A candidate for a single-chain magnet: [Mn3(OAc)6(py)2(H2O)2]n (OAc is acetate and py is pyridine) (2024)

Manganese carboxyl­ate cluster chemistry is now widely recognized as a fieldfrom which materials suitable for qubit-based computation could emerge. Thatchemistry was pioneered by George Christou and David Hendrickson, amongothers, with the fruitful Mn₁₂ line of compounds, including the emblematiccluster [Mn₁₂O₁₂(O₂CPh)₁₆(H₂O)₄] (Boyd et al., 1988;Sessoli et al., 1993). This mixed-valence cluster displays thetwoessential features required for a single-molecule magnet (SMM), i.e. ahigh-spin ground state and a large negative magnetic anisotropy. However, themost impressive behaviour reported for these clusters is the resonant quantumtunnelling of magnetization, which is a key property for molecular spintronics(Perenboom et al., 1998; Hill et al., 2003,2010). Like SMMclusters, some chain-shaped molecules also exhibit bis­tability and slowrelaxation of their magnetization. These compounds, termed single-chainmagnets (SCM), have their magnetic behaviour affected not only by the magneticanisotropy of the spins but also by their intra­chain magnetic inter­actions(Coronado et al., 2003; Brooker & Kitchen, 2009; Zhanget al.,2013).

Although many synthetic approaches have been described in the literature, therational design of a synthesis for a given manganese carboxyl­ate cluster ishindered by the fine tuning of the following parameters: the oxidation statesof the metal centres; the suitability of the symmetry of the ligand field forthe 3d orbitals centred on the metals; the coordination modes of thecarboxyl­ate ligands; and competition with ancillary ligands. Moreover, boththe nuclearity and dimensionality of the resulting structure are quiteunpredi­cta­ble, with a large range of possibilities, from isolated SMMs, whichmay be described as zero-dimensional nanoparticles, to three-dimensionalpolymeric architectures, which may be considered as bulk materials.

While working on the synthesis of such compounds, we obtained a one-dimensionalpolymer, namely [Mn₃(OAc)₆(py)₂(H₂O)₂]_n (OAc is acetate andpy is pyridine), (I), which is a candidate for a new SCM.

Although the synthesis of (I) was carried out under aerobic conditions, thecomplex is an Mn^II species with the formula[Mn₃(OAc)₆(H₂O)₂(py)₂]_∞, where OAc and py are acetate andpyridine ligands, respectively. One manganese ion, Mn1, is located on aninversion centre in the triclinic unit cell, and the other, Mn2, is located ina general position, close to an inversion centre. Three acetate ligands arebonded to Mn1, using three coordination modes (Fig. 1). The first acetate(atoms O1/O2) bridges independent Mn centres in the common 2.11syn–syn mode [for Harris coordination nomenclature, see Coxallet al. (2000)], found, for example, in the starting materialmanganesedi­acetate, Mn(OAc)₂.4H₂O (Bertaut et al., 1974; Nicolaïetal., 2001), and also in Co(OAc)₂.H₂O (Zhang et al.,2010) andother simple divalent metal salts. Acetate O5/O6 bridges the same metalcentres Mn1 and Mn2 in the 2.20 coordination mode, which is much lessfrequently observed. For Mn-based polymers, only two cases have been reportedto date including this coordination mode for OAc, viz.[Mn₃(OAc)₆(H₂O)₄].2H₂O (Cheng & Wang, 1991) and a complexpolymerwith a chain-like structure (Xu et al., 2009). Finally, thepolymericnature of (I) is fixed by the third acetate, O3/O4, in the 3.12 coordinationmode, between Mn1 and Mn2 in the asymmetric unit and a symmetry-related Mn2ⁱsite [symmetry code: (i) -x + 1, -y + 1, -z + 1]. Thisarrangement for polymerization is identical to that found in Mn(OAc)₂.4H₂O(Bertaut et al., 1974) and α-Mn(OAc)₂ (Lin et al.,2009), andin more complex polymeric compounds with various dimensionalities (e.g.Zartilas et al., 2008; Weng et al., 2008; Wanet al.,2010). These different µ₂- and µ₃-bridging modes are reflected inthe IRspectrum of (I): the ν_asym and ν_sym vibration modes for the COO^-groups are split into two or three bands around 1557 and 1414 cm^-1,respectively.

The arrangement of the acetate ligands completes the o­cta­hedral coordinationenvironment for the Mn1 centre. This metal centre presents an [MnO₆] ligandfield with the symmetry lowered from o­cta­hedral to C_i, because theMn1—O bond lengths for each acetate are significantly different: Mn1—O1 =2.1369(9), Mn1—O3 = 2.1997(8) and Mn1—O5 = 2.2277(8) Å. Thecoordinationenvironment for Mn2, which is in a general position, is completed with neutralligands, H₂O and pyridine, present in the reaction media. Mn2 thus has an[MnO₅N] coordination, with a ligand field probably slightly stronger thanthat for Mn1, and a large deviation from o­cta­hedral symmetry: the coordinationbond lengths for Mn2 are in the range 2.0995(9)–2.2946(10) Å and thetrans angles are in the range 164.84(4)–176.87(4)°.

The one-dimensional polymeric chain of (I), formed via inversioncentres, runs in the [010] direction (Fig. 2). Within the asymmetric unit, theMn1···Mn2 separation is 3.36180(18) Å and the Mn2···Mn2ⁱ distance,allowing polymerization, is longer, at 4.4804(3) Å. Both distances areunexceptional, considering that, in Mn–acetate-based polymers, they span alarge range, ca 2.5–5.2 Å, allowing a variety of magnetic groundstates for these materials. In the present case, long metal–metal separationsalternate along the chain with dimers of short separation. Parallel chains arepacked in the crystal structure, and the cohesion is maintained throughO—H···O hydrogen bonds of moderate strength, using the coordinated watermolecules as donors (Fig. 2 and Table 2). The graph set (Bernstein etal., 1995) resulting from two inversion-related inter­chain hydrogenbondsis R₂²(8), common in water–acetate systems. The metal–metaldistance in the R(8) [Should this beR₂²(8)?] ring is Mn2···Mn2ⁱⁱ = 5.3752(3) Å[symmetry code: (ii) -x + 2, -y + 1, -z + 1]. Since thisdistance is much longer than the metal–metal inter­actions along the chain,(I) should be considered as a true low-dimensional one-dimensional polymer,rather than a two-dimensional material. In spite of its low dimensionality,this material is a densely packed system, reaching a high Kitaigorodskipacking coefficient of C_K = 0.733 (PLATON; Spek, 2009).

The above-described geometric features are encouraging and make (I) acandidatefor being an Ising one-dimensional system behaving as a ferri- or ferromagnet.Assuming a weak enough crystal-field splitting, high-spin d⁵electronic configurations may be expected for the Mn^II centres. Theanisotropic trinuclear units (Mn1···Mn2···Mn2ⁱ) have distances between themagnetic sites that are suitable for ferromagnetic inter­actions. Indeed, thesedistances, of 3.36 and 4.48 Å, may be compared with those observed in thefirst heterometallic Mn^III–Ni^II polymeric SCM synthesized in 2002:Mn^III···Mn^III = 3.42 Å and Mn^III···Ni^II = 5.06 Å (Clérac etal., 2002). Thus, the actual nature of the magnetic behaviour forthetitle Mn^II polymer should be defined mainly by inter­chain contacts resultingfrom the R(8) [Should this be R₂²(8)?]ring motifs. A search of the Cambridge Structural Database (Version 5.35,updated May 2014; Allen, 2002) retrieved 139 similarR₂²(8) ringsin Mn compounds, with coordinated water molecules as donors and carboxyl­ate Oatoms as acceptors. Most of them are associated with first-levelcentrosymmetric patterns R(_a^a), as in (I), and the othersbelong to second-level patterns R(_a^b), in Motherwell'sgraph-set nomenclature (Motherwell et al., 2000). The importantfeatureto be considered, bearing the magnetic properties in mind, is the poorflexibility of this ring: for R(_a^a) patterns, the ring has achair conformation and the metal–metal inter­action is determined mainly bythe puckering parameters, while R(_a^b) patterns may adopt afolded conformation, allowing shorter metal–metal inter­actions. In the subsetof 139 hits, regardless of the Mn oxidation state and the dimensionality ofthe crystal structure, the Mn···Mn separation in R₂²(8) rings is inthe range 4.77–6.32 Å. Inter­estingly, the shortest separation was reportedfor a complex Mn₂₂ cluster, in which the R(8) [Should this beR₂²(8)?] ring links two molecules. This mixed-valencecompound is an SMM with quantum tunnelling of magnetization (Brockman etal., 2007). Thus, it may be inferred that the R(8)[Should thisbe R₂²(8)?] ring in (I) is not an efficientexchange pathway, making a strong anti­ferromagnetic inter­chain couplingunlikely.

Preliminary magnetic measurements of (I) are in agreement with this structuraldescription. At room temperature, the title complex shows a χ_mTvalue of 13.15 cm³ mol^-1 K, which is very close to that calculated forthree non-inter­acting S_i = 5/2 centres, 13.12 cm³ mol^-1 K,assuming a spin-only model (g_Mn = 2.00; Christian et al.,2004). A plot of M versus T shows different magnetizationpathwaysbelow a critical temperature T_C = 50 K, pointing to superparamagneticbehaviour [Might it help the reader to have this plot available in thesupporting information?]. On the other hand, a hysteresis loop in theM(H) plot is observed (Fig. 3), indicative of ferromagneticexchange inter­actions (Bertotti, 1998). Experimentally, an SCM showsbothsuperparamagnet-like properties with frequency-dependent out-of-phase signalsin AC susceptibility measurements, and hysteresis in M versus appliedDC field measurements (King et al., 2004; Brockman etal.,2007). Therefore, AC magnetic susceptibility measurements are currentlybeingcarried out, in order to characterize further the potential SCM behaviour ofthis new polymer.

All synthetic and post-synthetic work was carried out under aerobic conditions.The reagents and solvents were obtained from commercial sources and usedwithout further purification. IR analyses were performed using a Nicolet Nexus6700 FT–IR spectrometer in the 4000–600 cm^-1 range. Compound (I) wasprepared by the addition of excess pyridine (12.41 mmol, 1 ml) to a hotsolution of Mn(OAc)₂.4H₂O (1 mmol, 0.245 g) in EtOH (95%; 5 ml). Thesolution was stirred for 30 min and filtered, [and then layered withEt₂O?]. After 1 d, colourless crystals of (I) were obtained from thissolution by slow diffusion of Et₂O (yield 48%). Spectroscopic analysis: IR([Medium?], ν_max, cm^-1): 3252 (br), 2943 (s), 2878(s), 1557 (m), 1414 (m), 1004 (st), 877(m), 658 (m), 606 (m). Magnetic susceptibilitymeasurements on powdered crystals were carried out using a Quantum DesignSQUID magnetometer at 2 K under an applied field H of -50000 to 50000Oe, and in the temperature range 2–300 K at H = 1000 Oe. Correctionfor the diamagnetic contribution of the constituent atoms was applied usingPascal's constants.

Crystal data, data collection and structure refinement details are summarizedin Table 1. The collection of diffraction data (work done in Canada) and thefinal structure refinement (work done in Mexico) were routine works. AllC-bound H atoms were placed in calculated positions, with C—H = 0.95(pyridine) or 0.98 Å (methyl acetate). Methyl groups were considered asrigid tetra­hedral groups free to rotate about their C—C bonds. Isotropicdisplacement parameters for these H atoms were calculated as U_iso(H)= 1.2U_eq(C) for the pyridine molecule and U_iso(H) =1.5U_eq(C) for the methyl groups. Atoms H71 and H72 of the watermolecule were clearly detected in a difference map and were refined with freecoordinates and isotropic displacement parameters.