Incorporation of expanded organic cations in dysprosium(III) borohydrides for achieving luminescent molecular nanomagnets

Luminescent single-molecule magnets (SMMs) constitute a class of molecular materials offering optical insight into magnetic anisotropy, magnetic switching of emission, and magnetic luminescent thermometry. They are accessible using lanthanide(III) complexes with advanced organic ligands or metalloligands. We present a simple route to luminescent SMMs realized by the insertion of well-known organic cations, tetrabutylammonium and tetraphenylphosphonium, into dysprosium(III) borohydrides, the representatives of metal borohydrides investigated due to their hydrogen storage properties. We report two novel compounds, [n-Bu4N][DyIII(BH4)4] (1) and [Ph4P][DyIII(BH4)4] (2), involving DyIII centers surrounded by four pseudo-tetrahedrally arranged BH4– ions. While 2 has higher symmetry and adopts a tetragonal unit cell (I41/a), 1 crystallizes in a less symmetric monoclinic unit cell (P21/c). They exhibit yellow room-temperature photoluminescence related to the f–f electronic transitions. Moreover, they reveal DyIII-centered magnetic anisotropy generated by the distorted arrangement of four borohydride anions. It leads to field-induced slow magnetic relaxation, well-observed for the magnetically diluted samples, [n-Bu4N][YIII0.9DyIII0.1(BH4)4] (1@Y) and [Ph4P][YIII0.9DyIII0.1(BH4)4] (2@Y). 1@Y exhibits an Orbach-type relaxation with an energy barrier of 26.4(5) K while only the onset of SMM features was found in 2@Y. The more pronounced single-ion anisotropy of DyIII complexes of 1 was confirmed by the results of the ab initio calculations performed for both 1–2 and the highly symmetrical inorganic DyIII borohydrides, α/β-Dy(BH4)3, 3 and 4. The magneto-luminescent character was achieved by the implementation of large organic cations that lower the symmetry of DyIII centers inducing single-ion anisotropy and separate them in the crystal lattice enabling the emission property. These findings are supported by the comparison with 3 and 4, crystalizing in cubic unit cells, which are not emissive and do not exhibit SMM behavior.


10.
Best-fit parameters for various possible fittings of the field-and temperature-dependent magnetic relaxation times in 1@Y. (Table S2)  2. Alternate-current (ac) magnetic properties of 1, including dc-field-variable characteristics at 1.8 K, and temperature-variable characteristics at 1 kOe. ( Figure S4) Figure S4. Alternate-current (ac) magnetic properties of 1: dc-field-variable frequency dependences of the out-ofphase susceptibility, χ M " (a) and the in-phase susceptibility, χ M ' (b) collected in the 0-5 kOe range at T = 1.8 K, temperature-variable frequency dependences of the χ M " (c) and χ M ' (d) gathered in the 1.8-3.0 K range at H dc = 1 kOe, and the temperature dependence of the related ln(χ M "/χ M ') for the indicated frequencies of ac field (e). The solid lines in (a-d) are only to guide the eye. In (e), the black points represent the experimental data while the solid lines are the best curves for a simplified analysis approach elucidating two different Arrhenius-type relaxation routes operating at lower (blue lines) and higher (red lines) temperatures. The best-fit parameters are gathered in Table S1.

S6
3. Temperature-variable ac magnetic characteristics for 1 at H dc = 2.5 kOe. ( Figure S5) Figure S5. Temperature-variable ac magnetic characteristics for 1 at H dc = 2.5 kOe: frequency dependences of the out-of-phase susceptibility, χ M " (a) and the in-phase susceptibility, χ M ' (b) at the indicated temperatures, together with the related Argand plots (c), and the resulting temperature dependence of the relaxation time (d). The empty points in (a-c) represent the experimental data while the respective solid lines show the fitting according to the generalized Debye model. The black points in (d) represent the experimental data while the solid lines are the best fit-curves for two Arrhenius-type relaxation processes operating in the higher (red line) and lower (blue line) temperature regimes. The best-fit parameters are gathered in Table S1. together with the related temperature-dependence of the relaxation time, Figure S7). The best-fit parameters are gathered in Table S1. S8 5. Full temperature-dependent (ac) magnetic characteristics of 1@Y at H dc = 1 kOe. ( Figure S7) together with the related field-dependence of the relaxation time, Figure S6). The alternative fittings for the temperature-dependence of the relaxation time is presented in Figure S8. The best-fit parameters are gathered in Table S1.

S9
6. The comparison of the alternative fittings of the temperature dependence of magnetic relaxation time detected for 1@Y at the optimal dc field of 1 kOe. ( Figure S8) Figure S8. The comparison of the alternative fittings of the temperature dependence of magnetic relaxation time detected for 1@Y at the optimal dc field of 1 kOe: the experimental points (black circles) and the best-fit curves (coloured lines) obtained taking into account only Raman relaxation (a), the combination of Raman, direct, and QTM processes (b), only Raman relaxation with the fixed power n of 9, as expected for the Kramers ions (c), the combination of direct, QTM, and Raman relaxation with the fixed power n of 9 (d), and only Orbach relaxation (Arrhenius law) (e). The best-fit parameters are gathered in Table S2. S10 7. Alternate-current (ac) magnetic properties of 2, including dc-field-variable characteristics at 1.8 K, and temperature-variable characteristics at 1 kOe. ( Figure S9) Figure S9. Alternate-current (ac) magnetic properties of 2: dc-field-variable frequency dependences of the out-ofphase susceptibility, χ M " (a) and the in-phase susceptibility, χ M ' (b) collected in the 0-5 kOe range at T = 1.8 K, temperature-variable frequency dependences of the χ M " (c) and χ M ' (d) gathered in the 1.8-4.0 K range at H dc = 1 kOe. The solid lines in (a-d) are only to guide the eye. The experimental curves were not fitted due to the lack of χ M " maxima in the investigated frequency range. S11 8. Full alternate-current (ac) magnetic properties of 2@Y, including dc-fieldvariable characteristics at 1.8 K, and temperature-variable characteristics at 1 kOe. ( Figure S10) Figure S10. Full alternate-current (ac) magnetic properties of 2@Y: dc-field-variable frequency dependences of the out-of-phase susceptibility, χ M " (a) and the in-phase susceptibility, χ M ' (b) collected in the 0-5 kOe range at T = 1.8 K, temperature-variable frequency dependences of the χ M " (c) and χ M ' (d) gathered in the 1.8-2.2 K range at H dc = 1 kOe. The solid lines in (a-d) are only to guide the eye. The experimental curves were not fitted due to the lack of χ M " maxima in the investigated frequency range, only the simplified approach utilizing the temperature dependence of ln(χ M "/χ M ') was applied which is presented in Figure 2. S12 9. Best-fit parameters for the fittings of temperature-and/or dc-field-variable magnetic relaxation times for 1, 2, 1@Y, and 2@Y. (Table S1) Table S1. Best-fit parameters for the fittings of temperature-and/or dc-field-variable magnetic relaxation times for 1, 2, 1@Y, and 2@Y. compound type of fitting a simplified approach for the determination of the Arrhenius-type relaxation from the linear course of the ln(χ M "/χ M ') = f(T -1 ) plot S1-S2 the equation (1)    10. Best-fit parameters for various possible fittings of the field-and temperature-dependent magnetic relaxation times in 1@Y. (Table S2)   Table S2. Best-fit parameters for various possible fittings of the field-and temperature-dependent magnetic relaxation times in 1@Y ( Figure S8). Note that the parameters for the final fitting using Orbach, Direct, and QTM contributions were obtained within the simultaneous fit of both field-and temperature-dependences of relaxation time while for other alternative fittings, the parameters of Direct and QTM processes (if applied) were taken from that simultaneous fit, and used as the fixed values for the investigation of other possible relaxation routes.

Computational details
The ab initio calculations of CASSCF/RASSI/SINGLE_ANISO S3 type were carried out for 1, 2, 3, and 4, using the OpenMolcas quantum chemistry software package S4 . They were performed on the experimental geometries taken from powder X-ray diffraction analysis without optimization. Molecular clusters consisting of Dy 3+ central ion surrounded with BH 4 units were considered. The fragments of the crystal structures employed for analyses are presented in Figure S11. Three models with different basis sets were used: S -small with VDZP basis function quality, Llarge with VTZP basis, and Vvery large employing VQZP functions. Tables S3 and S8 contain contractions and labels of the basis sets for all the atoms. Scalar relativistic effects were taken into account by employing two-component second-order Douglas-Kroll-Hess (DKH2) Hamiltonian together with relativistic Atomic Natural Orbital basis sets -ANO-RCC type. S5-S6 In order to save disk space, the Cholesky decomposition of ERI-s (electron repulsion integrals) was used with the 1.0•10 -8 threshold. In the first step of the employed procedure, a State Average Multi-Configurational Self-Consistent Field (SA-CASSCF) calculation for 21 sextets, 224 quartets, and 490 doublets rising from different possible electron distributions for 4f 9 configuration was performed. The active space was composed out of 7 f-orbitals with 9 active valence electrons -CAS(9in7). In the next step, all sextets, 128 quartets, and 130 doublets optimized as spin-free states in the CASSCF step were mixed by the spin-orbit coupling within RASSI (Restricted Active Space State Interaction Program) S7 using meanfield spin-orbit (SO) integrals (AMFI) S8 resulting in 898 spin-orbit states In the final step, a SINGLE_ANISO S9 module was used to decompose spin-orbit states into states with a definite projection of the total momentum on the located quantization axis and to extract three components of the pseudo-g-tensor for eight ground Kramers doublets. The obtained energy splitting of the J = 15/2 manifold, together with the g x , g y , g z components of the pseudo-g-tensors within the basis of each doublet (̃= 1/2) and decomposition of the ground state into states with definite angular momentum on the quantization axis are presented in Tables S4-S7 (Table S4). The other compounds do not show an easy-axis type of magnetic anisotropy, thus the magnetic axes were not shown. Table S3. Description and contractions of the basis sets (two models: S -small, L -large) employed in the ab initio calculations of the Dy III crystal field in 1, 2, 3, and 4. Table S4. Summary of the energy splitting of the 6 H 15/2 multiplet of Dy III in 1 using models S and L with pseudo-gtensors of each Kramers doublet and the composition in the | J 〉 basis of the ground state.  3S2P1D  Table S5. Summary of the energy splitting of the 6 H 15/2 multiplet of Dy III in 2 using models S and L with pseudo-gtensors of each Kramers doublet and the composition in the | J 〉 basis of the ground state.

S L
Energy and pseudo-g-tensor components (g x , g y , g z ) of 8 ground Kramers doublets Energy / cm -1 Pseudo-g-tensor components Energy / cm -1 Pseudo-g-tensor components g x g y g z g x g y g z 0.000 9.  Table S6. Summary of the energy splitting of the 6 H 15/2 multiplet of Dy III in 3 using models S and L with pseudo-gtensors of each Kramers doublet and the composition in the | J 〉 basis of the ground state.

S L
Energy and pseudo-g-tensor components (g x , g y , g z ) of 8 ground Kramers doublets Energy / cm -1 Pseudo-g-tensor components Energy / cm -1 Pseudo-g-tensor components g x g y g z g x g y g z 0.000  Table S7. Summary of the energy splitting of the 6 H 15/2 multiplet of Dy III in 4 using models S and L with pseudo-gtensors of each Kramers doublet and the composition in the | J 〉 basis of the ground state.

S L
Energy and pseudo-g-tensor components (g x , g y , g z ) of 8 ground Kramers doublets Energy / cm -1 Pseudo-g-tensor components Energy / cm -1 Pseudo-g-tensor components g x g y g z g x g y g z 0.000  Table S9. Summary of the energy splitting of the 6 H 15/2 multiplet of Dy III in 1 and 2 using model V with pseudo-gtensors of each Kramers doublet and the composition in the | J 〉 basis of the ground state.