Short Radiative Lifetime and Non‐Triplet Sensitization in Near‐Infrared‐Luminescent Yb(III) Complex with Tripodal Schiff Base

Abstract

We prepared Ln(III) (Ln=Eu, Gd, and Yb) complexes with a tripodal Schiff base, tris[2‐(5‐methylsalicylideneimino)ethyl]amine (H₃L) and studied their photophysical properties. Upon ligand excitation, YbL showed Yb(III)‐centered luminescence in the near‐infrared region. While the overall quantum yield (0.60(1)%) of YbL in acetonitrile was moderate among the reported values for Yb(III) complexes, its radiative lifetime (0.33(2) ms) was significantly shorter than those reported previously. We propose that the ligand‐to‐metal charge‐transfer (LMCT) state mediated the sensitization in YbL. The emission and excitation spectra of EuL indicated the participation of the LMCT state in the sensitization. The radiative lifetime (0.84(7) ms) for EuL in the solid state was rather short compared to those of reported Eu(III) complexes. Our results show that the Yb(III) complex with the Schiff base ligand has two features: the short radiative lifetime and the non‐triplet sensitization path.

A seven‐coordinated Yb(III) complex with a tripodal Schiff base exhibited efficient luminescence in the near‐infrared region with an extremely short radiative lifetime. The sensitization process for the Yb(III) complex was explained by the ligand‐to‐metal charge‐transfer state mediated mechanism.

1. Introduction

The trivalent ytterbium (Yb(III)) ion has attracted significant interest for its application in several technologies, including biological imaging and solar‐energy conversion, due to its fascinating near‐infrared (NIR) luminescence property. ^[1] Thus far, Yb(III) complexes with light‐harvesting organic ligands have been extensively studied. The sensitization of the luminescence of Yb(III) via such ligands has been established as a promising strategy for overcoming the inefficiency of the direct excitation of Yb(III).

The overall quantum yield for the ligand‐sensitized luminescence of Yb(III) (${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ ) is defined in Equation 1:

$${\displaystyle {\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}=\eta \times {\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{Y}\mathrm{b}}=\eta \times \frac{{\tau }_{\mathrm{o}\mathrm{b}\mathrm{s}}}{{\tau }_{\mathrm{r}\mathrm{a}\mathrm{d}}}}$$ (1)

where η is the sensitization efficiency, ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{Y}\mathrm{b}}$ is the intrinsic quantum yield, τ _obs is the observed lifetime, and τ _rad is the radiative lifetime. Typically, ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ of Yb(III) complexes are very low (<1 %). ^[2] Although Yb(III) complexes with higher ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ (>1 %) have been recently reported,[ ³ , ⁴ , ⁵ , ⁶ , ^7c , ^7d ] to design an Yb(III) complex with a high ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ , it is still important to optimize η and ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{Y}\mathrm{b}}$ .

The value of η correlates with the sensitization process from the ligand to Yb(III). Similar to the case with other Ln(III) complexes, adjusting the lowest triplet (T₁) level slightly above the Yb(III) excited level (²F_5/2: ∼10350 cm⁻¹) facilitates a T₁‐mediated energy transfer, resulting in an increase in η. ^[6] In addition to the T₁‐mediated energy transfer, some Yb(III) complexes have a unique sensitization path related to the easy accessibility of the Yb(III) ion to the Yb(II) state. This mechanism was proposed by Horrocks et al, ^[8] who explained a sensitization in the Yb(III)‐containing protein with electron transfer involving an intermediate state consisting of Yb(II) and a tryptophan radical. This mechanism is applicable to small molecules in which ligand‐to‐metal charge‐transfer (LMCT) states formally act as the intermediate states for the electron transfer mechanism.[ ⁹ , ¹⁰ , ¹¹ ] The understanding and optimization of the sensitization through the T₁ or LMCT state is required to maximize the η of Yb(III) complexes.

Regarding the optimization of ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{Y}\mathrm{b}}$ , the removal of the X−H bonds (X=C, N, O) from the ligand considerably increases τ _obs, ^[7] resulting in an increase in ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{Y}\mathrm{b}}$ . Additionally, Seitz et al. recently showed that a decrease in τ _rad, which reflects the radiative decay of f‐f luminescence, also induces an increase in ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{Y}\mathrm{b}}$ . ^[7c] However, there are relatively few experimental τ _rad values for Yb(III) complexes. Therefore, very little is known about the correlation between the structure of the Yb(III) complexes and τ _rad.

To understand the sensitized Yb(III) luminescence and design the highly luminescent Yb(III) complex, further reports on the quantitative photophysical parameters in Equation (1) are required. Recently, we studied the synthesis of Ln(III) complexes with a tripodal Schiff base (tris[2‐(5‐methylsalicylideneimino)ethyl]amine, H₃L, Figure 1). ^[12] We focused on the ligand in its deprotonated form, as a sensitizer for Ln(III) luminescence, since the first and second coordination spheres do not have N−H or O−H bonds that induce the significant deactivation of the Yb(III) excited level. To date, few reports on the sensitized luminescence of analogous Ln(III) complexes are available. ^[13] Molloy et al. recently reported that Yb(III) complexes with analogous ligands show NIR luminescence upon ligand excitation. ^[14] However, quantitative experimental values such as τ _obs, ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}},$ and τ _rad are still lacking. Furthermore, the sensitization mechanism remains unclear. Here, we study the photophysical properties of a Yb(III) complex with a tripodal Schiff base (YbL) to understand its sensitized luminescence. In addition, we studied the photophysical properties of EuL to support the discussion on the presence of the LMCT state and τ _rad for this series of Ln(III) complexes. This is because the participation of the LMCT state in the sensitization process has been proposed for not only the Yb(III) ion but also the Eu(III) ion, which is the most easily reducible ion among the Ln(III) series. ^[15] Another reason is that the τ _rad of the Eu(III) ion, as well as that of the Yb(III) ion, can be obtained experimentally with ease. ^[16]

Figure 1.

Structure of tris[2‐(5‐methylsalicylideneimino)ethyl]amine.

2. Results and Discussion

2.1. Synthesis of LnL (Ln=Eu, Gd, and Yb)

We prepared YbL according to our previous report on the synthesis of TbL. ^[12] The Schiff‐base condensation in the presence of a Yb(III) salt provided a pale‐yellow powder. Recrystallization from a DMF/methanol mixed solvent yielded YbL ⋅ 3/2DMF; DMF was easily extracted during the drying process to produce YbL. Previously, ^[12] we only reported the crystal structure of the temporal product TbL ⋅ 3/2DMF but not of the final product, TbL. Herein, we successfully determined the crystal structure of both YbL ⋅ 3/2DMF and YbL as DMF molecules were removed from YbL ⋅ 3/2DMF while maintaining sufficient crystallinity. The crystal structure of YbL ⋅ 3/2DMF, isostructural to that of TbL ⋅ 3/2DMF, ^[12] is provided in the Supporting Information (Table S1, Figure S1).

The synthesis procedures for EuL and GdL were slightly different from the one described above. The recrystallization of Eu(III) and Gd(III) complexes from DMF in the presence of methanol led to the formation of impurities. In contrast, the slow evaporation of DMF under reduced pressure resulted in the formation of analytically pure samples of EuL and GdL. It has been reported that a solvent molecule coordinates to a Gd(III) center in addition to the analogous tripodal Schiff base ligand, resulting in an eight‐coordination geometry. ^[17] Similar eight‐coordinated complexes might be formed as impurities during the recrystallization in the presence of methanol and a trace amount of water in the organic solvent. The larger ionic radius of the Eu(III) and Gd(III) ions than that of the Yb(III) ion may facilitate the coordination of the solvent as an additional ligand.

2.2. Crystal Structure of YbL

The compound, YbL, crystallizes in the triclinic space group of P $\bar{1}$ . The crystal data are summarized in Table S1. Figure 2a shows the Oak Ridge thermal ellipsoid plot (ORTEP) diagram of YbL with numbering. The Yb(III) center was in a seven‐coordinated environment with three O atoms (O1, O2, O3), three N atoms (N1, N2, N3), and an apical N atom (N4). To describe the coordination geometry, we evaluated the shape factor (S). ^[18] The shape factor S is the minimal variance of dihedral angles along all edges of the polyhedron, defined as

$${\displaystyle S=min\sqrt{\left(\frac{1}{m}\right)\sum _{i=1}^m{({\delta }_i-{\theta }_i)}^2}}$$ (2)

Figure 2.

(a) Oak Ridge thermal ellipsoid plot (ORTEP) diagram (at the 50 % probability level), (b) coordination geometry (monocapped octahedron), and (c) crystal packing of YbL. Hydrogen atoms were omitted for clarity. The green balls represent Yb(III). Gray, blue, and red balls represent C, N, and O atoms, respectively.

where m, δ _i, and θ _i represent the number of edges, the dihedral angles between the two planes along the ith edge, and the corresponding dihedral angles along the i ^th edge for the ideal polyhedron, respectively. We calculated the S values for the representative seven‐coordinated geometries (Figure S2): a monocapped octahedron (C _3v) and a monocapped trigonal prism (C _2v). The detailed analysis is summarized in Table S2. The estimated S value for the monocapped octahedron (7.37) was significantly smaller than that for the monocapped trigonal prism (13.57). Therefore, the coordination geometry of YbL is best described as a distorted monocapped octahedron (Figure 2b).

The crystal structure did not contain any crystal solvents (Figure 2c). There were several intermolecular interactions, such as CH⋅⋅⋅π interactions (Table S3) and CH⋅⋅⋅O hydrogen bonds (Table S4) between the YbL molecules. The powder X‐ray diffraction (PXRD) patterns of EuL and GdL were similar to that of YbL (Figure S3), confirming the isostructurality of the LnL series.

2.3. Ultraviolet‐Visible (UV‐vis) Absorption Properties of LnL

The acetonitrile solutions of LnL ([LnL]=4.0×10⁻⁵ M, Ln=Eu, Gd, and Yb) showed absorption bands around 360 nm (Figure 3a), which were assigned to the ligand‐centered absorption (S₀→S₁). To study the LMCT state in EuL and YbL, the differences among the absorption spectra of LnL were carefully examined. The LMCT transitions of Eu(III) and Yb(III) complexes often overlap with the ligand‐centered absorption. ^[15] Furthermore, the absorption coefficient of the LMCT transition in solution is often significantly lower than that of its ligand‐centered absorption. ^[11a] Therefore, the LMCT transition for Eu(III) and Yb(III) complexes was confirmed by comparison of their absorption spectra with those of corresponding Gd(III) complexes, where Gd(III) is not easily reduced. ^[19] From Figure 3a, slight but meaningful differences were observed between the absorption spectra of LnL. The subtraction of the GdL spectrum from that of EuL and YbL showed bands at 400 and 380 nm for EuL and YbL, respectively (Figure S4). These bands were tentatively assigned as LMCT transitions. A solution of LnL was also measured at a relatively high concentration ([LnL]=1.80×10⁻³ M, Figure 3b). Shoulders were observed on the longer‐wavelength side of the ligand‐centered absorption corresponding to the tails of the LMCT transitions of EuL and YbL. This indicates that the LMCT states of EuL and YbL are close in energy to the S₁ state. The wavelength of the LMCT band of EuL was longer than that of YbL, showing that the LMCT energy level of EuL is lower than that of YbL.

Figure 3.

Absorption spectra of LnL (Ln=Eu, Gd, and Yb) in acetonitrile. (a) [LnL]=4.0×10⁻⁵ M and (b) [LnL]=1.80×10⁻³ M.

The solid‐state absorption properties of LnL were studied by diffuse reflectance (DR) spectroscopy (Figure 4). The spectra were mainly composed of broad bands, which can be ascribed to the ligand‐centered absorption. The ligand‐centered absorption band in the solid state was broadened, as compared to that of the solution (Figure 3a). It has been reported that the intermolecular interactions in the crystals of Ln(III) coordination compounds afford additional excited states, such as the inter‐ligand charge‐transfer (CT) state. ^[20] We attribute the broadening of the absorption bands of LnL in the solid state to the formation of the inter‐ligand CT state induced by their tight crystal packing.

Figure 4.

Diffuse reflectance spectra of LnL (Ln=Eu, Gd, and Yb).

In the solid state, shoulders on the longer‐wavelength side of the ligand‐centered absorption in EuL and YbL were clear. These shoulders were assigned to the tails of the LMCT transitions. The absorption coefficients of the solid‐state LMCT bands were moderately smaller than those of the ligand‐centered absorption and significantly higher than those of the solution LMCT bands. Ilichev et al. suggested that the LMCT transitions in Ln(III) complexes are strongly influenced by the surrounding environment. ^[11b] In this previous report, LMCT transitions of Yb(III) complexes are observed only in the solid state, but not in the solution. On the other hand, the results presented herein indicate the presence of the LMCT transition in both solution and solid state. Our study did not allow us to discuss the effect of the surrounding environment on the LMCT bands quantitatively. At least in terms of EuL and YbL, the absorption coefficients of the LMCT transition in the solid state were significantly higher than those observed in solution.

2.4. Luminescence Property of YbL

The acetonitrile solution of YbL exhibited NIR luminescence (Figure 5a) upon excitation at 350 nm. This luminescence was attributed to the ²F_5/2→²F_7/2 transition with Stark splitting related to the coordination geometry of the Yb(III) ion. The excitation spectra of the solution (Figure 6a) showed a ligand‐centered excitation band, suggesting that sensitization by the ligand occurred.

Figure 5.

Emission spectra of YbL (a) in acetonitrile ([YbL]=4×10⁻⁶ M) and (b) in the solid state, recorded at room temperature (λ _ex=350 nm). Both spectra were normalized at 1061 nm.

Figure 6.

Excitation spectra of YbL (a) in acetonitrile ([YbL]=4×10⁻⁶ M, normalized at 365 nm) and (b) in the solid state (normalized at 350 nm), recorded at room temperature (λ _em=972 nm).

The YbL powder showed NIR luminescence (Figure 5b). The peak wavelengths, which correlate to the Stark splitting, were almost consistent with those of the solution. This observation indicates that the coordination environment for the powder is similar to that for the solution. As compared to the luminescence spectrum of the solution (Figure 5a), the intensities at wavelengths shorter than 1014 nm for the powder were weak. This observation may be due to re‐absorption. The acetonitrile solution of YbL exhibited f‐f absorption (²F_7/2→²F_5/2, Figure S5, red line) in the NIR region. Overlap between the longer‐wavelength region of the f‐f absorption and shorter‐wavelength region of the f‐f luminescence (Figure S5) would allow for re‐absorption in the solid state. Re‐absorption of the Yb(III) luminescence in a Yb(III) coordination polymer was previously described. ^[21] The excitation spectra of the YbL powder (Figure 6b) displayed a ligand‐centered excitation band with a shoulder corresponding to the LMCT excitation at 450–500 nm. This result suggests that sensitization from the ligand also occurs in the YbL powder.

Both the NIR luminescence decay profiles of the solution and the powder were well‐fitted mono‐exponentially (Figure S6). The obtained lifetime values are listed in Table 1. The τ _obs value of the solution is very close to that of the powder. These results show that the coordination geometry of Yb(III) in the solution and in the powder were similar. This is consistent with the observation of the luminescent spectral shape. The τ _obs values of YbL are lower than those reported for Yb(III) complexes containing C−H bonds (typically below 50 μs).[ ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ ] The rate of the nonradiative deactivation of the Ln(III) excited state is inversely proportional to the distance between the Ln(III) and the C−H bond. ^[22] According to the crystal structure of YbL, there are methylene and methine C−H bonds with short Yb⋅⋅⋅H distances (3.4–4.4 Å). These C−H bonds may significantly deactivate the Yb(III) excited state and shorten τ _obs in YbL.

Table 1.

Observed and radiative lifetimes, intrinsic quantum yield, and sensitization efficiency of YbL at room temperature. Experimental values are given with an uncertainty of ±2σ.

Compounds	τ obs [a] [μs]	τ rad [ms]	${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{Y}\mathrm{b}}$ [%]	${\Phi }_{\mathrm{Y}b}^{\mathrm{L}}$ [%]	η [%]
YbL in acetonitrile	4.07(5)	0.33(2)[b]	1.23(8)[c]	0.60(1)[d]	49(3)[e]
YbL in the solid state	4.1(1)	–	–	–	–

[a] λ _ex=337.1 nm. [b] The value of τ _rad was calculated using Equation (3). [c] ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{Y}\mathrm{b}}$ =τ _obs/τ _rad. [d] This value was determined by the relative method using the toluene solution of Yb(TTA)₃phen as the standard. [e] η=${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ /${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{Y}\mathrm{b}}$ .

We determined the ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ of YbL in acetonitrile by the relative method using Yb(TTA)₃phen (HTTA: 4,4,4‐trifluoro‐1‐(2‐thienyl)‐1,3‐butanedione, Phen: 1,10‐phenanthroline, ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ =1.1 % in toluene) ^[23] as a standard. The ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ value for the solution was found to be 0.60(1)%. We could not obtain the ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ for the powder because of the limitation of our instrumentation. The derived value of ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ for YbL in acetonitrile is comparable to the reported ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ value for Yb(III) complexes with C−H bonds (typically below 1 %). ^[2]

We estimated the τ _rad value from the f‐f absorption of YbL in acetonitrile (Figure S5, red line) using Equation 3: ^[16]

$${\displaystyle \frac{1}{{\tau }_{\mathrm{r}\mathrm{a}\mathrm{d}}}=2303\times \frac{8\pi c{n}^2{{\tilde{\nu }}_{\mathrm{m}}}^2(2J+1)}{N_{\mathrm{A}}(2J{}^{\text{'}}+1)}\int ϵ\left(\tilde{\nu }\right)d\tilde{\nu }}$$ (3a)

$${\displaystyle {\tilde{\nu }}_{\mathrm{m}}=\frac{\int \tilde{\nu }·ϵ\left(\tilde{\nu }\right)d\tilde{\nu }}{\int ϵ\left(\tilde{\nu }\right)d\tilde{\nu }}}$$ (3b)

where c is the speed of the light in cm s⁻¹, n is the refractive index (n=1.344 for acetonitrile), N _A is Avogadro's constant, J and J’ are the quantum numbers for the ground and excited states, respectively, $ϵ\left(\tilde{\nu }\right)$ is the molar absorption coefficient in M⁻¹ cm⁻¹, $\tilde{\nu }$ is the wavenumber in cm⁻¹, and ${\stackrel{\sim }{\nu }}_{\mathrm{m}}$ is the barycenter of the transition defined in Equation (3b). The estimated value of τ _rad for the solution was 0.33(2) ms. To the best of our knowledge, this value of τ _rad is the shortest value among the reported τ _rad (0.5–1.3 ms) for Yb(III) complexes.[ ³ , ^7c , ²⁴ ] Generally, Ln(III) luminescence mainly consists of magnetic‐dipole (MD) and induced electric‐dipole (IED) transitions. The Ln(III) coordination environment affects the IED transition intensity. It has been shown that the Yb(III) luminescence consists of MD and IED transitions between the Stark sublevels of the excited state (²F_5/2) and those of the ground state (²F_7/2). ^[25] The underlying mechanism of the short τ _rad of YbL remains unclear, but the low‐symmetry coordination environment (C _3v) may enhance IED transition probability.

The obtained values of ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ , τ _obs, and τ _rad allowed us to calculate ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{Y}\mathrm{b}}$ and η for the solution using Equation (1). The values of ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ , τ _obs, τ _rad, ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{Y}\mathrm{b}}$ and η are summarized in Table 1. The value of η suggests that significant sensitization occurs in YbL. Although the η value for the powder is not provided in the present work, the emission (Figure 5b) and excitation spectra (Figure 6b) show that sensitization definitely occurs in the solid state.

We also studied the visible luminescence of YbL. The acetonitrile solution and YbL powder showed visible luminescence (Figure S7). The luminescent intensity of YbL was significantly lower than that of the ligand‐centered luminescence observed in GdL (Figure S8, blue lines). Although accurate values could not be obtained, the visible luminescence quantum yields of GdL and YbL in acetonitrile were <0.5 % and <0.04 %, respectively. As compared to the NIR luminescence of YbL (${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ =0.60(1)% for YbL in acetonitrile), the quantum yields of the visible luminescence of YbL are meaningfully low. These observations suggest that significant sensitization from the ligand occurs in YbL.

To discuss the sensitization pathway for YbL, we estimated the energies of the excited states of the ligand in acetonitrile and in the solid state. The S₁ energy levels (E(S₁)) were determined from the crossing point between the normalized emission and absorption of GdL (Figure S8). The T₁ energy levels (E(T₁)) were estimated from the shorter wavelength edges of the phosphorescence spectra of GdL (Figure S9). The calculated energies are listed in Table 2. According to Reinhoudt's empirical rule, ^[26] the intersystem crossing (ISC) from the S₁ to the T₁ states is effective when the energy gap between the S₁ and T₁ states (E(S₁−T₁)) is greater than 5000 cm⁻¹. The values of E(S₁−T₁) for the Schiff base ligand were lower than 4000 cm⁻¹. Therefore, the ISC efficiency of YbL may be rather low. The estimated T₁ levels of the ligand are much higher than the ²F_5/2 state of Yb(III) by approximately 10000 cm⁻¹. Since there is no spectral overlap between the ligand phosphorescence and the f‐f absorption (²F_7/2→²F_5/2), the sensitization of YbL cannot be accounted for the Förster and Dexter‐type energy transfer from the T₁ state of the ligand.

Table 2.

Estimated energy levels of the S₁ and T₁ states for the ligand in GdL.

Compounds	E(S1) [cm−1]	E(T1) [cm−1]	E(S1−T1) [cm−1]
GdL in acetonitrile	ca. 24500	ca. 20900	ca. 3600
GdL in the solid state	ca. 22800	ca. 19700	ca. 3100

It is possible that the LMCT state consisting of Yb(II) and L^2−. is related to sensitization in YbL. Horrocks et al. proposed an electron‐transfer sensitization mechanism for Yb(III) luminescence. ^[8] The LMCT states can act as the intermediate states for this electron transfer mechanism.[ ⁹ , ¹⁰ , ¹¹ ] Additionally, it has been implied that the ISC from the LMCT state to the ²F_7/2 state can sensitize the Yb(III) luminescence. ^[27] These LMCT‐mediated sensitization pathways may be operative for YbL.

Figure 7 shows the proposed sensitization mechanism for YbL. The LMCT state in YbL is located close to the ligand‐centered excited state. Ligand‐centered excitation (Figure 7a) followed by a conversion from the S₁ to LMCT state (Figure 7b) could feed the LMCT level. It should be noted that direct LMCT excitation (Figure 7c) is also possible in the solid state, as observed in the excitation spectrum (Figure 6b). In contrast, direct LMCT excitation in solution is not effective due to the extremely low absorption coefficient. In any case, electron transfer and/or ISC from the LMCT state may feed the ²F_7/2 state (Figure 7d) and allow subsequent Yb(III)‐centered luminescence (Figure 7e).

Figure 7.

Proposed sensitization pathway for YbL.

2.5. Luminescence Property of EuL

The radiative lifetime of the Eu(III) ion, as well as that of the Yb(III) ion, can be easily determined experimentally. ^[16] Furthermore, the participation of the LMCT states in the sensitization of the Ln(III) luminescence has been reported not only for Yb(III) complexes but also for Eu(III) complexes. ^[15] We studied the luminescence property of EuL to understand its τ _rad and the role of the LMCT state in the sensitization process.

We measured the luminescence spectra of EuL in the solid state at room temperature and 77 K. At room temperature, the EuL powder displayed weak Eu(III) luminescence (⁵D₀→⁷F_0–6) with a broad luminescence assigned to the ligand‐centered or LMCT transition (Figure 8a). The luminescence intensity was so weak that we could not determine τ _obs and ${\Phi }_{\mathrm{E}\mathrm{u}}^{\mathrm{L}}$ using our experimental setup. The low luminescence intensity and the presence of the residual luminescence suggest that the sensitization is inefficient in EuL. The energy gap between the T₁ and ⁵D₀ levels of EuL are within the ideal range ^[28] for energy transfer. It has been reported that LMCT states that lie close to the T₁ or ⁵D₀ states act as quenching channels in Eu(III) complexes. ^[15] Thus, quenching through the LMCT state may contribute to the low sensitization efficiency of EuL at room temperature. At 77 K, the residual broad luminescence disappeared (Figure 8b). This observation suggests that the sensitization efficiency increased upon reducing the temperature to 77 K.

Figure 8.

Emission spectra of EuL in the solid state (λ _ex=340 nm), recorded at room temperature (a) and 77 K (b). Both spectra were normalized at 613 nm (⁵D₀→⁷F₂).

We also measured the luminescence spectrum of EuL in acetonitrile. Upon excitation of the ligand, the acetonitrile solution of EuL showed no Eu(III) luminescence at room temperature. At 77 K, the acetonitrile solution displayed Eu(III) luminescence (Figure S10a). The spectral shape of the solution was clearly different from that of the powder (Figure S10b). This observation indicates that the coordination geometry in the solution at 77 K might be different from the seven‐coordinated C _3v geometry. In the following text, we mainly discuss the luminescent property of the EuL powder, whose coordination geometry is the same as that of YbL. The photophysical data for the acetonitrile solution at 77 K are given in Supporting Information (Figure S11 and Table S5).

To obtain the photophysical parameters of EuL, we conducted further studies on the luminescence of EuL in the solid state at 77 K. The luminescence spectrum without the broad luminescence (Figure 8b) allowed us to determine the τ _rad of EuL at 77 K using the following equation: ^[16]

$${\displaystyle \frac{1}{{\tau }_{\mathrm{r}\mathrm{a}\mathrm{d}}}=A_{\mathrm{M}\mathrm{D},0}{n}^3\left(\frac{I_{\mathrm{t}\mathrm{o}\mathrm{t}}}{I_{\mathrm{M}\mathrm{D}}}\right)}$$ (4)

where A _MD,0 is the spontaneous emission probability of the ⁵D₀→⁷F₁ transition (14.65 s⁻¹), n is the refractive index (1.5 for solid), and I _tot/I _MD is the ratio of the total area of the Eu(III) luminescence spectrum to the area of the ⁵D₀→⁷F₁ transition. The obtained value of τ _rad was 0.84(7) ms, which is lower than the typical values for Eu(III) complexes (1–14 ms). ^[15a] Hasegawa et al. reported that the low‐symmetric coordination environment and LMCT perturbation into the 4 f configuration enhance the radiative rate constant of seven‐coordinated Eu(III) complexes.[ ^18b , ^18c , ²⁹ ] Our present study did not provide further information on the τ _rad of EuL. However, we infer that the low‐symmetric coordination environment (C _3v) and the presence of the LMCT state may be related to the significantly short τ _rad. Consequently, we showed that the tripodal Schiff base enhanced the radiative rate constant for both Yb(III) and Eu(III).

We recorded the luminescence decay profile of the ⁵D₀→⁷F₂ transition of EuL in the solid state at 77 K (Figure S12). The decay profile was mono‐exponential, showing that τ _obs was 0.30(1) ms. We calculated ${\Phi }_{\mathrm{E}\mathrm{u}}^{\mathrm{E}\mathrm{u}}$ for EuL at 77 K using the τ _rad and τ _obs values. The values of τ _rad, τ _obs, and ${\Phi }_{\mathrm{E}\mathrm{u}}^{\mathrm{E}\mathrm{u}}$ are summarized in Table 3.

Table 3.

Observed and radiative lifetimes, and intrinsic quantum yield of EuL in the solid state at 77 K. Experimental values are given with an uncertainty of ±2σ.

Compound	τ obs a [ms]	τ rad b [ms]	${\Phi }_{\mathrm{E}\mathrm{u}}^{\mathrm{E}\mathrm{u}}$ c [%]
EuL in the solid state	0.30(1)	0.84(7)	36(3)

[a] λ _ex=337.1 nm. [b] The value of τ _rad was calculated using Equation (4). [c] ${\Phi }_{\mathrm{E}\mathrm{u}}^{\mathrm{E}\mathrm{u}}$ =τ _obs/τ _rad.

Finally, we measured the excitation spectra of EuL in the solid state with the emission of the ⁵D₀→⁷F₂ transition at room temperature and 77 K. Further information on the role of the LMCT state in the sensitization process was obtained from this measurement. At room temperature, the excitation spectrum of EuL showed a ligand‐centered excitation band with a shoulder assigned to the LMCT transition (Figure 9a). The intensities of these bands were weak and comparable with those of the Eu(III)‐centered transition. This observation shows that the sensitization by the ligand is ineffective in EuL at room temperature. No excitation bands corresponding to the transitions to the 4 f excited states at above the ⁵D₂ level were observed. This strongly suggests the presence of a quenching path. We attributed the quenching path in EuL to its LMCT state. At 77 K, the excitation spectrum showed ⁷F_0,1→⁵D₂, ⁷F₀→⁵D₃, and ⁷F₀→⁵L₆ excitation bands (Figure 9b), which were faint or absent in the excitation spectrum recorded at room temperature. According to previous reports,[ ^15b , ^15d ] 4 f excited states with energies above the LMCT energy level can be quenched via energy transfer to the LMCT state when the system has sufficient energy to overcome the activation barrier. The temperature dependence of the excitation spectra of EuL can be understood with the following consideration: decreasing the temperature suppresses the energy transfer from the ⁵D₂, ⁵D₃, and ⁵L₆ excited states to the LMCT state. Considering the above results, we assume that the LMCT state is located close to the ⁵D₂ level, and thereby near the T₁ state of the ligand (Figure 10). The occurrence of the T₁→LMCT energy transfer would suppress the T₁→⁵D₀ energy transfer, resulting in the low‐intensity Eu(III) luminescence for EuL. Overall, our data show that the LMCT state participates in the sensitization process in EuL.

Figure 9.

Excitation spectra of EuL in the solid state (λ _em=613 nm), recorded at room temperature (a) and 77 K (b). Both spectra were normalized at 580 nm.

Figure 10.

Proposed sensitization pathway for EuL.

3. Conclusions

We studied the photophysical properties of Ln(III) complexes with a tripodal Schiff base ligand (LnL, Ln=Eu, Gd, and Yb). Upon ligand excitation, YbL showed f‐f luminescence in the NIR region. Although the value of ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ for YbL is moderate among the reported values for Yb(III) complexes, it is significant that τ _rad is meaningfully lower than previously reported values. The sensitization mechanism for YbL was explained by the LMCT‐mediated process. The emission and excitation spectra of EuL indicate the participation of the LMCT state in the sensitization process for EuL. The τ _rad value for EuL is lower than those for most reported Eu(III) complexes, confirming that the tripodal Schiff base ligand enhanced the radiative rate constant for both Yb(III) and Eu(III). Our results suggest that Yb(III) complexes with tripodal Schiff base ligands are promising compounds for the design of highly NIR luminescent materials because of their extremely short τ _rad values. More detailed work is required to obtain further information on the sensitization mechanism and optimization of the η of the Yb(III) complex with the tripodal Schiff base ligands. In addition, we will attempt to minimize the deactivation path for the Yb(III) excited state to improve ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ .

Experimental Section

General Procedures

Europium(III) trifluoromethanesulfonate [Eu(CF₃SO₃)₃], Gd(III) trifluoromethanesulfonate [Gd(CF₃SO₃)₃], and Yb(III) trifluoromethanesulfonate [Yb(CF₃SO₃)₃] were purchased from Sigma Aldrich. Tris(2‐aminoethyl)amine (tren) and 5‐methylsalicylaldehyde were purchased from Tokyo Chemical Industry Co., Ltd. Elemental analysis was carried out using an NM‐10 system (J‐Science Lab Co., Ltd.). Powder X‐ray diffraction (PXRD) analysis was performed using a RINT‐2200VHF (Rigaku Corp.) instrument. The solid samples were gently ground using a mortar and pestle for the PXRD analysis.

Synthesis of EuL

A methanol solution (8 mL) containing 0.584 g (4 mmol) of tren was added to a hot methanol solution (40 mL) containing 1.198 g (2 mmol) of Eu(CF₃SO₃)₃. The resulting mixture was stirred for 10 min at ∼60 °C. A methanol solution (8 mL) containing 0.816 g (6 mmol) of 5‐methylsalicylaldehyde was subsequently added to the reaction mixture and stirred for 10 min at ∼60 °C. After cooling to room temperature, the product was collected by filtration and dried under reduced pressure. The crude product (0.593 g) was dissolved in 28 mL of hot DMF, and the resulting solution was filtered. The slow evaporation of DMF under reduced pressure resulted in the formation of pale‐yellow crystals. The crystals were collected by filtration and washed with methanol. After heating the obtained crystals to 100 °C under reduced pressure for 10 h, pale‐yellow crystals of EuL were obtained. Yield: 0.452 g (34.8 %). Elemental analysis calcd (%) for C₃₀H₃₃N₄O₃Eu: C 55.47, H 5.12, N 8.63; found: C 55.21, H 5.12, N 8.81.

Synthesis of GdL

The GdL complex was obtained using a procedure similar to the synthesis of EuL, with Gd(CF₃SO₃)₃ in place of Eu(CF₃SO₃)₃. Yield: 0.598 g (45.7 %). Elemental analysis calcd (%) for C₃₀H₃₃N₄O₃Gd: C 55.02, H 5.08, N 8.56; found: C 54.96, H 5.19, N 8.67.

Synthesis of YbL

A methanol solution (2 mL) containing 0.146 g (1 mmol) of tren was added to a hot methanol solution (13 mL) containing 0.310 g (0.5 mmol) of Yb(CF₃SO₃)₃. The resulting mixture was stirred for 10 min at ∼60 °C. A methanol solution (2 mL) containing 0.210 g (1.5 mmol) of 5‐methylsalicylaldehyde was subsequently added to the reaction mixture and stirred for 5 min at ∼60 °C. After cooling to room temperature, the product was collected by filtration and dried under reduced pressure. The crude product (0.286 g) was dissolved in 30 mL of hot DMF. To the resulting solution, 20 mL of methanol was added. The solution was cooled to 8 °C overnight, resulting in the precipitation of pale‐yellow crystals. The crystals were collected by filtration and washed with methanol. After heating the obtained crystals to 40 °C under reduced pressure for 16 h, pale‐yellow crystals of YbL were obtained. Yield: 0.160 g (47.9 %). Elemental analysis calcd (%) for C₃₀H₃₃N₄O₃Yb: C 53.73, H 4.93, N 8.36; found: C 53.73, H 4.86, N 8.54.

Optical Measurement

The absorption spectra were recorded using a UV‐1800 UV‐Vis spectrometer (SHIMADZU Corp.). The diffuse reflectance spectra were recorded using a V‐570 spectrophotometer (JASCO Corp.) equipped with an ISN‐470 integrating‐sphere (JASCO Corp.). The NIR emission and excitation spectra were recorded on a Fluorolog‐3 spectrometer (HORIBA Jobin Yvon Inc.). The value of ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ for YbL in acetonitrile was determined by a relative method with Yb(TTA)₃phen in toluene as the standard. Yb(TTA)₃phen was synthesized according to a previously reported procedure. ^[30] The value of ${\Phi }_{\mathrm{Y}\mathrm{b}}^{\mathrm{L}}$ for YbL was estimated using Equation 5:

$${\displaystyle {\Phi }_{\mathrm{x}}={\Phi }_{\mathrm{s}}\times \left(\frac{G_x}{G_s}\right)\times \left(\frac{{n_x}^2}{{n_s}^2}\right)}$$ (5)

where subscripts x and s denote the sample and the standard, respectively; $\Phi$ , the quantum yield; G, the slope from the plot of the integrated emission intensity vs. the amount of light absorbed at the excitation wavelength; and n, the refractive index. The visible emission and excitation spectra were recorded on an F‐7000 fluorescence spectrophotometer (Hitachi High‐Tech Corp.). The NIR luminescence lifetime was measured using an LSP‐1000 (UNISOKU Co., Ltd.), in which a pulsed nitrogen laser (337.1 nm, 3.5 ns) and an InGaAs photodiode were used as the light source and detector, respectively. In this measurement, the NIR luminescence was separated using an 800 nm long‐pass filter. The visible luminescence lifetime was measured using F‐7000. The luminescence decay profiles were fitted to mono‐exponential decay by the non‐linear least‐squares method. The lifetime and quantum yield are shown as averages of at least three independent measurements with an uncertainty of ±2σ. The powders that were ground gently using a mortar and pestle were subjected to solid‐state photophysical measurements.

Single‐Crystal XRD

The crystallographic data were collected on a Bruker AXS II CCD diffractometer with Mo−Kα radiation (λ=0.71073 Å), and they are shown in Table S1. The structure was solved using SHELXT‐2014 ^[31] and refined with SHELXL‐2017. ^[32] The non‐H atoms were refined anisotropically. All the H atoms were placed in a geometrically idealized position.

Deposition Numbers 1887108 and 1887109 contain the supplementary crystallographic data for this study. These data are provided free of charge by the joint Cambridge Crystallographic Data Centre and Fachinformationszentrum Karlsruhe Access Structures service www.ccdc.cam.ac.uk/structures.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

A. Masuya-Suzuki, S. Goto, T. Kambe, R. Karashimada, Y. Kubota, N. Iki, ChemistryOpen 2021, 10, 46.