Perfluorinated Aromatic Spacers for Sensitizing Eu(III) in Dinuclear Oligomers: Better than the Best by Chemical Design?

The unique optical characteristics of lanthanide ions (Ln^III) have driven their use in a wide range of applications, however the efficiency of populating directly their electronic states is limited so that the creation of an antenna to capture energy and generate excited Ln^III ions is broadly important. In this context, metal-organic frameworks, hybrid materials, and nanoparticles randomly doped with homo- or heterometallic mixtures of luminescent lanthanide cations, Ln^III, are intensively being investigated for engineering luminescent devices for bright white lighting, for upconversion, and as sensing agents.^[1] Although the exact location of the various metals in the final material is crucial for dual ligand-centered/metal-centered emission,^[1l,m] for upconversion^[1g] and for directional light-conversion^[2a] processes, the preparation of organized polymetallic 4f-4f oligomers and polymers remains rare and challenging.^[2] A statistical mechanics (Ising model) analysis suggests that standard repulsive nearest neighbor intermetallic interactions operating in linear polymers with regularly spaced binding sites should provide the targeted ordered …-Ln¹-Ln²-Ln¹-Ln²-… microstates.^[3,4] Pioneering work in this field has relied on the bulk electropolymerization of didentate 1,10-phenanthroline with thienyl spacers^[5] and the acyclic diene metathesis of tridentate 2,6-bis(benzimidazol-2-yl)pyridine,^[6] followed by reaction with Eu(β-diketonate)₃ or Eu(NO₃)₃ to yield red-emitting metallopolymers. A reliable exploitation of this concept for the development of luminescent materials, however, requires the efficient sensitization of the luminophore via the rational optimization of each photophysical step by using chemical tools.

As a first step toward this goal, the rigid segmental ligand strands L1–L3, made of two tridentate binding units separated by a rigid and electronically-tunable aromatic spacer, have been reacted with trivalent europium to give the dinuclear complexes [Eu₂(Lk)(hfac)₆] (Scheme 1).^[7] The use of a simple method for deciphering the various contributions to the sensitization mechanism clearly showed that [Eu₂(L3)(hfac)₆] displayed the largest global emission quantum yield ( ${\mathrm{\Phi }}_{\text{Eu}}^{\text{L}}=0.206(7)$, eq. 1) because of an efficient L3→Eu energy transfer step ( ${\eta }_{\text{en}.\text{tr}.}^{\text{L}\to \text{Eu}}=0.47(14)$ eq. 3, see the dark grey histograms in Fig. 1).^[7]

Scheme 1.

Chemical structures of complexes [Eu₂(Lk)(hfac)₆]

Figure 1.

This simplified Jablonski diagram for [Eu₂(Lk)(hfac)₆] (k = 2–4) shows the ligand-centered triplet-mediated sensitization mechanism of the two Eu³⁺ ions.^[10] The photophysical processes are described by the rate constants: $k_{\text{r}}^{\text{F}}$ for ligand fluorescence, $k_{\text{nr}}^{\text{F}}$ for ligand internal non-radiative conversion, $k_{\text{r}}^{\text{P}}$ for ligand phosphorescence, $k_{\text{nr}}^{\text{P}}$ for non-radiative relaxation from the ligand triplet state, $k_{\text{r}}^{\text{Eu}}$ for emission of Eu^*, $k_{\text{nr}}^{\text{Eu}}$ for nonradiative decay of Eu^*, k_ISC for ligand intersystem crossing, and $k_{\text{en}.\text{tr}.}^{\text{Eu}}$ for the ligand-to-metal energy transfer. Efficiencies of intersystem crossing (η_ISC), energy transfer ( ${\eta }_{\text{en}.\text{tr}.}^{\text{L}\to \text{Eu}}$), intrinsic quantum yield ( ${\mathrm{\Phi }}_{\text{Eu}}^{\text{Eu}}$), and quantum yield ( ${\mathrm{\Phi }}_{\text{Eu}}^{\text{L}}$) for the global ligand-mediated sensitization of Eu^III in [Eu₂(Lk)(hfac)₆] (solid-state, 293 K).

Theoretical considerations suggest that ${\eta }_{\text{en}.\text{tr}.}^{\text{L}\to \text{Eu}}$ could benefit from a shift of the L(³π^*) state to higher energy produced by the perfluorination of the central aromatic spacer in L4 (Scheme 2).^[8] Correspondingly, the expected decrease of $k_{\text{nr}}^{\text{F}}$ and $k_{\text{nr}}^{\text{Eu}}$, and of the so-called π-conjugation length A_π (eq. 5)^[9] in [Eu₂(L4)(hfac)₆],^[7] should optimize both intersystem crossing efficiency (η_ISC = 0.6(1), eq. 2) and intrinsic Eu-centered quantum yield ( ${\mathrm{\Phi }}_{\text{Eu}}^{\text{Eu}}=0.76(2)$, eq. 4) that were previously measured in [Eu₂(L3)(hfac)₆] (dark grey histograms in Fig. 1). Note that

Scheme 2.

Synthesis of the perfluorinated ligand L4.

$${\mathrm{\Phi }}_{\text{Eu}}^{\text{L}}={\eta }_{\text{ISC}}·{\eta }_{\text{en}.\text{tr}.}^{\text{L}\to \text{Eu}}·{\mathrm{\Phi }}_{\text{Eu}}^{\text{Eu}}$$ (1)

$${\eta }_{\text{ISC}}=\frac{k_{\text{ISC}}}{k_{\text{r}}^{\text{F}}+k_{\text{nr}}^{\text{F}}+k_{\text{ISC}}}=k_{\text{ISC}}{·}_{\text{L}}\left({{}^1\pi }^{\ast }\right)$$ (2)

$${\eta }_{\text{en}.\text{tr}.}^{\text{L}\to \text{Eu}}=\frac{2k_{\text{en}.\text{tr}.}^{\text{Eu}}}{k_{\text{r}}^{\text{P}}+k_{\text{nr}}^{\text{P}}+2k_{\text{en}.\text{tr}.}^{\text{Eu}}}=2k_{\text{en}.\text{tr}.}^{\text{Eu}}·{\tau }_{\text{L}}^{\text{Eu}}\left({{}^3\pi }^{\ast }\right)$$ (3)

$${\mathrm{\Phi }}_{\text{Eu}}^{\text{Eu}}=\frac{k_{\text{r}}^{\text{Eu}}}{k_{\text{r}}^{\text{Eu}}+k_{\text{nr}}^{\text{Eu}}}=k_{\text{r}}^{\text{Eu}}·{\tau }_{\text{Eu}}$$ (4)

and

$$k_{\text{r}}^{\text{F}}=k_{\text{nr}}^{\text{F}}·e^{A_{\pi }}$$ (5)

The centrosymmetrical perfluorinated ligand L4 is obtained through two successive Pd-catalyzed Suzuki-Myaura cross coupling reactions (Scheme 2, Figs. S1–S2). Stoichiometric mixing of L4 with [Ln(hfac)₃(diglyme)] (2.0 eq) in chloroform gives [Ln₂(L4)(hfac)₆] (Ln = Gd, Eu) with a 80% yield. Slow evaporation from concentrated acetonitrile/chloroform solutions provides X-ray quality prisms (Table S1) containing S-shaped centrosymmetrical neutral [Eu₂(L4)(hfac)₆] complexes, in which the Eu atoms are separated by 14.667(1) Å (Fig. 2). Each metal is nine-coordinate in a highly distorted monocapped square antiprismatic polyhedron, produced by the three nitrogen atoms of the bound tridentate aromatic unit and by the six oxygen atoms of the three didentate hexafluoroacetylacetonates, N1 occupying the capping position (see Appendix 1). All bond distances and bond angles are standard (Tables S2–S4) and the solid-state molecular structures of [Yb₂(L3)(hfac)₆] and [Eu₂(L4)(hfac)₆] are almost superimposable, except for the interannular phenyl-benzimidazole twist angle, which increases from 54.1(1)° to 66.2(1)° (Fig. S3 and Table S5).

Figure 2.

Perspective view of the molecular structure of [Eu₂(L4)(hfac)₆] as obtained from X-ray diffraction with numbering scheme. Thermal ellipsoids are represented at the 30% probability level and hydrogen atoms are omitted for clarity.

Irradiation into the allowed ligand-centered ¹π^*←¹π transition of [Eu₂(L4)(hfac)₆] at ν̄_exc = 28170 cm⁻¹ produces an intense long-lived red emission signal, arising from L4→Eu(III) energy transfer followed by Eu(⁵D₁) and Eu(⁵D₀)-centered luminescence (Fig. S4). The emission spectrum is dominated by the hypersensitive forced electric dipolar Eu(⁵D₀→⁷F₂) transition centered at 16340 cm⁻¹ leading to the largest global absolute quantum yields in this series ( ${\mathrm{\Phi }}_{\text{Eu}}^{\text{L}}=0.26(1)$, solid state, 293K, Table 1, entry 6).^[11] Using Einstein’s result for the spontaneous radiative emission rate,^[12] the radiative rate constant $k_{\text{r}}^{\text{Eu}}$ (Table 1, entry 2) is deduced from the I_tot/I_MD ratio, where I_tot is the integrated emission for the Eu(⁵D₀) level (⁵D₀ → ⁷F_J, J = 0 – 4) and I_MD is the integrated intensity of the magnetic dipolar Eu(⁵D₀→⁷F₁) transition (Table 1, entry 1). Combined with the characteristic lifetime τ_Eu = 0.90(1) ms (Table 1, entry 3), we calculate ${\mathrm{\Phi }}_{\text{Eu}}^{\text{Eu}}=0.77(1)$ for the intrinsic Eu-centered quantum yield (eq. 4, Table 1, entry 4), a value identical to that obtained for [Eu₂(L3)(hfac)₆], which implies that the gain in global quantum yields can be specifically assigned to the improved sensitization process ${\eta }_{\text{sens}}={\eta }_{\text{ISC}}·{\eta }_{\text{en}.\text{tr}.}^{\text{L}\to \text{Eu}}={\mathrm{\Phi }}_{\text{Eu}}^{\text{L}}/{\mathrm{\Phi }}_{\text{Eu}}^{\text{Eu}}=0.34(1)$ operating in [Eu₂(L4)(hfac)₆] (eq. 1, Table 1, entry 7). Given the experimental ligand-centered fluorescence lifetimes of only 30–60 ps that are measured for the [Gd₂(Lk)(hfac)₆] complexes (Lk = L1–L4, Table S6), one can assume that the energy migration processes in [Eu₂(Lk)(hfac)₆] is well described by the exclusive contribution of the triplet state as shown in Fig. 1.^[7,13] Considering that (i) the sum of the radiative and internal non-radiative conversion rate constants $k_{\text{r}}^{\text{F}}+k_{\text{nr}}^{\text{F}}$ controlling the relaxation of the ¹π^* excited state in the free ligand are the same in the gadolinium complex [Gd₂(L4)(hfac)₆] and that (ii) $k_{\text{ISC}}^{\mathbf{L}\mathbf{4}}\ll k_{\text{ISC}}^{\text{Gd}-\mathbf{L}\mathbf{4}}$ because of the paramagnetic and heavy atom effects generated by Gd(III),^[14] the introduction of the experimental characteristic ¹π^* lifetimes measured in L4 ( ${\tau }_{\text{L}}^{\mathbf{L}\mathbf{4}}\left({{}^1\pi }^{\ast }\right)=0.25(2)\text{ns}$) and in [Gd₂(L4)(hfac)₆] ( ${\tau }_{\text{L}}^{\text{Gd}-\mathbf{L}\mathbf{4}}\left({{}^1\pi }^{\ast }\right)=0.029(6)\text{ns}$, Table S6) into eq. 6 gives $k_{\text{ISC}}^{\text{Gd}-\mathbf{L}\mathbf{4}}=30(4){\text{ns}}^{-1}$ (Table 1, entry 8) and $k_{\text{r}}^{\text{F}}+k_{\text{nr}}^{\text{F}}=4.00(5){\text{ns}}^{-1}$ (Table 1, entry 9), from which η_ISC = 0.88(15) can be deduced with eq. 2 (Table 1, entry 10).^[7]

Table 1.

Experimental global ( ${\mathrm{\Phi }}_{\text{Eu}}^{\text{L}}$) and intrinsic ( ${\mathrm{\Phi }}_{\text{Eu}}^{\text{Eu}}$) quantum yields, luminescence lifetimes (τ_Eu) and calculated energy migration efficiencies (η_ISC, ${\eta }_{\text{en}.\text{tr}.}^{\text{L}\to \text{Eu}}$) and rate constants ( $k_{\text{r}}^{\text{Eu}},k_{\text{nr}}^{\text{Eu}},k_{\text{en}.\text{tr}.}^{\text{Eu}},k_{\text{r}}^{\text{P}}+k_{\text{nr}}^{\text{P}}$, k_ISC) for [Eu₂(Lk)(hfac)₆] in the solid state at 293 K (k = 2–4).^[10]

Compound	[Eu2(L2)(hfac)6]	[Eu2(L3)(hfac)6]	[Eu2(L4)(hfac)6]
Eu-centered luminescence
Itot/IMD	17.5(3)	18.2(3)	17.2(2)
$k_{\text{r}}^{\text{Eu}}/{\text{ms}}^{-1}$	0.86(2)	0.90(2)	0.85(1)
τEu/ms	0.88(4)	0.83(15)	0.90(1)
${\mathrm{\Phi }}_{\text{Eu}}^{\text{Eu}}$	0.76(4)	0.75(1)	0.77(1)
$k_{\text{nr}}^{\text{Eu}}/{\text{ms}}^{-1}$	0.21(1)	0.23(4)	0.200(4)
Global quantum yield and sensitization efficiency
${\mathrm{\Phi }}_{\text{Eu}}^{\text{L}}$	0.092(3)	0.206(7)	0.26(1)
${\eta }_{\text{ISC}}·{\eta }_{\text{en}.\text{tr}.}^{\text{L}\to \text{Eu}}$	0.122(7)	0.28(5)	0.34(1)
Energy migration and associated rate constants
$k_{\text{ISC}}^{\text{Gd}-\mathbf{L}}/{\text{ns}}^{-1}$	10.0(6)	9(2)	30(4)
$k_{\text{r}}^{\text{F}}+k_{\text{nr}}^{\text{F}}/{\text{ns}}^{-1}$	6.71(5)	6.25(3)	4.00(5)
η ISC	0.60(3)	0.59(12)	0.88(15)
${\eta }_{\text{en}.\text{tr}.}^{\text{L}\to \text{Eu}}$	0.20(2)	0.47(14)	0.39(6)
$2k_{\text{en}.\text{tr}.}^{\text{Eu}}/{\text{ms}}^{-1}$	2.1(3)	5.5(2.3)	5.7(2.1)
$k_{\text{r}}^{\text{P}}+k_{\text{nr}}^{\text{P}}/{\text{ms}}^{-1}$	8.1(5)	6.3(8)	9.1(2.5)
Reference	[7]	[7]	This work

$$\frac{1}{{\tau }_{\text{L}}^{\text{Gd}-\mathbf{L}}\left({{}^1\pi }^{\ast }\right)}-\frac{1}{{\tau }_{\text{L}}^{\mathbf{L}}\left({\pi }^{\ast }\right)}=\left(k_{\text{r}}^{\text{F}}+k_{\text{nr}}^{\text{F}}+k_{\text{ISC}}^{\text{Gd}-\mathbf{L}}\right)-\left(k_{\text{r}}^{\text{F}}+k_{\text{nr}}^{\text{F}}+k_{\text{ISC}}^{\mathbf{L}}\right)=k_{\text{ISC}}^{\text{Gd}-\mathbf{L}}-k_{\text{ISC}}^{\mathbf{L}}\approx k_{\text{ISC}}^{\text{Gd}-\mathbf{L}}$$ (6)

Finally, the energy transfer efficiency ${\eta }_{\text{en}.\text{tr}.}^{\text{L}\to \text{Eu}}={\eta }_{\text{sens}}/{\eta }_{\text{ISC}}=0.39(6)$ and the associated rate constant $2k_{\text{en}.\text{tr}.}^{\text{Eu}}=\left[{\eta }_{\text{en}.\text{tr}.}^{\text{L}\to \text{Eu}}/\left(1-{\eta }_{\text{en}.\text{tr}.}^{\text{L}\to \text{Eu}}\right)\right]\left(k_{\text{r}}^{\text{P}}+k_{\text{nr}}^{\text{P}}\right)=5.7(2.1){\text{ms}}^{-1}$ (eq. 3 with $\left(k_{\text{r}}^{\text{P}}+k_{\text{nr}}^{\text{P}}\right)=1/{\tau }_{\text{L}}^{\text{Gd}-\mathbf{L}\mathbf{4}}\left({{}^3\pi }^{\ast }\right)=9.1(2.5){\text{ms}}^{-1}$) calculated for [Eu₂(L4)(hfac)₆] (Table 1, entries 11 and 12) indicates no noticeable improvement of these parameters on changing from the difluorinated (L3) to the perfluorinated (L4) spacer, despite the 500–1000 cm⁻¹ energy blue shift of the ligand-centered ¹π^* (Fig. S5) and ³π^* (Fig. S6) excited states. Our simple method for dissecting the sensitization mechanism^[7] shows that the gain in the global quantum yield ${\mathrm{\Phi }}_{\text{Eu}}^{\text{L}}$, for the complex [Eu₂(L4)(hfac)₆] as compared to [Eu₂(L3)(hfac)₆], indeed results from an optimization of the intersystem crossing process η_ISC (Fig. 1).

Scrutiny of the various rate constants (Tables 1 and S7) reveals that the decrease of the radiative and internal conversion rate constant $k_{\text{r}}^{\text{F}}+k_{\text{nr}}^{\text{F}}$ for the ligand-centered L(¹π^*) state along the series L2 > L3 > L4 acts to improve η_ISC (eq. 2), but it is the remarkable increase of $k_{\text{ISC}}^{\text{Ln}-\mathbf{L}}$ in [Ln₂(L4)(hfac)₆], which eventually controls the overall intersystem crossing efficiency.

The physical origin of this beneficial effect can be traced back to the golden-rule expression for radiationless transitions:^[15]

$$k_{\text{ISC}}=\frac{2\pi }{\hslash }{\langle {{}^1\pi }^{\ast }\left|H_{\text{SO}}\right|{{}^3\pi }^{\ast }\rangle }^2\times \mathit{FCWDS}$$ (7)

where FCWDS is the Franck-Condon Weighted Density of States. It accounts for the density of vibrational states in the triplet and their vibrational overlap with the singlet vibrational state. A model that accounts for the thermal population of levels and uses a single quantum mode of frequency ω, is commonly associated with the Marcus-Levich-Jortner theory for electron transfer expressed in eq. 8:^[16]

$$\mathit{FCWDS}=\frac{exp(-S)}{\sqrt{4\pi RT}}\sum _{n=0}^{\infty }\frac{S^n}{n!}exp\left[-\frac{{(\mathrm{\Delta }E+n\hslash \omega +\lambda )}^2}{4\lambda RT}\right]$$ (8)

The spin-orbit coupling matrix element 〈¹π^* |H_SO| ³π^*〉 reaches a maximum for non-planar polyaromatic molecules containing heavy paramagnetic atoms in the molecular frame;^[16,17] two conditions which are fulfilled by all [Eu₂(Lk)(hfac)₆] complexes described in this study. We however note that the deviation from the planarity, as measured by the interplanar phenyl-benzimidazole angles, increases along the series L2 (25.26(4)°) < L3 (54.1(1)°) and L4 (66.2(1)°) in line with k_ISC (eq. 7 and Table 1). The Franck-Condon weighted density of states (FCWDS) depends on the singlet-triplet energy splitting ΔE = E(¹π^*)−E(³π^*) and on the reorganization energy λ, which corresponds to the energy difference between the triplet and the singlet state at its equilibrium geometry (eq. 8).^[18] In the limit of parabolic surfaces, this energy parameter, along with ΔE, provides the energy gap for n = 0. The successive fluorination of the ligands along the series [Eu₂(Lk)(hfac)₆] (k = 2–4) is known to significantly affect the frontier orbitals, hence λ and ΔE and FCWDS in eq. 8.^[7] While a quantitative understanding of the changes in λ requires sophisticated theoretical calculations of the vibrational coupling scheme controlling the Huang-Rhys factors (S),^[15,16] eqs (7)–(8) predict that the increasing ligand-centered energy gap ΔE observed along the series L2 (3550 cm⁻¹) ≈ L3 (3230 cm⁻¹) < L4 (5200 cm⁻¹) should lower k_ISC and η _ISC. The apparent contradiction with our experimental results (Table 1, entries 8 and 10) can be resolved by including higher lying triplet states ${{}^3\pi }_{n>1}^{\ast }$ in the model; an approach used successfully for oligothiophenes^[15] and helicenes.^[15b] This counter-intuitive correlation was empirically noticed for other polyaromatic chromophores, and it was suggested as a ‘rule-of-thumb’ that a singlet-triplet gap of E(¹π^*)−E(³π^*) ≥ 5000 cm⁻¹ warrants quantitative intersystem crossing processes in Tb- and Eu-complexes.^[19] Given that the singlet-triplet energy gap can be readily calculated by using DFT^[7] or semi-empirical^[8] methods, computations may be useful for identifying simple chemical and structural modifications that will enhance the quantum efficiency further.

In conclusion, the application of this simple method for analyzing the various contributions to Eu^III luminescence sensitization shows that perfluorination of the remote phenyl spacer in the rigid single-stranded dumbbell-shaped [Eu₂(L4)(hfac)₆] oligomer optimizes both intersystem crossing efficiency and intrinsic Eu^III quantum yield, thus maximizing the global quantum yields in these polyaromatic rigid complexes (red histograms in Fig. 1).