Rare-Earth Cyclobutadienyl Sandwich Complexes: Synthesis, Structure and Dynamic Magnetic Properties

: The potassium cyclobutadienyl [K 2 {  4 -C 4 (SiMe 3 ) 4 }] ( 1 ) reacts with MCl 3 (THF) 3.5 (M = Y, Dy) to give the first rare-earth cyclobutadienyl complexes, i.e. the complex anions [M{  4 - C 4 (SiMe 3 ) 4 }{  4 -C 4 (SiMe 3 ) 3 -  -(CH 2 SiMe 2 }] 2– , ( 2 M ), as their dipotassium salts. The tuck-in alkyl ligand in 2 M is thought to form through deprotonation of the ‘squarocene’ complexes [M{  4 -C 4 (SiMe 3 ) 4 } 2 ] – by 1 . Complex 2 Dy is a single-molecule magnet, but with prominent quantum tunnelling. An anisotropy barrier of 323(22) cm –1 was determined for 2 Dy in an applied field of 1 kOe, and S-shaped magnetic hysteresis loops were observed up to 7 K.

[a] Dr F.-S.  0.34 and 0.22 ppm in a 1:9:9:18 ratio, implying partial protonation of a Cb ligand (Figures S9-S11).Analysis by X-ray crystallography confirmed that the by-product is the potassium cyclobutenyl complex 3 (Scheme 1, Figure 3, Table S1).The structure of 3 consists of a potassium center coordinated to a planar  3 -cyclobutenyl ligand, with additional coordination by a molecule of toluene.The cyclobutenyl C-C bond lengths are 1.553(3) Å for C1-C2/C2′ and 1.431(3) Å for C2/C2′-C3, with K-C distances in the range 2.907(2)-3.001(3)Å.The SiMe3 substituent on C1 is disposed at an angle of 115.32(15)° relative to the cyclobutenyl ring, with the hydrogen on C1 engaging in an agostic interaction with potassium.The extended structure of 3 is a zigzag-type coordination polymer by virtue of CHK interactions with the SiMe3 groups of neighbouring molecules (Figure 3).The isolation of 3 supports the notion that a 'squarocene' complex of the type [M{ 4 -C4(SiMe3)4}2] -does indeed form, but also that it is subsequently deprotonated by 1, leading to the formation of 2M.
12a,13d] The magnetic properties of materials containing the {Dy( 4 -C4R4)n} building block should therefore be very interesting, and hence [K2(toluene)][2Dy] was studied in this context.In a static (D.C.) field of 5 kOe, the temperature dependence of MT, where M is the molar magnetic susceptibility, is typical of a monometallic Dy 3+ complex.The value of MT at 300 K is 13.54 cm 3 K mol -1 , which gradually decreases to 11.90 cm 3 K mol -1 at 25 K, before rapidly decreasing to a value of 1.95 cm 3 K mol -1 at 2 K (Figure S14).The field dependence of the magnetization, M(H), of [K2(toluene)][2Dy] at 1.8 and 5 K both show a rapid increase in M as the field is increased to 2 T before slowly increasing to saturation values of

Chemistry -A European Journal
This article is protected by copyright.All rights reserved.
5.01 and 4.96 B for 1.8 K and 5 K respectively (Figure S15).
Measuring the in-phase (') and the out-of-phase ('') susceptibility of [K2(toluene)][2Dy] using a small A.C. field of 1.55 Oe and zero D.C. field, as a function of frequency () and temperature, revealed slow magnetic relaxation properties, but with strong quantum tunneling of the magnetization (QTM).
Maxima in the ''() data were only found at very high frequencies (997 Hz) (Figure S17), making it difficult to extract an effective energy barrier to reversal of the magnetization (Ueff).However, measuring '() and ''() at 25 K in several DC fields in the range 200-1800 Oe (Figure S18) allowed the relaxation time () to be determined for each field by fitting of the associated Cole-Cole plots of ''(') (Figures S19-S27).These measurements allowed an optimum field of 1000 Oe to be determined (Figure S28, Table S2) and, hence, used to suppress the QTM for the in-field A.C. susceptibility measurements.Under the optimum conditions, clear maxima in ''() were observed in the temperature range 8-40 K (Figures 4, S29).The plot of ''(') (Figure S30, Table S3) was then fitted using a generalized Debye model and  parameters of 0.013-0.042,which indicate a very narrow distribution of relaxation times.A plot of the relaxation times as ln() vs. T -1 (Figure S31) is linear at T = 36-40 K, whilst at lower temperatures the data adopt a curved appearance.Since the application of a DC field mitigates the effects of QTM, the relaxation in [K2(toluene)][2Dy] can therefore be assigned to Orbach processes at higher temperatures and Raman processes at lower temperatures.Fitting the ln() vs. T -1 data to the hightemperature linear region using  −1 =  0 −1  − eff / B  yielded Ueff = 323(22) cm -1 and 0 = 1.83  10 -9 s.
Detailed insight into the magnetic relaxation in 2Dy was obtained using multireference ab-initio calculations (see SI for details). [15]The calculations incorporated three potassium ions involved in - 4 -bridging interactions with the Cb ligand, an  6toluene ligand, and two nearest neighbour complexes of 2Dy (Figure S32).The principal magnetic axis of the ground Kramers doublet within the 6 H15/2 ground multiplet is oriented towards the centres of the [Cb] 2-ligands (Figure 5), thus establishing an important magneto-structural property reminiscent of related dysprosium metallocene SMMs. [12]The calculated g-tensors (Table S4) and crystal field parameters (Table S5) [16] show that the ground doublet is highly axial, with the projection on the MJ = ±15/2 state being 0.949 (Table S5).However, the g-tensors consist of a small-but-significant transverse component, i.e. gx = 0.0076, gy = 0.0130, gz = 19.7338,which is sufficient to induce QTM in the absence of an external magnetic field, consistent with experimental observations.The excited doublets in 2Dy do not correspond to any definite projection, with the first-excited doublet consisting only of 0.591 MJ = ±13/2 character, indicating strong mixing by the non-axial components of the crystal field.Indeed, the off-diagonal crystal field parameters for this system are substantial, with, e.g.,  2 2 being much larger than  2 0 (Table S5), which is most likely due to the influence of the equatorial tuck-in ligand.The first-excited doublet lies at 284 cm -1 , which agrees qualitatively with the experimental barrier, taking into account experimental error and the omission from the calculations of electron correlation effects outside the 4f active space.Overall, the calculations point to the most probable relaxation mechanism occurring via the firstexcited doublet, with prominent QTM in the ground doublet, which is fully consistent with the experimental observations.Since the magnetic memory properties of SMMs have been used a basis for proposing nanoscale devices, [13a] the magnetization (M) vs. field (H) hysteresis properties of [K2(toluene)][2Dy] are also of interest.The M(H) data (Figure 4) show S-shaped loops which remain open up to 7 K (scan rate of 50 Oe s -1 ), but without any appreciable coercivity.In light of the prominent QTM in 2Dy, this observation is somewhat surprising, but can be attributed to the large Dy•••Dy separation of 10.45 Å, which is sufficiently large to mitigate against the effects of dipolar exchange.This hysteresis properties are broadly consistent with the divergence in the field-cooled and zero-field-cooled magnetic 10.1002/chem.201804776

Chemistry -A European Journal
This article is protected by copyright.All rights reserved.susceptibility, which occurs at 3.6 K (Figure S16).
In summary, we have described the synthesis and structures of the first rare-earth cyclobutadienyl complexes, 2M, which were obtained from the reaction of [K2{ 4 -C4(SiMe3)4}] (1) with MCl3 (M = Y, Dy).The 'squarocene' complexes 2M consist of a bis( 4 -cyclobutadienyl) coordination environment with the metal centers additionally bound to a tuck-in [CH2SiMe2] -ligand, the formation of which can be accounted for by deprotonation of the postulated sandwich complex [M{ 4 -C4(SiMe3)4}2] -by 1. Whilst 2Dy is an SMM but with prominent QTM, the large anisotropy barrier of 323(22) cm -1 and the open hysteresis loops observed for a system with a strong equatorial crystal field provide an enticement for pursuing a bis(cyclobutadienyl)dysprosium complex as a potentially outstanding SMM, which we are actively pursuing.Introducing the cyclobutadienyl ligand into rare-earth chemistry also furnishes new opportunities for studying the chemistry of these elements in a broader context, potentially including catalysis and small-molecule activation.

Figure 1 Figure 2
Figure 1 Molecular structure (left) and the extended structure (right) of 1. Thermal ellipsoids at 50% probability.

Figure 5
Figure 5 Principal axis of the g-tensor in the ground KD of 2Dy.