Reduced quenching effect of pyridine ligands in highly luminescent Ln(III) complexes: the role of tertiary amide linkers

Abstract

Luminescent Eu(III) and Tb(III) complexes were synthesised from octadentate ligands carrying various carbostyril sensitizing antennae and two bidentate picolinate donors. Antennae were connected to the metal binding site via tertiary amide linkers. Antennae and donors were assembled on a 1,4,7-triazacyclononane (tacn) platform. Solution- and solid-state structures were comparable to those of previously reported complexes with tacn architectures, with nine-coordinate distorted tricapped trigonal prismatic Ln(III) centres, and distinct from those based on 1,4,7,10-tetraazacyclododecane (cyclen) macrocycles. In contrast, the photophysical properties of these tertiary amide tacn-based complexes were more comparable to those of previously reported systems with cyclen ligands, showing efficient Eu(III) and Tb(III) luminescence. This represents an improvement over secondary amide-linked analogues, and is due to a greatly increased sensitization efficiency in the tertiary amide-linked complexes. Tertiary amide-linked Eu(III) and Tb(III) emitters were more photostable than their secondary amide-linked analogues due to the suppression of photoinduced electron transfer and back energy transfer.

---

Introduction

The luminescent properties of trivalent lanthanide ions (Ln) have long attracted attention due to their applications in lasers and lighting, encryption, barcoding, and as biological probes. Unlike organic fluorophores that often suffer from short emission lifetimes, broad emission bands, and small Stokes shifts, Ln luminescence is due to f–f transitions, and is therefore atomic-like, unique to the metal, and long-lived.^1,2 The inefficient absorption of Ln(III) centres can be overcome by allowing a light-harvesting antenna to absorb the excitation energy and transfer it to the metal ion. In addition to a high apparent absorption coefficient, such a strategy also yields a large apparent Stokes shift. A dazzling variety of antennae have been reported that encompass amino acids,^3,4 both small and complex heterocycles^5–21 transition metal complexes,^22–29 and even other Lns.^3,30

A crucial component of an emissive Ln coordination compound is the ligand. Stable complexes require the saturation of most of the 8–10 coordination sites of the Ln(III).³¹ Hepta-, octa-, and nonadentate ligands can be constructed from tetra- and triazamacrocyclic (cyclen and tacn, respectively, Fig. 1) building blocks upon the addition of mono- and bidentate donors, e.g. carboxylates, carbamides, or picolinates.³² While the primary role of the ligand is to create a well-defined, stable emitter, it has additional functions. Ln(III) luminescence is sensitive to nearby X–H oscillators.^3,33 While X can be O, N or, in the case of near infrared emitters, even C, the largest quenching effect is due to inner-sphere O–H-containing solvent molecules, such as water or MeOH.^34,35 A high-denticity ligand can displace most inner-sphere solvent molecules, and thus improve Ln(III) luminescence.

Fig. 1 (a) Previously reported tertiary amide-linked complexes with carbostyril antennae,³⁹ including cyclen-based benchmarks LnLc^t,R.⁴² (b) Secondary amide-linked carbostyril-sensitised LnLt^s,R consisting of a tacn ligand framework equipped with substituted picolinic acid donors.³⁸ (c) LnLt^t,R complexes studied here.

A further role of the ligand is to impose a coordination geometry. The shielded 4f-orbitals do not participate in directional bonding, and Lns bind preferentially hard donor atoms (such as O, F) through coulombic interactions. The ligand can thus determine the symmetry around the Ln(III), which in turn will govern the radiative lifetime (τ_rad) of the ion. The overall luminescence quantum yield (Φ_Ln) is the product of the efficiency with which the Ln(III) excited state is populated (η_sens) via absorption, intersystem crossing, energy transfer, etc., and the intrinsic quantum yield (Φ^Ln_Ln) of the ion (eqn (1)). The latter is equal to the proportion of observed (τ_obs) and radiative lifetimes.^36,37 All else being equal, shorter τ_rad yields higher Φ_Ln.

We have previously prepared both tacn and cyclen-based complexes with a variety of carbostyril and coumarin antennae (Fig. 1). Notably, the Eu(III) complexes of the octadentate tacn-based ligands (LnLt^s,R) had much shorter τ_rad (∼3 ms)³⁸ than those of their similarly octadentate cyclen-based analogues (e.g.LnLc^s,R, τ_rad ∼ 5 ms).^39–42

Ligands can control quenching processes other than X–H quenching. Photoinduced electron transfer (PeT) whereby the excited antenna reduces Ln(III) to Ln(II) is common in emitters with reducible Ln(III), such as Eu.^43,44 The re-oxidation of Eu(II) usually yields a quenched complex, although exceptions are known.^45,46 The detrimental effects of PeT in some systems is comparable to that of X–H oscillators.⁴⁰ Ligands that stabilise the Eu(III) oxidation state can increase the luminescence quantum yield.⁴⁰ The ligand may also directly quench the excited antenna.³⁹ This is the case in the picolinate-equipped LnLt^s,R (Fig. 1), where the electron-poor pyridines are reduced alongside Eu(III) by the photoexcited antenna. PeT to the pyridines was possible even in the non redox-active Tb(III) and Gd(III), and was more prominent for electron-poor pyridines.³⁸

Here, we report a series of tacn-based ligands and their Ln(III) complexes equipped with the same picolinate ligands that were previously shown to be detrimental to Ln(III) luminescence (Fig. 1). We show that changing the linker that connects the carbostyril antenna to the tacn fragment from a secondary (LnLt^s,R) to a tertiary amide (LnLt^t,R) can result in the recovery of Ln(III) luminescence, and afford emitters that are comparable to what was seen in cyclen-based systems. The complexes were structurally characterised by paramagnetic ¹H NMR spectroscopy and X-ray crystallography to enable comparison of the Ln(III) coordination environments in the secondary and tertiary amide-linked complexes. Steady-state and time-resolved luminescence spectroscopy was used to understand the sensitization and quenching pathways in these emitters, and revealed a significantly improved sensitization in the new set of picolinate-containing species compared to the previously reported ones. These results indicate that even efficient quenching processes can be interrupted by judiciously chosen structural changes.

Results and discussion

Synthetic procedures

The ligands Lt^t,R were synthesised from 4-methyl-, 4-methoxymethyl- or 4-trifluoromethyl-substituted 7-aminocarbostyrils 1a, 1b and 1c, via the tertiary chloroacetylamides 2a, 2b, and 2c, respectively (Scheme 1). The latter were prepared similarly to their tert-butyl ester analogues.³⁸ Di-substituted triazanonane 3 was monoalkylated with 2a–c to yield protected ligands 4a–c, which were deprotected under basic conditions. Heating Lt^t,R with LnCl₃ in a warm EtOH : H₂O mixture yielded LnLt^t,R after column chromatography on silica gel using ⁱPrOH : H₂O as the eluent in quantitative yield as white solids. Detailed experimental procedures and full chemical characterization for all new compounds are given in the ESI.†

Scheme 1 Preparation of LnLt^t,R and model compounds 5a–c.

X-ray crystallography

Single crystals suitable for X-ray diffraction analysis were obtained by vapor diffusion of glyme into concentrated aqueous solutions of GdLt^t,Me and TbLt^t,Me. The presence of the carboxylate group on the amide functionality of the ligand results in a coordination motif not seen for related complexes containing secondary amides.³⁸ The carboxylate group acts as a bridging ligand between LnLt molecules, forming 1D-polymeric chains. In addition, a secondary Ln centre is coordinated to a pyridine carboxylate C=O bond, acting as a bridge between 1D-polymeric chains resulting in a 2D-polymeric network (Fig. 2).

Fig. 2 Solid-state structures of TbLt^t,Me. H atoms, non-coordinating Cl⁻ counterions and water molecules omitted, triazacyclonane C atoms and antenna displayed as capped sticks and Ln2 co-ligands displayed as ball and stick for clarity. Ellipsoids displayed at 35% probability. Only major component of disordered sites displayed.

The Ln centre within the macrocylic cavity exhibits a nine-coordinate distorted tricapped trigonal prismatic geometry, comparable to that observed in related complexes. The trigonal prism is represented by three tacn N-donor atoms (N_3PL), and two pyridine N- and the antenna amide O-atoms (NNO_PL). The pyridine carboxylate groups and the bridging carboxylate arm of the tertiary amide serve as the capping ligands of the trigonal prism. The angle between the two planes is 115.1(3)° in Gd and 116.8(6)° in Tb, which compares well to those of related complexes. The deviation of the angle from 180° can be ascribed to the highly distorted geometry of the metal centre. The distance between the Ln centres and the NNO_PL are comparable to those of the corresponding secondary amide complexes (0.312(3) and 0.328(6) Å for Gd and Tb, respectively, versus average 0.312 Å), as are the distances to the N_3PL (Gd: 2.044(3) Å; Tb: 2.024(8) Å; versus 2.046 Å average). These metrics suggest that the tertiary amide group and the carboxylate bridging ligand have no significant impact on the direct coordination sphere of carbostyril-substituted tacn systems. The complexes are racemic in the solid-state with both Δ and Λ isomers present in the unit cell. The Ln–O (Gd: 2.388(4) Å; Tb: 2.376(6) Å), tacn Ln–N (Gd: 2.646(4) Å; Tb: 2.64(1) Å) and pyridine Ln–N (Gd: 2.548(6) Å; Tb: 2.538(8) Å) bond distances compare well to those of related complexes.

The secondary Ln centre is eight-coordinate and exhibits a square antiprismatic arrangement. Two picolinate carbonyl groups occupy flanking positions, while the remaining six sites are occupied by water ligands, and the charge balancing chloride counter-ions are non-coordinated and reside in the crystal lattice. The metal centre at this position is partially occupied (75%), which results in considerably disordered lanthanide bound ligands and imprecise bond lengths, which are not discussed here (see ESI† for further information).

NMR studies

Paramagnetic ¹H and ¹⁹F NMR spectroscopic analysis of EuLt^s and EuLt^t was performed in D₂O. Complexes with identical metal-binding sites had similar spectra (Fig. 3). At r.t. the signals were broad and indicative of the presence of several species that made the comparison of the data difficult. However, good quality spectra were obtained in all cases upon heating the samples at 80 °C. At high temperature the ¹H NMR spectra of the EuLt^s species contained ∼15 signals, consistent with the presence of a single species or of several species rapidly interconverting on the NMR timescale (Fig. S3–18†). The spectra of EuLt^t were much sharper than those of the corresponding EuLt^s even at room temperature, and further improved upon heating. The 80 °C-spectra of EuLt^t contained two sets of signals, which suggests that there are two stable conformers in the solution even at high temperatures. These conclusions were supported by the ¹⁹F NMR analysis of EuLt^s,CF3 and EuLt^t,CF3. At r.t. a large number of signals were observed for EuLt^s,CF3, and a broadened peak consisting of several peaks for EuLt^t,CF3. In line with the ¹H NMR data, at 80 °C the signals collapsed into a single peak (−61.9 ppm), while that of EuLt^t,CF3 into one major and two minor ones (around −62.7 ppm) (Fig. S11–18†).

Fig. 3 Variable temperature ¹H NMR spectra recorded in D₂O. (a) EuLt^s,CF3 at 353 K. (b) EuLt^s,CF3 at 298 K. (c) EuLt^t,CF3 at 353 K. (d) EuLt^t,CF3 at 298 K.

Photophysical studies

The photophysical properties of LnLt^t,R were measured under the same conditions as those of LnLt^s,R and LnLc^t,R,^38,42 in PIPES-buffered pH = 6.5 aqueous solution at r. t. The LnLt^t,R absorption spectra contained peaks in the 250–310 nm region that were absent from the spectra of LnLc^t,R, but were present in LnLt^s,R (Fig. 4, S22–27†). These peaks were assigned to pyridine π–π* transitions. The longest-wavelength absorption was determined by the antenna (Table 1), and the position of the maxima increased in the order LnLt^t,Me (λ_abs = 325 nm) < LnLt^t,MOM (λ_abs = 328 nm) < LnLt^t,CF3 (λ_abs = 338 nm). Upon antenna excitation weak antenna fluorescence was observed at λ_em = 366–391 nm; longer emission wavelengths were seen for the more red-absorbing antennae. While λ_abs for LnLt^t,R were at wavelengths that were 3–4 nm shorter than seen for the corresponding LnLt^s,R, this small difference all but disappeared in the emission spectra.

Fig. 4 Normalised absorption (left, grey). Excitation [left, blue, λ_em = 405 nm (Gd), λ_em = 542 nm (Tb), λ_em = 618 nm (Eu), 298 K], steady-state emission (middle, black, λ_ex = 325 nm, 298 K), steady-state emission [middle, purple, λ_ex = 325 nm (Gd), 77 K] and time-resolved emission (right, green (Tb), red (Eu), λ_ex = 325 nm) spectra of LnLt^t,Me complexes.

Table 1 Antenna photophysical properties of LnLt^s,R and LnLt^t,Ra

Compound	λ max [nm]	λ em [nm]	E 00(S1)^c [cm^−1]	E 00(T1)^c [cm^−1]
a In aqueous PIPES buffer (10 mM), pH 6.5, at 10 μM complex concentrations. b λ ex = 329 nm (GdLt^s,Me), 325 nm (GdLt^t,Me), 335 nm (GdLt^s,MOM), 328 nm (GdLt^t,MOM), 331 nm (GdLt^s,CF3), 330 nm (GdLt^t,CF3). c Calculated from the 0–0 transitions of the Gd-complexes recorded at 77 K.
GdLt^s,Me	329	366	28 900	22 900
GdLt^t,Me	325	365	29 200	23 000
GdLt^s,MOM	331	376	28 500	22 500
GdLt^t,MOM	328	375	28 900	22 800
GdLt^s,CF3	342	391	27 400	21 700
GdLt^t,CF3	338	391	27 500	22 000

---

The antenna first triplet excited states (T₁) were determined from the 0–0 transitions of the low-temperature fluorescence spectra (Table 1). The differences between the secondary and tertiary amide-linked antennae are small. The former are lower in energy, but only by 100–300 cm⁻¹. Oxygen sensitivity due to back energy transfer (BET) was previously seen in LnLt^s,CF3.⁴² To avoid BET the antenna T₁ should be at least 2000 cm⁻¹ above the Tb(III) excited state (20 400 cm⁻¹).^32,47 The 300 cm⁻¹ increase in T₁ energy to 22 000 cm⁻¹ in LnLt^t,CF3 is probably not sufficient to prevent BET.

Emission spectra were collected with λ_ex > 325 nm to avoid excitation of the pyridines. Antenna excitation yielded Tb(III) and Eu(III) emission from all the complexes. Spectral shapes and the ratios of peak intensities were similar in LnLt^s,R and LnLt^t,R, but differed from those observed in cyclen-based LnLc^t,R (Fig. S29–40†). Thus, LnLt^s,R and LnLt^t,R have similar coordination environments, which differ from that created by the DO3A-type ligands. As in the cyclen-based systems,⁴² the influence of the linkers on the coordination geometry is negligible. The Tb(III) and Eu(III) spectra consisted of 6 and 5 major signals corresponding to the ⁵D₄ → ⁷F_J (J = 6–0) the ⁵D₀ → ⁷F_J (J = 0–5) transitions, located at 488, 543, 582, 620, 652, 668, and 678 nm, and at 579, 593, 614, 649, 693, and 751 nm, respectively (Fig. 4).

The antenna and Ln(III)-based luminescence quantum yields (Φ_L and Φ_Ln, respectively) were determined by the optically dilute method using quinine sulfate (QS, Φ = 0.59)⁴⁸ as the external standard (Table 2). Gratifyingly, Φ_Ln obtained for the current series of Eu(III) and Tb(III) emitters (LnLt^t,R) were comparable to the high values previously obtained for the tert-amide-linked DO3A-based species (LnLc^t,R), a major improvement over the weak emission of LnLt^s,R. The highest Φ_Tb and Φ_Eu were measured for TbLt^t,MOM (42%) and EuLt^t,CF3 (13%), respectively. As before, Φ_L were low, below 7% in all cases. Intriguingly, there was no discernable pattern as to the order of Φ_L in the different ligand frameworks. In the non-photoactive and non-redox-active Gd(III) species, Φ_L decreases as follows: Lc^t,Me > Lt^t,Me > Lt^s,Me, Lt^s,MOM > Lc^t,MOM ≈ Lt^t,MOM, and Lt^s,CF3 > Lt^t,CF3 ≈ Lc^t,CF3.

Table 2 Antenna- and metal-based luminescence quantum yields of LnLt^t,R, with comparisons to LnLt^s,R (Ref. 38) and LnLc^t,R (Ref. 42). The values in parentheses show the change compared to LnLc^t,R

Complex	Φ L [%]^a	Φ Ln [%]^a
a Relative to QS (Φ = 0.59) in H2SO4 (0.05 M).^48 b Mean ± standard deviation for two independent measurements. c Mean ± standard deviation for three independent measurements.
GdLt^s,Me ^b	4.40 ± 0.10 (63%)	—
GdLt^t,Me ^b	5.47 ± 0.07 (79%)	—
GdLc^t,Me	6.8 (100%)	—
GdLt^s,MOM	6.42 ± 0.28 (131%)	—
GdLt^t,MOM	4.94 ± 0.06 (98%)	—
GdLc^t,MOM	5.1 (100%)	—
GdLt^s,CF3	4.53 ± 0.17 (147%)	—
GdLt^t,CF3	3.62 ± 0.02 (113%)	—
GdLc^t,CF3	3.2 (100%)	—
TbLt^s,Me	3.50 ± 0.10 (58%)	27.05 ± 1.05 (60%)
TbLt^t,Me	4.59 ± 0.10 (76%)	40.4 ± 0.40 (93%)
TbLc^t,Me	5.9 (100%)	43.4 (100%)
TbLt^s,MOM	4.92 ± 0.01 (109%)	28.05 ± 0.95 (64%)
TbLt^t,MOM	4.28 ± 0.03 (96%)	41.5 ± 0.50 (93%)
TbLc^t,MOM	4.5 (100%)	45.1 (100%)
TbLt^s,CF3	4.16 ± 0.24 (142%)	3.24 ± 0.14 (19%)
TbLt^t,CF3	3.33 ± 0.28 (116%)	19.0 ± 1.05 (113%)
TbLc^t,MF3	3.1 (100%)	15.9 (100%)
EuLt^s,Me	0.42 ± 0.03 (29%)	0.83 ± 0.02 (14%)
EuLt^t,Me	0.96 ± 0.24 (87%)	3.45 ± 1.10 (83%)
EuLc^t,Me	1.5 (100%)	6.0 (100%)
EuLt^s,MOM	0.16 ± 0.01 (6%)	2.44 ± 0.07 (28%)
EuLt^t,MOM	1.32 ± 0.06 (52%)	5.22 ± 0.15 (61%)
EuLc^t,MOM	2.5 (100%)	8.9 (100%)
EuLt^s,CF3	0.61 ± 0.02 (23%)	7.95 ± 0.42 (68%)
EuLt^t,CF3	2.81 ± 0.08 (100%)	13.0 ± 0.10 (108%)
EuLc^t,CF3	2.7 (100%)	11.6 (100%)

---

PeT from the antenna to the pyridines and to Eu(III) was calculated to be thermodynamically feasible (eqn (2)). Cyclic voltammetry yielded E_ox, the electron donor oxidation potential, as +1.91, +1.94, and +2.20 V (vs. NHE) for the antenna models 5a–c (Scheme 1, Fig. S19–21†), respectively. These models were prepared to enable cyclic voltammetric analysis of the antennae without interference from the other redox-active components of the complexes. E_red is the electron acceptor reduction potential. Pyridine E_red (−1.29 V vs. NHE) has been reported;³⁸E_red of Eu(III) was approximated with values found for a +1 charged cyclen-based complex (−0.80 V vs. NHE for a MOM-substituted complex).⁴⁰ This was necessary as the presence of the pyridines, which are reduced at similar potentials make Eu(III) E_red determination unreliable in this ligand framework. E_S is the antenna excited state energy, determined from the first vibronic band of the 77 K spectra as 3.60, 3.57, and 3.39 eV for GdLt^t,Me, GdLt^t,MOM, and GdLt^t,CF3 respectively. The last term is the coulombic stabilization of the charge-separated system, usually taken as ∼0.15 eV.⁴⁹

ΔG for PeT from the CF₃-substituted antenna to the pyridine was −0.54 eV. Values for the more electron-rich antennae were more negative, −0.98 eV and −1.04 eV for the MOM- and Me-substituted ones, respectively. Eu(III) reduction was even more favorable, with ΔG values of −0.95, −0.87, and −0.45 eV, for Me-, MOM- and CF₃-substituted complexes, respectively. In DO3A-complexes, sec-amide-linked antennae are more fluorescent than the tert-amide-linked ones.⁴² The overall Φ_L order is the result of the combination of these effects. In the Tb(III) and Eu(III) complexes the situation is further complicated by the photo- and, in the case of Eu(III), redox-activity of the metals.

The number of metal-coordinated water molecules (q) were determined from τ_obs of Eu(III) and Tb(III) (Table 3).^3,35 The luminescence decays were monoexponential for these complexes, with τ_obs = ∼0.52 ms and ∼1.34 ms for EuL and TbL, respectively. In all cases q was ∼1, which together with the octadentate ligand gives the expected nine-coordinate complexes. BET in TbLt^t,CF3 did not allow for the determination of q, however, it is most likely 1 by analogy with the other complexes.

Table 3 Ln(III) τ_obs and q-values in EuLt^t,R and TbLt^t,Ra

Complex	τ H2O [ms]	τ D2O [ms]	q
a Recorded in PIPES-buffered H2O or D2O solutions at pH = 6.5 at room temperature. b Calculated using q = (5 ms) (1/τH2O − 1/τD2O − 0.06 ms^−1) for Tb, and q = (1.2 ms) (1/τH2O − 1/τD2O − 0.25 ms^−1 − m 0.075 ms^−1) for Eu; m = number of nearby amide N–H oscillators.^3,35 c Real q value could not be determined.
TbLt^t,Me	1.37	2.13	0.99
EuLt^t,Me	0.51	1.43	1.18
TbLt^t,MOM	1.30	2.01	1.06
EuLt^t,MOM	0.52	1.43	1.15
TbLt^t,CF3	0.62	1.01	2.71^c
EuLt^t,CF3	0.52	1.43	1.15

---

The similar Φ_L and Φ_Ln values of pyridine-equipped LnLt^t,R and DO3A-based LnLc^t,R suggested that either PeT to the pyridines was less efficient in LnLt^t,R than in LnLt^s,R, or PET was successfully outcompeted by energy transfer to the Ln(III). The efficiency with which the antenna excited state populates the Ln(III) excited state (η_sens) is readily calculated for Eu(III) emitters using eqn (1) and (3) (Table 4).^36,37 Antenna excited state quenching affects η_sens. If complexes with identical antennae and similar geometries have different η_sens, this difference may be due to differences in PeT. Φ^Ln_Ln in eqn (1) can be calculated³⁶ from τ_rad of Eu(III), which is obtained from the luminescence spectrum using eqn (3). I_tot is the integrated full spectrum (530–800 nm), I_MD is the integrated ⁵D₀ → ⁷F₁ transition (582–603 nm), A_MD,0 is the spontaneous emission probability (14.65 s⁻¹), n is the refractive index of the aqueous medium (approximated with water), and τ_obs equals τ_H2O.

Table 4 τ _rad, Φ^Ln_Ln, and η_sens of EuL

Complex	τ rad [ms]	Φ Eu ^Eu [%]	η sens [%]
a Calculated using eqn (3).^36 b Calculated using eqn (1) and (3).^36
EuLt^s,Me	2.87	17.4	5
EuLt^t,Me	2.94	17.7	28
EuLc^t,Me	5.41	12.0	50
EuLt^s,MOM	2.86	17.5	14
EuLt^t,MOM	2.96	17.9	30
EuLc^t,MOM	5.40	12.2	73
EuLt^s,CF3	2.87	17.8	46
EuLt^t,CF3	2.99	17.4	75
EuLc^t,CF3	5.40	12.2	95

---

EuLt^t,R and EuLt^s,R have similar τ_rad (∼2.9 ms). This is in line with their having similar Ln(III) coordination environments, as suggested by their superimposable luminescence spectra (Fig. S53–55†). The slightly longer τ_obs of EuLt^t,R than of EuLt^s,R is presumably due to the removal of the N–H group from the amide linker. Overall, the effect of the tertiary amides on Φ^Ln_Ln is small compared to EuLt^s,R, and all complexes have Φ^Ln_Ln ∼ 17.5%. Compared with the Φ^Ln_Ln ∼ 12% obtained for EuLc^t,R, this is a significant improvement. The shortening of τ_rad to increase Φ^Ln_Ln is a rarely used strategy for improving Φ_Ln.⁵⁰

Although EuLt^t,R have higher Φ^Ln_Ln than EuLc^t,R their Φ_Ln values were similar or lower, due to the lower η_sens. EuLt^t,R shows marked improvement in η_sens compared to secondary amide-linked EuLt^s,R (e.g. 28% and 5% in EuLt^t,Me and EuLt^s,Me, respectively), just not enough to reach the values seen in EuLc^t,R (50% in EuLc^t,Me). The better sensitization in EuLt^t,R than in EuLt^s,R may be due to less efficient PeT to the pyridines. This is suggested by the fact that Φ_L in GdLt^t,Me is somewhat higher than in GdLt^s,Me even though in DO3A-based complexes secondary amide linked antennae have higher Φ_L than their tertiary amide-linked analogues.⁴² However, the low Φ_L values do not allow for a reliable comparison. A second alternative explanation for the observed results may be more efficient energy transfer due to a different antenna orientation. A third possibility is more efficient population of the Ln(III) feeding level, likely the antenna T₁, due to more efficient intersystem crossing in LnLt^t,R than in LnLt^s,R. The additional improvement seen in LnLc^t,R compared to LnLt^t,R is likely due to the absence of PeT to pyridines. As above, antenna orientation could also play a role. Both LnLc^t,R and Lnt^t,R have tert-amide-linked antennae, and the similar values of Φ_L are consistent with comparable levels of antenna S₁ quenching. Thus the higher η_sens in cyclen-based LnLc^t,R indicates more efficient energy transfer than in tacn-based Lnt^t,R.

Photostability of LnL. The photostability of a Ln(III) emitter is an important parameter for practical applications. Unlike those of organic fluorophores, Ln(III) excited states are not sensitive to quenching by atmospheric oxygen.⁵¹ Photodegradation is rather due to processes that compromise the integrity of the other functional units in the complex, often the antenna. BET from Ln(III)* to the antenna T₁ or slow energy transfer from T₁ to Ln(III) caused by too small or too large T₁-Ln(III)* gaps, respectively, both result in long T₁ lifetimes, and thus oxygen-sensitivity.

The photostabilities of the complexes reported herein (LnLt^t,R) were investigated and compared to those of LnLt^s,R and LnLc^t,R (Fig. 5). The Eu(III) complexes proved quite robust, and retained at least 90% of their luminescence after 2.5 h of irradiation. Small differences, however, could be noted. Complexes equipped with a CF₃-substituted antenna were the most resistant to photodegradation, while the least stable were the ones carrying the most electron-rich Me-substituted antenna. Cyclen-based EuLc^t,R were in all cases more stable than their tacn-based analogues EuLt^s,R and EuLt^t,R. Tert-amide-linked EuLt^t,R were somewhat more stable then sec-amide-linked EuLt^s,R. These trends are consistent with PeT opening up a degradation pathway for the complex. This pathway is more prominent for picolinate-carrying EuLt^s,R and EuLt^t,R, which contain both Eu(III) and pyridines as potential electron acceptors.

Fig. 5 Photostability of Tb and Eu complexes (10 μM) in aqueous PIPES buffer (10 mM), pH = 6.5, using 395 nm filter for Tb and 500 nm filter for Eu complexes, respectively. λ_ex(LnLc^t,Me) = 324 nm, λ_ex(LnLt^t,Me) = 325 nm, λ_ex(LnLc^t,MOM/CF3) = 327 nm, λ_ex(LnLt^s,Me/LnLt^t,MOM) = 328 nm, λ_ex(LnLt^s,MOM/(LnLt^s,CF3/LnLt^t,CF3) = 330 nm. Values normalised to integrated emission intensity at t₀.

In the case of the Tb(III) complexes fastest degradation was seen for those with the lowest antenna T₁. All three CF₃-substituted complexes rapidly lost Tb(III) luminescence efficiency. This loss was fastest for TbLt^s,CF3, with a sec-amide-linked antenna. For the other two antennae attaching them via a tert-amide, and the concomitant increase in T₁ was sufficient to protect the integrity of the emitters. These trends indicate that photolability is due to BET populating the antenna T₁, followed by T₁ reacting with atmospheric oxygen.

Conclusions

We investigated the effect of replacing a secondary amide linker with a methylcarboxylate-substituted tertiary amide one on the photophysical properties of tacn-based luminescent Ln(III) complexes. In solution the amide substituent was non-coordinating, and did not alter the coordination geometry of the complex compared to the parent species. This was indicated by the shapes of the Ln(III) luminescence spectra and the radiative lifetimes of the Eu(III) complexes, and was supported by the results of the paramagnetic NMR spectroscopic and single crystal X-ray crystallographic analyses in the solution and solid states, respectively.

Changing the linker from secondary to tertiary amide had a profound effect on the Eu(III) and Tb(III) luminescence. In tacn-based, pyridine-containing Ln(III) complexes with secondary amide linkers pyridine reduction by the photoexcited carbostyril competed efficiently with Eu(III) and Tb(III) sensitization. Upon linker replacement all of the emitters experienced an increase in luminescence quantum yield, with values close to, and in some cases even higher than, those obtained for cyclen-based systems lacking the quenching pyridines. An additional benefit of the tert-amide linkers is higher antenna T₁ energy, and thus decreased BET for Tb(III) emitters. Finally, tertiary amide linked complexes based on both tacn and cyclen frameworks showed high photostability, highlighting the importance of closing down PeT and BET quenching pathways to obtain robust emitters.

Author contributions

D. Kocsi and D. Kovacs: conceptualization, resources, investigation, data curation, visualization, formal analysis. J.A.L.W.: investigation, data curation, visualization, formal analysis. K.E.B.: conceptualization, funding acquisition, project administration, resources, formal analysis, visualization, supervision. All authors were involved in the writing of the original draft and took part in the reviewing and editing of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Swedish Research Council (project grant 2017-04077 for K.E.B.), the Knut och Alice Wallenbergs Foundation (Dnr: 2018.0066).

Notes and references

1. J.-C. G. Bünzli and S. V. Eliseeva, in Lanthanide Luminescence: Photophysical, Analytical and Biological Aspects, ed. P. Hänninen and H. Härmä, Springer Berlin Heidelberg, Berlin, Heidelberg, 2011, pp. 1–45.
2. A. de Bettencourt-Dias, in Luminescence of Lanthanide Ions in Coordination Compounds and Nanomaterials, John Wiley & Sons Ltd, 2014, pp. 1–48.
3. W. D. Horrocks and D. R. Sudnick, Acc. Chem. Res., 1981, 14, 384–392.
4. W. D. Horrocks, Jr., J. P. Bolender, W. D. Smith and R. M. Supkowski, J. Am. Chem. Soc., 1997, 119, 5972–5973.
5. A. S. Chauvin, F. Gumy, D. Imbert and J. C. G. Bünzli, Spectrosc. Lett., 2004, 37, 517–532.
6. A. de Bettencourt-Dias, P. S. Barber and S. Bauer, J. Am. Chem. Soc., 2012, 134, 6987–6994.
7. A. de Bettencourt-Dias, S. Viswanathan and A. Rollett, J. Am. Chem. Soc., 2007, 129, 15436–15437.
8. J. H. S. K. Monteiro, A. de Bettencourt-Dias and F. A. Sigoli, Inorg. Chem., 2017, 56, 709–712.
9. R. Xiong, D. Mara, J. Liu, R. Van Deun and K. E. Borbas, J. Am. Chem. Soc., 2018, 140, 10975–10979.
10. J. Laakso, G. A. Rosser, C. Szíjjártó, A. Beeby and K. E. Borbas, Inorg. Chem., 2012, 51, 10366–10374.
11. T. Zhang, X. Zhu, C. C. W. Cheng, W.-M. Kwok, H.-L. Tam, J. Hao, D. W. J. Kwong, W.-K. Wong and K.-L. Wong, J. Am. Chem. Soc., 2011, 133, 20120–20122.
12. Y. Ning, M. Zhu and J.-L. Zhang, Coord. Chem. Rev., 2019, 399, 213028.
13. Y. Ning, G.-Q. Jin and J.-L. Zhang, Acc. Chem. Res., 2019, 52, 2620–2633.
14. Y. Ning, J. Tang, Y.-W. Liu, J. Jing, Y. Sun and J.-L. Zhang, Chem. Sci., 2018, 9, 3742–3753.
15. D. Parker, P. K. Senanayake and J. A. G. Williams, J. Chem. Soc., Perkin Trans. 2, 1998, 2129–2139.
16. A. Beeby, S. Faulkner, D. Parker and J. A. G. Williams, J. Chem. Soc., Perkin Trans. 2, 2001, 1268–1273.
17. D. G. Smith, R. Pal and D. Parker, Chem. – Eur. J., 2012, 18, 11604–11613.
18. D. G. Smith, B. K. McMahon, R. Pal and D. Parker, Chem. Commun., 2012, 48, 8520–8522.
19. A. Foucault-Collet, C. M. Shade, I. Nazarenko, S. Petoud and S. V. Eliseeva, Angew. Chem., Int. Ed., 2014, 53, 2927–2930.
20. S. Petoud, G. Muller, E. G. Moore, J. Xu, J. Sokolnicki, J. P. Riehl, U. N. Le, S. M. Cohen and K. N. Raymond, J. Am. Chem. Soc., 2007, 129, 77–83.
21. S. Faulkner, M.-C. Carrie, S. J. A. Pope, J. Squire, A. Beeby and P. G. Sammes, Dalton Trans., 2004, 1405–1409.
22. C. Y. Chow, S. V. Eliseeva, E. R. Trivedi, T. N. Nguyen, J. W. Kampf, S. Petoud and V. L. Pecoraro, J. Am. Chem. Soc., 2016, 138, 5100–5109.
23. E. R. Trivedi, S. V. Eliseeva, J. Jankolovits, M. M. Olmstead, S. Petoud and V. L. Pecoraro, J. Am. Chem. Soc., 2014, 136, 1526–1534.
24. T. N. Nguyen, C. Y. Chow, S. V. Eliseeva, E. R. Trivedi, J. W. Kampf, I. Martinic, S. Petoud and V. L. Pecoraro, Chem. – Eur. J., 2018, 24, 1031–1035.
25. T. Lazarides, D. Sykes, S. Faulkner, A. Barbieri and M. D. Ward, Chem. – Eur. J., 2008, 14, 9389–9399.
26. T. Lazarides, N. M. Tart, D. Sykes, S. Faulkner, A. Barbieri and M. D. Ward, Dalton Trans., 2009, 3971–3979.
27. N. M. Shavaleev, L. P. Moorcraft, S. J. A. Pope, Z. R. Bell, S. Faulkner and M. D. Ward, Chem. Commun., 2003, 1134–1135.
28. N. M. Shavaleev, L. P. Moorcraft, S. J. A. Pope, Z. R. Bell, S. Faulkner and M. D. Ward, Chem. – Eur. J., 2003, 9, 5283–5291.
29. S. J. A. Pope, B. J. Coe and S. Faulkner, Chem. Commun., 2004, 1550–1551.
30. S. Faulkner and S. J. A. Pope, J. Am. Chem. Soc., 2003, 125, 10526–10527.
31. S. Cotton, in Lanthanide and Actinide Chemistry, 2006, pp. 35–60.
32. D. Parker, R. S. Dickins, H. Puschmann, C. Crossland and J. A. K. Howard, Chem. Rev., 2002, 102, 1977–2010.
33. W. D. Horrocks, Jr. and D. R. Sudnick, J. Am. Chem. Soc., 1979, 101, 334–340.
34. R. S. Dickins, D. Parker, A. S. de Sousa and J. A. G. Williams, Chem. Commun., 1996, 697–698.
35. A. Beeby, I. M. Clarkson, R. S. Dickins, S. Faulkner, D. Parker, L. Royle, S. A. S. de, J. A. G. Williams and M. Woods, J. Chem. Soc., Perkin Trans. 2, 1999, 493–504.
36. M. H. V. Werts, R. T. F. Jukes and J. W. Verhoeven, Phys. Chem. Chem. Phys., 2002, 4, 1542–1548.
37. K. Binnemans, Coord. Chem. Rev., 2015, 295, 1–45.
38. D. Kovacs, D. Kocsi, J. A. L. Wells, S. R. Kiraev and K. E. Borbas, Dalton Trans., 2021, 50, 4244–4254.
39. D. Kovacs, S. R. Kiraev, D. Phipps, A. Orthaber and K. E. Borbas, Inorg. Chem., 2020, 59, 106–117.
40. D. Kovacs, E. Mathieu, S. R. Kiraev, J. A. L. Wells, E. Demeyere, A. Sipos and K. E. Borbas, J. Am. Chem. Soc., 2020, 142, 13190–13200.
41. S. R. Kiraev, E. Mathieu, F. Siemens, D. Kovacs, E. Demeyere and K. E. Borbas, Molecules, 2020, 25, 5282.
42. D. Kovacs, D. Phipps, A. Orthaber and K. E. Borbas, Dalton Trans., 2018, 47, 10702–10714.
43. D. Parker and J. A. G. Williams, J. Chem. Soc., Perkin Trans. 2, 1996, 1581–1586.
44. D. Kovacs and K. E. Borbas, Coord. Chem. Rev., 2018, 364, 1–9.
45. A. K. R. Junker and T. J. Sorensen, Eur. J. Inorg. Chem., 2019, 2019, 1201–1206.
46. L.-M. Fu, X.-C. Ai, M.-Y. Li, X.-F. Wen, R. Hao, Y.-S. Wu, Y. Wang and J.-P. Zhang, J. Phys. Chem. A, 2010, 114, 4494–4500.
47. J.-C. G. Bunzli, Coord. Chem. Rev., 2015, 293–294, 19–47.
48. K. Suzuki, A. Kobayashi, S. Kaneko, K. Takehira, T. Yoshihara, H. Ishida, Y. Shiina, S. Oishi and S. Tobita, Phys. Chem. Chem. Phys., 2009, 11, 9850–9860.
49. A. Beeby, S. Faulkner and J. A. G. Williams, J. Chem. Soc., Dalton Trans., 2002, 1918–1922.
50. S. V. Eliseeva, D. N. Pleshkov, K. A. Lyssenko, L. S. Lepnev, J.-C. G. Bunzli and N. P. Kuzmina, Inorg. Chem., 2011, 50, 5137–5144.
51. A. Beeby, D. Parker and J. A. G. Williams, J. Chem. Soc., Perkin Trans. 2, 1996, 1565–1580.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details, synthesis and chemical characterisation of new compounds, additional photophysical and crystallographic characterization, ¹H, ¹³C, and ¹⁹F NMR spectra of new compounds. CCDC 2102702, 2102703. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/d1dt02893f

---