Rational construction of a library of M₂₉ nanoclusters from monometallic to tetrametallic

Significance

Precise alloying in nanoparticles with the structure retained has long been a major challenge. In addition, little has been achieved for alloy manipulation at the atomic level. Although several structures of bimetallic nanoclusters have been reported, the monodispersity of dopants remains challenging, not to mention the tri- or multimetallic nanoclusters. Here we present a maneuverable nanosystem based on an M₂₉ (M = Au/Ag/Pt/Pd/Cu) nanocluster with the tetra-stratified configuration M₁(center)@M₁₂(first shell)@M₁₂(SR)₁₈(second shell)@(M-PPh₃)₄(vertex). A rich library of 21 nanoclusters ranging from monometallic to tetrametallic compositions has been rationally constructed. The synergetic effects on optical properties and stability have been explicitly mapped out. Our methodology provides fundamental principles toward controllable alloying at the nanoscale with multimetallic compositions and atomically monodispersed dopants.

Abstract

Exploring intermetallic synergy has allowed a series of alloy nanoparticles with prominent chemical–physical properties to be produced. However, precise alloying based on a maintained template has long been a challenging pursuit, and little has been achieved for manipulation at the atomic level. Here, a nanosystem based on M₂₉(S-Adm)₁₈(PPh₃)₄ (where S-Adm is the adamantane mercaptan and M is Ag/Cu/Au/Pt/Pd) has been established, which leads to the atomically precise operation on each site in this M₂₉ template. Specifically, a library of 21 species of nanoclusters ranging from monometallic to tetrametallic constitutions has been successfully prepared step by step with in situ synthesis, target metal-exchange, and forced metal-exchange methods. More importantly, owing to the monodispersity of each nanocluster in this M₂₉ library, the synergetic effects on the optical properties and stability have been mapped out. This nanocluster methodology not only provides fundamental principles to produce alloy nanoclusters with multimetallic compositions and monodispersed dopants but also provides an intriguing nanomodel that enables us to grasp the intermetallic synergy at the atomic level.

It has been ca. 5,000 y since the Bronze Age in which “alloying” was first practiced. Alloying has indeed become the most important strategy in tailoring the physical/chemical performances of metal materials. In recent decades, colloidal nanoparticles have been of tremendous academic and industrial interest due to their extraordinary catalytic, optical, magnetic, and electrochemical properties, because alloying of metal nanoparticles with heterometals has proven to be a versatile strategy for improving the nanoparticle performances (1–5). In general, owing to the synergistic effect, alloys often display enhanced properties relative to the monometallic homologs, which largely broadens the applications of such nanomaterials (1–5). However, a detailed understanding of how the synergistic effect arises remains elusive for 2 main reasons: 1) nanoparticles that are uniform at the atomic level are difficult to prepare (usually, a distribution of size was obtained) and 2) their surface chemistry (e.g., metal–ligand interactions) is difficult to study (6). Nanochemists are often frustrated by the well-known fact that no 2 nanoparticles are the same, which precludes the fundamental study on many properties of colloidal nanoparticles in which the total structures must be known. Besides, the atomic-level tailoring of specific sites of nanoparticles with specific numbers of heterometals remains so far the least feasible. Atomic-level understanding of the synergism and atomically precise control over alloy compositions require precise molecular entities to serve as model nanosystems and precise molecular tools.

Atomically precise metal nanoclusters with ultrasmall sizes (e.g., <2-nm diameter of the metallic core) provide an exciting opportunity for investigating structure–property correlations at the atomic level, owing to their uniform size and precise structure (6–23). The quantum size effects of nanoclusters also lead to unprecedented intermetallic quantum synergisms (24–34), and such knowledge will serve as the key to fine tailoring of the physical/chemical performances of alloy nanoparticles. As to the composition/structure of alloy nanoclusters, due to the differences in size and electronegativity each type of heteroatom may prefer different site(s) in the parent nanocluster (35, 36). Taking the most researched Au₂₅(SR)₁₈ as an example, the Ag dopants are preferentially doped at the icosahedral shell, whereas the Cu dopants go to the staple motifs, albeit Cu and Ag are in the same group of gold (37, 38). By comparison, a single Pt or Pd heteroatom can be doped in the Au₂₅ nanocluster template, namely M₁Au₂₄(SR)₁₈, wherein the Pt or Pd dopant occupies the central position (39, 40). However, monodispersity of the dopants such as Ag and Cu remains extremely challenging in the M_xAu_25-x(SR)₁₈ (where x is often a range of Ag or Cu heteroatoms), and this phenomenon extends to other cluster templates. The attainment of monodispersity (i.e., x being a definite number, rather than a range) from bimetallic to multimetallic nanoclusters with the preserved structure is highly desirable for revealing the precise composition–property correlations, and such correlations will allow the controllable syntheses of new alloy nanomaterials with novel structure/composition and tailored performance.

In this work, we exploit the nanocluster system M₂₉(S-Adm)₁₈(PPh₃)₄ (where M = Ag/Au/Cu/Pt/Pd and S-Adm is adamantanethiolate) to prepare a rich library of multimetallic nanoclusters, with which the atomic-level tailoring has been accomplished in terms of the doping sites, the heterometal types, and the alloying contents. The M₂₉ template possesses a tetra-stratified configuration: M(center)@M₁₂(first shell)@M₁₂(SR)₁₈(second shell)@(Ag-PPh₃)₄(vertex). In particular, owing to the accessibility of each site with multichoices of metals, a rich library of 21 species of nanoclusters ranging from monometallic to tetrametallic compositions has been successfully prepared step by step via the combination of the in situ synthesis, the targeted metal exchange, and the forced metal exchange. The atomic monodispersity of each nanocluster in this M₂₉ library has been characterized by electrospray ionization mass spectrometry (ESI-MS) analysis. Among them, Pt₁Ag₁₂Cu₁₂Au₄(SR)₁₈(PPh₃)₄ and Pd₁Ag₁₂Cu₁₂Au₄(SR)₁₈(PPh₃)₄ are tetrametallic nanoclusters with atomic monodispersity. More importantly, the optical properties and thermal stability of these M₂₉ nanoclusters have been compared, and the synergetic effects on these properties have been mapped out at the atomic level. Overall, this M₂₉ (M = Ag/Au/Cu/Pt/Pd) nanosystem offers an ideal platform for unprecedented systematic evaluation of the synergistic effects, which provides impetus for future experimental and theoretical developments on controllable alloy nanomaterials with enhanced performance.

Results and Discussion

The M₂₉ Alloy Library.

A nanocluster methodology involving a stepwise approach (in situ synthesis–targeted metal exchange–forced metal exchange) has been exploited to alloy the monometallic Ag₂₉ nanocluster with different metals and to form a rich library of multimetallic nanoclusters with a unified configuration as M₁(kernel)@M₁₂(first shell)@M₁₂(SR)₁₈(second shell)@(M-PPh₃)₄(vertex) (M = Ag/Au/Pd/Pt/Cu) (Fig. 1). Specifically, bimetallic M₁Ag₂₈(SR)₁₈(PPh₃)₄ (M = Au/Pt/Pd) nanoclusters are all composed of 28 silver atoms and a single heteroatom. From each bimetallic M₁Ag₂₈ nanocluster, 4 trimetallic nanoclusters with different second-shell@vertex compositions have been prepared, for example from Pt₁Ag₂₈ to Pt₁Ag₂₄Cu₄, Pt₁Ag₂₄Au₄, Pt₁Ag₁₂Cu₁₆, and Pt₁Ag₁₆Cu₁₂. Note that the central-Pt/Pd/Au atom as well as the Ag₁₂ first shell in such bimetallic nanoclusters is retained in the subsequent alloying processes. Furthermore, the Au-doping operation (i.e., the forced metal-exchange method, discussed below) on the (Pt/Pd)₁Ag₁₂Cu₁₆ trimetallic nanoclusters creates the corresponding tetrametallic nanoclusters that follow an arrangement of tetra-stratified Pt/Pd(center)@Ag₁₂(first shell)@Cu₁₂(SR)₁₈(second shell)@(AuPPh₃)₄(vertex). Each nanocluster (except Pt/Pd doping) adopts a “+3” charge state, while the M₂₉ nanoclusters containing a single Pt or Pd heteroatom are in a “+2” charge state. Since the central Pt/Pd atom contributes no free valence electron, the nominal electron count of each nanocluster in this M₂₉ library is 8e.

Fig. 1.

The library of M₂₉ nanoclusters. (A) Structure of the tetrametallic Pd₁Ag₁₂Cu₁₂Au₄(S-Adm)₁₈(PPh₃)₄ nanocluster. (B) Structural anatomy of the center, first shell, second shell, and vertex. (C) Schematic illustration of the stepwise alloying process from monometallic to bi-, tri-, and tetrametallic nanoclusters on the basis of the M₂₉ template. The pink, orange, blue, green, and cerulean arrows indicate the alloying pathways of doping Pt, Au, Pd, Cu, and Ag, respectively. The gray, blue, orange, and purple backgrounds indicate the mono-, bi-, tri-, and tetrametallic compositions of each nanocluster, respectively. For clarity, the ligand parts (SR and PPh₃) are omitted in labels (the same below).

Controlling the Central Atom in the M₂₉ Template.

The parent Ag₂₉ and single-heteroatom doped M₁Ag₂₈ (M = Au/Pt/Pd) nanoclusters have been successfully prepared via an in situ synthetic procedure. The optical absorption, emission, and ESI-MS results of the obtained Pt₁Ag₂₈ are the same as those reported previously (41), where the single-Pt atom occupies the innermost center of the face-centered (fcc) Pt₁Ag₁₂ core (Fig. 2; also see SI Appendix, Fig. S1 for the structural anatomy of the M₁Ag₂₈ nanoclusters). Similarly, the central occupation of the single heteroatom has been observed in the crystal structure of Au₁Ag₂₈(SR)₁₈(PPh₃)₄. Considering 1) the homolog of these M₁Ag₂₈(SR)₁₈(PPh₃)₄ nanoclusters, 2) the central occupying tendency of the Pd heteroatom in the M₁₃ kernel (25, 40), and 3) the same Pd X-ray photoelectron spectroscopy (XPS) signal between Pd₁Ag₂₈(SR)₁₈(PPh₃)₄ and Pd₁Ag₂₄(S-PhMe₂)₁₈ (SI Appendix, Fig. S2), we thus propose the similar construction of the Pd₁Ag₂₈ with the determined structures of Pt₁Ag₂₈ and Au₁Ag₂₈ nanoclusters. In the fcc M₁Ag₁₂ core, the coordination number of the innermost metal is 12, the same as that in the fcc crystal lattice of the bulk Au/Ag/Pt/Pt or their alloys.

Fig. 2.

M₂₉ nanoclusters from monometallic to bimetallic. Structures and ESI-MS spectra of (A) Ag₂₉ and its central doped (B) Au₁Ag₂₈, (C) Pt₁Ag₂₈, and (D) Pd₁Ag₂₈ nanoclusters. (Insets) Experimental and simulated isotope patterns of each nanocluster. Color codes: cerulean sphere, Ag; orange sphere, Au; pink sphere, Pt; blue sphere, Pd; yellow sphere, S; purple sphere, P. For clarity, the carbon and hydrogen atoms are not shown.

ESI-MS was performed to validate the monodispersity of each nanocluster (Fig. 2, Bottom). First of all, the ESI-MS of each purified product shows a single intense peak, demonstrating the monodispersity of each nanocluster. In addition, the spacing of the mass peaks indicates a “3+” charge state of Ag₂₉ and Au₁Ag₂₈ nanoclusters but a “2+” charge state for the single Pt or Pd-doped M₁Ag₂₈, indicating no free valence electron contribution from the Pt or Pd dopant. Accordingly, these 4 nanoclusters all represent an 8-electron closed-shell electronic configuration (i.e., 29-18-3 = 8e for Ag₂₉ and Au₁Ag₂₈ and 28-18-2 = 8e for Pt₁Ag₂₈ and Pd₁Ag₂₈).

Controlling the Shell Atoms in the M₂₉ Template.

Transforming the M-Ag (M = Au/Pt/Pd) “precursor” into the M-Ag-Cu results in the generation of trimetallic M₁Ag_16-xCu_12+x nanoclusters with polydispersed x = 0–4. Considering that the number of the alternative occupying sites (i.e., 4) is the same as that of the vertex Ag-PPh₃ units in the Ag₂₉ template, these vertex sites are proposed to be the Ag-to-Cu exchange locations; in other words, the Ag₁₂(first shell)@Cu₁₂(second shell) constructions in such nanoclusters are nonswappable. Density functional theory (DFT) calculations also demonstrated that the metal-exchange location is the vertex, instead of sites on the second shell (SI Appendix, Figs. S3 and S4). Besides, by summing up the previous alloy nanoclusters that containing the Au/Ag and Cu metals (42–44), it is suggested that the first shell in this M₂₉ template is totally occupied by the 12 Ag atoms, and the second-shell metal is Cu (such arrangements have also been confirmed by the crystal structure of Pt₁Ag₁₂Cu₁₆(SR)₁₈(PPh₃)₄). Taking the Pt-centered nanosystem as an example (Fig. 3, Left), the ESI-MS spectrum of the in situ-prepared product displays 5 peaks centered at 3283.3, 3305.3, 3327.3, 3349.8, and 3371.8 Da, corresponding to the Pt₁Ag₁₂Cu₁₆(SR)₁₈(PPh₃)₄, Pt₁Ag₁₃Cu₁₅(SR)₁₈(PPh₃)₄, Pt₁Ag₁₄Cu₁₄(SR)₁₈(PPh₃)₄, Pt₁Ag₁₅Cu₁₃(SR)₁₈(PPh₃)₄, and Pt₁Ag₁₆Cu₁₂(SR)₁₈(PPh₃)₄, respectively (SI Appendix, Figs. S5 and S6). The mass spacing of adjacent peaks is 22 Da, which matches the expected gap between Ag and Cu atoms, that is, [M(Ag)-M(Cu)]/2(charge) = 22 Da. Consequently, the trimetallic Pt@Ag@Cu nanoclusters have been successfully synthesized, albeit with polydispersed dopant numbers (Fig. 3, Left).

Fig. 3.

M₂₉ nanoclusters with trimetallic compositions. Structures and ESI-MS spectra of Pt₁Ag₁₂Cu₁₂@Ag_xCu_4-x(SR)₁₈(PPh₃)₄ (x = 0–4) and its metal-exchanged products: trimetallic Pt₁Ag₁₆Cu₁₂(SR)₁₈(PPh₃)₄ and Pt₁Ag₁₂Cu₁₆(SR)₁₈(PPh₃)₄ nanoclusters with monodispersity. (Insets) Experimental and simulated isotope patterns of each nanocluster. Color codes: cerulean sphere, Ag; pink sphere, Pt; green sphere, Cu; red sphere, mixed Ag/Cu; yellow sphere, S; purple sphere, P. Carbon and hydrogen atoms are not shown. For clarity, the chemical formula of each nanocluster is only marked with the corresponding metal composition (the same below), because all nanoclusters in this M₂₉ system follow the same configuration as M₁@M₁₂@M₁₂(SR)₁₈@(M-PPh₃)₄ (M = Ag/Au/Pd/Pt/Cu).

The metal-exchange method has been extensively exploited to alloy the parent nanocluster with a retained framework (28, 35, 36). For example, the icosahedral-shell Au atoms are substituted by Ag atoms when alloying Au₂₅(SR)₁₈ with the Ag^I-SR complex (35). Besides, the reversible process has been observed by reacting the as-alloyed Ag_xAu_25-x(SR)₁₈ with the Au^I–SR complex (35). In this context, we are motivated to “focus” the distribution of Pt₁Ag_16-xCu_12+x(SR)₁₈(PPh₃)₄ into monodispersed nanoclusters (i.e., a single x value) by way of a targeted metal-exchange method. As depicted in Fig. 3, we indeed found that reaction of the Pt₁Ag_16-xCu_12+x(SR)₁₈(PPh₃)₄ with a large amount of Ag^I(PPh₃)NO₃ complex generates the monodispersed Pt₁Ag₁₆Cu₁₂(SR)₁₈(PPh₃)₄ nanocluster, which exhibits a single mass peak centered at 3,371.8 Da in the ESI-MS spectrum. Accordingly, the substitution of vertex Cu atoms by Ag atoms occurs in the parent nanoclusters. Furthermore, the detected [Cu^I(PPh₃)₂]⁺ (centered at 587.11 Da; SI Appendix, Fig. S7; ESI-MS spectra of the starting Pt₁Ag_16-xCu_12+x(SR)₁₈(PPh₃)₄ are shown in SI Appendix, Figs. S5 and S6) also validates the targeted metal-exchange process, which can be summarized as

$${\text{Pt}}_1{\text{Ag}}_{12}{\text{Cu}}_{12}{\left(\text{SR}\right)}_{18}{@\text{Ag}}_{\text{x}}{\text{Cu}}_{4-\text{x}}{\left({\text{PPh}}_3\right)}_4+\left(4-x\right){\hspace{0.17em}\text{Ag}}^{\text{I}}\left({\text{PPh}}_3\right){\text{NO}}_3\to {\text{Pt}}_1{\text{Ag}}_1{\text{Cu}}_{12}{\left(\text{SR}\right)}_{18}{@\text{Ag}}_4{\left({\text{PPh}}_3\right)}_4+\left(4-x\right){\hspace{0.17em}\text{Cu}}^{\text{I}}{\left({\text{PPh}}_3\right)}_2{\text{NO}}_3.$$

On the other hand, reacting the Pt₁Ag_16-xCu_12+x(SR)₁₈(PPh₃)₄ with Cu^I(PPh₃)₂Cl results in the monodispersed Pt₁Ag₁₂Cu₁₆(SR)₁₈(PPh₃)₄ nanocluster, whose composition and construction have been verified by the ESI-MS analysis and X-ray crystallography (Fig. 3, Lower Right and SI Appendix, Fig. S8). Besides, Pd-centered and Au-centered trimetallic nanoclusters have also been obtained with a procedure similar to that of the Pt-centered nanosystem (see SI Appendix, Figs. S9–S13 for the Pd-centered nanosystem and SI Appendix, Figs. S14–S18 for the Au-centered nanosystem).

Controlling the Vertex Atoms in the M₂₉ Template.

Based on the above understandings of the vertex metal atoms (bonded with the PPh₃ ligands) with exchangeable characteristic, we perceive a good opportunity to obtain the alloyed M₂₉ nanoclusters containing more than 3 types of metals. Recently, Negishi and coworkers (36) reported the synthesis of tetrametallic Pd₁Ag_xCu_yAu_24-x-y(SR)₁₈ nanoclusters by exploiting the metal-exchange rules that different metal dopants preferentially occupy different sites in the parent Au₂₅(SR)₁₈ nanocluster. However, alloying in the Au₂₅ template always results in a mixture product with distributed dopant numbers, for example Ag_xAu_25-x and Cu_xAu_25-x nanoclusters (35, 37, 38). In comparison, the M₂₉ nanosystem can act as a perfect model to obtain the multimetallic nanocluster in atomic monodispersity. As shown in Fig. 4 and SI Appendix, Figs. S19 and S20, the further metal-exchange operation on the trimetallic Pt₁Ag₁₂Cu₁₆(SR)₁₈(PPh₃)₄ has been performed by reacting the nanocluster with Au^I(PPh₃)Cl. On the one hand, the alloying outcomes follow an incompletely metal-exchanged pattern [even with the overdose of Au^I(PPh₃)Cl] and 5 peaks including Pt₁Ag₁₂Cu₁₂@Cu₄(SR)₁₈(PPh₃)₄, Pt₁Ag₁₂Cu₁₂@Cu₃Au₁(SR)₁₈(PPh₃)₄, Pt₁Ag₁₂Cu₁₂@Cu₂Au₂(SR)₁₈(PPh₃)₄, Pt₁Ag₁₂Cu₁₂@Cu₁Au₃(SR)₁₈(PPh₃)₄, and Pt₁Ag₁₂Cu₁₂@Au₄(SR)₁₈(PPh₃)₄ are detected in the ESI-MS (note that the maximum Au alloying number is 4, validating the vertex alloying mode in these M₂₉ nanosystems). On the other hand, reacting the mixture of Pt₁Ag₁₂Cu_16-xAu_x(SR)₁₈(PPh₃)₄ (x = 0–4) nanoclusters with only tiny amounts of Cu^I(PPh₃)₂Cl results in the monodispersed Pt₁Ag₁₂Cu₁₆(SR)₁₈(PPh₃)₄ nanocluster (Fig. 4). Combining the above 2 aspects, it is suggested that the bonding ability of Cu atoms in these 4 vertex sites is far stronger than that of the Au atoms. Considering that 1) the bonding mode of these vertex metals is μ₃-M (M = Cu/Au) and 2) μ₃-Cu has been reported several times (however, μ₃-Au has not been observed yet) (42–44), we speculate that the stability of such μ₃-Au is not as high as that of the μ₃-Cu in the M₂₉ nanosystem (SI Appendix, Fig. S21). The instability of vertex-Au nanoclusters has also been proved by the stability tests (discussed below) as well as DFT calculations (SI Appendix, Fig. S3).

Fig. 4.

M₂₉ nanoclusters from trimetallic to tetrametallic. Illustration of the synthetic procedures from Pt₁Ag₁₂Cu₁₆ to poly-dispersed Pt₁Ag₁₂Cu_16-xAu_x (x = 0–4), and then to monodispersed Pt₁Ag₁₂Cu₁₂Au₄ nanocluster. The blue curve represents the oxidation from Cu(I) to Cu(II) effected by H₂O₂, thus the noncoordinate ability of Cu(II) leads to the monodispersity of the final tetrametallic Pt₁Ag₁₂Cu₁₂Au₄ nanocluster. Color codes: cerulean sphere, Ag; pink sphere, Pt; orange sphere, Au; green sphere, Cu; red sphere, mixed Au/Cu; yellow sphere, S; purple sphere, P. For clarity, the carbon and hydrogen atoms are not shown.

For obtaining the tetrametallic Pt₁Ag₁₂Cu₁₂Au₄ nanocluster with monodispersity, the 4 vertex Cu atoms should be substituted by the Au dopants completely. By analyzing the ESI-MS results of metal-exchange processes from Pt₁Ag_16-xCu_12+x to Pt₁Ag₁₂Cu₁₆ (or Pt₁Ag₁₆Cu₁₂), no oxidation-reduction but just metal-exchange processes Ag(I)→Cu(I) or Cu(I)→Ag(I) occur, indicating that the vertex Cu atoms are almost in the Cu(I) valence state. By noting that adding the Cu^II(PPh₃)₂Cl₂ to the Pt₁Ag_16-xCu_12+x nanoclusters cannot convert them into monodispersed Pt₁Ag₁₂Cu₁₆ (SI Appendix, Fig. S22), we hypothesize that the Cu(II) atoms have no coordination ability to act as the μ₃-Cu at the vertex sites. Therefore, considering the competitive relationship of Au and Cu in these sites, the oxidation capability of H₂O₂ has been exploited to weaken the coordination ability of Cu atoms and this indeed allows us to prepare the monodispersed Pt₁Ag₁₂Cu₁₂Au₄ nanocluster (Fig. 4, Right). In this “forced metal-exchange” process, the addition of H₂O₂ turns the Cu(I) into the oxidized state [Cu(II), shown in SI Appendix, Fig. S23], and further addition of the Au^I(PPh₃)Cl guarantees the existence of the Au(I) that is able to occupy the vertex sites. The H₂O₂/Au^I(PPh₃)Cl addition is repeated 3 times to eliminate the Cu^I coordination completely, and then ESI-MS demonstrates the generation of the tetrametallic Pt₁Ag₁₂Cu₁₂Au₄(SR)₁₈(PPh₃)₄ with monodispersity (the Pd-centered procedure and results are in SI Appendix, Figs. S24–S26). This result is exciting since it not only produces a tetrametallic nanocluster with precise metallic occupancy but also evokes the possibility for revealing the intermetallic synergy in detail in a coherent monodispersed nanosystem from monometallic to bi-, tri-, and tetrametallic compositions. However, the obtained Au-vertex nanoclusters (Pt₁Ag₁₂Cu₁₂Au₄ or Pd₁Ag₁₂Cu₁₂Au₄) are less stable and decompose in solution (SI Appendix, Fig. S23), in agreement with the DFT results that the energy of Au-vertex nanoclusters is much higher than that of Cu-vertex nanoclusters (SI Appendix, Fig. S3).

Overview of the Unique M₂₉ Nanosystem.

As summarized in SI Appendix, Fig. S27, a unique nanosystem has been established based on the Pt-centered M₂₉ template (see SI Appendix, Fig. S28 for the Pd-centered nanosystem). With the controllable synthetic procedure (combining the in situ synthesis and metal-exchange method), the monometallic Ag₂₉ has been precisely alloyed to bimetallic Pt₁Ag₂₈, trimetallic Pt₁Ag₁₂Cu₁₆, and tetrametallic Pt₁Ag₁₂Cu₁₂Au₄ nanoclusters (SI Appendix, Fig. S27A). SI Appendix, Fig. S27B shows the structural anatomy of the Pt₁Ag₁₂Cu₁₂Au₄ nanocluster. Specifically, the single Pt atom occupies the innermost position and is further enclosed by 12 Ag atoms, constituting the Pt₁Ag₁₂ kernel. Then, four 3-fold symmetric Cu₃(S-Adm)₆ staple motifs wrap up the Pt₁Ag₁₂ kernel. Because these Cu₃(S-Adm)₆ motifs connect each other by sharing the end thiolates, a cage-like Cu₁₂(SR)₁₈ is formed. Finally, 4 vertex Au-PPh₃ units block the bare corners and the whole structure is constructed, formulated Pt₁Ag₁₂Cu₁₂Au₄(S-Adm)₁₈(PPh₃)₄ with a tetra-stratified configuration—Pt(center)@Ag₁₂(first shell)@Cu₁₂(S-Adm)₁₈(second shell)@(Ag-PPh₃)₄(vertex).

According to the aforementioned nanocluster methodology (including the in situ synthesis, targeted metal-exchange, and forced metal-exchange methods), a library of 21 species of M₂₉ (M = Ag/Cu/Pt/Pd/Au) nanoclusters has been successfully constructed. The “in situ synthesis” guarantees the single-metal occupation of the kernel, the first shell, and the second shell. Besides, a combination of “targeted metal-exchange” and “forced metal-exchange” methods guarantees the single-metal occupation of the vertex. SI Appendix, Fig. S29 generalizes the ESI-MS spectra and metal arrangements of these nanoclusters. ESI-MS results demonstrated the monodispersity of each nanocluster, which is of significance for the subsequent investigation on the metal synergistic effect at the atomic level. The monodispersity of each nanocluster was also verified by the elemental analysis and the inductively coupled plasma-atomic emission spectrometry results of these M₂₉ nanoclusters (SI Appendix, Tables S1 and S2). The NMR result of the M₂₉ nanoclusters (here we tested Pt₁Ag₂₈, Au₁Ag₂₈, and Pt₁Ag₁₂Cu₁₆ nanoclusters that had been crystallized) presented only one ³¹P peak, demonstrating the same chemical environment of the 4 PPh₃ ligands as well as the symmetry of the nanocluster’s configuration (SI Appendix, Fig. S30).

Seven subsystems have been established on the basis of the different metal arrangement, including Ag₂₅@M₄, Au₁Ag₂₄@M₄, Pt₁Ag₂₄@M₄, Pd₁Ag₂₄@M₄, Au₁Ag₁₂Cu₁₂@M₄, Pt₁Ag₁₂Cu₁₂@M₄, and Pd₁Ag₁₂Cu₁₂@M₄ subsystems. Each subsystem has the same M₁(center)@M₁₂(first shell)@M₁₂(SR)₁₈(second shell) metal configuration. For each subsystem, 3 different types of metals could be arranged (i.e., Ag, Cu, or Au) on the vertex sites, giving rise to 3 different nanocluster members in each subsystem (SI Appendix, Figs. S31–S37). Accordingly, altogether 21 nanoclusters (7 × 3 = 21) with monodispersity have been obtained for the M₂₉ library. Evidenced by the mass spectra, the nanocluster’s overall charge is “+3” when the kernel position is occupied by Ag or Au, whereas the charge is “+2” when Pt or Pd is at the innermost of the nanocluster (SI Appendix, Figs. S30–S37).

Synergy Effect on the Nanocluster Properties.

The intermetallic synergism has been experimentally and theoretically investigated based on alloy nanoclusters, and has been proved to be the decisive effect that influences the chemical and physical properties of nanoclusters (e.g., optical, catalytic, electrochemical properties, and stability) (24–26, 30, 32–34, 45–49). The available M₂₉ nanosystem now allows us to evaluate the synergistic effects in detail. Here we focus on the synergistic effects on the electronic structure of nanoclusters, and optical absorption, photoluminescence (PL), and stability.

First of all, we find that each nanocluster in the M₂₉ system possesses the 8e free-electronic structure. In this context, a comparison of the oxidation states of metals in different M₂₉ nanoclusters can help us infer the synergistic effects on electronic structures of these nanoclusters. SI Appendix, Figs. S38–S40 present the XPS results of different M₂₉ nanoclusters. From the XPS results, the substitution of Ag atoms in the second shell or vertex positions by Cu leads to the concentration of free valence electrons to the cluster’s kernel, and thus the XPS peaks of inner metals shift to the M(0) peak. Collectively, owing to the differences of the metal electronegativity, substituting the parent shell metal (Ag) with more active metal (Cu) would make the free electrons less bound to the outermost shell and concentrating to the kernel. Thus, the kernel metals would tend to present a metallic state (i.e., M⁰), and the shell metals will be in a more oxidized state (i.e., M^δ+).

SI Appendix, Figs. S41–S44 present the optical absorption and PL of the 21 nanoclusters, sorted by the metal occupation of the kernel. Some important structure-optical absorption correlations have been mapped out: 1) For the kernel, substituting the central Ag atom by a Pd heteroatom hardly changes the optical absorption, whereas doping a Pt/Au heteroatom into the kernel significantly red-shifts the absorption; 2) for the second shell, exchanging the Ag₁₂(SR)₁₈ into Cu₁₂(SR)₁₈ blue-shifts the optical absorption while maintaining their initial profiles; 3) for the vertex, the characteristic absorption profiles are almost retained when the vertex Ag atoms are substituted by Cu; by contrast, metal exchanging of these vertex atoms into Au not only red-shifts the initial absorption but also generates a new peak at the range of higher wavelength.

Interestingly, all nanoclusters in the M₂₉ nanosystem fluoresce when illuminated at 445 nm (SI Appendix, Figs. S41–S44). By comparing these PL spectra, synergy effects on PL characteristics in terms of PL intensity and emission wavelength are analyzed (Fig. 5 and SI Appendix, Figs. S45–S48). For the kernel, doping a Pd heteroatom into the center of M₂₉ weakens the PL intensity and blue-shifts the emission; on the contrary, enhanced PL intensity as well as red-shifted emission has been observed by substituting the center Ag with a Pt or Au atom (SI Appendix, Fig. S45). For the second shell, alloying the Ag₁₂(SR)₁₈ shell into Cu₁₂(SR)₁₈ blue-shifts the emission wavelength and slightly reduces the PL intensity (SI Appendix, Fig. S46). For the vertex, substituting the vertex Ag atoms by Au cannot only red-shift the emission but also significantly enhance the PL intensity; opposite phenomena have been observed in exchanging the vertex Ag atoms with Cu (SI Appendix, Fig. S47). According to the above observations, the Au₁@Ag₁₂@Ag₁₂(SR)₁₈@(Au-PPh₃)₄ displays the highest PL intensity and maximum emission wavelength (with quantum yield of 11.6% and emission of 715 nm) among the 21 nanoclusters (Figs. 5 and 6). Collectively, based on the abovementioned correlations between the metal arrangement in the M₂₉ template and PL properties, we find that controlling the nanocluster kernel or vertex metals with large electron affinity [note: metal electron affinity subsequence Au(2.309) > Pt(2.128) > Ag(1.302) > Cu(1.228) > Pd(0.562)] is in favor of preparing emissive nanoclusters with higher PL intensity. The obtained general trends of PL with different doping modes will provide guidelines for future work on fluorescent alloy nanoclusters.

Fig. 5.

Comparison of PL property of the M₂₉ nanoclusters. The rhomboid, triangular, circular, and square symbols represent the mono-, bi-, tri-, and tetrametallic nanoclusters, respectively.

Fig. 6.

(A–D) Illustration of the strategies to enhance the PL intensity of the M₂₉ nanocluster. (E) Au₅Ag₂₄(SR)₁₈(PPh₃)₄ nanocluster with the strongest PL intensity and the maximum emission wavelength, among these 21 nanoclusters of M₂₉.

Synergy effects on the thermal stability of these nanoclusters have also been investigated (SI Appendix, Figs. S48–S53). Time-dependent optical absorption of these nanoclusters (dissolved in CH₂Cl₂) was monitored at room temperature (SI Appendix, Figs. S48–S51). By comparing the variation trends of the absorptions, we conclude that 1) for the kernel, the stability sequence is Pd₁M₂₈ < Ag₁M₂₈ < Au₁M₂₈ ∼ Pt₁M₂₈; 2) for the second shell, alloying the Ag₁₂(SR)₁₈ shell into Cu₁₂(SR)₁₈ enhances the stability of nanoclusters; and 3) for the vertex, the stability sequence is M₂₅Cu₄ > M₂₅Ag₄ >> M₂₅Au₄ (SI Appendix, Figs. S52 and S53). The correlations between the metal occupation in vertex and the stability of nanoclusters agree with the energy changes of nanoclusters derived from the DFT calculations (SI Appendix, Fig. S3).

Conclusion

We have developed a methodology to prepare atomically precise alloy nanoclusters based on the M₂₉(S-Adm)₁₈(PPh₃)₄ template (M = Au/Ag/Pd/Pt/Cu) with a tetra-stratified configuration—M₁(center)@M₁₂(first shell)@M₁₂(SR)₁₈(second shell)@(M-PPh₃)₄(vertex). Owing to the easy maneuverability of each site in this M₂₉ template, a library of 21 nanoclusters with atomic monodispersity has been synthesized by exploiting the in situ synthesis, targeted metal-exchange, and forced metal-exchange methods, and these nanoclusters range from monometallic to bi-, tri-, and tetrametallic constitutions. The monodispersity of each nanocluster has been verified by the ESI-MS measurement. In addition, the precise structures of these M₂₉ nanoclusters enable us to uncover the intermetallic synergy at the atomic level, which provides guidelines for future work on alloy nanoclusters. Overall, the M₂₉ nanosystem presented in this work is of significance not only because it provides a platform to generate alloyed nanoclusters with multimetallic compositions and monodispersed dopants but also it offers atomic-level insight into the intermetallic synergy.

Methods

Synthesis of the Monometallic [Ag₂₉(S-Adm)₁₈(PPh₃)₄]³⁺.

For the nanocluster synthesis, AgNO₃ (30 mg, 0.18 mmol) was dissolved in CH₃OH (5 mL) and CH₃COOC₂H₅ (35 mL) with sonication. The solution was vigorously stirred (∼1,200 rpm) with magnetic stirring for 15 min. Then, Adm-SH (0.1 g) and PPh₃ (0.1 g) were added and the reaction was vigorously stirred (∼1,200 rpm) for another 90 min. After that, NaBH₄ (1 mL) aqueous solution (20 mg⋅mL⁻¹) was added quickly to the above mixture. The reaction was allowed to proceed for 36 h under a N₂ atmosphere. After that, the aqueous layer was removed, and the mixture in the organic phase was rotavaporated under vacuum. Then ∼15 × 3 mL of CH₃OH was used to wash the synthesized nanoclusters. The precipitate was dissolved in CH₂Cl₂, which produced the [Ag₂₉(S-Adm)₁₈(PPh₃)₄]³⁺ nanocluster. The yield is 16% based on the Ag element (calculated from the AgNO₃).

Synthesis of the Bimetallic [Pt₁Ag₂₈(S-Adm)₁₈(PPh₃)₄]²⁺.

Specifically, the metal source for synthesizing [Ag₂₉(S-Adm)₁₈(PPh₃)₄]³⁺ nanocluster (i.e., AgNO₃, 0.18 mmol) was altered to Ag/Pt mixture (AgNO₃/H₂PtCl₆·6H₂O = 0.17/0.01 mmol in the [Pt₁Ag₂₈(S-Adm)₁₈(PPh₃)₄]²⁺ synthesis) and the other conditions were not changed, then the [Pt₁Ag₂₈(S-Adm)₁₈(PPh₃)₄]²⁺ nanoclusters were obtained. The yield is 45% based on the Ag element (calculated from the AgNO₃).

Synthesis of Trimetallic [Pt₁Ag_12+xCu_16-x(S-Adm)₁₈(PPh₃)₄]²⁺ (x = 0–4).

Specifically, the metal source for synthesizing [Ag₂₉(S-Adm)₁₈(PPh₃)₄]³⁺ nanocluster (i.e., AgNO₃, 0.18 mmol) was altered to Ag/Cu/Pt mixture (AgNO₃/Cu^I(PPh₃)₂Cl/H₂PtCl₆·6H₂O = 0.1/0.07/0.01 mmol in the [Pt₁Ag_12+xCu_16-x(S-Adm)₁₈(PPh₃)₄]²⁺ synthesis) and the other conditions were not changed, then the [Pt₁Ag_12+xCu_16-x(S-Adm)₁₈(PPh₃)₄]²⁺ nanoclusters were obtained. The yield is 15% based on the Ag element (calculated from the AgNO₃).

Synthesis of Trimetallic [Pt₁Ag₁₂Cu₁₆(S-Adm)₁₈(PPh₃)₄]²⁺ and [Pt₁Ag₁₆Cu₁₂(S-Adm)₁₈(PPh₃)₄]²⁺.

A target metal-exchange method was exploited to “focus” the polydispersed [Pt₁Ag_12+xCu_16-x(S-Adm)₁₈(PPh₃)₄]²⁺ (x = 0–4) nanoclusters into the [Pt₁Ag₁₂Cu₁₆(S-Adm)₁₈(PPh₃)₄]²⁺ or [Pt₁Ag₁₆Cu₁₂(S-Adm)₁₈(PPh₃)₄]²⁺. Specifically, 0.1 mmol [Pt₁Ag_12+xCu_16-x(S-Adm)₁₈(PPh₃)₄]²⁺ was dissolved in 20 mL CH₂Cl₂; 1 mmol Cu^I(PPh₃)₂Cl was added to the above solution and the solution was further vigorously stirred (∼1,200 rpm) for 30 min. The organic phase was rotavaporated under vacuum and washed several times with CH₃OH. The precipitate was dissolved in CH₂Cl₂, which produced pure [Pt₁Ag₁₂Cu₁₆(S-Adm)₁₈(PPh₃)₄]²⁺. The yield is 80% based on the [Pt₁Ag_12+xCu_16-x(S-Adm)₁₈(PPh₃)₄]²⁺ nanoclusters. Changing the Cu^I(PPh₃)Cl into the Ag^I(PPh₃)NO₃ generated pure [Pt₁Ag₁₆Cu₁₂(S-Adm)₁₈(PPh₃)₄]²⁺ with a 70% yield.

Synthesis of Tetrametallic [Pt₁Ag₁₂Cu_16-xAu_x(S-Adm)₁₈(PPh₃)₄]²⁺ (x = 0–4).

A “target metal-exchange” method was exploited to alloy the trimetallic [Pt₁Ag₁₂Cu₁₆(S-Adm)₁₈(PPh₃)₄]²⁺ nanocluster into the tetrametallic [Pt₁Ag₁₂Cu_16-xAu_x(S-Adm)₁₈(PPh₃)₄]²⁺ (x = 0–4) nanoclusters. Specifically, 0.1 mmol [Pt₁Ag₁₂Cu₁₆(S-Adm)₁₈(PPh₃)₄]²⁺ was dissolved in 20 mL CH₂Cl₂; 0.5 mmol Au^I(PPh₃)Cl was added to the above solution and the solution was further vigorously stirred (∼1,200 rpm) for 60 min. The organic phase was rotavaporated under vacuum. The precipitate was dissolved in CH₂Cl₂, which produced the tetrametallic [Pt₁Ag₁₂Cu_16-xAu_x(S-Adm)₁₈(PPh₃)₄]²⁺ nanocluster. The yield is 85% based on the [Pt₁Ag₁₂Cu₁₆(S-Adm)₁₈(PPh₃)₄]²⁺.

Synthesis of Tetrametallic [Pt₁Ag₁₂Cu₁₂Au₄(S-Adm)₁₈(PPh₃)₄]²⁺.

The “forced metal-exchange” method was exploited to alloy the polydispersed [Pt₁Ag₁₂Cu_16-xAu_x(S-Adm)₁₈(PPh₃)₄]²⁺ (x = 0–4) nanoclusters into the monodispersed [Pt₁Ag₁₂Cu₁₂Au₄(S-Adm)₁₈(PPh₃)₄]²⁺ nanocluster. Specifically, 0.1 mmol [Pt₁Ag₁₂Cu_16-xAu_x(S-Adm)₁₈(PPh₃)₄]²⁺ was dissolved in 20 mL CH₂Cl₂; 200 μL H₂O₂ was added and the solution was further vigorously stirred (∼1,200 rpm) for 3 min. Then, 0.5 mmol Au^I(PPh₃)Cl was added to the above solution and further vigorously stirred for 30 min. The H₂O₂@Au^I(PPh₃)Cl addition was repeated 3 times to eliminate the Cu^I coordination completely. The organic phase was then precipitated with a large amount of CH₃OH. Finally, the precipitate was dissolved in CH₂Cl₂, which produced the monodispersed [Pt₁Ag₁₂Cu₁₆Au₄(S-Adm)₁₈(PPh₃)₄]²⁺ nanocluster. The yield is 20% based on the [Pt₁Ag₁₂Cu_16-xAu_x(S-Adm)₁₈(PPh₃)₄]²⁺.

Crystallization of Au₁Ag₂₈(S-Adm)₁₈(PPh₃)₄ and Pt₁Ag₁₂Cu₁₆(S-Adm)₁₈(PPh₃)₄.

Single crystals of Au₁Ag₂₈(S-Adm)₁₈(PPh₃)₄ and Pt₁Ag₁₂Cu₁₆(S-Adm)₁₈(PPh₃)₄ nanoclusters were grown at room temperature for 7 d in CH₂Cl₂/CH₃OH. Then, red crystals were collected and the structures of Au₁Ag₂₈(S-Adm)₁₈(PPh₃)₄ and Pt₁Ag₁₂Cu₁₆(S-Adm)₁₈(PPh₃)₄ were determined by X-ray crystallography.

Other than the nanoclusters in the Pt-centered system, details of synthesis of other cases are provided in SI Appendix.