English

• Coordination–driven self–assembly, a fascinating synthetic strategy in supramolecular chemistry, is widely used to prepare a variety of discrete metallosupramolecular architectures, including helicates, macrocycles/wheels and cages/capsules [1–6]. So far, the majority of the coordination–supramolecular architectures are based on transition metal ions [7–22], such as Pd^Ⅱ, Pt^Ⅱ, Fe^Ⅱ, Au^Ⅰ, Ag^Ⅰ and Zn^Ⅱ, while architectures containing lanthanide [23–27] are less explored. This is mainly due to the difficulty in controlling the coordination number and geometry of Ln^Ⅲ ions, which have variable and labile coordination habits. However, lanthanide supramolecular compounds have attractive optical and magnetic properties derived from the 4f electrons, which make them valuable for functional exploration. Our research group has been devoted to the synthesis and study of lanthanide organic polyhedrons (LOPs) since 2015 [24]. Along with other groups, we have successfully prepared several types of Ln₂L₃ helicates, Ln₄L₄ tetrahedrons and Ln₈L₆ cubes [23,24] with particular interests toward their luminescence, cell imaging, chirality, and catalysis [23,28–33]. However, all reported LOPs are synthesized by diamagnetic closed–shell ligands. Incorporating radical–containing ligands into the multi–component assemblies can provide a new platform to examine multiple spin–spin magnetic interactions and access novel functions that are unavailable in single radical systems [34–37]. To the best of our knowledge, no examples of such radical lanthanide polyhedrons exist.

  Organic radical species are prone to disproportion or intermolecular dimerization. It had been demonstrated that covalent approach was an effectively method to improve the stability of organic radicals by using bulky substituents for steric protection or introducing conjugated groups [38,39], polar substituents [40] or heteroatoms [41] to delocalize the spin density. To prevent the dimerization, non–covalent strategy, also known as supramolecular approach, mainly relies on host–guest chemistry [42–44] and non–covalent interactions, such as hydrogen bonds [45], electrostatic interaction [46], π–π interaction [47]. Apart from these, coordination bonds were found to enhance the organic radicals' stability. For example, Wang and co–works demonstrated that metal–coordination can synergistically stabilize the radical cations of the main group elements [48]. Moreover, recent reports have also revealed that the coordination of radical ligands to transition metal ions could generate impressive radical stability in metallacycles [49–51]. However, combining radical ligand with metal organic polyhedrons aimed at improving the radical stability was rarely investigated [52–54], in particular by coordination with lanthanide ions.

  Triphenylamine (TPA) radical cations and its derivatives, obtained by one–electron oxidation of triphenylamine, have been extensively applied in a wide range of areas, including organic and hybrid solar cells, organic redox catalysis, magnetic materials, etc. [54–66]. Herein, we report the first Ln₄(L^•+)₄–type (Ln = La^Ⅲ, Eu^Ⅲ and Tb^Ⅲ) radical–bridged lanthanide organic tetrahedrons synthesized from self–assembly of face–capping TPA–cored radical ligand with Ln ions (Scheme 1). TPA–cored cationic radical ligand L₁^•+ features three peripheral electron–donating pyridine diamide chelating sites at para position. For an facilitate–intuitive comparison, L₂ with a triazolyl group replacing inner amide group on L₁, and its Ln₄(L₂)₄ assemblies were also synthesized. However, it was observed that L₂ failed to oxidize into its radical form. All compounds have been well characterized by ¹H NMR, DOSY and ESI–TOF–MS. Moreover, the structure of Eu₄(L₁)₄, Eu₄(L₁^•+)₄ and Eu₄(L₂)₄ have been unambiguously confirmed by X–ray crystallographic analyses, wherein Eu₄(L₁^•+)₄ represents the first example of radical LOPs. The generation of radical cations, their stability and lanthanide–radical magnetic interaction were further studied by UV−vis−NIR, CV, EPR spectra, magnetic susceptibilities measurements, together with density functional theory (DFT) calculations. After coordination with lanthanide ions, the enhanced radical half–life time of L₁^•+ in Ln₄(L₁^•+)₄ assemblies was observed. Remarkably, Eu₄(L₁^•+)₄, Tb₄(L₁^•+)₄, and Tb₄(L₁^•+)₄ exhibit much longer radical half–life time than that of La₄(L₁^•+)₄ and Lu₄(L₁^•+)₄, likely due to the magnetic interactions between paramagnetic Ln^Ⅲ ions and L₁^•+.

  Scheme 1

  

  Scheme 1.  Synthetic route of TPA–bridging ligands L₁, L₂, L₁^•+ and self–assembly of Ln₄(L)₄ tetrahedral cages. (ⅰ) 6–(Isopropylcarbamoyl)picolinic acid, HATU, Et₃N, 70% yield, (ⅱ) NaNO₂/HCl, NaN₃, (ⅲ) 6–ethynyl–N–isopropylpicolinamide, SA, CuSO₄·5H₂O, 75% yield, (ⅳ) NOBF₄, 95% yield, (ⅴ) Ln(OTf)₃, 99% yield.

  DownLoad: Full-Size Img PowerPoint

  Ligand L₁ was readily synthesized by a one–step amide condensation reaction between 6–(isopropylcarbamoyl)picolinic acid and a C₃–symmetric tri(4–amino)triphenylamine (3NH₂–TPA) core. As a control, ligand L₂ was prepared by the Click reaction (Scheme 1, Scheme S1 and S2, Figs. S1–S4, S9 and S10 in Supporting information).

  When ligand was treated with 1 equiv. of Ln(OTf)₃ in acetonitrile solution at room temperature for serval minutes, the turbid suspension of the ligand quickly turned into a clear solution, indicating the generation of new species. Taking L₁ as an example, NMR and ESI–TOF–MS spectra show the formation of a high symmetric Eu₄(L₁)₄ species (Fig. 1, Figs. S5 and S6 in Supporting information). Compared to free L₁, characteristic upfield shifts for H_c-g on the chelating moieties of L₁ are clearly observed (Fig. 1b), while H_a and H_b on TPA core are shifting downfield, consistent with the coordination of chelating moieties with Eu^Ⅲ ions. Moreover, the signal of H_i on the terminal isopropyl splits into two doublets with equal integration, implying a rigid coordination configuration. The ¹H diffusion–ordered NMR spectroscopy (DOSY, Fig. 1c) displays the presence of sole product with a hydrodynamic diameter of 3.2 nm. As shown in Fig. 1d, high resolution ESI–TOF–MS analyses confirm the formation of the assembly with a chemical formula of Eu₄(L₁)₄(OTf)₁₂. A series of prominent peaks with charging states from 3+ to 6+ derived by the successive loss of OTf⁻ counterions and active hydrogen on the amide of Eu₄(L₁)₄(OTf)₁₂ are observed. For instance, peak with m/z = 898.8013 corresponds to the [Eu₄(L₁)₄(OTf)₁₂–H₄–(OTf)₉]⁵⁺ charged species (Fig. 1d, inset). After replacing Eu^Ⅲ ion with La^Ⅲ or Tb^Ⅲ ions, ¹H NMR spectrum and/or the ESI–TOF–MS also confirmed the formation of La₄(L₁)₄ and Tb₄(L₁)₄ (Figs. S7, S14 and S17 in Supporting information). The analogous Eu₄(L₂)₄ complex was also synthesized in a similar way for comparison purpose (Fig. S11 in Supporting information).

  Figure 1

  

  Figure 1.  ¹H NMR of (a) ligand L₁ in DMSO–d₆ and (b) tetrahedral cage Eu₄(L₁)₄(OTf)₁₂ in CD₃CN. (c) ¹H DOSY spectrum of Eu₄(L₁)₄(OTf)₁₂. (d) ESI–TOF–MS of Eu₄(L₁)₄(OTf)₁₂ with insets showing the observed (Obs.) and simulated (Sim.) isotopic patterns for the 5+ peaks. Single crystal X–ray structures of (e) Eu₄(L₁)₄ and (f) Eu₄(L₂)₄. Eu: Cyan, C: green, N: blue, O: red, H: white.

  DownLoad: Full-Size Img PowerPoint

  Single crystals of the Eu₄(L₁)₄ and Eu₄(L₂)₄ assemblies were grown by slow vapor diffusion of diethyl ether into their acetonitrile solution, confirming their tetrahedral geometry (Figs. 1e and f), which are isomorphic to our previous reported Eu₄L₄–type tetrahedral cages [67,68]. The four metal centers on the tetrahedron display cooperative ΔΔΔΔ or ΛΛΛΛ chirality, and both enantiomers coexist in crystals (Figs. S20–S22 in Supporting information), as they all crystallize in the achiral F_m-3 space group. In the crystal structures of Eu₄(L₁)₄ and Eu₄(L₂)₄, each Eu^Ⅲ centers are nine–coordinated and ligated by three chelating moieties from three different ligands.

  To investigate the redox properties of the free ligands and their assemblies, we initially oxidized the neutral ligand with nitrosonium tetrafluoroborate (NOBF₄), a common one–electron oxidant. Upon addition of NOBF₄, the solution color of L₁ in acetonitrile rapidly changed from yellow to green, which is a characteristic color of TPA–cored radical cation, indicating the generation of L₁^•+. The formation of radical cation was also unambiguously confirmed by the UV–vis and EPR spectra (vide infra). It should be noted that the ¹H NMR spectra of L₁^•+ were significantly broadened, due to the paramagnetic nature of the radical species on the TPA units (Fig. S4). Unfortunately, we failed to produce the radical cation L₂^•+ using several reported efficient oxidation reagents, including NOBF₄, Cu(ClO₄)₂ and AgBF₄, suggesting its high oxidation potential.

  To elucidate the distinction between L₁ and L₂, cyclic voltammetry (CV) was measured for both ligands in DMSO containing tetrabutylammonium hexafluorophosphate (ⁿBu₄NPF₆) at a scan rate of 100 mV/s (vs. Ag/AgCl) under an argon atmosphere. Free ligand L₁ exhibited one reversible oxidation wave around +0.82 V and an irreversible oxidation peak around +1.22 V, corresponding to the formation of L₁^•+ and L₁²⁺, respectively (Fig. 2a). Similar CV waves were observed for L₁^•+ (Fig. S48 in Supporting information), which further confirmed the assignment of the oxidation peaks and the stability of the radical cation form. However, only one reversible oxidation wave was observed around +1.23 V for ligand L₂ (Fig. 2b), which was even higher than the second oxidation potential of L₁. We also calculated the frontier orbitals of ligands L₁ and L₂ using DFT methods at the UB3LYP/6–31G(d) level of theory to explain the different behavior under electrochemical condition. Fig. 2c shows that the energy of the HOMO (HOMO = highest occupied molecular orbital) of L₂ (–5.54 eV) is lower than L₁ (–4.65 eV) by 0.89 eV [69]. This lower delocalized HOMO indicates weaker electron loss ability, which is in good agreement with the higher first reversible oxidation potential (E^ox) of L₂, consistent with the CV results (E^ox (L₁) = 0.82 V vs. E^ox (L₂) = 1.23 V). Moreover, we performed the CV experiments for Eu₄(L₁)₄ and Eu₄(L₂)₄ in acetonitrile. Compared with L₁^•+, only one broadened oxidation wave was observed for the cage Eu₄(L₁)₄ with a positive shift of the oxidation potential around +1.44 V, likely due to the complexation with Eu^Ⅲ and the polycationic nature (12+) of the Ln^Ⅲ–linked cage structure [52,70]. This behavior demonstrates the considerable stability of the Eu₄(L₁^•+)₄ radical cage as well as the absence of disassembly during the redox process. In sharp contrast, only one irreversible oxidation wave was observed around 1.37 V in the Eu₄(L₂)₄ acetonitrile solution.

  Figure 2

  

  Figure 2.  (a, b) Cyclic voltammetry of ligand L₁, Eu₄(L₁)₄ and L₂, Eu₄(L₂)₄ (1 mmol/L for ligands, 0.25 mmol/L for cages, 0.1 mol/L ⁿBu₄NPF₆, scan rate 100 mV/s) at room temperature. (c) Frontier orbitals of L₁ and L₂ calculated using DFT methods (B3LYP/6–31G(d)).

  DownLoad: Full-Size Img PowerPoint

  Next, we examined the assembly of L₁^•+ with Ln ions. We obtained a clear assembly solution by mixing the radical cation L₁^•+ with Eu(OTf)₃ in acetonitrile. The UV–vis spectrum of this solution showed three absorption peaks, two of which (under 500 nm) were similar to the absorption peaks of Eu₄(L₁)₄, indicating the formation of Eu₄(L₁^•+)₄ (Fig. S34 in Supporting information). It should be noted that a new shoulder peak at about 735 nm in Eu₄(L₁^•+)₄ was observed compared to Eu₄(L₁)₄. To verify this difference, we measured the UV–vis absorption of L₁ and L₁^•+ in the same solution (Fig. 3a). Unlike L₁, a typical radical absorption peak around 784 nm was found in L₁^•+, in agreement with the reported TPA radical cation [71]. Therefore, the new peak observed in Eu₄(L₁^•+)₄ could be attributed to the absorption band of the radical ligands L₁^•+ (Fig. S34), and its slight blue shift in Eu₄(L₁^•+)₄ may be caused by the coordination with Eu^Ⅲ ions. Similarly, the successful preparation of La₄(L₁^•+)₄ and Tb₄(L₁^•+)₄ was also confirmed by the UV–vis spectra (Fig. 3a and Fig. S34). We failed to obtain ESI–TOF–MS data for the radical assemblies Ln₄(L₁^•+)₄ (Ln = La, Eu and Tb), which may be due to their easy reduction under the ionization conditions.

  Figure 3

  

  Figure 3.  (a) Absorption spectra of L₁ (0.1 mmol/L, blue dot line), L₁^•+ (0.1 mmol/L, blue line), Tb₄(L₁)₄ (0.025 mmol/L, pink dot line), Tb₄(L₁^•+)₄ (0.025 mmol/L, pink line) in CH₃CN, Inset: the color view of L₁ and L₁^•+. (b) Excitation and emission spectra of Tb₄(L₁)₄ (blue line) and Tb₄(L₁^•+)₄ (pink line) in CH₃CN.

  DownLoad: Full-Size Img PowerPoint

  Interestingly, the greenish cubic single crystals (Fig. S21) of the radical assembly were grown overnight by the same vapor diffusion method as for the Eu₄(L₁)₄ crystal growth process. Like Eu₄(L₁)₄, this radical cage also crystalizes in the cubic space group F_m-3. The main difference lies in their C–N bond length on the TPA panels. This bond length in the radical cage (1.404 Å, Fig. S28 in Supporting information) is shorter by about 0.015 Å than that in cage Eu₄(L₁)₄ (1.419 Å), similar to the trends observed in the mono–radical cation of thiophene–bridged bis(TPA) [72], indicative of the radical feature of TPA panels in radical cage. The single crystal structure provides reliable evidence for the existence of radical centers in a tetrahedral cage. However, the crystal structure cannot confirm that each tetrahedral cage contains four radicals, due to the easy recovery feature of TPA-type radical during the crystal growth process. According to the calculated half–life time from UV–vis tracking experiment (vide infra) coupled with the crystal growth time, we speculate that about 2 radicals may be remained in a cage after the crystallization process.

  The spin density distribution of L₁^•+ was estimated by DFT calculations at the UB3LYP/6–31G(d) level of theory. Spin density of L₁^•+ was found to be mainly delocalized on the TPA core and three inner amido groups, with the highest Mulliken atomic spin density (M–ASD) of ~0.3 single electron on the central N atom (Fig. S33 and Table S4 in Supporting information for the spin density map and the M–ASD values of all atoms in L₁^•+). The delocalization effects resulted in partial double bond character for the central three C–N single bonds, which accounted for the shortened C–N bond length observed in the radical Eu₄(L₁^•+)₄ cage. Most of the organometallic assemblies reported in the literature were prepared by the reaction of a neutral or an anionic ligand with a cationic metal precursor, while those fabricated by a cationic ligand and metal ions were rare [52,73]. In some cases, transferring the neutral ligand into cationic state even leads to the disassembly of the assemblies [74]. The stability of our assembly may be attributed to the low positive charge density of the radical cation on the tridentate chelating claws, which reduced the cation–cation repulsion between the radical ligand and the Ln^Ⅲ ion.

  Luminescence measurements indicated that both L₁ and L₂ can sensitize Tb^Ⅲ (Fig. 3b and Fig. S46 in Supporting information). Taking Tb₄(L₁)₄ as an example, upon excitation at 315 nm, a cascade of characteristic emission bands at 618, 583, 543, 488 nm assigning to ⁵D₄ → ⁷F_J (J = 3 − 6) transitions is observed. The longest wavelength absorption maximum of Tb₄(L₁^•+)₄ appears at 740 nm, with a tail that extends beyond 800 nm. It is predictable that L₁^•+ cannot sensitize Tb₄(L₁^•+)₄ under the same test condition due to low energy levels of the radical species. As shown in Fig. 3b, a broadened band emission centered at 506 nm takes the place of the typical line−like Tb^Ⅲ green luminescence, which probably originated from those recovered neutral ligands in Tb's cage.

  The radical stability of L₁^•+, La₄(L₁^•+)₄ and Tb₄(L₁^•+)₄ was evaluated by monitoring the time–dependent UV–vis absorption spectra under ambient circumstance (Figs. 4a-c). The half–life times (t_1/2) of L₁^•+, La₄(L₁^•+)₄ and Tb₄(L₁^•+)₄ were determined to be 53 min, 482 min and 822 min, respectively. The assembly with Ln^Ⅲ ions significantly improved the stability of L₁^•+, with approximately 9−fold and 15−fold enhancements for La^Ⅲ and Tb^Ⅲ assemblies, respectively (Fig. 4d). The stability of the radical ligands L₁^•+ could be attributed to the covalent effect, as the incorporation of the ONO chelating moiety to the TPA core enabled the spin density to delocalize to the oxygen atom of inner amide (Fig. S33). Moreover, the coordination of the radical ligands with Ln^Ⅲ ions anchored the radical ligands on the four faces of the tetrahedron, resulting in reduced close stacking of the TPA core and thus enhancing the radical L₁^•+ stability in the cage. Furthermore, OTf^– and solvent molecules occupied the cavity of the cage and the space between the cages, which prevented the interactions with other reactive molecules, especially O₂, and slowed down decay reactions.

  Figure 4

  

  Figure 4.  Time–dependent UV–vis absorption spectra of (a) L₁^•+, (b) La₄(L₁^•+)₄ and (c) Tb₄(L₁^•+)₄ in CH₃CN recorded at ambient air and light conditions. (d) Plot of the absorbance at 784 nm (L₁^•+), 760 nm (La₄(L₁^•+)₄), 740 nm (Tb₄(L₁^•+)₄) with time, showing the half–life times were increased by 9–fold and 15–fold for La₄(L₁^•+)₄, Tb₄(L₁^•+)₄, respectively.

  DownLoad: Full-Size Img PowerPoint

  In addition, we found a 1.8-fold increase in t_1/2 from La₄(L₁^•+)₄ to Tb₄(L₁^•+)₄. To understand this difference, we first consider the influence of lanthanide paramagnetism, thereby EPR spectra under ambient conditions were tested in acetonitrile solution. As shown in Fig. 5a, the EPR spectrum of free ligand L₁^•+ showed a signal at g = 2.003 with a hyperfine structure originating from the nitrogen nuclei (I = 1, triplet), indicating that the higher spin density is mainly localized on the central nitrogen atom. Upon complexation with Ln^Ⅲ (Ln^Ⅲ = La^Ⅲ and Tb^Ⅲ) ions, distinct EPR signals were observed, which confirmed the formation of radical L₁^•+, since the diamagnetic behavior of La^Ⅲ ions and the fast spin–lattice relaxation feature of Tb^Ⅲ ions make both them EPR silent at room temperature [75]. This was further verified by the absence EPR signals for their non–radical counterparts La₄(L₁)₄ and Tb₄(L₁)₄. These results demonstrated that the radical ligands and assemblies were successfully prepared. Moreover, La₄(L₁^•+)₄ exhibited the same EPR hyperfine signals as the radical ligand L₁^•+, implying that there was no magnetic interaction between the radical and La^Ⅲ ions or among the four radical ligands themselves. This is reasonable because of the intrinsic diamagnetism of La^Ⅲ ions and the large separation of the central N atoms with higher spin density in the La₄(L₁^•+)₄ tetrahedron (the N–N distance in the Eu^Ⅲ analogue was 6.669 Å). In sharp contrast, six EPR peaks centered at g = 2.007 were observed for Tb₄(L₁^•+)₄, which differed from the radical ligand L₁^•+, suggesting that there was magnetic interaction between Tb^Ⅲ and the organic radical.

  Figure 5

  

  Figure 5.  (a) EPR spectra of ligand L₁ (light yellow line), L₁^•+ (yellow line, g = 2.003), La₄(L₁)₄ (light blue line), La₄(L₁^•+)₄ (blue line, g = 2.003), Tb₄(L₁)₄ (light pink line) and Tb₄(L₁^•+)₄ (pink line, g = 2.007) in acetonitrile solution at 293 K. (b) Temperature dependence of χ_mT for L₁^•+ (hexagon), Tb₄(L₁)₄ (triangle) and Tb₄(L₁^•+)₄ (square) measured at 5000 Oe.

  DownLoad: Full-Size Img PowerPoint

  To gain a deeper insight into the radical characteristic of L₁^•+ and the magnetic interaction between Tb^Ⅲ and the radical spins in Tb₄(L₁^•+)₄, direct current (DC) magnetic susceptibilities for the freshly prepared samples of L₁^•+, Tb₄(L₁)₄ and Tb₄(L₁^•+)₄ were recorded in the temperature range of 2 − 300 K. As shown in Fig. 5b, the χ_mT value of L₁^•+ at 300 K is 0.37 cm³ K/mol, very close to the expected value for a magnetically isolated system of one S = 1/2 with g = 2.0 (0.375 cm³ K/mol), confirming the open shell character of L₁^•+. The χ_mT values decrease continuously with lowing temperature, indicative of the presence of antiferromagnetic (AF) interaction between adjacent L₁^•+. For Tb₄(L₁^•+)₄, the experimental χ_mT value at room temperature is 47.13 cm³ K/mol, quite close to the theoretical value of 48.75 cm³ K/mol for four non−interacting Tb^Ⅲ ions (J = 6, g = 3/2) and four magnetically isolated radical spin units. Upon cooling, the χ_mT values undergo a smooth decrease until about 50 K, below which it decreases rapidly and reaches a minimum value of 16.30 cm³ K/mol at 2 K. The decrease in low temperature for Tb₄(L₁^•+)₄ is most likely caused by the thermal depopulation of Stark sublevels, but the contribution of AF exchange interactions cannot be ignored. In contrast, the χ_mT product of Tb₄(L₁)₄ maintains roughly constant with lowing temperature down to about 100 K, and then drops to a value of 21.26 cm³ K/mol at 2 K. This different decrease behavior in χ_mT vs. T curves unambiguously demonstrates the presence of AF coupling between Tb^Ⅲ ion and the radical spin carrier. To further prove the AF interactions, the χ_mT of Tb₄(L₁)₄ is subtracted from the χ_mT of Tb₄(L₁^•+)₄ to offset the contribution stemming from thermal depopulation of the Stark sublevels of Tb^Ⅲ ions, and the temperature dependence of Δχ_mT (Δχ_mT = χ_mT_{(Tb4(L1•+)4)} − χ_mT_(Tb4(L1)4)) is obtained (Fig. S51 in Supporting information). The Δχ_mT falls progressively with lowering temperature, confirming again the AF correlations between Tb^Ⅲ ion and radicals. The inherent unquenched orbital angular momentum of Tb^Ⅲ ion together with the delocalization behavior of radical electron make the quantitative estimation of the AF interaction magnitude tricky, which may deserve separate investigation in the future.

  Apart from the influence of lanthanide paramagnetism, the Ln radius may also exert effect on the stability of radical L₁^•+, as the Ln size can affect the rigidity of the ligands and the planarity of the central N atoms. To study this, the t_1/2 values of Eu₄(L₁^•+)₄, Gd₄(L₁^•+)₄ and Lu₄(L₁^•+)₄ were estimated, giving the t_1/2 values of 624 min, 1248 min and 347 min, respectively. By comparing the t_1/2 values with Ln radii (Tables S5 in Supporting information), it was found that the stability of the radical is not dependent on the Ln radius. However, we observed that the assemblies containing diamagnetic Ln ions (Ln = La^Ⅲ and Lu^Ⅲ) show lower t_1/2 values than those assemblies constructed by paramagnetic Ln assemblies (Ln = Eu^Ⅲ, Gd^Ⅲ, Tb^Ⅲ). Based on this observation, we propose that the existence of magnetic interaction between the radical ligands and paramagnetic Ln ions in Ln₄(L₁^•+)₄ (Ln = Eu^Ⅲ, Gd^Ⅲ and Dy^Ⅲ) increases the degree of radical electronic delocalization compared to that in Ln₄(L₁^•+)₄ (Ln = La^Ⅲ and Lu^Ⅲ) fabricated by diamagnetic Ln ions, and hence resulting in enhanced radical stability in Ln₄(L₁^•+)₄ (Ln = Eu^Ⅲ, Gd^Ⅲ and Dy^Ⅲ).

  In summary, we report the first example of a radical−bridged lanthanide organic tetrahedron constructed via coordination−driven self−assembly. The progressive stability enhancement of the radical panels was observed from the tetrahedron fabricated by diamagnetic Ln^Ⅲ ions to that by paramagnetic Ln^Ⅲ ions, as certificated by the UV−vis, EPR and magnetic susceptibility measurements. We infer that the enhancement in paramagnetic Ln^Ⅲ ions should be ascribed to the Ln^Ⅲ−radical magnetic interaction, which enables the radical electron more delocalization compared to the Ln^Ⅲ tetrahedron containing diamagnetic Ln^Ⅲ ions. Our findings not only offer some guidance for the preparation of stable radical systems, but also provide new candidates for smart lanthanide materials.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper

  Acknowledgments

  This work was supported by National Key Research and Development Program of China (Nos. 2021YFA1500400 and 2022YFA1503300), the National Natural Science Foundation of China (Nos. 21825107, 21971237, 22171264 and 22301301) and the Science Foundation of the Fujian Province (No. 2021J02016).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.109641.

  
  
• 1. [1]

     Y. Sun, C. Chen, J. Liu, et al., Chem. Soc. Rev. 49 (2020) 3889–3919. doi: 10.1039/D0CS00038H

  2. [2]

     E.G. Percastegui, T.K. Ronson, J.R. Nitschke, Chem. Rev. 120 (2020) 13480–13544. doi: 10.1021/acs.chemrev.0c00672

  3. [3]

     C.M. Hong, R.G. Bergman, K.N. Raymond, et al., Acc. Chem. Res. 51 (2018) 2447–2455. doi: 10.1021/acs.accounts.8b00328

  4. [4]

     M.D. Ward, C.A. Hunter, N.H. Williams, Acc. Chem. Res. 51 (2018) 2073–2082. doi: 10.1021/acs.accounts.8b00261

  5. [5]

     D.L. Caulder, K.N. Raymond, Acc. Chem. Res. 32 (1999) 975–982. doi: 10.1021/ar970224v

  6. [6]

     M. Fujita, M. Tominaga, A. Hori, et al., Acc. Chem. Res. 38 (2005) 369–378. doi: 10.1021/ar040153h

  7. [7]

     S. Bai, Y.F. Han, Acc. Chem. Res. 56 (2023) 1213–1227. doi: 10.1021/acs.accounts.3c00102

  8. [8]

     C.T. McTernan, J.A. Davies, J.R. Nitschke, Chem. Rev. 122 (2022) 10393–10437. doi: 10.1021/acs.chemrev.1c00763

  9. [9]

     C.B. Tian, Q.F. Sun, Chem. Eur. J. 29 (2023) e202300195. doi: 10.1002/chem.202300195

  10. [10]

      Y. Domoto, M. Fujita, Coord. Chem. Rev. 466 (2022) 214605–214617. doi: 10.1016/j.ccr.2022.214605

  11. [11]

      F. Wang, S. Bai, Q.W. Zhu, et al., CCS Chem. 5 (2023) 633–640. doi: 10.31635/ccschem.022.202201919

  12. [12]

      K. Acharyya, S. Bhattacharyya, H. Sepehrpour, et al., J. Am. Chem. Soc. 141 (2019) 14565–14569. doi: 10.1021/jacs.9b08403

  13. [13]

      P. Howlader, E. Zangrando, P.S. Mukherjee, J. Am. Chem. Soc. 142 (2020) 9070–9078. doi: 10.1021/jacs.0c03551

  14. [14]

      M. Pan, K. Wu, J.H. Zhang, et al., Coord. Chem. Rev. 378 (2019) 333–349. doi: 10.1016/j.ccr.2017.10.031

  15. [15]

      K. Wu, K. Li, S. Chen, et al., Angew. Chem. Int. Ed. 59 (2019) 2639–2643.

  16. [16]

      J. Zhu, X. Chen, X. Jin, et al., Chin. Chem. Lett. 34 (2023) 108002–108005. doi: 10.1016/j.cclet.2022.108002

  17. [17]

      Y.L. Lai, H.J. Zhang, J. Su, et al., Chin. Chem. Lett. 34 (2023) 107686–107690. doi: 10.1016/j.cclet.2022.07.029

  18. [18]

      C. Li, Y. Wang, Y. Lu, et al., Chin. Chem. Lett. 31 (2020) 1183–1187. doi: 10.1016/j.cclet.2019.09.034

  19. [19]

      D. Luo, B. Pan, J. Zhang, et al., Chin. Chem. Lett. 32 (2021) 1397–1399. doi: 10.1016/j.cclet.2020.11.002

  20. [20]

      R.Y. Chen, Y.P. He, J. Zhang, Polyoxometalates 1 (2022) 9140002–9140007. doi: 10.26599/POM.2022.9140002

  21. [21]

      R.Y. Chen, G.H. Chen, Y.P. He, et al., Chin. J. Struct. Chem. 41 (2022) 2201001–2201006.

  22. [22]

      X. Liu, X. Feng, K.R. Meihaus, et al., Angew. Chem. Int. Ed. 59 (2020) 10610–10618. doi: 10.1002/anie.202002673

  23. [23]

      D.J. Bell, L.S. Natrajan, I.A. Riddell, Coord. Chem. Rev. 472 (2022) 214786–214810. doi: 10.1016/j.ccr.2022.214786

  24. [24]

      X.Z. Li, C.B. Tian, Q.F. Sun, Chem. Rev. 122 (2022) 6374–6458. doi: 10.1021/acs.chemrev.1c00602

  25. [25]

      M.H. Du, D.H. Wang, L.W. Wu, et al., Angew. Chem. Int. Ed. 61 (2022) e202200537. doi: 10.1002/anie.202200537

  26. [26]

      X.L. Li, J. Wu, L. Zhao, et al., Chem. Commun. 53 (2017) 3026–3029. doi: 10.1039/C7CC00048K

  27. [27]

      X.L. Li, J. Wu, J. Tang, et al., Chem. Commun. 52 (2016) 9570–9573. doi: 10.1039/C6CC05326B

  28. [28]

      J. Wang, C. He, P. Wu, et al., J. Am. Chem. Soc. 133 (2011) 12402–12405. doi: 10.1021/ja2048489

  29. [29]

      D.E. Barry, D.F. Caffrey, T. Gunnlaugsson, Chem. Soc. Rev. 45 (2016) 3244–3274. doi: 10.1039/C6CS00116E

  30. [30]

      J. Hamacek, A. Vuillamy, Eur. J. Inorg. Chem. 2018 (2017) 1155–1166.

  31. [31]

      H.Y. Wong, W.S. Lo, K.H. Yim, et al., Chem 5 (2019) 3058–3095. doi: 10.1016/j.chempr.2019.08.006

  32. [32]

      S.J. Hu, X.Q. Guo, L.P. Zhou, et al., J. Am. Chem. Soc. 144 (2022) 4244–4253. doi: 10.1021/jacs.2c00760

  33. [33]

      S.J. Hu, X.Q. Guo, L.P. Zhou, et al., Chin. J. Chem. 41 (2023) 797–804. doi: 10.1002/cjoc.202200649

  34. [34]

      B.S. Dolinar, D.I. Alexandropoulos, K.R. Vignesh, et al., J. Am. Chem. Soc. 140 (2018) 908–911. doi: 10.1021/jacs.7b12495

  35. [35]

      D.I. Alexandropoulos, B.S. Dolinar, K.R. Vignesh, et al., J. Am. Chem. Soc. 139 (2017) 11040–11043. doi: 10.1021/jacs.7b06925

  36. [36]

      X. Liu, Y. Zhang, W. Shi, et al., Inorg. Chem. 57 (2018) 13409–13414. doi: 10.1021/acs.inorgchem.8b01981

  37. [37]

      X. Meng, W. Shi, P. Cheng, Coord. Chem. Rev. 378 (2019) 134–150. doi: 10.1016/j.ccr.2018.02.002

  38. [38]

      E.T. Seo, R.F. Nelson, J.M. Fritsch, et al., J. Am. Chem. Soc. 88 (1966) 3498–3503. doi: 10.1021/ja00967a006

  39. [39]

      D.H. Reid, Tetrahedron 3 (1958) 339–352. doi: 10.1016/0040-4020(58)80039-3

  40. [40]

      S. Kumar, M.R. Ajayakumar, G. Hundal, et al., J. Am. Chem. Soc. 136 (2014) 12004–12010. doi: 10.1021/ja504903j

  41. [41]

      P.P. Power, Chem. Rev. 103 (2003) 789–810. doi: 10.1021/cr020406p

  42. [42]

      H. Garcia, H.D. Roth, Chem. Rev. 102 (2002) 3947–4007. doi: 10.1021/cr980026x

  43. [43]

      V. Pushkara Rao, M.B. Zimmt, N.J. Turro, J. Photochem. Photobio. A: Chem. 60 (1991) 355–360. doi: 10.1016/1010-6030(91)90037-T

  44. [44]

      H. Hu, Y.Y. Zhang, H. Ma, et al., Angew. Chem. Int. Ed. 62 (2023) e202308513. doi: 10.1002/anie.202308513

  45. [45]

      E. Breinlinger, A. Niemz, V.M. Rotello, J. Am. Chem. Soc. 117 (1995) 5379–5380. doi: 10.1021/ja00124a029

  46. [46]

      Q. Song, F. Li, Z. Wang, et al., Chem. Sci. 6 (2015) 3342–3346. doi: 10.1039/C5SC00862J

  47. [47]

      E. Moulin, F. Niess, M. Maaloum, et al., Angew. Chem. Int. Ed. 49 (2010) 6974–6978. doi: 10.1002/anie.201001833

  48. [48]

      Z. Feng, S. Tang, Y. Su, et al., Chem. Soc. Rev. 51 (2022) 5930–5973. doi: 10.1039/D2CS00288D

  49. [49]

      G. Huo, X. Shi, Q. Tu, et al., J. Am. Chem. Soc. 141 (2019) 16014–16023. doi: 10.1021/jacs.9b08149

  50. [50]

      Q. Tu, G.F. Huo, X.L. Zhao, et al., Mat. Chem. Front. 5 (2021) 1863–1871. doi: 10.1039/D0QM00992J

  51. [51]

      S.K. Zhang, L.Z. Ma, W.Q. Ma, et al., Angew. Chem. Int. Ed. 61 (2022) e202209054. doi: 10.1002/anie.202209054

  52. [52]

      K. Yazaki, S. Noda, Y. Tanaka, et al., Angew. Chem. Int. Ed. 55 (2016) 15031–15034. doi: 10.1002/anie.201608350

  53. [53]

      Z. Lu, R. Lavendomme, O. Burghaus, et al., Angew. Chem. Int. Ed. 58 (2019) 9073–9077. doi: 10.1002/anie.201903286

  54. [54]

      G.F. Jin, Y.Z. Zhang, L. Yu, et al., Nano Res. 16 (2023) 10678–10683. doi: 10.1007/s12274-023-5690-2

  55. [55]

      J. Wang, K. Liu, L. Ma, et al., Chem. Rev. 116 (2016) 14675–14725. doi: 10.1021/acs.chemrev.6b00432

  56. [56]

      G. Tan, X. Wang, Acc. Chem. Res. 50 (2017) 1997–2006. doi: 10.1021/acs.accounts.7b00229

  57. [57]

      E. Moulin, J.J.T. Armao, N. Giuseppone, Acc. Chem. Res. 52 (2019) 975–983. doi: 10.1021/acs.accounts.8b00536

  58. [58]

      L. Mao, M. Zhou, X. Shi, et al., Chin. Chem. Lett. 32 (2021) 3331–3341. doi: 10.1016/j.cclet.2021.05.004

  59. [59]

      B. Huang, L. Mao, X. Shi, et al., Chem. Sci. 12 (2021) 13648–13663. doi: 10.1039/D1SC01618K

  60. [60]

      W.L. Jiang, Z. Peng, B. Huang, et al., J. Am. Chem. Soc. 143 (2021) 433–441. doi: 10.1021/jacs.0c11738

  61. [61]

      Z.X. Chen, Y. Li, F. Huang, Chem 7 (2021) 288–332. doi: 10.1016/j.chempr.2020.09.024

  62. [62]

      Ł. Skórka, J.M. Mouesca, J.B. Gosk, et al., J. Mater. Chem. C 5 (2017) 6563–6569. doi: 10.1039/C7TC01932G

  63. [63]

      M. Uebe, T. Kazama, R. Kurata, et al., Angew. Chem. Int. Ed. 56 (2017) 15712–15717. doi: 10.1002/anie.201709874

  64. [64]

      Y. Yokoyama, D. Sakamaki, A. Ito, et al., Angew. Chem. Int. Ed. 51 (2012) 9403–9406. doi: 10.1002/anie.201204106

  65. [65]

      X. Li, Y.L. Wang, C. Chen, et al., Chemistry 29 (2023) e202203242. doi: 10.1002/chem.202203242

  66. [66]

      X. Li, Y.L. Wang, C. Chen, et al., Nat. Commun. 13 (2022) 5367–5374. doi: 10.1038/s41467-022-33130-1

  67. [67]

      X.Q. Guo, L.P. Zhou, L.X. Cai, et al., Chem. Eur. J. 24 (2018) 6936–6940. doi: 10.1002/chem.201801132

  68. [68]

      L.L. Yan, C.H. Tan, L.P. Zhou, et al., J. Am. Chem. Soc. 137 (2015) 8550–8555. doi: 10.1021/jacs.5b03972

  69. [69]

      T. Lu, F. Chen, J. Comput. Chem. 33 (2012) 580–592. doi: 10.1002/jcc.22885

  70. [70]

      W.Y. Sun, T. Kusukawa, M. Fujita, J. Am. Chem. Soc. 124 (2002) 11570–11571. doi: 10.1021/ja0206285

  71. [71]

      K. Sreenath, C.V. Suneesh, V.K. Kumar, et al., J. Org. Chem. 73 (2008) 3245–3251. doi: 10.1021/jo800349n

  72. [72]

      S. Zheng, S. Barlow, C. Risko, et al., J. Am. Chem. Soc. 128 (2006) 1812–1817. doi: 10.1021/ja0541534

  73. [73]

      Y. Satoh, L. Catti, M. Akita, et al., J. Am. Chem. Soc. 141 (2019) 12268–12273. doi: 10.1021/jacs.9b03419

  74. [74]

      V. Croue, S. Goeb, G. Szaloki, et al., Angew. Chem. Int. Ed. 55 (2016) 1746–1750. doi: 10.1002/anie.201509265

  75. [75]

      J.E. McPeak, S.S. Eaton, G.R. Eaton, Methods Enzymol. 651 (2021) 63–101.

• 

  Scheme 1  Synthetic route of TPA–bridging ligands L₁, L₂, L₁^•+ and self–assembly of Ln₄(L)₄ tetrahedral cages. (ⅰ) 6–(Isopropylcarbamoyl)picolinic acid, HATU, Et₃N, 70% yield, (ⅱ) NaNO₂/HCl, NaN₃, (ⅲ) 6–ethynyl–N–isopropylpicolinamide, SA, CuSO₄·5H₂O, 75% yield, (ⅳ) NOBF₄, 95% yield, (ⅴ) Ln(OTf)₃, 99% yield.

  下载: 全尺寸图片 幻灯片
  

  Figure 1  ¹H NMR of (a) ligand L₁ in DMSO–d₆ and (b) tetrahedral cage Eu₄(L₁)₄(OTf)₁₂ in CD₃CN. (c) ¹H DOSY spectrum of Eu₄(L₁)₄(OTf)₁₂. (d) ESI–TOF–MS of Eu₄(L₁)₄(OTf)₁₂ with insets showing the observed (Obs.) and simulated (Sim.) isotopic patterns for the 5+ peaks. Single crystal X–ray structures of (e) Eu₄(L₁)₄ and (f) Eu₄(L₂)₄. Eu: Cyan, C: green, N: blue, O: red, H: white.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a, b) Cyclic voltammetry of ligand L₁, Eu₄(L₁)₄ and L₂, Eu₄(L₂)₄ (1 mmol/L for ligands, 0.25 mmol/L for cages, 0.1 mol/L ⁿBu₄NPF₆, scan rate 100 mV/s) at room temperature. (c) Frontier orbitals of L₁ and L₂ calculated using DFT methods (B3LYP/6–31G(d)).

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) Absorption spectra of L₁ (0.1 mmol/L, blue dot line), L₁^•+ (0.1 mmol/L, blue line), Tb₄(L₁)₄ (0.025 mmol/L, pink dot line), Tb₄(L₁^•+)₄ (0.025 mmol/L, pink line) in CH₃CN, Inset: the color view of L₁ and L₁^•+. (b) Excitation and emission spectra of Tb₄(L₁)₄ (blue line) and Tb₄(L₁^•+)₄ (pink line) in CH₃CN.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Time–dependent UV–vis absorption spectra of (a) L₁^•+, (b) La₄(L₁^•+)₄ and (c) Tb₄(L₁^•+)₄ in CH₃CN recorded at ambient air and light conditions. (d) Plot of the absorbance at 784 nm (L₁^•+), 760 nm (La₄(L₁^•+)₄), 740 nm (Tb₄(L₁^•+)₄) with time, showing the half–life times were increased by 9–fold and 15–fold for La₄(L₁^•+)₄, Tb₄(L₁^•+)₄, respectively.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  (a) EPR spectra of ligand L₁ (light yellow line), L₁^•+ (yellow line, g = 2.003), La₄(L₁)₄ (light blue line), La₄(L₁^•+)₄ (blue line, g = 2.003), Tb₄(L₁)₄ (light pink line) and Tb₄(L₁^•+)₄ (pink line, g = 2.007) in acetonitrile solution at 293 K. (b) Temperature dependence of χ_mT for L₁^•+ (hexagon), Tb₄(L₁)₄ (triangle) and Tb₄(L₁^•+)₄ (square) measured at 5000 Oe.

  下载: 全尺寸图片 幻灯片
•