Shuang-Yan
Lin
,
Chunlai
Wang
* and
Zhikun
Xu
*
Key Laboratory for Photonic and Electronic Bandgap Materials, Ministry of Education, School of Physics and Electronic Engineering, Harbin Normal University, Harbin 150025, P. R. China. E-mail: xuzhikunnano@163.com
First published on 10th October 2016
Three dinuclear Ln(III) compounds, [Ln2(H3L)2(PhCOO)6] (Ln = Sm (1), Dy (2)) and [Dy2(H4L′)2(PhCOO)4]·2CH3CN (3) have been synthesized and structurally and magnetically characterized. Structure analyses revealed distinct structural features in compounds with different symmetrical arms of Schiff-base ligands. With the symmetrical two-arm Schiff-base ligand H3L, dinuclear Ln(III) compounds 1 and 2 were synthesized, where nine-coordinated Ln(III) ions were bridged by two syn–syn η1:η1-µ2-benzoate groups. When using symmetrical Schiff-base ligand H5L′ with four arms, the dinuclear Dy(III) compound 3 was prepared, in which eight-coordinated Dy ions were bridged by two alkoxido oxygen atoms from additional arms of H5L′. Direct-current (DC) magnetic susceptibility studies revealed that the Dy(III) ions were very weakly coupled in dysprosium compounds. Alternating-current (AC) magnetic susceptibility studies for compounds 2 and 3 indicated that field-induced slow relaxation phenomenon occured for both compounds. Furthermore, two relaxation phases under the optimal field appeared in compound 3, which are probably associated with the existence of the anisotropic Dy(III) ion for the slow relaxation phase (SR) and significant quantum tunneling for the fast relaxation phase (FR).
The structure of the selected ligand plays a crucial role in forming novel frameworks and tuning magnetic properties.3a,9 Based on the theory of hard and soft acids and bases, lanthanide ions tend to coordinate with the O-donor ligands. Thus, it is reasonable to design the ligands containing multiple O-donors such as hydroxyl, carbonyl or carboxyl groups for constructing diverse lanthanide compounds. Among numerous ligands, the salen type with rich oxygen and nitrogen donors can stabilize Dy(III) ions in various coordination environments, which exhibit unique molecular structure displaying distinct mononuclear to tetranuclear anisotropic centers.10 A similar but more donor type ligand, DFMP (2,6-diformyl-4-methylphenol)-based symmetrical Schiff-base ligand, presents a more versatile coordination ability in lanthanide compounds.11 Herein, we synthesized such a symmetrical DFMP-based ligand (H3L) with two arms, and designed a new symmetrical ligand (H5L′) containing four arms by introduction of propanediol groups (Scheme 1). Three dinuclear compounds, [Ln2(H3L)2(PhCOO)6] (Ln = Sm (1), Dy (2)) and [Dy2(H4L′)2(PhCOO)4]·2CH3CN have been obtained. Magnetic property studies indicated that the Dy(III) ions are weakly coupled. It is noteworthy that the Dy2 compounds (compounds 2 and 3) displayed field-induced slow magnetic relaxation.
Elemental analyses for C, H and N were carried out on a Perkin-Elmer 2400 analyzer. IR spectra (4000–300 cm−1) were measured using KBr pellets by a Nicolet 6700 Fourier transform infrared spectrometer. All magnetization data were recorded on a Quantum Design MPMS-XL7 SQUID magnetometer equipped with a 7 T magnet. The variable-temperature magnetization was measured with an external magnetic field of 1000 Oe in the temperature range of 2–300 K. The experimental magnetic susceptibility data are corrected for the diamagnetism estimated from Pascal's tables and sample holder calibration.
Compound | 1 | 2 | 3 |
Empirical formula | Sm2C68H66N4O18 | Dy2C68H66N4O18 | Dy2C66H76N6O18 |
FW (g mol−1) | 1527.95 | 1552.25 | 1566.33 |
Crystal system | Triclinic | Triclinic | Monoclinic |
Space group |
P![]() |
P![]() |
P21/n |
a (Å) | 11.1124(5) | 11.1487(7) | 13.988(2) |
b (Å) | 11.4404(5) | 11.3872(7) | 15.402(2) |
c (Å) | 13.6416(7) | 13.6878(8) | 15.741(2) |
α (o) | 94.291(1) | 94.287(1) | 90 |
β (o) | 112.910(1) | 113.068(1) | 98.823(2) |
γ (o) | 97.524(1) | 97.695(1) | 90 |
V (Å3) | 1568.60(13) | 1568.86(17) | 3351.1(8) |
Z, ρcalcd (mg m−3) | 1, 1.618 | 1, 1.643 | 2, 1.552 |
F(000), Rint | 770, 0.0244 | 778, 0.0167 | 1580, 0.0520 |
R 1, wR2 [I > 2σ(I)] | 0.0338, 0.0721 | 0.0248, 0.0642 | 0.0374, 0.0830 |
R 1, wR2 (all data) | 0.0428, 0.0765 | 0.0272, 0.0659 | 0.0573, 0.0933 |
GOF | 1.024 | 1.045 | 1.029 |
Single crystal X-ray diffraction analyses revealed that compounds 1 and 2 were crystallographically isostructural, with the same dinuclear [Ln2(H3L)2(PhCOO)6] (Ln = Sm (1), Dy (2)) core; hence, the structure of compound 2 was selected and described as a representative (Fig. 1). Compound 2 crystallized in the triclinic space group P with Z = 1. The dinuclear dysprosium compound was composed of two [Dy(H3L) (PhCOO)2] units with the distance of Dy⋯Dy 5.3316(3) Å that was bridged by two syn–syn η1:η1-µ2-benzoate groups. The ligand H3L serves as a tridentate ligand and chelates the Dy atom through one phenoxido oxygen atom, one imine nitrogen atom and one alkoxido oxygen. It was remarkable that the terminal ligands of H3L, instead of bridging benzoate ligands, prevented the formation of a one-dimensional chain and resulted in the isolation of a discrete Dy2 compound, which was similar to the discrete linear Dy4 complex.14 Two η2 chelating benzoate ligands completed the coordination sphere of the Dy ion, generating a nine-coordinate center. The exact geometry was determined using the SHAPE 2.1 software15 and the result indicated that the nine-coordinate Dy center was situated in a distorted coordination sphere between a tricapped trigonal prism (i.e., the basal planes of atoms O1, O9A, N1 and O5, O8, O7 (symmetry code: 1 − x, −y, 1 − z)) and capped square antiprism (i.e., the basal planes made up of atoms N1, O9A, O8, O7 and O1, O4, O5, O6) (Table S1† and Fig. 2). In addition, the distances of Dy–O were in the range of 2.2817(19)–2.583(2) Å and the Dy–N bond length was 2.504(2) Å with all of these being normal coordination bonds.
![]() | ||
Fig. 2 Two possible coordination polyhedra for Dy ions observed in compound 2: tricapped trigonal prism (a); capped square antiprism (b). |
For compound 1 (Fig. S2†), the distance of Sm⋯Sm was 5.3991(4) Å. The distance of the Sm–N bond length was 2.555(3) Å and that of Sm–O were in the range of 2.314(3)–2.616(3) Å.
Examination of the crystal packing revealed that both molecules of 1 and 2 were in contact through π–π interactions and the separation distances of ring centroids Cg⋯Cg were 3.6569(2) and 3.6667(2) Å for compounds 1 and 2, respectively. Therefore, infinite supramolecular chains were generated with the shortest intermolecular Ln⋯Ln distances of 10.2063(5) and 10.2332(5) Å for 1 and 2, respectively (Fig. 3 and S3†).
![]() | ||
Fig. 3 Crystal packing of compound 2 showing the formation of a supramolecular chain through π–π interactions. The distance Cg3⋯Cg4 is 3.6667(2) Å. |
To explore the effect of additional alkoxido hydroxyl groups on the lanthanide compound, the symmetrical Schiff-base H5L′ with four arms was synthesized by the condensation of DFMP and 2-amino-2-methyl-1,3-propanediol (1:
2) in methanol. Similar to the procedure for compounds 1 and 2, the reaction of H5L′ with dysprosium(III) benzoate in the presence of Et3N produced compound 3.
As shown in Fig. 4, compound 3 crystallized in the monoclinic space group P21/n with Z = 2. The dinuclear dysprosium compound was composed of two [Dy(H4L′) (PhCOO)2] units bridged by two alkoxido oxygens from two additional arms. In contrast to the H3L with two arms in compounds 1 and 2, the ligand (H4L′)− with four arms served as a tetradentate ligand and chelated the Dy atom through one phenoxido oxygen atom, one imine nitrogen atom and two alkoxido oxygens. The alkoxido oxygens (O5 and O5a) from additional arms bridged two Dy ions, leading to a nearly rhombic Dy2O2 core with a Dy–O–Dy angle of 108.45(13)° and a Dy⋯Dy distance of 3.6850(6) Å. Similar to compounds 1 and 2, only one set of side arms of the ligand were coordinated, whereas the other side arms remain uncoordinated, which may be further used to build a distinct complex. In addition, one bidentate (in η2 form) chelating benzoate ligand and one monodentate (in η1 form) benzoate ligand completed the coordination sphere of the Dy ion, generating an eight-coordinate center. The exact geometry was also determined using the SHAPE 2.1 software.15 However, the results indicated that the nine-coordinate Dy center was not situated in a traditional coordination sphere listed in SHAPE 2.1 (Table S1† and Fig. 5). The coordination sphere of the Dy center may be close to a hula-hoop-like geometry, but a distorted hula-hoop. Moreover, the lengths of the Dy–O bonds were in the range of 2.265(3)–2.514(4) Å, and that of the Dy–N bond was 2.496(4) Å.
![]() | ||
Fig. 6 Plots of χMT vs. T for compounds 1 (top), 2 and 3 (bottom) in a dc field of 1000 Oe (2–300 K). |
For compound 1, the χMT value observed at room temperature (300 K) was 0.767 cm3 kmol−1, which is significantly larger than the theoretical one (0.178 cm3 kmol−1). This is because the first or even higher excited states (6H7/2, 6H9/2, …, 6H15/2) of the Sm ion can certainly be populated at room temperature.16 As the temperature was lowered, the χMT values of compound 1 decreased rapidly to 0.062 cm3 kmol−1 at 2.0 K (Fig. 6).
For compounds 2 and 3, the χMT values at room temperature were 28.30 and 28.28 cm3 kmol−1, which are very close to the expected value of 28.34 cm3 kmol−1 for two isolated Dy(III) ions (6H15/2, g = 4/3). On cooling, the χMT values of 2 and 3 decreased slightly from 300 to 80 K and then further decreased by different degrees. For 2, the χMT values decreased slowly to a minimum of 23.85 cm3 kmol−1 at 2 K. Comparing the magnetic properties of compound 2 with a noninteracting complex Dy4,14 we found that the χMT value of compound 2 almost agreed well with the half χMT value of complex Dy4 at low temperatures (½χMT = 23.9 cm3 kmol−1 at 2 K), indicating insignificant exchange interactions between the Dy(III) ions in compound 2. In contrast, the χMT product for compound 3 decreased more steadily than that of compound 2, reaching a minimum of 19.93 cm3 kmol−1 at 2 K, which suggests the presence of weak antiferromagnetic interactions in compound 3. In general, the decrease of χMT values was associated with the progressive depopulation of the excited mJ sublevels of the Ln(III) ions and/or the weak antiferromagnetic interactions between the Ln(III) ions.17 Thus, the decrease of the χMT values of compound 2 was mainly attributed to the progressive depopulation of the Dy(III) ions, whereas that of compound 3 was attributed to both factors.
To the best of our knowledge, Dy2 complexes with dialkoxo and dicarboxylato bridging groups (in η1:η1-µ2 coordination mode) have rarely been reported with SMM behavior, and the former complexes usually showed interesting magnetic properties (Table S2†).
To probe the low-temperature magnetic relaxation behaviour, alternating current (AC) magnetic susceptibility measurements were performed for compounds 2 and 3. Under a zero DC field, both compounds 2 and 3 did not exhibit significant slow magnetic relaxation since it is hard to observe an appreciable out-of-phase signal (χ″) at frequencies of up to 997 Hz and at temperatures down to 1.9 K (Fig. S4†). This may be attributed to a fast zero-field quantum tunnelling of magnetizations (QTM), which obliterates the advantage of anisotropic lanthanide ions. This was commonly reported for lanthanide-containing compounds. The high axial coordination geometry around the Dy(III) ions has much influence on the suppression of zero-field QTM and plays a key role for the SMMs with a high barrier, such as the approximate D4d symmetry in [TbPc2]18 and the axial hula hoop-like geometry in asymmetric Dy2 compounds19 and the Dy6 triangular prism.20 On the contrary, the fast QTM observed in compounds 2 and 3 indicated an absence of the highly axial nature of the Dy(III) ions. It is likely that the severely distorted geometry of nine-coordinate Dy in 2 and eight-coordinate Dy in 3 (Fig. 2 and 5) might result in greater importance of the transverse anisotropy terms,21 thus largely affecting the dynamic behaviour of both the compounds.
On the other hand, the application of a DC field may probably remove the degeneracy of the ±MJ energy levels and suppress the QTM from the +MJ state to the −MJ state.22 Indeed, field dependent signals for both compounds 2 and 3 were clearly observed at 997 Hz and 2 K with a significant peak around the optimal DC field, indicating that field-induced slow magnetic relaxation was operating (Fig. 7). The optimal fields were 300 and 1300 Oe for compounds 2 and 3, respectively. Specifically, the ratios of the intensities of χ″/χ′ were about 1:
14.3(0.07) and 1
:
4.9(0.2) at the optimal DC field for compounds 2 and 3, respectively. The latter ratio value shows almost a 3 fold over the former. Thus, the AC susceptibility was measured with the application of the optimal field (1300 Oe) for compound 3 to further investigate the relaxation behaviour.
![]() | ||
Fig. 7 Dependence of the out of phase signal of compounds 2 and 3 on applied DC field strength at 2.0 K and 997 Hz. |
From the temperature-dependent AC susceptibility under 1300 Oe (Fig. 8 and S5†), temperature-dependent χ′ and χ″ signals were observed with good peak shapes at low temperature, indicating the presence of slow magnetic relaxation under the optimal field, typical of field-induced SMM behaviour. On the other hand, the frequent-dependent AC susceptibility showed two maximal values at low temperature (<3.0 K) and one maximal value at high temperature (>3.0 K), which corresponded to two and one relaxation phases, respectively. The two relaxation phases at low temperature contained the high-frequency signal (fast relaxation phase, FR) and the low-frequency region (slow relaxation phase, SR).
Cole–Cole diagrams under 1300 Oe (Fig. S6†) showed two arc shapes at a low temperature, which were also indicative of the occurrence of the two relaxation processes. It should be noted that multiple relaxation processes for most lanthanide systems are due to the presence of crystallographically independent lanthanide centers and/or to significant quantum tunnelling. Hence, the SR was attributed to the presence of a unique crystallographic Dy ion in the dinuclear structure. The FR was indicative of quantum tunnelling, as the high-frequency peaks were almost temperature independent (shown in the upper part of Fig. 8). In addition, attempts to fit the data by the sum of two modified Debye functions were unsuccessful due to the fact that high frequency and low frequency peaks of the χ″ signals were not observed within the available frequency range of the AC measurements.
![]() | ||
Fig. 8 Temperature dependence (upper) and frequency dependence (bottom) of the out-of phase AC susceptibility of compound 3 below 8.0 K, under a 1300 Oe DC field. |
The above analysis confirms the importance of a correct analysis of structure–property relationships, particularly the coordination polyhedra on the anisotropy nature of lanthanide ions, which may have steered our efforts to achieve high magnetic axiality by designing the local environment. Furthermore, using the same type Schiff-base ligands with differences in some groups may be an advantageous synthetic methodology to synthesize a comprehensive (quasi)isostructural series of complexes and further analyse the structure–property relationships.
Footnote |
† Electronic supplementary information (ESI) available. CCDC 1487903–1487905. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c6ra16669e |
This journal is © The Royal Society of Chemistry 2016 |