Shruti
Moorthy
,
Ibtesham
Tarannum‡
,
Kusum
Kumari‡
and
Saurabh Kumar
Singh
*
Department of Chemistry, Indian Institute of Technology, Hyderabad, Kandi, Sangareddy, Telangana 502284, India. E-mail: sksingh@chy.iith.ac.in
First published on 15th May 2024
In the present work, we have explored a series of unsaturated hexa-18-crown-6 (U18C6) ligands towards designing highly anisotropic Dy(III) based single-ion magnets (SIMs) with the general formula [Dy(U18C6)X2]+ (where U18C6 = [C12H12O6] (1), [C12H12S6] (2), [C12H12Se6] (3), [C12H12O4S2] (4), [C12H12O4Se2] (5) and X = F, Cl, Br, I, OtBu and OSiPh3). By analysing the electronic structure, bonding and magnetic properties, we find that the U18C6 ligands prefer stabilising the highly symmetric eight-coordinated hexagonal bipyramidal geometry (HBPY-8), which is the source of the near-Ising type anisotropy in all the [Dy(U18C6)X2]+ complexes. Moreover, the ability of sulfur/selenium substituted U18C6 ligands to stabilize the highly anisotropic HBPY-8 geometry makes them more promising towards engineering the equatorial ligand field compared to substituted saturated 18C6 ligands where the exodentate arrangement of the S lone pairs results in low symmetry. Magnetic relaxation analysis predicts a record barrier height over 2700 K for [Dy(C12H12O6)F2]+ and [Dy(C12H12S6)X2]+ (where X = F, OtBu and OSiPh3) complexes, nearly 23% higher than those of the top performing Dy(III) based SIMs in the literature.
In general, stabilizing air-stable highly anisotropic Dy(III) complexes requires a large coordination number (CN > 7) and higher-order axial symmetry such as square antiprismatic/axially compressed octahedral (D4d/D4h), pentagonal bipyramidal (D5h) and hexagonal bipyramidal (D6h) environments with short Dy–Lax bonds.14–22,29–34 Some notable work includes the report of [Dy(OtBu)2(L)4]+ with a barrier height exceeding 2000 K in an axially compressed octahedral (D4h) environment,17 [Dy(OtBu)2(py)5]+ with a pentagonal bipyramidal (PBP) geometry (D5h)35 and R/S-[Dy(LN6)(OSiPh3)2] in a D6h environment showing barrier heights more than 1800 K.21 Except for a few, in most of these complexes, the equatorial positions are occupied by anionic polydentate ligands, offering an undesirable equatorial ligand field for oblate type Dy(III) ions.18,19,36–38 Alternatively, the neutral crown ether macrocyclic ligands are relatively weak-field ligands compared to anionic polydentate ligands and hence offer a way to engineer the equatorial ligand field by modulating the donor atoms and ring size (see Fig. 1). Numerous Dy(III) complexes with various 12-crown-4 ether, 15-crown-5 ether, and 18-crown-6 ether ligands are reported in the literature and among all these, the cavity of saturated 18-crown-6 ether (18C6) is found to be suitable for stabilizing highly anisotropic [Dy(18C6)X2]+ (where X is the axial ligand) complexes.39–43 Surprisingly, most of the reported [Dy(18C6)X2]+ complexes show only moderate barrier heights (∼100 cm−1) due to the stabilization of the low-symmetry environment around Dy(III) ions resulting from the high conformational flexibility (C–C bond rotations) in the 18C6 ligands except for a [Dy(OtBu)Cl(18C6)][BPh4] complex showing a barrier height of ∼1000 K.22,25,39–43 In contrast, the unsaturated hexa-18-crown-6 (U18C6) ligands with CC linkages offer conformation strain and prefer to stabilize the hexagonal bipyramidal environment around metal ions.44–46 Inspired by the exceptionally high coordinating capability of U18C6 towards the isolation of D6h geometry (preferable to generate an axial ligand field) and the suitable pore size of 18-crown-6 ligands to stabilize Ln(III) ions, here we investigated the electronic structure and magnetic properties in eleven Dy(III) complexes with the general formula [Dy(U18C6)X2]+ (where U18C6 = [C12H12O6] (1), [C12H12S6] (2), [C12H12Se6] (3), [C12H12O4S2] (4), [C12H12O4Se2] (5) and X = F, Cl, Br, I, OtBu and OSiPh3) (see Fig. 1) towards the search of new generation SIMs. All these complexes are named 1X–5X (where X = axial ligands, see Table S1†).
Conformational search analysis was carried out using xTB-CREST code to find all the conformers within an energy window of ∼10 kJ mol−1 (see the ESI† for details).53 All the obtained conformers were further optimised at the BP86 level of theory, and the lowest energy conformer was used for the bonding and magnetic property calculations. Single-point energy calculations were carried out using the hybrid PBE0 functional49 in ADF 2021 code to analyze the bonding interactions and energy decomposition analysis. Scalar relativistic effects were incorporated by using zeroth-order relativistic approximations (ZORA).54 The Slater-type all electron TZP basis set was used for the Dy atom and the DZP basis set for the remaining atoms, with “no frozen core” approximation (see the ESI† for details).
Complete-active space self-consistent field (CASSCF) calculations were performed on the DFT-optimized lowest energy structure to compute the magnetic properties.55 Here, we employed an all-electron SARC–DKH–TZVP basis set for the Dy(III) centre, the DKH-adapted version of def2–TZVP for O, S, Se, F, Cl, and Br atoms and Sapporo–DKH–TZP for the I atom.50,51,56 Using an active space of CAS(9,7), we computed 21 sextets and 224 quartets and performed the spin–orbit calculations using the spin–orbit mean field (SOMF-1X) operator. The computed spin-free energy and spin–orbit energies are provided in Tables S14–S19.†Ab initio ligand field theory (AILFT) calculations were performed at the CASSCF levels of theory to estimate the interelectronic repulsion in terms of Slater–Condon parameters and one electron energies to represent the f-orbital splitting.57 Next, we computed the effective demagnetisation barrier (Ueff) for the Orbach relaxation process as proposed earlier by Aravena et al.58 (see the ESI† for computational details).
Next, we computed the binding energy between the {DyX2}+ fragment and {U18C6} fragments in complexes 1x–5x using an energy decomposition analysis (EDA) approach within the scalar relativistic density functional theory (SR-DFT). SR-DFT calculations predict stabilizing interactions between the fragments for all the complexes, with oxa crown complexes relatively more stabilized than thia and selena crown complexes (see the ESI† for details). The decomposition of total interaction energy (ΔEint) suggests that the electrostatic (ΔEelstat) and orbital interactions (ΔEorb) are the most dominant contributions (50–80%) to the ΔEint (see Fig. S2†). Tables S4 and S5† show that as we move from oxa to thia/selena crown complexes, the ΔEelstat interaction decreases significantly due to a decrease in the electronegativity as we move down from O to S/Se atom (see Fig. 1C and Fig. S1† for computed MEPs). In contrast, the strength of the ΔEorb value marginally increases as we move from O to S/Se, which is attributed to weak lanthanide-ligand covalency (see Tables S6 and S7†). In addition, we have also compared the EDA analysis results of 1Cl with those of U18C6 and 18C6 ligands, which show relatively higher binding energy with 18C6 ligands compared to U18C6 ligands (see Tables S4 and S5†).
Next, we performed complete active space self-consistent field (CASSCF) calculations on the DFT-optimized geometry to compute the relevant spin Hamiltonian (SH) parameters (see the ESI† for computational details). Before discussing the analysis of SH parameters and SIM behaviour in complexes 1X–5X, we first compared the magnetic properties of Dy(III) complexes with those of U18C6 and 18C6 ligands. Here, we have analyzed the magnetic anisotropy of 1Cl with that of saturated and unsaturated crown ethers (see Fig. 2 and S3†). From a structural point of view, the [Dy(U18C6)Cl2]+ complex is relatively more symmetric and closer to the HBPY-8 geometry (CShM = 0.756) compared to the [Dy(18C6)Cl2]+ complex (CShM = 1.592). The large CShM value arises from a distorted equatorial plane formed by the saturated crown ether ligand. A close inspection of the structural parameters reveals that the axial Dy–Cl bonds are relatively shorter compared to Dy–O bonds in [Dy(U18C6)Cl2]+ (avg. Dy–Cl:avg. Dy–O = 1:1.04), while an opposite trend is observed for the [Dy(18C6)Cl2]+ complex (avg. Dy–Cl:avg. Dy–O = 1:0.96). This indicates that the strength of the equatorial ligand field is expected to be more significant in saturated crown ether than in unsaturated crown ether, which is also evident from EDA analysis. CASSCF calculations evidence these minor structural differences and predict the wide energy span of the low-lying eight KDs in [Dy(U18C6)Cl2]+ (1132 cm−1) compared to the [Dy(18C6)Cl2]+ complex (740 cm−1). The ground state g-values are highly axial for both the complexes, while low-symmetry around the Dy(III) ion in the [Dy(18C6)Cl2]+ complex results in a non-negligible transverse component (gxx and gyy) in the ground state g-values. The analysis of mJ states and g-tensor orientation predicts magnetic blocking up to the 4th excited Kramer doublet (KD) for the [Dy(U18C6)Cl2]+ complex, while the [Dy(18C6)Cl2]+ complex relaxes via the 3rd excited KD. Ab initio computed magnetic relaxation shows extremely weak QTM (1.61 × 10−6μB) values for the [Dy(U18C6)Cl2]+ complex compared to that of the [Dy(18C6)Cl2]+ complex showing a nearly 1000 times higher QTM value (2.52 × 10−3μB). The U18C6 ligand has an advantage over 18C6 ligands as they offer a weak equatorial ligand field and higher-order symmetry around the Dy(III) ion, which helps in the complete quenching of the QTM in the [Dy(U18C6)Cl2]+ complex. The computed Ucal value of 1036 cm−1 for the [Dy(U18C6)Cl2]+ complex is ∼2× times higher than that of the [Dy(18C6)Cl2]+ complex (535 cm−1), indicating a promising behaviour of the U18C6 ligand towards designing highly anisotropic Dy(III) based SMMs. From these exciting results, we moved to the next section and analyzed the magnetic anisotropy and magnetic relaxation mechanism in complexes 1X–5X.
Among all the complexes 1X–5X, we observed stabilization of mJ |±15/2〉 as the ground state, indicating the stabilization of the dominant axial ligand field. The computed ground state g-values (gzz ∼ 19.9 and gxx ∼ gyy ∼ 1 × 10−3) are highly axial and possess near-Ising type anisotropy for all the complexes. The computed gzz orientation of the ground state KD nearly passes through the X–Dy–X bond, indicating that the g-tensor orientation follows the highest order pseudo-C6 axis in all the complexes (see Fig. S5†). Among the studied 1X series, we observed that the first excited KD is 656.6, 381.3, 340.5 and 274.3 cm−1 above the ground state for 1F, 1Cl, 1Br and 1I, respectively (see Table S14†). The average axial to equatorial bond length ratios are 0.7 (1F), 0.9 (1Cl), 1.0 (1Br), and 1.1 (1I), indicating that lighter halides are structurally more favourable to generate an axial ligand field.59 In addition, we also observed an enormous negative charge on the F− ion compared to other halide ions (see Fig. S6†), which further leverages the axiality, resulting in the larger KD span for 1F. CASSCF computed ab initio ligand field theory (AILFT) predicts the following f-orbital manifold of 2815.3 (1F), 1626.9 (1Cl), 1315.3 (1Br), and 1035.2 (1I) cm−1 with giant f-orbital splitting for 1F, which corroborates the splitting of the KDs (see Fig. S7 and S8†). Calculations predict the magnetic relaxation via the 4th excited KD, i.e. mJ |±1/2〉 for all the complexes, setting the Ucal values of 1889, 1036.3, 844.6, and 644.8 cm−1 for 1F, 1Cl, 1Br and 1I, respectively (see Fig. S9†). Most importantly, the computed magnetic dipole matrix elements (kQTM) between ground state KDs are extremely weak (∼1 × 10−4μB), indicating quenching of quantum tunnelling of magnetization (QTM) in all the complexes. An extremely short Dy–F bond and sizeable negative charge on F− ions generate a more potent ligand field than heavier halides, resulting in a giant barrier height of 1889 cm−1 (2717 K) in 1F. Although complexes 1Br and 1I are not structurally favourable as the average axial/equatorial bond length ratio is ≥1, the isolation of the HBPY-8 geometry by neutral U18C6 ligands is the key for the highly anisotropic environment around the Dy(III) ion.
In order to fine-tune the magnetic anisotropy through ligand field modification, we have examined the complexes of sulfur and selenium-substituted crown ethers. For comparison, we have first examined the magnetic anisotropy of 2Cl and its saturated thia crown analogue. Unlike 2Cl where the endodentate arrangement of the S lone pairs forces the Dy(III) ion to sit in the middle of the ring, the exodentate nature of the S lone pair in saturated thia crown ligands throws the {DyCl2}+ fragment away from the center of the ring, stabilizing the low-symmetry environment (see Fig. 3). For complex 2Cl, we observed a near Ising-type ground state (gxx = gyy = 1 × 10−4 and gzz = 19.914) with magnetic blockade up to the 4th KD, resulting in a record Ucal of 1140 cm−1, a nearly 10% increase compared to that of 1Cl. On the other hand, a smaller Ucal value of 420.4 cm−1 is observed for the saturated analogue of 2Cl, which is nearly 3× times smaller than that of 2Cl. Moreover, the computed kQTM values are two orders of magnitude larger in the saturated analogue of 2Cl, suggesting strong QTM within the ground state. Comparative analysis indicates that ligand functionalization is not feasible in the saturated crown-ether based ligands as the exodentate arrangement of the S lone pair destroys the HBPY-8 arrangement, hence limiting the application of these ligands to engineer the equatorial ligand field.
Compared to the 1x series, we observed that complexes 2F–2I are highly symmetric, where six endodentate S atoms are arranged in a highly symmetric hexagonal manner. The computed CShM values are much smaller and closer to the ideal HBPY-8 geometry (see Table S2†). As a result, we observed a higher degree of axiality in the g-values and much smaller kQTM values compared to those of the 1X series (see Table 1). For 2X, we observed the following Ucal values: 1972.2 (2F), 1140.1 (2Cl), 934.6 (2Br), and 725.9 cm−1 (2I); these values are reasonably higher than those of the 1X series (see Table 1). The computed g-values and magnetic relaxation pattern for all 2F–2I are depicted in Fig. S9.† We noticed an increase of ∼100 cm−1 in the Ucal values as we moved from oxa crowns to thia crown complexes. Surprisingly, the increase in the barrier height is not colossal as the Dy-ligand covalency increases as we move from O to S, which counteracts the electrostatic effect, resulting in a modest increase in the Ucal value.
g zz | k QTM | U cal | U eff | T B | |
---|---|---|---|---|---|
1F | 19.908 | 3.04 × 10−5 | 1889.0 | 1710 | 67.5 |
1Cl | 19.989 | 1.61 × 10−6 | 1036.3 | 1087 | 37.0 |
1Br | 19.906 | 1.43 × 10−5 | 844.6 | 890 | 30.2 |
1I | 19.904 | 4.11 × 10−5 | 644.8 | 686 | 23.0 |
2F | 19.913 | 1.31 × 10−5 | 1972.2 | 1863 | 70.4 |
2Cl | 19.914 | 1.32 × 10−6 | 1140.1 | 1196 | 40.7 |
2Br | 19.913 | 8.96 × 10−7 | 934.6 | 1008 | 33.4 |
2I | 19.912 | 1.02 × 10−5 | 725.9 | 797 | 25.9 |
3Cl | 19.915 | 3.63 × 10−6 | 1133.3 | 1201 | 40.5 |
4Cl | 19.907 | 1.72 × 10−4 | 942.9 | 960 | 33.7 |
5Cl | 19.909 | 1.16 × 10−4 | 1043.0 | 1039 | 37.3 |
2OSiPh3 | 19.905 | 3.60 × 10−5 | 1903.3 | 1623 | 67.9 |
2OtBu | 19.913 | 2.86 × 10−6 | 2397.2 | 1886 | 85.6 |
Next, we investigated the selena crown complex (3Cl), where DFT predicts a substantially more bent ∠Cl–Dy–Cl bond angle than those for 2Cl and 1Cl, owing to the huge cavity provided by selena crowns (see Table S3†). As a result, we do not see any improvement in the g-values and magnetic relaxation pattern for 3Cl compared to those of 2Cl. Next, we studied magnetic anisotropy in 4Cl and 5Cl complexes arising from the oxa-thia (O4S2) and oxa-selena (O4Se2) ligands, and we observed stabilization of the mJ |±15/2〉 and reasonably large Ucal values of 942.9 cm−1 (1319 K) and 1043 cm−1 (1460 K) for complexes 4Cl and 5Cl, respectively. Among all the studied complexes, we observed that U18C6 prefers to stabilize the HBPY-8 environment around the Dy(III) ion, resulting in the magnetization blockade 4th KD and generating giant Ucal values in the range of ∼650–1900 cm−1 (∼900–2800 K) for these complexes. Next, we studied two model complexes of the best-performing 2X series, where the bulky OtBu and OSiPh3 ligands replace the halides at axial positions. The synthesis and isolation of these complexes often require a bulky group at the axial position to prevent multiple coordination and to maintain the axiality.32,61,62 DFT optimization predicts the stabilization of the HBPY-8 geometry around both complexes. Magnetic anisotropy and magnetic relaxation analyses indicate a giant barrier height of ∼2600 K and 3450 K for 2OSiPh3 and 2OtBu complexes, respectively, further highlighting the possibility of designing U18C6 ligand-based air-stable Dy(III) complexes (see Table S19 and Fig. S13†).
Next, we computed the effective demagnetization barrier (Ueff) using the ab initio computed KD's energy and transition dipole moments to analyze the Orbach magnetic relaxation pathways.58 For all the complexes, the temperature-dependent effective demagnetization barrier (Ueff) and relative contribution from each KD are provided in Table S12, Fig. S11 and S12,† and Fig. 4. A close inspection of the relaxation pattern suggests that at low temperatures (0–100 K), the Ueff values nearly remain zero, indicating all the population to be in the mJ |±15/2〉 ground state, which is blocked due to the minimal kQTM value (∼1 × 10−4μB). As the temperature increases beyond 100 K, the Ueff value increases by climbing to the higher excited KD with sequential absorption of the thermally available phonons, while the saturation in the Ueff plot at room temperature indicates rapid relaxation by spontaneously emitting phonons. For 2F, we noticed saturation in the Ueff with maximum contributions from the following three KDs : KD6 24% + KD5 18% + KD8 16% (see Table S12†). Most importantly, the Ueff values are nearly similar to the ab initio computed Ucal values for all the studied complexes, further confirming giant barrier height in these complexes (see Table 1 and Fig. 4). The computed Ueff value of ∼2700 K for 2F and 2OtBu is nearly ∼23% higher than the best reported Ucal value for the {DyCp2}+ family in the literature.
Footnotes |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4dt00632a |
‡ These authors contributed equally. |
This journal is © The Royal Society of Chemistry 2024 |