Synthesis, Structural Characterization, and DFT Investigations of [M_xM′_5–xFe₄(CO)₁₆]^3– (M, M′ = Cu, Ag, Au; M ≠ M′) 2-D Molecular Alloy Clusters

Abstract

Miscellaneous 2-D molecular alloy clusters of the type [M_xM′_5–xFe₄(CO)₁₆]^3– (M, M′ = Cu, Ag, Au; M ≠ M′) have been prepared through the reactions of [Cu₃Fe₃(CO)₁₂]^3–, [Ag₄Fe₄(CO)₁₆]^4– or [M₅Fe₄(CO)₁₆]^3– (M = Cu, Ag) with M′(I) salts (M′ = Cu, Ag, Au). Their formation involves a combination of oxidation, condensation, and substitution reactions. The total structures of several [M_xM′_5–xFe₄(CO)₁₆]^3– clusters with different compositions have been determined by means of single crystal X-ray diffraction (SC-XRD) and their nature in solution elucidated by electron spray ionization mass spectrometry (ESI-MS) and IR and UV–visible spectroscopy. Substitutional and compositional disorder is present in the solid state structures, and ESI-MS analyses point out that mixtures of isostructural clusters differing by a few M/M′ coinage metals are present. SC-XRD studies indicate some site preferences of the coinage metals within the metal cores of these clusters, with Au preferentially in corner sites and Cu in the central site. DFT studies give theoretical support to the experimental structural evidence. The site preference is mainly dictated by the strength of the Fe–M bonds that decreases in the order Fe–Au > Fe–Ag > Fe–Cu, and the preferred structure is the one that maximizes the number of stronger Fe–M interactions. Overall, the molecular nature of these clusters allows their structures to be fully revealed with atomic precision, resulting in the elucidation of the bonding parameters that determine the distribution of the different metals within their metal cores.

Short abstract

The distribution of coinage metals in 2-D molecular alloy clusters has been investigated by IR, UV−visible, ESI-MS, X-ray crystallography, and DFT methods, revealing substitutional and compositional disorder as well as some site preference.

1. Introduction

In the recent years, alloying has attracted considerable interest in the field of molecular metal clusters and nanoclusters from both a fundamental and applicative point of view.¹⁻⁶ Indeed, mixing different metals with atomic control may lead to nanomaterials with new physical or chemical properties.⁷⁻¹⁰ Different methods may be used for the synthesis of alloy clusters, including direct reduction, redox condensation, and metal exchange.^11,12 Redox condensation has been widely exploited in the field of molecular metal carbonyl clusters.¹³⁻¹⁸ Conversely, direct metal reduction and metal exchange have been mainly applied to Au nanoclusters. The former method often does not allow the structure of the resulting alloy nanoclusters to be controlled. Better results are obtained by selective metal exchange in preformed atomically precise Au nanoclusters.^1,2,9,11,12

Thiolate protected alloy nanoclusters and, generally speaking, Au-based alloy nanoclusters have been largely investigated in recent years, as part of the great interest for atomically precise and ultrasmall Au nanoparticles.¹⁹⁻²⁷ Within this framework, we reported some years ago a few examples of molecular Au nanoclusters protected by Fe-carbonyl fragments, that is, [Au₂₁{Fe(CO)₄}₁₀]^5–, [Au₂₂{Fe(CO)₄}₁₂]^6–, [Au₂₈{Fe(CO)₃}₄{Fe(CO)₄}₁₀]^8–, and [Au₃₄{Fe(CO)₃}₆{Fe(CO)₄}₈]^10–.²⁸ This may be viewed as an organometallic approach to metal nanoparticles, that includes also the [Ni₃₂Au₆(CO)₄₄]^6– and [Ni₁₂Au₆(CO)₂₄]^2– clusters containing a Au₆ core stabilized by Ni–CO moieties^29,30 and Pd clusters protected by Co-carbido-carbonyl fragments, that is, [H_4–nCo₂₀Pd₁₆C₄(CO)₄₈]ⁿ⁻ (n = 2–4), [H_3–nCo₁₅Pd₉C₃(CO)₃₈]ⁿ⁻ (n = 0–3), and [Co₁₃Pd₃C₃(CO)₂₉]⁻.³¹ Thus, planar organometallic species such as [M₃Fe₃(CO)₁₂]^3– (M = Cu, Ag, Au),^32,33 [M₄Fe₄(CO)₁₆]^4– (M = Ag, Au),^34,35 and [M₅Fe₄(CO)₁₆]^3– (M = Cu, Ag, Au)^28,32,34 may be viewed as 2-D molecular clusters, consisting of triangular M₃, square M₄, or centered rectangular M₅ 2-D cores stabilized by Fe(CO)₄ fragments. Moreover, Ag₄Au₄Fe₄(CO)₁₆(dppe)³⁶ and Cu₂Au₆Fe₄(CO)₁₆(dppe) (dppe = Ph₂PCH₂CH₂PPh₂)³⁷ represent examples of 2-D molecular alloy heteroleptic organomentallic clusters, which have been obtained by condensation of preformed anionic clusters with metal salts.

Within this context, 2-D molecular clusters are molecular planar clusters, that is, molecular clusters whose metal atoms lie on the same plane (Figure 1 and Scheme 1). In contrast, 3-D molecular clusters are molecular clusters whose metal atoms form a tridimensional metal core, such as tetrahedron, octahedron, icosahedron, and larger polyhedra or more complex and irregular nonplanar metal aggregates. 2-D molecular clusters are rather interesting,^38,39 since they represent an alternative to the more common 3-D growth of molecular clusters and nanoparticles. Moreover, they can be viewed as models of metal surfaces and monolayers. Nonetheless, their study is still rather limited compared to 3-D clusters, since their synthesis is not straightforward and, at the moment, general strategies are not available.

Figure 1.

Molecular structures of some representative (a–c) 2-D clusters and (d, e) 3-D clusters. (a) [M₃Fe₃(CO)₁₂]^3– (M = Cu, Ag, Au),^32,33 (b) [M₄Fe₄(CO)₁₆]^4– (M = Ag, Au),^34,35 (c) [M₅Fe₄(CO)₁₆]^3– (M = Cu, Ag, Au),^28,32,34 (d) [Ag₁₃Fe₈(CO)₃₂]^3–,^40,41 (e) [Au₂₈{Fe(CO)₃}₄{Fe(CO)₄}₁₀]^8–.²⁸

Scheme 1. Schematic Structures of the Clusters Discussed in the Paper.

(a) [Cu₃Fe₃(CO)₁₂]^3–. (b) [M₄Fe₄(CO)₁₆]^4– (M = Ag, Au). (c) [M₅Fe₄(CO)₁₆]^3– (M = Cu, Ag, Au). (d) [Ag₁₃Fe₈(CO)₃₂]^3– (only one Fe(CO)₄ is represented). (e) [AuFe₄(CO)₁₆]⁻. (f) Fe(CO)₄(AuPPh₃)₂. Carbonyl ligands are represented as lines.

Herein, we report a systematic study on [M_xM′_5–xFe₄(CO)₁₆]^3– (M, M′ = Cu, Ag, Au; M ≠ M′; x = 0–5) 2-D alloy molecular clusters obtained by a combination of oxidation, condensation, and substitution reactions. Their total structures have been determined by means of single crystal X-ray diffraction (SC-XRD) methods. In addition, they have been investigated by a combination of electron spray ionization mass spectrometry (ESI-MS) and IR and UV–visible spectroscopy. The relative stability of different isomers was rationalized on the basis of DFT calculations.

2. Results and Discussion

2.1. Synthesis of Ternary [M_xM′_5–xFe₄(CO)₁₆]^3– Clusters (M, M′ = Cu, Ag, Au; M ≠ M′; x = 0–5)

The reactions of [Cu₃Fe₃(CO)₁₂]^3– with increasing amounts of AgNO₃ resulted in the formation of the ternary [Ag_xCu_5–xFe₄(CO)₁₆]^3– clusters (Scheme 2 and Table 1), with x that increased as more AgNO₃ was added (up to 1 equiv). This point was evidenced by the fact that the ν_CO bands of [Cu₃Fe₃(CO)₁₂]^3– (1921 and 1843 cm^–1) were replaced by two intense ν_CO bands comprised between those of [Cu₅Fe₄(CO)₁₆]^3– (1940 and 1888 cm^–1) and [Ag₅Fe₄(CO)₁₆]^3– (1949 and 1878 cm^–1), depending on the Ag content.

Scheme 2. Synthesis of [Ag_xCu_5–xFe₄(CO)₁₆]^3– .

All of the reactions have been carried out in CH₃CN solution at room temperature. The reagents (AgNO₃, [Cu(CH₃CN)₄][BF₄], Cu(IMes)Cl) have been slowly added to the starting cluster solutions and the reactions monitored through IR spectroscopy. The stoichiometric ratios employed are summarized in Table 1. Complete details are given in the Experimental Section. The structure of “[Ag₆Fe₄(CO)₁₆]^2–” is represented in Scheme 3. Only one Fe(CO)₄ group is included in the schematic representation of [Ag₁₃Fe₈(CO)₃₂]^3–. The complete structure is reported in Figure 1. Carbonyl ligands are represented as lines.

Table 1. Experimental Conditions for the Synthesis of [NEt₄]₃[Ag_xCu_5–xFe₄(CO)₁₆] (x = 0–5)^a.

			composition of the reagents		composition of the products
entry		crystallization solvent	Ag	Cu	Ag	Cu
[Cu3Fe3(CO)12]3– + nAgNO3
1	n = 0.8	CH3CN	1.05	3.95	1.02	3.98
2	n = 1.3	CH3CN	1.51	3.49	5.00	0.00
3		dmf			4.25	0.75
4	n = 2.1	CH3CN	2.06	2.94	4.88	0.12
5	n = 2.3	dmf	2.17	2.83	4.92	0.08
[Cu5Fe4(CO)16]3– + 2.5 AgNO3
6		CH3CN	1.67	3.33	5.00	0.00
7		dmf			4.81	0.19
[Ag4Fe4(CO)16]4– + 1.06[Cu(CH3CN)4][BF4]
8		CH3CN	3.95	1.05	4.37	0.63
[Ag5Fe4(CO)16]3– + 3Cu(IMes)Cl (IMes = C3N2H2(C6H2Me3)2)
9		dmf	3.12	1.88	4.90	0.10
[Ag5Fe4(CO)16]3– + [Cu3Fe3(CO)12]3–
10		acetone	3.12	1.88	3.30	1.70
11		CH3CN			3.45	1.55

^aSee Scheme 2.

As a general procedure, the reactions were conducted with different stoichiometric amounts of the reagents ([Cu₃Fe₃(CO)₁₂]^3– and AgNO₃; [Cu₅Fe₄(CO)₁₆]^3– and AgNO₃; or other reagents as summarized in Schemes 2, 4, and 5 and Tables 1–3; complete details may be found in the Experimental Section). At the end of the reaction, the solvent was removed under reduced pressure, and the residue was washed with H₂O and toluene and extracted with solvents of increasing polarity (acetone, CH₃CN, dmf). The resulting solutions were analyzed by IR spectroscopy and, eventually, layered with an appropriate solvent (acetone/n-hexane; CH₃CN/n-hexane/di-iso-propyl-ether; dmf/isopropanol) in the attempt to obtain crystals suitable for X-ray diffraction. Yields have been determined on the isolated crystals, which have been, then, analyzed by SC-XRD, elemental analysis, microwave plasma-atomic emission spectrometry, ESI-MS, and IR and UV–visible spectroscopy.

Scheme 4. Synthesis of [Au_xCu_5–xFe₄(CO)₁₆]^3– .

All of the reactions have been carried out in CH₃CN solution at room temperature. The reagents (Au(PPh₃)Cl or Au(Et₂S)Cl) have been slowly added to the starting cluster solutions and the reactions monitored through IR spectroscopy. The stoichiometric ratios employed are summarized in Table 2. Complete details are given in the Experimental Section. The structure of “[Au₆Fe₄(CO)₁₆]^2–” is represented in Scheme 3. Carbonyl ligands are represented as lines.

Scheme 5. Synthesis of [Au_xAg_5–xFe₄(CO)₁₆]^3– .

All of the reactions have been carried out in CH₃CN solution at room temperature. The reagent (Au(Et₂S)Cl) has been slowly added to the starting cluster solutions and the reactions monitored through IR spectroscopy. The stoichiometric ratios employed are summarized in Table 3. Complete details are given in the Experimental Section. Carbonyl ligands are represented as lines.

Table 3. Experimental Conditions for the Synthesis of [NEt₄]₃[Au_xAg_5–xFe₄(CO)₁₆] (x = 0–5)^a.

			composition of the reagents		composition of the products
entry		crystallization solvent	Au	Ag	Au	Ag
[Ag4Fe4(CO)16]4– + 0.8Au(Et2S)Cl
21		CH3CN	0.83	4.17	0.64	4.36
22		acetone			0.81	4.19

^aSee Scheme 5.

Formation of [Ag_xCu_5–xFe₄(CO)₁₆]^3– (x = 0–5) under the conditions described above could be explained by considering a combination of oxidation (eq 1), condensation (eq 2), and substitution (eq 3–7) reactions. The occurrence of the substitution reactions was directly confirmed by reacting preformed [Cu₅Fe₄(CO)₁₆]^3– with increasing amounts of AgNO₃. Also, in these cases, formation of [Ag_xCu_5–xFe₄(CO)₁₆]^3– (x = 0–5) was observed, indicating that Ag⁺ ions can replace Cu⁺ ions within the structure of [Cu₅Fe₄(CO)₁₆]^3–. Oxidation/condensation reactions are likely to occur almost in parallel followed by substitution reactions. This point was assessed by ESI-MS analyses (see section 2.3), which clearly revealed the formation at first of [Cu₅Fe₄(CO)₁₆]^3– and [AgCu₄Fe₄(CO)₁₆]^3–, followed by [Ag_xCu_5–xFe₄(CO)₁₆]^3– species richer in Ag (x ≥ 2). Formation of [Cu₅Fe₄(CO)₁₆]^3– directly supports the oxidation reaction, as well as the appearance of a silver mirror on the reaction flask accompanied by some gas evolution. The fact that oxidation formally requires more AgNO₃ than condensation, in accord with eqs 1 and 2, may be explained assuming that there is competition between the two reactions which actually proceed in a parallel way. The stoichiometries depicted in eqs 1 and 2 are those required in the case that a single reaction occurred up to completion. Actually, both reactions start as soon as some AgNO₃ is added to [Cu₃Fe₃(CO)₁₂]^3–. Then, as some [Cu₅Fe₄(CO)₁₆]^3– and [AgCu₄Fe₄(CO)₁₆]^3– are formed, also substitution reactions take place. Assuming that they are all equilibria, the total amount of AgNO₃ added determines the final distribution of the products as the result of all equilibria. This usually results in one to three species as the main ones present in solution, as evidenced by ESI-MS analyses.

Oxidation:

1

Condensation:

2

Substitution:

3

4

5

6

7

By adding further AgNO₃ (>1 equiv per [Cu₃Fe₃(CO)₁₂]^3–), two new ν_CO bands at 1970 and 1890 cm^–1, attributable to [Ag₆Fe₄(CO)₁₆]^2–,^34,40 appeared besides those of [Ag_xCu_5–xFe₄(CO)₁₆]^3– (x = 0–5) and rapidly became the major ones. Formation of [Ag₆Fe₄(CO)₁₆]^2– was accompanied by the precipitation of an amorphous solid. Indeed, it was previously reported in the literature that [Ag₆Fe₄(CO)₁₆]^2– was actually a mixture of almost nonsoluble polymeric species (Scheme 3).^34,40 Their formation was due to different equilibria; some representative ones were depicted in eqs 8–11. Because of such equilibria, it was possible during the workup of these reaction mixtures to extract in polar solvents and, then, crystallize Ag-rich [Ag_xCu_5–xFe₄(CO)₁₆]^3– (x = 0–5) species. Indeed, [Ag₆Fe₄(CO)₁₆]^2– is in equilibrium with [Ag₅Fe₄(CO)₁₆]^3– through eqs 9–11, and in turn, [Ag₅Fe₄(CO)₁₆]^3– is in equilibrium with [Ag_xCu_5–xFe₄(CO)₁₆]^3– (x = 0–5) through eqs 3–7.

8

9

10

11

Scheme 3. Growth Scheme of the [M₆Fe₄(CO)₁₆]^2– (M = Cu, Ag, Au) Cluster in the Oligomeric Form: (A) [M₅Fe₄(CO)₁₆]^3– Unit; (B) [M{M₅Fe₄(CO)₁₆}₂]^5– Dimer; (C) [M₂{M₅Fe₄(CO)₁₆}₃]^7– Trimer.

Eventually, by employing 2.5 or more equivalents of AgNO₃ per mole of [Cu₃Fe₃(CO)₁₂]^3–, [Ag₁₃Fe₈(CO)₃₂]^3– was obtained as the final product.^40,41

12

Similar results were obtained by employing [Cu₅Fe₄(CO)₁₆]^3– instead of [Cu₃Fe₃(CO)₁₂]^3–, or Ag(dppe)(NO₃) instead of AgNO₃. During the workup of all of these reactions, it was possible to separate and crystallize several salts of the type [NEt₄]₃[Ag_xCu_5–xFe₄(CO)₁₆] (x = 0–5). As evidenced in Table 1, Ag very rapidly substituted Cu in such clusters, which were always richer in Ag than the reagents. These were the only ternary clusters obtained. For instance, by reacting [Cu₅Fe₄(CO)₁₆]^3– with 3 equiv of Ag(dppe)(NO₃), crystals of [Cu(dppe)₂]₃[Ag₁₃Fe₈(CO)₃₂] were obtained, that contained a binary Ag–Fe cluster. A few crystals of Cu₃Br₃(dppe)₃ were also obtained as a side product and mechanically separated from [Cu(dppe)₂]₃[Ag₁₃Fe₈(CO)₃₂]. The structure of Cu₃Br₃(dppe)₃ as acetone solvate as well as the isostructural Cu₃Cl₃(dppe)₃ were previously reported.⁴²

Ternary [Ag_xCu_5–xFe₄(CO)₁₆]^3– (x = 0–5) clusters could also be obtained from the reactions of [Ag₄Fe₄(CO)₁₆]^4– or [Ag₅Fe₄(CO)₁₆]^3– with Cu(I) salts such as [Cu(CH₃CN)₄][BF₄] and Cu(IMes)Cl (IMes = C₃N₂H₂(C₆H₂Me₃)₂), or by mixing together [Cu₃Fe₃(CO)₁₂]^3– and [Ag₅Fe₄(CO)₁₆]^3–. Some representative examples were reported in Table 1 and Scheme 2. [Cu(CH₃CN)₄][BF₄] resulted in being more reactive than Cu(IMes)Cl toward these clusters. Therefore, both reagents have been employed in order to better control the composition of the final [Ag_xCu_5–xFe₄(CO)₁₆]^3– (x = 0–5) clusters. This is a general strategy employed in this work, and other examples are AgNO₃/Ag(dppe)(NO₃) and Au(Et₂S)Cl/Au(PPh₃)Cl (for each couple, the first reagent is the more reactive one). Indeed, more reactive reagents (AgNO₃, [Cu(CH₃CN)₄][BF₄], Au(Et₂S)Cl) favor the substitution of the coinage metal present in the starting cluster with that present in the reagent, whereas substitution proceeds to a minor extent with less reactive reagents (Ag(dppe)(NO₃), Cu(IMes)Cl, Au(PPh₃)Cl).

It must be remarked that the fractionary indices of [Ag_xCu_5–xFe₄(CO)₁₆]^3–(x = 0–5) indicated that they were actually mixtures of species differing for a few Ag/Cu atoms. Indeed, in some cases, it was possible to separate species with slightly different compositions (entries 2–3, 6–7, and 10–11 in Table 1), by extraction with solvents of different polarities during workup of the same reaction.

The reactions of [Cu₃Fe₃(CO)₁₂]^3– and [Cu₅Fe₄(CO)₁₆]^3– with Au(I) salts were very similar to those employing Ag(I) salts (Scheme 4 and Table 2). Thus, ternary [Au_xCu_5–xFe₄(CO)₁₆]^3– (x = 0–5) clusters were formed at first, followed by [Au₆Fe₄(CO)₁₆]^2–. Then, further addition of Au(I) resulted in mixtures of “gold browns” such as [Au₂₁{Fe(CO)₄}₁₀]^5–, [Au₂₂{Fe(CO)₄}₁₂]^6–, [Au₂₈{Fe(CO)₃}₄{Fe(CO)₄}₁₀]^8–, [Au₃₄{Fe(CO)₃}₆{Fe(CO)₄}₈]^10–,²⁸ and, eventually, [AuFe₄(CO)₁₆]⁻ or Fe(CO)₄(AuPPh₃)₂,⁴³⁻⁴⁵ depending on the fact that an excess of Au(Et₂S)Cl or Au(PPh₃)Cl was used. Also, in this case, [Au_xCu_5–xFe₄(CO)₁₆]^3– (x = 0–5) were the only ternary species isolated.

Table 2. Experimental Conditions for the Synthesis of [NEt₄]₃[Au_xCu_5–xFe₄(CO)₁₆] (x = 0–5)^a.

			composition of the reagents		composition of the products
entry		crystallization solvent	Au	Cu	Au	Cu
[Cu3Fe3(CO)12]3– + nAu(PPh3)Cl
12	n = 1.4	acetone	1.59	3.41	1.15	3.85
13	n = 1.5	acetone	1.67	3.33	1.31	3.69
14		dmf			1.67	3.33
15	n = 3.0	dmf	2.5	2.5	2.48	2.52
[Cu3Fe3(CO)12]3– + nAu(Et2S)Cl
16	n = 0.7	CH3CN	0.95	4.05	2.18	2.82
17					2.73	2.27
18	n = 1.9	CH3CN	1.94	3.06	4.59	0.41
19					4.62	0.38
[Cu5Fe4(CO)16]3– + 1.2Au(Et2S)Cl
20		acetone	0.97	4.03	1.09	3.91

^aSee Scheme 4.

Replacement of Cu with Au in [Au_xCu_5–xFe₄(CO)₁₆]^3– (x = 0–5) was more gradual than in the related Ag–Cu–Fe clusters. As a consequence, it was possible to obtain species with a more continuous distribution of the two metals (Table 2). Moreover, Cu-substitution was favored by using Au(Et₂S)Cl compared to Au(PPh₃)Cl. This might be due to the fact that the phosphine complex was less reactive.

Ternary [Au_xAg_5–xFe₄(CO)₁₆]^3– clusters (x = 0–5) were, then, obtained from the reaction of [Ag₄Fe₄(CO)₁₆]^4– with Au(Et₂S)Cl as a representative example (Scheme 5 and Table 3).

The IR spectra of the ternary [M_xM′_5–xFe₄(CO)₁₆]^3– (x = 0–5; M, M′ = Cu, Ag, Au; M ≠ M′) clusters show two almost equally intense ν_CO bands in the regions 1942–1950 and 1878–1894 cm^–1 for [Ag_xCu_5–xFe₄(CO)₁₆]^3–, 1945–1951 and 1870–1887 cm^–1 for [Au_xCu_5–xFe₄(CO)₁₆]^3–, and 1946–1950 and 1874–1880 cm^–1 for [Au_xAg_5–xFe₄(CO)₁₆]^3– (see Figures S1–S9 in the Supporting Information). Similar spectral features have been reported for the related binary clusters, that is, 1940 and 1888 cm^–1 for [Cu₅Fe₄(CO)₁₆]^3–, 1949 and 1878 cm^–1 for [Ag₅Fe₄(CO)₁₆]^3–, and 1944 and 1861 cm^–1 for [Au₅Fe₄(CO)₁₆]^3–. All of the CO ligands are terminally bonded within Fe(CO)₄ groups. The presence of only two intense ν_CO stretching bands in these 2-D (planar) clusters containing an Fe(CO)₄ group has been explained by application of a spherical harmonic model (SHM) to their ν_CO vibrational spectra.⁴⁶ The reader may find in the cited literature full theoretical details on why SHM rather classical symmetry rules (group theory) is needed in order to interpret the IR spectra in the ν_CO region of these 2-D clusters. Overall, SHM correctly predicts the presence of two strong ν_CO bands in the IR spectra of such clusters. The same applies to the parent 2-D clusters [Cu₃Fe₃(CO)₁₂]^3– (1921 and 1843 cm^–1), [Ag₄Fe₄(CO)₁₆]^4– (1928 and 1852 cm^–1), and [Au₄Fe₄(CO)₁₆]^4– (1929 and 1864 cm^–1), as well as [Ag₆Fe₄(CO)₁₆]^2– (1970 and 1890 cm^–1) and [Au₆Fe₄(CO)₁₆]^2– (1976 and 1906 cm^–1). Thus, all of these 2-D clusters, regardless of the number of M atoms and Fe(CO)₄ groups, display very similar IR spectra on the ν_CO region, showing two intense bands that differ only in their frequencies. In turn, the observed stretching frequencies mainly depend on the ratio between the overall charge of the cluster and the number of metal atoms, and only to a lesser extent on the nature of the coinage metals present.

2.2. Molecular Structures of [M_xM′_5–xFe₄(CO)₁₆]^3– (M, M′ = Cu, Ag, Au; M ≠ M′; x = 0–5)

The molecular structure common to all ternary [M_xM′_5–xFe₄(CO)₁₆]^3– (x = 0–5; M, M′ = Cu, Ag, Au; M ≠ M′) clusters is based on a centered M₅ rectangle (Figure 2), as previously found also in the binary [M₅Fe₄(CO)₁₆]^3– (M = Cu, Ag, Au) clusters.^28,32,34,47 The positions occupied by the five coinage metals can be grouped into two sites: (a) the unique central position; (b) the four equivalent corner positions. This M₅ core is bonded to two μ-Fe(CO)₄ and two μ₃-Fe(CO)₄ groups on the shorter and longer edges of the rectangle, respectively. Overall, the M atom at the center forms four M_center–M_corner bonds and two M–Fe bonds, whereas M atoms at the corner sites form one M_center–M_corner bond, one M_corner–M_corner bond, and two M–Fe bonds. The Fe–M_corner–Fe coordination is almost linear, as expected for a d¹⁰ M(I) ion. Some sub van der Waals M–C(O) contacts are also present, but their nature in these and related clusters is rather debated.^48,49 Theoretical calculations reported in the literature mainly point out that the presence of such contacts in the solid state structures of Fe carbonyl clusters containing also coinage metals is essentially the consequence of steric requirements and not real bonds or even any (even weak) attraction. Thus, the preferential arrangement of the CO ligands about the Fe center brings the carbonyls in closer proximity to the M (Cu, Ag, Au) center, without any real interaction that could be spectroscopically evidenced.

Figure 2.

Molecular structure of the [M_xM′_5–xFe₄(CO)₁₆]^3– (x = 0–5; M, M′ = Cu, Ag, Au; M ≠ M′) clusters (purple, M in the center; orange, M in the corner positions; green, Fe; gray, C; red, O). M–C(O) contacts are represented as fragmented lines. Two different views are reported (a, b), as well as the metal core (c).

The fractional indices present in the formulas of [M_xM′_5–xFe₄(CO)₁₆]^3– (x = 0–5; M, M′ = Cu, Ag, Au; M ≠ M′) are indicative of compositional disorder; that is, the crystals contain mixtures of species. For instance, [Ag_4.25Cu_0.75Fe₄(CO)₁₆]^3– corresponds to a mixture of [Ag₄CuFe₄(CO)₁₆]^3– (75%) and [Ag₅Fe₄(CO)₁₆]^3– (25%), whereas [Au_2.48Cu_2.52Fe₄(CO)₁₆]^3– contains [Au₂Cu₃Fe₄(CO)₁₆]^3– (52%) and [Au₃Cu₂Fe₄(CO)₁₆]^3– (48%).

In addition, substitutional disorder is always observed, since M and M′ are (not equivalently) disordered over the central and corner positions (Tables 4–6). In the case of Au–Cu clusters, Au strongly prefers the corner sites and Cu the central position. This trend is observed, even if to a less extent, also in Ag–Cu clusters. In the case of Au–Ag clusters, even if the number of entries is very limited, it is possible to notice a Au preference for the corner sites.

Table 4. Composition of [NEt₄]₃[Ag_xCu_5–xFe₄(CO)₁₆] (x = 0–5).

	total		centre		corner
entrya	Ag	Cu	Ag	Cu	Ag	Cu
1	1.02	3.98	0.02	0.98	1.00	3.00
10	3.30	1.70	0.60	0.40	2.70	1.30
11	3.45	1.55	0.59	0.41	2.86	1.14
3	4.25	0.75	0.83	0.17	3.42	0.58
8	4.37	0.63	0.85	0.15	3.52	0.48
7	4.81	0.19	0.97	0.03	3.84	0.16
4	4.88	0.12	0.97	0.03	3.92	0.08
9	4.90	0.10	0.98	0.02	3.92	0.08
5	4.92	0.08	1.00	0.00	3.92	0.08
2,6	5.00	0.00	1.00	0.00	4.00	0.00

^aSee Table 1. Entries are listed in order of increasing Ag content.

Table 6. Composition of [NEt₄]₃[Au_xAg_5–xFe₄(CO)₁₆] (x = 0–5).

	total		centre		corner
entrya	Au	Ag	Au	Ag	Au	Ag
21	0.64	4.36	0.04	0.96	0.60	3.40
22	0.81	4.19	0.05	0.95	0.76	3.24

^aSee Table 3. Entries are listed in order of increasing Au content.

Table 5. Composition of [NEt₄]₃[Au_xCu_5–xFe₄(CO)₁₆] (x = 0–5).

	total		centre		corner
entrya	Au	Cu	Au	Cu	Au	Cu
20	1.09	3.91	0.00	1.00	1.09	2.91
12	1.15	3.85	0.00	1.00	1.15	2.85
13	1.31	3.69	0.00	1.00	1.31	2.69
14	1.67	3.33	0.00	1.00	1.67	2.33
16	2.18	2.82	0.00	1.00	2.18	1.82
15	2.48	2.52	0.00	1.00	2.48	1.52
17	2.73	2.27	0.00	1.00	2.73	1.27
18	4.59	0.41	0.59	0.41	4.00	0.00
19	4.62	0.38	0.61	0.39	4.00	0.00

^aSee Table 2. Entries are listed in order of increasing Au content.

The preference of Cu for the central position may be, at least partially, explained on the basis of the different ionic radii of the M(I) cations: Cu(I) 77 pm, Ag(I) 115 pm, and Au(I) 137 pm. Thus, the smallest Cu(I) ion prefers the central site. Moreover, corner positions display two strong M–Fe bonds with an almost linear Fe–M–Fe arrangement. This is the typical coordination found in M(I) complexes. The high affinity of Au for these corner sites indicates a stronger stability of such interactions in the case of Au compared to Cu and Ag.

As expected, M_center–M_corner [M(1)–M(2) in Figure 3], M_corner–M_corner [M(2)–M(2) in Figure 3], and M–Fe distances [M(1)–Fe(1), M(2)–Fe(1), and M(2)–Fe(2) in Figure 3] steadily increase moving from Cu-rich to Ag- or Au-rich clusters (Figure 3; Tables S1–S3 and Scheme S1 in the Supporting Information). This is well exemplified by comparing the M–M and M–Fe distances of ternary [M_xM′_5–xFe₄(CO)₁₆]^3– (x = 0–5; M, M′ = Cu, Ag, Au; M ≠ M′) and binary [M₅Fe₄(CO)₁₆]^3– (M = Cu, Ag, Au) clusters. Indeed, [Cu₅Fe₄(CO)₁₆]^3– displays considerably shorter distances, whereas these are almost identical in [Ag₅Fe₄(CO)₁₆]^3– and [Au₅Fe₄(CO)₁₆]^3–.

Figure 3.

M–M and M–Fe distances (Å) of (A) [NEt₄]₃[Ag_xCu_5–xFe₄(CO)₁₆]; (B) [NEt₄]₃[Au_xCu_5–xFe₄(CO)₁₆]; (C) [NEt₄]₃[Au_xAg_5–xFe₄(CO)₁₆]. Entries are reported in parentheses. M(1)–M(2), black; M(2)–M(2), red; M(1)–Fe(1), blue; M(2)–Fe(1), magenta; M(2)–Fe(2), green. See the scheme in the inset for the numbering.

2.3. ESI-MS Studies of [M_xM′_5–xFe₄(CO)₁₆]^3– (M, M′ = Cu, Ag, Au; M ≠ M′; x = 0–5)

In order to further elucidate the composition of ternary [M_xM′_5–xFe₄(CO)₁₆]^3– clusters, ESI-MS studies have been carried out on some of their [NEt₄]₃[M_xM′_5–xFe₄(CO)₁₆] crystals. The ESI-MS spectra recorded on CH₃CN solutions of [NEt₄]₃[Ag_1.02Cu_3.98Fe₄(CO)₁₆], [NEt₄]₃[Au_1.32Cu_3.68Fe₄(CO)₁₆], [NEt₄]₃[Au_2.48Cu_2.52Fe₄(CO)₁₆], [NEt₄]₃[Au_4.62Cu_0.38Fe₄(CO)₁₆], and [NEt₄]₃[Au_0.82Ag_4.18Fe₄(CO)₁₆] are reported in Figures S14–S37 in the Supporting Information (including calculated fits of the prominent peaks), and peak assignments are summarized in Tables S6–S10. For the sake of comparison, the ESI-MS spectra of binary [NEt₄]₃[Cu₅Fe₄(CO)₁₆] and [NEt₄]₃[Ag₅Fe₄(CO)₁₆] clusters are also reported (Figures S10–S13 and Tables S4 and S5 in the Supporting Information). It has not been possible to investigate [NEt₄]₃[Au₅Fe₄(CO)₁₆], since it is irreversibly oxidized during the dilution required for ESI-MS analyses. The peak assignments have been corroborated by comparing their experimental isotopic patterns with the theoretical ones based on the formulas.

Under ESI-MS conditions, the [M_xM′_5–xFe₄(CO)₁₆]^3– trianionic clusters were systematically detected as dianions. This might be explained assuming either oxidation of [M_xM′_5–xFe₄(CO)₁₆]^3– to [M_xM′_5–xFe₄(CO)₁₆]^2– or addition of H⁺ via electrostatic interaction resulting in [HM_xM′_5–xFe₄(CO)₁₆]^2– ions. Comparison of the experimental and calculated isotopic patterns of the most prominent peaks (Figures S10–S37 in the Supporting Information) suggests that formation of [HM_xM′_5–xFe₄(CO)₁₆]^2– adducts is more likely, even if the detection of a single hydrogen atom in such large clusters is at the limit of the precision of the instrument employed for these analyses. Formation of [HM_xM′_5–xFe₄(CO)₁₆]^2– adducts via electrostatic interaction is further supported by the fact that, in some cases, also the {[M_xM′_5–xFe₄(CO)₁₆][NEt₄]}^2– adducts involving [NEt₄]⁺ instead of H⁺ were present in the spectra.

The dianionic nature of all of these ions is confirmed by the systematic loss of m/z 14 units from the molecular ions, that corresponds to a CO ligand (28 amu) assuming z = 2. It must be remarked that CO loss is almost absent in the case of binary [Cu₅Fe₄(CO)₁₆]^3– and [Ag₅Fe₄(CO)₁₆]^3– clusters. Therefore, CO loss in the gas phase of ternary [M_xM′_5–xFe₄(CO)₁₆]^3– clusters may be viewed as an alloy effect due to the contemporary presence of two different coinage metals within their (M, M′)₅ core.

The ESI-MS spectrum (ES−) of [NEt₄]₃[Au_2.48Cu_2.52Fe₄(CO)₁₆] (Figures S24–S29 and Table S8 in the Supporting Information) displays two sets of peaks at m/z (relative intensities in parentheses): 759(30), 694(60), 680(10), 666(5), 652(10), 638(100), and 626(10) (set-1-Au₃Cu₂); 692(35), 628(60), 614(5), 600(30), 586(60), 572(80), and 558(15) (set-2-Au₂Cu₃). The peaks of set-1-Au₃Cu₂ originate from the [HAu₃Cu₂Fe₄(CO)₁₆]^2– ion (m/z = 694) by the stepwise loss of one to five CO ligands (m/z = 680, 666, 652, 638, 626), whereas the peak at m/z 759 corresponds to the {[Au₃Cu₂Fe₄(CO)₁₆][NEt₄]}^2– adduct. Similarly, the peaks of set-2-Au₂Cu₃ originate from the [HAu₂Cu₃Fe₄(CO)₁₆]^2– ion (m/z = 628) by the stepwise loss of one to five CO ligands (m/z = 614, 600, 586, 572, 558), whereas the peak at m/z 692 corresponds to the {[Au₂Cu₃Fe₄(CO)₁₆][NEt₄]}^2– adduct. Overall, the ESI-MS analysis indicates that two species of composition Au₃Cu₂ and Au₂Cu₃ are present in similar amounts, in agreement with the X-ray data which suggest that [NEt₄]₃[Au_2.48Cu_2.52Fe₄(CO)₁₆] contains 48% of [Au₃Cu₂Fe₄(CO)₁₆]^3– and 52% of [Au₂Cu₃Fe₄(CO)₁₆]^3–.

Similar considerations apply to [NEt₄]₃[Au_4.62Cu_0.38Fe₄(CO)₁₆] (Figures S30–S32 and Table S9 in the Supporting Information), for which the X-ray data suggest a mixture of 62% of [Au₅Fe₄(CO)₁₆]^3– and 38% of [Au₄CuFe₄(CO)₁₆]^3–. Indeed, the ESI-MS spectrum shows two sets of peaks attributable to a Au₅ (set-1-Au₅) and Au₄Cu (set-2-Au₄Cu) species. The first set (set-1-Au₅) displays peaks at m/z 893(30), 828(30), 814(45), 800(2), and 772(80) attributable to the {[Au₅Fe₄(CO)₁₆][NEt₄]}^2– adduct, the [HAu₅Fe₄(CO)₁₆]^2– parent ion, and the stepwise loss of one to four CO ligands. It must be remarked that the peak corresponding to the loss of three carbonyls is probably too weak to be detected. Conversely, the second set (set-2-Au₄Cu) shows peaks at m/z 826(30), 761(40), 747(20), 730(70), 719(10), and 705(100) attributable to {[Au₅Fe₄(CO)₁₆][NEt₄]}^2–, [HAu₅Fe₄(CO)₁₆]^2–, and the stepwise loss of one to four CO ligands.

The ESI-MS spectra of [NEt₄]₃[Ag_1.02Cu_3.98Fe₄(CO)₁₆], [NEt₄]₃[Au_1.32Cu_3.68Fe₄(CO)₁₆], and [NEt₄]₃[Au_0.82Ag_4.18Fe₄(CO)₁₆] (Figures S14–S23 and S33–S37 and Tables S6, S7, and S10 in the Supporting Information) contain three sets of peaks instead of two sets as above and suggest a slightly more complicated disorder (compositional) model. For instance, [NEt₄]₃[Ag_1.02Cu_3.98Fe₄(CO)₁₆] (Figures S14–S19 and Table S6 in the Supporting Information) could have been interpreted as a mixture of 2% [Ag₂Cu₃Fe₄(CO)₁₆]^3– and 98% [AgCu₄Fe₄(CO)₁₆]^3– on the basis of SC-XRD data. Conversely, its ESI-MS spectrum displays three sets of peaks (set-1-Ag₂Cu₃, set-2-AgCu₄, and set-3-Cu₅) attributable to the ionization of these two species as well as an additional [Cu₅Fe₄(CO)₁₆]^3– cluster. Indeed, set-1-Ag₂Cu₃ includes peaks at m/z 604(40), 539(60), 525(10), 510(50), and 497(30) attributable to {[Ag₂Cu₃Fe₄(CO)₁₆][NEt₄]}^2–, [HAg₂Cu₃Fe₄(CO)₁₆]^2–, and the stepwise loss of one to three carbonyls. Set-2-AgCu₄ shows peaks at m/z 582(60), 517(90), 503(10), 488(100), 474(65), 460(5), and 447(70) attributable to {[AgCu₄Fe₄(CO)₁₆][NEt₄]}^2–, [HAgCu₄Fe₄(CO)₁₆]^2–, and the stepwise loss of one to five CO ligands. Then, set-3-Cu₅ comprises peaks at m/z 560(35), 495(50), 482(15), 466(75), and 452(60) attributable to {[Cu₅Fe₄(CO)₁₆][NEt₄]}^2–, [HCu₅Fe₄(CO)₁₆]^2–, and the stepwise loss of one to three carbonyls. The presence of [Cu₅Fe₄(CO)₁₆]^3– in the mixture lends support to the occurrence of the initial oxidation reaction depicted in eq 1 (see section 2.1), whereas the presence of [AgCu₄Fe₄(CO)₁₆]^3– is mainly due to the condensation reaction (eq 2). Then, [Ag₂Cu₃Fe₄(CO)₁₆]^3– is formed via subsequent substitution (eq 3).

Similarly, the ESI-MS spectrum of [NEt₄]₃[Au_1.32Cu_3.68Fe₄(CO)₁₆] (Figures S19–S23 and Table S7 in the Supporting Information) suggests the presence of the three species [Au₂Cu₃Fe₄(CO)₁₆]^3–, [AuCu₄Fe₄(CO)₁₆]^3–, and [Cu₅Fe₄(CO)₁₆]^3–, as indicated by the presence of three sets of peaks. Three sets of peaks are present also in the case of [NEt₄]₃[Au_0.82Ag_4.18Fe₄(CO)₁₆] (Figures S33–S37 and Table S10 in the Supporting Information), suggesting a mixture of [Au₂Ag₃Fe₄(CO)₁₆]^3–, [AuAg₄Fe₄(CO)₁₆]^3–, and [Ag₅Fe₄(CO)₁₆]^3–. Also, in these cases, the distribution of the products is in agreement with the occurrence of oxidation, condensation, and substitution reactions.

On the basis of both SC-XRD and ESI-MS data, it is possible to conclude that the solid state structures of [M_xM′_5–xFe₄(CO)₁₆]^3– consist of mixtures of two or three species differing for one to two coinage metals in the metal core of the cluster.

It is also noteworthy that the [M_xM′_5–xFe₄(CO)₁₆]^3– clusters under ESI-MS conditions show a high tendency to lose one to four CO ligands, sometimes even five carbonyls. This suggests that in the gas phase their M_xM′_5–x cores could be stabilized by both Fe(CO)₄ and Fe(CO)₃ groups, whereas in the solid state only the former is observed. This is not the case for binary [M₅Fe₄(CO)₁₆]^3– clusters, suggesting that the tendency to lose CO ligands is somehow an alloy effect.

2.4. UV–Visible Studies of [M_xM′_5–xFe₄(CO)₁₆]^3– (M, M′ = Cu, Ag, Au; M ≠ M′; x = 0–5)

[M_xM′_5–xFe₄(CO)₁₆]^3– clusters have been studied by means of UV–visible spectroscopy in CH₃CN solution. The complete spectra can be found in Figures S38–S54 in the Supporting Information, whereas cumulative spectra are reported in Figures 4–6. The spectra of the binary [M₅Fe₄(CO)₁₆]^3– clusters have been recorded as references. A list of the main absorption bands of all of the [M_xM′_5–xFe₄(CO)₁₆]^3– clusters with extinction coefficients are reported in Tables S11–S13.

Figure 4.

UV–visible absorption spectra of [NEt₄]₃[Au_xCu_5–xFe₄(CO)₁₆] in CH₃CN at 298 K (concentration 1.25 × 10^–5 M). Cu5 = [Cu₅Fe₄(CO)₁₆]^3–; Au1.09Cu3.91 = [Au_1.09Cu_3.91Fe₄(CO)₁₆]^3–; Au1.15Cu3.85 = [Au_1.15Cu_3.85Fe₄(CO)₁₆]^3–; Au1.31Cu3.69 = [Au_1.31Cu_3.69Fe₄(CO)₁₆]^3–; Au1.67Cu3.33 = [Au_1.67Cu_3.33Fe₄(CO)₁₆]^3–; Au2.18Cu2.82 = [Au_2.18Cu_2.82Fe₄(CO)₁₆]^3–; Au2.48Cu2.52 = [Au_2.48Cu_2.52Fe₄(CO)₁₆]^3–; Au4.62Cu0.38 = [Au_4.62Cu_0.38Fe₄(CO)₁₆]^3–.

Figure 6.

UV–visible absorption spectra of [NEt₄]₃[Au_xAg_5–xFe₄(CO)₁₆] in CH₃CN at 298 K (concentration 1.25 × 10^–5 M). Ag5 = [Ag₅Fe₄(CO)₁₆]^3–; Au0.64Ag4.36 = [Au_0.64Ag_4.36Fe₄(CO)₁₆]^3–; Au0.82Ag4.18 = [Au_0.82Ag_4.18Fe₄(CO)₁₆]^3–.

[Cu₅Fe₄(CO)₁₆]^3– displays three weak features at 266, 331, and 408 nm; that at 331 nm is very weak. [Ag₅Fe₄(CO)₁₆]^3– displays a strong absorption at 326 nm accompanied by two weak shoulders at 283 and 399 nm. The strong feature at 326 nm seems to be predictive of the presence of Ag in the cluster (see below). [Au₅Fe₄(CO)₁₆]^3– is readily oxidized after the dilution necessary for UV–vis spectroscopy, and therefore, its spectrum has not been recorded.

In the case of ternary [Au_xCu_5–xFe₄(CO)₁₆]^3– clusters (Table S11, Figure 4), the UV–visible spectra show only weak features. As a general trend, the weak absorption at ca. 269 nm increases by increasing the Cu content, whereas that at 347–367 nm increases by increasing the Au content. A weak absorption at 428–435 nm appears when significant amounts of Cu are present (compositions from Cu_3.91Au_1.09 to Cu_2.52Au_2.48), whereas it is absent in the case of Cu_0.38Au_4.62. It must be remarked that all of these features are very weak and, therefore, sometimes it is not easy to clearly detect them.

In the case of [Ag_xCu’_5–xFe₄(CO)₁₆]^3– clusters (Table S12, Figure 5), there is a strong absorption at 326–337 nm whose intensity decreases by decreasing the Ag content, accompanied by two weaker features at lower and higher wavelengths. The strong feature at 326–337 nm is indicative of the presence of Ag in the clusters, as also indicated by the UV–visible spectra of [Au_xAg_5–xFe₄(CO)₁₆]^3– clusters (Table S13, Figure 6), whereas it is completely absent in those of [Au_xCu_5–xFe₄(CO)₁₆]^3–.

Figure 5.

UV–visible absorption spectra of [NEt₄]₃[Ag_xCu_5–xFe₄(CO)₁₆] in CH₃CN at 298 K (concentration 1.25 × 10^–5 M). Ag5 = [Ag₅Fe₄(CO)₁₆]^3–; Ag5Cu0 = [Ag₅Cu₀Fe₄(CO)₁₆]^3– (see entry 2 in Table 1); Ag4.37Cu0.63 = [Ag_4.37Cu_0.63Fe₄(CO)₁₆]^3–; Ag4.25Cu0.75 = [Ag_4.25Cu_0.753Fe₄(CO)₁₆]^3–; Ag3.45Cu1.55 = [Ag_3.45Cu_1.55Fe₄(CO)₁₆]^3–; Ag3.31Cu1.69 = [Ag_3.31Cu_1.69Fe₄(CO)₁₆]^3–; Ag1.02Cu3.98 = [Ag_1.02Cu_3.98Fe₄(CO)₁₆]^3–; Cu5 = [Cu₅Fe₄(CO)₁₆]^3–.

2.5. DFT Investigation on [M_xM′_5–xFe₄(CO)₁₆]^3– Clusters (M, M′ = Cu, Ag, Au; M ≠ M′; x = 0–5)

The general representation of the M–M and M–Fe bonds obtained from the analysis of the (3, –1) bond critical points (b.c.p.) is depicted in Scheme 6. Selected computed data at (3, –1) b.c.p. (electron density, ρ; potential energy density, V; energy density, E; Laplacian of electron density, ∇²ρ) for all of the possible isomers of [M_xM′_5–xFe₄(CO)₁₆]^3– clusters are collected in the Supporting Information (Tables S14–S25). The isomers are sketched with their acronyms in Scheme 7 for clarity. The small negative values of E and the positive values of ∇²ρ at b.c.p. agree in all of the cases with Bianchi’s definition of metal–metal bonds.⁵⁰

Scheme 6. General Representation of the M–M and M–Fe Bonds Obtained from the Analysis of the (3, –1) Bond Critical Points.

Types of interactions: (a) M_centre–M_corner; (b) M_corner–M_corner; (c) M_centre–Fe; (d and e) M_corner–Fe.

Scheme 7. Isomers of the [M_xM′_5–xFe₄(CO)₁₆]^3– Clusters with Acronyms.

Only the metal centers are sketched for clarity.

For what concerns the M–M interactions, the following trends are observed:

• (1)On considering type (a) interactions (M_center–M_corner), (3, –1) b.c.p. for Cu–Cu bonds are observed only without a Cu–Ag or Cu–Au bond at the same side of the molecule. (3, –1) b.c.p. for bonds involving other couples of attractors are always detected. On considering different Cu–Ag, Cu–Au, or Ag–Ag (3, –1) b.c.p. in the same molecule, the electron density (ρ) is lower (and the potential energy density V is less negative) when the other M–M bond at the same side of the molecule involves heavier atoms. Ag–Au and Au–Au b.c.p. properties are less affected by the adjacent bonds.
• (2)For what concerns type (b) interactions (M_corner–M_corner), (3, –1) b.c.p. for Cu–Cu and Cu–Ag bonds are never observed. Cu–Au, Ag–Ag, or Au–Au (3, –1) b.c.p. are detected only when the central metal is Cu. In all of the other cases, no (3, –1) b.c.p. is localized.

These data suggest a competition in the localization of electron density among type (a) and (b) bonds. The Cu–Cu bonds appear the least favorable, and type (a) (3, –1) b.c.p. were detected only in the absence of Cu–Ag or Cu–Au bonds close to Cu–Cu. Moreover, Cu as a central metal favors the localization of type (b) (3, –1) b.c.p., that are absent in the other cases. This last condition is however nonsufficient to localize (3, –1) Cu–Ag b.c.p.

The M–Fe bonds show lower variability, since 10 (3, –1) Fe–M b.c.p. are localized for all of the considered isomers. The general trends observed are the following:

• (1)M–Fe interactions with the same M and with Fe atoms bonded to the same coinage metals follow the strength order (e) > (d) > (c), as indicated by ρ and V values at (3, –1) b.c.p. This is in agreement with the fact that in (e) a Fe(CO)₄ group is bonded to two M sites, whereas in (c) and (d) the Fe(CO)₄ group is bonded to three M sites.
• (2)On comparing bonds of the same type, the M–Fe density values at (3, –1) b.c.p. are slightly lowered if Fe is involved in another interaction with a heavier atom, with this suggesting the preferential localization of electron density on Fe–Ag and Fe–Au bonds.

The comparison of the relative energy values among the isomers, reported in Figure 7, allows one to qualitatively recognize the dominant parameters determining the thermodynamic stability. In the case of [Cu_xAg_5–xFe₄(CO)₁₆]^3– and [Cu_xAu_5–xFe₄(CO)₁₆]^3– clusters, the most stable species are those where the number of type (e) and (d) Fe–Ag or Fe–Au bonds is maximized. The energy separations are more pronounced among the Au-containing species, probably because the energy difference between Fe–Cu and Fe–Au bonds is greater than that between Fe–Cu and Fe–Ag. The same behavior is observed for the isomers of [Ag_xAu_5–xFe₄(CO)₁₆]^3–, where the clusters containing the highest numbers of Fe–Au bonds have the lowest relative energy. In agreement with the previous observation, the energy separations are lower than those computed for the [Cu_xAu_5–xFe₄(CO)₁₆]^3– clusters.

Figure 7.

Relative energy values for the isomers of [M_xM′_5–xFe₄(CO)₁₆]^3– clusters. Dotted lines are drawn for clarity purposes.

The determination of the most stable species is doubtful only for [M₃M′₂Fe₄(CO)₁₆]^3– clusters, where three isomers (M₃M′₂-is2, M₃M′₂-is3, and M₃M′₂-is4) have roughly comparable energies, as expected because of the comparable Fe–M bond structure. Only in the case of [Cu₃Au₂Fe₄(CO)₁₆]^3–, the Cu₃Au₂-is3 isomer is appreciably less stable than Cu₃Au₂-is2 and Cu₃Au₂-is4.

To summarize, it appears that M–M and Fe–M_center interactions (a–c) play a role of secondary importance in determining the stability of the compounds with respect to the Fe–M_corner bonds (d and e). The nature of M in these last interactions is a key parameter, because Fe–M bonds with heavier group 11 atoms, gold in particular, are preferred.

3. Conclusions

The synthesis of several 2-D molecular alloy clusters of general formula [M_xM′_5–xFe₄(CO)₁₆]^3– (M, M′ = Cu, Ag, Au; M ≠ M′; x = 0–5) has been reported. These have been fully characterized by means of ESI-MS, IR, and UV–visible spectroscopies and their total structures determined through SC-XRD. Substitutional and compositional disorder are present in the solid state structures of these alloy molecular clusters. The latter refers to the fact that [M_xM′_5–xFe₄(CO)₁₆]^3– are actually mixtures of clusters with slightly different M/M′ composition. Conversely, substitutional disorder indicates that also mixtures of clusters with the same composition but different distributions of M and M′ (isomers) are present. ESI-MS analyses point out that also in solution mixtures of isostructural clusters differing by a few M/M′ coinage metals are present. SC-XRD studies indicate a preferential distribution of Cu, Ag, and Au on the central and corner sites of the clusters. Thus, Au prefers to occupy the corner sites, whereas Cu displays a pronounced preference for the central site. Ag is somehow in the middle, and it is found preferentially in the central site in combination with Au and the corner sites in Ag–Cu clusters. This SC-XRD experimental evidence has been further corroborated by means of DFT studies. The site preference of the different coinage metals is mainly dictated by the strength of the Fe–M_corner bonds that decreases in the order Fe–Au > Fe–Ag > Fe–Cu. The preferred structure is the one that maximizes the number of stronger Fe–M interactions. Thus, Au always prefers corner sites, Cu the central site, whereas Ag prefers corner sites in combination with Cu and the central site in combination with Au. Shared-shell and metallophilic M–M interactions have lower weight in the determination of the most stable structures.

Overall, organometallic fragments such as Fe(CO)₄ can be used in order to stabilize 2-D molecular alloy clusters. In turn, their molecular nature allows their structures to be fully revealed with atomic detail, resulting in the elucidation of the bonding parameters that determine the distribution of the different metals within their metal cores. This corroborates the idea that molecular clusters and nanoclusters may help our understanding of metal nanoparticles, metal aggregates, and surfaces.^{1−6,14−27,51}

4. Experimental Section

4.1. General Experimental Procedures

All reactions and sample manipulations were carried out using standard Schlenk techniques under nitrogen and in dried solvents. All of the reagents were commercial products (Aldrich) of the highest purity available and used as received, except for [NEt₄]₃[Cu₃Fe₃(CO)₁₂],³² [NEt₄]₃[Cu₅Fe₄(CO)₁₆],³² [NEt₄]₄[Ag₄Fe₄(CO)₁₆],³⁴ and [NEt₄]₃[Ag₅Fe₄(CO)₁₆],³⁴ which were prepared according to the literature. Analyses of C, H, and N were obtained with a Thermo Quest Flash EA 1112NC instrument. Analyses of Fe, Cu, Ag, and Au were performed by microwave plasma-atomic emission spectrometry on an Agilent 4210 MP-AES instrument. IR spectra were recorded on a PerkinElmer Spectrum One interferometer in CaF₂ cells. ESI mass spectra were recorded on a Waters Micromass ZQ4000 instrument using CH₃CN as solvent (source temperature = 150 °C; capillary voltage = 2.54 kV; infusion flow = 20 μL/min; cone voltage = 10 V). Absorption spectra were recorded at 298 K using an Agilent Cary 100 UV–vis spectrometer. Structure drawings have been performed with SCHAKAL99.⁵²

4.2. Synthesis of [Et₄N]₃[Ag_xCu_5–xFe₄(CO)₁₆] from [NEt₄]₃[Cu₃Fe₃(CO)₁₂] and AgNO₃

A variable volume (V_AgNO3, see list below) of a solution of AgNO₃ (0.230 g, 1.36 mmol) in CH₃CN (10 mL) was added dropwise to a solution of [NEt₄]₃[Cu₃Fe₃(CO)₁₂] (0.468 g, 0.432 mmol) in CH₃CN (15 mL). The mixture was stirred for 1 h at room temperature, and then, the solvent was removed under reduced pressure. The residue was washed with H₂O (2 × 15 mL) and toluene (2 × 15 mL) and extracted with solvents of increasing polarity (acetone, CH₃CN, dmf). The resulting solutions were analyzed by IR spectroscopy and, eventually, layered with an appropriate solvent (acetone/n-hexane; CH₃CN/n-hexane/di-iso-propyl-ether; dmf/isopropanol) in the attempt to obtain crystals suitable for X-ray diffraction. The results were the following:

V_AgNO3 = 2.55 mL (0.344 mmol)

Crystals of [Et₄N]₃[Ag_1.02Cu_3.98Fe₄(CO)₁₆] were obtained by slow diffusion of isopropanol on the dmf fraction (yield 0.150 g, 31% based on Ag, 32% based on Cu).

C₄₀H₆₀Ag_1.02Cu_3.98Fe₄N₃O₁₆ (1425.44): calcd (%): C 33.71, H 4.24, N 2.95, Fe 15.67, Cu 17.75, Ag 7.72; found: C 33.92, H 4.38, N 2.76, Fe 15.86, Cu 17.59, Ag 7.47. IR (nujol, 293 K) ν_CO: 1943(s), 1905(m), 1854(s), 1829(sh) cm^–1. IR (acetone, 293 K) ν_CO: 1939(s), 1919(sh), 1897(m), 1845(s) cm^–1. IR (CH₃CN, 293 K) ν_CO: 1943(s), 1922(sh), 1878(m), 1824(sh) cm^–1.

V_AgNO3 = 4.13 mL (0.562 mmol)

Crystals of [Et₄N]₃[Ag_4.25Cu_0.75Fe₄(CO)₁₆] were obtained by slow diffusion of isopropanol on the dmf fraction (yield 0.074 g, 36% based on Ag, 3% based on Cu).

C₄₀H₆₀Ag_4.25Cu_0.75Fe₄N₃O₁₆ (1568.63): calcd (%): C 30.63, H 3.86, N 2.68, Fe 14.24, Cu 3.04, Ag 29.23; found: C 30.41, H 4.04, N 2.39, Fe 13.07, Cu 2.91, Ag 29.44. IR (nujol, 293 K) ν_CO: 1943(s), 1897(m), 1867(sh), 1851(s) cm^–1. IR (CH₃CN, 293 K) ν_CO: 1980(sh), 1965(ms), 1950(s), 1895(sh), 1879(vs), 1835(w) cm^–1.

V_AgNO3 = 6.67 mL (0.907 mmol)

Crystals of [Et₄N]₃[Ag_4.88Cu_0.12Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane and di-iso-propyl-ether on the CH₃CN fraction (yield 0.081 g, 27% based on Ag, 0.5% based on Cu).

C₄₀H₆₀Ag_4.88Cu_0.12Fe₄N₃O₁₆ (1596.34): calcd (%): C 30.10, H 3.79, N 2.63, Fe 13.99, Cu 0.48, Ag 32.98; found: C 30.38, H 3.59, N 2.97, Fe 14.12, Cu 0.69, Ag 32.75. IR (nujol, 293 K) ν_CO: 1944(s), 1892(m), 1869(sh), 1853(s) cm^–1. IR (CH₃CN, 293 K) ν_CO: 1947(s), 1882(m) cm^–1.

V_AgNO3 = 7.31 mL (0.994 mmol)

Crystals of [Et₄N]₃[Ag_4.92Cu_0.08Fe₄(CO)₁₆] were obtained by slow diffusion of isopropanol on the dmf fraction (yield 0.095 g, 29% based on Ag, 0.4% based on Cu).

C₄₀H₆₀Ag_4.92Cu_0.08Fe₄N₃O₁₆ (1598.11): calcd (%): C 30.06, H 3.78, N 2.63, Fe 13.98, Cu 0.32, Ag 33.21; found: C 29.86, H 4.05, N 2.32, Fe 13.78, Cu 0.21, Ag 33.02. IR (nujol, 293 K) ν_CO: 1947(s), 1891(s) cm^–1. IR (dmf, 293 K) ν_CO: 1942(s), 1878(s) cm^–1.

4.3. Synthesis of [Et₄N]₃[Ag_xCu_5–xFe₄(CO)₁₆] from [NEt₄]₃[Cu₅Fe₄(CO)₁₆] and AgNO₃

A solution of AgNO₃ (0.144 g, 0.850 mmol) in CH₃CN (5 mL) was added dropwise to a solution of [NEt₄]₃[Cu₅Fe₄(CO)₁₆] (0.468 g, 0.340 mmol) in acetone (15 mL). The mixture was stirred for 1 h at room temperature, and then, the solvent was removed under reduced pressure. The residue was washed with H₂O (2 × 15 mL) and toluene (2 × 15 mL) and extracted with solvents of increasing polarity (acetone, CH₃CN, dmf). The resulting solutions were analyzed by IR spectroscopy and, eventually, layered with an appropriate solvent (acetone/n-hexane; CH₃CN/n-hexane/di-iso-propyl-ether; dmf/isopropanol) in the attempt to obtain crystals suitable for X-ray diffraction.

Crystals of [Et₄N]₃[Ag₅Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane and di-iso-propyl-ether on the CH₃CN fraction (yield 0.038 g, 14% based on Ag).

C₄₀H₆₀Ag₅Cu₀Fe₄N₃O₁₆ (1601.66): calcd (%): C 30.00, H 3.78, N 2.62, Fe 13.95, Ag 33.67; found: C 30.31, H 3.51, N 2.99, C 30.00, H 3.78, N 2.62, Fe 14.12, Ag 33.49. IR (nujol, 293 K) ν_CO: 1944(m), 1890(sh), 1867(sh), 1853(s) cm^–1. IR (CH₃CN, 293 K) ν_CO: 1949(s), 1894(sh), 1878(s), 1833(w) cm^–1. IR (dmso, 293 K) ν_CO: 1944(s), 1895(sh), 1872(vs), 1833(m) cm^–1.

Crystals of [Et₄N]₃[Ag_4.81Cu_0.20Fe₄(CO)₁₆] were obtained by slow diffusion of isopropanol on the dmf fraction (yield 0.44 g, 16% based on Ag, 0.3% based on Cu).

C₄₀H₆₀Ag_4.81Cu_0.20Fe₄N₃O₁₆₃ (1593.01): calcd (%): C 30.14, H 3.79, N 2.63, Fe 14.02, Cu 0.80, Ag 32.55; found: C 29.95, H 3.61, N 2.82, Fe 14.25, Cu 0.67, Ag 32.74. IR (dmf, 293 K) ν_CO: 1945(vs), 1844(s) cm^–1.

4.4. Synthesis of [Et₄N]₃[Ag_4.37Cu_0.63Fe₄(CO)₁₆] from [NEt₄]₄[Ag₄Fe₄(CO)₁₆] and [Cu(CH₃CN)₄][BF₄]

[Cu(CH₃CN)₄][BF₄] (0.095 g, 0.306 mmol) was added as a solid in small portions to a solution of [NEt₄]₄[Ag₄Fe₄(CO)₁₆] (0.468 g, 0.289 mmol) in CH₃CN (15 mL). The mixture was stirred for 1 h at room temperature, and then, the solvent was removed under reduced pressure. The residue was washed with H₂O (2 × 15 mL) and toluene (2 × 15 mL) and extracted with CH₃CN (15 mL). Crystals of [Et₄N]₃[Ag_4.37Cu_0.63Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane and di-iso-propyl-ether on the CH₃CN fraction (yield 0.121 g, 29% based on Ag, 16% based on Cu).

C₄₀H₆₀Ag_4.37Cu_0.63Fe₄N₃O₁₆ (1573.73): calcd (%): C 30.53, H 3.84, N 2.67, Fe 14.19, Cu 2.54, Ag 29.95; found: C 30.37, H 4.02, N 2.83, Fe 14.02, Cu 2.68, Ag 30.14. IR (CH₃CN, 293 K) ν_CO: 1948(s), 1894(sh), 1878(s), 1833(w) cm^–1.

4.5. Synthesis of [Et₄N]₃[Ag_4.90Cu_0.10Fe₄(CO)₁₆] from [NEt₄]₃[Ag₅Fe₄(CO)₁₆] and Cu(IMes)Cl

Cu(IMes)Cl (0.354 g, 0.880 mmol) was added as a solid in small portions to a solution of [NEt₄]₃[Ag₅Fe₄(CO)₁₆] (0.468 g, 0.293 mmol) in CH₃CN (15 mL). The mixture was stirred for 1 h at room temperature, and then, the solvent was removed under reduced pressure. The residue was washed with H₂O (2 × 15 mL) and toluene (2 × 15 mL) and extracted with CH₃CN (15 mL). Crystals of [Et₄N]₃[Ag_4.90Cu_0.10Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane and di-iso-propyl-ether on the CH₃CN fraction (yield 0.160 g, 34% based on Ag, 1.1% based on Cu).

C₄₀H₆₀Ag_4.90Cu_0.10Fe₄N₃O₁₆ (1597.22): calcd (%): C 30.08, H 3.79, N 2.63, Fe 13.99, Cu 0.40, Ag 33.09; found: C 29.88, H 4.03, N 2.85, Fe 14.15, Cu 0.54, Ag 32.89. IR (CH₃CN, 293 K) ν_CO: 1950(s), 1895(sh), 1879(s), 1834(m) cm^–1.

4.6. Synthesis of [Et₄N]₃[Ag_xCu_5–xFe₄(CO)₁₆] from [NEt₄]₃[Ag₅Fe₄(CO)₁₆] and [NEt₄]₃[Cu₃Fe₃(CO)₁₂]

A solution of [NEt₄]₃[Cu₃Fe₃(CO)₁₂] (0.317 g, 0.293 mmol) in CH₃CN (15 mL) was added dropwise to a solution of [NEt₄]₃[Ag₅Fe₄(CO)₁₆] (0.468 g, 0.293 mmol) in CH₃CN (15 mL). The mixture was stirred for 1 h at room temperature, and then, the solvent was removed under reduced pressure. The residue was washed with H₂O (2 × 15 mL) and toluene (2 × 15 mL) and extracted with solvents of increasing polarity (acetone, CH₃CN, dmf). The resulting solutions were analyzed by IR spectroscopy and, eventually, layered with an appropriate solvent (acetone/n-hexane; CH₃CN/n-hexane/di-iso-propyl-ether; dmf/isopropanol) in the attempt to obtain crystals suitable for X-ray diffraction.

Crystals of [Et₄N]₃[Ag_3.30Cu_1.70Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane on the acetone fraction (yield 0.320 g, 47% based on Ag, 41% based on Cu).

C₄₀H₆₀Ag_3.30Cu_1.70Fe₄N₃O₁₆ (1526.30): calcd (%): C 31.48, H 3.96, N 2.75, Fe 14.64, Cu 7.08, Ag 23.32; found: C 31.32, H 4.12, N 2.96, Fe 14.48, Cu 6.87, Ag 23.57. IR (acetone, 293 K) ν_CO: 1942(s), 1898(m) cm^–1.

Crystals of [Et₄N]₃[Ag_3.45Cu_1.55Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane and di-iso-propyl-ether on the CH₃CN fraction (yield 0.100 g, 15% based on Ag, 12% based on Cu).

C₄₀H₆₀Ag_3.45Cu_1.55Fe₄N₃O₁₆ (1532.72): calcd (%): C 31.34, H 3.95, N 2.74, Fe 14.57, Cu 6.43, Ag 24.28; found: C 31.62, H 3.79, N 2.95, Fe 14.38, Cu 6.29, Ag 24.57. IR (CH₃CN, 293 K) ν_CO: 1946(m), 1894(sh), 1847(s) cm^–1.

4.7. Synthesis of [Cu(dppe)₂]₃[Ag₁₃Fe₈(CO)₃₂] from [NEt₄]₃[Cu₅Fe₄(CO)₁₆] and Ag(dppe)(NO₃)

Ag(dppe)(NO₃) (0.482 g, 0.850 mmol) was added as a solid in small portions to a solution of [NEt₄]₃[Cu₅Fe₄(CO)₁₆] (0.468 g, 0.340 mmol) in acetone (15 mL). The mixture was stirred for 1 h at room temperature, and then, the solvent was removed under reduced pressure. The residue was dissolved in dmf (10 mL), and further, Ag(dppe)(NO₃) (0.482 g, 0.850 mmol) was added as a solid. The mixture was stirred for 1 h at room temperature, and then, a saturated solution of [NEt₄]Br in H₂O (40 mL) was added up to complete precipitation of the products. The solid was recovered by filtration, washed with H₂O (2 × 15 mL), toluene (2 × 15 mL), thf (15 mL), acetone (15 mL), and CH₃CN (15 mL) and eventually extracted with dmf (10 mL). Crystals of [Cu(dppe)₂]₃[Ag₁₃Fe₈(CO)₃₂] were obtained by slow diffusion of isopropanol on the dmf solution (yield 0.104 g, 15% based on Ag, 3.4% based on Cu).

While repeating this reaction, a few crystals of Cu₃Br₃(dppe)₃ were also obtained as a side product and mechanically separated from [Cu(dppe)₂]₃[Ag₁₃Fe₈(CO)₃₂].

C₁₈₈H₁₄₄Ag₁₃Cu₃Fe₈O₃₂P₁₂ (5323.39): calcd (%): C 42.29, H 2.72, Fe 8.39, Cu 3.58, Ag 26.33; found: C 42.67, H 2.35, Fe 8.53, Cu 3.29, Ag 26.47. IR (nujol, 293 K) ν_CO: 1994(s), 1912(m) cm^–1. IR (dmf, 293 K) ν_CO: 1999(vs), 1916(ms) cm^–1.

4.8. Synthesis of [Et₄N]₃[Au_xCu_5–xFe₄(CO)₁₆] from [NEt₄]₃[Cu₃Fe₃(CO)₁₂] and Au(PPh₃)Cl

A variable amount (m_Au(PPh3)Cl, see list below) of Au(PPh₃)Cl was added as a solid to a solution of [NEt₄]₃[Cu₃Fe₃(CO)₁₂] (0.468 g, 0.432 mmol) in CH₃CN (15 mL). The mixture was stirred for 1 h at room temperature, and then, the solvent was removed under reduced pressure. The residue was washed with H₂O (2 × 15 mL) and toluene (2 × 15 mL) and extracted with solvents of increasing polarity (acetone, CH₃CN, dmf). The resulting solutions were analyzed by IR spectroscopy and, eventually, layered with an appropriate solvent (acetone/n-hexane; CH₃CN/n-hexane/di-iso-propyl-ether; dmf/isopropanol) in the attempt to obtain crystals suitable for X-ray diffraction. The results were the following:

m_Au(PPh3)Cl = 0.284 g (0.605 mmol)

Crystals of [Et₄N]₃[Au_1.15Cu_3.85Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane on the acetone fraction (yield 0.160 g, 20% based on Ag, 31% based on Cu).

C₄₀H₆₀Au_1.15Cu_3.85Fe₄N₃O₁₆ (1534.11): calcd (%): C 31.33, H 3.94, N 2.74, Fe 14.57, Cu 15.95, Au 14.77; found: C 31.08, H 4.13, N 2.53, Fe 14.29, Cu 16.11, Au 14.89. IR (acetone, 293 K) ν_CO: 1946(s), 1884(s), 1831(w) cm^–1. IR (dmf, 293 K) ν_CO: 1946(s), 1881(s).

m_Au(PPh3)Cl = 0.305 g (0.648 mmol)

Crystals of [Et₄N]₃[Au_1.31Cu_3.69Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane on the acetone fraction (yield 0.168 g, 22% based on Au, 31% based on Cu).

C₄₀H₆₀Au_1.31Cu_3.69Fe₄N₃O₁₆ (1555.46): calcd (%): C 30.90, H 3.89, N 2.70, Fe 14.37, Cu 15.08, Au 16.60; found: C 30.77, H 4.04, N 2.56, Fe 14.19, Cu 14.87, Au 16.91. IR (acetone, 293 K) ν_CO: 1947(s), 1882(m) cm^–1.

Crystals of [Et₄N]₃[Au_1.67Cu_3.33Fe₄(CO)₁₆] were obtained by slow diffusion of isopropanol on the dmf fraction (yield 0.112 g, 18% based on Au, 18% based on Cu).

C₄₀H₆₀Au_1.67Cu_3.33Fe₄N₃O₁₆ (1602.83): calcd (%): C 29.97, H 3.77, N 2.62, Fe 13.94, Cu 13.20, Au 20.52; found: C 30.14, H 4.02, N 2.36, Fe 13.81, Cu 13.49, Au 20.21. IR (dmf, 293 K) ν_CO: 1945(s), 1878(m) cm^–1.

m_Au(PPh3)Cl = 0.610 g (1.30 mmol)

Crystals of [Et₄N]₃[Au_2.48Cu_2.52Fe₄(CO)₁₆] were obtained by slow diffusion of isopropanol on the dmf fraction (yield 0.130 g, 14% based on Au, 15% based on Cu).

C₄₀H₆₀Au_2.48Cu_2.52Fe₄N₃O₁₆ (1710.90): calcd (%): C 28.08, H 3.53, N 2.46, Fe 13.06, Cu 9.36, Au 28.55; found: C 28.27, H 3.31, N 2.62, Fe 13.21, Cu 9.08, Au 28.27. IR (dmf, 293 K) ν_CO: 1951(2), 1887(2) cm^–1.

4.9. Synthesis of [Et₄N]₃[Au_xCu_5–xFe₄(CO)₁₆] from [NEt₄]₃[Cu₃Fe₃(CO)₁₂] and Au(Et₂S)Cl

A variable volume (V_Au(Et2S)Cl, see list below) of a solution of Au(Et₂S)Cl (0.348 g, 1.08 mmol) in CH₃CN (10 mL) was added dropwise to a solution of [NEt₄]₃[Cu₃Fe₃(CO)₁₂] (0.468 g, 0.432 mmol) in CH₃CN (15 mL). The mixture was stirred for 1 h at room temperature, and then, the solvent was removed under reduced pressure. The residue was washed with H₂O (2 × 15 mL) and toluene (2 × 15 mL) and extracted with solvents of increasing polarity (acetone, CH₃CN, dmf). The resulting solutions were analyzed by IR spectroscopy and, eventually, layered with an appropriate solvent (acetone/n-hexane; CH₃CN/n-hexane/di-iso-propyl-ether; dmf/isopropanol) in the attempt to obtain crystals suitable for X-ray diffraction. The results were the following:

V_Au(Et2S)Cl = 2.80 mL (0.302 mmol)

Crystals of [Et₄N]₃[Au_2.18Cu_2.83Fe₄(CO)₁₆] and [Et₄N]₃[Au_2.73Cu_2.28Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane and di-iso-propyl-ether on the CH₃CN fraction (yield 0.098 g, 47% based on Au, 11% based on Cu).

C₄₀H₆₀Au_2.73Cu_2.28Fe₄N₃O₁₆ (1743.59): calcd (%): C 27.53, H 3.47, N 2.41, Fe 12.80, Cu 8.30, Au 30.82; C₄₀H₆₀Au_2.18Cu_2.83Fe₄N₃O₁₆ (1670.21): calcd (%): C 28.74, H 3.62, N 2.52, Fe 13.36, Cu 10.76, Au 25.69; found (mixture): C 28.25, H 3.34, N 2.63, Fe 13.05, Cu 9.12, Au 28.55. IR (nujol, 293 K) ν_CO: 1948(s), 1911(m), 1864(s) cm^–1. IR (acetone, 293 K) ν_CO: 1999(w), 1974(m), 1951(vs), 1922(s), 1864(s) cm^–1. IR (CH₃CN, 293 K) ν_CO: 1950(s), 1915(s), 1864(s) cm^–1.

V_Au(Et2S)Cl = 7.60 mL (0.821 mmol)

Crystals of [Et₄N]₃[Au_4.59Cu_0.42Fe₄(CO)₁₆] and [Et₄N]₃[Au_4.62Cu_0.38Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane and di-iso-propyl-ether on the CH₃CN fraction (yield 0.104 g, 29% based on Au, 1.6% based on Cu).

C₄₀H₆₀Au_4.62Cu_0.38Fe₄N₃O₁₆₃ (1995.77): calcd (%): C 24.09, H 3.03, N 2.11, Fe 11.20, Cu 1.21, Au 45.53; C₄₀H₆₀Au_4.59Cu_0.42Fe₄N₃O₁₆₃ (1991.77): calcd (%): C 24.11, H 3.03, N 2.11, Fe 11.21, Cu 1.34, Au 45.36; found (mixture): C 24.21, H 2.86, N 2.29, Fe 11.34, Cu 1.16, Au 45.61. IR (nujol, 293 K) ν_CO: 1954(s), 1869(s) cm^–1. IR (acetone, 293 K) ν_CO: 1956(s), 1930(ms), 1902(m) cm^–1.

4.10. Synthesis of [Et₄N]₃[Au_1.09Cu_3.91Fe₄(CO)₁₆] from [NEt₄]₃[Cu₅Fe₄(CO)₁₆] and Au(Et₂S)Cl

A solution of Au(Et₂S)Cl (0.131 g, 0.408 mmol) in acetone (5 mL) was added dropwise to a solution of [NEt₄]₃[Cu₅Fe₄(CO)₁₆] (0.468 g, 0.340 mmol) in acetone (15 mL). The mixture was stirred for 1 h at room temperature, and then, the solvent was removed under reduced pressure. The residue was washed with H₂O (2 × 15 mL) and toluene (2 × 15 mL) and extracted with acetone. Crystals of [Et₄N]₃[Au_1.09Cu_3.91Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane on the acetone fraction (yield 0.32 g, 56% based on Ag, 48% based on Cu).

C₄₀H₆₀Au_1.09Cu_3.91Fe₄N₃O₁₆ (1525.44): calcd (%): C 31.49, H 3.96, N 2.75, Fe 14.64, Cu 16.29, Au 14.07; found: C 31.79, H 4.12, N 2.54, Fe 14.39, Cu 16.47, Au 13.87. IR (acetone, 293 K) ν_CO: 1945(s), 1880(m) cm^–1.

4.11. Synthesis of [Et₄N]₃[Au_xAg_5–xFe₄(CO)₁₆] from [NEt₄]₄[Ag₄Fe₄(CO)₁₆] and Au(Et₂S)Cl

A solution of Au(Et₂S)Cl (0.075 g, 0.231 mmol) in CH₃CN (4 mL) was added dropwise to a solution of [NEt₄]₄[Ag₄Fe₄(CO)₁₆] (0.468 g, 0.289 mmol) in CH₃CN (15 mL). The mixture was stirred for 1 h at room temperature, and then, the solvent was removed under reduced pressure. The residue was washed with H₂O (2 × 15 mL) and toluene (2 × 15 mL) and extracted with solvents of increasing polarity (acetone, CH₃CN, dmf). The resulting solutions were analyzed by IR spectroscopy and, eventually, layered with an appropriate solvent (acetone/n-hexane; CH₃CN/n-hexane/di-iso-propyl-ether; dmf/isopropanol) in the attempt to obtain crystals suitable for X-ray diffraction.

Crystals of [Et₄N]₃[Au_0.81Ag_4.20Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane on the acetone fraction (yield 0.099 g, 21% based on Au, 21% based on Ag).

C₄₀H₆₀Au_0.81Ag_4.20Fe₄N₃O₁₆ (1673.38): calcd (%): C 28.68, H 3.61, N 2.51, Fe 13.34, Ag 27.05, Au 9.53; found: C 29.03, H 3.39, N 2.24, Fe 13.49, Ag 26.85, Au 9.78. IR (acetone, 293 K) ν_CO: 1946(vs), 1927(m), 1896(ms), 1874(s), 1835(sh) cm^–1.

Crystals of [Et₄N]₃[Au_0.64Ag_4.36Fe₄(CO)₁₆] were obtained by slow diffusion of n-hexane and di-iso-propyl-ether on the CH₃CN fraction (yield 0.180 g, 30% based on Au, 41% based on Ag).

C₄₀H₆₀Au_0.64Ag_4.36Fe₄N₃O₁₆ (1658.68): calcd (%): C 28.96, H 3.65, N 2.53, Fe 13.47, Ag 28.35, Au 7.60; found: C 29.31, H 3.39, N 2.22, Fe 13.22, Ag 28.04, Au 7.85. IR (CH₃CN, 293 K) ν_CO: 1967(sh), 1950(s), 1930(sh), 1880(s), 1835(sh) cm^–1.

4.12. X-ray Crystallographic Study

Crystal data and collection details for [NEt₄]₃[M_xM′_5–xFe₄(CO)₁₆] (x = 0–5; M, M′ = Cu, Ag, Au; M ≠ M′) are reported in Tables S26–S29. The diffraction experiments were carried out on a Bruker APEX II diffractometer equipped with a PHOTON100 detector using Mo Kα radiation. Data were corrected for Lorentz polarization and absorption effects (empirical absorption correction SADABS).⁵³ Structures were solved by direct methods and refined by full-matrix least-squares based on all data using F².⁵⁴ Hydrogen atoms were fixed at calculated positions and refined by a riding model. All non-hydrogen atoms were refined with anisotropic displacement parameters, unless otherwise stated. [NEt₄]₃[M_xM′_5–xFe₄(CO)₁₆] (x = 0–5; M, M′ = Cu, Ag, Au; M ≠ M′) are isomorphous with [NEt₄]₃[M₅Fe₄(CO)₁₆] (M = Cu, Ag, Au).^28,32,34 The salts analyzed displayed some disorder of the CO ligands and/or [NEt₄]⁺ cations. This, in some cases, lowered the symmetry of the space group from P4₂/mnm to P4̅2₁m. In the case of the lower symmetry P4̅2₁m space group, the crystals were refined as racemic twins. The positions occupied by M(1) and M(2) are disordered Cu/Ag, Cu/Au, or Ag/Cu.

4.13. Computational Details

Geometry optimizations of the clusters were performed in the gas phase without symmetry constraints using the PBEh-3c method, a modified version of PBE0 (with 42% HF exchange) that uses a split-valence double-ζ basis set (def2-mSVP) and adds three corrections that consider dispersion, basis set superposition, and other basis set incompleteness effects.⁵⁵ The “restricted” DFT approach was always applied. Calculations were performed with ORCA 4.2.0.⁵⁶ The output, converted in .molden format, was elaborated with the software Multiwfn, version 3.5.⁵⁷ Cartesian coordinates of all of the DFT-optimized structures have been included in the Supporting Information as an .xyz file.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.inorgchem.0c02443.

• Crystallographic data in CIF format, IR spectra, ESI-MS spectra, UV–visible absorption spectra, selected computed data for all of the isomers, and crystals and experimental details (PDF)

• Optimized coordinates in XYZ format (XYZ)

Accession Codes

CCDC 2016042–2016064 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by emailing data_request@ccdc.cam.ac.uk, or by contacting The Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223 336033.

The authors declare no competing financial interest.