μ3-Methoxido-κ3 O:O:O-tris(μ-l-p-tyrosinato-κ3 N,O:O)tris(l-p-tyrosinato-κ2 N,O)trinickel(II,III) methanol tetrasolvate

A trinuclear nickel complex, [Ni3(C9H10NO3)6(CH3O)]·4CH4O, was synthesized and characterized as a neutral cluster containing the incomplete cubane {Ni3(μ1-O)(μ2-O)2(μ3-O)} core of 2M3–1 topology. The three nickel cations show similar octahedral coordination, {Ni(μ1-O)(μ2-O)2(μ3-O)(μ1-N)2}; the positive charge is balanced by six tyrosinate ligands and one methoxide ion. The mean oxidation state of each NiII ion is therefore +2.33. The common coordination modes, chelating (via the amino N and the carboxylate O atoms) and bridging (via the carboxylate O atom), are exhibited by the tyrosinates. Three interligand (intracluster) N—H⋯O hydrogen-bonding interactions stabilize the incomplete cubane-type moiety. Additional N—H⋯O, O—H⋯O and C—H⋯O interactions are formed between clusters, and between the clusters and methanol molecules to regulate the spatial orientation of the tyrosinate and the assembly of the clusters in the crystal. The approximate equilateral triangular arrangement of the three nickel cations in the incomplete cubane-type moiety suggests the possible magnetic frustration, and the proximity of these metal cations indicates weak metallic bonds. The structure contains approximately 39% solvent-accessible volume between the clusters. This is filled with 17 molecules of disordered methanol and was modelled with SQUEEZE [Spek (2009 ▶). Acta Cryst. D65, 148–155]; the reported unit-cell characteristics do not take these molecules into account. The H atoms of the solvent molecules have not been included in the crystal data.

A trinuclear nickel complex, [Ni 3 (C 9 H 10 NO 3 ) 6 (CH 3 O)]Á-4CH 4 O, was synthesized and characterized as a neutral cluster containing the incomplete cubane {Ni 3 ( 1 -O)( 2 -O) 2 ( 3 -O)} core of 2M3-1 topology. The three nickel cations show similar octahedral coordination, {Ni( 1 -O)( 2 -O) 2 ( 3 -O)( 1 -N) 2 }; the positive charge is balanced by six tyrosinate ligands and one methoxide ion. The mean oxidation state of each Ni II ion is therefore +2.33. The common coordination modes, chelating (via the amino N and the carboxylate O atoms) and bridging (via the carboxylate O atom), are exhibited by the tyrosinates. Three interligand (intracluster) N-HÁ Á ÁO hydrogen-bonding interactions stabilize the incomplete cubane-type moiety. Additional N-HÁ Á ÁO, O-HÁ Á ÁO and C-HÁ Á ÁO interactions are formed between clusters, and between the clusters and methanol molecules to regulate the spatial orientation of the tyrosinate and the assembly of the clusters in the crystal. The approximate equilateral triangular arrangement of the three nickel cations in the incomplete cubane-type moiety suggests the possible magnetic frustration, and the proximity of these metal cations indicates weak metallic bonds. The structure contains approximately 39% solvent-accessible volume between the clusters. This is filled with 17 molecules of disordered methanol and was modelled with SQUEEZE [Spek (2009). Acta Cryst. D65, [148][149][150][151][152][153][154][155]; the reported unit-cell characteristics do not take these molecules into account. The H atoms of the solvent molecules have not been included in the crystal data.
The asymmetric unit of 1 contains a neutral cluster of ninety-one non-hydrogen atoms comprising three Ni ions, six tyrosinate ligands, a methoxide ion and four methanol molecules (Fig. 1) The six tyrosinate ligands exhibit the common chelating mode of coordination, using the amino N atoms and either the carboxylate µ 2 -η 2 :η 0 O atoms (O4, O10, O16) or the carboxylate µ 1 -η 1 :η 0 O atoms (O1, O7, O13). These generate two five-membered chelate rings about each Ni center. These coordination modes are commonly found in the tyrosinate ligands (Pei & Wang, 2006;Wojciechowska et al., 2011Wojciechowska et al., , 2012. Curiously, none of the phenolic groups of the tyrosine ligands within 1 are coordinated to the metal, despite conditions sufficiently basic to produce methoxide. The presence of the coordinated methoxide should invalidate any assumption on the presence of any extra-framework species with positive charges. The positive charge of Ni ions is therefore balanced by six tyrosinate ligands and one methoxide ion, resulting in the mean oxidation state of each nickel to be +2.33 (possibly a combination of two Ni II and one Ni III ). The solid conclusion may be derived by a magnetic study of 1.
According to previous literature, three inter-ligand (intra-cluster) hydrogen bonding interactions of N-H···O type were reported to be important in stabilizing the incomplete cubane structure (Ama et al., 2000). This seems to be partially true The clusters are arranged by the 2 1 screw axis into layers in the xz plane. These layers are stacked in an ABAB arrangement parallel to b. There exist hydrogen bonds between the clusters, both within the layers and between them.
This packing arrangement of clusters is rather inefficient and the structure contains large voids centred on the origin such that approximately 39% of the structure is solvent accessible volume. Methanol molecules within these regions were poorly located and the reflection data were treated with the SQUEEZE algorithm (Spek, 2009) to model electron density within these regions. These calculations reveal that each void contains around 279 electrons consistent with around 17 molecules of methanol, giving an overall composition for 1 of [Ni 3 (C 9 H 11 NO 3 ) 6 (OCH 3 )].21CH 3 OH. The methanol is lost very quickly when crystals are removed from solvent and this has prevented extensive analysis of the properties of 1.
The presence of methoxide suggests that it should be possible to obtain similar structures with other weakly coordinating anions. Similarly, replacement of method by other, bulkier and less volatile solvents, may enable further studies on similar compounds, in particular magnetic measurements. Clusters of this type therefore may be suitable for fundamental magnetic studies by variation of ligand bulk, or may prove suitable nodes in the construction of framework solids by appropriate ligand choice.

Experimental
Ni(NO 3 ) 2 .6H 2 O (0.0148 g, 0.5 mmol; 98% Alfa Aesar) and 4-(HO)C 6 H 4 CH 2 CH(NH 2 )CO 2 H (L-tyrosine; 0.0185 g, 0.10 mmol; 98% Sigma-Aldrich) were dissolved in methanol (5.0 cm 3 ; 99.8% Fisher Scientific) using a glass vial (vial A). A few drops of HCl (37% Fisher Scientific) were necessary to completely dissolve the L-Tyr. To a smaller glass vial, ca. 0.2 cm 3 of (C 2 H 5 ) 3 N (triethylamine; 99% Fisher Scientific) was added and the vial closed using lid with a small pin hole (vial B). Vial B was then inserted in vial A, which was then closed tightly. After ca 4 months, a few blue blocks crystallized from the solution, and were isolated for X-ray diffraction data collection. order to take the contribution of the disordered methanol into account, the ′SQUEEZE option′ in the program PLATON (Spek, 2009) was implemented. This resulted in an improvement of the R and wR from 0.073 and 0.212 to 0.040 and 0.097, respectively.

Figure 1
The asymmetric unit of 1 showing atom-labeling scheme and with 50% probability displacement ellipsoids. Hydrogen atoms are omitted for clarity.    Refinement. Refinement of F 2 against ALL reflections. The weighted R-factor wR and goodness of fit S are based on F 2 , conventional R-factors R are based on F, with F set to zero for negative F 2 . The threshold expression of F 2 > σ(F 2 ) is used only for calculating R-factors(gt) etc. and is not relevant to the choice of reflections for refinement. R-factors based on F 2 are statistically about twice as large as those based on F, and R-factors based on ALL data will be even larger.  (11)