Six-Coordinate Ln(III) Complexes with Various Coordination Geometries Showing Distinct Magnetic Properties

Abstract

The syntheses, structural characterization, and magnetic properties of three lanthanide complexes with formulas [Ln(L¹)₃] (Ln = Dy (1Dy); Er (1Er)); and [Dy(L²)₂] (2Dy) were reported. Complexes 1Dy and 1Er are isostructural with the metal ion in distorted trigonal-prismatic coordination geometry, but exhibit distinct magnetic properties due to the different shapes of electron density for Dy^III (oblate) and Er^III (prolate) ions. Complex 1Dy shows obvious SMM behavior under a zero direct current (dc) field with an effective energy barrier of 31.4 K, while complex 1Er only features SMM behavior under a 400 Oe external field with an effective energy barrier of 23.96 K. In stark contrast, complex 2Dy with the octahedral geometry only exhibits the frequency dependence of alternating current (ac) susceptibility signals without χ″ peaks under a zero dc field.

1. Introduction

Single-molecule magnets (SMMs), with the individual molecules acting as tiny magnets, are the most appealing candidates to develop memory devices with ultra-high density and spintronic devices [1,2,3,4,5,6]. The realization of those potential applications firstly depends on the high magnetic blocking temperature (T_B) and large effective energy barrier (U_eff), which have attracted the interests of chemists, physicists, and theorists [7,8,9]. The effective energy barrier (also called the anisotropy barrier, ∆E) is the potential barrier hampering the reversal of the magnetization. For polynuclear transition metal SMMs, such as Mn₁₂ [10] and Mn₄ [11], the anisotropy barrier can be expressed as ∆E = DS² or D(S² − 1/4) for integer and half-integer spins, in which D is the axial zero-field splitting parameter and S is the spin of ground state. The magnetic blocking temperature (T_B) can refer to the highest temperature at which the M(H) hysteresis loop is observed. It is worth noting that T_B strongly depends on the field-sweep rate. Usually, a complex with a high anisotropy barrier will not have a guaranteed high blocking temperature, which may be attributed to the fast quantum tunneling of magnetizations. Up to now, the records of blocking temperature and effective energy barrier have been achieved by the complex {[(Cp^ttt)₂Dy][B(C₆F₅)₄]} (Cp^ttt = 1,2,4-tri(tertbutyl)cyclopentadienide) [12,13], as reported by the Layfield and Mills groups, with T_B = 60 K and U_eff = 1837 K, in which the value of U_eff is higher by more than a factor of 30 than that of the Mn₁₂ [10], the first SMM with U_eff = 61 K. The remarkable SMM properties of {[(Cp^ttt)₂Dy][B(C₆F₅)₄]} should mainly benefit from a perfectly axial crystal field realized by the bis(cyclopentadienyl) ligand, which demonstrates that the coordination environment of lanthanide ions plays a critical role in designing and modifying the SMMs, with the exception of complex N₂³⁻–Ln₂ [14,15], where strong lanthanide-radical magnetic exchange coupling hinders zero-field fast relaxation pathways, and the asymmetric Dy₂(ovph)₂ [16,17] with the Ising exchange interaction between Dy^III ions, which efficiently suppresses fast QTM [18].

Considering the various coordination geometries and numbers of lanthanide complexes, choosing the particular coordination geometry is of vital importance in order to obtain better SMMs. For Dy^III ions with the oblate-shaped electron density, a crystal field in which the ligand electron density is concentrated above and below the equatorial plane is desired to enhance the magnetic anisotropy [19]. Furthermore, this kind of crystal field could lead to a highly efficient dysprosium SIM (single-ion magnet), such as D_4d [20,21,22] or D_5h [23,24,25]. Moreover, the low-coordinate lanthanide complexes are superior at controlling the coordinated environment and understanding the magneto-structural relationship [26,27,28,29,30].

SIMs with six-coordination geometries are still rare in the previously reported Ln-based complexes, compared with other SIMs with high coordination numbers [31,32]. Recently, Gao and co-workers reported that the complex [(L^CO)Dy(N^*)₂] (L^COH = {N-[(2-MeO)–C₆H₅]}N = C(Me)CH = C(Me)N(H){N′-[(2-MeO)C₆H₅]} and HN* = HN(SiMe₃)₂) with trigonal-prismatic coordination geometry, exhibited a high energy barrier U_eff = 190 K under a zero dc field [33]. It is crucial to note that the magnetic axis is approximately collinear to the direction of [N*]⁻–N, resulted from the strong axial-ligand field improved by the short Dy–N([N*]⁻) bond lengths of 2.296 Å. However, complexes Dy(H₂BPz^Me₂)₃ (HPz = pyrazole) [34] and Dy(Bc^Me)₃ ([Bc^Me]⁻ = dihydrobis(methylimidazolyl)borate) [35] both exhibit only field- or dilution-induced slow relaxation of magnetization, in which cases the averaged bond distances of Dy–N and Dy–C are 2.477 and 2.577 Å, respectively. Those cases have demonstrated that the occurrence of SIMs’ properties not only depends on the coordination geometry, but also on the strong axial-ligand field. In contrast, the six-coordinate Ln-based complexes possessing octahedral geometries show no SIM properties under a zero dc field, such as {(H₂O)[Ln(NA)₂]·H₂O}_n (H₂NA = 5-hydroxynicotinic acid) [28] and [Yb(H₃L)₂]Cl·5CH₃OH·H₂O (H₃L = tri(((2-hydroxy-3-methoxybenzyl)amino)ethyl)amine) [36], which can be attributed to the fact that the cubic O_h symmetry does not have second-order uniaxial anisotropy parameter $B_2^0$ [37,38,39].

Herein, we report three six-coordinate lanthanide complexes, [Ln(L¹)₃] (Ln = Dy (1Dy); Er (1Er)); and [Dy(L²)₂] (2Dy), where HL¹ = 2-(((2,6-diisopropylphenyl)imino)methyl)-phenol and H₂L² = 6,6′-(2-(dimethylamino)ethylazanediyl)bis(methylene)bis(2,4-di-tert-butylphenol) (Scheme 1). Complexes 1Ln present the distorted trigonal-prismatic coordination geometry, while complex 2Dy shows distorted octahedral coordination geometry. The ac susceptibility data reveal the SIM behavior of 1Dy under a zero dc field, but the field-induced SIM behavior of 1Er. Without exception, complex 2Dy only shows the frequency dependence of ac signals without χ″ peaks under a zero dc field, as reported previously [28,40]. Therefore, this series of complexes sheds light on the magneto-structural correlation of six-coordinate complexes with different geometries. Importantly, complex 1Dy is the second SIM with trigonal-prismatic coordination geometry in the absence of an external field.

2. Results and Discussion

2.1. Crystallography

Single-crystal X-ray diffraction investigation revealed that complexes 1Dy, 1Er (Figure 1), and 2Dy (Figure 2) crystallize in the triclinic P$\overline{1}$ space group with Z = 2, in which complexes 1Dy and 1Er are isostructural. Herein, the crystal structure of 1Dy is described representatively. Details of the crystallographic data and the structure solution of three complexes are summarized in

Table 1. Crystallographic data and structure refinement details of complexes 1Dy, 1Er, and 2Dy.

	1Dy	1Er	2Dy
Formula	C57H66DyN3O3	C57H66ErN3O3	C68H109DyN4O4 c
FW, g·mol−1	1003.62	1008.38	1209.09 c
crystal system	Triclinic	Triclinic	Triclinic
space group	P$\overline{1}$	P$\overline{1}$	P$\overline{1}$
T, K	293(2)	293(2)	293(2)
λ, Å	0.71073	0.71073	0.71073
a, Å	10.8822(15)	10.9407(11)	13.7983(12)
b, Å	11.4229(15)	11.4509(11)	16.6655(15)
c, Å	20.823(3)	20.735(2)	18.6715(17)
α, °	89.735(3)	90.066(2)	70.7220(10)
β, °	88.681(3)	91.458(2)	77.406(2)
γ, °	76.170(2)	103.853(2)	86.857(2)
V, Å3	2512.7(6)	2521.3(4)	3954.7(6)
Z	2	2	2
reflns collected	16100	15710	23959
unique reflns	10012	9686	15601
Rint	0.0300	0.0605	0.0408
GOF on F2	1.036	0.996	1.080
R1 a, wR2 b (I ≥ 2σ(I))	0.0381, 0.0766	0.0584, 0.1014	0.0606, 0.1659
R1, wR2 (all data)	0.0476, 0.0822	0.0956, 0.1212	0.0878, 0.1857
CCDC number	1586924	1586925	1586926

^a R₁ = ∑||F_o| − |F_c||/∑|F_o|; ^b wR₂ = [∑w(F_o² − F_c²)²/∑w(F_o²)²]^1/2; ^c The formula and the formula weight of 2Dy do not include the squeezed solvents.

. Selected bond distances and angles are listed in Table S1 (see Supplementary Materials). The asymmetric unit of 1Dy contains one Dy^III ion with a [N₃O₃] coordination environment, which comes from three [L¹]⁻ ligands. The coordination geometry around the Dy^III ion is similar to the trigonal-prismatic geometry, which has been proven by the SHAPE 2.1 software [41,42,43], revealing that the Dy^III ion is located in a distorted trigonal prism with a deviation of 2.36 from the ideal D_3h symmetry (Table S2, Supplementary Materials). The up and down basal planes are constructed by atoms O1, O3, and N3 and O2, N1, and N2, respectively, in which the θ angle between the two planes is 13.12° (Figure S1, Supplementary Materials). The bond distances of Dy–O and Dy–N are in the range of 2.144(3)–2.167(3) Å and 2.446(3)–2.606(3) Å, respectively. The angles of O1–Dy–O2 and O2–Dy–O3 are 142.05° and 132.32°, respectively. The packing arrangement along the c axis (Figure S2, Supplementary Materials) demonstrates that the shortest Dy···Dy distance is 10.88 Å. For complex 1Er, the bond distances of Er–O and Er–N are in the range of 2.123(5)–2.146(5) Å and 2.411(4)–2.567(5) Å, respectively, which are shorter than those of 1Dy. The angles of O1–Er–O2 and O2–Er–O3 are 140.62° and 132.89°, respectively. The θ angle between two planes is 11.56° (Figure S3, Supplementary Materials), and the shortest Er···Er distance is 10.94 Å (Figure S4, Supplementary Materials).

The Dy^III ion of complex 2Dy locates in a [N₄O₂] coordination environment, in which O1 and O2 come from one H₂L² ligand, and O3, O4, N3, and N4 come from a second H₂L² ligand. The six-coordinate Dy^III ion is in a distorted octahedral arrangement with a deviation of 1.70 from the ideal O_h symmetry (Table S2, Supplementary Materials). Two phenol O atoms (O3 and O4) are axially coordinated to Dy^III with a Dy–O3 distance of 2.275(3) and a Dy–O4 distance of 2.170(3) Å, and O3–Dy–O4 bond angle of 155.23°. The other bond distances of Dy–O are 2.182(3) and 2.208(3) Å, which are longer than those of Dy–O4. The bond distances of Dy–N are 2.515(4) and 2.662(4) Å. Furthermore, the shortest distance of the neighboring Dy^III ions is 10.36 Å (Figure S5, Supplementary Materials).

2.2. Magnetic Properties

The variable-temperature magnetic susceptibility data of complexes 1Dy, 1Er, and 2Dy were collected on the polycrystalline samples under an applied magnetic field of 1 kOe. At room temperature, the χ_MT values (Figure 3) of 1Dy, 1Er, and 2Dy are 13.77, 11.38, and 13.21 cm³·K·mol⁻¹, respectively, which are slightly lower than the expected value for free Dy^III (14.17 cm³·K·mol⁻¹) and Er^III (11.48 cm³·K·mol⁻¹) ions. As the temperature decreased, the χ_MT products decrease slowly down to 2 K, reaching values of 11.60 and 11.15 cm³·K·mol⁻¹ for 1Dy and 2Dy, respectively. For complex 1Er, upon cooling, the χ_MT products slightly decrease over the range of 300–100 K, followed by an obvious decrease until 2 K with a minimum of 7.99 cm³·K·mol⁻¹. The decrease of χ_MT values can be attributed to the Stark level splitting of lanthanide ions with large unquenched orbital moment. The field-dependent magnetizations for complexes 1Dy and 2Dy show the same tendency (Figures S6 and S8, Supplementary Materials), as observed in most dysprosium complexes reported [44]. The magnetizations increase rapidly up to a field of 10 kOe, followed by an almost constant rise to 70 kOe, reaching values of 5.57 and 4.91 μ_B at 1.9 K for 1Dy and 2Dy, respectively. However, a residual slope for complex 1Er (Figure S7, Supplementary Materials) is observed even at high field, and the magnetization finally reaches a value of 5.07 μ_B at 1.9 K. The non-saturation of the field-dependent magnetizations at high field (70 kOe) for three complexes reveal the presence of magnetic anisotropy caused by the crystal-field effects and/or low-lying excited states.

Alternating current susceptibility measurements were also conducted for three complexes under zero and 400 Oe dc field to further probe the dynamics of magnetization. In the absence of an applied dc field, both in-phase (χ′) and out-of-phase (χ″) ac susceptibilities for complex 1Dy exhibit frequency (Figure 4) and temperature (Figure S9, Supplementary Materials) dependency. However, no maximum peaks of temperature dependence of the out-of-phase (χ″) signal are observed in the range of 1–1488 Hz, which may be caused by the quantum tunneling of the magnetization (QTM), as also indicated by strong temperature-independent peaks below 9 K showed in Figure 4. To evaluate the effective barrier of magnetic relaxation, the relaxation times (τ) were extracted from the plot χ″ versus υ using the Debye model [45]. The τ versus T⁻¹ plot (Figure 5) shows a crossover from a linear increase of thermally activated to a temperature independent regime of QTM, which suggests the presence of more than one relaxation pathway. The plot was fitted using Equation (1), yielding effective energy barriers U_eff of 31.40 K with a τ₀ = 3.56 × 10⁻⁴ s, where the τ⁻¹_QTM, AH²T, CTⁿ, and ${\mathsf{\tau }}_0^{-1}$exp(−U_eff/k_BT) represent quantum tunneling, direct, Raman, and Orbach relaxation processes, respectively. For complex 1Dy, the direct process is excluded since the corresponding contribution is nullified at zero dc field. Other parameters obtained from the fitting are given in Table S3 (see Supplementary Materials).

$${{\mathsf{\tau }}^{-1}}_{obs}={{\mathsf{\tau }}^{-1}}_{QTM}+A{H}^{2}T+C{T}^n+{\mathsf{\tau }}_0^{-1}\exp \left(\raisebox{1ex}{$-U_{eff}$}\!\left/ \!\raisebox{-1ex}{$k_{B}T$}\right.\right)$$

(1)

For complex 1Er, no out-of-phase (χ″) signals (Figure S10, Supplementary Materials) were observed above 1.9 K at 997 Hz, which may be attributed to the fast quantum tunneling of the magnetization at zero dc field. The rather different magnetic behaviors of 1Dy and 1Er are correlated with the axial ligand field of trigonal-prismatic coordination geometry, as Dy^III is oblate and Er^III is prolate [19]. In order to suppress the QTM process, the ac magnetic susceptibility measurements were also performed under a dc field (Figure S11, Supplementary Materials). The non-zero frequency- and temperature-dependent χ′ and χ″ signals (Figure 4 and Figure S12, Supplementary Materials) were observed at low temperature, indicating the field-induced SMM behavior. The relaxation times (τ) of 1Er were extracted from the plot χ″ versus υ using the Debye model. The τ versus T⁻¹ plot (Figure 5), showing a smooth increase as the temperature was lowered, corroborates that the QTM is suppressed to a certain extent. The plot was fitted using Equation (1), yielding effective energy barriers U_eff of 23.96 K with a τ₀ = 5.46 × 10⁻⁸ s, and other parameters are listed in Table S3 (see Supplementary Materials). To avoid overparametrization, the direct process is canceled.

The Cole-Cole plots (Figure 6) of 1Dy and 1Er both show an asymmetrical semicircular shape, which can be fitted by the generalized Debye model [45], giving a series of α parameters below 0.11 from 1.9 to 13 K for 1Dy and 0.13 from 1.9 to 3.7 K for 1Er, respectively, which indicates a narrow distribution of the relaxation time for both complexes.

For complex 2Dy, both in-phase (χ′) and out-of-phase (χ″) ac susceptibilities exhibited frequency (Figure 4) and temperature (Figure S13, Supplementary Materials) dependency under a zero dc field, indicative of slow relaxation of magnetization. However, no peaks of χ″ were observed, indicating the presence of fast QTM relaxation.

Compared with 1Ln with trigonal-prismatic geometry, the six-coordinated 2Dy with octahedral geometry demonstrates inferior magnetic properties, verifying that the coordination geometry around lanthanide ions directly affects the SMM performance. The closer the distribution of the ligands to spherical symmetry, such as the environment of cubic symmetry (octahedron, etc.), the smaller the crystal-field (CF) splitting [46]. Therefore, the trigonal prismatic geometry can improve a relatively axial ligand field compared with the octahedron, which means that the six-coordinated lanthanide-based complexes, especially dysprosium complexes, located in trigonal-prismatic geometry are more likely to show SMM behavior in principle, which coincides with our experimental results. In order to further explore the magnetic properties of these two six-coordinate Dy-based complexes, the Magellan program [47] was used to calculate the magnetic anisotropy axes of complex 1Dy (Figure 7). The results reveal that the orientation of the magnetic axis of complex 1Dy is found to be almost collinear to Dy–O2 with an angle of 3.707°. Apparently, the negative charges on the O atoms are much larger than those on the N atoms for 1Dy, where it is more capable of stabilizing the ground doublet. As the ground-state wave function of complex [(L^CO)Dy(N*)₂] [33] featuring similar trigonal-prismatic coordination geometry around Dy^III ion with complex 1Dy shows a dominant M_J = ±15/2 doublet, the magnetic axis orientation might prefer the negative charge dense direction for the Dy^III ion with the oblate shaped electron density of <±15/2> doublet. Herein, the calculation was based on an electrostatic model, and a further quantitative evaluation of the anisotropy axis through ab initio calculation is definitely needed.

3. Materials and Methods

All manipulations of air- and moisture-sensitive complexes were performed under a nitrogen atmosphere using standard Schlenk techniques and a glovebox. THF (tetrahydrofuran) and toluene were distilled under nitrogen over sodium and sodium benzophenone. Pentane and hexane were distilled under nitrogen over CaH₂ (calcium hydride). Ln(N(SiMe₃)₂)₃ [48], ligand HL¹ [49], and ligand H₂L² [50,51] were synthesized according to previously published procedures under ambient conditions. All other starting materials were commercially available and used without further purification. FTIR spectra were obtained using a Nicolet 6700 Flex FTIR spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) equipped with smart iTR^TM attenuated total reflectance (ART) sampling accessory in the range from 500 to 4000 cm⁻¹. Elemental analysis for C, N, and H was carried out via a Perkin-Elmer 2400 analyzer (Perkin Elmer, Waltham, MA, USA).

X-ray crystal structure determinations. Single-crystal X-ray data of the title complexes were collected using a Bruker Apex II CCD diffractometer (Bruker, Billerica, MA, USA) equipped with graphite-monochromatized Mo Kα radiation (λ = 0.71073 Å). Data processing was completed using the SAINT processing program (Bruker, Billerica, MA, USA). The structures were solved by direct methods and refined by full-matrix least-squares methods on F² using SHELXTL-2014 [52,53,54]. All non-hydrogen atoms were determined from the difference Fourier maps and anisotropically refined. Hydrogen atoms were introduced at the calculated positions and refined with fixed geometry with respect to their carrier atoms. Further details may be obtained from the Cambridge Crystallographic Data Centre on quoting the depository numbers 1586924–1586926 (http://www.ccdc.cam.ac.uk).

Magnetic susceptibility measurements were recorded on a Quantum Design MPMS-XL7 SQUID magnetometer (Quantum Design, San Diego, CA, USA) equipped with a 7 T magnet. The variable-temperature magnetization was measured in the temperature range of 1.9–300 K with an external magnetic field of 1000 Oe. The dynamics of the magnetization were investigated in a 3.0 Oe ac oscillating field at different frequencies ranging from 1 to 1500 Hz. Diamagnetic corrections for the complexes were made with the Pascal’s constants [55] for all the constituent atoms as well as the contributions of the sample holder.

Synthesis of [Dy(L¹)₃] (1Dy). Dy(N(SiMe₃)₂)₃ (0.1 mmol) in 5 mL of toluene was added to a solution of HL¹ (0.3 mmol) in 10 mL toluene at room temperature. The solution was stirred at room temperature for 6 h, and then filtered. After removal of toluene under reduced pressure, recrystallization of the residue in hexane at ambient temperature gave 1Dy as yellow crystals after several days. Yield: ~70%. Selected IR (cm⁻¹): 2960 (s), 2926 (m), 1603 (s), 1583 (s), 1536 (s), 1464 (m), 1443 (s), 1381 (m), 1359 (w), 1254 (w), 1200 (w), 1167 (m), 1143 (m), 1106 (w), 921 (m), 851 (m), 792 (m), 750 (m), 739 (m), 593 (w). Anal. Calcd. for [Dy(L¹)₃] (C₅₇H₆₆DyN₃O₃, MW = 1003.62): C, 68.15%; H, 6.57%; N, 4.18%. Found: C, 68.21%; H, 6.67%; N, 4.26%.

Synthesis of [Er(L¹)₃] (1Er). 1Er were synthesized using a procedure similar to that for 1Dy with the replacement of Dy(N(SiMe₃)₂)₃ by Er(N(SiMe₃)₂)₃. Yield: ~70%. Selected IR (cm⁻¹): 2959 (s), 2926 (m), 1603 (s), 1583 (s), 1536 (s), 1464 (m), 1444 (s), 1381 (m), 1360 (w), 1342 (s), 1320 (m), 1254 (w), 1200 (w), 1167 (m), 1143 (m), 1106 (w), 921 (m), 851 (m), 792 (m), 750 (m), 739 (m), 594 (w). Anal. Calcd. for [Er(L¹)₃] (C₅₇H₆₆ErN₃O₃, MW = 1008.38): C, 67.83%; H, 6.55%; N, 4.16%. Found: C, 67.72%; H, 6.79%; N, 4.08%.

Synthesis of [Dy(L²)₂] (2Dy). A solution of Dy(N(SiMe₃)₂)₃ (0.5 mmol) in pentane was added dropwise to a solution of H₂L² (1 mmol) in pentane at −78 °C. The solution was stirred overnight at room temperature and then filtered. The final filtrate was left unperturbed at room temperature, X-ray quality crystals of 2Dy were obtained after few days. Yield: ~60%. Anal. Calcd. for [Dy(L²)₂] (C₆₈H₁₀₉DyN₄O₄, MW = 1209.09): C, 67.48%; H, 9.02%; N, 4.63%. Found: C, 67.52%; H, 9.09%; N, 4.59%.

4. Conclusions

In summary, we have synthesized and characterized three Ln-based mononuclear complexes, 1Dy, 1Er, and 2Dy, which show distinct magnetic properties. In complexes 1Dy and 1Er, the trigonal-prismatic coordination geometry provides an axial ligand field, which is in favor of the oblate Dy^III ion rather than the prolate Er^III ion; therefore, 1Dy exhibits the better magnetic properties with U_eff = 31.4 K under a zero dc field. 1Er shows field-induced SIM properties with U_eff = 23.96 K under a 400 Oe dc field. Complex 2Dy with the O_h symmetry only displays the frequency dependence of ac signals without χ″ peaks under the zero dc field, indicating the presence of fast QTM relaxation. For six-coordinate dysprosium complexes, the trigonal-prismatic coordination geometry is much more favorable to designing effective SIMs.