Single-molecule magnet properties of a monometallic dysprosium pentalene complex

The pentalene-ligated dysprosium complex [(  8 -Pn † )Dy(Cp*)] (1 Dy ) (Pn † = [1,4-( i Pr 3 Si) 2 C 8 H 4 ] 2– ) and its magnetically dilute analogue are single-molecule magnets, with energy barriers of 188(11) cm –1 and 245(28) cm –1 , respectively. Whilst the [Cp*] – ligand in 1 Dy provides a strong axial crystal field, the overall axiality of this system is attenuated by the unusual folded structure of the [Pn † ] 2– ligand. Single-molecule

Single-molecule magnets (SMMs) are coordination compounds that display a magnetic memory effect and an effective energy barrier (Ueff) to flipping of their magnetic dipoles. 1 Such materials have, thus far, proven to be of significant fundamental interest, however some SMMs have been incorporated into nanoscale molecular spintronic devices. 2 Ligand design continues to be a key strategy for addressing the properties of SMMs, particularly increasing the temperature at which slow magnetic relaxation can be observed.Synthetic approaches to the design of d-and f-block SMMs are dominated by Werner-type coordination chemistry, 3 however the organometallic approach to SMMs has led to some eyecatching recent examples. 4,5Within the context of lanthanide SMMs, well-known organometallic ligands such as cyclopentadienide, [Cp] -, 6,7 cyclooctatetraenide, [COT] 2-, 8 and cycloheptatrienide, [C7H7] 3-, 9 have been used to influence the properties of dysprosium-and erbium-containing SMMs.In several notable examples, theoretical studies have provided detailed insight into how the properties of these organometallic ligands impact upon the electronic structure of the Ln 3+ cation, leading to striking increases in the magnetic blocking temperature (TB) and Ueff.
In light of the advances made to date using organometallic chemistry, considerable scope remains for exploring other non-classical ligands in the context of single-molecule magnetism, hence we now turn our attention to the dianion of pentalene, i.e. [C8H6] 2-or [Pn] 2-, an aromatic bicyclic ligand consisting of two fused C5 rings.Pentalene coordination chemistry 10 is considerably underdeveloped relative to that of more established -organometallic ligands such as cyclopentadienide.However, important developments in the synthesis of pentalene pro-ligands have enabled the study of many pentalene complexes, which, in addition to the fundamental interest in their chemistry, have applications in catalysis and small-molecule activation, 11 and as models for metal-containing polymers. 127][8] In particular, the formal dianionic charge and the  8 -coordination mode of pentalene, combined with the fold angle between the two fused rings, 10 provide a unique platform on which to construct new magnetic materials.We now describe the SMM properties of [( 8 -Pn † )Dy(Cp*)] (1Dy) (Pn † = [1,4-( i Pr3Si)2C8H4] 2-) and its magnetically dilute analogue, which were synthesized according to Scheme 1.The addition of one stoichiometric equivalent of K2Pn † to MCl3 (M = Y, Dy) in thf, followed by one equivalent of NaCp*, produced orange solutions from which crystals of 1Dy and 1Y were isolated in yields of 35% and 30%, respectively.X-ray crystallography confirmed the expected isostructural nature of 1Dy (Fig. 1) and 1Y (Fig. S4) (Tables S1, S2), with the metal centres being bound to an  8 -Pn † ligand and an  5 -Cp* ligand.The Dy-Pncent distances of 2.235(3) Å are significantly shorter than the analogous Cpcent distance of 2.344(5) Å ('cent' denotes the centroid of a C5 ring).The Dy-C distances to the pentalene bridgehead carbon atoms C(4) and C(5) are 2.359(7) Å and 2.371(7) Å, whereas the distances to the wingtip carbons C( 2) and C( 7) are considerably longer at 2.749(6) Å and 2.731(6) Å, respectively.The Dy-C distances to the carbon atoms in the intermediate positions lie in the range 2.600(6)-2.640(6)Å and the pentalene fold angle is 26.9( 4) (Fig. S5).The range of Dy-C distances to the Cp* ligand is 2.610( 9)-2.643(12) Å (average 2.62 Å).The two Pncent-Dy-Cpcent angles are 152.47(11)° and 153.05( 11)°.The dysprosium centre in 1Dy resides 0.200(2) Å above the plane of the three centroids, resulting in a pyramidal coordination environment with approximate Cs symmetry.The shortest intermolecular DyDy distance is 8.8313(8) Å.The solid-state molecular structures of 1Dy and 1Y are also consistent with the solution-phase structure of diamagnetic 1Y, as confirmed by 1 H, 13 C and 29 Si NMR spectroscopy (Fig. S1-S3).
The magnetic properties of 1Dy, which were measured in a static (D.C.) field of 5000 Oe, are typical of a monometallic Dy 3+ complex with a 6 H15/2 ground term.Thus, the value of MT, where M is the molar magnetic susceptibility, is 13.51 cm 3 K mol -1 at 300 K (Fig. S6), which is close to theoretical value of 14.17 cm 3 K mol -1 .A gradual decrease in MT was observed down to about 20 K, at which point a precipitous drop occurred and a value of 7.60 cm 3 K mol -1 was reached at 2 K.The overall temperature dependence of the susceptibility is indicative of depopulation of higher-lying crystal field states of Dy 3+ as the temperature is lowered, followed by the onset of magnetic blocking at very low temperatures.At 1.8 K and 5 K, the magnetization (M) of 1Dy increases rapidly up to fields of about 10 kOe, followed by a more gradual increase at higher fields and reaching values of 5.0 B at 70 kOe (Fig. S6).
The SMM properties of 1Dy were revealed through measurements of the in-phase (') and out-of-phase ('') A.C. magnetic susceptibility as a function of frequency () (Figs 2 and S7).The ''() plot shows a series of well-defined maxima in the temperature range 2-41 K, with the position of the maxima shifting to higher frequencies as the temperature is i.e.Dy@1Y produced  parameters of 0-0.39 and Ueff = 245(28) cm -1 , 0 = 4.14  10 -8 s, C = 0.00639 s -1 K -n , n = 3.62 and QTM = 4.63 s.Magnetic hysteresis in 1Dy was observed by measuring the field-dependence of the magnetization with a sweep rate of 6.6 Oe s -1 .Waist-restricted hysteresis loops were observed up to 2.4 K, although without any coercivity owing to prominent QTM processes.Similar measurements on Dy@1Y allowed wider loops with small (e.g. 100 Oe at 1.8 K) coercive fields to be observed up to 3.0 K, which is consistent with the reduced significance of QTM in the diluted sample.To provide further insight into the magnetic properties, the electronic structure of 1Dy was studied by multi-reference ab initio calculations. 13The coordinates of the heavy atoms were used in the calculations as determined by X-ray crystallography, and the positions of H atoms were optimized at the DFT level (see ESI for details).The experimental and calculated MT(T) agree well (Fig. S6), with the calculated MT value at 300 K being 13.80 cm 3 mol -1 K, compared to the experimental value of 13.51 cm 3 mol -1 K.The deviation is not large (~2%) and most likely results from neglecting electron correlation outside the 4f orbital space in the CASSCF calculations.The most important qualitative feature of the plot, namely, the gradual decrease in MT upon decreasing the temperature, is correctly produced.The calculated M(H) plots are in very good agreement with experiment (Fig. S6).
The energies of the eight lowest Kramers' doublets within the ground 6 H15/2 multiplet of 1Dy, along with the principal components of the respective g-tensors and the angles between the ground and excited doublets are listed in Table S3.The principal magnetic axis of the ground doublet in 1Dy passes through the centre of the Cp* ligand and the midpoint of the fused pentalene C-C bond (Fig. 1).The ground doublet is almost axial, with a large gz component and small transverse components, hence the QTM is completely blocked in the ground doublet.The angles between the magnetic axes of the ground doublet and the first three excited doublets are small, and then quickly become perpendicular in the higher doublets.The first excited doublet lies at 197 cm -1 , which is quite close to the experimentally observed barrier height of 188 cm -1 in 1Dy.In the first excited state, the transverse components of the g tensor are still small, but not vanishingly so, and the QTM process is not completely blocked.Based on the experimental evidence, the QTM in this doublet is significant enough such that thermally activated QTM via the first excited doublet is the dominant relaxation mechanism.
The splitting of the 6 H15/2 multiplets in 1 was studied further by calculating the ab initio crystal field (CF) parameters, 14 which are listed in Table S4.The decomposition of the SO-RASSI wave-functions of the sixteen lowest states (eight lowest doublets) into squared projections onto |  ⟩ states (where J = 15/2) is given in Table S5.The states in the lowest doublet have large squared projections (0.925) on the MJ = ±15/2 states, as is usual for Dy 3+ SMMs. 6,7The MJ states become increasingly mixed as one moves to higher doublets.The first excited doublet has a squared projection of 0.888 on the MJ = ±13/2 state and therefore still approximates to the MJ = ±13/2 states.In higher doublets the correspondence of the SO-RASSI states with a single given MJ state is lost.
The mechanisms for the relaxation of magnetization in 1Dy was studied by constructing the qualitative relaxation barrier using a previously proposed method. 15Plotting the energies of the lowest states against their respective magnetic moments, with the states being connected by their transition magnetic moment matrix elements, provides the relaxation route corresponding to a pathway traced by the largest matrix elements.The resulting plot (Fig. 4) retains its "barrier-like" structure up to the sixth doublet.Based on the calculations, the most probable relaxation route in 1Dy is an Orbach mechanism via the third excited Kramers doublet at 498 cm -1 .However, the experimental data for 1Dy show that the relaxation takes place via the first excited doublet.The QTM in this doublet is weak (roughly an order of magnitude stronger than in the ground doublet), but strong enough to overcome the Orbach route.In addition, the increase in the barrier of approximately 60 cm -1 upon dilution implies that Orbach-type relaxation is not the only mechanism, and hence that a Raman process should also contribute.
The anisotropy barriers and hysteresis properties determined for 1Dy and Dy@1Y are reminiscent of those found in the series of dysprosium metallocene SMMs reported by some of us, 6,7a which have very similar Dy-C(Cp) distances to 1Dy.Since Cp ligands in axial positions are known to promote SMM properties in complexes of Dy 3+ , the bridgehead pentalene carbon atoms, which occupy axial positions and are much closer to the metal centre, should also enhance the magnetic axiality.However, it is noticeable that the other Dy-C(Pn) distances -and the positions of the carbon atoms with respect to the metal centre -are similar to those found in dysprosium complexes of the [COT] 2-ligand.Since [COT] 2-is thought to diminish the magnetic axiality of the prolate Dy 3+ ion in, e.g., [Dy(COT)2] -, 8 we can propose that the nonbridgehead pentalene carbons in 1Dy provide a non-negligible equatorial field and therefore produce an effect similar, yet stronger, to that of COT.Furthermore, the Pncent-Dy-Cpcent in 1Dy angles are 152.47(11)° and 153.05(11)°, hence they are very similar to the Cp-Dy-Cp angle of 152.845(2)° in [(Cp ttt )2Dy] + , an SMM with a barrier of 1277 cm -1 and a TB of 60 K. 7a Since the properties of [(Cp ttt )2Dy] + arise from the exceptional axiality of the ligand environment, the two opposing C5 rings in the pentalene ligand of 1Dy effectively compete with each other in a way that diminishes the axiality.Hence, 1Dy is an SMM but with a modest barrier and waistrestricted hysteresis.The large, non-axial  2 2 parameter (Table S4) also explains the significant mixing of the higher-lying Kramers doublets.The principal reason for the magnetic axiality in 1Dy is therefore the relatively large negative axial crystal field parameter  2 0 , whereas the other important axial parameters  4 0 and  6 0 are smaller.In summary, the magnetic properties of the first pentaleneligated SMM have been described.Large anisotropy barriers were determined for 1Dy and its magnetically dilute analogue, the origins of which were assigned to the strong axial field provided by the [Cp*] -ligand and the bridgehead carbon atoms of the [Pn † ] 2-ligand.The dominant relaxation process in 1Dy is thermally activated process via the second Kramers doublet.The appreciable equatorial field provided by the nonbridgehead carbon atoms attenuates the Ueff value and results in magnetic hysteresis occurring without coercivity.In terms of magneto-structural correlations, the folded nature of the pentalene ligand provides a unique coordination chemistry strategy for addressing the electronic structure of Ln 3+ cations, and our on-going research will apply this in the design of magnetic, spintronic and optical materials.

Fig. 1
Fig. 1 Left: Thermal ellipsoid representation (50% probability) of the molecular structure of 1Dy (with H-atoms and i Pr groups omitted for clarity).Right: the principal axis of the g-tensor in the ground Kramers doublet of 1Dy.

Fig. 2
Fig. 2 Frequency dependence of  in zero applied field for 1Dy.The solid lines are a guide for the eye.

Fig. 3
Fig. 3 Temperature dependence of  for 1Dy (circles) and Dy@1Y (squares).Solid red lines represent fits of the data using the parameters stated in the text.

Fig. 4
Fig. 4 Calculated magnetic relaxation barrier for 1Dy.Darker arrows indicate the largest matrix elements, indicating the most probable relaxation route.