Nitronyl Nitroxide Biradical-Based Binuclear Lanthanide Complexes: Structure and Magnetic Properties

: Employing a new nitronyl nitroxide biradical NITPhPzbis(NITPhPzbis = 5-(1-pyrazolyl)-1,3-bis(1’-oxyl-3’-oxido-4’,4’,5’,5’-tetramethyl-4,5-hydro-1 H -imidazol-2-yl)benzene), a series of 2 p -4 f complexes [Ln 2 (hfac) 6 (H 2 O)(NITPhPzbis)] (Ln III = Gd 1 , Tb 2 , Dy 3 ; hfac = hexaﬂuoroacetylacetonate) were successfully synthesized. In complexes 1 – 3 , the designed biradical NITPhPzbis coordinates with two Ln III ions in chelating and bridging modes to form a four-spin binuclear structure. Direct-current magnetic study of Gd analogue indicates that ferromagnetic exchange exists between the Gd ion and the radical while antiferromagnetic coupling dominates between two mono-radicals. Dynamic magnetic data show that the χ ” signals of complex 3 exhibit frequency dependence under zero ﬁeld, demonstrating slow magnetic relaxation behavior in complex 3 . And the estimated values of U e ﬀ and τ 0 are about 8.4 K and 9.1 × 10 − 8 s, respectively.


Synthetic Aspect and Crystal Structures
The specific synthesis process of NITPhPzbis is shown in Scheme S1. Three binuclear Ln-based complexes were successfully prepared by using NITPhPzbis. The elemental analysis results of complexes 1-3 match well with their theoretical values. To further determine the phase purity of the crystal samples, PXRD tests on complexes 1-3 were conducted ( Figure S1). It can be seen that the experimental spectra are consistent with the simulated spectrum of complex 3, which provides an important guarantee for the next study of magnetic property. Figure S2 presents the IR spectra of 1-3. The peaks observed at about 1605 cm −1 , 1506 cm −1 , and 1353 cm −1 are severally attributed to the C=C, C=N, and N−O stretching in the NITPhPzbis ligand while the absorption peaks at about 1648 cm -1 (ν C=O ), 800 cm −1 (δ C-O ), 1247 cm −1 , 1131 cm −1 (ν C-F ), and 658 cm −1 (δ C-F ) originate from the co-ligand hfac − . The weak absorption bands at about 3575cm −1 are due to the coordinated H 2 O molecules.
Crystallographic study proves that complexes 1-3 are isomorphous (Figure 1 and Figures S3 and S4) and all belong to the triclinic Pī system. The crystallographic data and refinement parameters for 1-3 are included in Table 1, and the important parameters are presented in Table 2. Here, the structure of complex 1 will be a representative for description. As shown in Figure 1a, the NITPhPzbis ligand chelates one Gd(hfac) 3 unit through its two adjacent NO groups, and coordinates with another Gd(hfac) 3 unit through another NO group to form a four-spin binuclear structure. Both Gd1 and Gd2 are in eight-coordinated environment, and the eighth coordination sphere of Gd2 is completed with a H 2 O molecule. There is weak hydrogen bond interaction between the pyrazolyl-N (N6) and the coordinated water molecule in which the distance of O17 water (H-donor)···N6 pyrazol (H-acceptor) is 2.811 Å and the angle of O17 water -H···N6 pyrazol is 148.91 • . According to SHAPE analysis [32,33], the coordination geometry of Gd1 is C 2v symmetry while that of Gd2 is D 2d (Figure 1b, Table S2). In complex 1, the bond lengths of the two Gd1-O rad bonds are 2.383(8) Å and 2.472(7) Å, respectively, and the distance of Gd2-O rad is 2.454(7) Å. The bond lengths here are consistent with those existed in other nitronyl nitroxide biradical-Gd III complexes [26][27][28]. The length of the Gd-O hfac bond ranges from 2.343(9) to 2.412(7) Å. The bond distance of Gd2-O H2O is 2.378(7) Å. The Gd1-O-N-C torsion angles are 61.3 (16) • and −58.5 (17) • , respectively, while the Gd2-O-N-C torsion angle is −64.8 (14) • . The two dihedral angles between benzene ring and the ON-C-NO of mono-radicals are 37.7(5) • and 31.4(4) • while the dihedral angle formed by pyrazole ring and benzene ring is 30.5(4) • . In the binuclear structure of 1, the distance between the adjacent Gd1 and Gd2 is 8.351 Å. Figure 2 shows the crystal packing diagram of 1. The closest separation between intermolecular Gd III ions is 6.265 Å, and the shortest distance between uncoordinated NO···NO is 11.976 Å. The two dihedral angles between benzene ring and the ON-C-NO of mono-radicals are 37.7(5)° and 31.4(4)° while the dihedral angle formed by pyrazole ring and benzene ring is 30.5(4)°. In the binuclear structure of 1, the distance between the adjacent Gd1 and Gd2 is 8.351 Å. Figure 2 shows the crystal packing diagram of 1. The closest separation between intermolecular Gd III ions is 6.265 Å, and the shortest distance between uncoordinated NO•••NO is 11,.976 Å.     2.343(9)-2.412 (7) 2.378 (7) 2.323 (5)

Magnetic Properties
The dc magnetic susceptibility data of three complexes measured in 1000 Oe magnetic field are shown in Figures 3 and 4. For complex 1, the room temperature value of χMT is 16.99 cm 3 Kmol -1 (Figure 3), slightly higher than the value(16.51 cm 3 Kmol −1 ) expected for two Gd III ions(Gd III : 8 S7/2, g = 2, C = 7.88 cm 3 Kmol −1 ) and two mono-radicals (radical: S = 1/2, g = 2, C = 0.375 cm 3 Kmol −1 ) which are magnetically uncorrelated. On lowering the temperature, the χMT value increases steadily to a maximum of 21.35 cm 3 Kmol −1 at 2K, indicating that ferromagnetic exchange is dominant in the system. The field dependent magnetization curves shown in Figure 3 (inset, 2K).The observed M value of 16.15 Nβ at 7T is in agreement with the theoretical saturation value of 16 Nβ. Based on the molecular structure, magnetic exchange pathway in 1 is shown in Scheme 2, in which J1 is the exchange coupling between the Gd III ion and the ligated NO group and J2 represents the magnetic coupling between two mono-radicals through the Gd III ion and/or m-phenylene ring.

Magnetic Properties
The dc magnetic susceptibility data of three complexes measured in 1000 Oe magnetic field are shown in Figures 3 and 4. For complex 1, the room temperature value of χ M T is 16.99 cm 3 Kmol −1 (Figure 3), slightly higher than the value (16.51 cm 3 Kmol −1 ) expected for two Gd III ions (Gd III : 8 S 7/2 , g = 2, C = 7.88 cm 3 Kmol −1 ) and two mono-radicals (radical: S = 1/2, g = 2, C = 0.375 cm 3 Kmol −1 ) which are magnetically uncorrelated. On lowering the temperature, the χ M T value increases steadily to a maximum of 21.35 cm 3 Kmol −1 at 2K, indicating that ferromagnetic exchange is dominant in the system. The field dependent magnetization curves shown in Figure 3 (inset, 2K). The observed M value of 16.15 Nβ at 7T is in agreement with the theoretical saturation value of 16 Nβ. Based on the molecular structure, magnetic exchange pathway in 1 is shown in Scheme 2, in which J 1 is the exchange coupling between the Gd III ion and the ligated NO group and J 2 represents the magnetic coupling between two mono-radicals through the Gd III ion and/or m-phenylene ring.
A quantitative analysis has been performed using PHI software [34,35]. The spin Hamiltonian of this system is H = −2J 1 Ŝ Gd2Ŝrad1 +Ŝ rad1ŜGd1 +Ŝ Gd1Ŝrad2 − 2J 2 Ŝ rad1Ŝrad2 . The experimental dc susceptibility and magnetization data can be well reproduced and the best fitting parameters are g rad = 2.00 (fixed), g Gd = 2.02, J 1 = 0.74 cm −1 , and J 2 = −3.24 cm −1 . The observed Gd-NO ferromagnetic exchange is expected, which can be ascribed to the electron transfer from π* orbital of the nitronyl nitroxide radical to 6s/5d empty orbitals of Gd III ion [36,37]. The obtained magnitude of J 1 is comparable with those in other Gd-nitronyl nitroxide complexes [38]. For J 2 , two kinds of magnetic pathways are active: one is via the Gd III ion, in which antiferromagnetic coupling is usually generated [39]; the other is through m-phenylene ring, which results in ferromagnetic interaction based on spin polarization mechanism [29] (Scheme 3). Consequently, the observed negative J 2 value is the competitive result of these two kinds of magnetic exchange, implying that antiferromagnetic exchange dominates here; this phenomenon has been observed in the similar biradical-Gd compounds [40]. It should be noted that the Gd-NO and NO-NO magnetic interactions observed here are weaker than those in the tetranuclear [Gd 4 (hfac) 12   A quantitative analysis has been performed using PHI software [34,35]. The spin Hamiltonian  A quantitative analysis has been performed using PHI software [34,35]. The spin Hamiltonian   A quantitative analysis has been performed using PHI software [34,35]. The spin Hamiltonian of this system is ( ) ( ) (NITPhO-3Pybis = 5-(3-pyridinyloxy)-1,3-bis(1'-oxyl-3'-oxido-4',4',5',5'-tetramethyl-4,5-hydro-1H-imidazol-2-yl) [28] that contains similar NIT bridged Ln dimer and these can be attributed the differences of Gd-O-N-C torsion angles (61.3(16)°, −58.5(17)° and −64.8(14)° for 1 and −60.2(15)°, 59.5(14)°, and −135.0(13)° for Gd4) in these two complexes which can affect the magnetic exchange between the Gd ion and NO unit [41].  Figures S7 and S8. The M value is 12.72 Nβ and 12.57 Nβ at 8T for 2 and 3, respectively, which is lower than expected saturation value, suggesting the presence of a significant magnetic anisotropy in the systems [42,43].
Due to the presence of magnetic anisotropic Tb III or Dy III ions in complexes 2 and 3, spin dynamic properties were studied using ac susceptibility measurements under zero dc field. As displayed in Figures S9 and 5, there is no non-zero out-of-phase signal for complex 2 while complex 3 exhibits clear temperature-dependent χ" components, indicating slow relaxation of magnetization behavior in 3. However, no visible peaks of the χ" signals are found for complex 3, which can be attributed to the fast quantum tunneling of magnetization (QTM) [44,45]. Accordingly, frequency-dependent ac magnetic susceptibilities for 3 were investigated at 2 K under dc field range of 200-4900 Oe to determine the optimized external field for suppressing the QTM. As seen, no maxima of χ"(v) components can be observed in the measurement window of our apparatus ( Figure S11), suggesting that the QTM cannot be effectively suppressed in the applied dc field, thus the optimum dc field cannot be achieved. To obtain the estimation of reversal barrier (Ueff) and characteristic relaxation time (τ0), the equation Bartolomé et al. [46,47] was employed. The best fitting afforded Ueff ≈ 8.4 K and τ0 ≈ 9.1 × 10 −8 s ( Figure   5b) which is in line with the expected characteristic relaxation time 10 −6 -10 −11 s for SMMs [48]. In Scheme 3. Spin polarization mechanism for the magnetic coupling mediated by m-phenylene ring between two mono-radicals.
For complexes 2 and 3, the found χ M T values at 300K is 24.37 and 28.81 cm 3 Kmol −1 (Figure 4), respectively, which is close to the theoretical values (24.39 cm 3 Kmol −1 for 2 and 29.09 cm 3 Kmol −1 for 3) reckoned from an uncoupled system consisting of two Ln III ions (Tb III : 7 F 6 , g = 3/2, C = 11.82 cm 3 Kmol −1 ; Dy III : 6 H 15/2 , g = 4/3, C = 14.17 cm 3 Kmol −1 ) and two mono-radicals (one radical: S = 1/2, g = 2, C = 0.375 cm 3 Kmol −1 ). For both complexes, the value of χ M T gradually declines when the temperature is lowered, reaching the minimum values of 22.23 cm 3 Kmol −1 at 9K for 2 and 27.22 cm 3 Kmol −1 at 24K for 3, which is most probably governed by the depopulation of the Tb III or Dy III ions Stark levels. Then upon further cooling, the χ M T value rapidly increases to the maxima of 22.91 cm 3 Kmol −1 at 3K for 2 and 30.20 cm 3 Kmol −1 at 2K for 3, which can be attributed to the ferromagnetic NO-Tb/Dy interaction. For 2, below 3K, the χ M T value decreases rapidly and reaches a value of 22.28 cm 3 Kmol −1 at 2 K. The M versus H curves in the 0-8 T field range at 2K for complexes 2 and 3 are depicted in Figures S7 and S8. The M value is 12.72 Nβ and 12.57 Nβ at 8T for 2 and 3, respectively, which is lower than expected saturation value, suggesting the presence of a significant magnetic anisotropy in the systems [42,43].
Due to the presence of magnetic anisotropic Tb III or Dy III ions in complexes 2 and 3, spin dynamic properties were studied using ac susceptibility measurements under zero dc field. As displayed in Figure S9 and Figure 5, there is no non-zero out-of-phase signal for complex 2 while complex 3 exhibits clear temperature-dependent χ" components, indicating slow relaxation of magnetization behavior in 3. However, no visible peaks of the χ" signals are found for complex 3, which can be attributed to the fast quantum tunneling of magnetization (QTM) [44,45]. Accordingly, frequency-dependent ac magnetic susceptibilities for 3 were investigated at 2 K under dc field range of 200-4900 Oe to determine the optimized external field for suppressing the QTM. As seen, no maxima of χ"(v) components can be observed in the measurement window of our apparatus ( Figure S11), suggesting that the QTM cannot be effectively suppressed in the applied dc field, thus the optimum dc field cannot be achieved. To obtain the estimation of reversal barrier (U eff ) and characteristic relaxation time (τ 0 ), the equation ln(χ /χ ) = ln(ωτ 0 ) + U e f f /k B T proposed by J. Bartolomé et al. [46,47] was employed. The best fitting afforded U eff ≈ 8.4 K and τ 0 ≈ 9.1 × 10 −8 s (Figure 5b) which is in line with the expected characteristic relaxation time 10 −6 -10 −11 s for SMMs [48]. In order to explain the dynamic magnetic behavior of complex 3, magnetic axis directions of Dy1 and Dy2 centers in the binuclear structure were analyzed using the Magellan program [49]. Easy axes on the two Dy III ions are obviously unparallel ( Figure 6); an angle of 87.65 • is formed between them. This angle is not conducive to the formation of large anisotropy in the system, which may be the reason for the poor magnetic relaxation behavior of 3 [50,51]. As seen, the previously reported tetranuclear [Dy 4 (hfac) 12 (NITPhO-3Pybis) 2 ] complex presents superior slow magnetic relaxation behavior than the present [Dy 2 (hfac) 6 (H 2 O)(NITPhPzbis)] complex, which should be attributed to the different coordination environments of Dy ions in the two compounds. The coordination geometries of Dy ions are ascertained as C 2v and D 2d in [Dy 2 (hfac) 6 (H 2 O)(NITPhPzbis)] while those as C 2v and D 4d in [Dy 4 (hfac) 12 (NITPhO-3Pybis) 2 ]. As known, D 4d symmetry is in favor of suppressing QTM effect, thus resulting in different magnetic relaxation [52]. order to explain the dynamic magnetic behavior of complex 3, magnetic axis directions of Dy1 and Dy2 centers in the binuclear structure were analyzed using the Magellan program [49]. Easy axes on the two Dy III ions are obviously unparallel ( Figure 6); an angle of 87.65° is formed between them. This angle is not conducive to the formation of large anisotropy in the system, which may be the reason for the poor magnetic relaxation behavior of 3 [50,51].As seen, the previously reported tetranuclear [Dy4(hfac)12(NITPhO-3Pybis)2] complex presents superior slow magnetic relaxation behavior than the present [Dy2(hfac)6(H2O)(NITPhPzbis)]complex, which should be attributed to the different coordination environments of Dy ions in the two compounds. The coordination geometries of Dy ions are ascertained as C2v and D2d in [Dy2(hfac)6(H2O)(NITPhPzbis)] while those as C2v and D4d in [Dy4(hfac)12(NITPhO-3Pybis)2]. As known, D4d symmetry is in favor of suppressing QTM effect, thus resulting in different magnetic relaxation. [52]

Materials and Characterizations
The nitronyl nitroxide biradical NITPhPzbis and Ln(hfac)3•2H2O were obtained according to literature methods [53][54][55][56]. The C, N, H content analyses for complexes 1-3 were completed on a Perkin-Elmer 240 elemental analyzer. A Bruker-Vector 22 Spectrometer was employed to collect the IR spectra of all complexes between 4000-400 cm −1 . The PAN alytical X'Pert Powder X-ray diffractometer(Cu/Kα radiation, 40 kV × 30 mA) was used to record the PXRD data of 1-3 at room temperature. Magnetic studies of all complexes were conducted on a Quantum Design SQUID VSM and a PPMS magnetometer. Diamagnetic contributions were estimated using Pascal's constants [57]. order to explain the dynamic magnetic behavior of complex 3, magnetic axis directions of Dy1 and Dy2 centers in the binuclear structure were analyzed using the Magellan program [49]. Easy axes on the two Dy III ions are obviously unparallel ( Figure 6); an angle of 87.65° is formed between them. This angle is not conducive to the formation of large anisotropy in the system, which may be the reason for the poor magnetic relaxation behavior of 3 [50,51].As seen, the previously reported tetranuclear [Dy4(hfac)12(NITPhO-3Pybis)2] complex presents superior slow magnetic relaxation behavior than the present [Dy2(hfac)6(H2O)(NITPhPzbis)]complex, which should be attributed to the different coordination environments of Dy ions in the two compounds. The coordination geometries of Dy ions are ascertained as C2v and D2d in [Dy2(hfac)6(H2O)(NITPhPzbis)] while those as C2v and D4d in [Dy4(hfac)12(NITPhO-3Pybis)2]. As known, D4d symmetry is in favor of suppressing QTM effect, thus resulting in different magnetic relaxation. [52]

Materials and Characterizations
The nitronyl nitroxide biradical NITPhPzbis and Ln(hfac)3•2H2O were obtained according to literature methods [53][54][55][56]. The C, N, H content analyses for complexes 1-3 were completed on a Perkin-Elmer 240 elemental analyzer. A Bruker-Vector 22 Spectrometer was employed to collect the IR spectra of all complexes between 4000-400 cm −1 . The PAN alytical X'Pert Powder X-ray diffractometer(Cu/Kα radiation, 40 kV × 30 mA) was used to record the PXRD data of 1-3 at room temperature. Magnetic studies of all complexes were conducted on a Quantum Design SQUID VSM and a PPMS magnetometer. Diamagnetic contributions were estimated using Pascal's constants [57].

Materials and Characterizations
The nitronyl nitroxide biradical NITPhPzbis and Ln(hfac) 3 ·2H 2 O were obtained according to literature methods [53][54][55][56]. The C, N, H content analyses for complexes 1-3 were completed on a Perkin-Elmer 240 elemental analyzer. A Bruker-Vector 22 Spectrometer was employed to collect the IR spectra of all complexes between 4000-400 cm −1 . The PAN alytical X'Pert Powder X-ray diffractometer(Cu/Kα radiation, 40 kV × 30 mA) was used to record the PXRD data of 1-3 at room temperature. Magnetic studies of all complexes were conducted on a Quantum Design SQUID VSM and a PPMS magnetometer. Diamagnetic contributions were estimated using Pascal's constants [57].

Preparation of [Ln 2 (hfac) 6 (H 2 O)(NITPhPzbis)]
The specific process was as follows: a solution of Ln(hfac) 3 ·2H 2 O (0.02 mmol) in n-heptane (20 mL) was refluxed with stirring for 6 h. Subsequently, 10 mL CH 2 Cl 2 solution containing 0.0045 g (0.01 mmol) NITPhPzbis ligand was slowly added. After refluxing for 25 min, the mixture was cooled and filtered, and the filtrate was left to evaporate at room temperature. About 5-6 days later, dark-purple block-like crystals were isolated.

X-ray Crystallography
Crystal diffraction data of 1-3 were collected on a Rigaku Saturn CCD X-ray diffractometer (Mo-Kα, λ = 0.71073 Å) with a scanning mode of ω-ϕ and a test temperature of 113(2) K. Direct methods were utilized to solve crystal structures of 1-3 using SHELXS-2014 and SHELXL-2014 [58,59]. All non-H atoms were refined anisotropically, and the H atom positions were given geometrically. To rationalize some disordered F and C atoms on the hfac − co-ligand, commands of ISOR, SIMU, and DELU were used.

Conclusions
In this contribution, three novel biradical Ln-based complexes have been successfully prepared by utilizing the 5-(1-pyrazolyl) phenyl modified nitronyl nitroxide biradical NITPhPzbis. Complexes 1-3 feature a four-spin binuclear structure in which the NITPhPzbis serves as a bridging and chelating ligand. Magnetic studies show that complex 3 exhibits obvious slow magnetic relaxation behavior. The nitronyl nitroxide radical-Ln complex has long been an attractive strategy for constructing molecular nanomagnets. But research on nitronyl nitroxide biradical-Ln chemistry is still limited. This work discloses that nitronyl nitroxide biradical is effective building block for constructing Ln-based complexes. In the follow-up work, we will try to develop novel molecular magnetic materials presenting intriguing topologies and magnetic properties by changing the functional group in biradical and/or introducing 3d spin.