}} \right)}}{\longleftrightarrow}}[{\text{Ta}}_{2}^{5 + } {\text{Ta}}^{4 + } {\text{Ta}}^{5 + } ]^{19 + } $$

The results of the present study demonstrate that under self assembly conditions in a thermodynamically controlled high-temperature synthesis, the lighter chalcogen enters into a position of maximum connectivity. This agrees well with the observations made for the triangular clusters with the {M3(μ3-Q)(μ-Q2)3} core [4, 5; A. L. Gushchin and M. N. Sokolov, unpublished results]. Moreover, in the structure of Cs4[Re6S9.45Se3.55] the μ3 positions in the cluster {Re6(μ3-Q)8} are exclusively occupied by the sulfur atoms while Se participates in the μ-Q2 bridges between the octahedral clusters [13]. In high–temperature syntheses of cuboidal Re clusters, {Re4(μ3-Q)4} from the elements, when two different chalcogens are introduced in the system, the lighter chalcogen invariably occupies the inner ligand site in the cluster, as for example in Re4S4Te4 and in Re4S4Cl8(TeCl2)4 [14, 15]. Thus the preference of S over Se and Te for the bridging position of maximum connectivity seems to be a general rule. A possible explanation for this systematic preference may be that the smallest size of the S maximizes the M–M bonding which is important for the overall cluster stability.

Supporting Information Available

Crystallographic data in CIF format have been deposited at Fachinformationszentrum Karlsruhe under ICSD number 419404 and can be retrieved be request (crystaldata@FIZ-Karlsruhe.de).