Gas-phase studies of chemically synthesized Au and Ag clusters

Metal clusters composed of fewer than a few hundred atoms are located in terms of size between the nanoparticles and atoms of the corresponding metal (Scheme 1) and have attracted the attention of scientists over the last four decades.¹ The central issue since the early stage of research has been observing the finite-size effects on the physical properties of metal clusters and understanding their origins from a microscopic viewpoint. Versatile and efficient production methods for naked clusters (laser vaporization^2,3 and magnetron sputtering⁴ methods) and ultrasensitive characterization methods coupled with mass spectrometry have enabled us to address these fundamental questions.⁵ Magic numbers have been searched for by mass spectrometry.⁵ Electronic structures have been studied by photoelectron spectroscopy (PES)^6–11 and photodissociation (PD) spectroscopy,¹² whereas geometric structures (overall motif and atomic packing) have been studied by ion mobility mass spectrometry (IM MS),^13,14 electron diffraction (ED),¹⁵ and vibrational spectroscopy.¹⁶ Chemical properties and binding strength have been studied by collision-induced dissociation/reaction mass spectrometry (CID/CIR-MS).^17,18 These experimental results together with theoretical results have led to the discovery of a variety of remarkable size-specific phenomena and physicochemical properties. For example, observations of magic numbers in the size distributions have led to the concept of the electron shell model based on the jellium model.¹⁹ It is widely recognized that various physicochemical properties of metal clusters deviate significantly from those of their bulk counterparts due to the unique geometric and electronic structures (Scheme 1) and evolve dramatically as a function of size, as exemplified by the metal–insulator transition.²⁰ Rapid progress in the research under the catchphrases “small is different”^21,22 and “every atom counts”^23,24 has convinced the community that metal clusters are promising functional units of novel materials.

The chemical synthesis of metal clusters has been an interesting challenge in materials science on the nanoscale, as evidenced by the explosive growth in the materials science of nanocarbons after the large-scale production of C₆₀ by Smalley and Krätschmer.²⁵ Gold clusters have provided ideal platforms to tackle this challenge because of their robustness against oxidation under ambient conditions. The first requirement for the chemical synthesis is to stabilize individual Au clusters against aggregation so that they can be treated as conventional chemical compounds. This is achieved by passivation of the Au cluster surface with organic ligands L to yield monolayer-protected Au clusters (Au MPCs) with the formula of [Au_xL_y]^z or by decorating with polymers. The ligands include thiolates (SR),²⁶ alkynyls (C≡CR),²⁷ phosphines (PR₃),^28, N-heterocyclic carbenes (NHC),²⁹ and halides (X), whereas the polymers include polyvinylpyrrolidone (PVP). The second requirement is to define the chemical formula of [Au_xL_y]^z with atomic and molecular precision. Figure 1(1) shows the general approach for the atomically precise synthesis of Au MPCs. Rapid reduction of Au(I)L_n complexes produces an atomically polydispersed mixture of [Au_xL_y]^z. Subsequent etching by an excess amount of L transforms metastable clusters into thermodynamically and chemically stable clusters (size focusing). Then, fractionation using electrophoresis, chromatography, or reprecipitation yields atomically precise [Au_xL_y]^z. By using this approach, a large variety of Au/Ag-based MPCs [(Au/Ag)_xL_y]^z with atomically defined sizes in the range of x < ∼500 have been synthesized.^30–33 The nonbulk-like atomic packing of the Au cores has been elucidated by single-crystal x-ray diffraction (SCXRD) analysis and transmission electron microscopy.³⁴ The discrete nature of the electronic structures has been demonstrated by UV–vis absorption spectroscopy, photoluminescence spectroscopy, and voltammetry.

Various gas-phase methods can be applied to Au/Ag MPCs by introducing them in the gas phase using electrospray ionization (ESI) or the matrix-assisted laser desorption/ionization (MALDI) method [Figs. 1(2) and 1(3)]. These studies will provide novel and complementary information on their intrinsic structures, structures that cannot be obtained by conventional characterization methods. This Perspective summarizes the recent progress and discusses the future prospects of the gas-phase studies on Au/Ag MPCs.

Early mass spectrometric studies using laser-desorption ionization by Whetten have revealed the existence of a series of magic sizes in thiolate-protected Au clusters Au_x(SR)_y.^26,35–37 Later mass spectrometric studies by Tsukuda and Murray identified the currently well-known magic clusters [Au₂₅(SR)₁₈]⁻, Au₃₈(SR)₂₄, and Au₁₄₄(SR)₆₀.^38–40 Recently, nearly atomically precise synthesis up to Au_∼2000(SR)_∼290 was achieved by Dass.⁴¹ Currently, mass spectrometry plays an essential role in studies on synthesizing [(Au/Ag)_xL_y]^z: (1) determination of the chemical formula including net charge (x, y, z), which are the key descriptors of the clusters and (2) detection and identification of intermediate species transiently formed in solution during the formation or transformation reactions of the clusters. The basic requirement for MS is to ionize the chemically synthesized clusters in the intact form for which conventional methods such as ESI³⁸ and MALDI⁴² have been used. [(Au/Ag)_xL_y]^z intrinsically charged (z ≠ 0) in solution are directly introduced into the mass spectrometer as a continuous beam by desolvation in the ESI source. The ESI method can be applied to neutral clusters (z = 0) either by protonation or deprotonation of the ligands before ESI or attachment of cationic species such as Cs⁺ during ESI. The ESI method also allows us to sample intermediate and transient species nascently formed in the solution. In contrast, a pulsed beam of [(Au/Ag)_xL_y]^z can be generated by the MALDI method: solidified cluster samples with a matrix (most typically DCTB)^42,43 are irradiated with a pulsed laser light. The MALDI method is also used to desorb Au clusters stabilized by the polymer.⁴⁴ Portions of the continuous ion beam from the ESI source or the pulsed ion beam from the MALDI source are injected into the time-of-flight (TOF) mass spectrometer by applying a pulsed electric field. The ESI/MALDI MS also acts as an interface for the gas-phase measurements listed in Fig. 1(3). This section showcases typical examples of mass spectrometric characterization of Au/Ag MPCs and mass spectrometric detection of reaction intermediates.

The primary purpose of mass spectrometry is to determine the chemical formulas (x, y) and charge states (z) of the synthesized [(Au/Ag)_xL_y]^z, as exemplified in Fig. 2. Figure 2(a) shows the first ESI mass spectrum of the glutathione (GSH)-protected Au₂₅ cluster, [Au₂₅(SG)_18−x(SG⁻)_x]⁻, which is identified as the currently well-known magic composition.³⁸ The same composition was observed using other thiolates such as [Au₂₅(SC₂H₄Ph)₁₈]⁻ [Fig. 2(b)]⁴² and Ag analog [Ag₂₅(SC₆H₃Me₂)₁₈]⁻ [Fig. 2(c)].^45, Figure 2(d) illustrates that the composition of thiolate-protected Au nanoparticles with the diameter of 2.9 nm can be determined with the accuracy of Au_940±20(SC₂H₄Ph)_160±4.⁴⁶ The MALDI mass spectra in Fig. 2(e) demonstrate that pure and single atom doped Au₃₄ magic clusters can be synthesized using PVP as a stabilizer.^47,48

Another important application of MS is the detection of transient species produced during complex processes in solution, which will provide mechanistic insights at the molecular level.^67–74, In situ ESI MS has been applied to a variety of solution processes. For example, Xie detected 29 intermediate species during the formation of [Au₂₅(SG)₁₈]⁻ by slow reduction of Au(I) precursors with CO.⁶⁷ A 2e⁻ reduction growth mechanism was proposed based on the sequential appearance of intermediates with even-numbered valence electrons. In spontaneous alloying between [Au₂₅(SR)₁₈]⁻ (7) and [Ag₂₅(SR′)₁₈]⁻ (11),⁷⁰ Pradeep elucidated the formation of dianionic species [Ag₂₅Au₂₅(SR)₁₈(SR′)₁₈]²⁻ at the initial stage [Fig. 3]. In situ ESI-MS has been also successfully applied to the seed-mediated growth of [Au₂₅(SR)₁₈]⁻ (7) to [Au₄₄(SR)₂₆]⁻,⁷¹ the alloying process between [Au₂(SR)₂Cl]⁻ and [Ag₄₄(SR′)₃₀]⁴⁻ (14),⁷² the ligand-exchange induced size transformation of Au₃₈(SC₂H₄Ph)₂₄ to Au₃₆(SC₆H₄t-Bu)₂₄,⁷⁵ and the intercluster reaction between [Au₂₅(SR)₁₈]⁻ (7) and [Ag₄₄(SR′)₃₀]⁴⁻ (14).⁶⁸ 

Mass analysis of fragmentation pathways of Au/Ag MPCs provides molecular-level information on the thermal stability and the relative binding affinities of different ligands with respect to the clusters. There are three methods to study dissociation of Au/Ag MPCs in the gas phase [Fig. 1(3)]: collision-induced dissociation mass spectrometry (CID-MS), photodissociation mass spectrometry (PD-MS), and surface-induced dissociation mass spectrometry (SID-MS). In CID-MS, the dissociation is induced by the collision with the atmospheric molecules in an ESI source or in a gas cell. In PD-MS, Au/Ag MPCs undergo dissociation via the electronically excited state upon photoabsorption of UV-vis light. A unique feature of PD-MS is that the energy deposited to the clusters is tuned precisely and can be larger than that by CID. As a result, the PD with the UV light results in richer fragmentation patterns compared to those by the CID method. In SID-MS, Au/Ag MPCs are allowed to collide with the solid surface. Since it has been demonstrated that ∼10% of the collisional energy can be deposited into a projectile by collision with a solid surface on a timescale of femtoseconds,⁷⁷ collision with the surface can impart much larger energy than that by PD and CID and promote further fragmentation.

A typical setup is schematically shown in Fig. 1(3). The continuous beam of chemically purified Au/Ag MPC ions from the ESI source is introduced into a differentially pumped area in which the typical pressure is ∼100 Pa. Isolated Au/Ag MPC ions undergo dissociation upon collision with the buffer gas. The voltage applied to the electrodes in the differentially pumped area [V_CID in Fig. 1(2)] is adjusted to control the nominal collision energy of Au/Ag MPC ions with the background gas: the collision energy increases with V_CID.

Table III lists the formulas and the n^* values of fragment ions and small fragments observed in CID processes of representative Au/Ag MPCs. In the CID of [Au₁₁(PR₃)₈X₂]⁺ (3), the loss of the Au(PR₃)X units competes with the loss of PR₃ ligands [Fig. 5(a), entry 1].⁷⁸ The numbers of released PR₃ and Au(PR₃)X units increased with the increase in the collision energy [Fig. 5(a)],while retaining the n^* values of the fragment ions to be 8. In the CID of [Au₂₅(SR)₁₈]⁻ (7), [Au₂₁(SR)₁₄]⁻ was formed as a major fragment ion by losing a stable unit of Au₄(SR)₄ with a ring structure (entry 2a).^79,80 The loss of an Au₄(SR)₄ unit is commonly observed not only in the CID of other [Au_x(SR)_y]⁻ clusters⁸¹ but also in the MALDI of Au_x(SR)_y recorded under high laser fluence.⁴² Theoretical calculations predicted that the dissociation of [Au₂₅(SCH₃)₁₈]⁻ into [Au₂₁(SCH₃)₁₄]⁻ and Au₄(SCH₃)₄ is exothermic only by 0.82 eV and proceeds via complex rearrangement of the intracluster chemical bonds.⁸² When the collision energy was further increased, loss of the second Au₄(SR)₄ and a R₂ unit was observed (entry 2b). Thiolate-protected Ag clusters exhibit different CID patterns from the Au analogues. [Ag₂₅(SR)₁₈]⁻ (11) undergoes sequential loss of a stable unit of Ag₃(SR)₃ [Fig. 5(b), entries 3a and 3b].⁸³ Multiply charged Ag clusters [Ag₂₉(SRS)₁₂]³⁻ (13) [Fig. 5(c), entries 4a and 4b] and [Ag₄₄(SR)₃₀]⁴⁻ (14) (entry 5) undergo CID while releasing small ionic species [Ag₅(SR)₃]⁻ and [Ag_n(SR)_n+1]⁻ (n = 1, 2), respectively. In summary, the n^* values in all the fragment ions in entries 1–5 are retained to be 8 or 18. These results revealed that the CID pathways are governed by the electronic stability of the fragment ions.

The CID pattern also provides information on the relative binding affinity of the different ligands.^29,84,85 In the CID of [Au₁₁(PR₃)₇(NHC)X₂]⁺ (4), the loss of NHC was suppressed and that of Au(NHC)X or Au(PR₃)X was dominant (entry 6).²⁹ This result demonstrates that NHC has significantly higher binding affinity to Au than PR₃ through the strong Au—C bond.²⁹ The difference of binding affinities has been supported theoretically.⁸⁴ The stability against CID processes decreased in the order of [Au₂₅(p-SC₆H₄CO₂H)₁₈]⁻ > [Au₂₅(m-SC₆H₄CO₂H)₁₈]⁻ > [Au₂₅(o-SC₆H₄CO₂H)₁₈]⁻.⁸⁵ This trend was explained in terms of weakening of the Au—S bonds by the steric effect of the carboxyl group at the ortho position.

Elimination of anionic ligands in the form of neutral molecules such as H₂ and (C≡CR)₂ results in the reduction of the fragment ions (reductive elimination). Pradeep found that CID of [Ag₁₈(PR₃)₁₀H₁₆]²⁺ resulted in the reductive elimination of H₂ and the formation of Ag₁₇H₁₄⁺(2e) and Ag₁₇⁺(16e) (entry 7).⁸⁶ Tsukuda showed that the CID of [MAu₂₄(C≡CR)₁₈]²⁻ (M = Pd (10a); Pt (10b)) mainly afforded [MAu₂₄(C≡CR)_18−2n]²⁻ (n = 1–6) having (8 + 2n) electrons via sequential reductive elimination of (C≡CR)₂ [Fig. 6(a), entry 8].⁸⁷ Theoretical calculations on a model system [MAu₂₄(C≡CCF₃)₁₆]²⁻ predicted that the increased electrons are not accommodated in the 1D orbital of the M@Au₁₂ core but are localized at the Au₂(C≡CCF₃)₁ sites formed from the original Au₂(C≡CCF₃)₃ motif by desorption of (C≡CCF₃)₂. Thus, [MAu₂₄(C≡CR)_18−2n]²⁻ can be viewed as novel assemblies of stable clusters with 8e and 2e. The desorption step of (C≡CR)₂ continued to n = 6, leading to the formation of M@Au₁₂[Au₂(C≡CR)₁]₆ [Fig. 6(b)]. The above CID process corresponds to a half reaction of homo-coupling of terminal alkynes HC≡CR by regarding [MAu₂₄(C≡CR)₁₈]²⁻ as reaction intermediates. In this regard, the CID-MS of Au/Ag MPCs will give deeper insight into the mechanism of cluster-catalyzed reactions.

Johnson conducted high-resolution UV–vis absorption spectroscopy of [Au₉(PR₃)₈]³⁺ (1) at cryogenic temperature in the gas phase.^89–91 Condensation of inert gases (He, N₂) as a tag can freeze the internal motion of the Au₈ core without noticeable influence on the stable structures. In addition, the tags are inevitably (with 100% quantum yield) released upon absorption of UV–vis laser by the tagged MPCs. As a result, the action spectrum of the depletion of the intensity of tagged MPCs corresponds to their optical absorption spectra at extremely low temperature. The UV–visible spectrum of [Au₉(PR₃)₈]³⁺ (1) thus recorded exhibits much sharper profiles than that in the solution phase, thanks to the suppression of thermal broadening [Fig. 8(a)].^89–91 The spectra at cryogenic temperature can be compared directly with the theoretical spectra [Fig. 8(b)].⁹² It was suggested that hydride and halide affect the electronic structure of [Au₉(PR₃)₈]³⁺ in similar ways.⁹⁰ 

Whetten found novel PD processes of [Au₂₅(SC₆H₄CO₂H)₁₈]⁻(NH₄⁺)₃ at 193 nm.⁹³ The initial stage of the PD was the sequential loss of ammonium salts leading to [Au₂₅(SC₆H₄CO₂H)₁₅]²⁺, followed by two competing fragmentations of S(C₆H₄CO₂H)₂ and (C₆H₄CO₂H)₂. These PD processes led to nearly complete removal of the thiolates, while retaining the total number of Au atoms. This phenomenon is similar to the observation of [Au₂₅S_∼12]⁻ in the laser desorption process of [Au₂₅(SR)₁₈]⁻ (7).⁹⁴ 

IM MS determines the collision cross section (CCS) of Au/Ag MPCs, which directly reflects the geometrical motif including the ligand layer. The experimental CCS value is determined using the drift time in the gas cell estimated from the arrival time of the injected ions at the mass spectrometer. Thus, IM MS also allows us to identify structural isomers if present and to monitor an isomerization process induced by collisional excitation.

The ATDs of [Au₉(PR₃)₈]³⁺ (1) and [PdAu₈(PR₃)₈]²⁺ (2) with crown motifs introduced by ESI exhibit single peaks: the CCSs were determined to be 442 and 422 Ų, respectively.⁹⁶ Their ATDs were monitored by increasing the collision energy by reducing the He pressure in the gas cell to test the possibility of detecting structural isomerization.⁵¹ Both [Au₉(PR₃)₈]³⁺ (1) and [PdAu₈(PR₃)₈]²⁺ (2) isomerized to smaller species with the CCS values of 404 and 402 Ų, respectively. These observations were interpreted in terms of the collision-induced transformation of the ligand layer structures from the disordered phase into the densely packed phase due to CH–π and π–π interactions found in the singe crystal. This observation suggests that the ESI allows the isolation of the clusters while retaining the structures including the ligand layers in dispersing media, which are determined by the subtle balance between ligand–solvent and ligand–ligand interactions.

Recently, high resolution IM MS of [Au₂₅(SR)₁₈]^z (z = −1, 0) has revealed a topological isomer having a larger CCS value than that of the well-known crystal structure [7, Fig. 10(b)], as shown in Fig. 10(a).¹⁰¹ The CCS value experimentally determined for the new peak agrees well with that of a less stable (ΔE = 0.5 eV) isomer predicted theoretically [Fig. 10(c)] in which two terminal thiolates of the Au₂(SR)₃ oligomer units are bonded to adjacent Au atoms on the icosahedral Au₁₃ core:¹⁰² the CCS values of structure 7 and the new isomer were calculated to be 473 and 583 Ų, respectively. The relative population of the isomer with respect to the ground-state structure was enhanced by in-source activation. This behavior suggests that the new isomer is connected topologically to the ground-state structure via a simple rotation of the Au core without breaking any Au—S bonds [Fig. 10(d)].

The above results not only illustrate the structural fluxionality of Au/Ag MPCs but also suggest the risk that they may be isomerized to another structure from the crystallized structures during the ESI process.

The electronic structures of Au/Ag MPCs have been conventionally investigated by UV–vis absorption spectroscopy. The absorption spectra provide information on the energy differences between the quantized levels, such as the HOMO–LUMO gap.⁵⁷ Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) have been used to determine the redox potentials in solution. The results not only explain the redox behaviors in electrocatalysis but also determine the electrochemical HOMO–LUMO gap. Gas-phase PES on Au/Ag MPC anions allows us to directly probe the energy levels of occupied states with respect to the vacuum level and the density of state, such as electron affinities (EAs). PES also gives novel information such as the repulsive Coulomb barrier (RCB) of multiply charged anions and novel phenomena such as thermionic emission.

Figure 11 shows the PE spectrum of [Au₂₅(SC₁₂H₂₅)₁₈]⁻ recorded at 355 nm.¹⁰³ The spectrum exhibits a small hump (band A) at the onset and an intense band (band B) at higher energy, which are assigned to the photodetachment from the 1P orbital and Au 5d bands, respectively. Since SCXRD results showed that the structural difference between [Au₂₅(SR)₁₈]⁻ (7) and [Au₂₅(SR)₁₈]⁰ is small,¹⁰⁴ the spectral onset determines the adiabatic electron affinity (AEA) of [Au₂₅(SC₁₂H₂₅)₁₈]⁰. The AEA value of [Au₂₅(SC₁₂H₂₅)₁₈]⁰ thus determined was 2.2 eV. These experimental AEA values are significantly smaller than those of a model system [Au₂₅(SCH₃)₁₈]⁰ calculated theoretically (3.0¹⁰⁵ and 3.17 eV¹⁰⁶). The AEA values of [Au₂₅(SR)₁₈]⁰ and [Ag₂₅(SR)₁₈]⁰ were determined to be 2.36 ± 0.01 and 2.02 ± 0.01 eV, respectively, which are comparable to that of [Au₂₅(SC₁₂H₂₅)₁₈]⁰. These results indicate that the electronic structure of the Au₁₃ core is similar to that of the Ag₁₃ core and is not affected seriously by the structures of the thiolates.

Ligated metal chalcogenide molecular clusters such as Co₆Se₈L₆ (L = passivating ligands) have recently received renewed attention as “superatomic” building blocks of novel materials [Fig. 12(a)].^107–109 One of the attractive features of this family of clusters is that the characteristics can be tuned by selecting the ligands appropriately. Khanna et al. recently theoretically predicted that the sequential replacement of PEt₃ ligands of Co₆Se₈(PEt₃)₆ to CO increases the AEA value and causes transformation from a superatomic alkali metal to a superatomic halogen.¹⁰⁹ Bowen applied gas-phase anion PES to test this prediction. The clusters Co₆S₈(PEt₃)_6−x(CO)_x (x = 0–3) were synthesized by ligand substitution of Co₆S₈(PEt₃)₆ with CO. Their anions were formed using an infrared desorption/laser photoemission supersonic helium expansion source.¹¹⁰ The PES spectra of [Co₆S₈(PEt₃)_6−x(CO)_x]⁻ (x = 0–3) measured at 3.49 eV are shown in Fig. 12(b).¹¹¹ The AEAs of Co₆S₈(PEt₃)_6−x(CO)_x monotonically increase from 1.1 to 1.8 eV as x is increased from 0 to 3. These results will provide a guiding principle when designing programmable superatomic building units.

The electron binding mechanism in multiply charged anions is significantly affected by the repulsive Coulomb barrier (RCB) that emerges due to electrostatic repulsion between the detached electron and the remaining anion, as demonstrated by PES studies on C₇₀²⁻ and o-, m-, and p-C₆H₄(CO₂⁻)₂.^112,113 PES showed that the electron binding energies (E_B) of dianions [MAg₂₄(SR)₁₈]²⁻ [M = Pd (12a); Pt (12b)] are smaller than that of [Ag₂₅(SR)₁₈]⁻ (11) by ∼1 eV [Fig. 13(a)].¹¹⁴ The spectral cutoffs in E_B > 2.4 eV for [MAg₂₄(SR)₁₈]²⁻ (12) are due to the suppression of the electron detachment from the orbitals with E_B > hν − E_RCB, where E_RCB represents the height of the RCBs. The E_RCB value for [MAg₂₄(SR)₁₈]²⁻ (12) was determined to be 1.1 eV based on the cutoff positions in the PE spectra recorded at 532 and 355 nm. The significant reduction of the E_B values of [MAg₂₄(SR)₁₈]²⁻ (12) as compared to [Ag₂₅(SR)₁₈]⁻ (11) was ascribed to the weaker binding of valence electrons in M@(Ag⁺)₁₂ (M = Pd, Pt) compared to that in Ag⁺@(Ag⁺)₁₂ due to the reduction in the formal charge of the core potential and the upward shift of the apparent vacuum level by the presence of a RCB for M@(Ag⁺)₁₂ [Fig. 13(b)]. A recent PES study on [Ag₄₄(SR)₃₀]⁴⁻ (14) by Tsukuda and Häkkinen revealed that AEA of [Ag₄₄(SR)₃₀]³⁻ was negative [Figs. 13(c)],¹¹⁵ indicating that electron detachment from [Ag₄₄(SR)₃₀]⁴⁻ to form [Ag₄₄(SR)₃₀]³⁻ is energetically favorable. However, the tetra-anion in the gas phase was stable against electron autodetachment even under CID conditions. This observation was explained by the energy barrier associated with the electron detachment mainly due to the RCB being larger than that for fragmentation [Fig. 13(d)]. Electron detachment through RCB by tunneling in a multiply charged anion [Au₂₅(SG)₁₂(SG⁻)₆]⁷⁻ has been observed after photoabsorption of 200–300 nm.¹¹⁶ 

Veenstra and Kappes conducted femtosecond time-resolved, two-color, pump–probe photoelectron spectroscopy on free [Ag₂₉(SRS)₁₂]³⁻ (13) to examine the mechanism of phosphorescence in the absence of the solvents.¹¹⁷ The relaxation processes of the photoexcited state pumped by a 490-nm laser were probed by an 800-nm laser. Figure 14(a) shows a contour plot of the PE spectra vs electron kinetic energy (EKE) as a function of pump–probe delay. Two features are observed: (i) EKE decreases rapidly within ∼100 fs [inset of Fig. 14(a)], indicating ultrafast decay of the initially populated state, and (ii) a long-lived state with EKE of 2.1 eV remains populated at delay times greater than 100 ps. The peak region of the observed long-lived state is explained to be roughly consistent with the photoluminescent peak of [Ag₂₉(SRS)₁₂]³⁻ (13), whereas the former excited state is not observed in solution. The relaxation mechanism [Ag₂₉(SRS)₁₂]³⁻ (13) was explained by using a Jablonski diagram constructed by using AEA of [Ag₂₉(SRS)₁₂]²⁻ and E_RCB to be 0.9 ± 0.1 and 1.7 ± 0.1 eV, respectively, determined by one-photon PES [Fig. 14(b)]. A singlet charge transfer state, S_n, created by the pump light rapidly relaxes via a cascade of internal conversion (IC) steps to S₁. Then, either rapid ISC to a long-lived triplet state (T₁) or IC to the vibrationally excited ground state occurs.

The PE spectrum of [Au₂₅(SR)₁₈]⁻ (7) recorded at 266 nm [Fig. 15(a)] does not show the band structures observed at 355 nm [Fig. 11] but is dominated by a slow electron emission observed at E_B > 4.0 eV.¹¹⁸ The slow electron emission at 266 nm is assigned to thermionic emission (TE) from vibrationally excited anionic states, as observed in the PES of naked cluster anions of W, Nb, and Pt.¹¹⁹ The PES profile is reproduced by the kinetic energy distribution simulated at 1.6 × 10³ K (black dotted lines in Fig. 15(a), which is estimated by assuming that the photon energy absorbed (4.66 eV) is equally distributed to the vibrational degrees of freedom of the Au₁₃ core. The PD-MS of [Au₂₅(SR)₁₈]⁻ (7) at 266 nm revealed that the TE is a major decay pathway and that PD is negligible. The relaxation pathways of [Au₂₅(SR)₁₈]⁻ upon photoexcitation at 266 nm are explained by a Jablonski diagram [Fig. 15(b)]. Initially, [M₂₅(SR)₁₈]⁻ is photoexcited selectively to [M₂₅(SR)₁₈]^−*, which is embedded in the photodetachment continuum [shaded area in Fig. 15(b)]. Electronic transitions within the Au₁₃ core and metal-to-ligand transitions¹²⁰ may be involved in the electronic transitions at 266 nm. Electronically excited [Au₂₅(SR)₁₈]^−* quickly undergoes IC to form vibrationally excited [Au₂₅(SR)₁₈]⁻ followed by TE leaving internal energy in the remaining neutral [Au₂₅(SR)₁₈]⁰. Photodissociation is almost completely suppressed even though the photon energy (4.66 eV) exceeds the energy required for dissociation into [Au₂₁(SR)₁₄]⁻ and Au₄(SR)₄ (2.9 eV).⁸² Protection of the Au₁₃ core by stiff Au₂(SR)₃ units¹²¹ may contribute to the efficient IC process of [Au₂₅(SR)₁₈]^−* by retarding the nuclear motion toward the dissociation. TE was also observed in the photoirradiation of [Ag₂₅(SR)₁₈]⁻ (11).¹¹⁸ 

This article summarized the recent progress in the gas-phase characterizations of Au/Ag MPCs, [(Au/Ag)_xL_y]^z, synthesized in solution with atomic precision.

Mass spectrometry (MS) coupled with soft ionization such as electrospray ionization (ESI) and matrix-assisted laser desorption/ionization (MALDI) methods has been routinely used to determine the chemical formulas of Au/Ag MPCs. Another important application of MS is the in situ detection of transient species produced in solution, which will help the molecular-level understanding of the mechanism of complex processes in solutions: seed- or hydride-mediated growth processes, alloying processes, ligand exchange reactions, ligand-exchange induced size transformation, and metal exchange reactions between the clusters. Novel intermediates such as hydride-doped clusters and dimers of the clusters were detected.

Dissociation MS induced by photoirradiation (PD-MS) and collision with buffer gas (CID-MS) or surface (SID-MS) allows us to investigate the fragmentation patterns of Au/Ag MPCs upon energy deposition. The fragmentation patterns provide information on the thermal stability of Au/Ag MPCs, the relative binding affinities of different ligands with respect to the cores, and the effect of the ligand structure on their binding affinities. Photodissociation is also used for action spectroscopy to measure absorption spectra in the gas phase, showing fine structures due to the absence of solvents.

Ion mobility (IM) MS determines the collision cross section (CCS) of Au/Ag MPCs, which directly reflects the geometrical motif including the ligand layer. IM MS also allows us to identify structural isomers if present and to monitor an isomerization process induced by collisional excitation.

Photoelectron spectrometry (PES) on Au/Ag MPC anions allows us to directly probe the electronic structures of occupied states: energy levels with respect to the vacuum level and density of states, the height of the repulsive Coulomb barrier for multiply charged anions, and the relaxation dynamics and thermionic emission of photoexcited electrons.

Further characterization of Au/Ag MPCs in the gas phase will elucidate their intrinsic geometric and electronic structures in the absence of perturbations from the environments. Future studies in combination with the gas-phase methods and the conventional methods such as single-crystal x-ray crystallography and x-ray absorption spectroscopy¹²² will deepen our understanding of the correlations between the structures and properties of Au/Ag MPCs and thus contribute to the development of the materials science of Au/Ag MPCs.

We thank Dr. Shinjiro Takano for the assistance in preparing Tables I and II. This research was financially supported by the Grant-in-Aid for Scientific Research (Grant Nos. JP17H01182 and 20H00370) from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT).