Heteroleptic [CrIIIN6] Chromophores as Partners for Lanthanide-Based Light Conversion in d-f Molecular Complexes

Abstract The connection of a dianionic 2,2’-biimidazolate (biim 2−) bridging unit to cis-[Cr(N ∩N) 2] 3+ (N ∩N is a chelating didentate ligand) or cis-[Cr(N ∩N ∩N ∩N)] 3+ building blocks (N ∩N ∩N ∩N is a chelating tetradentate ligand) produces heteroleptic pseudo-octahedral [CrN 6] + chromophores. Their reduced cationic charge is compatible with the subsequent complexation of trivalent lanthanides (Ln 3+) to give d-f {[(N ∩N) 2Cr(biim)] nLn} (3+n)+ ( n = 1–4), {[(N ∩N) 2Cr(biim)]Ln(Tp) 2} 2+ and {[(N ∩N ∩N ∩N)Cr(biim)]Ln(Tp) 2} 2+ adducts (Tp is tri(1 H-pyrazol-1-yl)-λ 4-borate). Moving from polyaromatic N ∩N (1,10 phenanthroline) to saturated N ∩N ∩N ∩N polyamine (cyclam) receptors controls the photophysical properties and leads to tunable light conversion in the target heterometallic complexes when Eu(III) is exploited as the activator for downshifting and Er(III) as the activator for upconversion. 1. Introduction The well-known Tanabe–Sugano diagram established for pseudo-octahedral [MX 6] scaffolds possessing an open-shell d 3 electronic configuration [ 1, 2, 3] applies to [Cr IIIX 6] chromophores [ 4, 5, 6] ( Figure 1), and justifies the challenging efforts undertaken in the last two decades for combining them with open-shell 4f-block trivalent cations. Firstly, the ground-state atomic term Cr( 4A 2) is magnetically active and behaves as a ‘pure’ multiple spin center ( S = 3/2) with (i) negligible magnetic anisotropy, (ii) modulable zero-field splitting, and (iii) long electronic relaxation time [ 7]. When Cr(III) centers are integrated into polynuclear assemblies exhibiting intermetallic ferromagnetic exchange coupling, the magnetic moment of Cr(III) can be exploited for the design of slow-relaxing single molecular magnets with accessible blocking temperatures [ 8, 9, 10]. Some specific arrangements of ferromagnetically coupled d-f chromium–lanthanide wheels and butterflies gave access to rare magneto-structural maps [ 11, 12, 13, 14, 15, 16]. Alternatively, antiferromagnetic Cr(III)-Gd(III) coupling is highly desired for the preparation of isotropic ultra-dense assemblies inducing large magneto-caloric cooling effects [ 17, 18, 19]. The concomitant long electronic relaxation time makes the basis for (i) electron paramagnetic resonance imaging (EPRI) [ 20], the counterpart of (nuclear) magnetic resonance imaging (MRI), and (ii) molecular quantum bits with long coherence times [ 21]. Finally, the ground-state d 3 electronic configuration is associated with the largest ligand-field reorganization energies accompanying ligand exchange processes in octahedral complexes [ 22, 23]. [Cr IIIX 6] units are therefore kinetically inert, a rare situation for 3d metal complexes, but which is highly desirable for the rational design of heterometallic d-f assemblies with no statistical scrambling in solution [ 24, 25, 26]. For pseudo-octahedral Cr(III) complexes, there are two types of excited states. The first category corresponds to quartet excited states ( S = 3/2) resulting from the promotion of one or more electrons from the ground t 2 orbitals into the anti-bonding e * orbitals with no spin-flip (full traces in Figure 1a). Upon electromagnetic excitation, the latter transitions obey the spin rule, but their absorption cross-sections are still limited by the parity (Laporte) rule (d↔d transitions are magnetically allowed but electric-dipole transitions are forbidden). Interestingly, the resulting distorted quartet excited states, when their lifetimes are long enough, can be exploited to induce photochemical transformations [ 27, 28]. The second category refers to doublet excited states ( S = ½), where one electron undergoes a spin-switch upon light excitation (dashed traces in Figure 1a). The three lowest-energy spin-flip transitions with E/ B LnCr 2 > LnCr 3 > LnCr 4 > Cr because the electro-attractive character of Ln 3+ is reduced upon successive binding to several anionic biimidazolate binding units. One, however, notices that the absorption spectra extracted for LaCr 4, LaCr 3 and LaCr 2 ( Figure S2c) are slightly more red-shifted compared to the corresponding ones for LuCr 4, LuCr 3 and LuCr 2 species ( Figure 4c) because La 3+ is a weaker electron attractor than Lu 3+ due to its larger ionic radius ( Table 1, entry 7). For comparison purposes, similar titrations using Zn(OTf) 2, which is also diamagnetic but less charged and smaller than La 3+ and Lu 3+, showed that a maximum of three [(phen) 2Cr(biim)] + complex-as-ligands can be coordinated to a single Zn 2+ center ( Table 1, column 3 and Figure S2). The interpretation of the association constants β 1 , n M , C r (M = Lu, La and Zn) relies on the site-binding model where β 1 , n M , C r = K s t a t ୍ଠ K c h e m [ 69]. Kstat accounts for the purely statistical contribution produced by the change in rotational entropies [ 70]. It is obtained using the symmetry number method (see Supplementary File S1). Kchem refers to the chemical processes involved in the association process, which can be split into the intermolecular affinity Δ G M , C r = − R T l n f M , C r between the metal (M) and the complex-as-ligand (Cr), and the intramolecular interaction between the two complex-as-ligands Δ E C r , C r = − R T l n u C r , C r bound to the same metallic center (La 3+, Lu 3+ or Zn 2+). Assuming that Δ ECr,Cr is constant for a given metal, no matter the geometric arrangement of the ligands in the selected MCr n complex, one can write: β 1 , n M , C r = K s t a t ୍ଠ f M , C r n ୍ଠ ( u C r , C r n 2 − n 2 (6) The exponents n and ( n2 − n)/2 in Equation (6) represent the number of metal–ligand interactions and ligand–ligand interactions in the MCr n complex, respectively. Transformed into free energy, Equation (6) becomes: − R T l n β 1 , n M , C r − l n K s t a t = n Δ G M , C r + n 2 − n 2 Δ E C r , C r (7) which can be solved for Δ GM,Cr and Δ ECr,Cr with M = Zn (Equation (8)) or M = Ln (Equation (9)) using matrix formulations and multilinear least-squares methods ( Table 1, entries 5,6): − R T l n β 1 , 1 Z n , C r l n β 1 , 2 Z n , C r l n β 1 , 3 Z n , C r − ln 24 ln 120 ln 64 = Δ G Z n , C r 1 2 3 + Δ E C r , C r 0 1 3 (8) − R T l n β 1 , 1 L n , C r l n β 1 , 2 L n , C r l n β 1 , 3 L n , C r l n β 1 , 4 L n , C r − ln 32 ln 288 ln 702 ln 224 = Δ G L n , C r 1 2 3 4 + Δ E C r , C r 0 1 3 6 (9) For all metals, Δ GM,Cr is negative and represents the driving force of the intermolecular complexation reaction, while Δ ECr,Cr is positive, which points to stepwise anti-cooperative connections of cationic [(phen) 2Cr(biim)] + complex-as-ligands to the central M z+ centers. In fact, Δ ECr-Cr increases when the ionic radius of the metal decreases, but these values remain one order of magnitude smaller than the absolute value of Δ GM,Cr and are comparable with thermal energy at room temperature. Apart from the spectrophotometric titrations, the MCr n complexes lack any additional characterizations. ESI-MS studies demonstrate that the cationic {Ln[(biim)Cr(phen) 2] n} (3+n)+ aggregates do not survive in (or cannot be transferred into) the gas-phase. In solution, the slow-relaxing electron-spin paramagnetic Cr(III) centers induce fast nuclear-spin relaxation and severe line broadening, rendering the NMR signals of nearby nuclei undetectable; this is in line with the experimental 1H-NMR spectrum of [Cr(phen) 2(biim)] + being featureless [ 64]. Furthermore, the association constants deduced from spectrophotometric titrations show that intricate mixtures of different MCr n species are present in solution at millimolar concentrations, and this is regardless of the M:Cr stoichiometry ( Figures S3–S5). This drastically complicates selective crystallization processes, and attempts to obtain single crystals of these assemblies suitable for X-ray diffraction (XRD) were unsuccessful ( Figure 5, center). With this in mind, our efforts were redirected toward the preparation of Ln-Cr dyads, where both Ln and Cr centers are coordinatively saturated ( Figure 5 top and bottom). 2.2. Reaction of [(phen) 2Cr(biim)] + Complex-as-Ligand with Ln(Tp) 2(OTf) (Ln = Eu, Y) To isolate and characterize 1:1 [LnCr] dyads, we planned to react [(phen) 2Cr(biim)] + guests with a lanthanide host complex bearing ancillary ligands. The starting salts [Ln(hfac) 3dig] (hfac = 1,1,1,5,5,5-hexafluoro-pentane-2,4-dione) have been widely used in our group for this purpose and were considered as suitable precursors for the formation of {[(phen) 2Cr(biim)]Ln(hfac) 3} + [ 71]. However, these [Ln(hfac) 3] units proved to be unstable in polar solvents such as acetonitrile due to ligand scrambling [ 71]. This prevented clean reactions with [(phen) 2Cr(biim)] +, which is only soluble in polar solvents ( Figure 5, top). To overcome this limitation, we turned our attention toward the alternative lanthanide [Ln(Tp) 2(OTf)] host (Tp = hydridotris(1 H-pyrazol-1-yl)borate, Figure 5, bottom) [ 72] that Kaizaki and coworkers reacted with [(acac) 2Cr(ox)] − complex-as-ligand in polar solvents to give dinuclear [(acac) 2Cr(ox)Ln(Tp) 2] dyads (Ln = La, Nd, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, acac − = pentane-2,4-dionate, Figure 2b) [ 57, 60, 73, 74]. [Ln(Tp) 2(OTf)] (Ln = Eu, Y) were thus reacted with [(phen) 2Cr(biim)] + in acetonitrile to form the heterometallic complexes [(phen) 2Cr(biim)Ln(Tp) 2] 2+, which could be characterized in solution ( 1H-NMR, HRMS and spectrophotometry) and in the solid state (XRD). Contrary to the slow-relaxing paramagnetic LnCr n assemblies discussed above, the fast-relaxing [Eu(Tp) 2] +, or even diamagnetic [Y(Tp) 2] + moiety, possesses protons for which the 1H-NMR signal can be easily tracked. Upon stepwise titrations with slow-relaxing paramagnetic [(phen) 2Cr(biim)] + complex-as-ligand, the observed distance-dependent increase in line broadening of the 1H NMR signals along the series Cr···H4 Y > Er, in agreement with the standard lanthanide contraction trend [ 77]. The Ln-N distances are anecdotally shorter with the bound Tp − ligand than with the biimidazolate bridge ( Table S14), and the average Ln···Cr separation amounts to 5.86(2) Å (Ln = Eu, Y, Table S15). The trivalent chromium center adopts the well-known pseudo-octahedral coordination sphere distorted by the formation of three five-membered chelate rings, as previously discussed for [(phen) 2Cr(H nbiim)] (n+1)+ [ 64]. Interestingly, the Cr-N biim distances of 2.041(5) Å in [(phen) 2Cr(biim)Ln(Tp) 2] are slightly larger than 2.024(7) Å reported for the fully protonated [(phen) 2Cr(H 2biim)] 4+ complex ( Table S16). This fixes the electro-attracting effect of the trivalent lanthanide [Ln(Tp) 2] + bound to the biimidazolate bridge in the dyads to be comparable, or even slightly larger, than that of two protons in the solid state. In solution, the absorption spectra recorded for [(phen) 2Cr(biim)Ln(Tp) 2] 2+ (Ln = Y, Eu; Figure 7a) reveal that the ligand-to-ligand charge transfer (LLCT) band of the [(phen) 2Cr(biim))] + moiety undergoes a blue shift upon coordination to the [Ln(Tp) 2] + moiety comparable to the fixation of only a single proton to the biimidazolate bridge, as found in [Cr(phen) 2(Hbiim)] 2+ (compare the red trace with the blue and green traces in Figure 7a). Consequently, the weak spin-forbidden chromium-centered absorption bands Cr( 2E← 4A 2) and Cr( 2T 1← 4A 2), expected around 720–750 nm in [(phen) 2Cr(biim)Ln(Tp) 2] 2+ dyads, are masked by the more intense band foot of the LLCT transitions, and could not be identified. As a logical issue, upon excitation in the UV, the searched NIR emission bands arising from Cr( 2E) and Cr( 2T 1) excited states in [(phen) 2Cr(biim)Ln(Tp) 2] 2+ are completely quenched by non-radiative energy transfer toward the broad non-emissive LLCT levels at room temperature ( Figure 7b). At 77 K, the non-radiative processes slow down, and the three complexes [(phen) 2Cr(biim)Ln(Tp) 2] + (Ln = Y, Eu, Er) become emissive, showing Cr( 2E→ 4A 2) phosphorescence at λ max = 750 nm (13333 cm −1, Figure 7c). We conclude that the spin-flip excited states Cr( 2E) and Cr( 2T 1) in [(phen) 2Cr(biim)Ln(Tp) 2] 2+ dyads relax too fast at room temperature to act as donors that efficiently transfer their energy to the lanthanide neighbor working as the energy acceptor. This restricts their potential to be used as sensitizers at room temperature for both light downshifting and light upconversion. The required removal of the deleterious quenching by the low-energy LLCT levels was therefore considered as a priority. We thus decided to investigate similar Cr(III)-Ln(III) dyads, but replacing the phenanthrolines with the saturated cyclam ligand (1,4,8,11-tetraazacyclotetradecane) lacked accessible low-energy π* orbitals. 2.3. Synthesis and Characterization of [cyclam)Cr(biim)Ln(Tp) 2](OTf) (Ln = Eu, Y, Er) Dyads The synthesis of the heteroleptic [(cyclam)Cr(H 2biim)] 3+ scaffold was performed using the Kane–Maguire strategy [ 24] and following a literature procedure [ 78]. The starting Cr(III) salt CrCl 3·6H 2O was reacted with 0.9 equivalent of cyclam in boiling DMF, producing cis-[Cr(cyclam)Cl 2]Cl as a purple powder ( Figure 8). The cis configuration is retained thanks to a small excess of CrCl 3·6H 2O, which prevents base-catalyzed isomerization to the trans isomer [ 79]. Anion exchange was then performed to convert cis-[(cyclam)CrCl 2]Cl into cis-[(cyclam)Cr(OTf) 2]OTf using triflic acid and releasing HCl gas as a side product. The labile triflates could finally be substituted with a didentate H 2biim ligand to give the target [(cyclam)Cr(H 2biim)](OTf) 3 building block as an orange solid ( Figure 8). Recrystallization by vapor diffusion of Et 2O into a methanolic solution produced crystals suitable for XRD studies ( Supplementary File S3, Tables S3 and S4 and Figures S15 and S17). As observed with [(phen) 2Cr(H 2biim)] 3+ ( Figure 3) [ 64], the acidic protons of the biimidazole ligand in [(cyclam)Cr(H 2biim)] 3+ can be removed in basic media. Reaction with one equivalent of NaOH produced [Cr(cyclam)(Hbiim)](OTf) 2, which could be isolated in good yield as an orange solid ( Figure 8) and recrystallized for crystal structure determination ( Supplementary File S3, Tables S5 and S6 and Figures S16 and S18). Using an excess of triethylamine in CH 3CN causes the immediate precipitation of the doubly deprotonated [(cyclam)Cr(biim)]OTf salt as an insoluble orange solid in quantitative yield ( Figure 8). Its insolubility in both water and organic solvents prevented the preparation of single crystals suitable for X-ray diffraction. Similarly, it was also impossible to determine the p Ka’s for the protons of [Cr(cyclam)(H 2biim)] 3+ for solubility reasons, but one can reasonably assume values like those measured for [(phen) 2Cr(H 2biim)] 3+ (p Ka 1 = 4.67(3), p Ka 2 = 8.59(11) in Figure 3) [ 64]. Heterogeneous mixing of the insoluble salt [Cr(cyclam)(biim)]OTf with [Ln(Tp) 2(OTf)] in acetonitrile led to their complete dissolution and the formation of the soluble heterometallic complex [(cyclam)Cr(biim)Ln(Tp) 2] 2+. The complex could be isolated by precipitation with Et 2O or by evaporation of the solution to give [(cyclam)Cr(biim)Ln(Tp) 2](OTf) 2 in good yield ( Figure 8). Recrystallization provided crystals measurable by XRD ( Supplementary File S3: Ln = Y, Figure S22 and Table S9; Ln = Eu, Figures S23 and S25 and Table S10; Ln = Er, Figures S24 and S26 and Table S11). As expected, the Cr(III) metal keeps its distorted pseudo-octahedral [CrN 6] coordination sphere and Er(III) is coordinated by eight nitrogen atoms in [(cyclam)Cr(biim)Ln(Tp) 2](OTf) 2 dyads ( Figure 8, dCr-Ln = 5.89(1) Å, Table S15) as previously discussed for [(phen) 2Cr(biim)Ln(Tp) 2](OTf) 2 ( Figure 5, dCr-Ln = 5.86(2) Å, Table S15). In acetonitrile solution, the 1H-NMR spectra of the heterometallic complexes [(cyclam)Cr(biim)Ln(Tp) 2] 2+ (Ln = Y and Eu) display measurable broad peaks for the protons of the Tp − ligands, because they are sufficiently remote from the slow-relaxing paramagnetic Cr(III) center ( Figures S32 and S33), as previously described for the [(phen) 2Cr(biim)Ln(Tp) 2] 2+ dyads. The 1H NMR data thus supports that the Cr(III)-containing unit is coordinated to the [Ln(Tp) 2] + moiety in solution. In addition, millimolar solutions of [(cyclam)Cr(biim)Ln(Tp) 2] 2+ analyzed by HRMS (Ln = Y, Eu, Er; Figures S34–S36) systematically exhibit intense signals corresponding to {[(cyclam)Cr(biim)Ln(Tp) 2]OTf} + ( Figures S34–S36). 2.4. Photophysical and Light Conversion Properties of [cyclam)Cr(biim)Ln(Tp) 2](OTf) 2 (Ln = Eu, Y, Er) Dyads Having [(cyclam)Cr(H nbiim)] (n+1)+ ( n = 0–2) and [(cyclam)Cr(biim)Y(Tp) 2] 2+ at hand ( Figure 8), the Cr(III)-centered photophysical properties, in the absence of any intermetallic Cr↔Ln energy transfers, are accessible. This is not the case for the [ErN 8] site, which is only found close to Cr(III) in [(cyclam)Cr(biim)Er(Tp) 2] 2+. It was therefore highly desirable to access the photophysical properties of erbium complexes possessing similar coordination spheres, but in the absence of communication with the Cr(III) center. To accurately mimic the coordination environment of the Ln(III) ion, the [(cyclam)Cr(biim)] + unit in [(cyclam)Cr(biim)Ln(Tp) 2] 2+ was replaced with the didentate 1,1′-dimethyl-1 H,1′ H-2,2′-biimidazole (Me 2biim) ligand to give [(Me 2biim)Ln(Tp) 2]OTf ( Figure 9 and Supplementary File S3: Ln = Y, Figures S27 and S29 and Table S12; Ln = Er, Figures S28 and S30 and Table S13). In acetonitrile, the 1H-NMR spectrum of diamagnetic [(Me 2biim)Y(Tp) 2]OTf demonstrated the selective formation of the heteroleptic complex in solution ( Supplementary File S4, Figure S37). Additionally, the HRMS of solutions of [(Me 2biim)Ln(Tp) 2]OTf (Ln = Y, Figure S38 and Ln = Er, Figure S39) display peaks of [(Me 2biim)Ln(Tp) 2] +, thus confirming the formation of the target complexes in solution. The complementary [(Me 2biim)Eu(Tp) 2]OTf adduct based on Eu(III) was prepared in situ by stoichiometric mixing of Me 2biim and [Eu(Tp) 2(OTf)] in acetonitrile. The absorption spectra of the complexes [(cyclam)Cr(H 2biim)] 3+ (yellow traces in Figure 10) and [(cyclam)Cr(Hbiim)] 2+ (blue traces in Figure 10) could be recorded in acetonitrile, but [(cyclam)Cr(biim)]OTf was too insoluble in all available solvents to access its photophysical properties in solution. The UV part of the absorption spectra displays intense absorption bands (ε > 10 3 M −1·cm −1), which correspond to π*←p transitions centered on the aromatic biimidazole ring, completed with charge transfer transitions (LMCT or MLCT, Figure 10, top). Compared with [(phen) 2Cr(Hbiim)] 2+ ( Figure 7a), the absence of low-energy LLCT charge transfer bands in [(cyclam)Cr(Hbiim)] 2+ ( Figure 10) cleared the 450-800 nm visible domain from intense transitions, thus giving access to the easy detection of the weaker metal-centered spin-allowed Cr( 4T 2← 4A 2) transition around 500 nm (ε < 500 M −1·cm −1, Figure 10, top) and spin-forbidden Cr( 2T 1, 2E← 4A 2) transition around 700 nm (0.55 M −1·cm −1 < ε < 0.67 M −1·cm −1, Figure 10, bottom). The [(cyclam)Cr(biim)Y(Tp) 2] 2+ dyad exhibits an absorption spectrum like the one of [Cr(cyclam)(Hbiim)] 2+, thus confirming that the [Ln(Tp) 2] + group has an electro-attracting strength on the biimidazolate bridge comparable to a single H +. These characteristics match those previously observed for [(phen) 2Cr(biim)Ln(Tp) 2] 2+ assemblies ( Section 2.2). However, the maximum of the spin-flip Cr( 2T 1, 2E← 4A 2) absorption bands is blue-shifted for [Cr(cyclam)(H 2biim)] 3+ ( λmax = 689 nm) and for the related cyclam-based [(cyclam)Cr(biim)Y(Tp) 2] 2+ dyads ( λmax = 700 nm, Figure 10, bottom), compared to their [(phen) 2Cr(biim)Ln(Tp) 2] 2+ analogs at 77 K ( λmax = 750 nm in Figure 7c). This results from the weaker nephelauxetic effect induced by the aliphatic cyclam ligand, compared to that produced by the polyaromatic phenanthroline units. The complex [(cyclam)Cr(biim)Er(Tp) 2] 2+ additionally exhibits the typical Er( 2S+1LJ← 4I 15/2) f-f absorption bands beyond 600 nm ( Figure 10, bottom, and Figure S40), which exactly fit those recorded for [(Me 2biim)Er(Tp) 2] + ( Figure S40). In line with the crystal structures collected in the solid state, one concludes that the didentate ligand Me 2biim mimics satisfyingly the local environment around the Er(III) ion found upon connection of [(cyclam)Cr(biim)] + complex-as-ligand to Er(Tp) 2] + for the dyad in solution. Finally, the radiative rate constants krad for the Cr( 2E/ 2T 1← 4A 2) and Er( 2S+1LJ← 4I 15/2) transitions, which are crucial for estimating intrinsic emission quantum yields (vide infra), could be estimated from the absorption spectra using the Strickler-Berg Equation (11) [ 80, 81, 82]. They are collected in Table 2 and Table S17: k r a d = 2303 ୍ଠ 8 π c n 2 ν ~ 2 g G S N A g E S ∫ ε ν ~ d ν ~ (11) Here, c is the speed of light in vacuum (cm·s −1), n is the refractive index of the solvent, N A is Avogadro’s number (mol −1), gGS is the degeneracy of the ground state, gES is the degeneracy of the excited state, ν ~ is the barycenter of the transition in wavenumber (cm −1) and ∫ ε ν ~ d ν ~ is the area under the absorption spectrum of each transition (M −1·cm −2). Upon excitation in the UV, the complexes [Cr(cyclam)(H 2biim)] 3+ and [Cr(cyclam)(Hbiim)] 2+ emit Cr( 2E, 2T 1→ 4A 2) phosphorescence at 690 nm (14,500 cm −1) and 701 nm (14,300 cm −1) respectively ( Figure 11a). The emission bands are isoenergetic with the absorption bands of the transition 2E← 4A 2, indicating negligible Stokes shifts ( Figure 10). The measured total emission lifetime τtot reaches 1.13 μs and 0.42 μs for [Cr(cyclam)(H 2biim)] 3+ and [Cr(cyclam)(Hbiim)] 2+, respectively, from which the intrinsic quantum yields ( Φintrinsic) could be estimated using Equation (12), where krad is the radiative rate constant ( Table 2): ɸ i n t r i n s i c = k r a d k r a d + k n o n r a d = k r a d · τ t o t (12) Φintrinsic = 1.6·10 −4 found for [Cr(cyclam)(H 2biim)] 3+ and Φintrinsic = ୫.୭ ୍ଠ ୧୦ −5 for [Cr(cyclam)(Hbiim)] 2+ are low compared to many other Cr(III) complexes for which the emission quantum yield can be higher than 0.1 at room temperature [ 31, 32, 35, 83]. This implies efficient non-radiative relaxation pathways tentatively ascribed to the presence of the high-energy N-H bond oscillators of the cyclam ring close to the Cr(III) center [ 26, 32, 84]. To test this hypothesis, we tried to synthesize the same complex with the fully N-methylated variant of cyclam: Me 4cyclam, but all our attempts to complex this bulky ligand to Cr(III) proved to be unsuccessful ( Supplementary File S5 in the ESI). UV-Visible excitations of the complexes [(cyclam)Cr(biim)Ln(Tp) 2] 2+ (Ln = Y, Eu, Er) at room temperature ( Figure 11b), or at 77K ( Figure S44), result in spin-flip Cr( 2E/ 2T 1→ 4A 2) emissions at λmax = 700 nm together with a trace of the forced electric-dipole hypersensitive Eu( 5D 0→ 7F 2) band at 618 nm for the Cr-Eu pair ( Figure 11b). The total Cr( 2E) emission lifetimes of 0.23 ≤ τtot ≤ 0.25 ms recorded for Cr-Y and Cr-Eu dyads are shorter than those gathered for [Cr(cyclam)(H 2biim)] 3+ and [Cr(cyclam)(Hbiim)] 2+ ( Table 2), pointing to a global increase in the non-radiative deexcitation rate constant of the latter excited state in the dyads. The further reduction in the lifetime to reach τ(Cr( 2E)) tot = 0.1 ms in Cr-Er, combined with the observation of (i) a dual NIR Cr( 2E/ 2T 1→ 4A 2) ( Figure 12a) and IR Er( 4I 13/2→ 4I 15/2) luminescence ( Figure 12b) and (ii) a biexponential decay of the Cr( 2E) donor level ( Table 2 and Figure S48), suggests the operation of a reversible energy transfer (ET) process connecting the two metallic centers ( Figure 12c). The latter assumption is corroborated by (i) the excitation spectrum upon analyzing the Er( 4I 13/2→ 4I 15/2) emission at 1550 nm, which shows the spin-allowed Cr( 4T 2→ 4A 2) transition at 525 nm acting as the main sensitizing channel ( Figure S45) and (ii) the successful sensitization of the Er( 4I 13/2→ 4I 15/2) emission by selective excitation of the spin-forbidden Cr( 2E/ 2T 1← 4A 2) transition using a 698 nm laser beam ( Figure S46). Following an initial pulsed excitation, the solution of the differential Equation (13) modeling the biexponential time-dependent relaxation of the Cr *Er level (referred to as 1 in the associated energy diagram of Figure 12d) is obtained using the Lagrange-Sylvester formula ( Supplementary File S6) [ 59, 85]. Reasonably fixing kCr as the inverse of the total emission lifetime recorded for Cr-Y ( kCr = 1/τ tot = ୪.୦ ୍ଠ ୧୦ 6 s −1, Table 2), the three other constants kEr = ୧.୨ ୍ଠ ୧୦ 6 s −1, kET = ୪.୩ ୍ଠ ୧୦ 6 s −1 and kBET = ୨.୫ ୍ଠ ୧୦ 6 s −1 could be estimated by a non-linear least-square fit of the relative population N(|1⟩( t) (computed with the Lagrange-Sylvester formula) to the experimental biexponential emission trace of the Cr( 2E, 2T 1) excited level observed for the Cr-Er dyad ( Figure S48 in Supplementary File S6). The Cr→Er energy transfer rate constant kET = ୪.୩ ୍ଠ ୧୦ 6 s −1 extracted with this method for [(cyclam)Cr(biim)Er(Tp) 2] 2+ ( dCr-Er = 5.9 Å, Table S15) is circa 50 times larger than the value of kET = ୮.୯୬ ୍ଠ ୧୦ 4 s −1 previously reported for {[(dqp)Cr(L2)] 3Er} 6+ ( dCr-Er = ୧୪ ସ୍ଖ, Figure 2c) [ 61]. The intrinsic energy transfer efficiency η E T C r → E r can now be estimated from kCr and kET (ignoring the back-energy transfer): η E T C r → E r = k E T k C r + k E T = 4.3 ୍ଠ 10 6 4.0 ୍ଠ 10 6 + 4.3 ୍ଠ 10 6 = 0.52 (14) It reaches η E T C r → E r = 52% efficiency, which is favorable (optimum = 50%) for boosting energy transfer upconversion (ETU) in molecular complexes [ 59, 86]. 2.5. Looking for Light Upconversion in [(Me 2biim)Er(Tp) 2]OTf Complex and in [(cyclam)Cr(biim)Er(Tp) 2](OTf) 2 Dyad Using high-excitation-power lasers, it is possible to excite [(Me 2biim)Er(Tp) 2] + and [(cyclam)Cr(biim)Er(Tp) 2] 2+ through the Laporte-forbidden intrashell f-f transitions of the Er(III) center at 801 nm (Er( 4I 9/2← 4I 15/2)) and at 966 nm (Er( 4I 11/2← 4I 15/2)). Standard one-photon light downshifting processes result in ultimate Er( 4I 13/2⟶I 15/2) emission at 1550 nm ( Figure 13Figures S49 and S50). Surprisingly, upon UV excitation at 250 nm, the excited levels centered on either Tp − or Me 2biim ligands bound to Er(III) in mononuclear [(Me 2biim)Er(Tp) 2] only show broad ligand-centered π*⟶π emission bands ( Figure S51) with no trace of Er-centered emission induced by the antenna effect ( Figure 13a). This contrasts with the successful indirect UV-based sensitization of Er(III) centers observed in the [(cyclam)Cr(biim)Er(Tp) 2] 2+ dyad, where the doubly deprotonated bridging ligand biim 2− ensures ultimate sensitization of Er( 4I 13/2→ 4I 15/2) via energy transfer ( Figure 12c). Using highly focused 801 nm and 966 nm lasers, beyond inducing the light downshifting response described above ( Figure 13a), the excitation light is intense enough to induce light upconversion in solution for both [(Me 2biim)Er(Tp) 2] + and [(cyclam)Cr(biim)Er(Tp) 2] 2+ at room temperature ( Figure 14). The upconversion signals are, however, very weak using the 801 nm laser (close to noise level, see Figure 14a), but become slightly more intense when using the 966 nm laser ( Figure 14b) because (i) our 966 nm laser can reach higher excitation power and (ii) the Er( 4I 11/2← 4I 15/2) transition absorbs more efficiently the initial photon densities than the Er( 4I 9/2← 4I 15/2) transition (see Figure S40). The upconversion emission spectra display two bands corresponding to the blue-green Er( 4S 3/2→ 4I 15/2) and Er( 2H 11/2→ 4I 15/2) transitions. Except for excitation of [(cyclam)Cr(biim)Er(Tp) 2] 2+ at 801 nm, which provided upconverted signals that were too weak to be safely recorded at low incident powers, the integrations of the upconverted emission bands upon 966 nm excitation for both complexes, [(Me 2biim)Er(Tp) 2] + and [(cyclam)Cr(biim)Er(Tp) 2] 2+, are strong enough to be accessible ( Figures S52 and S53). They show linear log( I) vs log( P) plots, with an experimental slope close to 2, which confirms the absorption of two successive photons according to the single-center Excited State Absorption (ESA) mechanism depicted in Figure 15 and previously established for closely related mononuclear Er(III) complexes in solution [ 59, 87]. Finally, chromium-centered excitation into the Cr( 2E, 2T 1← 4A 2) at 698 nm for [(cyclam)Cr(biim)Er(Tp) 2] 2+ gave only faint Er( 4S 3/2→ 4I 15/2) and Er( 2H 11/2→ 4I 15/2) upconverted emission bands at any temperature ( Figure S54), but induced standard light-downshifting via efficient intramolecular Cr→Er energy transfer detailed in Section 2.4 ( Figure 13b). One concludes that no reasonable Energy Transfer Upconversion (ETU) mechanism could be implemented for this specific Cr-Er dyad in solution. To summarize, the upconversion signals induced upon excitation at λexc = 698, 801 or 966 nm of [(Me 2biim)Er(Tp) 2] + and [(cyclam)Cr(biim)Er(Tp) 2] 2+ remain very weak (close to the detection limit of our setup), while other Er(III) complexes reported in the literature upconvert much more efficiently [ 86, 88, 89, 90]. Furthermore, we found that the complex [(Me 2biim)Er(Tp) 2] +, which contains no chromium sensitizer, upconverts more efficiently than the chromium-containing [(cyclam)Cr(biim)Er(Tp) 2] 2+ dyad upon all the light excitations explored in this work. 3. Materials and Methods 3.1. Solvents and Starting Materials Chemicals were purchased from Sigma-Aldrich (Burlington, MA, United States) and Acros (Switzerland) and used without further purification unless otherwise stated. Dry reagents were either purchased as packed under inert atmosphere with acroseal and molecular sieves or distilled by appropriate procedures. Dry solvents were distilled over either calcium hydride or metallic sodium. Silica-gel plates (Merck (Germany), 60 F 254) were used for thin-layer chromatography, and SiliaFlash ପ୍ପ (Quebec, Canada) silica gel P60 (0.04–0.063 mm) and Acros (Stabio, Switzerland) silica gel 60 (0.035–0.07 mm) were used for preparative column chromatography. The complex [Cr(phen) 2(biim)]OTf was synthesized according to the published procedure [ 64]. Potassium hydrotris(1-pyrazolyl)borate (KTp) and its Ln(III) complexes Ln(Tp) 2OTf (Ln = Y, Eu) were adapted from the literature procedure [ 72]. 3.2. Spectroscopic and Analytical Measurements 1H and 13C NMR spectra were recorded at 298 K on a Bruker Avance (Karlsruhe, Germany) 400 MHz spectrometer equipped with BCU temperature control for variable temperature measurements. Chemical shifts are given in ppm with respect to TMS. Pneumatically-assisted electrospray (ESI-MS) mass spectra were recorded from ~1 × 10 −4 M (ligands) and ~1 × 10 −3 M (complexes) solutions on an Applied Biosystems (Foster City, CA, United States) API 150EX LC/MS System equipped with a Turbo Ionspray source. Elemental analyses were performed by K. L. Paglia from the Microchemical Laboratory of the University of Geneva. Electronic spectra in the UV-Vis region were recorded at 293 K from solutions in CH 3CN with a Perkin-Elmer Lambda 1050 (Waltham, MA, USA) using quartz cells of 0.1 or 1.0 mm path length. Solid-state absorption spectra were recorded with a Perkin-Elmer Lambda 900 using capillaries. The emission spectra were recorded using a Fluorolog (Horiba Jobin-Yvon) instrument equipped with an iHR320 imaging spectrometer, a 450 W xenon lamp illuminator (FL-1039A/40A) and a Peltier-cooled photomultiplier tube (PMT Hamamatsu R928P, Hamamatsu, Japan). The emission spectra were corrected for the wavelength-dependent sensitivity of the PMT. Time-resolved data were collected using a digital oscilloscope (Tektronix MDO4104C, Portland, OR, United States) coupled to a water-cooled photomultiplier tube (PMT Hamamatsu R928P) or to a time-gated photomultiplier module (Hamamatsu H11526-20-NF, Hamamatsu, Japan). Pulsed excitation at 355 nm was achieved using the third harmonic of a pulsed Nd:YAG laser (Quantel Q-Smart 850, Lumibird, France) or MPL-F-355nm-100mW-DB22’005 from CNI Laser (10kHz repetition rate). Variable temperature measurements were done using a closed-cycle cryosystem (Janis, CCS-900/204N, Woburn, MA, United States) with the sample sitting in the exchange gas (helium) to achieve efficient cooling. Complexes of known corrected molecular weight were dissolved in acetonitrile to obtain ~1 mM solutions that were immobilized in 2 mm-diameter cylindrical quartz cuvettes. The cuvettes, sealed with fast-drying silver agar gel, were mounted on a metallic copper sample holder. The ultrafast time-gated experiments used a pulsed laser as an excitation light source (355 nm, CNI Laser MPL-F-355, Changchun, China). Time-correlated single-photon counting was performed with a quTAG MC (quTools), and the detector of the FluoroLog (R5509-73, Horiba Scientific, Kyoto, Japan) was used as the detector. 3.3. X-Ray Crystallography Summary of crystal data intensity measurements and structure refinements for complexes [(cyclam)Cr(H 2biim)]OTf 3, [(cyclam)Cr(Hbiim)]OTf 2, [(phen) 2Cr(biim)Ln(Tp) 2] (Ln = Y, Eu) [(cyclam)Cr(biim)LnTp) 2]OTf 2 (Ln = Y, Eu, Er) and [Ln(Tp) 2(Me 2biim)]OTf complexes (Ln = Y, Er) are collected in Supplementary File S3 with pertinent bond lengths and bond angles. ORTEP views with numbering schemes. The crystals were mounted on Hampton cryoloops with protection oil. X-ray data collections were performed with an XtaLAB Synergy-S diffractometer (Rigaku, Japan) equipped with a hybrid pixel array “hypix arc 150” detector. The structures were solved by using dual-space methods [ 91]. Full-matrix least-squares refinements on F 2 were performed with SHELXL-2020 [ 92]. CCDC 2550803-2550811 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Center via https://www.ccdc.cam.ac.uk/structures/ (accessed on 11 May 2026). 3.4. Synthesis 3.4.1. Synthesis of Potassium hydrotris(1-pyrazolyl)borate (KTp) [ 72] In a flask, 20 g of pyrazole (291 mmol) and 4 g of KBH 4 (74.2 mmol) were mixed, and the resulting solid was heated at 190° for 2 h under reflux, until no more hydrogen evolved. The hot mixture was precipitated in toluene under stirring, then filtered. The white solid was recrystallized in anisole. A total of 13.620 g (54.0 mmol, 73%) of KTp was obtained. 1H NMR (400 MHz, CD 3CN) δ 7.54 (d, J = 2.1, 3H), 7.44 (d, J = 1.6 Hz, 3H), 6.07 (t, J = 1.9 Hz, 3H), and 4.71 (m, 1H). 3.4.2. Synthesis of [Y(Tp) 2OTf] [ 72] In a flask, 1.290 g (2.41 mmol) of Y(OTf) 3 was dissolved in 50 ml of dry THF. Molecular sieves were added to the solution to remove possible traces of water remaining in the solvent and in the salt. A total of 1.215 g (4.82 mmol) of KTp was added, and the solution was stirred at room temperature for 1 h. The THF was removed under reduced pressure, and the residue was dried under vacuum. Toluene was added, and the suspension was filtered to remove insoluble KOTf. The solution was evaporated under reduced pressure, then dried under vacuum at 100 °C for 30 min. To remove the last traces of toluene, the white solid was dissolved in 250 mL of Et 2O, and the solution was evaporated under reduced pressure. The white solid was dried under vacuum overnight. A total of 1.359 g (2.05 mmol, 85%) of YTp 2OTf was obtained. 1H NMR (400 MHz, CD 3CN) δ 7.84 (d, J = 2.1 Hz, 6H), 7.16 (d, J = 2.1 Hz, 6H), 6.15 (t, J = 2.1 Hz, 6H), 4.74 (very br m, 2H). Anal. calc. for C 19H 20B 2F 3N 12O 3SY: C 34.37, H 3.04, N 25.31. Found: C 34.13, H 3.16, N 25.29. 3.4.3. Synthesis of Eu(Tp) 2OTf [ 72] In a flask, 1.141 g (1.90 mmol) of Eu(OTf) 3 was dissolved in 50 ml of dry THF. Molecular sieves were added to the solution to remove possible traces of water remaining in the solvent and in the salt. A total of 0.960 g (3.81 mmol) of KTp was added, and the solution was stirred at room temperature for 1 h. The THF was removed under reduced pressure, and the flask was dried under vacuum. Toluene was added, and the suspension was filtered to remove insoluble KOTf. The solution was evaporated under reduced pressure, then dried under vacuum at 100 °C for 30 min. To remove the last traces of toluene, the white solid was dissolved in 250 mL of Et 2O, and the solution was evaporated under reduced pressure. The white solid was dried under vacuum overnight. A total of 1.003 g (1.38 mmol, 72%) of EuTp 2OTf was obtained. 1H NMR (400 MHz, CD 3CN) δ 13.93 (s, 6H), 3.04 (d, J = 2.1 Hz, 6H), 0.46 (s, 6H), −1.06–−2.19 (very br m, 2H). Anal. calc. for C 19H 20B 2F 3N 12O 3SEu: C 31.39, H 2.77, N 23.12. Found: C 31.37, H 2.72, N 23.21. 3.4.4. Synthesis of [(Me 2biim)Y(Tp) 2]OTf In a flask, 25 mg (0.15 mmol) of Me 2biim and 100 mg (0.15 mmol) of YTp 2OTf were dissolved in dry MeCN. The solution was evaporated under reduced pressure, yielding 111 mg (0.13 mmol, 89%) of [(Me 2biim)Y(Tp) 2]OTf as a white powder. Vapor diffusion of pentane into a THF solution of [(Me 2biim)Y(Tp) 2]OTf led to the formation of block-shaped crystals suitable for XRD. Anal. calc. for C 27H 30B 2F 3N 16O 3SY: C 39.25, H 3.66, and N 27.12. Found: C 38.76, H 3.48, N 26.78. m/z (ESI-HRMS): calcd. for [Y(Tp) 2(Me 2biim)] + (C 26H 30B 2N 16Y +): 677.209, found: 677.208. 3.4.5. Synthesis of [(Me 2biim)Er(Tp) 2]OTf In a flask, 26 mg (0.16 mmol) of Me 2biim and 120 mg (0.16 mmol) of ErTp 2OTf were dissolved in dry MeCN. The solution was stirred for 5 min at room temperature. The solution was then evaporated under reduced pressure, yielding 124 mg (0.14 mmol, 85%) of [(Me 2biim)Er(Tp) 2]OTf as a white powder. Vapor diffusion of pentane into a THF solution of [(Me 2biim)Er(Tp) 2]OTf led to the formation of block-shaped crystals suitable for XRD. Anal. calc. for C 27H 30B 2F 3N 16O 3SEr·1.2H 2O: C 35.01, H 3.53, and N 24.20. Found: C 34.84, H 3.29, and N 23.97. m/z (ESI-HRMS): calcd. for [Er(Tp) 2(Me2biim)] + (C 26H 30B 2N 16Er +): 754.235, found: 754.235. 3.4.6. Synthesis of [Cr(cyclam)Cl 2]Cl [ 78, 79] A total of 1.00 g (5.00 mmol) of cyclam and 1.50 g (5.63 mmol) of CrCl 3·6H 2O were dissolved in DMF (30mL). The solution was refluxed for 20 min. The solution was cooled and filtered, and the purple precipitate was washed with 20 mL of acetone and dried. A total of 1.57 g (4.37 mmol, 88%) of cis-[Cr(cyclam)Cl 2]Cl was obtained as a purple powder. 3.4.7. Synthesis of [Cr(cyclam)OTf 2]OTf [ 93] A total of 1.568 g (4.37 mmol) of cis-[Cr(cyclam)Cl 2]Cl was added to 2 mL (22.8 mmol) of trifluoromethanesulfonic acid in a Schlenk tube connected to an N 2 flux and a trap bubbler with a Na 2CO 3 aqueous solution to quench gaseous HCl. The solution was stirred for 1 h at room temperature, then 1 h at 100 °C. The mixture was cooled to 5 °C and was carefully poured into 40 mL of Et 2O. A pink precipitate started to form with a little bit of scratching to help the precipi

www.mdpi.com

Zum Originalartikel