Synthesis and crystal structure of Dy2VN@Ih(7)-C80.
Dy2VN@Ih(7)-C80 was synthesized by a direct arc-discharge method and purified by high-performance liquid chromatographic (HPLC) separation. The analytical HPLC profile and high-resolution mass spectrum of Dy2VN@Ih(7)-C80 are shown in Figure 1. The electronic properties of Dy2VN@Ih(7)-C80 were investigated by vis-NIR spectroscopic and cyclic voltametric (CV) studies. The absorption spectrum (Fig. 1c) of Dy2VN@Ih(7)-C80 shows four absorption peaks at 465, 581, 663, and 773 nm with the absorption onset at ~ 1424 nm, corresponding to an optical bandgap of 0.87 eV. These absorption bands have significant red shifts compared to those of the reported Dy2ScN@Ih(7)-C80 (absorption bands: 406, 565, 676, and 705 nm)33, suggesting the crucial role of the entrapped V atom on the electronic structure of the whole EMF molecule. The recorded CV curves are shown in Fig. 1d and the characteristic redox potentials are listed in Table S4. Dy2VN@Ih(7)-C80 has one reversible oxidation process (0.05 V) and four reduction processes (-0.81, -1.53, -1.81, and -2.35 V), in which the first, third and fourth reduction processes are all reversible, but the second reduction process is irreversible. The first oxidation potential and the first reduction potential are 0.05 V and -0.81 V, respectively, leading to an electrochemical bandgap of 0.86 eV, which is in perfect agreement with its optical bandgap.
Molecular structure of Dy2VN@Ih(7)-C80 and interactions of the mixed metal atoms inside the cage are studied by X-ray crystallography (Fig. 2). Inside the cage, the N atom is fully ordered which locates on the crystallographic symmetric plane. Albeit of some disorder for Dy and V (Fig. S2 and Table S2), the major sites can be clearly distinguished. Dy1 with an occupancy value of 0.43 is the major Dy site and Dy1A generated by crystallographic operations from Dy1 is the other. The major V site locates on the crystallographic mirror plane with the occupancy value of 0.44. Due to the much smaller ion radius of V (0.64 Å) than that of Dy (0.91 Å), their metal-cage and metal-N distance show remarkable differences. The distance between Dy1 or Dy1A and the nearest cage carbon atom is 2.230 Å, whereas the distance between V1 and the nearest cage carbon atom is 2.041 Å. The relative orientation between the inner metals and the cage resembles that of Sc2VN@Ih(7)-C80 which also contains the transition metal V34, but is obviously different from those of the reported lanthanide metal-based EMFs, such as MSc2N@Ih(7)-C80 (M=La, Ce, Gd, Tb)35–37 where the large lanthanide metal atom resides under the centers of hexagons while the small Sc atoms are close to an intersection between a hexagon and a pentagon. This dramatic difference demonstrates the crucial role of the entrapped transition metal V in the metal-cage interactions.
In the Dy2VN cluster, the bond length of Dy1-N is 2.090 Å while the V-N bond is much shorter (1.841 Å). These clear differences allow us to unambiguously identify V from Dy which is difficult to distinguish in other metallofullerenes containing mixed metals such as Sc3-xVxN@C80(x =1-2)34,38. The included angles of Dy1-N-V1 and Dy1-N-Dy2 are 119.66° and 120.50°, respectively, so that the sum of the three angles is 359.82° (≈360°), suggesting a planar structure of the Dy2VN cluster. With the spatial confinement effect of the fullerene cage on the endohedral Dy2VN cluster, the Dy-Dy distance has been significantly suppressed to 3.629 Å, which is the shortest among all the reported values for 3d-4f SMMs, as summarized in Table 1. The strong cage confinement is further reflected by the ultrashort Dy-V distance of 3.401 Å, which represents one of the smallest 3d-4f distances among all the reported 3d-4f SMMs. The compressed nature of the Dy2VN cluster inside the cage is further verified by DFT calculations (Figure S7). The optimized geometry of Dy2VN@Ih(7)-C80 resembles the crystal structure and in particularly confirms the ultrashort metal-metal distances. The calculated Dy-Dy and Dy-V distances are 3.631 Å and 3.491/3.331 Å, respectively, which are comparable to the recorded values in crystal structures (vide supra). Additional theoretical analysis shows weak Dy-Dy and Dy-V bonding with respective Wiberg bond order of 0.05 and 0.12/0.15, respectively. Such a unique cage confinement effect on the metals strengthens the interactions between the anisotropic lanthanide metals, and enhances the 3d-4f magnetic coupling, slowing down the magnetic relaxation and improving the SMM performance (vide infra).
Table 1. Dy-Dy and 3d-4f distances in same representative 3d-4f SMMs.
3d-4f SMMs
|
Dy-Dy distance / Å
|
3d-4f distance / Å
|
ref.
|
Dy2VIIIN
|
3.629
|
3.401
|
this work
|
DyCuII5
|
N/A
|
3.904~3.964
|
[39]
|
Dy2CuII10
|
4.291
|
3.899~3.965
|
[39]
|
Dy2CrIII2
|
4.105
|
3.290~3.302
|
[40]
|
Dy2FeIII4
|
3.952
|
3.431~3.450 / 5.766~5.973
|
[41]
|
DyFeII2
|
N/A
|
3.520~3.599
|
[22]
|
Dy2CoII2
|
6.183
|
3.489~3.496
|
[23]
|
Dy2MnII2
|
6.292
|
3.553~3.613
|
[24]
|
DyVIVO
|
N/A
|
3.474
|
[42]
|
Dy2VIIIN = Dy2VIIIN@C80; DyCuII5=[DyCuII5(quinha)5(sal)2(py)5]-(CF3SO3)·py·4H2O;39
Dy2CuII10 = [Dy2CuII10(quinha)10 (sal)2(OH)(py)9] (CF3SO3)3·2py·2CH3OH· 2H2O;39
Dy2CrIII2 = CrIII2Dy2(OMe)2(O2CPh)4(mdea)2(NO3)2;40
Dy2FeIII4 = [FeIII4Dy2(μ3-OH)2 (mdea)6(SCN)2(NO3)2(H2O)2]·4H2O·2MeCN;41
DyFeII2 = [FeII2Dy(L)2(H2O)]ClO4·2H2O, L = 2,2′,2′′-(((nitrilotris(ethane-2,1-diyl))tris(azanediyl))tris(methylene))tris(4-chlorophenol);22
Dy2CoII2 = [CoII2Dy2(L)4(NO3)2(THF)2] ·4THF,H2L = (E)-2-(2-hydroxy-3-methoxybenzylideneamino)phenol;23 Dy2MnII2 = Dy2MnII2(L)4(NO3)2(DMF)2, H2L = (E)-2-ethoxy-6-(((2-hydroxyphenyl)imino)methyl)phenol;24
DyVIVO = Dy(VIVO)L(NO3)3(H2O), H2L = N, N′-bis(1-hydroxy-2-benzylidene-6-methoxy)-1,7-diamino-4-azaheptane.42
Magnetic Properties.
The magnetic properties of Dy2VN@Ih(7)-C80 were measured on a Quantum Design MPMS3 SQUID magnetometer (Fig. 3 and Fig. S3-S6). The temperature dependent susceptibility data were collected under 1 kOe direct current (dc) field on a power sample in the temperature range of 2-300 K. On cooling, the χmT values slightly go up from 26.4 cm3 K mol-1 at 300 K to 27.02 cm3 K mol-1 at 100 K and increase sharply with further cooling, reaching a maximum of 34.4 cm3 K mol-1 at around 10 K (Fig. 3a). The significant increase of χmT values with temperature decreasing implies ferromagnetic interaction within the cluster. Below 10 K, the χmT values drop rapidly due to the depopulation of the Stark levels and magnetic anisotropies as the temperature decreases. The χmT value (27.02 cm3 K mol-1) at 300 K is slightly smaller than the theoretical value of 29.34 cm3 K mol-1 for two uncoupled Dy (III) ions (the ground state 6H15/2 and gJ = 4/3) and one V (III) ion (S = 1 and g = 2).
Fig. 3. Magnetic properties of Dy2VN@Ih(7)-C80. a, Temperature dependence of χmT products under 1 kOe dc field and the fitted temperature-dependent susceptibilities using Lines model. The experimental values were first scaled by a factor of 1.09 to make them consistent with the theoretical values at room temperature before fitting. b, Zero-field-cooled (ZFC) magnetization and field-cooled (FC) magnetization under 2 kOe dc field at sweep rate of 3 K min-1. c, Magnetic hysteresis at different temperatures with the sweep rate of 200 Oe s-1. The inset shows that the opening hysteresis can be observed at up to 12 K. d, Comparison of the blocking temperature (TB, ZFCFC and TB, loop) and coercive field (Hc) between Dy2VN@Ih(7)-C80 and typical 3d-4f SMMs23,39–41. The solid symbol represents TB, ZFCFC while the hollow symbol corresponds to TB, loop. The pentacle represents Dy2VN@Ih(7)-C80 in this work. The triangle stands for the SMMs in Table S6 for which no related values are given. The circles stand for the other SMMs shown in Table S6.
The zero-field-cooled (ZFC) magnetization and field-cooled (FC) magnetization (ZFC-FC) data were collected in heating mode under 2 kOe dc field. The temperature of bifurcation in the ZFC-FC curve (TB, ZFCFC) of Dy2VN@Ih(7)-C80 is determined to be 9.5 K (Fig. 3b), which offers the highest TB in all reported 3d-4f SMMs (Fig. 3d and Table S6). In accordance with its TB, ZFCFC, Dy2VN@Ih(7)-C80 exhibits open magnetic hysteresis up to 12 K (TB, loop), as described in Fig. 3c. The hysteresis is quite broad with a coercive field of 2.73 T at 2 K (Fig. 3d and Fig. S5). To the best of our knowledge, this is among the largest coercive fields in all reported 3d-4f SMMs30-32, 50-53(Fig. 3d and Table S6).
Spin dynamics of Dy2VN@Ih(7)-C80 were also characterized by time-dependent dc measurements. Zero-field magnetization relaxation times τ below 15 K were determined by free-order exponential fitting of magnetization decay curves recorded after being magnetized by 1 kOe dc field (Fig. S6 and Table S7). The fitting of relaxation times τ vs. T-1 could be accomplished by combining Orbach and QTM processes using equation Eq1:
The best fit gives the Orbach barrier of U1 = 70.7 K (Fig. 4a), which is actually the exchange barrier. And the fitted QTM relaxation time is τQTM = 1249.8 s (Fig. 4a), indicating that QTM is effectively inhibited in Dy2VN@Ih(7)-C80. TB can also be obtained directly from the relaxation times. Sessoli et al. have suggested a more universal SMM characteristic, TB, 100, the temperature at which the relaxation time is 100 s43. However, for most of the reported SMMs, the relaxation times were usually much shorter than 100 s within the detectable temperature range. To the best of our knowledge, there is only one 3d-4f SMM [Dy2CuII10(quinha)10(sal)2(OH)(py)9]3+ whose TB, 100 is ca. 2 K39. For Dy2VN@Ih(7)-C80, TB, 100 deduced from the temperature dependence of τ reaches to ca. 7.2 K.
Fig. 4. Exchange energy barriers of Dy2VN@Ih(7)-C80. a, Plots of the logarithm of the relaxation times (lnτ) vs. reciprocal temperature (T-1). The circles represent the relaxation times τ extracted from magnetization decay curves. The blue solid line indicates the fitting with the combination of Orbach and QTM processes, while the red line represents the result of only using Orbach process above 6 K (see in Supporting Information). b, Energy levels obtained from Lines model. The black lines represent the pseudo-doublets as a function of their magnetic moments along the magnetic axis. Inset: low-lying energy levels of the exchange part. The numbers at each arrow stand for the mean absolute value of the corresponding matrix element of the transition magnetic moment. The red arrow corresponds to the deduced relaxation pathway (ULines = 63.4 K).
Theoretical Analysis.
The performance of polynuclear SMMs is affected not only by the magnetic anisotropy of single ion, but also the interactions between paramagnetic centers. Therefore, exploration of magnetic interactions in polynuclear systems utilizing theoretical calculation is helpful to reveal the origin of magnetic relaxation. The studied molecule Dy2VN@Ih(7)-C80 is a three-center spin system and the magnetic states can be described with the total spin Hamiltonian (Eq2):
where the first two terms describe the crystal-field (CF) effect of the Dy centers and the third stands for the zero-filed splitting (ZFS) of the V center. The fourth and fifth terms describe the interactions between the magnetic centers while the last term represents the Zeeman interaction. The interactions between different Dy centers and V center are considered to be identical as the two Dy centers are structurally identical.
Before proceeding to the discussion of the interactions within the cluster and understanding the energy-level spectrum simulated with the Hamiltonian (Eq2), single-ion CF parameters for Dy center and ZFS parameters for V center need to be treated properly. Ab initio calculations were performed at the CASSCF/SO-RASSI level of theory. The computational results show that both Dy centers have strong uniaxial magnetic anisotropy and the ground state easy-axis is aligned along the metal-nitrogen bond, as shown in Fig. S8. The calculated ground state g tensors of Dy centers are very close to the values of the Ising limit states and the overall crystal-filed splitting (CFS) is up to 1500 cm-1, with the first excited state of higher than 400 cm-1 (Table S10). In contrast, the calculated E/D for V center is 0.31, thus it could be treated as isotropic with an average giso of 1.9. The calculated high magnetic excited states of both single Dy center and V center indicate that the much lower U1 may stem from the magnetic coupling of spin centers within the fullerene cage.
Table 2. The magnetic interactions obtained using Lines model.
|
Dy-Dy
|
Dy-V
|
Jdip / cm-1
|
6.25
|
0.80
|
Jexch / cm-1
|
-25
|
52.50
|
Jtotal / cm-1
|
-18.75
|
53.30
|
With the corresponding magnetic properties of the mononuclear fragments in hand, the magnetic interactions were explored to further elucidate the magnetic relaxation of Dy2VN@Ih(7)-C80. The temperature-dependent susceptibilities of Dy2VN@Ih(7)-C80 were fitted based on Eq2 and only the data above 20 K were considered during the fitting (See in Methods, Fig. 3a). Given the fitting result from Table 2, the interaction between Dy centers is characterized as antiferromagnetic while the interaction between Dy and V is ferromagnetic. The exchange-coupled levels and g-tensor values of Dy2VN@Ih(7)-C80 were also calculated and summarized in Table S11. These levels are grouped into pseudo-doublets split by the tunneling gap (∆t) because of the even number of unpaired electrons in Dy2VN@Ih(7)-C80. The calculated ground exchange state features an extremely small ∆t, indicating that QTM is suppressed and large Hc could be observed. On the contrast, the third excited pseudo-doublet possesses a relatively large ∆t, thus the relaxation pathway could be deduced as indicated with red arrows in Fig. 4b combining with the transition probabilities. The inferred exchange barrier gives ULines = 63.4 K and is well consistent with the energy barrier fitted from the data of demagnetization (U1 = 70.7 K). Although the splitting of the energy states is much lower than the crystal splitting of an individual Dy(III), the strong coupling between Dy(III) and V(III) can effectively suppress QTM, leading to large coercivity and high TB.