Synthesis, crystal structure and selected properties of K₂[Ni(dien)₂]{[Ni(dien)]₂Ta₆O₁₉}·11 H₂O

2.1 General synthetic aspects

All chemicals except K₈{Ta₆O₁₉}·16 H₂O were purchased from commercial sources and used without further purification: Diethylenetriamine (dien) >98%, Sigma Aldrich; Ta₂O₅: 99% Ta, Alfa Aesar; Cu(NO₃)₂·3H₂O: >99%, Merck; Ni(NO₃)₂·6H₂O: >98%, Merck; KOH: 85%, abcr, DMSO: 99%, Grüssing). The water soluble precursor K₈{Ta₆O₁₉}·16 H₂O was prepared according to ref. [40] and the synthetic method has been described in literature [31].

2.2 Synthesis of K₂[Ni(dien)₂]{[Ni(dien)]₂Ta₆O₁₉}·11 H₂O

0.15 mmol Ni(NO₃)₂·6H₂O and 0.15 mmol diethylenetriamine (dien) were dissolved in a solution of 4 mL DMSO and 1 mL H₂O and subsequently transferred into a 20 mL glass tube. Then 4 mL of a 3:1 DMSO: H₂O mixed solvent and a solution of 0.05 mmol K₈{Ta₆O₁₉}·16 H₂O (pH 12.0) in 4 mL H₂O were added slowly one after the other into the tube. The tube was capped with parafilm and was left undisturbed at room temperature. After a few weeks, light blue crystals were obtained. The crystals were filtered off and washed with mother liquor only. The yield was about 15% based on Ta. – Elemental analysis: Calcd. for C₁₆H₇₄K₂N₁₂Ni₃O₃₀Ta₆: C 8.5, H 3.3, N 7.5; found C 8.2, H 3.4, N 7.2%.

2.3 Characterization methods

The powder X-ray diffraction (PXRD) measurements were done in transmission geometry on a STOE STADI–P diffractometer (CuKα ₁ radiation (λ = 1.540598 Å), Ge (111) monochromator, MYTHEN 1 K detector. X-ray powder patterns were measured as flat samples, except for the compounds obtained from the TG experiments. These were sealed in glass capillaries (Hilgenberg, 0.2 mm outer diameter) and data was collected in Debye–Scherrer geometry. The PXRD pattern of the phase pure title compound is shown in Figure S1 (Supplementary material available online).

Indexing (using singular value decomposition [41]), Pawley fits, structure solution and Rietveld refinements were carried out with Topas (academic version 6.0) [42].

A sample of the partly rehydrated sample was obtained by heating the title compound to 170 °C in the DTA-TG apparatus and allowing it to age for 7 days. Indexing of all reflections in the powder pattern leads only to low figures of merit. However, after omitting all reflections assigned to the title compound, indexing yielded numerous triclinic unit cells with reasonable cell volumes and good figures of merit. A cell with similar metrics as that of the title compound was chosen (a = 10.5135(3), b = 10.3331(3), c = 12.9878(4) Å, α = 66.880(2)°, β = 75.803(2)°, γ = 75.893(2)°). The structure was solved in real space using the global optimization method of simulated annealing as implemented in Topas [43]. It was assumed that the major structural motifs remained intact and only water was removed, and rigid bodies were set up to describe the [Ni(dien)₂]²⁺, [Ni(dien)]²⁺ and {Ta₆O₁₉}^8– moieties, the K⁺ cation being translated unrestricted. The [Ta₆O₁₉]^8– anion was placed on the inversion center of the chosen space group P 1 ‾ , since only one formula unit can be present in the unit cell. After the first run the Ni²⁺ cation of the [Ni(dien)₂]²⁺ complex was found on a special position, further reducing the number of free parameters. After the second run the [Ni(dien)]²⁺ fragment was located which is attached to the {Ta₆O₁₉}^8– moiety. In the next step, the positions of the rigid bodies and the K⁺ ion were subjected to a Rietveld refinement. Inspection of the Fourier map did not allow assigning the residual electron density to meaningful positions for further oxygen atoms (i.e., water molecules). This is not surprising taking into account the low scattering power of oxygen compared to tantalum and the potential disorder regarding the orientation of the complexes and water molecules when taking into account the synthesis conditions.

CCDC 2106440 (partial rehydrated sample), contain the supplementary crystallographic data for this paper. These data can be obtained free charge from the Cambridge Crystallographic Data Centre via http://www.ccdc.cam.ac.uk/data_request/cif.

Infrared (IR) spectra were recorded on an ALPHA-P FT-IR spectrometer (Bruker) in a range of 375–4000 cm⁻¹. CHN analysis was made with an EURO EA 3000 elemental analyzer from EURO VECTOR instruments and related software. For collection of UV/Vis diffuse-reflectance spectra an UV/Vis two-channel spectrometer Cary 5 (Carian Techtron Pty., Darmstadt) was used with BaSO₄ as white standard.

DTA-TG curves were measured on a Netzsch STA 409 CD instrument with a heating rate of 4 K min⁻¹ in air. For the calibration, standard reference materials were used.

Water sorption experiments were carried out with a Belsorp_max instrument (BEL JAPAN INC.) used at T = 298 K. Prior to the measurement the crystal water of the sample was removed by heating at 170 °C under reduced pressure for 5 h.

2.4 Single-crystal structure determination

The intensity data was collected on a STOE IPSD-2 diffractometer with MoKα radiation (λ = 0.71073 Å) at T = 200(2) K. The structure was solved with Shelxs-97 [44] and refined against F ² using Shelxl-2014 [45]. A numerical absorption correction was performed (T _min/_max: 0.093/0.216). All atoms were on general positions and all non-hydrogen atoms were refined anisotropically. C–H and N–H hydrogen atoms were positioned with idealized geometry and were refined isotropically by using a riding model (U _iso(H) = 1.2 U _eq(C, N)). The water H atoms were not located, but considered in the calculation of the molecular formula and the molecular mass. One C atom of an ethylene diamine ligand as well as some of the water O atoms were found to be disordered and were refined using a split model. Selected crystallographic data and refinement results are listed in Table S1 (Supplementary material).

CCDC 2106439 (K₂[Ni(dien)₂]{[Ni(dien)]₂Ta₆O₁₉}·11 H₂O), contain the supplementary crystallographic data for this paper. These data can be obtained free charge from the Cambridge Crystallographic Data Centre via http://www.ccdc.cam.ac.uk/data_request/cif.

3.1 Description of the crystal structure

K₂[Ni(dien)₂]{[Ni(dien)]₂Ta₆O₁₉}·11 H₂O crystallizes in the triclinic space group P 1 ‾ with one formula unit and 11 lattice water molecules per unit cell. The main structural motif is the Lindqvist-type anion [Ta₆O₁₉]^8–, which is composed of six TaO₆ octahedra sharing common edges (Figure 1). Three different types of Ta–O bonds are identified: Ta–µ _t-O (1.787–1.793 Å); Ta–µ ₂-O (1.950–1.976 Å), and Ta–µ ₆-O (2.363–2.389 Å (Tab. S2; Supplementary material), with all geometrical data in agreement with literature values [16, 26, 31, 37, 46, 47]. Results of bond valence sum calculations [48], [49], [50] are in agreement with Ta being in the +V and Ni in the +II oxidation state (Tab. S3; Supplementary material).

Figure 1:

View of the cluster motif of K₂[Ni(dien)₂]{[Ni(dien)]₂Ta₆O₁₉}·11 H₂O (left) and representation of the octahedral coordination of both Ni²⁺ cations. All H atoms as well as the disorder of Ni1 (see Fig. S2; Supplementary material) are omitted for clarity.

The shell of the [Ta₆O₁₉]^8– anion is expanded by two symmetry related [Ni(dien)]²⁺ complexes (Ni1) which are covalently bonded to three µ ₂-bridging O²⁻ anions of the {Ta₆O₁₉} core to form a {[Ni(dien)]₂Ta₆O₁₉}^4– anion. The Ni–µ ₂-O bonds (Ni–O: 2.128(5)–2.152(5) Å) are slightly longer than the sum of ionic radii (Ni²⁺ _CN6 = 0.69 Å; O²⁻ = 1.35 Å) (Figure 1) [51]. Three N donor atoms of dien to form a distorted NiN₃O₃ octahedron complete the coordination environment of Ni1. The Ni–N bond lengths range from 2.062(6) to 2.087(6) Å, and both Ni–O and Ni–N bond lengths agree with literature data [21, 52], [53], [54], [55], [56]. Two symmetry related isolated [Ni(dien)₂]²⁺ complexes (Ni2) are observed in which the Ni²⁺ cation is octahedrally surrounded by six N atoms from two dien ligands (Figure 1). The axial Ni–N bonds of 2.094(7) Å are shorter than the equatorial Ni–N bonds of 2.135(7)–2.142(7) Å (Tab. S4; Supplementary material) [22, 57, 58]. In both complexes, the dien ligands adopt the s-fac conformation. The Ta–μ ₂O–Ni angles are between 96.17(19)° and 98.09(18)°, the N–Ni–µ ₂O angles range from 89.0(2)° to 173.9(2)° and N–Ni–N angles from 91.9(3)° to 180° (Tab. S5; Supplementary material).

The {[Ni(dien)]₂Ta₆O₁₉}^4– moieties form rods along [100], [010] and [001] (Figure 2). The empty space between the rods is filled by [Ni(dien)₂]²⁺ and K⁺ cations and a solvent accessible void of 166.3 Å³ [59] is occupied by crystal water molecules (Fig. S3; Supplementary material). The K⁺ cation is surrounded by μ ₂-O^2– anions of the [Ta₆O₁₉]^6– cluster and O atoms of crystal water molecules. The K–O distances (2.780(6)–3.354(9) Å) are longer than the sums of the ionic radii (K⁺ _CN6: 1.38 Å; K⁺ _CN7: 1.46 Å; K⁺ _CN8: 1.51 Å; O²⁻: 1.35 Å) [51] (Tab. S6; Supplementary material), but they are shorter than the sum of van der Waals radii of 4.24 Å [60] or 4.27 Å [61]. Common values for K–O bond lengths have values up to 3.2 Å as observed for KOH·4H₂O [62] or K₈{Ta₆O₁₉}·16 H₂O [46].

Figure 2:

Arrangement of rods of the {[Ni(dien)]₂Ta₆O₁₉}^4– moieties viewed along [111]. Note that only the anions are displayed. Color code see Figure 1.

For the title compound a clear gap in the K–O distance distribution appears at >3.12 Å, and using this value the K⁺ cation is surrounded by eight O²⁻ anions to form an irregular KO₈ polyhedron. Two KO₈ polyhedra build a K₂O₁₄ moiety by sharing a common edge acting as connecting unit between the {[Ni(dien)]₂Ta₆O₁₉}^4– moieties and forming chains along [010] (Figure 3). The chains are linked by hydrogen bonds (O⋯O: 2.52–2.92 Å) to H₂O molecules along [100] (Tab. S7; Supplementary material). The H₂O molecules are located in channels along [010] and form discrete units categorized as D2 groups [63, 64], which join two K₂O₁₄ moieties via K–O_water⋯H–O interactions generating a layer-like structure extending in the (001) plane (Figure 4).

Figure 3:

View of the environment of the K⁺ cation forming KO₈ polyhedra appearing as dinuclear units K₂O₁₄ units. View of the connection of the {[Ni(dien)]₂Ta₆O₁₉}^4– anions through K₂O₁₄ moieties and the resulting chain structure along the c axis (color code is in Figure 1, beige for: K₂O₁₄ ²⁺).

Figure 4:

View of the arrangement of the {[Ni(dien)]₂Ta₆O₁₉}^4– anions and [Ni(dien)₂]²⁺ complexes along [100] (top). View of the crosslinking crystal water molecules forming layers in the (001) plane (bottom). The dien ligands and H atoms are not shown for clarity (Color code see Figure 1, beige: K).

The layers are separated by 3.612 Å between the cluster cores along [010] and by 6.070 Å between diagonally arranged anions (Fig. S4; Supplementary material). The [Ni2(dien)₂]²⁺ complexes are located between the clusters and stabilize the layered structure via N–H⋯O_cluster interactions (distances: 2.920(7)–3.139(8) Å) thus generating a 3D network (Fig. S5; Supplementary material).

3.2 Spectroscopic properties

The combined MIR-FIR spectrum of the title compound (Fig. S6; Supplementary material) shows a broad and nearly featureless band between 3000 and 3500 cm⁻¹, which is typical for the O–H vibrations of H₂O involved in hydrogen bonding. The three weak signals at 3308, 3182 and 3140 cm⁻¹ are caused by the asymmetric and symmetric N–H stretching modes. The absorptions related to the CH₂ groups are observed as very weak bands at 2927, 2891 and 2860 cm⁻¹. The bending vibration of –NH₂ is located at 1600 cm⁻¹. An unambiguous assignment of further vibrations in the region from 900 to 1500 cm⁻¹ is not possible. The different stretching vibrations (Ta=O_t and Ta–O–Ta) of the cluster ion are located between 836 and 480 cm⁻¹, and the Ni–N vibrations are found at 388 cm⁻¹. All values are in good agreement with literature data [65]. The UV/Vis spectrum (Fig. S7; Supplementary material) shows the characteristic transitions for Ni²⁺ complexes in octahedral environment. The bands at 1.31 eV (10561 cm⁻¹), 1.90 eV (15324 cm⁻¹) and 3.48 eV (27825 cm⁻¹) are assigned to the spin-allowed d–d transitions ³A_2g(F) → ³T_2g(F), ³A_2g(F) → ³T_1g (F) and ³A_2g(F) → ³T_1g (P) [66]. The resulting crystal field splitting parameters D _q (1056 cm⁻¹) and B (887 cm⁻¹) are in accordance to those found in similar Ni²⁺ systems [66]. A small shoulder at 2.16 eV in the second band may be caused by the Ni–O bonds as observed for Ni²⁺ complexes with O²⁻ ligands [67, 68].

3.3 Thermoanalytical investigations

Heating of K₂[Ni(dien)₂]{[Ni(dien)]₂Ta₆O₁₉}·11 H₂O (Figure 5) leads to a weak endothermic event with the first weight loss up to about 200 °C (T _peak = 115 °C) of 8.9% corresponding to the removal of ∼10 H₂O molecules. A plateau extends from 175 to 255 °C followed by an exothermic thermal event (T _peak = 340 °C) accompanied by a mass loss of 17.2% most likely due to the removal of dien molecules. The PXRD characterization of the product recovered at 1000 °C indicates a mixture of well crystalline NiTa₂O₆ and K₆Ta_10.8O₃₀ (Fig. S8; Supplementary material).

Figure 5:

DTA-TG-DTG curves of K₂[Ni(dien)₂]{[Ni(dien)]₂Ta₆O₁₉}·11 H₂O, measured in air with 4 K min⁻¹ heating rate.

In a second DTA-TG experiment, heating was stopped at 170 °C, and the PXRD pattern of the thus dehydrated phase was not in accordance with that of the starting material. After storing the dehydrated sample for 7 days under ambient conditions, the PXRD pattern was still different compared to the pattern of the pristine material (Figure 6). This observation suggests that some H₂O was absorbed from the atmosphere, but the original structure had only partially recovered. This sample may be regarded as an intermediate containing fewer crystal water molecules and/or represents a mixture of a fully rehydrated material and a sample containing less H₂O.

Figure 6:

Comparison of PXRD patterns measured for the pristine sample, after the TG experiment stopped at 170 °C and after storing this sample on air. Dotted lines are guides for the eyes.

A Rietveld refinement of the PXRD pattern (for details see Experimental Section) of this sample assuming the presence of a mixture of the fully rehydrated phase and an intermediate with unknown water content (Figure 7) yields a slightly larger unit cell volume V _rehyd = 1318 Å³ compared to V _{single crystal} = 1295 Å³ (Tab. S1; Supplementary material). This difference can be explained by the fact that the single crystal data were measured at T = 200 K. For the intermediate compound V = 1241 Å³ is obtained which is 77 Å³ smaller than for the rehydrated sample (Figure 8). Because atomic positions of O atoms of the intermediate could not be located during Rietveld refinement, the volume occupied of H₂O in the pristine compounds was calculated to be about 15 Å³ [59]. Assuming that H₂O requires an identical space in the structure of the intermediate it contains ≈5–6 H₂O molecules per molecular formula. The lower water content is also accompanied by significant differences of the crystallographic axes and angles (Figure 8, insets), while the overall topology of the arrangement of the constituents is similar.

Figure 7:

Difference plot after the final two-phase Rietveld refinements of the diffraction pattern of the less water containing intermediate and the fully rehydrated phase of the title compound. Vertical bars are the Bragg reflections with the color code shown in the inset.

Figure 8:

Comparison of the structures of the fully rehydrated compound and of the intermediate. The insets show the unit cell parameters (in units of Å and deg) of the two compounds.

3.4 Water sorption properties

The water sorption properties of the title compound were investigated by performing an absorption and desorption experiment (Figure 9). At low pressures up to 0.2 p/p ₀ a steep increase of H₂O absorption is observed leading to an uptake of about 10 H₂O molecules. Further increase of the partial pressure results in a weak and nearly linear H₂O absorption of about 1 H₂O up to p/p ₀ ≈ 0.7. In the final step up to p/p ₀ ≈ 0.92 further H₂O molecules are absorbed yielding totally ≈14 H₂O, which is slightly more than the experimentally determined content. This observation may be explained by textural porosity, i.e. water is stored in cavities between crystallites. In the desorption branch a hysteresis is observed and an almost linear decrease of the H₂O content down to 11.8 H₂O at p/p ₀ ≈ 0.1 occurs, and further H₂O removal at lower partial pressures appears to be kinetically hindered.

Figure 9:

Water sorption isotherms of the title compound after activation at 170 °C.

The PXRD patterns of the title compound and of the samples fully dehydrated, after H₂O uptake and after H₂O desorption are displayed in Figure 10. The pattern of the fully dehydrated compound is significantly different to that of the pristine material and exhibits a very good crystallinity as well as long-range order as evidenced by sharp and intense reflections. After the H₂O uptake, the pattern is virtually identical with that of the starting material. This finding is an indication that the integration of excess H₂O observed at the highest H₂O partial pressure (Figure 9) is not incorporated in the crystal framework. Finally, the pattern of the sample after water desorption does not differ from that of the title compound. A Pawley fit evidences only minor differences of the unit cell parameters, i.e. the structure of the title compound is retained during the desorption process (Fig. S9; Supplementary material).

Figure 10:

The PXRD patterns collected at different stages of crystal water content.