Novel 3-Ethoxysalicylaldehyde Lanthanide Complexes Obtained by Decomposition of Salen-Type Ligands

Abstract

Three new asymmetrically coordinated lanthanide derivatives based on the bicompartmental salen-type ligands N,N′-bis(3-ethoxysalicylidene)propylene-1,3-diamine (H₂EtOsalpr) and 3-ethoxysalicylaldehyde (HEtvain) have been synthesized and structurally and photophysically characterized. All the compounds show dimeric structures of the general formula [Ln(H₂EtOsalpr)(NO₃)₂(Etvain)]₂ (Ln = Nd, Eu, Dy), with each salen-type ligand bridging two lanthanide ions. The Etvain ligand comes from the H₂EtOsalpr decomposition being coordinated to the corresponding lanthanide. The Nd(III) derivative shows fluorescence emission in the NIR region, but for the Eu(III) and Dy(III) compounds, only a broad band, attributed to the ligand emission, was observed.

1. Introduction

Lanthanide(III) ions are a series of elements that exhibit a number of interesting luminescent and magnetic properties with a wide range of applications [1,2]. The photophysical properties and high quantum yields exhibited by these cations are due to electronic transitions between the 4f orbitals, whose emission lines span much of the electromagnetic spectrum from the UV/Vis to the infrared and can give both phosphorescence and fluorescence events. Although the luminescence of these ions makes them very promising elements for their use in the synthesis of materials with different properties, they have a drawback: f-f transitions are forbidden by the selection rules, which results in high quantum yields but poor luminescence in materials with lanthanides. This disadvantage can be overcome by the so-called antenna effect, which consists of the indirect excitation of a cation through the absorption of energy by the metal environment and the transfer of this energy to the lanthanide so that the choice of an environment capable of absorbing more energy and acting as an ‘antenna’ will result in more efficient luminescent materials [3,4,5,6,7,8,9]. The ligand must be able to transfer this energy to the corresponding lanthanide, usually from its triplet state (T₁), since they have longer lifetimes and a more suitable energy for transfer to the lanthanide orbitals. This implies that the ligand must have an intersystem crossing (ISC) of as high as possible and a fluorescence quantum yield of as low as possible, otherwise the transfer will not be effective.

Within of the broad group of Schiff bases, the family of N,N′-bis(salicylidene)ethylene-1,2-diamine (salen) derivatives are of great interest [10,11,12]. These ligands are easy to modify, so it is possible to change their electronic nature, their chirality, their stability, their flexibility and the number of functional compartments in their structures [13]. In turn, these ligands possess chromophore groups capable of absorbing energy, making them an interesting choice for their use as antenna ligands in the synthesis of luminescent materials. In addition to their photophysical properties, salen-type ligands allow a large number of modifications in their structure, which in turn permits their coordination to a wide range of atoms ranging from transition metals to lanthanides and even small molecules as chelating ligands. Then, an enormous structural variety of complexes is possible. While the presence of nitrogen donor atoms generates a tendency toward the coordination of transition metals [14,15], the introduction of groups with oxygen atoms in specific positions allows binding to metals with larger sizes and oxophilic characters, as is the case of lanthanides [15].

For that reason, the study of complexes of trivalent lanthanide ions with salen-type ligands has increased in recent years in order to enhance their desirable luminescent properties. In this way, the lanthanide complexes obtained by reaction with the bicompartmental salen-type ligand N,N′-bis(3-methoxysalicylidene)propylene-1,3-diamine (H₂MeOsalpr) have been widely studied (Figure 1). These compounds show notable structural diversity and several derivatives, with different dimensionalities that have been reported. These structures include examples of monomers [16], dimers [17], tetramers [18] or coordination polymers [19]. In contrast, there are no references of lanthanide structures with the very similar ligand N,N′-bis(3-ethoxysalicylidene)propylene-1,3-diamine (H₂EtOsalpr). With this aim in mind, we have studied the reactions of the H₂EtOsalpr ligand with several lanthanide ions with the intention of using the salen-type ligand as antenna complex to determine the luminescent and structural properties. This combination would yield new materials with potential uses in different fields such as medicine, biology or the lighting industry [19,20,21].

2. Results and Discussion

2.1. Synthesis

New lanthanide compounds 1–3 (Figure 2) were synthesized by reaction between the H₂EtOsalpr ligand and the corresponding lanthanide nitrate, forming a double phase and employing a methanol/acetonitrile mixture, as indicated in the experimental section. Compounds 1 and 2 were obtained as single crystals, which were suitable for single-crystal X-ray diffraction. Compound 3 precipitated as an orange powder.

All the compounds are insoluble in common organic solvents. Due to this fact, crystallization was carried out by the slow diffusion of the solutions of the respective reagents. Prolonged quiescence of the compounds resulted in the partial hydrolysis of the original ligand, giving rise to the starting reagents, the diamine and 3-ethoxysalicylaldehyde, the latter deprotonated. This process has previously been observed in Cu(II) and Fe(III) compounds coordinated to analogous salen derivatives [22,23]. 3-ethoxysalicylaldehyde (HEtvain), thus formed, can coordinate to the lanthanide ion due to the oxophilic character of these metal atoms via the carbonyl and phenoxy groups, resulting in the compounds finally obtained (Figure 2). In order to check this assumption, a reaction of dysprosium salt, the salen-type ligand and the 3-ethoxysalicylaldehyde in a 1:1:1 molar ration was made. The isolated compound showed the same infrared spectrum as that of compound 1.

2.2. Infrared Spectroscopy

All the studied compounds showed similar solid-state infrared spectra, as they all have the same structure and, therefore, a minimum shift of the main bands was expected (Figure S1). Each of the spectra showed a strong band at around 1644 cm⁻¹, attributed to the ν(C=N) stretching mode of the H₂EtOsalpr ligand. This band shifted to higher frequencies relative to the free ligand because of its coordination to the metal ion. In the same way, the ν(C-O_phen) band, appearing at 1220 cm⁻¹, was moved to lower frequencies due to the coordination of the lanthanide ion to the phenolic oxygen atom. At 1619 and 1605 cm⁻¹, two additional bands were observed that corresponded to the carbonyl ν(C=O) vibration mode of the metal-coordinated Etvain ions. In all the complexes studied, the bands corresponding to the nitrate groups appeared at around 1440, 1300 and 1030 cm⁻¹. The values of the frequencies of the compounds obtained suggest that the nitrates were coordinated in a bidentate fashion to the metal ion [24], as confirmed by the single-crystal structure detailed below.

2.3. Single-Crystal X-Ray Diffraction Analysis

The Dy(III) complex 1 crystallized in the triclinic P-1 (2) space group. The molecular unit consisted of dimers of the formula [Dy(H₂EtOsalpr)(NO₃)₂(Etvain)]₂ (Figure 3), in which the Dy(III) atoms were bridged through two H₂EtOsalpr ligands showing a Dy(III)···Dy(III) distance of 9.8247(5) Å. The asymmetric unit corresponded to the formula unit, since both halves of the dimer were related through an inversion center. The bridging salen-type ligand coordinated in a bidentate way to one of the dysprosium ions using the phenoxy and ethoxy oxygen atoms of one end, while at the other end, it was linked to the other metal ion in a monodentate fashion through the phenoxy group. The metal–oxygen bond distances were 2.246(3) and 2.283(3) Å for the phenoxy oxygen atoms and 2.714(4) Å for the weakly coordinated ethoxy group. The cation coordination was completed by the four oxygen atoms of two bidentate nitrate groups and two additional oxygen atoms corresponding to the phenoxy and carbonyl groups of a deprotonated 3-ethoxysalicylaldehyde species. The metal–oxygen bond distances in this case were 2.401(4) and 2.233(3) Å, corresponding to the carbonyl and phenoxy oxygen atoms, respectively; these distances are comparable with those found in similar compounds [25]. The coordinative environment of the dysprosium ion corresponded to nine oxygen atoms that surround the metal, forming a distortedly capped square antiprism (Figure 4).

The deprotonated 3-ethoxysalicylaldehyde species appeared in the reaction due to the cleavage of the original ligand, as has been previously commented. The H₂EtOsalpr ligands were neutral, with the phenolic hydrogen located on the imino nitrogen and forming an intramolecular hydrogen bond to the phenoxy oxygen atom with the distances of N-H = 0.860(5) Å and O···H = 1.988(5) Å and an N-H···O angle of 132.0(1)°. This hydrogen bond corresponded to the phenolic hydrogen migration to the imine nitrogen, as is usually observed in this type of compound [9]. The presence of this H-bond precluded any possible interactions between the lanthanide ions and the nitrogen atoms, since the Dy-N distances ranged between 4.27 and 4.5 Å, a fact expected from the oxophilicity of the lanthanide ions. The two aromatic parts of the ligand formed a torsion angle of 123.3(1)° in order to accommodate the bridge between the metal ions. Each of these dimeric units was isolated, and they did not show significant interactions between them (Figure 5).

In a similar way, the equivalent europium complex 3 resulted in being isostructural to the Dy and Nd derivatives due to the similarity of both the infrared spectra and the powder diffraction patterns (Figure S2).

2.4. Luminescent Properties

The lanthanide compounds described in this work were expected to show emission in different spectral regions. Thus, the Nd(III) usually shows three emission bands in the NIR, centered at 880 nm, due to the fluorescence ⁴F_3/2 → ⁴I_9/2 transition, at 1060 nm (⁴F_3/2 → ⁴I_11/2) and around 1300 nm (⁴F_3/2 → ⁴I_13/2), the second one being much more intense than the other two. On the other hand, the Dy(III) and Eu(III) show phosphorescence phenomena. The emission of the former appears in the orange region, with a band centered at around 570 nm due to the ⁴F_9/2 → ⁶H_{15/2 →. 5/2} transition, while the emission of the latter appears in the red region, with a most significant maximum at 620 nm due to the ⁵D₀ → ⁷F₂ transition.

The emission properties of the three compounds were measured in the solid state, using an excitation wavelength of 373 nm since the ligands absorbed in this region. For the emission spectra of neodymium complex 2, a long-pass filter was always placed before detection to reduce the effects of the scattering produced by the solid. In this case, the antenna effect was clearly observed, since an emission maximum at 1060 nm was found (Figure 6). Along with this band, a second emission at around 1340 nm can be observed, as expected.

The emission spectra of the Dy and Eu complexes (1 and 3) showed a broad band, centered at 550 nm, that is attributed to the H₂EtOsalpr ligand (see below), while the typical emission bands of the lanthanide ions were not observed (Figure 7). If we analyze these spectra, it is not clear whether the lanthanide emission occurred, since the observed band, due to the ligand, is very broad. However, the values of the half-life times for the lanthanide ions appeared in the expected range. Since these measurements are more sensitive than the emission intensities, we assumed that the lanthanide luminescence promoted by the antenna effect of the salen-type ligand was overlapped by the ligand emission. This fact prevents us from identifying the emission of the lanthanide ions, in contrast to what has been reported for the luminescence studies of europium and dysprosium complexes with analogous ligands [26,27].

The presence of the broad emission band at 550 nm led us to study the luminescent properties of the ligands involved: in particular, H₂EtOsalpr, which, in theory, is a ligand able to allow an energy transfer to the corresponding lanthanide ion.

The absorption and emission spectra of the H₂EtOsalpr ligand are shown in Figure 8. The ligand absorbs in the UV region and emits in the visible. The absorption spectrum shows typical bands at 224, 262 and 334 nm, corresponding, respectively, to the transitions σ → σ*, π → π* and n → π*.

The cut-off point between the excitation (or absorption) spectra and the emission spectra gives information about the bandgap between the S₁ and S₀ states. In this case, the bandgap of the H₂EtOsalpr ligand in acetonitrile was 3.60 eV (345 nm).

Figure 9 shows the excitation and emission spectra of an ethanol solution of the H₂EtOsalpr and HEtvain species. As can be observed, both ligands absorbed and emitted in similar regions, a fact that could be detrimental to the transfer process to the lanthanide ion, as it would be difficult to selectively excite only the H₂EtOsalpr ligand, which is the ligand responsible for the antenna effect.

It should be noted that the H₂EtOsalpr ligand in the polar solvents favored the charge transfer state and also stabilized the transitions of π → π* and n → π*; hence, the cut-off point between the excitation and emission spectra was shifted to 418 nm (2.97 eV). This value is lower than expected for the bandgap between the S₁ and S₀ states and more probably corresponds to a charge transfer state into the ligand. In the HEtvain case, the bandgap was approximately 396 nm (3.14 eV).

The fluorescence lifetimes of the H₂EtOsalpr and HEtvain species were also measured (Figure S3 and Table S1). These lifetimes are typical for fluorescence, and there was a slightly longer one for the H₂EtOsalpr ligand. The ¹O₂ production of the H₂EtOsalpr ligand was measured in order to estimate the intersystem crossing (ISC). This value can give us information on how much the T₁ level was populated, but no ¹O₂ has been observed (Figure S4). This result may indicate either that there was no production of ¹O₂ or that the quantum yield of the ISC was very low.

Figure 10 shows the emission spectra of the H₂EtOsalpr ligand at 77 K collected at different time scales. The steady-state emission spectrum shows a high overlap of fluorescence and phosphorescence. Collecting the spectra at different time scales, it was possible to separate the fluorescence and phosphorescence emissions. The fluorescence maximum appeared at 516 nm (2.41 eV), while the phosphorescence maximum was observed at 544 nm. The triplet energy was estimated at the wavelength where the intensity was 10% with respect to the phosphorescence maximum value. Since this maximum appeared at 544 nm (2.29 eV), we can estimate the triplet energy at around 467 nm (2.68 eV).

At 77 K, the lifetime was estimated at the emission maximum (

Table 1. Fluorescence lifetimes of H₂EtOsalpr in ethanol at 298 K (λ_exc = 373 and λ_em = 479 nm) and emission lifetimes of H₂EtOsalpr in ethanol at 77 K (λ_exc = 373 and λ_em = 518 nm).

t (ns)	λem = 479 nm, 298 K	λem = 518 nm, 77 K Fluorescence Lifetimes	λem = 518 nm, 77 K Phosphorescence Lifetimes
τ1	6.2	4.9	3090
τ2	2.4	1.8	363
τint	4.0	3.4	2200
τamp	3.3	2.7	900

and Figure S4).

At 77 K, we were able to observe both fluorescence and phosphorescence and determine their corresponding lifetimes, and we have been able to determine the energy of T₁ in ethanol. In principle, the T₁ energy of the ligand is sufficient to produce fluorescence on the neodymium(III) ion and phosphorescence both in the dysprosium(III) and europium(III) ions, but with a low ligand ISC, this transfer will not occur efficiently.

3. Materials and Methods

3.1. Materials

The synthesis of these compounds was carried out using products and solvents from Sigma-Aldrich (Diegem, Belgium) and ThermoScientific (Waltham, MA, USA). The preparation of the ligand N,N′-bis(3-ethoxysalicylidene)propylene-1,3-diamine (H₂EtOsalpr) was carried out by a reaction between propylenediamine and 3-ethoxysalicylaldehyde (HEtvain) according to the literature [14,28].

3.2. Physical Measurements

The infrared spectra were collected in the solid state using a Fourier transform spectrophotometer with the ATR Perkin-Elmer accessory Spectrum 100 in the range of 4000–550 cm⁻¹ with a 4 cm⁻¹ resolution (UCM, Madrid, Spain). The CHN elemental analysis was performed on a LECO CHNS-932 at the Microanalysis Unit of the UCM (Madrid, Spain). Single-crystal data for [Dy(H₂EtOsalpr)(NO₃)₂(Etvain)]₂ (1) were collected at room temperature on a Bruker APEX-II CCD diffractometer working with graphite monochromated Cu-K_alpha radiation (λ = 1.54184 Å) at the X-ray Diffraction Unit of the UCM (Madrid, Spain). The cell parameters were determined and refined by the least-squares fit of all the reflections collected, and a semiempirical absorption correction was applied to the reduced data using the SADABS program [29]. The structures were solved by intrinsic phasing using the SHELXT solution program [30] and refined by full-matrix least-squares on F² using the SHEXL refinement package [31] running in the Olex2 environment [32]. All non-hydrogen atoms were refined anisotropically. The hydrogen atoms were included with fixed isotropic contributions at their calculated positions determined by molecular geometry. The methyl carbon atoms were modelized as disordered between two alternate positions, with fixed occupancy factors of 0.5. The ultimate-difference Fourier maps revealed no peaks of chemical significance. Information on the data collection parameters, crystallographic data and final agreement parameters appears in

Table 2. Crystal data and structure refinement for compound 1.

Empirical Formula	C60H66Dy2N8O26
Formula weight	1640.2
Temperature	293(2) K
Wavelength	1.54184 Å
Crystal system	Triclinic
Space group	P–1 (2)
Unit cell dimensions	a = 10.1607(3) Å b = 10.6719(3) Å c = 17.0363(5) Å α = 74.543(2)° β = 79.487(2)° γ = 80.239(2)°
Volume	1736.30(9) Å3
Z	1
Crystal size	0.11 × 0.08 × 0.02 mm
Density(calculated) μ	1.569 g/cm3
	12.102 mm–1
Theta(max)	68.31°
Goodness-of-fit on F2 Refinement	0.994 R[F2 > 2σ(F2)] = 0.0471 wR(F2) = 0.1255
Rint	0.18
Largest diff. peak and hole	1.644, −0.878 eA−3

. The CCDC deposit number for compound 1 is 2395092. X-ray powder diffraction (XRPD) measurements were performed in a Bruker D8 Advance diffractometer with monochromatized CuKα radiation. Simulated XRPD patterns from the single-crystal data were obtained with the Mercury 4.0 software [33].

The absorption, emission and emission lifetime spectra of all the studied substances were measured in HPLC-grade acetonitrile solutions using quartz cells with a 1 cm optical path length. The UV/Vis absorption spectra were recorded on a single-beam Varian Cary 50 spectrophotometer (UCM, Madrid, Spain). The fluorescence, excitation and phosphorescence spectra were recorded using a FluoTime 300 time-correlated single-photon-counting (TCSPC) spectrofluorometer (PicoQuant) equipped with a 300 W Xe-lamp as a steady-state excitation system with a spectral range from 250 nm to 2600 nm and a Peltier cooled photomultiplier tube (PMT, PMA-C 192-M type, 230–920 nm) (UCM, Madrid, Spain). The emission lifetimes were determined using, as an excitation source, a diode laser head emitting at 373 nm (LDH-P-C-375, temperature-stabilized laser emitting collimated picosecond pulses), and the emitted photons were detected, keeping the count rate below 1% to avoid artifacts such as the pile-up effect (PMA-C 192-M PMT and TimeHarp 260 PICO Single TCSPC PC plugin board for independent PCIe bus channels with 25 ps time resolution). The solid-emission measurements were performed by introducing a fine powder of the compound between two glass coverslips.

As the T₁ of the ligands was involved in the antenna effect, all measurements in the solutions were performed by purging with Ar for at least 15 min, since the atmospheric oxygen was a fluorescence quencher, but above all, it was very effective in quenching the T₁ state and, thus, phosphorescence.

The emission spectra and emission lifetimes under an inert atmosphere were measured with solvent-saturated Ar-purged solutions. The fluorescence and phosphorescence measurements of the H₂EtOsalpr ligand at 77 K were performed in HPLC-grade ethanol.

The ¹O₂ production was determined by directly detecting the lifetime of the ¹O₂ phosphorescence using the 373 nm laser, measuring at 1270 nm and using an 850 nm long-pass filter to eliminate the possible harmonics of the emission and excitation.

The EasyTau 2 program was used to acquire the emission and excitation spectra and to measure the emission lifetimes, and the extension of this program, Fluofit, was used to obtain the emission lifetimes.

3.3. Synthesis of the Compounds

All the compounds obtained were synthesized using the same method. As an example, the synthesis of the dysprosium derivative is described:

[Dy(H₂EtOsalpr)(NO₃)₂(Etvain)]₂ (1): A solution of 0.225 g (0.61 mmol) of the ligand H₂EtOSalpr in 8 mL of acetonitrile was placed in a test tube. Over this solution, 0.224 g (0.61 mmol) of Dy(NO₃)₃·H₂O dissolved in 8 mL of methanol was slowly added to obtain a double layer, which was kept at rest. After several days, single crystals suitable for X-ray diffraction were obtained. Analysis calculated for C₃₀H₃₅DyN₄O₁₃: C 43.83; H 4.29; N 6.81%. Found C, 43.77; H, 4.41; N, 6.81%. Most significant IR bands (cm⁻¹): 1644, 1618, 1605 ν(C=N) and ν(C=O); 1215 ν(C-O)_phen and 1053 ν_s(C-OEt); 1456 ν(N=O); 1297 ν_as(NO₂); 1032 ν_s(NO₂).

[Nd(H₂EtOsalpr)(NO₃)₂(Etvain)]₂ (2): Analysis calculated for C₃₀H₃₅NdN₄O₁₃: C 44.82; H 4.39; N 6.97%. Found C, 44.08; H, 4.36; N, 7.13%. Most significant IR bands (cm⁻¹): 1642, 1619, 1606 ν(C=N) and ν(C=O); 1218 ν(C-O)_phen; 1071 ν_s(C-OEt); 1440 ν(N=O); 1299 ν_as(NO₂); 1032 ν_s(NO₂).

[Eu(H₂EtOsalpr)(NO₃)₂(Etvain)]₂ (3): Analysis calculated for C₃₀H₃₅EuN₄O₁₃: C, 44.40; H, 4.35; N, 6.90%; C, 43.67; H, 4.31; N, 6.98%. Significant IR bands (cm⁻¹) 1644, 1618, 1605 ν(C=N) and ν(C=O); 1219 ν(C-O)_phen; 1075 ν_s(C-OEt); 1442 ν(N=O); 1306 ν_as(NO₂); 1032 ν_s(NO₂).

Dysprosium compound 1 was also synthesized by an alternate method: A solution of 0.225 g (0.61 mmol) of the ligand H₂EtOsalpr and 0.101 g (0.61 mmol) of HEtvain in 10 mL of acetonitrile was stirred in a round-bottom flask. Over this solution, 0.224 g (0.61 mmol) of Dy(NO₃)₃·H₂O dissolved in 5 mL of methanol were added. After two hours of stirring, a yellow solid precipitated. The solid was filtered, washed with methanol and vacuum-dried. The yield was 67%.

4. Conclusions

The reaction of N,N′-bis(3-ethoxysalicylidene)propylene-1,3-diamine (H₂EtOsalpr) with nitrate salts of dysprosium, neodymium and europium took place with the partial hydrolysis of salen-type ligands and coordination of deprotonated 3-ethoxysalicylaldehyde (Etvain) to the lanthanide ion. The final obtained derivatives were dimers with the metal ions bridged by two asymmetrically coordinated H₂EtOsalpr ligands. The luminescent properties of these compounds are clearly dominated by the emission of the salen-type ligand that presents the singlet and triplet excited states very closely in energy. This fact can favor the thermally activated delayed fluorescence of the ligand in detriment of its antenna effect and, therefore, restrains the energy transfer to the lanthanide ions. For this reason, no neat emissions were observed for the dysprosium and europium compounds, whose phosphorescence signal should have overlapped with the more intense ligand emission. On the contrary, the neodymium ion presented a neat fluorescence emission signal in the NIR region, distant enough from the ligand emission, making this compound a good candidate for use in medical applications due to the usual biocompatibility observed in salen ligands [34,35].