Slow magnetic relaxation in a {Co II Co III2 } complex containing a high magnetic anisotropy trigonal bipyramidal Co II centre †

We report a trinuclear mixed-valence {Co II Co III2 } complex, where the Co II centre adopts a trigonal bipyramidal geometry, leading to a large, easy-plane magnetic anisotropy and ﬁ eld-induced slow magnetic relaxation with a Raman-like relaxation process. Single-Molecule Magnets (SMMs) show promise in a number of technological applications such as molecular spintronics, 1 high-density information storage, 2 and qubits for quantum When the magnetic properties arise from a paramagnetic ion in an appropriate ligand field, referred to as Single-Ion Magnets (SIMs). The of SMM and is the an Δ E k B that the reorientation of magnetisation the M S = ± S components of the ground state S . Magnetic anisotropy, which on the spin – orbit di ﬃ cult to in

Single-Molecule Magnets (SMMs) show promise in a number of technological applications such as molecular spintronics, 1 high-density information storage, 2 and qubits for quantum computation. 3 When the magnetic properties arise from a single paramagnetic ion in an appropriate ligand field, then these molecules are often referred to as Single-Ion Magnets (SIMs). The origin of SMM and SIM behaviour is the presence of an energy barrier (ΔE/k B ) that prevents the reorientation of magnetisation between the M S = ±S components of the ground state S. Magnetic anisotropy, which depends on the spin-orbit contribution, 4 is difficult to control in polymetallic systems, and hence recent research has been focused on monometallic systems. Monometallic Co II complexes have been found to exhibit both large negative (easy-axis) and large positive (easyplane) magnetic anisotropy when found in tetracoordinate, pentacoordinate and hexacoordinate geometries. 5 More specifically, Co II ions with an ideal trigonal bipyramidal geometry are expected to exhibit a high easy-plane anisotropy. 6 We are working towards the synthesis of coordination complexes containing at least one high magnetic anisotropy centre, such as trigonal bipyramidal (TBP) Co II , in order to boost the barrier ΔE/k B . Herein, we report the complex [Co II Co III 2 (µ 3 -OH)(µ-pz) 4 (DBM) 3 ]·2MeCN (1·2MeCN) (Fig. 1), which is a new solvate of a previously reported {Co 3 } complex, obtained by a different synthetic procedure. 7 Complex 1 is a mixed-valence isosceles triangle of Co II /Co III ions synthesised from the reaction of CoCl 2 ·6H 2 O with Hpz ( pyrazole) and HDBM (dibenzoylmethane) in MeOH/MeCN in the presence of NEt 3 (see the ESI †). The crystallographic data can be found in Table S1 (see the ESI †). The Co II centre (Co1) is five-coordinate adopting a slightly distorted TBP geometry, while the two diamagnetic Co III ions (Co2 and Co3) adopt an octahedral geometry. The oxidation states of Co II and Co III were confirmed using Bond Valence Sum (BVS) analysis. 8 Continuous shape measures (CShMs), 9 which provide an estimate of the distortion from the ideal TBP geometry for Co1, give a value of 0.33 (where 0 corresponds to the ideal polyhedron), confirming a small distortion (Table S2 and Fig. S3 ESI †). The crystal packing is shown in Fig. S4. † Intermolecular interactions are present through hydrogen-π and π-π interactions between the phenyl and pyrazolate rings of neighbour-ing molecules, while the shortest intermolecular Co1⋯Co1′ distance is ∼9.3 Å.
Direct current (dc) magnetic susceptibility measurements were performed on a polycrystalline sample of 1 restrained in eicosane in the 290-2 K temperature range in an applied magnetic field of 1000 Oe (Fig. 2). The χ M T value at room temperature (2.45 cm 3 mol −1 K) corresponds to a high-spin Co II ion and indicates a spin-orbit coupling contribution (the Co III ions are diamagnetic so for g = 2, S = 3/2, χ M T = 1.88 cm 3 mol −1 K). Upon cooling, χ M T decreases slowly until ∼50 K to reach 2.16 cm 3 mol −1 K and then decreases rapidly below ∼50 K to reach 1.40 cm 3 mol −1 K at 2 K, indicating zero-field splitting of the ground state. Additionally, magnetisation versus field plots at 2, 4 and 6 K did not saturate at the highest available field of 5 T, a further indication of the presence of magnetic anisotropy (Fig. 2 inset).
Microanalysis and powder X-ray diffraction carried out on ground and non-ground samples ( Fig. S1 and S2 †) show that the lattice solvent is easily lost. Such desolvation could cause changes in the crystal packing, resulting in small changes of the local cobalt environment and hence small changes to the Co II g values and zero-field splitting (ZFS) parameters. High frequency EPR studies on 1 (Fig. S5 †) are in agreement and suggest the presence of two discrete species within the microcrystalline powder sample in an ∼50 : 50 ratio, having distinct ZFS parameters. Analysis of the EPR data gives the parameters for the two species as: g x = g y = 2.18, g z = 2.07 with an E/D ratio ∼0. 13 and g x = g y = 2.23, g z = 2.08 with an E/D ratio ∼0.17 (see the ESI †). Using the average of these two sets, g x = g y = 2.205, g z = 2.075, the dc magnetic susceptibility data and the magnetisation curves of 1 were fitted simultaneously using the program PHI 10 (Fig. 2), as described by the following effective Hamiltonian equation (1): The first and second terms represent the axial and rhombic ZFS terms, parameterised by D and E, respectively, Ŝ is the spin operator with components Ŝ i (i = x, y, z), and the final term denotes the Zeeman interaction with the local magnetic field,B, parameterised through the Landé g $ tensor.
Fixing the values of g x = g y = 2.205 and g z = 2.075, and χ TIP = 0.0009 cm 3 mol −1 , where χ TIP stands for the contribution of temperatureindependent paramagnetism arising from two Co III and one TBP Co II , 11 we were able to extract the ZFS parameters D = +23.85 (±0.17) cm −1 and E = +4.04 (±0.09) cm −1 . The E/D ratio extracted from the fitting of the magnetic data is ∼0.17, consistent with the EPR studies. The magnitude of D is also consistent with previously reported Co II centres in TBP geometry with easy-plane anisotropy. 12,13 The relatively high value of the transverse anisotropy indicates significant mixing of the M S = ±3/2 and ±1/2 levels, and can be attributed to the small deviation from the ideal TBP geometry and/or the different nature of the ligands in the equatorial positions. 13 Only a few examples of Co II in TBP geometry with easy-plane anisotropy have been found to show slow relaxation of the magnetisation. 13 Therefore, we performed alternating current (ac) magnetic susceptibility measurements. In zero applied dc field, 1 does not display any out-of-phase ac signals, due to efficient zero-field quantum tunnelling. However, by using an applied dc field to suppress tunnelling, compound 1 does display slow magnetic relaxation at low temperature. Variable dc fields (500-5000 Oe) were applied to 1 at 2 K in order to obtain the optimum dc field at which the characteristic relaxation time of the magnetisation (τ) possesses the largest value (Fig. S7, ESI †). The characteristic relaxation times for each field were calculated using CC-FIT, 14 and the τ max value was obtained at 1000 Oe. The frequency dependence of the in-phase and out-of-phase magnetic susceptibility was measured under the optimum dc field for the range of temperatures 1.8-8 K (Fig. 3).
The fitting of the Cole-Cole plot (out-of-phase versus inphase signals) for 1 was performed using CC-FIT 14 (Fig. 4), resulting in small values of the Cole-Cole parameter α (0.08-0.02) indicative of a relatively narrow distribution of relaxation times. The τ values were used to construct an Arrhenius plot for the temperatures 1.8-5 K, from which the relaxation parameters of ΔE/k B (energy barrier) and τ 0 (preexponential factor) at higher temperatures were extracted for 1 (Fig. S8, ESI †). Fitting within the linear region (Orbach relaxation mechanism), the values ΔE/k B = 23.18 (±2.2) K and τ 0 = 1.14 × 10 −7 s were extracted. However, the value of ΔE/k B is smaller than the calculated energy difference between the ground and first excited state , a clear indication that other relaxation processes need to be considered. Using eqn (2), we attempted to fit the τ −1 versus T data but we were not able to extract reasonable values. The terms are the direct, tunnelling, Raman and Orbach contributions, in that order. 15 In order to avoid over-parameterisation we attempted to fit the τ versus field (H) plot using only the terms for direct and tunnelling processes (which are dependent on the field) to extract the parameters A, B 1 , and B 2 . However, all efforts were unsuccessful, an indication that there is a more complicated dependence of τ with the field. Therefore, using only the tunnelling (expressed as the parameter B) and Raman contributions (see eqn (3)) we were able to fit the τ −1 versus T plot (Fig. 4) affording the values B = 926 s −1 , C = 2.3 K −n s −1 and n = 6.6. The exponent factor n in the Raman process should be equal to 9 for Kramers' ions, or 5 in the presence of low-lying states. However, lower values for n have been reported in cases where acoustic and optical phonons are involved. 15,16 In conclusion, [Co II Co III 2 (µ 3 -OH)(µ-pz) 4 (DBM) 3 ] is the only reported example of a mixed-valence Co II /Co III polynuclear complex containing a single trigonal bipyramidal Co II centre that gives rise to slow magnetic relaxation. 17 In the case of 1, this arises from a large, easy-plane magnetic anisotropy. To obtain the zero-field splitting parameters, we used high frequency EPR measurements to extract the g factors. These were then fixed in the simultaneous fitting of the dc magnetic susceptibility and magnetisation data to give the parameters D = +23.85 (±0.17) cm −1 and E = +4.04 (±0.09) cm −1 . Furthermore, it has been demonstrated recently how the magnetic anisotropy of octahedral Co II is transferred to the overall magnetic anisotropy of a polymetallic {Cr 7 Co} system. This is of interest for quantum information processing, especially in relation to molecules with a large spin ground state that is characterised by a large, easy-plane anisotropy. 18 Hence, the next step is to develop a route to incorporate high magnetic anisotropy trigonal bipyramidal Co II centres into exchangecoupled polymetallic systems that contain multiple paramagnetic centres.

Conflicts of interest
There are no conflicts to declare.