Coordination Mode Engineering of Silver Cages: From ACQ to AIE for Sequential Artificial Light Harvesting and Near-White Light Emission Authors: Pranay Kumar Maitra, Anusmita Rout, Partha Sarathi Mukherjee Published: 2026 ABSTRACT Silver­(I)-based metallacages are gaining interest in supramolecular chemistry owing to their tunable properties and diverse applications. Herein, we report a sequential artificial light-harvesting system using a silver-based cage. Two metallacages were synthesized: one coordinated through nitrogen (MC 1 , N–Ag–N) and the other through carbene carbon (MC 2 , C–Ag–C), each with newly constructed aggregation-induced emissive (AIE-active) ligands. MC 1 presents a flexible structure and exhibits AIE properties in the aggregate form, allowing for efficient three-stage sequential energy transfer from the cage (MC 1 ) to eosin-Y (ESY), to sulforhodamine 101 (SR101), and finally to Cyanine-5-maleimide (Cy5). Notably, we can easily achieve bright white-light emission with a high quantum yield of 93%. In sharp contrast, MC 2 with a rigid structure exhibits high fluorescence in the solution state, while significant fluorescence quenching occurs in its aggregated state, thereby prohibiting its application in light-harvesting systems. This work focuses on the crucial function of structural flexibility in silver-based metallacages in energy-transfer processes and promotes silver-based metallacages as promising new candidates for advanced sequential artificial light-harvesting. Introduction Natural light-harvesting systems (LHSs) are crucial to photosynthesis that underpins life on the Earth by transforming sunlight into chemical energy. Plants, algae, fungi and bacteria generate the base of the energy pyramid, converting sunlight into essential compounds like carbohydrates, proteins, fatty acids and cellulose. − The explanation for photosynthetic processes motivated researchers to construct synthetic LHS systems that replicate the natural processes associated with sunlight to optimize the operation of solar resources. − In natural LHSs, chromophores (donors) clustered in closely packed surfaces absorb sunlight and funnel incoming energy to the acceptor via very efficient pathways. Mimicking this process in artificial systems requires overcoming key challenges like: (i) donor chromophores must aggregate without fluorescence decay, (ii) there should be an overlap between donor emission and acceptor absorption band, (iii) donor–acceptor separation must ensure balance of through-space interactions and avoid collisional quenching. − To fulfill these requirements, various platforms such as polymers, metal–organic frameworks, dendrimers, covalent organic cages, DNA, and supramolecular systems have been studied by researchers. − These systems can be categorized into covalent as well as noncovalent architectures. Covalent systems generally exhibit high stability but suffer from synthetic limitations because that reduce the number of donors per acceptor, leading to ACQ and low efficiency of energy transfer. − While noncovalent systems utilize dynamic interactions such as π–π stacking, hydrogen bonding, and ion–dipole interactions, these enable flexible donor–acceptor arrangements, providing an improved energy transfer capability. − Supramolecular coordination complexes (SCCs) with aggregation-induced emission (AIE) features are an emerging platform for the construction of artificial LHSs. The uniqueness of AIE systems lies in their enhanced emission upon the aggregation process, which mimics the close packing of chromophores utilized in natural photosynthesis. Since Würthner et al. in 2006 reported the SCC-based artificial LHS, the field is relatively unexplored compared to other applications of SCCs, including catalysis, sensing, and biomedical diagnostics. − Most studies on SCCs based LHSs have concentrated on single-step energy transfer. − In comparison, sequential energy transfer (SET) is the characteristic property of natural LHSs in which energy is transmitted stepwise through intermediate acceptors. SET systems have unique advantages, such as large Stokes shifts, and the possibility to transfer energy over distances beyond the Förster radius where no direct spectral overlaps are made. − Nevertheless, SET-based artificial LHSs are still rare, with no existing studies based on silver metallacages. Silver­(I)-based metallacages, being a category of SCCs, are well-defined three-dimensional architectures formed through self-assembly. The design harnesses the flexible coordination chemistry of silver ions, which can produce either rigid or flexible structures depending on their ligands. − Silver-based cages have been well recognized for their applications in catalysis, drug delivery, and host–guest chemistry; however, their utilization in artificial LHSs is mostly unexplored. − For example, in 2023, we described the first silver-carbene complex in solution for artificial LHSs; however, that system was unable to support two-step transfer of energy. This study bridges the gap by demonstrating the first sequential artificial light-harvesting approach using a silver-based metallacage. Two different cages were synthesized: MC 1 , a flexible nitrogen coordinated cage; and MC 2 , a more rigid carbene-carbon coordinated cage. In the aggregated form, MC 1 exhibited strong AIE, performing three-step sequential energy distribution from eosin Y (ESY) to sulforhodamine 101 (SR101) and finally to cyanine-5-maleimide (Cy5) (Scheme ). On the other hand, the fluorescence in a dilute solution of MC 2 is very high; however, fluorescence emission of MC 2 is quenched in the aggregate state, thus not serving as a light-harvesting material. These results highlight the crucial role of flexibility in the structural designs of effective energy transfer and position silver-based metallacages extensively as materials for sequential artificial light harvesting. Schematic Representation of the Synthesis of MC 1 and MC 2 Results and Discussion The new ligands L 1 and L 2 were synthesized through a multistep synthetic strategy (Schemes and S1). Both ligands were fully characterized by NMR spectroscopy and ESI-MS analyses (Figures S1–S12). Schematic Representation of the Synthesis of L 1 , L 2 The reaction of tetraimidazolium ligand L 2 with Ag2O in acetonitrile yielded the tetranuclear silver-octacarbene metallacage MC 2 . The product was isolated in 89% yield by precipitating the acetonitrile solution of the reaction mixture with diethyl ether. Similarly, the reaction of tetraimidazole ligand L 1 with AgSbF6 in a 2:1 mixture of acetonitrile and dimethyl sulfoxide (DMSO) afforded the tetranuclear metallacage MC 1 , which was isolated in 90% yield by precipitation with diethyl ether. For MC 2 , the disappearance of the N–CH–N proton resonance at 8.56 ppm in the 1H NMR spectrum (Figure d) indicated silver­(I)-carbene bond formation. Additional evidence was provided by an upfield shift of two imidazolium protons, which appeared merged at 7.0 and 6.70 ppm, compared to their positions in L 2 . The 13C NMR spectrum revealed a characteristic peak at 181.28 ppm corresponding to the N–C–N carbene carbon (Figure S21). For MC 1 , 1H NMR spectrum showed a downfield shift of the imidazolium proton to 8.16 ppm, which is expected due to silver­(I)–nitrogen bond formation (Figure b). Peak assignments were validated using 1H–1H COSY correlations (Figure S14). Moreover, single diffusion bands with log D values at −9.33 and −10.36 in the 1H DOSY NMR (Figures S22 and S16) of MC 2 and MC 1 , respectively, are in agreement with the formation of a single product in both the cases. (a, b) 1H NMR spectra of L1 and MC1 (DMSO-d 6, 298 K). (c, d) 1H NMR spectra of L2 and MC2 (CD3CN, 298 K). While these NMR results validated the silver­(I)-carbene coordination, the precise assembly structure was confirmed by ESI-MS of MC 2 [(C 124 H 116 N 16 )­(Ag) 4 (PF 6 ) 4 ; MW = 2836.4339] (Figure S23). Isotopically resolved distribution patterns for [MC 2 -2PF6]2+, [MC 2 -3PF6]3+, and [MC 2 -4PF6]4+ fragments confirmed the formation of a tetranuclear silver­(I)-octacarbene complex (Figure a, c). ESI-MS analysis was employed to confirm the composition of MC 1 [(C 100 H 84 N 16 )­(Ag) 4 (SbF 6 ) 4 ; MW = 2875.9038] (Figure S17). Well-resolved isotopic patterns for [MC 1 -2SbF6]2+ and [MC 1 -3SbF6]3+ fragments established the formation of a tetranuclear silver­(I)–nitrogen complex (Figure e, g). Isotopic distribution patterns: (a) experimental (black) and (b) theoretical (red) patterns of the [MC2-4PF6 ] 4+ fragment; (c) experimental (black) and (d) theoretical (red) patterns of the [MC2-3PF6 ] 3+ fragment; (e) experimental (black) and (f) theoretical (red) patterns of the [MC1-3SbF6 ] 3+ fragment; (g) experimental (black) and (h) theoretical (red) patterns of the [MC1-2SbF6 ] 2+ fragment. Computational Method Despite extensive efforts to grow diffraction-quality crystals of MC 1 and MC 2 using various solvent systems, no suitable crystal for X-ray diffraction was obtained. To elucidate their structures, geometry optimization was carried out using density functional theory (DFT) (Figure ). The calculations employed B3LYP functional for electronic structure optimization in the ground state, with the LANL2DZ basis set applied to all atoms (C, H, N, and Ag). This approach allowed us to predict the structural configurations and understand the differences in the photophysical behavior of MC 1 and MC 2 . (a, b) DFT optimized structures of (a) MC1 and (b) MC2 . Color codes: carbon (gray), nitrogen (blue), silver (yellow). H atoms are omitted for clarity. Frontier molecular orbitals of MC1 (left) and MC2 (right). For MC 2 , the silver­(I) ions exhibit a rigid linear coordination geometry due to the Ag–C bonds formed with N-heterocyclic carbene (NHC) donors, which are relatively strong, highly directional, and possess significant covalent character because of the strong σ-donating ability of the carbene ligands. Such robust Ag–C coordination restricts ligand rotation and suppresses structural rearrangement, thereby generating a comparatively rigid cage framework. Each silver­(I) ion is symmetrically coordinated by two N-heterocyclic carbene (NHC) donors from separate L 2 tetracarbene ligands. The optimized structure shows a highly symmetric arrangement, with dihedral angles between the double-bond plane and phenyl rings as 92.44° and 26.93° respectively, which indicate that two phenyl rings on opposing sides oriented almost perpendicular to the ligand plane. This is attributed to the restricted rotational freedom of the phenyl rings that contributes to fluorescence enhancement in solution and quenching in the aggregate state, as rigidity hampers the dissipation of energy in a luminescent pathway. In contrast, for MC 1 , the silver­(I) ions adopt a more flexible linear coordination geometry owing to more dynamic Ag–N coordination environment, which permits greater rotational freedom and reversible bond reorganization in solution, leading to enhanced conformational adaptability and a more flexible cage architecture. Each silver­(I) ion is linearly coordinated between two nitrogen donors from separate L 1 ligands. Here, dihedral angles between the double-bond plane and phenyl rings are 45.62° and 36.22°, respectively. This flexibility enables the retention of aggregation-induced emission (AIE) in the aggregate state, resulting in enhanced fluorescence due to efficient energy dissipation via luminescent pathways. This computational study highlights the pivotal role of coordination geometry and ligand dynamics in determining the photophysical properties of silver-based metallacages. The frontier molecular orbital analysis further supports these observations. The electron density associated with both the HOMO and LUMO is predominantly localized on the ligand backbone, with negligible contribution from the silver­(I) centers. This ligand-centered electronic structure is particularly significant from a photophysical standpoint. In many reported palladium­(II)-or platinum­(II)-based coordination cages, substantial metal participation in the frontier orbitals enables metal-to-ligand or ligand-to-metal charge transfer transitions. − Such interactions often enhance spin–orbit coupling and facilitate nonradiative decay processes, leading to pronounced fluorescence quenching due to heavy-atom effects. In contrast, the minimal involvement of Ag­(I) centers in the frontier orbitals preserves the intrinsic ligand-centered excited states. Consequently, coordination to silver does not significantly quench emission, allowing the complexes to retain their luminescent properties. This electronic decoupling between the metal centers and the emissive ligand framework represents a key advantage of silver-based metallacages over analogous Pd­(II) or Pt­(II) systems, particularly for the development of luminescent supramolecular architectures. Photophysical Properties After acquiring the important structural information, photophysical studies of the metallacages MC 1 and MC 2 were performed both in solution and in the aggregate state. DMF solutions of L 2 and MC 2 were excited at 359 nm. While L 2 was less emissive, the formation of the Ag­(I)-coordinated metallacage MC 2 led to a 5-fold enhancement in emission (Figure b), with the fluorescence quantum yield increasing from 5% to 41%. This significant enhancement is attributed to coordination-induced emission (CIE), wherein the coordination with Ag­(I) rigidifies the structure and restricts intramolecular motion of the phenyl rings (Figure b). The resulting restriction in intramolecular motion (RIM) suppresses nonradiative decay pathways, favoring radiative energy dissipation. However, aggregation of MC 2 in the water-DMF mixture led to emission quenching (Figure c). Although the methylene-based imidazole linker is flexible, but formation of the cage renders the structure highly rigid. This rigidity results in lower emission in the aggregated state. In the case of MC 1 , Ag­(I)-coordination resulted in a modest 2-fold enhancement in emission due to its flexible coordination, which permits phenyl ring rotation, limiting the rigidity-induced emission unlike in MC 2 , where Ag-carbene bonds are more rigid leading to higher coordination induced emission. Fluorescence quantum yield was increased from 4% to 9% upon coordination in MC 1 . Interestingly, unlike MC 2 (which showed ACQ), with the increase of water fraction in the DMF solution of MC 1 aggregation-induced emission (AIE) was very prominent. This enhancement is attributed to increased restriction of phenyl ring rotation in the aggregate state of MC 1 , as confirmed by DFT-optimized structures (Figure a). In a 90% water/DMF mixture, MC 1 exhibited a 5.6-fold enhancement in emission intensity compared to its dilute solution and fluorescence quantum yield enhanced from 9% to 72% (Figure d). The morphology of the MC 1 supramolecular aggregate was examined by TEM and SEM, which suggested the formation of nanospheres (Figures S39–S40). MC 1 in 90% water/DMF mixture, which provided maximum emission intensity, was used for energy transfer studies. (a) UV–vis absorption spectra of L1 , MC1 , L2 and MC2 in DMF ([MC1], [MC2] = 10–5 M; [L1], [L2] = 2 × 10–5 M). (b) Fluorescence emission spectra of L2 and MC2 in DMF. (c) Fluorescence emission spectra of MC2 in DMF with increasing water fractions. (d) Fluorescence emission spectra of L1 in DMF (black line), MC1 in DMF (red line) and MC1 in a 9:1 (v/v) H2O/DMF mixture (blue spectrum) ([MC1], [MC2] = 10–5 M; [L1], [L2] = 2 × 10–5 M). Notably, the emission spectrum of the MC 1 assembly overlapped significantly with the absorption spectrum of ESY (Figure S41), which indicates its potential as an efficient donor in light-harvesting systems (LHS). To further investigate energy transfer, ESY was selected as the first acceptor. Upon addition of ESY to the MC 1 assembly, the emission intensity of ESY at 551 nm increased progressively, accompanied by a decrease in the emission intensity of MC 1 at 477 nm when excited at 357 nm (Figure a). Control experiments confirmed negligible emission of ESY in the absence of MC 1 (Figure S43). The antenna effect (AE) in the MC 1 -ESY system was determined to be 62.62. A color change from blue (MC 1 ) to yellowish green (MC 1 -ESY) was observed, demonstrating efficient energy transfer from MC 1 to ESY. The energy transfer efficiency (ΦET) in the MC 1 -ESY system reached 51.07% at a donor-to-acceptor ratio of 2000:20 in 90% water/DMF. To better understand the energy transfer phenomenon, fluorescence lifetime analysis was performed on donor compound MC 1 in the presence of ESY (Figure S45). Fluorescence lifetime was found to be 1.3 ns for donor compound MC 1 , which decreased to 0.8 ns upon the addition of ESY. The decrease in lifetime indicates an effective transfer of energy from MC 1 to ESY. Inspired by the sequential energy transfer (SET) observed in natural photosynthesis, a SET system was constructed using MC 1 -ESY. (a) Fluorescence spectra of MC1 with gradual addition of ESY (λex = 357 nm) in a 90% H2O/DMF mixture (DEX is direct excitation of ESY at 520 nm) (Inset: color change from cyan blue to yellowish green). (b) Fluorescence spectra of MC1 -ESY with gradual addition of SR101 (λex = 357 nm) in a 90% H2O/DMF mixture (DEX is direct excitation of SR101 at 588 nm) (Inset: color change from yellowish green to red). (c) Fluorescence spectra of MC1 -ESY-SR101 with gradual addition of Cy5 (λex = 357 nm) in a 90% H2O/DMF mixture (DEX is direct excitation of Cy5 at 648 nm) (Inset: color change from red to warm white). (d) 1931 CIE chromaticity coordinate of (MC1 -ESY-SR101-Cy5) = (2000:20:20:5) ratio in 90% H2O/DMF mixture [10–5 M solutions of each component were mixed]. SR101 was selected as the second acceptor due to its absorption spectrum overlapping with the emission spectrum of the MC 1 -ESY assembly (Figure S46). Upon adding SR101 to the MC 1 -ESY system, the emission band at 606 nm, attributed to SR101, increased, while the emission peak of MC 1 -ESY at 551 nm decreased (Figure b). This was accompanied by a visible color change from yellow (MC 1 -ESY) to light red (MC 1 -ESY-SR101). The second-step energy transfer efficiency and antenna effect in the MC 1 -ESY-SR101 system were determined to be 67.28% and 30.55, respectively, at a donor-to-acceptor ratio of 2000:20:20 (Figures S47–S48). Again, fluorescence lifetime measurements were carried out for donor MC 1 -ESY in the presence of SR101 (Figure S50). Fluorescence decay revealed that in the presence of SR101, the average lifetime of the MC 1 emissions at 477 nm decreased from τav = 0.8 ns to τav = 0.6 ns and the average lifetime of the ESY emissions at 551 nm decreased from τav = 1.6 ns to τav = 1.2 ns, which suggests an efficient energy transfer from MC 1 -ESY to SR101. To construct a three-step sequential light-harvesting system, Cy5 was introduced as a third FRET acceptor. Cy5 was chosen because its absorption spectrum overlaps with the emission spectrum of the MC 1 -ESY-SR101 assembly (Figure S51). Upon the addition of Cy5, the emission band at 673 nm, attributed to Cy5, increased, while the SR101 emission peak at 606 nm decreased (Figure c). At a donor-to-acceptor ratio of 2000:20:20:20, the overall energy transfer efficiency and antenna effect were calculated to be 71.18% and 8.95, respectively (Figures S52–S54). Lifetime measurements again confirmed the sequential energy transfer. In the presence of Cy5, the average lifetime of MC 1 emission at 477 nm decreased from τav = 0.6 ns to τav = 0.5 ns, ESY emission at 551 nm decreased from τav = 1.21 ns to τav = 0.75 ns, and SR101 emission at 606 nm decreased from τav = 3.6 ns to τav = 1.2 ns (Figure S55). These results demonstrate efficient stepwise energy transfer from the MC 1 -ESY-SR101 assembly to Cy5, confirming the successful fabrication of a three-step FRET-based light-harvesting system. Notably, during this three-step sequential energy transfer process, strong white light emission was achieved at a molar ratio of [MC 1 ]:[ESY]:[SR101]:[Cy5] = 2000:20:20:5 (Figure c) with the fluorescence quantum yield of 93% (Figure S58). The emission color coordinate of the MC 1 -ESY-SR101-Cy5 system was calculated to be (0.318, 0.327), which is very close to the ideal white point (0.33, 0.33) (Figure d). This result highlights the potential application of the present light-harvesting assembly in the development of white-light emitting materials. For control experiment, selective excitation at 520 nm (corresponding to the absorption maximum of ESY; Figure S53b) resulted in negligible emission from the ESY-SR101-Cy5 system. This observation confirms that energy transfer within cage MC 1 proceeds in a strictly stepwise manner rather than through direct excitation pathways. The donor–acceptor ratio was optimized to achieve efficient sequential energy transfer while minimizing concentration-dependent quenching. In each case, the emission intensity increased with increasing acceptor concentration up to 20 μL. Further addition of the acceptor dye beyond this point led to a decrease in emission intensity due to self-quenching and increased nonradiative decay pathways caused by excessive dye loading. Accordingly, a donor–acceptor ratio of 2000:20:20:20 was selected as the optimal condition for efficient light-harvesting and sequential energy-transfer studies. We wanted to explore if the ligand L 1 alone can show similar light-harvesting behavior and its efficiency compared to that of MC 1 . Ligand L 1 exhibits aggregation-induced emission behavior in a water/DMF mixture. In its aggregated state (90% water in DMF), the emission spectrum of L 1 shows substantial overlap with the absorption band of ESY, suggesting the possibility of energy transfer from aggregated L 1 to ESY (Figure S56). At a molar ratio of 2000:20 for [L 1 ]:[ESY], excitation of the system at 358 nm resulted in a decrease in the emission intensity of L 1 at 482 nm. In contrast, the emission intensity of ESY increased only slightly (λex = 550 nm). This weak enhancement of ESY emission suggests that energy transfer from aggregated L 1 to ESY is negligible. We further examined the possibility of sequential light-harvesting using L 1 and ESY together in the presence of SR101. Under these conditions, no emission peak corresponding to SR101 was observed after excitation at 358 nm. This result indicates that sequential light-harvesting does not occur efficiently in case of free ligand L 1 (Figure S57). Overall, these observations demonstrate that MC 1 serves as a significantly more effective donor platform than ligand L 1 for constructing multistep FRET-based light-harvesting systems. Finally, the stability of the metallacage MC 1 in the presence of Eosin Y was investigated to exclude any cage decomposition during the light-harvesting process. Accordingly, 1H NMR titration experiments were carried out by gradual addition of Eosin Y to a solution of MC 1 (Figure S59). Upon addition of the dye, continuous chemical-shift changes of the cage proton signals were observed, indicating host–guest interactions under fast exchange conditions. Importantly, the overall spectral pattern of MC 1 remained unchanged during the titration, confirming that the metallacage framework remains stable in the presence of Eosin Y. Nonlinear fitting of the titration data using the BindFit program supported a 1:1 host–guest binding model with an apparent association constant (K a) of 1.06 × 102 M–1 (Figure S59a). In contrast, control experiments using the free ligand (L 1 ) and Eosin Y showed no noticeable changes in the 1H NMR spectra, indicating negligible interactions without the cage cavity (Figure S59b). These results suggest that the confined cavity of MC 1 plays an important role in facilitating host–guest interactions between MC 1 and Eosin Y, which promotes effective dipole–dipole interactions and energy-transfer processes responsible for the observed white-light emission. Conclusions We have successfully demonstrated a sequential artificial light-harvesting system with a silver­(I)-based metallacage. By employing rational coordination-mode engineering, two structurally distinct cages were synthesized from two AIE-active ligands. One of them is a flexible cage (MC 1 ) with N–Ag–N coordination mode, and the other one is a rigid cage (MC 2 ) with C–Ag–C coordination bonds. Two cages are emissive in solution but differ significantly in their aggregation behavior. While MC 1 retains aggregation-induced emission and acts as a highly efficient multichromophoric donor, MC 2 exhibits aggregation-caused quenching and is therefore not suitable for light-harvesting purposes. The flexible architecture of MC 1 allows the effective three-step sequential energy transfer from cage to eosin Y, sulforhodamine 101, and Cyanine-5 dyes, resulting in near-ideal white-light emission with high overall energy-transfer efficiency. Comparative experiments using the free ligand demonstrated that metallacage construction is essential for a multistep energy transfer process in the present case. These results indicate that structural flexibility and coordination geometry are key parameters that govern the aggregation behavior and energy-transfer pathways in silver-based supramolecular assemblies. This work extends the concept of artificial light-harvesting systems to silver metallacages, providing a design strategy for developing advanced multistep energy transfer platforms and white-light-emitting materials. Furthermore, these findings open opportunities for the future development of new Ag­(I)-N-based coordination cages incorporating AIE-active ligands and confined cavities, where energy-transfer processes involving different donor–acceptor dye combinations can be explored for advanced artificial light-harvesting and supramolecular luminescent systems. Supplementary Material