Tuning the photophysical properties of lanthanide(III)/zinc(II) ‘encapsulated sandwich’ metallacrowns emitting in the near-infrared range

Abstract

A family of Zn₁₆Ln(HA)₁₆ metallacrowns (MCs; Ln = Yb^III, Er^III, and Nd^III; HA = picoline- (picHA²⁻), pyrazine- (pyzHA²⁻), and quinaldine- (quinHA²⁻) hydroximates) with an ‘encapsulated sandwich’ structure possesses outstanding luminescence properties in the near-infrared (NIR) and suitability for cell imaging. Here, to decipher which parameters affect their functional and photophysical properties and how the nature of the hydroximate ligands can allow their fine tuning, we have completed this Zn₁₆Ln(HA)₁₆ family by synthesizing MCs with two new ligands, naphthyridine- (napHA²⁻) and quinoxaline- (quinoHA²⁻) hydroximates. Zn₁₆Ln(napHA)₁₆ and Zn₁₆Ln(quinoHA)₁₆ exhibit absorption bands extended into the visible range and efficiently sensitize the NIR emissions of Yb^III, Er^III, and Nd^III upon excitation up to 630 nm. The energies of the lowest singlet (S₁), triplet (T₁) and intra-ligand charge transfer (ILCT) states have been determined. Ln^III-centered total (Q^L_Ln) and intrinsic (Q^Ln_Ln) quantum yields, sensitization efficiencies (η_sens), observed (τ_obs) and radiative (τ_rad) luminescence lifetimes have been recorded and analyzed in the solid state and in CH₃OH and CD₃OD solutions for all Zn₁₆Ln(HA)₁₆. We found that, within the Zn₁₆Ln(HA)₁₆ family, τ_rad values are not constant for a particular Ln^III. The close in energy positions of T₁ and ILCT states in Zn₁₆Ln(picHA)₁₆ and Zn₁₆Ln(quinHA)₁₆ are preferred for the sensitization of Ln^III NIR emission and η_sens values reach 100% for Nd^III. Finally, the highest values of Q^L_Ln are observed for Zn₁₆Ln(quinHA)₁₆ in the solid state or in CD₃OD solutions. With these data at hand, we are now capable of creating MCs with desired properties suitable for NIR optical imaging.

---

Introduction

The numerous advantages of non-invasive near-infrared (NIR) optical imaging for biological applications^1–4 have stimulated significant advances in the creation of novel probes for cell, tissue and organ visualization. Among the different classes of NIR imaging probes,^5,6 Ln^III-based molecular compounds can offer complementary advantages,⁷ in particular, sharp emission bands with defined positions, large energy difference between excitation and emission wavelengths, and high photostability. Apart from biological applications, the NIR emission arising from Ln^III complexes can be used in material sciences, telecommunications, light-emitting diodes, night vision displays, security inks, and solar energy conversion.^8,9 However, for the design of highly NIR-emitting Ln^III molecular compounds, several challenges have to be addressed.^10,11 The emission efficiency of Ln^III complexes is quantified by the following equation:where Q^L_Ln is the total quantum yield obtained under ligand excitation, η_sens is the sensitization efficiency of the organic ligands taking into account the energy migration processes occurring both within the ligands and from the ligand to the Ln^III excited state, Q^Ln_Ln is the intrinsic quantum yield or quantum yield upon direct f–f excitation, which reflects the relative impact of radiative and non-radiative processes occurring around Ln^III, and τ_rad is the radiative lifetime or the luminescence lifetime in absence of non-radiative deactivation processes. In view of practical applications, the brightness of a Ln^III compound is defined as ε × Q^L_Ln, where ε is the molar absorption coefficient.

The first requirement to obtain high Q^L_Ln is to optimize energy migration processes, in particular η_sens. This parameter is mainly dependent on the nature of the organic ligands and the relative positions of their electronic levels (lowest singlet S₁, triplet T₁ or charge transfer (CT) states) with respect to the accepting levels of Ln^III.^10,11 In this respect, several phenomenological rules have been developed for visible-emitting Ln^III complexes.^12–15 On the other hand, no clear correlations have been established for coordination compounds incorporating NIR-emitting Ln^III while the choice of chromophores is much wider and includes molecules absorbing light in the visible and the NIR ranges.^16–20 A second important challenge is to minimize sources of non-radiative deactivation mechanisms (overtones of high-energy vibrations, CT states, back energy transfer processes) that decrease observed luminescence lifetimes and, in turn, the corresponding intrinsic quantum yields. Avoiding sources of non-radiative deactivation taking places through overtones of not only O–H and N–H, but also C–H vibrations that are widely present in complexes formed with organic ligands, is of particular importance for NIR-emitting Ln^III because of the small energy gaps between the emitting and the receiving states. The most widely used strategy to overcome this limitation is by halogenation^{18,19,21–23} or by deuteration^24–26 of the organic ligands. Another approach to minimize non-radiative deactivations is to create a protective hydrophobic environment around the Ln^III ion as in the case of complexes formed with imidodiphosphonate ligands.^27,28 In addition, according to eqn (1), one more parameter that can be tuned and that lead to the enhancement of Q^L_Ln is the radiative lifetime. The validity of this approach has been demonstrated in several studies.^26,29–31

Recently, our group has pioneered a new approach toward the design of NIR-emitting Ln^III complexes that take advantage of metallacrown (MC) scaffolds.^32–38 MCs are a class of inorganic macrocycles forming repeating [metal–N–O] subunits.³⁹ These MC complexes possess oxygen atoms oriented toward the center of the ring allowing an efficient pre-organization for cation encapsulation in a manner similar to classical organic crown ethers.

We created a family of NIR-emitting Zn₁₆Ln(HA)₁₆ MCs assembled using picoline- (picHA²⁻),³³ pyrazine- (pyzHA²⁻),³⁷ and quinaldine-(quinHA²⁻)³² hydroximate ligands (Fig. 1) and demonstrated their promising applications for NIR cell imaging.^37,38 The unique ‘encapsulated sandwich’ structure of this family of MCs not only allows for highly efficient sensitization of the characteristic emissions from Yb^III, Nd^III and Er^III but also locates the luminescent Ln^III ion at a relatively long distance (∼7 Å) from C–H oscillators present on the organic ligands, thereby minimizing the effect of non-radiative deactivation. We have demonstrated that the nature of the hydroxamate ligand affects not only the luminescence but also the functional properties, such as the water solubility obtained for MCs formed with the pyzHA²⁻ chromophoric building-block. However, the absorption maxima of all previously reported Zn₁₆Ln(HA)₁₆ MCs have remained in the UV range. An excitation light in the UV domain is detrimental for practical applications such as biological optical imaging due to strong interaction with biological tissues or fluids, resulting in their disturbance or damage.

Fig. 1 (Left) Molecular structure of the Zn₁₆Yb(quinoHA)₁₆ MC obtained by single crystal X-ray diffraction. Color code: green, Yb; yellow, Zn; red, O; blue, N; grey, C. H atoms and solvents have been omitted for clarity. (right) Structures of the hydroxamic acids used to assemble ‘encapsulated sandwich’ MCs.

Therefore, in this work, we expand the Zn₁₆Ln(HA)₁₆ family by creating MCs with two new ligands, naphthyridine- (napHA²⁻) and quinoxaline- (quinoHA²⁻) hydroximate (Fig. 1) following the specific goal of shifting the absorption wavelength further into the visible region. In addition, this comprehensive study that includes Zn₁₆Ln(HA)₁₆ MCs with a full range of substitution patterns addresses the general goal of bringing a rationalization for the smart tuning of the functional and photophysical properties, with a specific focus on NIR emission, on the basis of the nature of the hydroximate ligand. Synthesis and characterization of new Zn₁₆Ln(napHA)₁₆ and Zn₁₆Ln(quinoHA)₁₆ MCs formed with several lanthanide cations (Ln = Yb^III, Er^III, Nd^III, Gd^III) and Y^III, including the structural analysis of Zn₁₆Nd(quinoHA)₁₆ and Zn₁₆Yb(quinoHA)₁₆, are reported. Absorption, excitation, and emission spectra, total and intrinsic quantum yields, sensitization efficiencies, observed and radiative lifetimes were acquired and extensively analyzed herein for the NIR-emitting Zn₁₆Ln(napHA)₁₆ and Zn₁₆Ln(quinoHA)₁₆ (Ln = Yb^III, Er^III, Nd^III) in the solid state and in CH₃OH and CD₃OD solutions. Moreover, we completed the detailed photophysical studies for Zn₁₆Ln(picHA)₁₆, Zn₁₆Ln(pyzHA)₁₆ and Zn₁₆Ln(quinHA)₁₆ by recording Q^Ln_Ln for Nd^III and Er^III analogues or τ_rad for Yb^III. Our approach allows the quantification of all relevant parameters from eqn (1), in particular τ_obs, τ_rad, η_sens and Q^Ln_Ln, and their impact on the total quantum yield Q^L_Ln within the series of Zn₁₆Ln(HA)₁₆ MCs created with the different hydroximate ligands.

Experimental

All syntheses were performed under aerobic conditions using chemicals and solvents as received unless otherwise stated. ESI-MS spectra were collected with a Micromass LCT time-of-flight electrospray mass spectrometer in a negative ion mode at a cone voltage of −40 V on samples dissolved in CH₃OH. Samples were injected via a syringe pump. Data were processed with the program MassLynx 4.0. Elemental analyses were performed by Atlantic Microlabs Inc.

Synthesis of quinoxalinehydroxamic acid (H₂quinoHA)

In a three-neck flask, 2-quinoxalinecarboxylic acid (3.1 g, 18.0 mmol) was added into a solution of N-methylmorpholine (2.4 mL, 21.8 mmol) in dichloromethane (50.0 mL). The mixture was stirred under N₂ for 15 minutes and cooled down to 0 °C. Ethylchloroformate (2.0 mL, 21.0 mmol) was added dropwise under stirring. The mixture was stirred for an additional 1 hour and filtered to obtain a brown solution. Meanwhile, a fresh hydroxylamine solution was prepared by reacting hydroxylamine hydrochloride (1.9 g, 27.0 mmol) with potassium hydroxide at 85% (1.8 g, 27.0 mmol) in methanol (16.0 mL) at 0 °C. The solution was stirred for 10 minutes, filtered, and the filtrate obtained was added dropwise into the brown solution while stirring at 0 °C. The reaction mixture was stirred for 2 hours, filtered, and the filtrate was left for slow evaporation to induce the precipitation of a white solid. The solid was recrystallized in hot water, collected by filtration, washed with a copious amount of water, and dried under vacuum to yield the pure powder of H₂quinoHA. Yield ∼26% (0.90 g), ESI-MS, calc. for [M − H⁺], C₉H₆N₃O₂, 188.05; found 188.0. ¹H NMR (400 MHz, DMSO-d₆) δ ppm 11.78 (s, 1H), 9.40 (s, 1H), 9.37 (s, 1H), 8.20–8.15 (m, 2H), 8.00–7.95 (m, 2H). ¹³C NMR (400 MHz, DMSO-d₆) δ ppm 161.0, 144.7, 143.6, 142.8, 139.8, 131.7, 131.2, 129.4, 129.1. Elem. anal., calc. (found) for (C₉H₇N₃O₂), C: 57.14 (57.42), H: 3.73 (3.85), N: 22.21 (22.06).

Synthesis of naphthyridinehydroxamic acid (H₂napHA)

H₂napHA was synthesized in two steps.

Step 1: to a stirred solution of 1,6-naphthyridine-2-carboxylic acid (0.20 g, 1.1 mmol) in 20.0 mL of CH₃OH, 1.0 mL of H₂SO₄ was added dropwise. The solution was then refluxed for 6 hours. After cooling down to room temperature, the solution was neutralized by a CH₃OH solution of NaOH until the formation of a solid which was filtered, washed with H₂O, and dried under vacuum to obtain methyl naphthyridine-2-carboxylate. Yield: ∼56% (0.12 g). ¹H NMR (400 MHz, CDCl₃) δ ppm 9.37 (s, 1H), 8.84 (d, 1H), 8.46 (d, 1H), 8.31 (d, 1H), 8.10 (d, 1H), 4.09 (s, 3H).

Step 2: without further purification, 0.095 g (0.50 mmol) of methyl naphthyridine-2-carboxylate was added to a 15.0 mL solution of CH₃OH containing 0.40 g of hydroxylamine hydrochloride. An amount of 0.32 g of NaOH was added and the resulting solution was stirred at room temperature overnight. CH₃OH was then removed under vacuum to give a solid, which was washed by a dilute HCl solution to give pure H₂napHA. Yield: ∼53% (0.050 g). ESI-MS, calc. for [M − H⁺], C₉H₆N₃O₂, 188.05; found 188.0. ¹H NMR (400 MHz, DMSO-d₆) δ ppm 11.71 (s, 1H), 9.50 (s, 1H), 9.32 (s, 1H), 8.82 (d, 1H), 8.76 (d, 1H), 8.21 (d, 1H), 7.95 (d, 1H). Elem. anal., calc. (found) for (C₉H₇N₃O₂), C: 57.14 (57.26), H: 3.73 (3.70), N: 22.21 (22.10).

Synthesis of Zn₁₆Ln(quinoHA)₁₆ and Zn₁₆Ln(napHA)₁₆ (Ln = Nd^III, Gd^III, Er^III, Yb^III) and the Y^III analogues

A general synthesis for these compounds is described below for the Zn₁₆Nd(quinoHA)₁₆ complex. Complexes with Y^III, other Ln^III and napHA²⁻ were prepared by substituting the appropriate metal ion for Nd^III and hydroxamic acid for H₂quinoHA. Deuterated pyridine (py-d₅) was used in the syntheses of the Y^III analogues.

[Zn₁₆Nd(quinoHA)₁₆(py)₈](OTf)₃ (Zn₁₆Nd(quinoHA)₁₆). H₂quinoHA (0.060 g, 0.32 mmol) was added to a solution containing 1.0 mL pyridine, 5.0 mL DMF, and 5.0 mL H₂O. The solution was stirred for 5 minutes in order to completely dissolve the solid. Zinc triflate (0.12 g, 0.32 mmol) and neodymium triflate (0.024 g, 0.040 mmol) were added and the orange solution was stirred at 80 °C for 2 hours. The solution was then cooled down to room temperature, and filtered. The filtrate was layered with H₂O to give red–brown crystals after three days. The crystals were collected by filtration, washed with H₂O, and dried in air. Yield ∼28% (31 mg). ESI-MS (CH₃OH), calc. for [M]³⁺, Zn₁₆NdC₁₄₄H₈₀N₄₈O₃₂, 1392.79; found 1393.0. Elem. anal., calc. (found) for (Zn₁₆NdC₁₈₇H₁₂₀N₅₆O₄₁F₉S₃)(H₂O)₁₂, C: 40.98 (40.67), H: 2.65 (2.38), N: 14.31 (14.24).

[Zn₁₆Gd(quinoHA)₁₆(py)₈](OTf)₃ (Zn₁₆Gd(quinoHA)₁₆). Yield ∼35% (38 mg). ESI-MS (CH₃OH), calc. for [M]³⁺, Zn₁₆GdC₁₄₄H₈₀N₄₈O₃₂, 1398.13; found 1396.8. Elem. anal., calc. (found) for (Zn₁₆GdC₁₈₇H₁₂₀N₅₆O₄₁F₉S₃)(H₂O)₁₂, C: 40.88 (40.68), H: 2.64 (2.26), N: 14.28 (14.13).

[Zn₁₆Er(quinoHA)₁₆(py)₈](OTf)₃ (Zn₁₆Er(quinoHA)₁₆). Yield ∼24% (26 mg). ESI-MS (CH₃OH), calc. for [M]³⁺, Zn₁₆ErC₁₄₄H₈₀N₄₈O₃₂, 1400.79; found 1400.3. Elem. anal., calc. (found) for (Zn₁₆ErC₁₈₇H₁₂₀N₅₆O₄₁F₉S₃)(H₂O)₁₂, C: 40.81 (40.85), H: 2.64 (2.43), N: 14.25 (14.38).

[Zn₁₆Yb(quinoHA)₁₆(py)₈](OTf)₃ (Zn₁₆Yb(quinoHA)₁₆). Yield ∼23% (25 mg). ESI-MS (CH₃OH), calc. for [M]³⁺, Zn₁₆YbC₁₄₄H₈₀N₄₈O₃₂, 1403.46; found 1403.4. Elem. anal., calc. (found) for (Zn₁₆YbC₁₈₇H₁₂₀N₅₆O₄₁F₉S₃)(H₂O)₁₂, C: 40.76 (40.90), H: 2.63 (2.39), N: 14.24 (14.26).

[Zn₁₆Y(quinoHA)₁₆(py-d₅)₈](OTf)₃ (Zn₁₆Y(quinoHA)₁₆). ESI-MS (CH₃OH), calc. for [M]³⁺, Zn₁₆YbC₁₄₄H₈₀N₄₈O₃₂, 1375.12; found 1373.4.

[Zn₁₆Nd(napHA)₁₆(py)₈](OTf)₃ (Zn₁₆Nd(napHA)₁₆). Yield ∼27% (30 mg). ESI-MS (CH₃OH), calc. for [M]³⁺ Zn₁₆NdC₁₄₄H₈₀N₄₈O₃₂, 1392.79; found 1393.0. Elem. anal., calc. (found) for (Zn₁₆NdC₁₈₇H₁₂₀N₅₆O₄₁F₉S₃)(H₂O)₁₂, C: 40.98 (40.70), H: 2.65 (2.52), N: 14.31 (14.10).

[Zn₁₆Gd(napHA)₁₆(py)₈](OTf)₃ (Zn₁₆Gd(napHA)₁₆). Yield ∼32% (35 mg). ESI-MS (CH₃OH), calc. for [M]³⁺, Zn₁₆GdC₁₄₄H₈₀N₄₈O₃₂, 1398.13; found 1397.7. Elem. anal., calc. (found) for (Zn₁₆GdC₁₈₇H₁₂₀N₅₆O₄₁F₉S₃)(H₂O)₁₂, C: 40.88 (40.67), H: 2.64 (2.65), N: 14.28 (14.30).

[Zn₁₆Er(napHA)₁₆(py)₈](OTf)₃ (Zn₁₆Er(napHA)₁₆). Yield ∼24% (26 mg). ESI-MS (CH₃OH), calc. for [M]³⁺, Zn₁₆ErC₁₄₄H₈₀N₄₈O₃₂, 1400.79; found 1401.1. Elem. anal., calc. (found) for (Zn₁₆ErC₁₈₇H₁₂₀N₅₆O₄₁F₉S₃)(H₂O)₁₂, C: 40.81 (40.71), H: 2.64 (2.66), N: 14.25 (14.22).

[Zn₁₆Yb(napHA)₁₆(py)₈](OTf)₃ (Zn₁₆Yb(napHA)₁₆). Yield ∼27% (30 mg). ESI-MS (CH₃OH), calc. for [M]³⁺, Zn₁₆YbC₁₄₄H₈₀N₄₈O₃₂, 1403.46; found 1403.0. Elem. anal., calc. (found) for (Zn₁₆YbC₁₈₇H₁₂₀N₅₆O₄₁F₉S₃)(H₂O)₁₂, C: 40.76 (40.60), H: 2.63 (2.71), N: 14.24 (14.11).

[Zn₁₆Y(napHA)₁₆(py-d₅)₈](OTf)₃ (Zn₁₆Y(napHA)₁₆). ESI-MS (CH₃OH), calc. for [M]³⁺, Zn₁₆YC₁₄₄H₈₀N₄₈O₃₂, 1375.12; found 1375.1.

X-ray crystallography

Orange blocks of Zn₁₆Nd(quinoHA)₁₆ and Zn₁₆Yb(quinoHA)₁₆ were grown from a dimethylformamide/water solution of the compound at 22 °C. A crystal of dimensions 0.20 × 0.19 × 0.14 mm for the former, and 0.12 × 0.11 × 0.08 mm for the latter, was mounted on a Rigaku AFC10K Saturn 944+ CCD-based X-ray diffractometer equipped with a low temperature sample conditioning device and a Micromax-007HF Cu-target micro-focus rotating anode (λ = 1.54187 A) operated at 1.2 kW power (40 kV, 30 mA). The X-ray intensities were measured at 85(1) K with the detector placed at a distance 42.00 mm from the crystal.

For Zn₁₆Nd(quinoHA)₁₆, a total of 4125 images were collected with an oscillation width of 1.0° in ω. Exposure times were 5 s and 25 s for the low- and high-angle images respectively. Rigaku d*trek images were exported to CrysAlisPro for the processing of the collected data and corrected for absorption. The integration of the experimental data yielded a total of 472 686 reflections to a maximum 2θ value of 148.15° of which 12 397 were independent and 11 558 were greater than 2σ(I). The final cell constants (Table S1†) were based on the xyz centroids of 129 624 reflections described above 10σ(I). The analysis of the data showed a negligible intensity decay during data collection. The structure was solved and refined with the Bruker SHELXTL (version 2014/6) software package⁴⁰ and using the space group P4/ncc with Z = 4 for the formula C₁₈₇H₁₂₀N₆₂O_57.5F₉S₃Zn₁₆Nd. All non-hydrogen atoms were refined anisotropically with the hydrogen atoms placed in idealized positions. Full matrix least-squares refinements based on F² converged at R₁ = 0.0550 and wR₂ = 0.1902 [based on I > 2σ(I)], R₁ = 0.0572 and wR₂ = 0.1943 for all data. The SQUEEZE subroutine of the PLATON program suite⁴¹ was used to address highly disordered solvent and two triflate anions present in the structure. Additional details are presented in Table S1.†

For Zn₁₆Yb(quinoHA)₁₆, a total of 2289 images were collected with an oscillation width of 1.0° in ω. The exposure times were 5 s and 25 s for the low- and high-angle images. The integration of the experimental data yielded a total of 386 106 reflections to a maximum 2θ value of 136.49° of which 11 168 were independent and 9078 were greater than 2σ(I). The final cell constants (Table S2†) were based on the xyz centroid 143 200 reflections above 10σ(I). Analysis of these data showed negligible intensity decay during data collection. They were processed with CrystalClear 2.0 and corrected for their absorption. The structure was solved and refined with the Bruker SHELXTL (version 2014/6) software package,⁴⁰ using the space group P4/ncc with Z = 4 for the formula C₁₈₇H_124.5N₅₆O₄₇F₉S₃Zn₁₆Yb. All non-hydrogen atoms were refined anisotropically with the hydrogen atoms placed on idealized positions. Full matrix least-squares refinement based on F² converged at R₁ = 0.0687 and wR₂ = 0.2087 [based on I > 2σ(I)], R₁ = 0.0768 and wR₂ = 0.2193 for all data. The SQUEEZE subroutine of the PLATON program suite⁴¹ was used to address highly disordered solvent and two triflate anions present in the structure. Additional details are presented in Table S2.†

Photophysical measurements

Data were collected on samples in the solid state or on freshly prepared solutions placed in 2.4 mm i.d. quartz capillaries or quartz Suprasil cells. Absorption spectra were acquired on a Varian Cary 100Bio or a Jasco V670 spectrophotometer in absorbance mode. Steady-state emission and excitation spectra were measured on a Horiba-Jobin-Yvon Fluorolog 3 spectrofluorimeter equipped with either a visible photomultiplier tube (PMT) (220–800 nm, R928P; Hamamatsu), a NIR solid-state InGaAs detector cooled to 77 K (800–1600 nm, DSS-IGA020L; ElectroOptical Systems, Inc., USA), or a NIR PMT (950–1650 nm, H10330-75; Hamamatsu). Phosphorescence spectra of Gd^III MCs were acquired in the solid state at 77 K on a Fluorolog 3 spectrofluorimeter in a time-resolved mode. All spectra were corrected for the instrumental functions. Luminescence lifetimes were determined under excitation at 355 nm provided by a Nd:YAG laser (YG 980; Quantel). Signals in the NIR range were detected by the NIR H10330-75 PMT connected to an iHR320 monochromator (Horiba Scientific). The output signals generated from the detectors were fed into a 500 MHz bandpass digital oscilloscope (TDS 754C; Tektronix) and transferred to a PC for data processing with the Origin 8® software. Luminescence lifetimes are averages of three or more independent measurements. Quantum yields were determined with a Fluorolog 3 spectrofluorimeter based on the absolute method using an integration sphere (GMP SA). Each sample was measured several times. Experimental error for the determination of quantum yields is estimated as ∼10%.

Results

Synthesis

Zn₁₆Ln(napHA)₁₆ and Zn₁₆Ln(quinoHA)₁₆ were prepared using a modified synthetic procedure in respect to the one reported previously for Zn₁₆Ln(HA)₁₆ (HA = picHA²⁻, pyzHA²⁻, quinHA²⁻).^32,33,37 The self-assembly reaction between H₂napHA or H₂quinoHA, Zn^II and Ln^III triflates was performed in a DMF/H₂O/pyridine solution. The use of an excess of Ln^III triflates (molar ratio Zn(OTf)₂ : Ln(OTf)₃ 8 : 1) ensures the formation of pure products. All reactions can be performed at room temperature but significantly higher yields were obtained when the reaction mixtures were heated to 80 °C. All synthesized MCs were systematically characterized by ESI mass-spectrometry (ESI†) and elemental analysis. According to the elemental analysis, the general composition of the MCs can be presented as follows [Zn₁₆Ln(HA)₁₆(py)₈](OTf)₃(H₂O)₁₂, in which each MC molecule contains 1 Ln^III, 16 Zn^II, 16 naphthyridine- (napHA²⁻) or quinoxaline- (quinoHA²⁻) hydroximates, 8 pyridine molecules coordinated to Zn^II ions, three triflate anions as counter-ions and 12 water molecules. An ESI-MS analysis has been performed in a positive mode, so only [Zn₁₆Ln(HA)₁₆]³⁺ species could be detected.

Single-crystal X-ray diffraction

X-ray quality single crystals were obtained for Nd^III and Yb^III analogues of Zn₁₆Ln(quinoHA)₁₆. The molecular structure of Zn₁₆Yb(quinoHA)₁₆ as obtained from X-ray single crystal diffraction is shown in Fig. 1, while that of Zn₁₆Nd(quinoHA)₁₆ is in Fig. S5.† Both compounds crystallize in the tetragonal space group P4/ncc. The structures of both MCs reveal the encapsulation of a Ln^III ion within the Zn₁₆ shell, with the complexes having the general formula of Ln^III[12-MC_{Zn(II), quinoHA}-4]₂[24-MC_{Zn(II), quinoHA}-8]. The 6-coordinate Zn^II ions of the [24-MC_{Zn(II), quinoHA}-8] ring are crystallographically equivalent, and each of them is coordinated by three O atoms and two N atoms from the hydroximate ligands and one N atom from the pyridine. The Zn^II ions of the two [12-MC_{Zn(II), quinoHA}-4] rings are pyridine-free, 5-coordinated and adopt a distorted square pyramidal geometry. The Ln^III ion is sandwiched between two [12-MC_{Zn(II), quinoHA}-4] rings and further encapsulated by a [24-MC_{Zn(II), quinoHA}-8]. The MC molecule has a crystallographic S₈ symmetry. Each Ln^III ion is coordinated by eight O atoms from two [12-MC_{Zn(II), quinoHA}-4] rings. No solvent molecule is coordinated directly to Yb^III ion that adopts a square-antiprismatic coordination geometry around it, whilst 1.5 water molecules bound to Nd^III are observed and its coordination geometry is best described as a (bi)capped square antiprism.

NMR spectroscopy

Because of the paramagnetic nature of most Ln^III ions that induces fast relaxation and often causes strong line broadening in the NMR spectra, diamagnetic Y^III analogues (Zn₁₆Y(HA)₁₆) were prepared with both napHA²⁻ and quinoHA²⁻ families of ligands. Deuterated pyridine was used to prevent the observation of ¹H signals from the coordinated pyridine solvent molecules. ¹H and COSY-NMR spectra of Zn₁₆Y(quinoHA)₁₆ collected in DMSO-d₆ are shown in Fig. 2; the ones of Zn₁₆Y(napHA)₁₆ are provided in the ESI (Fig. S4†).

Fig. 2 (top) ¹H-NMR spectrum and (bottom) ¹H–¹H COSY-NMR spectrum of Zn₁₆Y-quinoHA acquired in DMSO-d₆.

As can be seen from the crystal structure of Zn₁₆Yb(quinoHA)₁₆, crystallographically inequivalent quinoHA²⁻ ligands are present in the molecular architecture. The first group includes eight ligands of the two [12-MC-4] and the second group contains the other eight ligands of the [24-MC-8]. The ¹H NMR spectrum of Zn₁₆Y(quinoHA)₁₆ displays, therefore, the presence of both types of quinoHA²⁻ ligands in the ratio of 1 : 1. The NMR peaks were assigned as represented in Fig. 2 (top), in agreement with the presence of a total of 10 protons belonging to the two quinoHA²⁻. This peak assignment was confirmed by the acquisition of an ¹H–¹H COSY-NMR spectrum Fig. 2 (bottom), which allows the determination of neighboring protons exhibiting correlations due to spin–spin coupling. Identical NMR spectra were obtained for the same samples after one week, confirming the long-term stability in solution of the Zn₁₆Ln(HA)₁₆ complexes.

Photophysical properties

Ligand-centered photophysical properties. H₂quinoHA and H₂napHA hydroxamic acids exhibit absorption bands in the UV region which extend to the visible range up to 450 nm for the latter. These bands can be mainly attributed to π → π* transitions (Fig. 3, black traces) on the basis of their extinction coefficients. It is important to note that, taking into account the heteroaromatic nature of the chromophoric moieties, the contribution of n → π* transitions to the absorption bands should not be excluded.⁴² From the edge of the absorption spectra, the positions of the lowest singlet state (S₁) were found to be located at 380 nm (26 320 cm⁻¹) and 447 nm (22 370 cm⁻¹) for H₂quinoHA and H₂napHA, respectively (Table S4†). The deprotonation and the formation of Zn₁₆Ln(HA)₁₆ complexes induce a blue shift of the π → π* absorption bands and the apparition of a new intra-ligand charge transfer (ILCT) band in the (near-)visible range which extends up to 490 nm and 510 nm for Zn₁₆Ln(napHA)₁₆ and Zn₁₆Ln(quinoHA)₁₆, respectively (Fig. 3, red traces). The positions of the apparent maxima appear at 387 nm (ε₃₈₇ = 68 290 M⁻¹ cm⁻¹) for Zn₁₆Ln(napHA)₁₆ and at 402 nm (ε₄₀₂ = 73 370 M⁻¹ cm⁻¹) for Zn₁₆Ln(quinoHA)₁₆. The molar absorption coefficients of Zn₁₆Ln(HA)₁₆ assembled using picHA²⁻, pyzHA²⁻, and quinHA²⁻ ligands are summarized in Table S3,† while a comparison of absorption spectra is presented in Fig. 8.

Fig. 3 Absorption spectra of hydroxamic acids (multiplied by a factor of 16) and of Zn₁₆Gd(HA)₁₆ (HA = quinoHA²⁻, napHA²⁻) recorded in CH₃OH at room temperature.

Fig. 4 (left) Excitation (λ_em = 980 nm) and (right) emission (λ_ex = 320–420 nm) spectra of Zn₁₆Yb(HA)₁₆ measured in the solid state (solid traces), or in 200 μM solutions in CH₃OH (HA = picHA²⁻, quinHA²⁻, napHA²⁻, quinoHA²⁻) or H₂O (HA = pyzHA²⁻) (dashed traces) at room temperature.

Fig. 5 (left) Excitation (λ_em = 1064 nm) and (right) emission (λ_ex = 320–420 nm) spectra of Zn₁₆Nd(HA)₁₆ collected on solid state samples (solid traces), in 1 mg mL⁻¹ solutions in methanol (HA = picHA²⁻, quinHA²⁻, napHA²⁻, quinoHA²⁻) or in water (HA = pyzHA²⁻) (dashed traces) at room temperature.

Fig. 6 (left) Excitation (λ_em = 1525 nm) and (right) emission (λ_ex = 320–420 nm) spectra of Zn₁₆Er(HA)₁₆ collected on solid state samples (solid traces), or in 1 mg mL⁻¹ solutions in methanol (HA = picHA²⁻, quinHA²⁻, napHA²⁻, quinoHA²⁻) or in water (HA = pyzHA²⁻) (dashed traces) at room temperature.

Fig. 7 Absorption spectra of Zn₁₆Yb(HA)₁₆ MCs measured in the energy range of the ²F_5/2 ← ²F_7/2 transition in CH₃OH (HA = picHA²⁻, quinHA²⁻, napHA²⁻, quinoHA²⁻) or in H₂O (HA = pyzHA²⁻) at room temperature (1.25–5.1 mM).

Fig. 8 Comparison of absorption spectra of Zn₁₆Yb(HA)₁₆ (HA = picHA²⁻, quinHA²⁻, napHA²⁻, quinoHA²⁻) or H₂O (HA = pyzHA²⁻) collected in CH₃OH solutions at room temperature.

Energy diagram. Since the sensitization of the characteristic emission of Ln^III ions in coordination compounds usually occurs through the ‘antenna effect’, the positions of ligand-centered electronic levels, in particular the lowest singlet (S₁), triplet (T₁) and ILCT states in respect to the accepting levels of Ln^III, are of particular importance.^43–46 The specific energies of the ILCT states were determined from the lower energy edge of the absorption spectra of Zn₁₆Ln(HA)₁₆ and were found to be located at 19 685 cm⁻¹ (508 nm) for Zn₁₆Ln(quinoHA)₁₆ and 20 830 cm⁻¹ (480 nm) for Zn₁₆Ln(napHA)₁₆. Corresponding Gd^III complexes are often used to determine the position of the T₁ state since the energy of the ⁶P_7/2 level (32 100 cm⁻¹)⁴⁷ is too high to be populated through the energy levels of most organic ligands. Moreover, the heavy-atom and paramagnetic effects of Gd^III facilitate intersystem-crossing. The corresponding Zn₁₆Gd(HA)₁₆ complexes were synthesized and their emission spectra in the solid state were analyzed. It was observed that the excitation into ILCT bands does not produce any detectable emission at room temperature or at 77 K. Higher energy excitation wavelengths at 290–320 nm, upon application of a time delay after the excitation flash, allowed us to collect phosphorescence spectra (Fig. S7†) and to determine the positions of the T₁ states: 24 150 cm⁻¹ (414 nm) for Zn₁₆Ln(quinoHA)₁₆ and 25 510 cm⁻¹ (392 nm) for Zn₁₆Ln(napHA)₁₆. (Table S4†). The diagram with the relative positions of the singlet, triplet and ILCT states for Zn₁₆Ln(HA)₁₆ as well as of the accepting energy levels of Yb^III, Nd^III and Er^III is shown in Fig. 9.

Fig. 9 Schematic energy diagram showing the energy positions of singlet states in H₂HA, as well as triplet and ILCT levels in Zn₁₆Gd(HA)₁₆ in respect to the accepting energy levels of NIR-emitting Yb^III, Nd^III, Er^III ions.

Ln^III-centered photophysical properties. Luminescence spectra were recorded for Zn₁₆Ln(HA)₁₆ (Ln = Yb^III, Nd^III, Er^III; HA = quinoHA²⁻, napHA²⁻) in the solid state and in CH₃OH and CD₃OD solutions under excitation in the range of 320–420 nm. All studied MCs exhibit intense Ln^III-centered emission in the NIR range arising from the ²F_5/2 → ²F_7/2, ⁴F_3/2 → ⁴I_J (J = 9/2, 11/2, 13/2), and ⁴I_13/2 → ⁴I_15/2 transitions for Yb^III (Fig. 4, right), Nd^III (Fig. 5, right) and Er^III (Fig. 6, right) analogues, respectively. Excitation spectra of Zn₁₆Yb(HA)₁₆ MCs measured in the solid state and in solution upon monitoring the emission at 980 nm are dominated by ligand-centered broad bands. As Yb^III does not possess any electronic levels located in the UV-visible range, this result indicates the presence of an efficient sensitization of Yb^III emission through the electronic structure of the MC scaffold (Fig. 4, left). In the excitation spectra of Zn₁₆Nd(HA)₁₆ (Fig. 5, left) and Zn₁₆Er(HA)₁₆ (Fig. 6, left) MCs collected in the solid state, characteristic sharp intraconfigurational f–f transitions could also be observed in addition to broad ligand-centered bands. The intensities of f–f transitions are similar to those of ligand-centered bands for Nd^III and Er^III MCs formed with the quinoHA²⁻ ligand but are weaker for the corresponding Zn₁₆Ln(napHA)₁₆ MCs. Such observations point to the possibility of sensitizing the characteristic Nd^III and Er^III emission not only through the MC scaffold but also through the direct excitation of f–f transitions. In solution, f–f transitions are undetectable and the excitation spectra of Zn₁₆Nd(HA)₁₆ (Fig. 5, left) and Zn₁₆Er(HA)₁₆ (Fig. 6, left) MCs are dominated by broad ligand-centered bands. For all analyzed MCs, profiles of the excitation spectra collected in solutions are matching those of the corresponding absorption spectra (Fig. 7 and S6†) while in the solid state, a broadening combined with an extension of the bands to the red domain up to 550 nm for Zn₁₆Ln(napHA)₁₆ and 630 nm for Zn₁₆Ln(quinoHA)₁₆ is observed. Expanded excitation ranges for solid state samples compared to those obtained in solution can be explained by saturation effects.^48,49

Ln^III-centered quantum yields (Q^L_Ln) and observed luminescence lifetimes (τ_obs) (Table 1) were collected under ligand excitation for all Zn₁₆Ln(HA)₁₆ in the solid state and in CH₃OH and CD₃OD solutions at room temperature. Using phenomenological equations^50,51 and values of τ_obs obtained in CH₃OH and CD₃OD solutions, we can confirm that no solvent molecules are directly coordinated to Yb^III or Nd^III in Zn₁₆Ln(quinoHA)₁₆ and in Zn₁₆Ln(napHA)₁₆. This result suggests that in the case of Zn₁₆Nd(quinoHA)₁₆ in CH₃OH and CD₃OD solutions, water molecules coordinated to Nd^III are dissociated while the coordination of larger solvent molecules is sterically hindered.

Table 1 Photophysical parameters for Zn₁₆Ln(HA)₁₆ in the solid state and in 200 μM solutions in methanol (HA = picHA²⁻, quinHA²⁻, napHA²⁻, quinoHA²⁻) or water (HA = pyzHA²⁻)^a

Compound	State/solvent	τ obs (μs)	τ rad (μs)	Q ^LnLn (%)	Q ^LLn (%)	η sens (%)
a Data at room temperature. Standard deviation (2σ) between parentheses; estimated relative errors: τobs, ±2%; τrad, ±10–12%; Q^LLn, ±10%; Q^LnLn, ±10–12%; ηsens, ±22%. b Under excitation at 355 nm. c See text for details. d Under excitation at 320–420 nm. e Average lifetime calculated from τ1 = 1.12(1) μs (74%); τ2 = 0.39(1) μs (26%).
Zn16Yb(picHA)16	Solid	34.5(1)	410	8.4	0.40(2)	4.8
	CH3OH	12.1(1)	524	2.3	0.13(1)	5.6
	CD3OD	133(1)	526	25.3	1.60(3)	6.3
Zn16Nd(picHA)16	Solid	1.18(2)	293	0.40(9)	0.40(1)	∼100
	CH3OH	0.90(1)	375	0.24	0.22(2)	92
	CD3OD	3.53(1)	377	0.94	0.98(1)	∼100
Zn16Er(picHA)16	Solid	3.47(4)	11.8 × 10^3	2.95(9) × 10^−2	1.34(4) × 10^−2	45
	CH3OH	1.02(1)	15.1 × 10^3	6.75 × 10^−3	1.30(3) × 10^−3	19
	CD3OD	10.4(1)	15.2 × 10^3	6.8 × 10^−2	5.24(6) × 10^−2	77
Zn16Yb(pyzHA)16	Solid	45.6(3)	399	11.4	0.659(4)	5.8
	H2O	5.57(1)	499	1.12	1.12(7) × 10^−2	1.0
	D2O	81.3(1)	508	16	0.257(3)	1.6
Zn16Nd(pyzHA)16	Solid	1.71(1)	133	1.82(3)	0.444(9)	34
	H2O	0.214(4)	166	0.13	7.7(1) × 10^−3	6.0
	D2O	1.29(1)	169	0.76	6.17(9) × 10^−2	8.1
Zn16Er(pyzHA)16	Solid	4.96(4)	3.82 × 10^3	0.130(6)	9.5(2) × 10^−3	7.3
	D2O	7.02(2)	4.91 × 10^3	0.14	1.70(4) × 10^−3	1.2
Zn16Yb(quinHA)16	Solid	47.8(4)	530	9.0	2.44(4)	27
	CH3OH	14.88(1)	675	2.2	0.25(1)	11.3
	CD3OD	150.7(2)	679	22	2.88(2)	13.0
Zn16Nd(quinHA)16	Solid	1.79(2)	166	1.54(4)	1.13(4)	∼100
	CH3OH	1.16(1)	212	0.55	0.38(1)	69
	CD3OD	4.11(2)	214	1.9	1.35(1)	70
Zn16Er(quinHA)16	Solid	5.73(2)	3.58 × 10^3	0.16(1)	4.2(1) × 10^−2	26.3
	CH3OH	1.25(1)	4.58 × 10^3	2.7 × 10^−2	9.9(3) × 10^−4	3.6
	CD3OD	11.40(3)	4.6 × 10^3	0.25	3.6(1) × 10^−2	14.5
Zn16Yb(quinoHA)16	Solid	49.7(2)	250	19.9	0.637(5)	3.2
	CH3OH	22.6(1)	320	7.1	0.135(1)	1.9
	CD3OD	124.0(1)	322	38	0.88(2)	2.3
Zn16Nd(quinoHA)16	Solid	1.04(1)^e	202	0.51(1)	0.153(2)	30
	CH3OH	0.79(1)	259	0.31	7.1(1) × 10^−2	23
	CD3OD	1.50(1)	261	0.57	0.160(5)	28
Zn16Er(quinoHA)16	Solid	5.89(2)	7.46 × 10^3	7.9(1) × 10^−2	1.11(5) × 10^−2	14.1
	CH3OH	1.32(1)	9.55 × 10^3	1.3 × 10^−2	3.7(3) × 10^−4	2.7
	CD3OD	10.8(1)	9.61 × 10^3	0.11	9.8(2) × 10^−3	8.7
Zn16Yb(napHA)16	Solid	41.3(2)	351	11.8	1.10(2)	9.3
	CH3OH	20.0(1)	454	4.4	0.156(3)	3.5
	CD3OD	159(1)	456	35	1.57(3)	4.5
Zn16Nd(napHA)16	Solid	1.13(1)	139	0.82(2)	0.34(1)	41
	CH3OH	0.993(4)	178	0.56	0.168(3)	30
	CD3OD	3.73(3)	180	2.1	0.731(3)	35
Zn16Er(napHA)16	Solid	3.96(1)	4.21 × 10^3	9.4(3) × 10^−2	1.29(1) × 10^−2	13.7
	CH3OH	1.27(1)	5.39 × 10^3	2.3 × 10^−2	5.2(4) × 10^−4	2.2
	CD3OD	11.7(1)	5.41 × 10^3	0.22	1.88(2) × 10^−2	8.7

---

In addition, for all Nd^III and Er^III samples of the Zn₁₆Ln(HA)₁₆ family, their intrinsic quantum yields (Q^Ln_Ln) were measured in the solid state under direct excitation into the ⁴F_7/2 ← ⁴I_9/2 transition at 750 nm for Zn₁₆Nd(HA)₁₆ and the ⁴F_9/2 ← ⁴I_15/2 transition located at 650 nm for Zn₁₆Er(HA)₁₆. Taking into account Q^L_Ln, Q^Ln_Ln and τ_obs for Nd^III and Er^III samples in the solid state, the values of τ_rad and η_sens were calculated according to eqn (1). It has been shown previously that these MCs remain intact in solution. Therefore, τ_rad were adjusted for the differences in refractive indexes (n), i.e. τ_rad(solution) = τ_rad(solid state) × (n_{solid state}/n_solution)³.²⁹ In the solid state, the refractive index was set to be equal to 1.5, while the n_solution values are admitted to be the ones of the neat solvents, e.g. n(CD₃OD) = 1.326; n(CH₃OH) = 1.329; n(D₂O) = 1.328; n(H₂O) = 1.34.

For Zn₁₆Yb(HA)₁₆ MCs, intrinsic quantum yields could not be measured experimentally. However, Yb^III electronic transitions generating the luminescence terminate in the ground level. Therefore, if the absorption spectrum corresponding to the emission spectrum is known, τ_rad can be calculated using the modified Einstein's equation (eqn (2a) and (2b)):⁵²

where c is the speed of light in centimeters per second, n is the refractive index, N_A is Avogadro's number, J and J′ are the quantum numbers for the ground and the excited states, respectively, is the integrated spectrum of the ²F_5/2 ← ²F_7/2 transition, [small nu, Greek, tilde]_m is the barycenter of the transition. The corresponding absorption spectra were measured for all Zn₁₆Yb(HA)₁₆ (Fig. 7). The resulting τ_obs values were adjusted to the media based on the difference in refractive indexes, as explained before, and used to calculate Q^Ln_Ln and η_sens according to eqn (1).

All quantitative photophysical parameters are summarized in Table 1 and presented on Fig. 10.

Fig. 10 Comparison of total and intrinsic quantum yields, sensitization efficiencies, observed and radiative lifetimes for Zn₁₆Ln(HA)₁₆ (Ln = Yb^III, Nd^III, Er^III; HA = picHA²⁻ (A), pyzHA²⁻ (B), quinHA²⁻ (C), quinoHA²⁻ (D), napHA²⁻ (E)) in the solid state (black bars) or solution in CH₃OH or H₂O (red bars) and CD₃OD or D₂O (blue bars).

Discussion

MCs with the general formula, Ln^III[12-MC_Zn(II),HA-4]₂[24-MC_Zn(II),HA-8] (Zn₁₆Ln(HA)₁₆, Ln^III = Yb, Nd, Er, Gd & Y; HA = picHA²⁻, pyzHA²⁻, quinHA²⁻, quinoHA²⁻, napHA²⁻) can be obtained by self-assembly reactions between the corresponding hydroxamic acid, Zn^II and Ln^III triflates in an appropriate molar ratio. Although the general synthetic strategy is the same, the solvents and the base used in the reactions have to be adjusted according to the nature of the hydroxamic acid. In the case of MCs formed with picHA²⁻ and quinHA²⁻, an addition of sodium hydroxide or trimethylamine is necessary to ensure the formation of the desired Zn₁₆Ln(HA)₁₆ (ref. 32 and 33) while MCs formed with pyzHA²⁻,³⁷ quinoHA²⁻ and napHA²⁻ can be obtained in the presence of pyridine in H₂O or in a H₂O/DMF mixture, respectively. The crystal structures of all studied Zn₁₆Ln(HA)₁₆ MCs show that Ln^III ion is sandwiched between two [12-MC_Zn(II),HA-4] and are further encapsulated by a [24-MC_Zn(II),HA-8] unit. The Ln^III ion is bonded to four hydroximate oxygens of each of the [12-MC_Zn(II),HA-4] motifs, forming a square antiprismatic coordination geometry. No solvent molecule is coordinated directly to the Ln^III ion except for Zn₁₆Nd(quinoHA)₁₆ where 1.5 water molecules connected to Nd^III were observed. Nevertheless, in general, the differences in the electronic structure of the hydroximate ligands have a minimal effect on the first coordination sphere around the Ln^III ion, thus providing an opportunity to study systematically the effect of the nature of the ligand on the resulting photophysical properties of Zn₁₆Ln(HA)₁₆ MCs. In particular, we will discuss how the nature of the hydroximate ligands forming the Zn₁₆Ln(HA)₁₆ scaffold affects absorption, excitation and emission spectra as well as the different quantitative photophysical parameters for the newly reported NIR-emitting Zn₁₆Ln(quinoHA)₁₆ Zn₁₆Ln(napHA)₁₆ and the previously described Zn₁₆Ln(picHA)₁₆,³³ Zn₁₆Ln(pyzHA)₁₆ (ref. 37) and Zn₁₆Ln(quinHA)₁₆.³²

First, we demonstrated that, by varying the nature of hydroxamic acids, i.e. the number of nitrogen atoms in the aromatic rings and the degree of conjugation, it is possible to tune the maxima of the low-energy ILCT bands in the absorption spectra of Zn₁₆Ln(HA)₁₆ from 330 nm (HA = picHA²⁻), 360 nm (HA = picHA²⁻), 380 nm (HA = pyzHA²⁻) to 387 nm (HA = napHA²⁻) and 402 nm (HA = quinoHA²⁻) (Fig. 8).

In view of the sensitization of Ln^III emission, the mutual energy positions of the ligand-centered donating states and the accepting levels of Ln^III are of particular importance.^20,43–46 In general, the sensitization process involves an energy transfer originating from the triplet state of the chromophore unit. However, the population of Ln^III excited energy levels through singlet and charge transfer states located at lower energy should also be considered. The corresponding energy diagram for Zn₁₆Ln(HA)₁₆ and Yb^III, Nd^III and Er^III is presented in Fig. 9. Several features are important to note. In the case of MCs formed with picHA²⁻ and quinHA²⁻, the positions of the T₁ and the ILCT states are located very closely to each other with an energy difference which is less than 600 cm⁻¹, while the singlet state is located at least 6500 cm⁻¹ higher in energy. For Zn₁₆Ln(pyzHA)₁₆, Zn₁₆Ln(quinoHA)₁₆ and Zn₁₆Ln(napHA)₁₆, the T₁ state is positioned at a higher energy than the ILCT and the energy difference between these states is larger than 2800 cm⁻¹. Singlet states in Zn₁₆Ln(pyzHA)₁₆, Zn₁₆Ln(quinoHA)₁₆ are located higher in energy than the T₁ by values of 3535 cm⁻¹ and 2170 cm⁻¹, respectively, while in Zn₁₆Ln(napHA)₁₆, the energy of S₁ state is lower than the one of T₁. The last observation can be explained by the different nature of these electronic states, i.e. nπ* vs. ππ*. It should also be mentioned that the positions of the singlet states were determined from the edge of the absorption spectra of the corresponding hydroxamic acids and that their energies may shift in Zn₁₆Ln(HA)₁₆. Nevertheless, one can see that for all studied Zn₁₆Ln(HA)₁₆ MCs, the ligand-centered levels are located higher in energy than the ²F_5/2, ⁴F_3/2 and ⁴I_13/2 levels of Yb^III, Nd^III and Er^III, respectively, so all can act as potential sensitizers of their characteristic NIR emission.

Excitation spectra of all the studied Zn₁₆Yb(HA)₁₆ MCs collected in the solid state and in solution upon monitoring the Yb^III emission at 980 nm are dominated by broad ligand-centered bands indicating an efficient sensitization of Yb^III emission through the electronic structure of the MC scaffold (Fig. 4, left). The excitation spectra of Zn₁₆Nd(HA)₁₆ (Fig. 5, left) and Zn₁₆Er(HA)₁₆ (Fig. 6, left) MCs in the solid state reveal the presence of sharp bands that correspond to f–f transitions in addition to the broad ligand-centered bands. Their intensities depend on the nature of the hydroxamic acid. The excitation spectrum of Zn₁₆Nd(quinoHA)₁₆ is dominated by the sharp bands arising from the f–f transitions while their intensities are significantly less important for Zn₁₆Nd(quinHA)₁₆. Among Er^III MCs, f–f transitions have the lowest intensities for Zn₁₆Er(picHA)₁₆ and Zn₁₆Er(quinHA)₁₆ while they have the highest intensities for Zn₁₆Er(pyzHA)₁₆ and Zn₁₆Er(quinoHA)₁₆. Such observations indicate that the sensitization of the characteristic Nd^III and Er^III emissions for Zn₁₆Ln(HA)₁₆ in the solid state may not only occur through the MC scaffold but also through direct excitation of f–f transitions. In solution, the excitation spectra of all Zn₁₆Nd(HA)₁₆ and Zn₁₆Er(HA)₁₆ MCs exhibit only the presence of ligand-centered bands and no f–f transitions could be detected.

Upon excitation into the ILCT bands in the range of 320–420 nm, all the studied Zn₁₆Ln(HA)₁₆ MCs exhibit characteristic Ln^III emission in the NIR range. For Yb^III MCs, emission bands are observed in the range of 950–1060 nm and originate from the ²F_5/2 → ²F_7/2 transition (Fig. 4, right), while for Er^III bands appear in the range of 1450–1650 nm arising from the ⁴I_13/2 → ⁴I_15/2 transition (Fig. 6, right). The emission spectra of Zn₁₆Nd(HA)₁₆ are dominated by the ⁴F_3/2 → ⁴I_11/2 transition in the range of 1050–1120 nm, while two additional bands with lower intensities appear at 850–930 nm and 1300–1420 nm. They are assigned to ⁴F_3/2 → ⁴I_9/2 and ⁴F_3/2 → ⁴I_13/2 transitions, respectively (Fig. 5, right). It should be noted that for a specific Ln^III ion, the obtained emission spectra are almost independent of the nature of the hydroxamic acid forming the MC, reflecting the similarity of coordination environments around Ln^III which is in full agreement with the structural data (vide supra). Moreover, the emission spectra recorded for the Zn₁₆Ln(HA)₁₆ samples in the solid state and solutions are very similar by their bandwidths and crystal-field splitting as an additional indication that MCs remain intact in solution.

Experimental luminescence decay curves collected for most of the studied MCs in the solid state and solutions are best fitted with mono-exponential functions, reflecting the presence of only one emissive Ln^III-containing species. The sole exception has been observed with Zn₁₆Nd(quinoHA)₁₆ in the solid state for which a bi-exponential luminescence decay was observed with corresponding ‘long’ (1.12(1) μs) and ‘short’ (0.39(1) μs) τ_obs values. The ‘short’ component corresponds most probably to MCs containing 1.5 water molecules directly coordinated to Nd^III (as observed in the crystal structure of this MC, vide supra), while the ‘long’ component correspond to the Zn₁₆Nd(quinoHA)₁₆ that doesn't contain H₂O molecules directly coordinated to Nd^III. Luminescence decays collected on CH₃OH and CD₃OD solutions of Zn₁₆Nd(quinoHA)₁₆ are mono-exponential, indicating that coordinated to Nd^III water molecules are dissociated while the coordination of other solvent molecules is sterically hindered. Observed luminescence lifetimes (τ_obs, Table 1) significantly decrease when going from the solid state samples to solutions in CH₃OH or H₂O (Fig. 10, black and red bars) reflecting a detrimental effect of O–H and C–H vibrations on non-radiative deactivation of the NIR-emitting Ln^III ions.¹² The latter is also revealed in a pronounced (2–15 times) lengthening of τ_obs values for solutions of MCs in deuterated solvents (Fig. 10, blue bars). The use of phenomenological equations^50,51 with the values of τ_obs recorded in CH₃OH and CD₃OD, or in H₂O and D₂O solutions for Zn₁₆Yb(HA)₁₆ and Zn₁₆Nd(HA)₁₆ confirmed that no solvent molecules are directly coordinated to either Yb^III or Nd^III within the whole series of studied MCs, indicating that shortening of τ_obs values is induced by non-radiative deactivations through second-sphere interactions with overtones of O–H and C–H vibrations.

Values of τ_obs recorded on solutions of Zn₁₆Yb(HA)₁₆ and Zn₁₆Nd(HA)₁₆ in deuterated solvents are longer compared to these collected in CH₃OH and H₂O which results in a comparable enhancement of the Q^Ln_Ln and Q^L_Ln values. On the other hand, for Zn₁₆Er(HA)₁₆ in CD₃OD solution the τ_obs and Q^Er_Er values are 8–10 times larger compared to the values obtained in CH₃OH, while the Q^L_Er values increase by 26–40 times. An explanation can be found in the values of sensitization efficiencies (η_sens) that are similar for solutions of Yb^III and Nd^III MCs in protic and deuterated solvents while at least a 3-times enhancement is observed for the values of Er^III collected in CD₃OD. In general, the lowest values of Q^L_Ln are obtained for solutions of Zn₁₆Ln(HA)₁₆ in CH₃OH and in H₂O. The comparison of Q^L_Ln values for Zn₁₆Ln(HA)₁₆ samples in the solid state and in solution in deuterated solvents (Fig. 10, black and blue bars) revealed different trends depending on the nature of the hydroxamic acid. Similar values of Q^L_Ln are observed for Zn₁₆Ln(quinHA)₁₆ and Zn₁₆Ln(quinoHA)₁₆, since the higher values of τ_obs and Q^Ln_Ln in CD₃OD solutions are compensated by a comparable enhancement of the sensitization efficiency in the solid state. For Zn₁₆Ln(pyzHA)₁₆, the values of η_sens are 3.6–6.1 times larger in the solid state than in D₂O solutions leading to a significantly improved value of Q^L_Ln for the former. In the case of Zn₁₆Ln(picHA)₁₆ and Zn₁₆Ln(napHA)₁₆, Q^L_Ln are lower for samples in the solid state vs. solution in CD₃OD since the values of η_sens are similar or lower in the solid state. Thus, the non-radiative deactivation of Ln^III NIR emission through O–H and C–H vibrations in solutions of MCs in protic solvents has a dominating role and negatively impact the main quantitative characteristics. On the other hand, the values of Q^L_Ln for samples of Zn₁₆Ln(HA)₁₆ in the solid state and in deuterated solvents depend on the nature of the hydroximate ligand and Ln^III ion.

Several points should be highlighted if we consider further the effect of the nature of the hydroximate ligands on the quantitative photophysical parameters of Zn₁₆Ln(HA)₁₆ in the solid state. For a specific Ln^III ion, the values of τ_obs are within a relatively narrow range, i.e. 34.5–49.7 μs for Yb^III, 1.04–1.79 μs for Nd^III and 3.47–5.89 μs for Er^III MCs. Considering the similarity of the crystal structures, such variations are mainly caused by the coordination of water molecules (for Zn₁₆Nd(quinoHA)₁₆) or by the presence of sources of non-radiative deactivations through co-crystallized solvent molecules. On the other hand, the effect of the nature of the hydroximate ligand on τ_rad values is more pronounced. Changes as significant as >3-fold are detected in the case of Zn₁₆Er(picHA)₁₆ (11.8 ms) vs. Zn₁₆Er(quinHA)₁₆ (3.58 ms) or Zn₁₆Er(pyzHA)₁₆ (3.82 ms). Among Nd^III MCs, the longest τ_rad value is observed for Zn₁₆Nd(picHA)₁₆ (293 μs) and the shortest for Zn₁₆Nd(pyzHA)₁₆ (133 μs). For Yb^III MCs, the variation of the τ_rad values is also significant, from 530 μs for Zn₁₆Yb(quinHA)₁₆ to 250 μs for Zn₁₆Yb(quinoHA)₁₆. So, despite the highly similar coordination environments (through eight oxygen atoms) and symmetry (square antiprism) around the Ln^III ions in Zn₁₆Ln(HA)₁₆, the nature of the hydroximate ligand appears to impact radiative lifetimes. An enhanced understanding of the factors that control this parameter is important for the design of highly luminescent Ln^III compounds since the shorter τ_rad leads to a higher Q^Ln_Ln (Eq. (1)). The validity of this approach has been demonstrated in several studies.^26,29–31 Indeed, values of Q^Ln_Ln are the highest for Nd^III (1.5–1.8%) and Er^III (0.13–0.16%) MCs formed with pyzHA²⁻ and quinHA²⁻, while in the case of Yb^III for MCs formed with quinoHA²⁻ (19.9%), all of them correspond to the lowest values of τ_rad. Another important parameter that is largely affected by the nature of the hydroximate ligand is the sensitization efficiency. It varies from 4.8 to 27%, from 30 to 100%, and from 7.3 to 45% for Zn₁₆Yb(HA)₁₆, Zn₁₆Nd(HA)₁₆ and Zn₁₆Er(HA)₁₆, respectively. The highest values of η_sens have been found in Yb^III MCs formed with quinHA²⁻ and in Zn₁₆Er(picHA)₁₆. The scaffolds of both MCs sensitize equally well the characteristic emission of Nd^III with a η_sens of ∼100%. Here, a correlation with the mutual positions of the ILCT and the triplet states can be followed (vide supra). In particular, these energy states are located very close to each other in Zn₁₆Ln(picHA)₁₆ and in Zn₁₆Ln(quinHA)₁₆ while for other MCs, the energy of T₁ is higher than the one of the ILCT state. On the basis of the values of η_sens, a high proximity in energy positions of T₁ and ILCT should be preferred for the sensitization of the NIR-emitting Yb^III, Nd^III and Er^III. Finally, all these variations lead to the largest values of Q^L_Ln observed for Zn₁₆Ln(quinHA)₁₆ in the solid state.

Conclusions

In this work, we have extended the Zn₁₆Ln(HA)₁₆ ‘encapsulated sandwich’ family by creating MCs with two new ligands, naphthyridine- and quinoxaline-hydroximates. We have demonstrated how the nature of the hydroximate ligand impacts the photophysical properties of these MCs. In particular, by varying ligands from picoline-, pyrazine-, quinaldine-, naphthyridine- and quinoxaline-hydroximates, the low-energy ILCT bands present in the absorption spectra of Zn₁₆Ln(HA)₁₆ can be shifted from the UV to the visible range up to 510 nm and the corresponding maxima from 330 nm to 402 nm. MC scaffolds sensitize well the NIR characteristic emissions of Yb^III, Nd^III and Er^III ions in all studied Zn₁₆Ln(HA)₁₆ with total quantum yield values and observed luminescence lifetimes that are comparable to the highest values reported so far for Ln^III coordination compounds. The extensive detailed analysis of the quantitative photophysical parameters revealed that the values of Q^L_Ln, Q^Ln_Ln, η_sens and τ_obs depend on (i) the media, (ii) the nature of the Ln^III ion and (iii) the hydroximate ligands. The values of τ_obs are the longest collected for Zn₁₆Ln(HA)₁₆ in deuterated solvents followed by those measured in the solid state and in protic solvents. Considering that no solvent molecule is directly coordinated to Ln^III in solution, a negative effect of second-sphere non-radiative deactivation of Ln^III ions induced by overtones of O–H and C–H vibrations is significant. For a specific Ln^III ion, the measured values of Q^L_Ln are the highest for Zn₁₆Yb(quinHA)₁₆ and Zn₁₆Nd(quinHA)₁₆ in all media, for Zn₁₆Er(quinHA)₁₆ in the solid state and for Zn₁₆Er(picHA)₁₆ in solution. With few exceptions, the sensitization is more efficient in Zn₁₆Ln(HA)₁₆ samples in the solid state compared to the values measured in solutions and η_sens values as high as 100% are observed for Zn₁₆Nd(picHA)₁₆ and Zn₁₆Nd(quinHA)₁₆. Notably, for Zn₁₆Ln(picHA)₁₆ and Zn₁₆Ln(quinHA)₁₆, the energies of the triplet and the ILCT states are fairly similar (ΔE < 600 cm⁻¹) while E(T₁) is higher than E(ILCT) for other MCs. Despite the similarities of the coordination environments and symmetries around Ln^III within the Zn₁₆Ln(HA)₁₆ family, radiative lifetimes of a specific Ln^III ion are not constant and are largely affected by the nature of hydroximate ligand. Finally, having all of these comprehensive data in hand, it is possible to rationally design and create Zn₁₆Ln(HA)₁₆ with desired photophysical properties suitable for NIR imaging applications. More detailed studies that combine time-resolved measurements with theoretical calculations are required to finalize the precise assignment of energy levels and to obtain a deeper level of understanding of intricate energy transfer processes occurring in Zn₁₆Ln(HA)₁₆ which will be the subject of future reports.

Author contributions

S. V. E.: conceptualization, photophysical data acquisition and analysis, writing – original draft, editing, funding acquisition; T. N. N.: conceptualization, synthesis and characterization, writing – review, editing; J. F. K.: crystallographic data acquisition; E. R. T.: synthesis and characterization, writing – review, editing; V. L. P.: conceptualization, supervision, funding acquisition, writing – review, editing; S. P.: conceptualization, funding acquisition, writing – review, editing.

Data availability

The experimental data that support the findings of this study are provided in ESI† and available online or upon reasonable request from the corresponding authors.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This research was supported in part by the National Science Foundation under grant CHE-1664964, La Ligue Contre le Cancer and La Région Centre-Val de Loire. S. P. acknowledges support from Institut National de la Santé et de la Recherche Médicale (INSERM).

Notes and references

1. S. He, J. Song, J. Qu and Z. Cheng, Chem. Soc. Rev., 2018, 47, 4258–4278.
2. Kenry, Y. Duan and B. Liu, Adv. Mater., 2018, 30, e1802394.
3. Q. Miao and K. Pu, Adv. Mater., 2018, 30, e1801778.
4. C. Wang, Z. Wang, T. Zhao, Y. Li, G. Huang, B. D. Sumer and J. Gao, Biomat, 2018, 157, 62–75.
5. Y. Fan and F. Zhang, Adv. Opt. Mater., 2019, 7, 1801417.
6. I. Martinić, S. V. Eliseeva and S. Petoud, J. Lumin., 2017, 189, 19–43.
7. Y. Y. Ning, M. L. Zhu and J. L. Zhang, Coord. Chem. Rev., 2019, 399, 213028.
8. J.-C. G. Bünzli and S. V. Eliseeva, J. Rare Earths, 2010, 28, 824–842.
9. L. Wang, Z. Zhao, C. Wei, H. Wei, Z. Liu, Z. Bian and C. Huang, Adv. Opt. Mater., 2019, 7, 1801256.
10. S. Comby and J.-C. G. Bünzli, in Handbook on the Physics and Chemistry of Rare Earths, ed. K. A. Gschneidner Jr, J.-C. G. Bünzli and V. K. Pecharsky, Elsevier Science B.V., Amsterdam, 2007, ch. 235, vol. 37, pp. 217–470.
11. J.-C. G. Bünzli and S. V. Eliseeva, in Comprehensive Inorganic Chemistry II, ed. V. W.-W. Yam, Elsevier B.V., Amsterdam, 2013, ch. 8.08, vol. 8, pp. 339–398.
12. J.-C. G. Bünzli and S. V. Eliseeva, in Springer Series on Fluorescence. Lanthanide Luminescence: Photophysical, Analytical and Biological Aspects, ed. P. Hänninen and H. Härmä, Springer Verlag, Berlin, 2011, vol. 7, pp. 1–45.
13. M. Latva, H. Takalo, V. M. Mukkala, C. Matachescu, J.-C. Rodriguez-Ubis and J. Kankare, J. Lumin., 1997, 75, 149–169.
14. R. D. Archer, H. Y. Chen and L. C. Thompson, Inorg. Chem., 1998, 37, 2089–2095.
15. S. Sato and M. Wada, Bull. Chem. Soc. Jpn., 1970, 43, 1955–1962.
16. M. H. V. Werts, R. H. Woudenberg, P. G. Emmerink, R. van Gassel, J. W. Hofstraat and J. W. Verhoeven, Angew. Chem., Int. Ed., 2000, 39, 4542–4544.
17. M. H. V. Werts, J. W. Hofstraat, F. A. J. Geurts and J. W. Verhoeven, Chem. Phys. Lett., 1997, 276, 196–201.
18. J.-Y. Hu, Y. Ning, Y.-S. Meng, J. Zhang, Z.-Y. Wu, S. Gao and J.-L. Zhang, Chem. Sci., 2017, 8, 2702–2709.
19. Y. Y. Ning, J. Tang, Y. W. Liu, J. Jing, Y. S. Sun and J. L. Zhang, Chem. Sci., 2018, 9, 3742–3753.
20. Z. Zhang, Y. Zhou, H. Li, T. Gao and P. Yan, Dalton Trans., 2019, 48, 4026–4034.
21. I. Hernández and W. P. Gillin, in Handbook on the Physics and Chemistry of Rare Earths, ed. J.-C. G. Bünzli and V. K. Pecharsky, Elsevier, 2015, ch. 269, vol. 47, pp. 1–100.
22. G. Mancino, A. J. Ferguson, A. Beeby, N. J. Long and T. S. Jones, J. Am. Chem. Soc., 2005, 127, 524–525.
23. W. Wu, X. Zhang, A. Y. Kornienko, G. A. Kumar, D. Yu, T. J. Emge, R. E. Riman and J. G. Brennan, Inorg. Chem., 2018, 57, 1912–1918.
24. C. Bischof, J. Wahsner, J. Scholten, S. Trosien and M. Seitz, J. Am. Chem. Soc., 2010, 132, 14334–14335.
25. C. Doffek, N. Alzakhem, M. Molon and M. Seitz, Inorg. Chem., 2012, 51, 4539–4545.
26. M. F. K. Trautnitz, C. Doffek and M. Seitz, ChemPhysChem, 2019, 20, 2179–2186.
27. D. Davis, A. J. Carrod, Z. Guo, B. M. Kariuki, Y.-Z. Zhang and Z. Pikramenou, Inorg. Chem., 2019, 58, 13268–13275.
28. P. B. Glover, A. P. Bassett, P. Nockemann, B. M. Kariuki, R. Van Deun and Z. Pikramenou, Chem.–Eur. J., 2007, 13, 6308–6320.
29. J.-C. G. Bünzli, A.-S. Chauvin, H. K. Kim, E. Deiters and S. V. Eliseeva, Coord. Chem. Rev., 2010, 254, 2623–2633.
30. C. Doffek and M. Seitz, Angew. Chem., Int. Ed., 2015, 54, 9719–9721.
31. S. V. Eliseeva, D. N. Pleshkov, K. A. Lyssenko, L. S. Lepnev, J.-C. G. Bünzli and N. P. Kuzmina, Inorg. Chem., 2011, 50, 5137–5144.
32. 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.
33. J. Jankolovits, C. M. Andolina, J. W. Kampf, K. N. Raymond and V. L. Pecoraro, Angew. Chem., Int. Ed., 2011, 50, 9660–9664.
34. 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.
35. J. C. Lutter, S. V. Eliseeva, J. W. Kampf, S. Petoud and V. L. Pecoraro, Chem.–Eur. J., 2018, 24, 10773–10783.
36. T. N. Nguyen, C. Y. Chow, S. V. Eliseeva, E. R. Trivedi, J. W. Kampf, I. Martinić, S. Petoud and V. L. Pecoraro, Chem.–Eur. J., 2018, 5, 1031–1035.
37. I. Martinić, S. V. Eliseeva, T. N. Nguyen, V. L. Pecoraro and S. Petoud, J. Am. Chem. Soc., 2017, 139, 8388–8391.
38. I. Martinić, S. V. Eliseeva, T. N. Nguyen, F. Foucher, D. Gosset, F. Westall, V. L. Pecoraro and S. Petoud, Chem. Sci., 2017, 8, 6042–6050.
39. G. Mezei, C. M. Zaleski and V. L. Pecoraro, Chem. Rev., 2007, 107, 4933–5003.
40. G. Sheldrick, Acta Crystallogr., 2015, C71, 3–8.
41. A. Spek, Acta Crystallogr., 2015, C71, 9–18.
42. K. K. Innes, I. G. Ross and W. R. Moomaw, J. Mol. Spectrosc., 1988, 132, 492–544.
43. J.-C. G. Bünzli, Coord. Chem. Rev., 2015, 293–294, 19–47.
44. W. S. Lo, W. T. Wong and G. L. Law, RSC Adv., 2016, 6, 74100–74109.
45. Y. Hasegawa, Y. Kitagawa and T. Nakanishi, NPG Asia Mater., 2018, 10, 52–70.
46. A. D'Aleo, F. Pointillart, L. Ouahab, C. Andraud and O. Maury, Coord. Chem. Rev., 2012, 256, 1604–1620.
47. W. T. Carnall, P. R. Fields and K. Rajnak, J. Chem. Phys., 1968, 49, 4443–4446.
48. H. Wei, Z. Zhao, C. Wei, G. Yu, Z. Liu, B. Zhang, J. Bian, Z. Bian and C. Huang, Adv. Funct. Mater., 2016, 26, 2085–2096.
49. M. Ganapathi, S. V. Eliseeva, N. R. Brooks, D. Soccol, J. Fransaer and K. Binnemans, J. Mater. Chem., 2012, 22, 5514–5522.
50. A. Beeby, I. M. Clarkson, R. S. Dickins, S. Faulkner, D. Parker, L. Royle, A. S. de Sousa, J. A. G. Williams and M. Woods, J. Chem. Soc., Perkin Trans. 2, 1999, 493–503.
51. S. Faulkner, A. Beeby, M.-C. Carrié, A. Dadabhoy, A. M. Kenwright and P. G. Sammes, Inorg. Chem. Commun., 2001, 4, 187–190.
52. M. H. V. Werts, R. T. F. Jukes and J. W. Verhoeven, Phys. Chem. Chem. Phys., 2002, 4, 1542–1548.

---

Footnotes

† Electronic supplementary information (ESI) available: ¹H- and ¹³C NMR spectra, mass spectra, crystallographic parameters, including cif files, absorption and phosphorescence spectra, supplementary tables with photophysical parameters. CCDC 19988811998882. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/d1sc06769a

‡ These authors contributed equally.

§ Current address: Helen Scientific Research and Technological Development Co., Ltd, Ho Chi Minh City, Vietnam.

¶ Current address: Department of Chemistry, Oakland University, Rochester, Michigan 48309, USA.

---