Assembly of silver Trigons into a buckyball-like Ag₁₈₀ nanocage

Significance

Here we present a striking outcome from the alliance between chemistry and mathematics in the design, synthesis, and characterization of a silver cage, Ag₁₈₀. In principle, the design replaces each carbon atom of C₆₀ with a triplet of argentophilicity-bonded silver atoms to produce a 3.4.6.4 (1,1) polyhedron with sixty 3-gons, ninety 4-gons, twelve 5-gons, and twenty 6-gons. Results from mass spectroscopy suggest an assembly mechanism in solution based on such triplets––the Silver-Trigon Assembly Road (STAR). Indeed, the STAR mechanism may be a general synthetic pathway toward even larger silver polyhedral cages. Besides its fundamental appeal, this synthetic cage may be considered for use as a molecular luminescent thermometer.

Abstract

Buckminsterfullerene (C₆₀) represents a perfect combination of geometry and molecular structural chemistry. It has inspired many creative ideas for building fullerene-like nanopolyhedra. These include other fullerenes, virus capsids, polyhedra based on DNA, and synthetic polynuclear metal clusters and cages. Indeed, the regular organization of large numbers of metal atoms into one highly complex structure remains one of the foremost challenges in supramolecular chemistry. Here we describe the design, synthesis, and characterization of a Ag₁₈₀ nanocage with 180 Ag atoms as 4-valent vertices (V), 360 edges (E), and 182 faces (F)––sixty 3-gons, ninety 4-gons, twelve 5-gons, and twenty 6-gons––in agreement with Euler’s rule V − E + F = 2. If each 3-gon (or silver Trigon) were replaced with a carbon atom linked by edges along the 4-gons, the result would be like C₆₀, topologically a truncated icosahedron, an Archimedean solid with icosahedral (I_h) point-group symmetry. If C₆₀ can be described mathematically as a curling up of a 6.6.6 Platonic tiling, the Ag₁₈₀ cage can be described as a curling up of a 3.4.6.4 Archimedean tiling. High-resolution electrospray ionization mass spectrometry reveals that {Ag₃}_n subunits coexist with the Ag₁₈₀ species in the assembly system before the final crystallization of Ag₁₈₀, suggesting that the silver Trigon is the smallest building block in assembly of the final cage. Thus, we assign the underlying growth mechanism of Ag₁₈₀ to the Silver-Trigon Assembly Road (STAR), an assembly path that might be further employed to fabricate larger, elegant silver cages.

Symmetry is a consequence of the self-assembly of many beautiful molecules (1–8), with assembly of C₆₀ from a cooling carbon plasma (9) one of the best-known examples. In nature, living organisms take advantage of molecular self-assembly to construct many complicated macromolecules. For example, COPII protein assembles into two Archimedean solids, the cuboctahedron and the icosidodecahedron (10), ferritin forms a rhombic dodecahedron (11), and spherical viruses form icosahedral cages (12, 13). Inspired by C₆₀ and such biological macromolecules, scientists face the challenge of mimicking and synthesizing comparably impressive single molecules (14, 15) from large numbers of subcomponents. However, the synthesis of molecules as impressive as C₆₀ has been rare––but see As₂₀ (2), [{Cp*Fe(P₅)}₁₂{CuCl}₁₀{Cu₂Cl₃}₅{Cu(CH₃CN)₂}₅] (3), Mo₃₆₈ (16), Ti₄₂ (17), Cd₆₆ (18), Pd₁₄₅ (19), Al₅₆ (20), Ln₁₀₄ (21), [{Cp^BIGFe(P₅)}₁₂{CuBr}₉₂] (22), [{Cp*Fe(P₅)}₁₂{CuCl}₂₀] (23), [{Cp^BnFe(P₅)}₁₂{CuCl}₂₀] (24), Pd₃₀ (25), and Ag₃₇₄ (26), all Platonic solids, Archimedean solids, or combinations of them.

Nanosized silver clusters have attracted increasing research interest because of their fascinating structures and potential applications (27–29). Although silver clusters with up to 490 metal atoms in the core are known (30), these lack high symmetry. Silver atoms are prone to form polygons like 3- and 4-gons with the aid of an argentophilic interaction (31) and/or ligation of thiolates. Here we report fabrication of these polygonal building blocks common in Platonic and Archimedean solids into a highly symmetric cage.

Results and Discussion

In our effort to use thiolates as capping ligands to construct well-ordered silver clusters, we fabricated a nanocage with 180 silver atoms (Fig. 1 A and B and SI Appendix, Table S1) based on abundant argentophilic interactions (Fig. 1C and SI Appendix, Table S2). The {Ag₁₈₀}⁴⁶⁺ (hereafter, Ag₁₈₀ for short) cage has these characteristics: (i) With 180 metal atoms, it is a very large silver cage, not a cluster. (ii) It is a silver cage with icosahedral (I_h) topology. (iii) It comprises 3-, 4-, 5-, and 6-gons (Fig. 1D). (iv) It has high solution stability.

Fig. 1.

Molecular structure of the nanocage Ag₁₈₀. (A) Space-filling model of the structure of 1 (Ag₁₈₀) established by X-ray crystallography. H atoms are omitted for clarity. Atoms colored green, red, yellow, and gray correspond to Ag, O, S, and C. (B) Ball-and-stick model of the structure of 1 (Ag₁₈₀) corresponding to the space-filling model in A that omits H atoms. (C) Ball-and-stick model of only the Ag atoms (purple balls). The 180 Ag-Trigon (purple) bonds are within the length range for argentophilic interactions (SI Appendix, Table S2). The 180 (black) edges are longer. (D) Silver 3-, 4-, 5-, and 6-gonal faces in different colors on the Ag₁₈₀ nanocage.

Solvothermal reaction (65 °C) of [Ag(ⁱPrS)]_n (ⁱPr = isopropyl, C₃H₇) and CH₃SO₃Ag (silver methanesulfonate) in methanol (see SI Appendix for detailed methods) gives yellow, octahedron-shaped crystals of complex 1 (SI Appendix, Fig. S1). Complex 1 was fully characterized by high-resolution electrospray ionization mass spectroscopy (HR-ESI-MS), ¹H NMR on acid-digested solutions of crystals, thermogravimetric analysis, and single-crystal X-ray diffraction (Fig. 1 A and B and SI Appendix, Table S1), to assess solution behavior and assembly mechanism, organic ligand ratio, thermal stability, and its average structure.

Because the crystals are fragile and unstable on leaving the mother liquor, we collected X-ray diffraction data from a single crystal protected by Paraton oil at 100 K on a Bruker APEX II single-crystal diffractometer. X-ray crystallographic results reveal that complex 1 crystallizes in the highly symmetric cubic space group Fm-3 with a large unit-cell volume of ∼110,000 ų (SI Appendix, Table S1). The electron-density map provides a landscape of the overall connectivity of a giant cage with 180 Ag atoms (Fig. 1C), 90 ⁱPrS⁻, and 44 CH₃SO₃⁻, and its asymmetric unit contains only 1/24th of the cage (SI Appendix, Fig. S2).

The exact molar ratio of ⁱPrS⁻ and CH₃SO₃⁻ in 1 was further deduced from the ¹H NMR spectra of a DCl-digested solution of the crystals of 1 (SI Appendix, Fig. S3). Proton resonances with the predicted positions and coupling patterns are clearly detected in the expected regions. By integrating resonance peak intensities, we find that the ⁱPrS⁻ and CH₃SO₃⁻ are in the proportion 1:1.005, indicating that they are equimolar in 1. Thermogravimetric analysis of polycrystalline samples of 1 (SI Appendix, Fig. S4) show that the complex loses guest methanol (total percentage mass loss = 3.08%) from 30 °C to 69 °C and stabilizes to 143 °C before decomposition. Based on the above data, the reliable formula of 1 can be deduced to be {[Ag₁₈₀(ⁱPrS)₉₀(CH₃SO₃)₄₄]·(CH₃SO₃)₄₆·34CH₃OH} (also denoted as 1a·46CH₃SO₃·34CH₃OH).

This Ag₁₈₀ cage has 360 Ag···Ag edges (Fig. 1C). Of these, 180 are in the 3.00–3.14-Å range (SI Appendix, Table S2; numbering of Ag atoms in Fig. 2A and SI Appendix, Fig. S2) that is characteristic of genuine argentophilic interactions, and these bonds (purple edges in Fig. 1C) are found exclusively in the silver Trigons (purple-filled triangles in Fig. 1D). Another 180 Ag···Ag edge lengths, 3.19–3.67 Å (black edges in Fig. 1C) (SI Appendix, Table S2), represent very weak argentophilic interactions (32). (Due to four disordered silver atoms, the Ag···Ag distances related to these Ag atoms are discussed based on only the main orientation.) On the basis of these argentophilic interactions, the surface of the Ag₁₈₀ cage can be divided into four kinds of all-silver polygons: sixty 3-gons, ninety 4-gons, twelve 5-gons and twenty 6-gons (Fig. 1D).

Fig. 2.

STAR. (A) ORTEP plot for the crystal structure of 1a (Ag₁₈₀) with numbered crystallographically independent Ag atoms (green). C and H atoms are omitted for clarity. Atoms, bonds, and edges in front obscure their mirror counterparts in back. S atoms from exterior ⁱPrS⁻ (isopropyl sulfide –C₃H₇S⁻) ligands “above” 4-gons are colored yellow; S atoms from interior ⁱPrS⁻ ligands “below” 4-gons are colored blue. Exterior CH₃SO₃ ligands are shown as SO₃ groups above 5-gons and 6-gons with O atoms in red and S atoms in yellow. Interior CH₃SO₃ ligands are shown as SO₃ groups below 5-gons with O atoms in red and S atoms in black. (B) Ultimately, the cage is assembled from silver Trigons, the colored triangles. We suppose that five Trigons assemble into a pentaTrigon, {Ag₃}₅, that includes five Trigons, pairs of argentophilic bonds that create five linking 4-gons, each with its coordinating ⁱPrS⁻ ligand on the outside (thus marked by a yellow disk), and a 5-gon. As shown in these Schlegel (plane) diagrams of Ag₁₈₀, pentaTrigons could further link to each other with pairs of bonds that create a 4-gon with its ⁱPrS⁻ ligand on the inside (thus marked by a blue disk). (C) Assembly by decaTrigons {Ag₃}₁₀, each composed of two linked pentaTrigons.

All ⁱPrS⁻ ligands––appearing as just (yellow and blue) S atoms in Fig. 2A––were ligated on the 4-gons in μ₄ fashion with Ag–S distances of 2.29–2.54 Å (SI Appendix, Table S2). Among the 90 ⁱPrS⁻ ligands, 60 coordinate from positions outside the Ag₁₈₀ cage. These are marked in Fig. 2A by S atoms colored yellow above all of the five 4-gons surrounding each of the twelve 5-gons. The remaining 30 coordinate from positions inside the Ag₁₈₀ cage and are marked by S atoms colored blue. Each of the twenty 6-gons share three 4-gons with another 6-gon, and it is these 30 shared 4-gons (20 × 3/2 = 30) that have the interior ⁱPrS⁻ ligands marked by blue S atoms. Thus, the 4-gons around each 6-gon are capped by alternating exterior (yellow) and interior (blue) ⁱPrS⁻ ligands.

In addition, there are 44 CH₃SO₃⁻ ligands, shown as just SO₃ groups in Fig. 2A. Each of the twenty 6-gons is capped by one CH₃SO₃⁻ ligand on the outside of the cage, marked by a yellow S atom and three (red) O atoms. Each of the twelve 5-gons is capped by two CH₃SO₃⁻ ligands, one outside the cage, also marked by a yellow S atom and three (red) O atoms, and one inside, marked by a black S atom and three (red) O atoms. Due to the large steric hindrance from the ⁱPrS⁻ ligands capping the surrounding 4-gons, no CH₃SO₃⁻ ligands cap the silver 3-gons.

Due to the coordination interactions between Ag atoms and O and S donors, eight crystallographically unique Ag atoms can be divided into two groups without the consideration of Ag···Ag interactions. Based on interatomic distances (SI Appendix, Table S2) and as shown in Fig. 2A and SI Appendix, Fig. S2, Ag2–Ag8 coordinate with two sulfur atoms and one oxygen atom, whereas Ag1 coordinates with two sulfur atoms and two oxygen atoms.

The Ag₁₈₀ has a spherical shape with a bare silver cage diameter of 2.5 nm and an overall diameter (including the ligand shell) of 3.0 nm, whereas the diameter of the inner void is 1.9 nm, thus an accessible volume of the inner cavity calculated as 3,769 ų. Of the total unit-cell volume, 43.7% is occupied by disordered methanol molecules and CH₃SO₃⁻ that are not imaged by X-ray crystallography but are revealed by the following HR-ESI-MS experiments.

The HR-ESI-MS of crystals of complex 1 dissolved in methanol (Fig. 3A) shows three sequentially charged parent ion species centered at m/z 5,781.11, 6,960.63, and 8,713.51 that we assign to {1a·40CH₃SO₃·15CH₃OH·2H₂O}⁶⁺ ({Ag₃}₆₀⁶⁺, m/z Calc. = 5,780.84), {1a·41CH₃SO₃·14CH₃OH·5H₂O}⁵⁺ ({Ag₃}₆₀⁵⁺, Calc. m/z = 6,960.41), and {1a·42CH₃SO₃·15CH₃OH·H₂O}⁴⁺ ({Ag₃}₆₀⁴⁺, Calc. m/z = 8,714.25), respectively, indicating that the large Ag₁₈₀ cage retains its structural integrity in methanol.

Fig. 3.

HR-ESI-MS. (A) Positive ion mode HR-ESI-MS of the crystals of 1 dissolved in methanol. (B) Positive ion mode HR-ESI-MS of the reaction mixture after solvothermal reaction. Molecular species with the same number of silver atoms are distinguished by Latin characters α, β, γ, δ, etc.

We attribute three dominant fragment peaks below m/z = 3,000 to {[Ag₃₀K(ⁱPrS)₁₃(CH₃SO₃)₁₂(OH)₃]}³⁺ ([{Ag₃}₁₀-α]³⁺, Exp. m/z = 1,814.70; Calc. m/z = 1,814.74), {[Ag₃₀K(ⁱPrS)₁₃(CH₃SO₃)₁₃(OH)₃]}²⁺ ([{Ag₃}₁₀-β]²⁺, Exp. m/z = 2,769.54; Calc. m/z = 2,769.60), and {[Ag₃₀(ⁱPrS)₈(CH₃SO₃)₁₉Cl(CH₃OH)₂]}²⁺ ([{Ag₃}₁₀-γ]²⁺, Exp. m/z = 2,871.48; Calc. m/z = 2,871.50). We assign detailed molecular formulae for these species based on the experimental and simulated isotopic distributions (SI Appendix, Fig. S5 A–C). We found no other fragments, suggesting a coordination–disassociation equilibrium predominantly between Ag₃₀ and Ag₁₈₀ species in solution, which is further evidenced by successful recrystallization of 1 in methanol.

In addition, the X-ray structure shows argentophilic interactions that are distributed only in silver Trigons, suggesting an assembly mechanism based on the silver Trigons. In further support of this mechanism, the HR-ESI-MS of the solution after solvothermal reaction (Fig. 3B) reveals a series of oligomeric silver Trigons species in the m/z = 500–3,000 range with different charge states. The most dominant peak is {Ag₃}₁¹⁺. The others correspond to integer multiples of the Ag₃ Trigon, including {Ag₃}₂¹⁺, {Ag₃}₄²⁺, {Ag₃}₅²⁺, {Ag₃}₆²⁺, {Ag₃}₇²⁺, {Ag₃}₈²⁺, and [{Ag₃}₁₀-δ]²⁺. We assign detailed molecular formulae for these species as well based on the experimental and simulated isotopic distributions (SI Appendix, Fig. S5 D–K). The HR-ESI-MS of the solution after solvothermal reaction (Fig. 3B) also shows three Ag₁₈₀ species with diminishing positive charges, [{Ag₃}₆₀-α]⁶⁺, [{Ag₃}₆₀-β]⁵⁺, and [{Ag₃}₆₀-γ]⁴⁺, almost the same to those observed in ESI-MS of dissolved crystals of 1 in methanol but with different numbers of solvent molecules.

The rich {Ag₃}_n fragments and some nascent {Ag₃}₆₀ species in the assembly process before the final crystallization of the Ag₁₈₀ cage suggest that the {Ag₃} building block plays a special role in formation of the Ag₁₈₀ cage. We thus assign the underlying growth mechanism of Ag₁₈₀ to the Silver-Trigon Assembly Road (STAR), in which silver Trigons (purple triangles in the Schlegel diagram of the Ag₁₈₀ cage shown in Fig. 2B) act as the smallest building block that is gradually aggregated by bridging ⁱPrS⁻ and CH₃SO₃⁻ to form larger {Ag₃}_n species.

We further suppose that “pentaTrigons” are assembled from five (purple) Trigons linked to each other by ⁱPrS⁻ ligands to form 4-gons (Fig. 2B). These 4-gons, marked by yellow disks as in Fig. 2A, have ⁱPrS⁻ ligands that will end up on the “outside” of the cage (Fig. 1 A and B). We then suppose that the Ag₁₈₀ cage assembled itself from 12 pentaTrigons linked to each other by ⁱPrS⁻ ligands to form additional 4-gons. These latter 4-gons have ⁱPrS⁻ ligands that end up on the “inside” of the cage and are thus marked by blue disks in Fig. 2B (as in Fig. 2A). An interesting feature of this path is that the formation of pentaTrigons and their attachments to make Ag₁₈₀ appear not to require independent generation of hexagons and their surrounding 4-gons. However, the prominence of {Ag₃}₁₀ in cages dissolved in methanol suggests a path that relies on initial assembly of pentaTrigons into decaTrigons {Ag₃}₁₀, as marked by adjacent pairs of pentaTrigons with Trigons of the same color in Fig. 2C) and then assembly of six of the latter into the Ag₁₈₀ cage.

In the Ag₁₈₀ cage, Ag···Ag edge lengths in 3-gons, 4-gons, and 5-gons are not exactly equal; 5-gons are not equiangular, and 6-gons centered on the presumed threefold axis are also not equiangular (SI Appendix, Table S2). For these reasons, although the Ag₁₈₀ cage has mirror planes––as can be appreciated from Figs. 1 and 2A, where the “front” half of the cage is in perfect register with and thus occludes the “back” half of the cage––it does not have all of the symmetries of a geometrically icosahedral structure. Nonetheless, the overall topology of the Ag₁₈₀ cage is icosahedral. Likewise, C₆₀ has two different bond lengths (33), longer in the five bonds within pentagons and shorter in the three remaining bonds within 6-gons, so it resembles but is not identical to the truncated icosahedron.

The similarities go deeper. Replacing each atom of the truncated icosahedron (Fig. 4A, heavier lines) with a triangle (representing a silver Trigon) produces an icosahedral cage with three times as many (3 × 60 = 180) vertices, sixty 3-gons, ninety 4-gons, twelve 5-gons, and twenty 6-gons, a cage with the same (icosahedral) topology as Ag₁₈₀ (Fig. 4A, lighter lines; Fig. 4B). Indeed, both the truncated icosahedron (or C₆₀) and the Ag₁₈₀ cage are the (1,1) members of a series of icosahedral cages, where the (1,1) Goldberg indices specify the triangle that is glued to each of the 20 triangles of an icosahedron to produce the final icosahedral cage (12, 34–36) (Fig. 4 C and D; see “Goldberg indices and Goldberg cages” in SI Appendix).

Fig. 4.

Relationship between cages like Ag₁₈₀ and icosahedral fullerenes. (A) The relationship between the C₆₀ buckyball (thick lines) and the Ag₁₈₀ cage (thin lines). Replacing each silver 3-gon by a carbon atom and each silver 4-gon by an edge produces the same topology as C₆₀, the truncated icosahedron. (B) All of the 4-gons surrounding a 5-gon are marked with small yellow disks, marking coordination with an ⁱPrS⁻ ligand on the outside of the cage. Each triangle (or Trigon in Fig. 3A) is surrounded by two 4-gons marked by these small yellow disks and one 4-gon marked by a small blue disk, the latter indicating coordination with an ⁱPrS⁻ ligand on the inside of the cage. (C) Equilateral triangular cutouts with various (i, j) indices and T numbers drawn over 6.6.6 tilings. (D) Equilateral triangles with various (i, j) indices and T numbers (T = i² + ij + j²) drawn over 3.4.6.4 tilings. The letter P in the 6-gons at the corners of the Goldberg triangle are a reminder that those 6-gons in the tiling become Pentagons in the cage. The large yellow disks mark special hexagons that are surrounded by only yellow 4-gons. These are seen only in larger 3.4.6.4 cages with T divisible by 3, like T = 9 and 12. The red disks mark 4-gons that the rules require have their ⁱPrS⁻ ligand on both the inside and the outside, thus a forbidden configuration.

As noted, in the Ag₁₈₀ cage all of the 4-gons surrounding a 5-gon coordinate with an ⁱPrS⁻ ligand on the outside of the cage, so those 4-gons are marked by yellow disks (Fig. 4B). By contrast, of the six 4-gons surrounding a 6-gon, yellow alternates with blue: Three of the 4-gons are shared with a 5-gon and are already marked by yellow disks, whereas the other three are shared between 6-gons, coordinate with an ⁱPrS⁻ ligand on the inside of the cage, and are marked by blue disks. Thus, there appear to be two rules for assembly of Ag₁₈₀. (i) All of the 4-gons surrounding a 5-gon are marked by yellow disks; (ii) To minimize steric hindrance, each triangle is surrounded by two 4-gons marked by yellow disks and one 4-gon marked by a blue disk. Therefore, for the triangles surrounding the 5-gon, already surrounded by two “yellow 4-gons,” the third 4-gon must be blue.

Examination of Fig. 4D suggests that only 3.4.6.4 cages with T (= i² + ij + j²) divisible by 3, e.g., T = 3, 9, and 12 [(i, j) = (1,1), (3,0), and (2,2)], obey both rules. By contrast, for T = 1 [(i,j) = (1,0)], all three 4-gons surrounding a 3-gon are yellow (outer ⁱPrS⁻ ligand). For T = 4 [(i,j) = (2,0)], the first rule causes some 3-gons to be surrounded by one yellow 4-gon and two blue 4-gons (inner ⁱPrS⁻ ligand). For T = 7 [(i,j) = (2,1)], the first rule requires some 4-gons (marked by red) to be impossibly both yellow (outside) and blue (inside). For T = 9 and 12, the second rule requires that some special hexagons––those in the (1,1) position with respect to a pentagon or another special hexagon––be surrounded by all yellow 4-gons. There is no rule against that arrangement, and these cages pass, as would all cages with T divisible by 3. However, these larger cages could not be built exclusively from pentaTrigons. Instead, they could be built from a combination of pentaTrigons and hexaTrigons (SI Appendix, Fig. S6).

However, we have not seen these larger cages with 9 × 60 = 540 or 12 × 60 = 720 silver atoms. We supposed that steric hindrance was responsible, currently set “just right” by the ⁱPrS⁻ for Ag₁₈₀. However, when we replaced the propyl group in the ⁱPrS⁻ ligand with an ethyl group, we still produced Ag₁₈₀, as indicated by the similar unit-cell parameters. Therefore, we suggest that STAR with pentaTrigons and decaTrigons may be the dominant factor in determining the outcome of the synthesis.

Although assembly of Ag₁₈₀ relies on aggregation of Trigons, breakdown appears to follow a different path. The crystal of 1 can dissolve in water, but except for very small fragments like Ag₆ (SI Appendix, Fig. S5L), Ag₁₈₀-related molecular ion peaks are not detected. Dissolution of 1 into other solvents shows that the Ag₁₈₀ cage can be stabilized in many other solvents but with different disassociation degrees. These solvents include n-propanol, ethanol, acetone, and dichloromethane (SI Appendix, Fig. S7 A–D), but these solutions also show increasing amounts of Ag₁₃₀. Indeed, for 1 dissolved in the aprotic solvents dimethylformamide and acetonitrile (SI Appendix, Fig. S7 E and F), the fragment peaks are solely Ag₁₃₀ species with different charge states. The number 130 is not divisible by three, so the structure of Ag₁₃₀ cannot contain silver Trigons exclusively. Of course, it is possible to imagine Ag₁₃₀ structures with a spherical shape or conversely with a bowl shape. Further work would be required to elucidate its actual structure.

The solid-state diffuse reflectance UV/Vis spectra of a crystalline sample of 1 exhibits a main absorption band centered at 346 nm and a shoulder peak at 423 nm tailed to 700 nm (SI Appendix, Fig. S8). The UV (346 nm) absorption peaks can be attributed to the n → π* transition of ⁱPrS⁻, and the visible region (423 nm with its tail) can be attributed to the charge-transfer transition from the S 3p to Ag 5s orbitals.

We investigated the solid-state luminescence of 1, which reveals weak near-infrared emission with a maximum λ_em = 723 nm for excitation at 365 nm at room temperature (Fig. 5A). The emission of 1 may be due to the ligand-to-metal charge transfer, with charge transfer from S 3p to Ag 5s orbitals, a transition perturbed by Ag···Ag interactions^37. For checking possible thermochromic luminescence of 1, we collected its emission spectra from room temperature down to 93 K. As 1 is cooled from 293 to 93 K, the maximum emission wavelength shifted from 723 to 693 nm (Δ_em = 30 nm), and the intensity of luminescence increased nearly 10-fold (Fig. 5 A and B). The emission color evolution of a selected crystal at the different temperatures is also shown in Fig. 5C. The increased intensity of emission upon cooling may be caused by reduction of the nonradiative decay at low temperature, whereas the 30-nm hypsochromic-shifted emission may be ascribed to the significantly restricted swing of ⁱPr groups at low temperature (37). Moreover, 1 shows a linear correlation between maximum emission intensity (I_max) and temperature (T) in the range of 93–193 K. The excellent linearity between I_max and T could provide a calibration curve for a molecular luminescent thermometer at low temperature.

Fig. 5.

Emission spectroscopic properties of 1. (A) Luminescence spectra as a function of temperature from 93 to 293 K in the solid state for excitation at 365 nm. (B) Variation of maximum emission intensity (black circles; the red solid line is a linear fit in the range of 93–193 K) and peak emission wavelength (blue diamonds) from 93 to 293 K. (C) Photographs of a selected crystal of 1 irradiated with 365-nm UV light at the temperatures indicated.

For comparison, we also measured luminescent spectra for the [Ag(ⁱPrS)]_n precursor from 293 to 93 K (λ_ex = 365 nm; SI Appendix, Fig. S9A). The data reveal a maximum emission peak at 545 nm at room temperature but no obvious shift in peak wavelength with cooling to 93 K (SI Appendix, Fig. S9B). The data also reveal a sixfold enhancement of emission intensity from 293 to 93 K. The smaller shift in peak emission wavelength (Δ_em < 5 nm) and the smaller intensity enhancement makes [Ag(ⁱPrS)]_n less useful than 1 as a thermochromic luminescent material. Moreover, the linearity between I_max and T of [Ag(ⁱPrS)]_n is acceptable only from 173 to 293 K (SI Appendix, Fig. S9B), which makes [Ag(ⁱPrS)]_n less attractive as a potential molecular luminescent thermometer in the lower temperature range.

We have presented the synthesis and characterization of a remarkable silver cage, Ag₁₈₀. This achievement may be regarded as the product of an alliance between chemistry and mathematics, which suggested that we paste the argentophilicity-bonded silver Trigons in the arrangement of the vertices of the 6.6.6 (1,1) Archimedean polyhedron (with twelve 5-gons and twenty 6-gons) to form a 3.4.6.4 (1,1) polyhedron (with sixty 3-gons, ninety 4-gons, twelve 5-gons, and twenty 6-gons). Moreover, the cage’s structural features, combined with the mass spectroscopy analysis, suggest the growth mechanism––the STAR––in solution for this nanocage. Indeed, the STAR mechanism may be a general synthetic pathway toward larger silver polyhedral cages like Ag₅₄₀ and Ag₇₂₀. Besides the fundamental interest of this synthetic cage, we foresee applications including its use as a molecular luminescent thermometer.

Materials and Methods

We prepared the precursors of (AgⁱPrS)_n according to the literature but used ⁱPrSH instead (38). All of the chemicals and solvents we used in the syntheses were of analytical grade, and we used them without further purification. See SI Appendix for further detailed methods.