Structured copper-hydride nanoclusters provide insight into the surface-vacancy-defect to non-defect structural evolution

Abstract

Exploring the structural evolution of clusters with similar sizes and atom numbers induced by the removal or addition of a few atoms contributes to a deep understanding of structure–property relationships. Herein, three well-characterized copper-hydride nanoclusters that provide insight into the surface-vacancy-defect to non-defect structural evolution were reported. A surface-defective copper hydride nanocluster [Cu₂₈(S-c-C₆H₁₁)₁₈(PPh₂Py)₃H₈]²⁺ (Cu₂₈-PPh₂Py for short) with only one C₁ symmetry axis was synthesized using a one-pot method under mild conditions, and its structure was determined. Through ligand regulation, a 29^th copper atom was inserted into the surface vacancy site to give two non-defective copper hydride nanoclusters, namely [Cu₂₉(SAdm)₁₅Cl₃(P(Ph-Cl)₃)₄H₁₀]⁺ (Cu₂₉-P(Ph-Cl)₃ for short) with one C₃ symmetry axis and (Cu₂₉(S-c-C₆H₁₁)₁₈(P(Ph-^pMe)₃)₄H₁₀)⁺ (Cu₂₉-P(Ph-Me)₃ for short) with four C₃ symmetry axes. The optimized structures show that the 10 hydrides cap four triangular and all six square-planar structures of the cuboctahedral Cu₁₃ core of Cu₂₉-P(Ph-Me)₃, while the 10 hydrides cap four triangular and six square-planar structures of the anti-cuboctahedral Cu₁₃ core of Cu₂₉-P(Ph-Cl)₃, with the eight hydrides in Cu₂₈-PPh₂Py capping four triangular and four square planar-structures of its anti-cuboctahedral Cu₁₃ core. Cluster stability was found to increase sequentially from Cu₂₈-PPh₂Py to Cu₂₉-P(Ph-Cl)₃ and then to Cu₂₉-P(Ph-Me)₃, which indicates that stability is affected by the overall structure of the cluster. Structural adjustments to the metal core, shell, and core–shell bonding model, in moving from Cu₂₈-PPh₂Py to Cu₂₉-P(Ph-Cl)₃ and then to Cu₂₉-P(Ph-Me)₃, enable the structural evolution and mechanism responsible for their physicochemical properties to be understood and provide valuable insight into the structures of surface vacancies in copper nanoclusters and structure–property relationships.

---

1. Introduction

The nanoclusters continue to receive attention because of their promising catalysis, biosensing, and luminescence applications.^1–5 Owing to the quantum size effect and discrete electronic energy levels, the removal or addition of a very small number of atoms from the core or shell structure induces structural defects or structural distortions in the nanocluster that trigger significant changes in performance.^6–9 Atom-by-atom regulation can provide insight into the structural-evolution process and the mechanism responsible for how structural change affects performance.^10–12 For example, the removal of the center S atom from the Ag₁₄S core in [Ag₆₂S₁₃(SBu^t)₃₂]⁴⁺ (0e) results in the formation of a complete FCC Ag₁₄ unit in [Ag₆₂S₁₂(SBu^t)₃₂]²⁺ (4e) (defective structure with a hollow core), in which the presence of free valence electrons leads to photoluminescence quenching.^13–15 The traceless removal of two kernel atoms from Au₄₈(m-MBT)₂₆ to give Au₄₆(m-MBT)₂₆ facilitates PL emission (surface defects),¹⁶ while a single 28^th Ag atom inserted into the interstice (hole vacancy) of the outermost metal layer of the [Ag₂₇(S^tBu)₁₄(S)₂(CF₃COO)₉(DMAc)₄]·DMAc nanocluster gives photoluminescent [Ag₂₈(SAdm)₁₄(S)₂(CF₃COO)₁₀(H₂O)₄].¹⁷ Consequently, nanoclusters have been used as effective platforms to dissect the structural evolution of a nanomaterial from the perspective of defect chemistry and to understand how this evolution changes its properties, thereby providing rich insight into the relationship between structure and performance.^18–23

Nowadays, well determined copper-hydride nanoclusters have also been reported, such as the [Cu₈H₆(μ-dppm)₅](PF₆)₂,²⁴ [Cu₉H₇(μ-dpmppm)₃]X₂, [Cu₁₆H₁₄(μ-dpmppm)₄]X₂,²⁵ [Cu₂₀H₁₁{E₂P(OiPr)₂}₉ (E = S or Se),^26,27 [Cu₂₃(PhSe)₁₆(Ph₃P)₈(H)₆]·BF₄,²⁸ [Cu₂₅H₂₂(PR₃)₁₂]Cl,^29,30 [Cu₂₈H₁₅{S₂CN(nPr)₂}₁₂](PF₆),³¹ [Cu₂₈H₂₀(S₂P(OⁱPr)₂)₉]⁻,³² [Cu₂₈H₁₆(dppp)₄(RS)₄(CF₃CO₂)₈],³³ [Cu₃₀H₁₈{E₂P(OR)₂}₁₂](E = S or Se; R = ⁿPr,ⁱPr or ⁱBu),³⁴ Cu₃₂H₂₀{S₂PR₂}₁₂],^35,36 [Cu₅₃(RCOO)₁₀(C≡C^tBu)₂₀Cl₂H₁₈]⁺,³⁷ [Cu₆₁(S^tBu)₂₆S₆Cl₆H₁₄]⁺ (ref. 38) and [Cu₈₁(PhS)₄₆(tBuNH₂)₁₀(H)₃₂]³⁺.³⁹ And the copper-hydride nanoclusters have drawn increasing research interest because they have aesthetically fascinating molecular structures and are potentially useful in catalysis, hydrogen storage, and photovoltaics applications.^8,37,40–43 In contrast to the numerous reports on the structural evolution of gold/silver clusters, examining the structural evolution of copper-hydride clusters remains challenging owing to difficulties associated with synthesizing and crystallizing copper-hydride nanoclusters. A reversible transformation between [Cu₇(H){S₂CR}₆] and [Cu₈(H){S₂CR}₆](PF₆), a defect to defect-growth-based copper hydride cluster was reported by the Liu group.⁴⁴ Metal exchanging Pt⁴⁺ ions into [Cu₃₂(PET)₂₄Cl₂H₈]²⁻ resulted in [Pt₂Cu₃₄(PET)₂₂Cl₄]²⁻, an internal-defective nanocluster.^45,46 The surface vacancy defective [Cu₃₆H₁₀(PET)₂₄(PPh₃)₆Cl₂] nanoclusters and the hypothetical non-defective half-cubic copper hydride [Cu₃₈H₁₀(SC₆H₃F₂)₂₆(PPh₃)₈]²⁺ nanocluster were considered to be other examples of copper-hydride structural evolution that revealed surface metal atom vacancies, ligand defects, and skeletal distortion.⁴⁷ Although studies on defect clusters have been reported, there are few studies on the effects of surface defects and regrowth on structure and properties because only a few series of copper-hydride clusters are well determined, let alone studies on structural evolution driven in an atom-by-atom manner in a copper-hydride.^7,8,44,48,49

Herein, we report three novel well-determined Cu-hydride nanoclusters that provide insight into surface-vacancy-defect to non-defect structural evolution: surface-defective [Cu₂₈(S-c-C₆H₁₁)₁₈(PPh₂Py)₃H₈]²⁺ (Cu₂₈-PPh₂Py), and non-defective [Cu₂₉(SAdm)₁₅Cl₃(P(Ph-Cl)₃)₄H₁₀](PF₆) (Cu₂₉-P(Ph-Cl)₃) and Cu₂₉(S-c-C₆H₁₁)₁₈(P(Ph-^pMe)₃)₄H₁₀(BF₄) (Cu₂₉-P(Ph-Me)₃). Structural analyses of the surface vacancy-defective Cu₂₈-PPh₂Py and non-defective Cu₂₉-P(Ph-Cl)₃/Cu₂₉-P(Ph-Me)₃ provide deep insight into how surface vacancy defects structurally evolve into non-defects, including differences in the metal core, number of hydrogen ligands, packing model, metal shell, and core–shell bonding model. Interestingly, Cu₂₉-Ph-^pMe has a virtually identical structure to that of the MAg₂₈(SR)₁₈(PR′₃)₄ silver nanocluster in terms of the number of metal atoms, the thiol and phosphine ligand counts, and the atomic arrangement, despite the presence of hydrogen.⁵⁰ The structural evolution provides an understanding of how surface vacancy defects structurally evolve and provides valuable insight into the structures of surface vacancies in copper nanoclusters as well as structure–property relationships.

2. Methods

2.1 Chemicals

Tetrakis(acetonitrile)copper(I) tetrafluoroborate ([Cu(CH₃CN)₄]·BF₄, 98%), tetrakis(acetonitrile)copper(I) hexafluorophosphate ([Cu(CH₃CN)₄]·PF₆, 98%), diphenyl-2-pyridylphosphine (PPh₂Py, 97%), tri(p-tolyl)phosphine (P(Ph-^pMe)₃, 98%), tris(4-chlorophenyl)phosphine(P(Ph-^pCl)₃, 98%), cyclohexyl mercaptan (C₆H₁₂S, ≥99%), 1-adamantanethiol (1-AdmSH, 98%), sodium borohydride (NaBH₄, 98%), sodium borodeuteride (NaBD₄, 98%), acetonitrile (CH₃CN), chloroform (CHCl₃), chloroform-d (CDCl₃), methanol (MeOH), ethanol (EtOH), dichloromethane (CH₂Cl₂), and n-hexane (Hex) were obtained from commercial sources and used as received.

2.2 Synthesis of [Cu₂₈(C₆H₁₁S)₁₈(PPh₂Py)₃(H)₈][BF₄]₂ nanoclusters (Cu₂₈-PPh₂Py-H)

[Cu₂₈(C₆H₁₁S)₁₈(PPh₂Py)₃(H)₈][BF₄]₂ was prepared by dissolving [Cu(CH₃CN)₄]·BF₄ (160 mg, 0.51 mmol) in 1 : 1 (v/v) CH₃CN/CHCl₃ (10 mL). After 10 min, PPh₂Py (100 mg, 0.40 mmol) was added to the solution, followed by C₆H₁₁SH (22 μL, 0.18 mmol), and the mixture was vigorously stirred. After 20 min, a freshly prepared solution of NaBH₄ (30 mg, 0.8 mmol, dissolved in 2 mL MeOH) was added dropwise to the aforementioned solution, which immediately turned brown. The mixture was stirred for 5 h at room temperature, at which time a bright orange precipitate of the product had formed at the bottom of the flask. The product was collected and recrystallized from CH₂Cl₂/Hex over several days at room temperature.

2.3 Synthesis of [Cu₂₈(C₆H₁₁S)₁₈(PPh₂Py)₃(D)₈][BF₄]₂ nanoclusters (Cu₂₈-PPh₂Py-D)

Cu₂₈-PPh₂Py-D was synthesized similarly to Cu₂₈-PPh₂Py-H, with the exception that NaBD₄ (40 mg, 1 mmol) was used as the reducing agent.

2.4 Synthesis of [Cu₂₉(SAdm)₁₅Cl₃(P(Ph-Cl)₃)₄H₁₀](PF₆) nanoclusters (Cu₂₉-P(Ph-Cl)₃-H)

Cu₂₉-P(Ph-Cl)₃-H was prepared by dissolving [Cu(CH₃CN)₄]·PF₆ (100 mg, 0.27 mmol) in 1 : 1 (v/v) CH₃CN/CHCl₃ (10 mL) After 10 min, P(Ph-^pCl)₃ (100 mg, 0.27 mmol) was added to the solution, followed by 1-AdmSH (50 mg, 0.30 mmol) and the mixture was stirred vigorously. After 20 min, a freshly prepared solution of NaBH₄ (60 mg, 1.6 mmol, dissolved in 2 mL MeOH) was added. The mixture was stirred at room temperature for 10 h. The product was collected and recrystallized from CH₂Cl₂/Hex over several days at room temperature. Although no chloride was used in the synthesis, the capping Cl atoms in Cu₂₉-P(Ph-Cl)₃ can be reasonably considered to be derived from the solvent CHCl₃, similar to previous reports of some Cl ligand-containing clusters.^51,52

2.5 Synthesis of [Cu₂₉(SAdm)₁₅Cl₃(P(Ph-Cl)₃)₄D₁₀](PF₆) nanoclusters (Cu₂₉-P(Ph-Cl)₃-D)

The synthesis of Cu₂₉-P(Ph-Cl)₃-D is similar to that of Cu₂₉-P(Ph-Cl)₃-H, except that the reducing agent is changed to 64 mg NaBD₄ (1.6 mmol).

2.6 Synthesis of [Cu₂₉(SC₆H₁₁)₁₈((Ph-^pMe)₃P)₄H₁₀][BF₄] nanoclusters (Cu₂₉-P(Ph-Me)₃-H)

Cu₂₉-P(Ph-Me)₃-H was prepared following the procedure used to synthesize Cu₂₈-PPh₂Py-H, with the exception that tri(p-tolyl)phosphine ((Ph-^pMe)₃P) (116 mg, 0.38 mmol) was used instead of PPh₂Py (100 mg, 0.40 mmol). Red crystals were obtained by recrystallization from CH₂Cl₂/Hex over one week.

2.7 Synthesis of [Cu₂₉(C₆H₁₁S)₁₈((Ph-^pMe)₃P)₄D₁₀][BF₄] nanoclusters (Cu₂₉-Ph-^pMe-D)

Cu₂₉-P(Ph-^pMe)₃-D was synthesized similarly to Cu₂₉-P(Ph-Me)₃-H, with the exception that NaBD₄ (40 mg, 1 mmol) was used instead of NaBH₄ (30 mg, 0.8 mmol).

2.8 Characterization

All UV-vis spectra of the nanoclusters were recorded on a Metash UV-6000PC UV-vis spectrophotometer, and the samples were dissolved in CH₂Cl₂ whose background correction was made using a CH₂Cl₂ blank. X-ray photoelectron spectroscopy (XPS) measurements were performed using a Thermo ESCALAB 250 configured with a monochromated Al Kα (1486.8 eV) 150 W X-ray source, a 0.5 mm circular spot size, a flood gun to counter charging effects, and an analysis chamber base pressure lower than 1 × 10⁻⁹ mbar, and the data were collected with an FAT of 20 eV. Electrospray ionization (ESI) mass spectrum was obtained on a Waters XEVO G2-XS QT of mass spectrometer. The samples are dissolved in a mixture solution of CH₂Cl₂/CH₃OH (v/v = 1 : 1), which is directly infused into the chamber at 10 μL min⁻¹ with positive mode. Nuclear magnetic resonance (NMR) analysis was performed on a Bruker Avance spectrometer operating at 400 MHz. CD₂Cl₂ was used as the solvent to dissolve the crystals of copper-hydride clusters; the residual solvent peak (i.e., 1H at 5.32 ppm) was used as the reference. The clusters were dissolved in CH₂Cl₂ for the ²H NMR spectra. Data collection for single-crystal X-ray diffraction was carried out using a Stoe Stadivari diffractometer under a liquid nitrogen flow at 150 K, using graphite-monochromatized Cu Kα radiation (λ = 1.54186 Å). Data reductions and absorption corrections were performed using SAINT and SADABS programs.

3. Results and discussion

The Cu₂₈-PPh₂Py, Cu₂₉-P(Ph-Cl)₃, and Cu₂₉-P(Ph-Me)₃ series of copper nanoclusters were directly synthesized using a one-pot method under mild conditions. Briefly, Cu₂₈-PPh₂Py was synthesized by dissolving [Cu(CH₃CN)₄]BF₄ in a mixture of acetonitrile (CH₃CN) and chloroform (CHCl₃) at room temperature, after which C₆H₁₁SH and PPh₂Py were added, and NaBH₄ was used as the reducing agent. The solution turned dark red. The orange crystals were characterized by single-crystal X-ray crystallography (SC-XRC) (Fig. S1†). Cu₂₈-PPh₂Py crystallized in the monoclinic P21/a space group (Table S1†). On the other hand, Cu₂₈-PPh₃ (Fig. S2†) crystallized in the monoclinic I2/a space group (Table S2†). Cu₂₉-Ph-^pMe was obtained when the P(Ph-^pMe)₃ ligand was used instead of PPh₂Py under the same conditions (Fig. S3†). Red Cu₂₉-Ph-^pMe crystallized in the trigonal R3̄ space group (Table S3†).

The structures of Cu₂₈-PPh₂Py, Cu₂₉-P(Ph-Cl)₃, and Cu₂₉-P(Ph-Me)₃ are shown in Fig. 1. Cu₂₈-PPh₂Py/PPh₃ contains 28 Cu atoms, 18 S-c-C₆H₁₁ ligands, eight hydrides, and three PPh₂Py ligands (Fig. 1A and B). A surface vacancy defect located in the Cu₃(SR)₆ motif is observed in Cu₂₈-PPh₂Py (Fig. 1C). The copper skeletons of Cu₂₉-P(Ph-Cl)₃ and Cu₂₉-P(Ph-Me)₃ formed (Fig. 1D–I) when the 29^th copper atom was inserted into the interstice (surface vacancy site) (Fig. 1F and I). The formed Cu₂₉-P(Ph-Cl)₃ possesses 29 Cu atoms, 15-SAdm ligands, 3 Cl atoms, 10 hydrides, and four (P(Ph-Cl)₃). Similarly, Cu29-P(Ph-Me)₃ possesses 29 Cu atoms, 18 S-c-C₆H₁₁ ligands, 10 hydrides, and four P(Ph-^pMe)₃ ligands, with four P(Ph-^pMe)₃ units occupying the four vertices of the tetrahedron. As shown in Fig. S4–S6,† the Cu₂₈-PPh₂Py framework possesses a C₁ symmetry axis, whether observed from the position of the three phosphine ligands or the vacancy (Fig. S4†). On the other hand, the Cu₂₉-P(Ph-Cl)₃ framework possesses one C₃ symmetry axis, indicative of improved symmetry (Fig. S5†), while the Cu₂₉-P(Ph-Me)₃ framework forms a standard tetrahedral configuration with four C₃ symmetry axes (Fig. S6†). Once again, the Cu₂₉-Ph-^pMe cluster exhibits significantly improved symmetry, affecting the copper-hydride nanocluster's stability. The novel changes observed for the metal core, the number of hydrogen ligands, the packing model, the metal shell, and the core/shell bonding model facilitate a detailed understanding of the surface-vacancy-defect to non-defect structural evolution.

Fig. 1 The overall structures of the series of copper nanoclusters. (A–C) Cu₂₈-PPh₂Py, (D–F) Cu₂₉-P(Ph-Cl)₃, and (G–I) Cu₂₉-P(Ph-Me)₃ and the Cu^29th sites. Color scheme: copper = Cu, yellow = S, purple = P, blue-purple = N, green = Cl, and gray = C. H atoms are omitted for clarity.

The changes observed during structural evolution can be understood in detail by splitting the structure. The structural anatomy of surface-defective Cu₂₈-PPh₂Py is shown in Fig. 2A, which reveals that the Cu₁₃ copper kernel has an anti-cuboctahedral structure. The Cu–Cu distances in the anti-cuboctahedron vary between 2.459 and 2.787 Å, with an average value of 2.605 Å. The anti-cuboctahedron Cu₁₃ is surrounded by a Cu₁₅(SR)₁₈P₃ cage shell, which can be viewed as an assembly of a single Cu₃(SR₆) (Motif 1a) and three Cu₄S₆P₁ motifs (Motifs 2a–4a). While Motifs 2a–4a share the same molecular formula, their geometries differ. In contrast, Motif 1a has surface vacancy defects. Non-defective Cu₂₉-P(Ph-Cl)₃ capped by P(Ph-Cl)₃, SAdm-, and Cl⁻ and H⁻ ligands was obtained (Fig. 2B) when the 29^th copper atom was inserted into the surface vacancy site of Motif 1a. An anti-cuboctahedral Cu₁₃ core, similar to the anti-cuboctahedral Ag₁₃ core in Ag₁₉ and Ag₂₅ (ref. 53) was observed for Cu₂₉-P(Ph-Cl)₃ by comparison with the structure of Cu₂₈-PPh₂Py. The anti-cuboctahedral Cu₁₃ core is further capped by a Cu₁₆(SR)₁₅P₄ shell assembled by using Cu₄(SR)₆P₁ (Motif 2b) and three Cu₄(SR)₅ClP₁ units (Motifs 2a, 2c, and 2d). Because the AdmSH ligand is very sterically hindering, another thiol ligand cannot be accommodated; hence, the small Cl acts as a protective ligand to maintain the stability of the Cu₂₉-P(Ph-Cl)₃ structure. Another non-defective Cu₂₉-P(Ph-Me)₃ nanocluster capped by P(Ph-Me)₃, S-c-C₆H₁₁- and H⁻ ligands is shown in Fig. 2C. In contrast to the anti-cuboctahedral Cu₁₃ core of Cu₂₈-PPh₂Py and Cu₂₉-P(Ph-Cl)₃, the Cu₁₃ core in Cu₂₉-P(Ph-Me)₃ is cuboctahedral in structure. The cuboctahedral Cu₁₃ framework was also observed in [Cu₁₃(S₂CNnBu₂)₆(C≡CR)₄](PF₆).^54,55 The Cu–Cu distances in the cuboctahedron vary between 2.457 and 2.878 Å, with an average value of 2.651 Å. The cuboctahedral Cu₁₃ core is surrounded by a Cu₁₆(SR)₁₈P₄ shell assembled using four identical Cu₄S₆P₁ motifs (Motifs 1c–4c). The structure of Cu₂₉-P(Ph-Me)₃ is consistent with that of a standard tetrahedron, in which the Cu₂₉-P(Ph-Me)₃ frame has the same structure as MAg₂₈(SR)₁₈(PR′₃)₄ (Fig. S7†).

Fig. 2 Structural anatomies of the series of copper nanoclusters. (A) Cu₂₈-PPh₂Py, (B) Cu₂₉-P(Ph-Cl)₃, and (C) Cu₂₉-P(Ph-Me)₃. Color scheme: copper = Cu, yellow = S, purple = P, and green = Cl.

Furthermore, surface-kernel structural transfer provides diverse bonding patterns during structural evolution. First, both the cuboctahedral and anti-cuboctahedral Cu₁₃ cores have eight triangular (Cu₃) surfaces and quadrilateral (Cu₄) surfaces (Fig. S8†). As shown in Fig. 3A, Cu₂₈-PPh₂Py has one Cu₄(SR)₆P₁ motif (marked with a red circle) that covers one triangular Cu₃ surface, two Cu₄(SR)₆P₁ motifs (marked with green circles) that cover quadrilateral (Cu₄) surfaces, as well as one Cu₃(SR)₆ motif (marked with a yellow circle) containing the surface vacancy defect that also covers a quadrilateral (Cu₄) surface. A vertex Cu–P resulting from the spatial position of the thiol ligand located within the Cu₃(SR) ring is the difference between the Cu₄(SR)₆P₁ units (marked with green circles) and the Cu₃(SR)₆ moiety (marked with a yellow circle). As shown in Fig. S9,† when the carbon tails of the three thiol ligands in the Cu₃(SR)₆ ring are aligned outwards, the reserved space can support a vertex Cu–P staple insertion to form Cu₄(SR)₆P₁, while there is insufficient reserved space to support a vertex Cu–P staple when one thiol-ligand carbon tail is arranged inward, resulting in surface vacancy defects. This observation reveals that the orientation of the ligand carbon tail affects the formation of surface-vacancy defects. In addition, its bonding mode includes three quadrilateral Cu₄ units capped by Cu₂(SR)₅ shells (marked with blue circles) and three triangular Cu₃ units capped by Cu₂(SR)₅ shells (marked with purple circles). For non-defective Cu₂₉-P(Ph-Cl)₃, the Cu₄(SR)₆P₁ motif marked with a red circle also covers a triangular Cu₃ surface, and the three Cu₄(SR)₅ClP₁ motifs marked with green circles also cover quadrilateral Cu₄ surfaces (Fig. 3B). Three quadrilateral Cu₄ moieties capped by Cu₂(SR)₅ shells (marked with blue circles) and three triangular Cu₃ units capped by Cu₂(SR)₄Cl shells (marked with purple circles) are also present, as observed in the structure of Cu₂₈-PPh₂Py. For Cu₂₉-P(Ph-Cl)₃, the capping Cl and AdmSH ligands endow enough space to support the insertion of Cu–P to form non-defective clusters (Motif 1a vs. Motif 1b in Fig. 2). Structural evolution from the surface-vacancy-defective Cu₂₈-PPh₂Py to non-defective Cu₂₉-P(Ph-Cl)₃ involves the growth of defect motifs, with the bonding mode between the core and the motif maintained to a certain degree. The other non-defective Cu₂₉-P(Ph-Me)₃ structure contains four Cu₄(SR)₆P₁ motifs that cover four triangular surfaces (Cu₃) of cuboctahedral Cu₁₃ (marked with red circles). The spatial arrangement of thiol ligands can better support the insertion of Cu–P, forming four identical Cu₄S₆P₁ motifs. All six Cu₂(SR)₅ shells cover the quadrilateral Cu₄ unit (marked with blue circles) (Fig. 3C).

Fig. 3 The detailed core/shell bonding modes. The bonding modes in (A) Cu₂₈-PPh₂Py, (B) Cu₂₉-P(Ph-Cl)₃, and (C) Cu₂₉-P(Ph-Me)₃. Structures marked with circles of the same color are of the same type. Color scheme: copper = Cu, yellow = S, purple = P, and green = Cl.

Structural evolution involves the growth of surface vacancy defects from surface-vacancy-defective Cu₂₈-PPh₂Py to non-defective Cu₂₉-P(Ph-Cl)₃, accompanied by core and bond pattern maintenance (Fig. 3Avs.Fig. 3B). Structural evolution involves the transformation of the Cu₂₉-P(Ph-Cl)₃ core and remodeling of the Cu₂₉-P(Ph-Cl)₃ shell, respectively, to form the other non-defective Cu₂₉-P(Ph-Me)₃ (Fig. 3Bvs.Fig. 3C). These results provide atomically precise insights into the defect-induced readjustment of the local structure.

Single-crystal X-ray crystallography (SCXRC) revealed both the intramolecular structure and intermolecular packing mode of the metal nanoclusters during structural evolution. Cu₂₈-PPh₂Py and Cu₂₉-P(Ph-Cl)₃ only exhibit 2H arrangements with “AB” packing sequences (Fig. S10 and S11†), while the crystalline unit cell of Cu₂₉-P(Ph-Me)₃ shows a special “ABCDEF” stacking sequence. The Cu₂₉-Ph-^pMe nanoclusters in the stacking layer are uniformly arranged and each nanocluster is surrounded by six identical nanoclusters with the same tropism, as shown in Fig. S12.† Such a 6H pattern was previously only observed in Au₆₀ reported by Wu et al.⁵⁶

The UV-vis spectra of Cu₂₈-PPh₂Py, Cu₂₉-P(Ph-Cl)₃, and Cu₂₉-P(Ph-Me)₃ in CH₂Cl₂ are shown in Fig. S13.† Several weak absorptions are observed at 256 and 400 nm for Cu₂₈-PPh₂Py, at 270, 326, and 415 nm for Cu₂₉-P(Ph-Cl)₃, and at 410, 325, and 266 nm for Cu₂₉-P(Ph-Me)₃. Because determining the number of hydrogen atoms by SC-XRD is difficult, careful ESI-MS was used to determine the valence and molecular formula of each cluster and the number of hydrides. Positive-mode ESI-MS data for Cu₂₈-PPh₂Py, Cu₂₉-P(Ph-Cl)₃, and Cu₂₉-P(Ph-Me)₃ are shown in Fig. 4. The spectrum of Cu₂₉-Ph-^pMe shows a prominent peak corresponding to a +1 charge at m/z = 5144.69 Da, which is attributable to [Cu₂₉(S-c-C₆H₁₁)₁₈(P(Ph-^pMe)₃)₄H₁₀]⁺ (Fig. 4A). The experimentally observed isotope pattern for [Cu₂₉(S-c-C₆H₁₁)₁₈(P(Ph-^pMe)₃)₄H₁₀]⁺ is in good agreement with the calculated pattern (Fig. S14A†). Cu₂₉-P(Ph-Me)₃-D was synthesized and subjected to ESI-MS to further confirm the presence and number of hydrides. A prominent peak corresponding to a +1 charge was observed at m/z = 5154.66 Da, which is attributable to [Cu₂₉(S-c-C₆H₁₁)₁₈(P(Ph-^pMe)₃)₄D₁₀]⁺ (Fig. 4B). Accordingly, 10 hydrides are present in the Cu₂₉-P(Ph-Me)₃ cluster. Similarly, the ESI-MS spectrum of Cu₂₉-P(Ph-Cl)₃ confirmed that it is complex (Fig. 4C and D). First, the prominent peak corresponding to a charge of +1 observed at m/z = 4468.27 Da belongs to [Cu₂₉(SC₁₀H₁₅)₁₅Cl₃H₁₀]⁺ derived from the removal of four P(Ph-Cl)₃ ligands from the complete [Cu₂₉(SC₁₀H₁₅)₁₅Cl₃(P(Ph-Cl)₃)₄H₁₀]⁺, and it is fully consistent with the calculated result (Fig. S14B†). Neighboring peaks at 4433.32 and 4396.33 Da are attributable to [Cu₂₉(SC₁₀H₁₅)₁₅Cl₃H₁₀-Cl] and [Cu₂₉(SC₁₀H₁₅)₁₅Cl₃H₁₀-2Cl], respectively. The difference of 10 Da between the m/z values of Cu₂₉-P(Ph-Cl)₃-H and Cu₂₉-P(Ph-Cl)₃-D indicates the presence of 10 hydrogen atoms. By comparison, the ESI-MS spectrum of [Cu₂₈(S-c-C₆H₁₁)₁₈(Ph₂PyP)₃H₈]²⁺ was more complicated, with peaks corresponding to charges of +1, +2, and +3 observed. Specifically, the +1 signal at m/z = 4726.34 Da is attributable to [Cu₂₈(C₆H₁₁S)₁₈(C₁₇H₁₄NP)₃(H)₈ + Cl + CH₃CN]⁺, while that corresponding to a charge of +2 at m/z = 2390.55 Da is attributable to [Cu₂₈(C₆H₁₁S)₁₇(C₁₇H₁₄NP)₃(H)₈ + Cl + (CH₃CN)(CH₂Cl₂)₂]²⁺, consistent with the one thiol moiety in [Cu₂₈(C₆H₁₁S)₁₈(C₁₇H₁₄NP)₃(H)₈]²⁺ replaced by Cl. The +3 peak at m/z = 1583.35 Da is attributable to [Cu₂₈(S-c-C₆H₁₁)₁₈(C₁₇H₁₄NP)₃H₈ + CuCl]³⁺ (Fig. 4E–H and isotopic patterns in Fig. S14C–E†). We also synthesized and characterized Cu₂₈-PPh₂Py-D. The intervals between the 3+, 2+, and 1+ peaks of Cu₂₈-PPh₂Py-H and Cu₂₈-PPh₂Py-D were found to be 2.66, 4, and 8, respectively, consistent with the presence of eight hydrogen atoms (i.e., 2.66 × 3/4 × 2/8 × 1). Although no BF₄⁻ counterion was found in the unit cell of Cu₂₈-PPh₂Py, its presence was confirmed by ESI-MS (Fig. S25†). So, combining these experimental results, none of the three copper hydride clusters have free electrons (28(Cu 4s¹) − 18 (SR) − 8 (H-) − 2 (charge) = 0e for Cu₂₈-PPh₂Py, 29 (Cu 4s¹) − 15 (SR) − 3 (Cl) − 10 (H–) − 1 (charge) = 0e for Cu₂₉-P(Ph-Cl)₃ and 29 (Cu 4s¹) − 18 (SR) − 10 (H–) − 1 (charge) = 0e for Cu₂₉-P(Ph-Me)₃).

Fig. 4 ESI-MS of the series of copper nanoclusters. Positive-mode ESI-MS spectra of (A and B) Cu₂₉-P(Ph-Me)₃, (C and D) Cu₂₉-P(Ph-Cl)₃, and (E–H) Cu₂₈-PPh₂Py.

X-ray photoelectron spectroscopy (XPS) was used to analyze the chemical compositions and copper-charge states in Cu₂₈-PPh₂Py, Cu₂₉-P(Ph-Cl)₃, and Cu₂₉-P(Ph-Me)₃ (Fig. S20–S24†), with the Cu_2p, S_2p, P_2p, F_1s, and N_1s XPS spectra shown in Fig. S20–S23.† The F_1s signals are ascribable to the BF₄⁻/PF₆⁻ counter ions. The presence of “BF₄⁻” in the Cu₂₈-PPh₂Py nanocluster was also confirmed by ESI-MS (Fig. S25†). Simple XPS cannot accurately determine the specific valence states of copper in copper nanoclusters because the Cu 2p_3/2 binding energies of Cu₀ and the Cu⁺ in Cu₂S are identical (932.6 eV);⁵⁷ Therefore, Cu Auger-electron spectroscopy is required to further determine the specific valence states of copper. The Auger spectra of Cu₂₈-PPh₂Py, Cu₂₉-P(Ph-Cl)₃, and Cu₂₉-P(Ph-Me)₃ show only one main Cu LMM peak at 570.69 eV (Fig. S24†), which corresponds to the Cu +1 oxidation state. Moreover, the three clusters exhibited no observable signal near 943 eV, consistent with the absence of Cu(II) in Cu₂₈-PPh₂Py, Cu₂₉-P(Ph-Cl)₃, and Cu₂₉-P(Ph-Me)₃.

The structural evolution from Cu₂₈-PPh₂Py to Cu₂₉-P(Ph-Cl)₃ and then to Cu₂₉-P(Ph-Me)₃ may lead to changes in the positions of the hydrogen atoms. Trying to grow single crystals of suitable quality for neutron diffraction is challenging.^26b,27,31,35 So, DFT calculations and HNMR spectra were often used for validating possible plausible locations for hydrogen.^{26a,38,39,45,47} To determine the hydride locations in Cu₂₈-PPh₂Py, Cu₂₉-P(Ph-Cl)₃, and Cu₂₉-P(Ph-Me)₃, density functional theory (DFT) calculations to optimize their positions based on the X-ray crystallographic structures were performed first.^{26a,38,39,45,47,58} In previous reports, it was found that for copper clusters, the positions of hydrogens in structures with similar moieties have a certain similarity, for example, the positions of 6 hydrogens in Cu₃₂H₂₀ are found to be similar to their structural positions in Cu₂₀H₁₁.^8,26b,35 Inspired by these research reports, and given the number of hydrides and the similarity in Cu@Cu₁₂ kernel structures as well as the motifs of Cu₂₈-PPh₂Py, Cu₂₉-P(Ph-Cl)₃, Cu₂₉-P(Ph-Me)₃ and [Cu₂₅(SPhCl₂)₁₈H₁₀]³⁻ (anti-cuboctahedral Cu₂₅), the locations of the hydrides in Cu₂₈-PPh₂Py, Cu₂₉-P(Ph-Cl)₃, and Cu₂₉-P(Ph-Me)₃ were determined referring to that of [Cu₂₅(SPhCl₂)₁₈H₁₀]³⁻, while the positions of hydrogens in Cu₂₅ (Fig. 5A), which served as references, were carefully determined by the Zheng group via HNMR spectroscopy and DFT calculations.⁵⁹ These optimized structures can remain stable, indicating that the determined hydrogen positions are reasonable. Fig. 5 shows that all hydrides are located around the M₁₃ core. Half of the eight triangular structures and all six square-planar structures of the Cu₁₃ core surface in Cu₂₉-P(Ph-Me)₃ are capped with one hydride each (Fig. 5B). Meanwhile, Cu₂₈-PPh₂Py shares the same icosahedral structure as Cu₂₅, and its eight hydrides are separated into two groups: four capping the triangular structure and another four capping the square-planar structure of the Cu₁₃ core surface (Fig. 5C). Another Cu₂₉-P(Ph-Cl)₃ cluster was obtained, whose Cu@Cu₁₂ core is the same as that of the Cu₂₅ core and Cu₂₈-PPh₂Py, which are both cuboctahedral. The ten H atoms of Cu₂₉-P(Ph-Cl)₃ cover all square faces and half of the triangular faces (Fig. 5D). The maintained frameworks of the proposed structures of the aforementioned Cu₂₉-P(Ph-Me)₃, Cu₂₉-P(Ph-Cl)₃, and Cu₂₈-PPh₂Py clusters (denoted as Cu₂₉-P(Ph-Me)₃^T, Cu₂₉-P(Ph-Cl)₃^T, and Cu₂₈-PPh₂Py^T) following geometry optimization using the GGA: PBE/DND method and the Dmol³ package^60–62 confirmed that these structures are stable (Fig. S19†). The geometric details of the Cu@Cu₁₂ kernels of Cu₂₉-P(Ph-Me)₃^T, Cu₂₉-P(Ph-Cl)₃^T, and Cu₂₈-PPh₂Py^T are listed in Table S4.† The positions of the hydrogens in these three nanoclusters are more intuitively compared in Fig. S15–S18.† All four μ₃-H-Cu₃ (μ₃-H-1/2/3/4-Cu₃) units are located in the Cu₆(SR)₆ ring in Cu₂₅, and three μ₄-H-Cu₄ (μ₃-H-5/6/7-Cu₄) moieties are capped by three Cu₂(SR)₅ motifs, and three μ₄-H-Cu₄ (μ₃-H-8/9/10-Cu₄) units are capped by three Cu₃(SR)₆ units (Fig. S15†). The four μ-H-Cu₃ (μ-H-1/2/3/4-Cu₃) units in Cu₂₉-P(Ph-Me)₃ and Cu₂₈-PPh₂Py are also located in their Cu₆(SR)₆ rings, while three μ-H-Cu₃ (μ-H-1/2/3/-Cu₃) moieties are located in the Cu₆(SR)₆ ring of Cu₂₉-P(Ph-Cl)₃, and one μ-H-Cu₃ moiety is located in the Cu₆(SR)₃Cl₃ ring. Furthermore, six μ-H-Cu₄ (μ-H-5/6/7/8/9/10-Cu₄) moieties are capped by six Cu₂(SR)₅ motifs in Cu₂₉-P(Ph-Me)₃ (Fig. S16†), three μ-H-Cu₄ (μ-H-5/6/7-Cu₄) units are capped by three Cu₂(SR)₅ motifs, and three μ-H-Cu₄ (μ-H-8/9/10-Cu₄) moieties are capped by three Cu₃(SR)₅Cl units in the case of Cu₂₉-P(Ph-Cl)₃ (Fig. S17†), while three μ-H-Cu₄ (μ-H-5/6/7-Cu₄) units are capped by three Cu₂(SR)₅ motifs, and only one μ-H-Cu₄ (μ-H-8-Cu₄) is capped by Cu₃(SR)₆ in the case of Cu₂₈-PPh₂Py (Fig. S18†). These observations provide a reference for predicting the number and locations of the hydrogen atoms in the nanocluster framework.

Fig. 5 Structural anatomies of the three nanoclusters: (A) Cu₂₅, (B) Cu₂₉-P(Ph-Me)₃^T, (C) Cu₂₈-PPh₂Py^T, and (D) Cu₂₉-P(Ph-Cl)₃^T. All carbon, hydrogen, and nitrogen atoms are omitted for clarity. Color scheme: copper = Cu, blue = Cu, yellow = S, pink = S, purple = P, white = H, and green = Cl.

Furthermore, the HNMR spectra of these three copper-hydride nanoclusters were carefully obtained via dissolving crystals in the solvent (Fig. S26–S28†). For the ²HNMR of Cu₂₈-PPh₂Py-D (Fig. S26B†), four different D atom signal peaks can be observed, with intensity ratios of about 3 : 1 : 1 : 3. Broadened signal peaks can be observed for the D signal in the ²HNMR spectra of Cu₂₉-P(Ph-Me)₃-D. And there also were about 4D in the upfield region (low ppm), and about 6D in the low-field region (high ppm) (Fig. S27B†). The deuterium signal ratio at upfield and low-field was 6 : 4.13 for Cu₂₉-P(Ph-Cl)₃-D (Fig. S28B†). Combining the optimized geometry, ²HNMR data, and the peak position of the D in Cu₂₅H₁₀:⁵⁹ (i) for Cu₂₈-PPh₂Py-D, the 4D (3D_a + 1D_b) in the upfield region (low ppm) can be attributed the H-1, H-2, H-3, and H-4 and the 4D (1D_c + 3D_d) in the low-field region (high ppm) can be attributed to H-5, H-6, H-7, and H-8 (Fig. S17† and the inset of Fig. S26B†); (ii) for Cu₂₉-P(Ph-Me)₃-D, the 4D (4D_a) in the upfield region (low ppm) can be attributed the H-1, H-2, H-3, and H-4 and the 6D (6D_d) observed in the low-field region (high ppm) can be attributed to H-5, H-6, H-7, H-8, H-9 and H10 (Fig. S16† and the inset of Fig. S27B†). In contrast to four different D atoms (D_a, D_b, D_c, and D_d) of Cu₂₈-PPh₂Py-D, Cu₂₉-P(Ph-Me)₃-D has two D atoms, viz., D_a and D_d, due to the high symmetry; (iii) for Cu₂₉-P(Ph-Cl)₃-D, the 4D (3D_a + 1D_b) in the upfield region (low ppm) can be attributed the H-1, H-2, H-3, and H-4 and the 6D (3D_c + 3D_d) in the low-field region (high ppm) can be attributed to H-5, H-6, H-7, H-8, H-9 and H-10 (Fig. S18† and the inset of Fig. S28B†). From the information of ²HNMR spectra, the signal peak of hydrogen of can be better attributed. The 46H located between 7 ppm and 9 ppm were assigned to hydrogen from PPh₂Py (3 × 14) and 4 H_c/d in the cluster. Due to the influence of N on PPh₂Py, part of the hydrogen is significantly shifted, i.e. the hydride resonance (3H) at 8.76 ppm. The signal in the alkane region (206H) was due to the thiol and H-ligands (Cal. 18 × 11 + 4(H_a/H_b) = 204H, and Δ = 2H) (Fig. S26†). In the ¹HNMR spectrum of Cu₂₉-P(Ph-Me)₃-H, 52H in the aromatic region belonged to the –C₆H₄R of P(Ph-Me)₃ (4 × 4 × 3) and 6 H_d in the cluster, and the 237.63H were attributed to the thiol, H-ligands and –CH₃ in P(Ph-Me)₃ (Cal. 18 × 11 + 4(H_a) + 4 × 3 × 3 = 238H) (Fig. S27A†). Similarly, 52H in the aromatic region belonged to the –C₆H₄– of P(Ph-Cl)₃ and 6 H_c/d in the cluster, and the 228.12H were attributed to the thiol, H-ligands (Cal. 15 × 15 + 4 (H_a/b) = 229H) (Fig. S28A†).Based on the this information, we can reasonably consider that the hydrogen located in the upfield was the hydrogen (on the cluster surface) covering the Cu₃ surface of the Cu₁₃ core, and the hydrogens encapsulated in trigonal prismatic (tp) cages were located in the low-field region, which in turn showed the rationality of predicting the hydrogen positions in the three clusters.

Surface-defect growth and enhanced overall symmetry affect the stability of copper-hydride nanoclusters. The thermal stabilities of these copper-hydride nanoclusters were explored by storing CH₂Cl₂ solutions of Cu₂₉-P(Ph-Me)₃, Cu₂₉-P(Ph-Cl)₃, and Cu₂₈-PPh₂Py at room temperature for a week (Fig. S29†). Fig. S29A† reveals that Cu₂₈-PPh₂Py starts to decompose after three days, accompanied by precipitation, while some flocs began to form on the fifth day in the Cu₂₉-P(Ph-Cl)₃ solution (Fig. S29B†). In contrast, the solution of Cu₂₉-P(Ph-Me)₃ remained stable for 5 days (Fig. S29C†). Furthermore, a solution of Cu₂₉-P(Ph-Me)₃ in CHCl₃ was stable for more than 5 h in an oil bath at 50 °C, while these conditions destroyed both Cu₂₉-P(Ph-Cl)₃ and Cu₂₈-PPh₂Py (Fig. S30†). In this context, Cu₂₉-P(Ph-Me)₃ is the most stable, followed by Cu₂₉-P(Ph-Cl)₃ and Cu₂₈-PPh₂Py. The metastability exhibited by Cu₂₈-PPh₂Py is ascribable to surface vacancy defects that result in incomplete capping of the nanocluster surface. In comparison, Cu₂₉-P(Ph-Me)₃, which has a standard tetrahedral structure, is more stable than Cu₂₉-P(Ph-Cl)₃.

4. Conclusions

We synthesized and identified three copper hydride nanoclusters through ligand engineering, including [Cu₂₈(S-c-C₆H₁₁)₁₈(PPh₂Py)₃H₈](BF₄), a surface-vacancy-defective nanocluster, and [Cu₂₉(SAdm)₁₅Cl₃(P(Ph-Cl)₃)₄H₁₀](PF₆) and [Cu₂₉(S-c-C₆H₁₁)₁₈(P(Ph-^pMe)₃)₄H₁₀](BF₄), which are both non-defective nanoclusters. Because there are few reports on the series of the defect- and defect-growth-based copper-hydride clusters, the series of well-characterized copper nanoclusters prepared in this study provides insight into surface-vacancy-defect to non-defect structural evolution. Structural adjustments were observed in the metal core, hydrogen ligand, metal shell, core/shell bonding pattern, and packing mode during the evolution of surface-vacancy-defective Cu₂₈-PPh₂Py to non-defective Cu₂₉-P(Ph-Cl)₃ and then to the other Cu₂₉-P(Ph-Me)₃. Specifically: (i) the anti-cuboctahedral Cu₁₃ core is maintained between Cu₂₈-PPh₂Py and Cu₂₉-P(Ph-Cl)₃, but changes to a cuboctahedral Cu₁₃ structure in Cu₂₉-P(Ph-Me)₃. (ii) The doubly charged Cu₂₈-PPh₂Py is embedded with eight hydrides, while singly charged Cu₂₉-P(Ph-Cl)₃/Cu₂₉-P(Ph-Me)₃ is embedded with 10 hydrides; however, the positions of the hydrogens are very similar in both clusters. (iii) The Cu₁₆(SR)₁₈P₄ shell in Cu₂₉-P(Ph-Me)₃ contains four identical Cu₄S₆P₁ motifs assembled by sharing six SR ligands, in which four Cu₄S₆P₁ motifs cover the four triangular surfaces (Cu₃) of the cuboctahedral Cu₁₃. On the other hand, the Cu₁₆(SR)₁₈P₄ shell in Cu₂₉-P(Ph-Cl)₃ is an assembly of one Cu₄(SR)₆P₁ motif that covers the triangular surfaces (Cu₃) and three Cu₄(SR)₅ClP₁ motifs that cover the quadrilateral surface (Cu₄) of the anti-cuboctahedral Cu₁₃. As was observed for Cu₂₉-P(Ph-Cl)₃, Cu₂₈-PPh₂Py is assembled from one Cu₄(SR)₆P₁ motif that covers the triangular surfaces (Cu₃), two Cu₄(SR)₆P₁ motifs that cover the quadrilateral surface (Cu₄), and one defective Cu₃(SR)₆ unit that covers the quadrilateral surface (Cu₄), which leads to a complex bonding environment. (iv) Cu₂₉-P(Ph-Me)₃ shows a 6H pattern in the crystal lattice, whereas Cu₂₈-PPh₂Py and Cu₂₉-P(Ph-Cl)₃ only show 2H arrangements. (v) Stability improved in moving from Cu₂₈-PPh₂Py to Cu₂₉-P(Ph-Cl)₃ and then to Cu₂₉-P(Ph-Me)₃, which indicates that surface defects, as well as the symmetry of the overall structure, affect the stability of the copper-hydride cluster. The structural evolution observed in moving from surface-vacancy-defective Cu₂₈-PPh₂Py to non-defective Cu₂₉-P(Ph-Cl)₃ and then to the other non-defective Cu₂₉-P(Ph-Me)₃ enables the structural evolution and the mechanism responsible for the physicochemical properties to be understood and provides valuable insight into the structural surface vacancies in copper nanoclusters and their structure–property relationships.

Data availability

All the data supporting this article have been included in the main text and the ESI.†

Author contributions

Y. B. and B. Y. carried out the experiments, analyzed the data, and wrote the manuscript. X. K. revised the manuscript. X. W. and H. Y. completed DFT calculations. Z. L. and H. D. assisted in the synthesis of nanoclusters. M. Z., S. C., and S. J. designed the project, analyzed the data, and revised the manuscript.

Conflicts of interest

The authors declare no competing financial interests.

Acknowledgements

We acknowledge the financial support provided by the National Natural Science Foundation of China (21901001, 21871001, and 21631001), Ministry of Education and Education Department of Anhui Province, andAnhui Provincial Natural Science Foundation (No. 1908085QB54), and acknowledge the technical support of the high-performance computing platform of Anhui University.

Notes and references

1. R. Jin, C. Zeng, M. Zhou and Y. Chen, Chem. Rev., 2016, 116, 10346–10413.
2. I. Chakraborty and T. Pradeep, Chem. Rev., 2017, 117, 8208–8271.
3. R. Jin, G. Li, S. Sharma, Y. Li and X. Du, Chem. Rev., 2021, 121, 567–648.
4. S. Y. Qian, Z. P. Wang, Z. X. Zuo, X. M. Wang, Q. Wan and X. Yuan, Coord. Chem. Rev., 2022, 451, 214268.
5. Y. Xiao, Z. Wu, Q. Yao and J. Xie, Aggregate, 2021, 2, 114–132.
6. X. Wei, H. L. Shen, C. Xu, H. Li, S. Jin, X. Kang and M. Z. Zhu, Inorg. Chem., 2021, 60, 5931–5936.
7. M. R. Friedfeld, J. L. Stein, A. Ritchhart and B. M. Cossairt, Acc. Chem. Res., 2018, 51, 2803–2810.
8. R. S. Dhayal, W. E. Van Zyl and C. W. Liu, Acc. Chem. Res., 2016, 49, 86–95.
9. Q. Z. Li, B. Y. Huang, S. Yang, H. Zhang, J. S. Chai, Y. Pei and M. Z. Zhu, J. Am. Chem. Soc., 2021, 143, 15224–15232.
10. Z. Gan, J. Chen, L. Liao, H. Zhang and Z. K. Wu, J. Phys. Chem. Lett., 2018, 9, 204–208.
11. D. Crasto, S. Malola, G. Brosofsky, A. Dass and H. Häkkinen, J. Am. Chem. Soc., 2014, 136, 5000–5005.
12. W. J. Du, S. Deng, S. Chen, S. Jin, Y. Zhen, Y. Pei and M. Z. Zhu, J. Phys. Chem. Lett., 2021, 12, 6654–6660.
13. G. Li, Z. Lei and Q.-M. Wang, J. Am. Chem. Soc., 2010, 132, 17678–17679.
14. S. Jin, S. X. Wang, Y. B. Song, M. Zhou, J. Zhong, J. Zhang, A. D. Xia, Y. Pei, M. Chen, P. Li and M. Z. Zhu, J. Am. Chem. Soc., 2014, 136, 15559–15565.
15. S. Jin, S. X. Wang, L. Xiong, M. Zhou, S. Chen, W. J. Du, A. D Xia, Y. Pei and M. Z. Zhu, Chem. Mater., 2016, 28, 7905–7911.
16. Y. Zhou, L. W. Liao, S. L. Zhuang, Y. Zhao, Z. B. Gan, W. M. Gu, J. Li, H. T. Deng, N. Xia and Z. K. Wu, Angew. Chem., Int. Ed., 2021, 60, 8668–8672.
17. J.-S. Yang, Z. Han, X.-Y. Dong, P. Luo, H.-L. Mo and S.-Q. Zang, Angew. Chem., Int. Ed., 2020, 59, 11898–11902.
18. Y. Negishi, T. Nakazak, S. Malol, S. Takano, Y. Niihor, W. Kurashige, S. Yamazoe, T. Tsukuda and H. Häkkine, J. Am. Chem. Soc., 2015, 137, 1206–1212.
19. Y. W. Li and R. Jin, J. Am. Chem. Soc., 2020, 142, 13627–13644.
20. X. Kang, Y. Li, M. Zhu and R. Jin, Chem. Soc. Rev., 2020, 49, 6443–6514.
21. Y. W. Li, M. Zhou, Y. B. Song, T. Higaki, H. Wang and R. C. Jin, Nature, 2021, 594, 380–384.
22. Q. F. Yao, Z. N. Wu, Z. H. Liu, Y. Z. Lin, X. Yuan and J. P. Xie, Chem. Sci., 2021, 12, 99–127.
23. Y.-M. Su, Z. Wang, C.-H. Tung, D. Sun and S. K. Schein, J. Am. Chem. Soc., 2021, 143, 13235–13244.
24. K. Nakamae, M. Tanaka, B. Kure, T. Nakajima, Y. Ura and T. Tanase, Chem.–Eur. J., 2017, 23, 9457–9461.
25. K. Nakamae, T. Nakajima, Y. Ura, Y. Kitagawa and T. Tanase, Angew. Chem., Int. Ed., 2020, 59, 2262–2267.
26. (a) R. S. Dhayal, J.-H. Liao, Y.-R. Lin, P.-K. Liao, S. Kahlal, J.-Y. Saillard and C. W. Liu, J. Am. Chem. Soc., 2013, 135, 4704–4707; (b) J.-H. Liao, R. S. Dhayal, X. P. Wang, S. Kahlal, J.-Y. Saillard and C. W. Liu, Inorg. Chem., 2014, 53, 11140–11145.
27. R. S. Dhayal, J.-H. Liao, X. Wang, Y.-C. Liu, M.-H. Chiang, S. Kahlal, J.-Y. Saillard and C. W. Liu, Angew. Chem., Int. Ed., 2015, 54, 13604–13608.
28. R.-W. Huang, J. Yin, C. Dong, P. Maity, M. Nejib Hedhili, S. Nematulloev, B. Alamer, A. Ghosh, O. F. Mohammed and O. M. Bakr, ACS Mater. Lett., 2021, 3, 90–99.
29. T. A. Nguyen, Z. R. Jones, B. R. Goldsmith, W. R. Buratto, G. Wu, S. L. Scott and T. W. A. Hayton, J. Am. Chem. Soc., 2015, 137, 13319–13324.
30. A. L. Chen, X. Kang, S. Jin, W. J. Du, S. X. Wang and M. Z. Zhu, J. Phys. Chem. Lett., 2019, 10, 6124–6128.
31. A. J. Edwards, R. S. Dhayal, P.-K. Liao, J.-H. Liao, M.-H. Chiang, R. O. Piltz, S. Kahlal, J.-Y. Saillard and C. W. Liu, Angew. Chem., Int. Ed., 2014, 53, 7214–7218.
32. X. H. Liu, H. Shen, Y. Gao, G. C. Deng, H. W. Deng, Y.-Z. Han, B. K. Teo and N. F. Zheng, Chem. Commun., 2022, 58, 7670–7673.
33. Q. L. Guo, B.-L. Han, C.-F. Sun, Z. Wang, Y. W. Tao, J.-Q. Lin, G.-G. Luo, C.-H. Tung and D. Sun, Nanoscale, 2021, 13, 19642–19649.
34. R. Kumar Barik, S.-C. Huo, C.-Y. Wu, T.-H. Chiu, J.-H. Liao, X. P. Wang, S. Kahlal, J.-Y. Saillard and C. W. Liu, Chem.–Eur. J., 2020, 26, 10471–10479.
35. R. S. Dhayal, J.-H. Liao, S. Kahlal, X. Wang, Y.-C. Liu, M.-H. Chiang, W. E. van Zyl, J.-Y. Saillard and C. W. Liu, Chem.–Eur. J., 2015, 21, 8369–8374.
36. R. S. Dhayal, H.-P. Chen, J.-H. Liao, W. E. van Zyl and C. W. Liu, ChemistrySelect, 2018, 3, 3603–3610.
37. P. Yuan, R. Chen, X. Zhang, F. Chen, J. Yan, C. Sun, D. Ou, J. Peng, S. Lin, Z. Tang, B. K. Teo, L. S. Zheng and N. F. Zheng, Angew. Chem., Int. Ed., 2019, 58, 835–839.
38. A. Ghosh, R.-W. Huang, B. Alamer, E. Abou-Hamad, M. Nejib Hedhili, O. F. Mohammed and O. M. Bakr, ACS Mater. Lett., 2019, 1, 297–302.
39. R. W. Huang, J. Yin, C. Dong, A. Ghosh, M. J. Alhilaly, X. Dong, M. Hedhili, E. Abou-Hamad, B. Alamer, S. Nematulloev, Y. Han, O. Mohammed and O. M. Bakr, J. Am. Chem. Soc., 2020, 142, 8696–8705.
40. A. Baghdasaryana and T. Bürgi, Nanoscale, 2021, 13, 6283–6340.
41. A. Sagadevan, A. Ghosh, P. Maity, O. F. Mohammed, O. M. Bakr and M. Rueping, J. Am. Chem. Soc., 2022, 144, 12052–12061.
42. Q. Tang, Y. J. Lee, D.-Y. Li, W. Choi, C. W. Liu, D. Lee and D.-E. Jiang, J. Am. Chem. Soc., 2017, 139, 9728–9736.
43. C.-Y. Liu, S.-F. Yuan, S. Wang, Z.-J. Guan, D.-E. Jiang and Q.-M. Wang, Nat. Commun., 2022, 13, 2082.
44. P.-K. Liao, C.-S. Fang, A. Edwards, S. Kahlal, J.-Y. Saillard and C. W. Liu, Inorg. Chem., 2012, 51, 6577–6591.
45. S. Lee, M. S. Bootharaju, G. Deng, S. Malola, W. Baek, H. Häkkinen, N. F. Zheng and T. Hyeon, J. Am. Chem. Soc., 2020, 142, 13974–13981.
46. S. Lee, M. S. Bootharaju, G. Deng, S. Malola, H. Häkkine, N. F. Zheng and T. Hyeon, J. Am. Chem. Soc., 2021, 143, 12100–12107.
47. C. W. Dong, R. Wu Huang, C. L. Chen, J. Chen, S. Nematulloev, X. R. Guo, A. Ghosh, B. Alamer, M. N. Hedhili, T. T. Isimjan, Y. Han, O. F. Mohammed and O. M. Bakr, J. Am. Chem. Soc., 2021, 143, 11026–11035.
48. R. P. B. Silalahi, Q. Wang, J.-H. Liao, T.-H. Chiu, Y.-Y. Wu, X. P. Wang, S. Kahlal, J.-Y. Saillard and C. W. Liu, Angew. Chem., Int. Ed., 2022, 61, e202113266.
49. K. K. Chakrahari, J. P. Liao, R. P. Brocha Silalahi, T.-H. Chiu, J.-H. Liao, X. P. Wang, S. Kahlal, J.-Y. Saillard and C. W. Liu, Small, 2021, 17, 2002544.
50. X. Kang, M. Zhou, S. X. Wang, G. D. Sun, M. Z. Zhu and R. C. Jin, Chem. Sci., 2017, 8, 2581–2587.
51. F. Hu, J.-J. Li, Z.-J. Guan, S.-F. Yuan and Q.-M. Wang, Angew. Chem., Int. Ed., 2020, 59, 5312–5315.
52. X. J. Zou, Y. Lv, X. Kang, H. Z. Yu, S. Jin and M. Z. Zhu, Inorg. Chem., 2021, 60, 14803–14809.
53. S. F. Yuan, P. Li, Q. Tang, X.-K. Wan, Z.-A. Nan, D.-E. Jiang and Q.-M. Wang, Nanoscale, 2017, 9, 11405–11409.
54. K. K. Chakrahari, J.-H. Liao, S. Kahlal, Y.-C. Liu, M.-H. Chiang, J.-Y. Saillard and C. W. Liu, Angew. Chem., Int. Ed., 2016, 55, 14704–14708.
55. K. K. Chakrahari, R. P. Brocha Silalahi, J.-H. Liao, S. Kahlal, Y.-C. Liu, J.-F. Lee, M.-H. Chiang, J.-Y. Saillard and C. W. Liu, Chem. Sci., 2018, 9, 6785–6795.
56. Z. B. Gan, J. S. Chen, J. Wang, C. M. Wang, M.-B. Li, C. H. Yao, S. L. Zhuang, A. Xu, L. L. Li and Z. K. Wu, Nat. Commun., 2017, 8, 14739.
57. P. Liu and E. J. Hensen, J. Am. Chem. Soc., 2013, 135, 14032–14035.
58. S. Wang, Z. Wu, S. Dai and D.-E. Jiang, Angew. Chem., Int. Ed., 2021, 60, 12289–12292.
59. C. Sun, N. Mammen, S. Kaappa, P. Yuan, G. Deng, C. Zhao, J. Yan, S. Malola, K. Honkala, H. Häkkinen, B. K. Teo and N. F. Zheng, ACS Nano, 2019, 13, 5975–5986.
60. J. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
61. B. Delley, J. Chem. Phys., 1990, 92, 508–517.
62. B. Delley, J. Chem. Phys., 2000, 113, 7756–7764.

---

Footnotes

† Electronic supplementary information (ESI) available: Details of the synthesis process, characterization, and X-ray analysis, and Fig. S1–S30 and Tables S1–S6 offer more details on the nanoclusters. CCDC 2109287, 2109289, and 2166571. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d2sc03239b

‡ These authors contributed equally.

---