Introduction and review

One of the most complex categories of coordination compounds are polynuclear coordination clusters (CCs) that incorporate multiple metal ions into a single molecular entity and are linked by bridging ligands [1]. These entities are of great interest for their aesthetically beautiful structures [2–4], unexpected transformations [5–7] and potential applications in magnetism [8–10], luminescence [11–15], catalysis [16–18], etc. Paramagnetic transition metal CCs are of intense interest and have attracted a vast amount of attention since the discovery that some CCs behave as single-molecule magnets (SMMs) [19–21]. The NiII (d8) ion has second-order orbital angular momentum, and zero-field splitting (ZFS) which can result in significant single-ion anisotropy and potentially in molecules exhibiting interesting magnetic properties [22, 23]. The interest in polynuclear Ni(II) coordination chemistry was first captured when the first Ni(II)-based SMM, a Ni12 complex, was reported in 2001 by Cadiou et al. [24]. Ever since, there have been a number of homometallic polynuclear NiII CCs with high nuclearities including, Ni5 [7, 25], Ni6 [26], Ni7 [27, 28], Ni8 [29–32], Ni9 [33, 34], Ni11 [35], Ni12 [7], Ni13 [36], Ni14 [37], Ni20 [38], Ni21 [39], Ni24 [40] and Ni26 [41], and many of these display interesting magnetic properties including ferromagnetic, ferrimagnetic coupling, diamagnetism and SMM behaviour.

Clusters of this size incorporate simple modified ligands with a wide variety of coordination modes for bridging, such as diethanolamine [42], Schiff base [43], carbide [44, 45] and carboxylate [46]. The introduction of bridging groups can increase the nuclearity of a CC. Carbonate anions offer a diverse range of bridging modes within cluster type molecules. A number of high-nuclearity CCs have been based on carbonate moieties [47, 48]. While Ni(II) CCs with bridging carbonate ligands are known, structural factors and magnetic exchange within these clusters greatly vary due to the large number of coordination modes of the CO3 2− anion [39, 49, 50]. Some interesting examples include, a Ni6 containing a carbonato bridge [51] and an Ni12 where four Ni4O4 units are templated around a central CO3 2− anion core [52]. On the other hand, the first reported mixed NiII/NaII CC was reported in 1976 by Jonas for potential small-molecule activation [53]. Since then, a number of NiII/NaII CCs have been reported, targeting for high-nuclearity clusters and interesting magnetic properties: Ni4Na2 [54], Ni4Na5 [55], Ni4Na3 [56], Ni4Na4 [57, 58] and others [59–62]. However, since 2007, NiII/NaI CCs with nuclearity over 10 have been reported far more frequently. The highest reported of these is a Ni18Na6 cluster [63] and the second Ni16Na4 [64], both synthesised by calix [4] arene-type ligands and the third is a Ni16Na2 cluster. In addition, two Ni12Na2 clusters were reported by Christou et al. [37, 65]. In all cases, similar anti-ferromagnetic behaviour was observed.

The diprotic Schiff base ligand (E)-2-(2-hydroxy-3-methoxybenzylideneamino)-phenol (H2L1, Scheme 1) initially reported in 1971 to capturing UO2 [66] can be synthesised in almost quantitative yields [67] and has two pockets that can coordinate to metal centres. Previously, this ligand has been involved in the synthesis of homometallic [68–71] and heterometallic CCs [72–75]. We recently employed this ligand in 3d/4f chemistry to synthesise a family of homogeneous efficient catalysts towards a domino reaction [76]. Interestingly, when H2 L1 was employed in Ni(II) chemistry, a tetranuclear Ni4 CC exhibiting ferromagnetic interactions at low temperatures was isolated. [71] With the interest of introducing carbonate anions into a system, there are three key methods: direct addition of carbonate or bicarbonate [14], atmospheric fixation of carbon dioxide [77] and in situ decomposition of ligands [78].

Scheme 1
scheme 1

Protonated form of the organic ligand H2 L1 used in this study

Full size image

Having all these in mind, in this article, we study the influence of the presence of Na cations and CO3 2− anions on the nuclearity of the given chemical system Ni(II)/H2L1 and we report two compounds formulated [Ni II3 Na(L1)3(HL1)(MeOH)2] (1) and [Ni II6 NaI(L1)5(CO3)(MeO)(MeOH)3(H2O)3]·4(MeOH) 2(H2O) [2 4(MeOH) 2(H2O)]. Topological aspects and magnetic properties of these compounds are further discussed.

Experimental

Materials

Chemicals (reagent grade) were purchased from Sigma-Aldrich and Alfa Aesar. All experiments were performed under aerobic conditions using materials and solvents as received.

Instrumentation

IR spectra of the samples were recorded over the range of 4000–650 cm−1 on a Perkin-Elmer Spectrum One FT-IR spectrometer fitted with a UATR polarization accessory.

X-ray diffraction

Crystallography

Data for 1 and 2·4(MeOH) were collected at the National Crystallography Service, University of Southampton [79] using a Rigaku Saturn 724+ area detector mounted at the window of an FR-E+ rotating anode generator with a Mo anode (λ = 0.71075 Å) under a flow of nitrogen gas at 150(2) K for 1 and 100(2) K for 2·4(MeOH). Both structures were determined using Olex2 [80], solved using either Superflip [81] or SHELXT [82, 83] and refined with SHELXL-2014 [84]. All non-H atoms were refined with anisotropic thermal parameters, and H atoms were introduced at calculated positions and allowed to ride on their carrier atoms. For 1, attempts to model the lattice solvents (2MeOH) were unsuccessful despite multiple data collections. Therefore, the solvent mask function of Olex2 [80] was employed to remove the contribution of the electron density associated with those molecules from the intensity data. Geometric/crystallographic calculations for all structures were performed using PLATON [85], Olex2 [80] and WINGX [82] packages; graphics were prepared with Crystal Maker [86] CCDC 1476870-1476871 (Tables 1, 2, 3).

Table 1 Crystallographic data for compounds 1 and 2
Full size table
Table 2 Selected bond distances and angles for compound 1
Full size table
Table 3 Selected bond distances and angles for compound 2
Full size table

Magnetic studies

Magnetic susceptibility measurements were carried out on polycrystalline samples with a MPMS5 Quantum Design susceptometer working in the range 30–300 K under external magnetic field of 0.3 T and under a field of 0.03 T in the 30–2 K range to avoid saturation effects. Diamagnetic corrections were estimated from Pascal tables. The quality of the fitting is parameterised as the factor R = (χ M T exp−χ M T calc)2/(χ M T exp)2.

Synthesis of H2L1

o-Vanillin (0.025 mol, 3.35 g) and 2-amino-phenol (0.025 mol, 2.73 g) were dissolved in MeOH (5 mL). The suspension was refluxed for 1 h, during which time a bright orange solid precipitated. After cooling to room temperature, the solid was filtered off and washed with cold MeOH and Et2O. The solid was dried in vacuo. Yield 99 %. 1H NMR (500 MHz, DMSO-d6) δ 9.75–9.71 (m, 1H), 8.95 (s, 1H), 7.36 (dd, J = 8.0, 1.6 Hz, 1H), 7.21–6.99 (m, 3H), 6.96 (dd, J = 8.1, 1.4 Hz, 1H), 6.91–6.81 (m, 2H), 3.80 (s, 3H); 13C NMR (126 MHz, DMSO-d6) δ 162.01, 152.27, 151.42, 148.63, 134.95, 128.46, 124.26, 120.05, 119.96, 119.70, 118.37, 116.97, 115.71, 56.34, 40.62, 40.53, 40.45, 40.36, 40.28, 40.19, 40.11, 40.03, 39.95, 39.86, 39.69, 39.53.

Synthesis of 1

Ni(ClO4)2.6H2O (0.1 mmol, 37 mg), H2 L1 (0.1 mmol, 24 mg) and Na2CO3 (0.1 mmol, 10 mg) were suspended in MeOH (20 mL) and stirred for 1 h. The solution was filtered and the filtrate left for slow evaporation. After 14 days, small brown crystals of 1 were collected and washed with Et2O. [Yield = 45 % calculated based on Ni(II)]. IR (ν, cm−1) = 3402, 3291, 1780, 1609, 1552, 1454, 1388, 1291, 1224, 1182, 1073, 1033, 964, 820, 733, 635. CHN [Ni3Na(C14H11NO3)3(C14H12NO3)(CH3OH)2]; observed C-56.63 %; H-4.47 %, N-4.43 % (expected); C-56.76 %; H-4.35 %; N-4.56 %.

Synthesis of [2 4(MeOH) 2(H2O)]

A similar synthetic procedure to 1 was followed; however, after filtration, the solution was placed in a vial which in turn was placed in a larger vial that contained a saturated aqueous solution of Na2CO3. A few drops of concentrated HCl were added to the saturated solution, and the vial was immediately sealed. After 5 days, large brown block-like crystals of compound [2 4(MeOH)] were collected. (Yield = 60 % calculated based on Ni(II)). IR (ν, cm−1) = 3288, 1604, 1552, 1451, 1388, 1294, 1229, 1182, 1075, 1037, 966, 818, 736, 637. CHN [Ni6Na(C14H11NO3)5(CO3)(CH3O)(CH3OH)3(H2O)3] (observed); C-49.51 %; H-4.11 %; N-3.79 %; (expected) C-49.52 %; H-4.21 %; N-3.85 %.

Results and discussion

Two NiII/NaII CCs were synthesised. The aim to use Na2CO3 as the base to introduce carbonate bridges to the system results in the presence of one NaI in both CCs, while both compounds have different structures to the previously reported homometallic NiII CC with H2L1 [69]. In 1, the use of Na2CO3 did not result in the introduction of carbonate bridging groups; however, in 2, where the reaction solution was exposed to a high CO2 atmosphere during slow evaporation, a single CO3 2− anion is observed. 1 and 2 are new additions to a large family of previously reported NiII/NaI CCs.

1 crystallises in the monoclinic space group P21/n and contains one molecule in the asymmetric unit cell. Compound 1 contains three NiII ions, one NaI ion, three doubly deprotonated (L1) ligands, one mono deprotonated (HL1 −) ligand and three coordinated CH3OH solvent. Each NiII ion coordinates to six heteroatoms and displays a distorted octahedral geometry. There are three observed coordination modes for the doubly deprotonated ligands (L1) (Scheme 2); in each, a NiII ion occupies the ONO pocket between the phenoxido and imido atoms. In the singly protonated (HL1) ligand coordination mode (Scheme 2) [69, 71–74, 87–91], a NiII ion is observed occupying the OO pocket between the methoxido and phenoxide ligand atoms. Interestingly, L1 demonstrates a twisted out-of-plane geometry between the two aromatic rings with a torsion angle of −100.679° along the C = C-N = C bond, whilst the L1 Ligand aromatic rings are close to being in-plane with torsion angles between −166.840° and 174.23°. In addition to this, both NiII and the NaI are observed bridging from the phenoxido atoms. The NaI ion displays a distorted octahedral geometry with average Na–O bond distances ranging from 2.272 to 2.648A. Then, three NiII ions all display distorted octahedral geometry with average Ni–O bind distances ranging from 1.981 to 2.343A. Four μ3-Ophenoxo donors are placed in the corners of the cube linking the four cations. Ni–O–Ni bridges form three faces of the cube with bond angles of 97.0(1)°/97.7(1)°, 97.18(9)°/99.87(9)° and 93.38(9)°/102.7(1)°.

Scheme 2
scheme 2

Coordination modes of the organic ligand H2 L1 seen in 1 (upper) and 2 (middle and lower)

Full size image

According to a nomenclature, developed by some of us [92], the decorated core of this compound is assigned 3M4-1 (Fig. 1). An extended literature survey reveals only three previously reported complexes with Ni3Na core topologies, each of which are synthesised from similar Schiff base ligands and coordination pockets to H2 L1 [93, 94]. The most recently reported of these Zhang et al. [93] shares the same 3M4-1 topology as 1, making 1 the second example of a Ni II3 /NaI cubane.

Fig. 1
figure 1

(Upper) The crystal structure of compound 1, (lower left) the main Ni3Na core, (lower right) the 3M4-1 topology. Colour code Ni (green), Na (light yellow), O (red), N (blue), C (yellow) (Color figure online)

Full size image

Compound 2 crystallises in the triclinic P-1 space group and contains two molecules in the asymmetric unit. The molecule of 2 contains six NiII ions, one NaI ion, five doubly deprotonated L1 ligands, one bridging (µ5-CO3), one deprotonated MeOH solvent molecule with three CH3OH and three H2O molecules which fulfil the coordination geometry of the metal ions.

All six NiII ions have a distorted octahedral geometry. The L1 demonstrates four different bridging modes (Scheme 2) with NiII ions observed, occupying the (ONO) and (OO) pocket as well as bridging from the phenoxido, methoxido and imine groups. The NaII ion is seven-coordinated, and the coordination geometry can best be described as a capped trigonal prism determined by SHAPE software. The NaI ion coordinates to the (OO) pocket between the methoxide and phenoxido atoms of three ligands with the seventh position fulfilled by a H2O molecule. The CO3 2− entity bridges Ni2, Ni4, Ni5 and Ni6 through a (μ5-CO3) ion. Na–O bond distances range from 2.282 to 2.625A. The Ni–O and Ni–N distances vary between 1.983(6) and 2.321(4) Å, while the Ni–O–Ni angles mediated by the Ophenoxo donors vary from 88.88(17) and 101.8(2)°, those mediated by the μ3-OMe bridge, that links Ni1, Ni2 and Ni3, exhibit bond angles ranging between 95.74(19) and 100.2(2)° and Ni5-O22-Ni6, mediated by one Ocarbonate atom reach 130.4(2)°.

Adopting our topological approach, the core of compound 2 can be enumerated as 2,3,4M7-1 [92]. The 2,3,4M7-1 core topology can be viewed as two lozenges fused with a shared apex vertices (Fig. 2 lower); this can be otherwise described as a “butterfly” motif, where the shared vertices are the body and the protruding lozenges the “wings”. The first reported compound of this topology was a mixed valence MnII/MnIII complex in 1991 [95]. This was subsequently followed by a number of homometallic CCs, including MnIII [96] and PbII [97]. The first heterometallic complex of this topology was reported in 2012 with the core topology of Zn5Na2 [98] with the ((2,3-dihydroxypropylamino)methyl)phenol ligand (Scheme 3 H3L5). The metal ion positions of this core topology differ to the reported compound 2. The NaI ions occupy the two apex vertices of the two fused lozenges with all other positions occupied by the 3d ZnII ions. In both examples, the fused vertice is occupied by a 3d ion (NiII/ZnII). Metal ion geometries widely differ in both complexes, with NaI ions possessing severely distorted octahedral geometry and ZnII ions tetrahedral. Both of these complexes are formed from Schiff base ligands incorporating the 2-hydroxy-benzaldehyde moiety, which provides a similar coordination environment, with further bridging attributed to the third hydroxyl group in H3L5. To the best of our knowledge, 2 is the first example of an NiII/NaI CC exhibiting the 2,3,4M7-1 topology. Moreover, ignoring the existence of the Na ion, the motif of the hexanuclear Ni6 unit can be enumerated as 1,2,2,3,3M6-1 (Figure S3), which has been reported only once before in Zinc chemistry [99].

Fig. 2
figure 2

(Upper) The crystal structure of compound 2. Lattice molecules and H atoms are omitted for clarity. Colour code Ni (green), Na (light yellow), O (red), N (blue), C (yellow). (lower) The 2,3,4M7-1 topology of compound 2 (Color figure online)

Full size image
Scheme 3
scheme 3

Protonated forms of the organic ligands used for the synthesis of 3d and 3d/Na CCs possessing the 2,3,4M7-1 topology

Full size image

Magnetic studies

The χ MT value at room temperature for compound 1 is 3.97 cm3 Kmol−1, increase in cooling to a maximum value of 4.50 cm3 Kmol−1 at 13 K, prior to decrease at low temperature, indicating a dominant weak ferromagnetic interaction. Fit of the experimental data was performed with PHI [100] program and applying the Hamiltonian:

$$H = - 2J_{1} \left( {S_{1} \cdot S_{2} + S_{1} \cdot S_{3} } \right) \, {-}2J_{2} \left( {S_{2} \cdot S_{3} } \right)$$

including a D ion parameter. Considering the D ion parameter, an excellent fit of the experimental data was obtained for the parameters J 1 = 3.6 cm−1, J 2 = −2.3 cm−1, D = 1.2 cm−1, g = 2.26 and R = 4.2·10−5 but this was not the case when the isotropic Hamiltonian were applied. The shape of the χ MT plot corresponds to a weak ferromagnetic coupling with a strong decrease at low temperature due to combination of the AF component and the ZFS of the same order of magnitude than J.

The χ MT value at room temperature for 2 is 7.1 cm3 Kmol−1 at room temperature and on cooling increases slight up to a rounded maximum of 7.34 cm3 Kmol−1 at 45 K. Below this temperature, the χ MT value decreases down to 5.94 cm3 Kmol−1 at 2 K. From the structural data, compound 2 consists of a Ni4 butterfly subunit linked by means of different kinds of bridges to two additional NiII cations, and thus, the corresponding Hamiltonian to analyse the experimental data must be:

$$H = - J_{1} \left( {S_{1} \cdot S_{2} + S_{1} \cdot S_{3} + S_{2} \cdot S_{4} + S_{3} \cdot S_{4} } \right) \, {-}J_{2} \left( {S_{2} \cdot S_{3} } \right) \, {-}J_{3} \left( {S_{4} \cdot S_{5} } \right) \, {-}J_{4} \left( {S_{2} \cdot S_{6} + S_{4} \cdot S_{6} } \right) \, {-}J_{5} \left( {S_{5} \cdot S_{6} } \right)$$

This expression is clearly overparameterised, and different solutions can be obtained assuming all constants or simplifying the Hamiltonian (in all cases R factor in the 10−5–10−6 range). This feature is a consequence of the topology of that complex: The butterfly subunit can give any local spin between S = 0 and S = 4 as function of the J 1 and J 2 values, and in addition, the two additional NiII cations form triangular subunits in which there are competitive interactions, and thus, only a qualitative analysis can be made.

It is well established that the FM/AF border for Ni–O–Ni bridges is placed around 96° for hydroxo bridges and values slightly larger for alkoxo or phenoxo ones. The most of the Ni–O–Ni bond angles in compound 2 lies around this border and thus, only weak ferro or anti-ferromagnetic interactions can be expected. In contrast, the very large Ni5-O22-Ni6 bond angle mediated by one of the O-donors of the carbonate ligand must give a clear AF interaction. In good agreement, all obtained fits gave a large dispersion of values for J 1–J 4, but in all cases, J 5 gave a value around −7 cm−1 (Scheme 4). In light of this data, we must assume that two S = 1 local spins are cancelled, whereas the remainder four spins are related by weak interactions.

Scheme 4
scheme 4

Schematic representation of the exchange couplings seen in compound 2

Full size image

The presence of competitive interactions is reinforced by magnetisation experiments. The unsaturated value under a field of 5 T is 6.25 Nμβ and can be reasonably fitted as an isolated S = 3 spin with D = 0.89 cm−1 and g = 2.14 (Fig. 3).

Fig. 3
figure 3

(left) χ MT versus T plots for the complexes 2 (circles) and 1 (diamonds). Solid lines show the fit of the experimental data for 1 and one of the fits obtained for 2. (right) The magnetisation plot for compound 2

Full size image

Conclusions

Our synthetic strategy to introduce NaI and carbonate, as co-ligand, along with H2L1 in Ni chemistry, resulted in two new Ni/Na compounds possessing a rare (for 1) and an unseen (for 2) topology. Magnetic studies reveal the presence of weak ferromagnetic interactions at low temperature for both compounds. Assumptions for the structural relationship of 1 and 2 can be attempted, but more synthetic studies are required to support it. Our future studies will be focused in two different directions: a) to extend the systematic synthetic study by varying metal ligand ratios and co-ligands aiming to achieve higher nuclearity CCs and b) to further develop our topological approach [101] by incorporating all centres as nodes.