Lanthanide complexes based on a conjugated pyridine carboxylate ligand: structures, luminescence and magnetic properties

Three lanthanide compounds have been synthesized, namely, {[Dy2(bpda)3(H2O)3]4·2H2O}(Dy-1), {[Sm(bpda)2·(H2O)]·H2O}n (Sm-2) and {[Tb2(bpda)3(H2O)3]4·2H2O} (Tb-3) (H2bpda = 2,2′-bipyridine-6,6′-dicarboxylic acid). Their structures were determined by single crystal X-ray diffraction and characterized by elemental analysis, infrared spectroscopy and thermogravimetric analysis. Dy-1 and Tb-3 are isostructural with a conjugate bimolecular four-nuclear cluster structure constructed with intramolecular hydrogen bonds and they form a 3D supramolecular structure with intermolecular hydrogen bonding. Sm-2 is a one-dimensional chain structure and is further connected by intricate hydrogen bonds into a three-dimensional supramolecular structure. These three compounds exhibit significant characteristic luminescence from the ligand to the central Ln(iii) ion, which is found by solid-state photoluminescence measurement. Sm-2 exhibits a long luminescence lifetime and high fluorescence quantum yield. A slow relaxation phenomenon is observed for the dysprosium compound by measuring the alternating-current susceptibility at low temperature and the underlying mechanism was further confirmed by theoretical calculations.


Introduction
Through the past decades, lanthanide compounds including europium(III), terbium(III) and samarium(III) compounds have attracted much attention in the elds of biological analysis, 1 magnetism, 2,3 chemosensors, 4-6 electroluminescent devices and laser systems. [7][8][9][10][11] However, lanthanide(III) ions usually have very low optical transition absorption coefficients, which greatly limits their practical applications. This disadvantage can be overcome by using highly absorbent ligands to efficiently sensitize lanthanide ions. Aer the introduction of suitable organic ligands, lanthanide compounds exhibit unique photophysical properties, such as strong luminescence, high quantum yield, long luminescence life and large Stokes shis. 12,13 Experiments show that H 2 bpda ligand is an efficient sensitizer for the lanthanide ions because the energy gaps between the ligand and lanthanum ions is suitable for the effective ligand-to-metal energy transfer. Moreover, H 2 bpda ligand can be coordinate to lanthanide ions in various modes, such as unidentate coordinating, bidentate chelating and bridging coordination since the carboxyl group of the ligand can be partially or completely dehydrogenated by adjusting the pH value. 14 Thereby, the rich diversity of structures have the potential to have some fascinating properties. Meanwhile, the magnetic properties of lanthanide compounds are also an attractive eld 15,16 and they have provided an opportunity to shed light on tuning of the magnetic properties of Ln(III) compound. 17 Single-molecule magnets (SMM) exhibit magnetic bistability and quantum magnetic properties due to the existence of magnetic anisotropic energy barrier, which makes it a candidate material for ultra-high density information storage, quantum computing and molecular spintronic. 18 The magnetic measurements in this article reveal that Dy-1 displays weak magnetic relaxation under a zero dc eld. Combined with the ab initio calculations, the magnetic anisotropy and magnetic dynamic of Dy-1 were studied.

Materials and physical measurements
H 2 bpda and other raw materials are analytical reagents, purchased from commercial channels, without further purication. Elemental analysis of carbon, hydrogen and nitrogen was performed on a Vario EL III elemental analyzer. Fouriertransform infrared (FT-IR) spectra (4000-400 cm À1 ) were collected in the solid state on an Avatar™ 360 E. S. P. IR spectrometer using KBr pellet. Using SDT 2960 thermogravimetric analyzer, the temperature rise rate is 10 C min À1 (Al 2 O 3 ceramic disc is the support) when nitrogen ow is 40 mL min À1 in the range of 30-800 C, and thermogravimetric analysis (TGA) is carried out. Solid-state luminescence spectra, luminescence lifetimes and luminescence quantum yield (QY) of the three compounds were measured with an Edinburgh instrument FLS1000 uorescence spectrometer at room temperature. Luminescence QY was also collected by the same Edinburgh FLS1000 which equipped with an integrating sphere. Magnetic susceptibility were measured with Quantum Design PPMS-XL9 VSM. DC variable-temperature magnetic susceptibilities were measured under a 0.1 T applied magnetic eld in 2-300 K. The diamagnetic contribution calculated by Pascal constants was used to correct all the data.

Syntheses
H 2 bpda (0.036 g, 0.15 mmol), 0.5 mL Ln(NO 3 ) 3 (Ln ¼ Dy, Sm, Tb) (0.1 mol L À1 ) and 12 mL deionized water were mixed together and stirred for several minutes, adjusting the pH value to 4 with hydrochloric acid. The mixture was placed in a 25 mL Teonlined autoclave, heated at 150 C autogenic pressure for 3 days, and then cooled to room temperature at a rate of 1 C h À1 . Aer ltration, washing and drying, crystals of 1-3 were obtained suitable for X-ray diffraction analysis.

X-ray crystallography
The X-ray intensity data of compounds were collected on an Oxford Diffraction Super Nova area-detector diffractometer using mirror optics monochromatic MoKa radiation (l ¼ 0.71073 A) at 296(8) K. Structures have been solved with olex2.solve and rened with ShelXL (2014) and olex2.rene. 19 Crystal analytical data are shown in Table 1, and selected bond lengths and bond angles are shown in Table S1 of ESI. †

Infrared spectroscopy and absorption spectra
Within the range of 4000-400 cm À1 , the IR spectra of 1-3 compounds and H 2 bpda were determined. As shown in Fig. 1, the similarity of the complexes 1 and 3 spectra suggested that they had similar coordination structures. The broad bands at 3070-3500 cm À1 are assigned to O-H stretching vibrations in Dy-1 and Tb-3, while this band is not obvious in Sm-2, indicating that there is hardly any hydroxyl group in this complex.
In the IR spectrum of H 2 bpda ligand, the bands at 1692 cm À1 and 1264-1325 cm À1 could be attributed to stretching vibration (n(C]O)) and bending vibration (d(O-H)) of carboxylic acid, respectively. These two bands disappeared in complexes 1-3 and two new ones of 1545-1667 cm À1 and 1378-1457 cm À1 appeared. The bands at 1545-1667 cm À1 could be attribute to the asymmetric stretching vibrations of carboxylate (n as (COO À )), and the bands at 1378-1457 cm À1 could be ascribed to the symmetric stretching vibrations of carboxylate (n s (COO À )) in the complexes. 20 The separation (Dn) between n as (COO À ) and n s (-COO À ) can be used to explain the coordination types of carboxyl groups in ligand. Therefore, the Dn values of 167-210 cm À1 in the spectra of compounds 1-3 suggest that the carboxylate groups may coordinate to the lanthanide ions via monodentate and bidentate coordination modes. 21,22 The UV-vis absorption spectra of the ligand and the three complexes were measured in DMF solvent. As shown in Fig. 2, the absorption peaks of H 2 bpda appeared at 302 nm, which could be ascribed to the n-p* or p-p* absorption of conjugate pyridine ring of the ligand. This peak disappears in the complexes and new absorption peaks appeared at 305-317 nm. Relative to the ligand, the peak red shi in the complexes indicated that the ligand coordinated to the lanthanide ions and lower the energy. 23 In addition, compared with the ligand, the absorption spectra of lanthanide compounds changing, indicated that the coordination of the lanthanide ions signicantly inuence the energy levels of the ligands. 24

Structural descriptions
Complexes Dy-1 and Tb-3 are isostructural, hence only the structure of Dy-1 is discussed detailed as a representative. As shown in Fig. 3a, the structure unit of Dy-1 consists of two Dy(III) ion, three bpda 2À ligand, three coordinated water molecules. Dy1 is eight-coordinated of which four oxygen atoms and four nitrogen atoms come from two bpda 2À ligands, respectively. Dy2 is also eight coordinated, where two oxygen atoms and two nitrogen atoms come from the same bpda 2À ligand and three oxygen atoms from three water molecules and the rest of the oxygen atom from another bpda 2À ligand. Dy-O bond distances are in the range of 2.277(2)-2.390(3) A and Dy-N lengths in the range of 2.457(2)-2.501(3) A, which indicate that oxygen atom has stronger coordination capacity than nitrogen atom. The adjacent dysprosium ions Dy1 and Dy2 are connected by O9 of the carboxyl group. The distance between Dy1 and Dy2 is 6.278 A. The [Dy 2 (bpda) 3 (H 2 O) 3 ] units are connected by hydrogen bonds between lattice water molecules and coordinated water molecules (Fig. 3c), forming a three-dimensional supramolecular structure.
Complex 2 crystallize in monoclinic system, P2(1)/n space group. As shown in Fig. 3b, Sm1 is nine-coordination geometry with an O5-N4 donor set containing four oxygen atoms and four nitrogen atoms from two bpda 2À ligands and another oxygen atom from one water molecule, resulting in a distorted tri-  Table S1, ESI † for details). Adjacent samarium ions are bridged through the carboxyl group in the ligand to form one-dimensional chain structure (Fig. 3d). Fig. 4 is the thermogravimetric analysis diagram of complexes 1-3. Dy-1 and Tb-3 are heterogeneous isomorphism and their TGA curves similar, and then take Dy-1 as an example for analysis. As shown in Fig. 4, Dy-1 starts the rst decompose at below 242 C. The observed weight loss of 9.91% is consistent with the calculated value of 9.7%, which can assign to the decomposition of coordinated water molecules and lattice water molecule. The main mass loss occurs in the temperature range of 460-544 C with the loss of 54.41%, corresponding to the decomposition of residual organic components of the compound (calculated value, 54.92%), which is consistent with the crystal structure analysis results. The mass percentage of the residue is 34.91%, which is basically consistent with the theoretical calculation value of 34.83% of the oxide, so the nal residue is considered as Dy 2 O 3 . For Sm-2, the TG curve exhibits an initial mass loss of 4.74% over the temperature range 55-135 C, corresponding to the department of the lattice water molecule and coordinated water molecule (calculated. 4.10%). The second main mass loss (58.43%) occurs in 395-491 C corresponding to the decomposition of the residual organic components of the compound (calculated, 58.5%). The residue is thought to be Sm 2 O 3 (found 31.05%; calculated 31.43%).

Photoluminescence properties
The solid-state emission spectra of the compounds 1-3 were measured at ambient temperature and shown in Fig. 5. As   Fig. 1 The IR spectra of H 2 bpda ligand and complex 1-3. shown in Fig. 5a, the visible region emission from Dy-1 consists of four transitions, 4 F 9/2 / 6 H 15/2 (magnetic-dipole), 4 F 9/2 / 6 H 13/2 , 4 F 9/2 / 6 H 11/2 (hypersensitive, electric-dipole) and 4 F 9/2 / 6 H 9/2 , corresponding emission peaks are 487, 546, 577 and 662 nm, respectively. The emission band at 577 nm ( 4 F 9/2 / 6 H 11/2 ) is the strongest among the four bands, which is strongly inuenced by the local environment prevailing around Dy(III) ions. 25 The emission intensity ratio of 4 F 9/2 / 6 H 11/2 vs. 4 F 9/2 / 6 H 15/2 is 2.6 and so high value indicates that this complex lacks a centre of symmetry, 26,27 since 4 F 9/2 / 6 H 11/2 is probed to determine coordination symmetry around Dy(III) system. 28 To Sm-2, the sharp peaks at 590, 616, 652 and 700 nm should be ascribed to the Sm(III) transitions of 4 G 2/5 / 6 H J (J ¼ 5/2, 7/2, 9/2 and 11/2, respectively). The emission band at 616 nm ( 4 G 5/2 / 6 H 7/2 ) is the strongest among the four bands (Fig. 5c). The luminescent emission spectrum of Tb-3 (Fig. 5e) shows four typical bands at about 488, 547, 588 and 623 nm, which correspond to the transitions of the excited state 5 D 4 to the ground states 7 F J (J ¼ 3, 4, 5, and 6) of the Tb(III) ions, respectively. 29,30 Among the four bands, the emission band at 547 nm is the strongest, emitting green light visible to the naked eye under laser lamp. Fig. 5b, d and f are corresponding luminescence decay curves of complexes 1-3 and they are measured in the condition of the strongest emission peak. The corresponding uorescence lifetime value (s) obtained by tting curves of the decay curves on FLS1000 Photoluminescence Spectrometer are listed in Table 2. As shown in Table 2, the experimentally tted value of s for 2 is 847 ms at 616 nm and it has the longest uorescence lifetime among the three complexes. It is also a rather better value with comparison to some previous Sm-complex. 31,32 The data of solid quantum efficiency measured in the condition of maximum emission for 1-3 are reported in Table 3. The quantum yield of the samarium compound is calculated to be 21.4%, which is much higher than those of the samarium complexes reported in the literature (typically in the range of 1-20%), 13,14,31 though a small amount of water molecules are involved in the coordination sphere.

Magnetic properties
The direct current (dc) magnetic susceptibility of Dy-1 was studied in an applied magnetic eld of 2000 Oe and the temperature range 300-2 K and plotted as c MT vs. T in Fig. 6. For Dy-1, the observed c MT value is 51.55 cm 3 K mol À1 at 300 K, which is lower than the expected value of 56.68 cm 3 K mol À1 for four uncoupled Dy(III) ions (S ¼ 5/2, L ¼ 5, 6 H 15/2 , g ¼ 4/3). Upon cooling, c MT gradually decreases until 12 K and then drops rapidly to reach a minimum of 42.67 cm 3 K mol À1 at 2 K. The susceptibility c M (blue line in Fig. 6) increases slowly with decreasing temperature and then dramatically increases to 21.37 cm 3 mol À1 at 2 K. This behavior is attributed to antiferromagnetic exchange interaction between the Dy(III) ions and Stark energy level degeneracy caused by spin-orbit coupling. 33,34 The variation of alternating current (ac) susceptibility with frequency and temperature for Dy-1 under 2.2 Oe ac oscillating eld was studied, and the magnetization kinetics was  Fig. 4 The TGA curves of 1-3.
This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 6192-6199 | 6195 investigated ( Fig. 7 and 8). The results show that there is magnetic relaxation in complex Dy-1, which is the typical characteristic associated with single molecular magnet behaviour.
The ac susceptibility experiments were carried out in the range 2-19 K and the frequencies selected were 10, 50, 100, 500, 800 and 1000 Hz, respectively. As displayed in Fig. 7 and S1, ESI, † the relaxation time at different temperatures was obtained  by tting the c 00 or c 0 vs. frequency curves. As shown in Fig. 7, below 8 K, the values of c 0 and c 00 keep increasing on cooling, which indicates that there is a slow relaxation of magnetization phenomenon in Dy-1 expected for a single-molecule magnet. Furthermore, the phenomenon of magnetic relaxation becomes more obvious with the increase of magnetic eld intensity and the decrease of temperature. However, slow relaxation of magnetization for Dy-1 is observed experimentally only in a narrow temperature range, and no maximum of c 00 is observed in the temperature window, in which the energy barrier and corresponding relaxation time could not be calculated. Alternatively, a method employed by G. Y. Yang et al. 35 can be used to evaluate roughly the energy barrier E a and relaxation time s 0 based on the following relationship (eqn (1)): By nonlinear tting the experimental ln(c 00 /c 0 ) vs. 1/T at different frequencies, we obtained an estimate of the activation energy E a /k B ¼ 1.14 K and s 0 ¼ 1.2 Â 10 À6 s. A more precise result must wait for very low temperature measurements (T < 1 K) by using a micro-SQUID.
As described in 3.2 structural descriptions section, Dy1 is coordinated to four nitrogen atoms and four oxygen atoms while Dy2 is connected to two nitrogen atoms and six oxygen atoms. Due to the bond lengths are different aer Dy(III) coordinating to nitrogen or oxygen atoms, the symmetry and intensity of Dy(III) coordination eld are different, which may strongly impact on magnetic anisotropy, leading to distinct dynamic behavior. 36,37 It is visible that magnetic relaxation mainly results from Dy(III) anisotropy that is very sensitive to changes of the coordination geometry.
Additionally, magnetization data (M) for Dy-1 were collected in the eld range 0-70 kOe and at 2, 3 and 5 K, as shown in the Fig. 8, the magnetization measurements of Dy-1 increase rapidly for low elds then increases gently to a value of 20.89 Nb for Dy-1 at 2 K and 70 kOe. This value is lower than the expected saturation value of 40 mB, but close to four uncorrelated Dy ions' magnetic moments (4 Â 5.23 mB), which is likely due to crystal-eld effects and the low-lying excited states. 38 At 2 K, the M versus H data of Dy-1 exhibit slim buttery-shaped hysteresis loops without a remanence and a coercive eld (Fig. S2, ESI †). This lack is due to the slow sweep rate of the loop compared with the fast zero-eld relaxation. 39,40 The non-superposition of the M versus HT À1 curves obtained (Fig. 8) suggests the signicant magneto-anisotropy and a low-lying excited state present in Dy-1. 41-43

Conclusion
This article reports the syntheses, crystal structures, uorescence and magnetic properties of three Ln(III) compounds base on H 2 bpda ligand. Complexes 1, 2 and 3 exhibit strong f-f transition and long luminescent lifetime and high quantum efficiency, which indicated that the ligand H 2 bpda was a good organic chelator to absorb and transfer energy to Dy(III), Sm(III), Tb(III) and could be considered as promising candidate in the design of photoluminescence devices. Dy-1 exhibit slow magnetization relaxation and single molecular magnet behaviour except showing obvious yellow emission, which it may be a candidate of photo-magnetic functional.