Advances for Triangular and Sandwich-Shaped All-Metal Aromatics

Abstract

Much experimental work has been contributed to all-metal σ, π and δ-aromaticity among transition metals, semimetallics and other metals in the past two decades. Before our focused investigations on the properties of triangular and sandwich-shaped all-metal aromatics, A. I. Boldyrev presented general discussions on the concepts of all-metal σ-aromaticity and σ-antiaromaticity for metallo-clusters. Schleyer illustrated that Nucleus-Independent Chemical Shifts (NICS) were among the most authoritative criteria for aromaticity. Ugalde discussed the earlier developments of all-metal aromatic compounds with all possible shapes. Besides the theoretical predictions, many stable all-metal aromatic trinuclear clusters have been isolated as the metallic analogues of either the σ-aromatic molecule’s [H₃]⁺ ion or the π-aromatic molecule’s [C₃H₃]⁺ ion. Different from Hoffman’s opinion on all-metal aromaticity, triangular all-metal aromatics were found to hold great potential in applications in coordination chemistry, catalysis, and material science. Triangular all-metal aromatics, which were theoretically proved to conform to the Hückel (4n + 2) rule and possess the smallest aromatic ring, could also play roles as stable ligands during the formation of all-metal sandwiches. The triangular and sandwich-shaped all-metal aromatics have not yet been specifically summarized despite their diversity of existence, puissant developments and various interesting applications. These findings are different from the public opinion that all-metal aromatics would be limited to further applications due to their overstated difficulties in synthesis and uncertain stabilities. Our review will specifically focus on the summarization of theoretical predictions, feasible syntheses and isolations, and multiple applications of triangular and sandwich shaped all-metal aromatics. The appropriateness and necessities of this review will emphasize and disseminate their importance and applications forcefully and in a timely manner.

1. Introduction

The definition for aromaticity was originally proposed in the 19th century when August Kekulé proposed the typical structure of benzene, C₆H₆ [1,2,3]. Since then, aromaticity has become a popular concept and was wildly used in organic chemistry to demonstrate molecules possessing conjugated (4n + 2) π electrons. After vigorous and long development, this concept has been extended to numerous unsaturated organic species including the smallest cyclopropenyl cation and much larger polycyclic aromatics [4]. All-metal aromatics are a group of metallic rings that also conform to the Hückel (4n + 2) electrons rule. Depending on the type of electrons (s, p or d) delocalized among the metallic core, these all-metal aromatics possess σ, π or δ-aromaticity, respectively. More recently, the voice that the concept of aromaticity should not be extended beyond carbon-based systems is present because they believe that aromatic organometallics are very likely to be fragile. However, the discovery of all-metal aromaticity started from the theoretical predictions of aromaticity for metallocyclopentadienyls [5,6,7,8,9,10]. Later on, several organometallic compounds consisting of metallic aromatic rings have been isolated with fully qualified characterizations [11,12,13]. Before the vivant investigations on triangular all-metal aromatics, A. I. Boldyrev presented general standards to identify the presence of all-metal aromaticity and antiaromaticity [3,14,15]. Meanwhile, Schleyer introduced Nucleus-Independent Chemical Shifts (NICS) in several papers and established a mature system to predict or prove aromaticity by this method [16,17,18]. Ugalde reviewed the earlier developments of all-metal aromatic compounds and encouraged their promising prospects for further utilization [19]. Later on, the advances in all-metal aromaticity were driven forcefully by several groups [20,21,22,23]. However, we are going to pay more attention to triangular all-metal aromatics and specifically summarize these group of organometallics due to its multiple theoretical predictions and existences, rapid developments and exciting application potentials [24,25].

Triangular all-metal aromatics are a group of organometallic complexes that conform to the Hückel (4n + 2) π electrons rule and contain the smallest metallic ring core. In fact, before the successful syntheses and isolation for solid-state triangular all-metal complexes, a great amount of work on theoretical calculations has already predicted the existence of triangular all-metal σ, π or δ-aromaticity among transition metals, semimetallics and other metals, even before 2010 [26,27,28,29,30,31,32,33]. Besides the theoretical predictions, all-metal aromaticity in stable trinuclear clusters has been disclosed by many chemists and these metallic clusters were proved to be either analogues of the σ-aromatic molecule’s [H₃]⁺ ion [34] or analogues of the π-aromatic molecule’s [C₃H₃]⁺ ion (Scheme 1) [35]. Besides the triangles, the discovery of aromaticity in four-membered all-metal clusters, like [Al₄]²⁻, also showed considerable research activities [36,37].

Here, in this review, we focus on the theoretical prediction and feasible synthesis of all types of present triangular and sandwich-shaped all-metal aromatics and their rapidly developing applications. The appropriateness of this summarization will boost the developments and applications of this new family of all-metal aromatics and debate Hoffman’s opinion on all-metal aromaticity [38]. We found all-metal aromatics showed a range of potential applications which differ from the public’s misunderstanding that they would be only the laboratory curiosity with gorgeous X-ray structures and limited to further utilizations due to their exaggerated complicated syntheses and low stabilities. As we know, even with the largest ring strain, the simplest aromatic molecular cyclopropenium ion [C₃H₃]⁺ is quite stable and most of its metallic analogues showed comparable stabilities. All the known triangular all-metal aromatics will be presented and discussed below, mainly including the syntheses, characterizations and theoretical calculations of isolated triangular all-metal aromatic clusters, predictions of aromaticity in structures theoretically proposed, and their varieties of applications in coordination chemistry, catalytic reactions, and material sciences.

2. Experimental Findings of Triangular All-Metal Aromatic Clusters

2.1. Triangular π-Aromatic [M₃]²⁻ (M = B, Al, Ga)

Group 13 triangular all-metal aromatics, which all possess three-centre-two-electron cores, are the earliest investigated triangular systems with aromaticity. In 1995, Robinson and co-workers reported the preparation of cyclogallene, Na₂[(Mes₂C₆H₃)Ga]₃ (Mes = 2,4,6-Me₃C₆H₂), which was the first isolated triangular all-metal aromatics. In this large organometallics, three gallium atoms were located on the three termini of the triangular ring core (Scheme 2) [39]. According to the crystal structure solved by X-ray diffraction, Na₂Ga₃ exhibited as a peculiar metallic trigonal bipyramid. The metallic core Na₂Ga₃ was arranged as two sodium atoms positioned above and beneath the [Ga₃]²⁻ plane with the bond lengths measured as Ga–Ga: 2.441(1)Å and Ga–Na: 3.220(2) Å, respectively. The authors claimed that the gallium atoms are mainly sp² hybridized according to the electronic properties of cyclogallenes. There was one unoccupied p orbital left for each gallium atom without participating in the formation of a sigma bond. So, each of the two sodium atoms could donate one electron to the unoccupied p orbitals of the gallium atoms to provide the two 2π-electrons delocalized on three centres which are required for Hückel’s (4n + 2) rule for the determination of aromaticity. Later, more groups found that group 13 M(I) metals could form cationic clusters easily via disproportionation reactions [40].

Besides experimental proof, theoretical calculations have been implemented in order to prove the existence of metalloaromaticity in the metallic ring dianion [Ga]₃²⁻. Through investigation into their electron clouds (Figure 1) [41], the authors certificated that cyclogallenes were metallic analogues of the smallest main group triangular aromatic structure: cyclopropenium cation. The HOMO-1 of the cyclogallene in Na₂[Ga₃H₃] presented extreme similarities to the π-electron cloud of the triangular aromatic [C₃H₃]⁺ and also possessed the same number of valence electrons. Given the fully recognized “aromatic” property of the cyclopropenium cation, the authors believed that it was justified to claim that triangular “all-metal aromaticity” appeared in these cyclogallene dianions. Combining these experimental findings with their theoretical investigations, Robinson concluded that their isolated Na₂[Ga₃H₃] was the first triangular all-metal π-aromatic structure.

Shortly after the introduction of Na₂[(Mes₂C₆H₃)Ga]₃, Robinson and co-workers reported the efficient synthesis and full characterizations of its analogous potassium complex K₂[(Mes₂C₆H₃)Ga]₃ (Equation (1)) [42] through the reaction of the corresponding gallium resource with potassium in 2007. K₂Ga₃ also contains a similar [Ga₃]²⁻ aromatic core and bears typical organic ligands. The crystal structure showed that the mean Ga–Ga–Ga bond angle is nearly 60.0° and the Ga–Ga bond distances ranged from 2.4187(5) to 2.4317(5) Å.

3(Mes₂C₆H₃)GaCl₂ + 8K → K₂[((Mes₂C₆H₃)Ga)]₃ + 6KCl

(1)

Quantum mechanical self-consistent field and density functional theory (DFT) were authoritative theoretical calculation tools and were exploited to theoretically examine the structural properties of the anionic cores [Ga₃]²⁻, [GaH]₃²⁻, Na₂[GaH]₃ and K₂[GaH]₃. Through full optimizations of their well-defined π molecular orbitals, they found these organometallic species were aromatic, as their calculated Independent Chemical Shifts (NICS) indices were largely negative. With solid experimental and theoretical results, they proved that the triangular metallic cyclogallene dianion, [Ga₃]²⁻, exhibited definite aromatic behavior [43]. Robinson’s group also proposed that when these three-membered rings were composed of mixed carbon, silicon atoms with gallium atoms could contribute heterocyclic 2π-electron cyclogallene systems which should also be synthetically accessible, stable, and aromatic [44]. Then, based on their huge amount of work in gallium-involved compounds, Robinson summarized short reviews in this field in which quite scientific comments were given on the structures and properties of triangular aromatic organometallic clusters [45].

In 2006, the Na₂Al₃ complex, which was an aluminium analogue for the cyclogallenes M₂Ga₃ (M = Na, K), was successfully isolated and theoretically demonstrated [46]. Through the rapid reduction process of RAlI₂ (R = Mes₂C₆H₃) using sodium metal, the three-aluminium core Na₂Al₃ was formed in diethyl ether and was successfully isolated after regular work up. According to the Kohn–Sham orbital representation for the delocalized HOMO-2 of Na₂Al₃ (Figure 2), the three aluminium atoms in Na₂Al₃ could also be considered as sp² hybridized, which was the same mode for the gallium atoms in cyclogallenes. X-ray structure of Na₂Al₃ showed the bond length of Al–Al was equal to 2.5202(2) Å on average in the aluminium ring core. Thus, Na₂Al₃ was the first characterized aromatic cycloaluminium complex.

Besides the heavier elements in group 13, the synthesis for the triboracyclopropenyl dianion was also reported in 2015. The boron-based analogue of the aromatic cyclopropenyl cation belongs to the prototypical Hückel π romatics. [B₃(NCy₂)₃]²⁻ was isolated as its dimeric Na⁺ salt (Na₄[B₃(NCy₂)₃]₂ (Figure 3A) [47], which was obtained through reduction of Cl₂BNCy₂ by sodium metal. Cyclic voltammetry measurements presented an extremely high oxidation potential (E_pc = −2.42 V), which was contributed by its good reactivity. The Hückel-type π aromatic character of the [B₃(NCy₂)₃]²⁻ dianion was characterized by single-crystal X-ray diffraction and verified by various theoretical methods. DFT measurements clearly indicated the π aromaticity of the [B₃]²⁻ core in a similar magnitude to that in cyclopropenium and benzene. In addition, the different ligands stabilized boron-based analogue (B₃N₃Ph₆)²⁻ was synthesized by the simple reduction of B₃N₃Ph₆ using either potassium or rubidium in the 18-crown-6 ether. Similarly, DFT indicated that two electrons delocalized over the three boron atoms in the triangular (B₃N₃Ph₆)²⁻, and the tri-boron core exhibited (π, σ)-mixed homoaromaticity. The reduced (B₃N₃Ph₆)²⁻ could also act as a robust two-electron reductant for unsaturated compounds in reduction reaction (Figure 3B) [48].

Besides the π-aromatic core [B₃]²⁻, Frenking’s group reported another smallest π-aromatic species B₃⁺ in the [B₃(NN)₃]⁺ and [B₃(CO)₃]⁺ complexes in 2016. For comparison of the two representative 2π-aromatic systems, cyclic group-13 cation complex [B₃L₃]⁺ featuring L→B dative bonds, and cyclic group-13 dianion [B₃R₃]²⁻ featuring R-B electron-sharing bonds (Figure 4A) [49]. Spectroscopic identification (mass spectrum and experimental IR spectrum) was obtained for these boron–nitrogen cation complexes. The mass for these cations was formed through pulsed laser vaporization of a boron-11. Fully quantum chemical bonding analysis showed that the ligated adducts were mainly stabilized by the σ-donation of L→[B₃L₂]⁺ (Figure 4B). This finding was a new field for ligand-stabilized boron complexes (B₃L₃), which could also be extended to main-group adducts with other atoms and ligands in three-membered cycles.

Furthermore, a noble-gas-supported aromatic B₃⁺ cluster was also obtained by the contribution of strong covalent noble gas–boron bonds [50]. The triangular ion [B₃]⁺ could also trap rare gases forming circular cationic compounds B₃Rg_n⁺ [51]. In fact, besides the [B₃]²⁻ and [B₃]⁺ core, fully theoretical evidence confirmed the aromaticity in X₃⁻ (X = B, Al, Ga) species via Boldyrev’s method [52]. Theoretical studies for the electronic structures, stabilities, and aromaticity in H₂B₂XH (X = N, P) molecules and the möbius aromatic planar metallaborocycles were also examined [53,54]. The above all-boron aromatic clusters could play important roles as potential novel ligands or building blocks in chemistry [55].

2.2. Triangular π-Aromatic [Si]₃⁺, [Ge]₃⁺, [Si₂C]⁺

In addition to group 13 elements, cyclic cations [A₃H₃]⁺ (A = C, Si, Ge, Sn, and Pb) were theoretically calculated with curiosities about whether these group of organometallics were going to show any evidence for aromaticity [56,57,58]. The experimental observation of three-membered cations made of the group 14 elements was predicted to be possible [59,60]. Soon, Sekiguchi and co-workers first reported the isolation and characterizations of new tris(tri-tert-butylsilyl)cyclotrigermenium tetraphenylborate [t-Bu₃SiGe]₃⁺BPh₄⁻, abbreviated as [Ge₃]⁺ [61]. This [Ge₃]⁺ system was a free germyl cation with 2π-electrons incorporated in the cyclic core. This [Ge₃]⁺ complex could be isolated as a type of yellowish solid in the inert atmosphere after the reaction of tetrakis(tri-tert-butylsilyl)cyclotrigermene with trityl tetraphenylborate (Scheme 3A). The single-crystal X-ray crystallographic analysis for [Ge₃]⁺ elucidated that Ge–Ge bond lengths range from 2.321(4) to 2.333(4) Å and the Ge–Ge–Ge bond angles vary from 59.8(1)° to 60.3(1)°. The three Ge atoms were localized in a nearly equilateral triangle. These crystallographic data resemble the cyclopropenium cation. The cyclotrigermenium cation [Ge₃]⁺ prefers a planar existence due to the electronic properties and steric hindrance caused by the bulky t-Bu₃Si groups.

In 2005, the same group introduced the first isolation of the persilaaromatic structure [(t-Bu₃Si₂)₂(t-Bu₂MeSi₂)]⁺BAr₄⁻, abbreviated as [Si₃]⁺ [62]. Due to the lack of proper precursors, the initial attempts for their syntheses were not successful. With modified methods in the presence of triphenylmethylium tetraarylborate, the desired cyclotrisilene was easily transformed into the targeted cyclotrisilenylium cation (Scheme 3B). According to the crystallographic data, the bond angles in the cyclotrisilenylium cation [Si₃]⁺ range from 59.76(10)° to 60.20(10)° and the Si–Si bond lengths vary from 2.211(3) to 2.221(3) Å. The internal [Si₃]⁺ core presented a nearly equilateral triangle. The three substituted silicon atoms are in the same plane within 0.39Å. These experimental characters of [Si₃]⁺ hinted that it was another triangular all-metal aromatics made of semimetallics.

It is worth mentioning that the synthetic strategies for the robust heteroaromatic cluster [Si₂C]⁺ were reported by Sekiguchi and co-workers based on the other group’s previous investigations on their analogues [63]. Using (t-Bu₃Si)₂SiLi₂ and 1-adamantanecarbonyl chloride in polar organic solvent formed one yellow solid 1,1,2-tris(tri-tert-butylsilyl)-3-(1-adamantyl)disilacyclopropene, [Si₂C], which was examined by X-ray crystallography. The hybridized [Si₂C] included one Si=C double bond which was equal to 1.745(2) Å and these data resemble the simple cyclopropane. The reaction of [Si₂C] with triphenylmethylium tetraarylborate gave 1,2-bis(tri-tert-butylsilyl)-3-(1-adamantyl) disilacyclopropenylium ion, [Si₂C]⁺, with great isolated yield (Scheme 4). The tetraarylborate salt of [Si₂C]⁺ was isolated as yellow crystals and was quite sensitive to air and water. The silicon–carbon hybridized [Si₂C]⁺ complex represented the first isolable derivative of cyclopropenylium ion, [C₃]⁺. The mixed heteroaromatic [Si₂C]⁺ showed an isosceles triangle in crystal structures; similarly, both the cyclopropenyl cation [C₃]⁺ and its silyl analogue [Si₃]⁺ were arranged in equilateral triangular cores.

The first three-membered TiSn₂ ring core was also isolated and fully characterized in 2014 by Kuwabara and coworkers [64]. Normally, the Ti–Sn single bonds in coordinated complexes range from 2.842 to 2.984 Å. However, they found that the Ti–Sn bond lengths in TiSn₂ ranged from 2.6867(16) to 2.7254(17) Å for Cp₂Ti[SnC₄Et₄]₂, which was obviously shorter. In addition, the regular bond lengths for Sn=Sn bonds range from 2.575(4) to 2.85126(19) Å; however, the distance between the two Sn atoms in TiSn₂ was equal to 3.0576(14) Å, which was obviously much longer. All these crystal results implied the possibility of the presence of electron delocalization and aromaticity in the three-membered TiSn₂ core. Indeed, their following complementary theoretical analysis confirmed that their triangular TiSn₂ was σ-aromatic. Further, another niobium-necked cluster [As₃Nb(As₃Sn₃)]³⁻ was also experimentally and theoretically examined, and confirmed that the Sn₃²⁻ core was aromatic [65].

2.3. Triangular σ-Aromatic [Au]₃⁺

Using powerful N-heterocyclic carbene (NHC) 1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene as the ligand, the first NHC-coordinated triangular tri-gold cluster [LAu]₃⁺ was synthesized by Sadighi’s group in 2012 [66]. The preparation of [LAu]₃⁺ was quite simple, with two steps (Scheme 5). First, the monomer [(LAu)₃CO₃]⁺OTf⁻ was obtained when [(LAu)₂OSiMe₃]⁺OTf⁻ reacted with [LAuOSiMe₃] under the pure CO₂ atmosphere. Secondly, in the presence of CO, [(LAu)₃CO₃]⁺OTf⁻ underwent reduction and gave the reduced cyclic trimer, [(LAu)₃]⁺OTf⁻. Crystallographic analysis for triangular [LAu]₃⁺ cation showed approximate D₃ symmetry with canted imidazolylidene rings (Figure 5a). The [Au]₃⁺ ring was nearly equilateral, the Au–Au–Au bond angles ranged from 59.603(9)° to 60.331(8)° and the length of the Au–Au bond ranged from 2.6438(5) to 2.6633(5) Å. The three directly connected carbon atoms on the NHC are basically coplanar with the triangular [Au]₃⁺ core.

Full valence delocalization along the cationic tri-gold core was discovered according to their complementary DFT analysis. The HOMO was composed mainly of the 6s orbitals for gold and the LUMO was composed of a degenerate pair. According to the frontier-orbital energy-level diagram, the calculated HOMO–LUMO gap was quite large (5.42 eV), which hinted at potential inert properties (Figure 5b). Theoretical parameters explained the great stabilities of the complex [LAu]₃⁺OTf⁻. The plots of the highest occupied (HOMO) and the lowest unoccupied Kohn–Sham orbitals (LUMOs) showed their corresponding electron densities (Figure 5c). Their orbital pictures were analogous to that of aromatic [H₃]⁺ and their frontier orbitals were echoed by those in [H₃]⁺. So, this σ-aromatic [Au₃]⁺ core with only 6s orbitals contributed to the necessary delocalized three-centre-two-electron mode was isolobal to σ-aromatic [H₃]⁺.

In 2014, Bertrand’s group improved the syntheses of aromatic [LAu]₃⁺ species with different NHC ligands and attempted applying these peculiar tri-gold clusters in the regular catalytic carbonylation of amines in the presence of CO under mild conditions and presented somehow ordinary reactivities [67]. Based on the known Sadighi’s synthetic strategies, Bertrand introduced another two straightforward synthetic routes for triangular aromatic gold complexes either ligated by three cyclic (alkyl)(amino)carbenes (CAAC) or by mixed ligands (two CAACs plus one phosphine). The first method was treatment of the (CAAC_c-Hex)AuCl with Ag₂O in the presence of NaBF₄ in a polar organic solvent. The intermediate complex (CAAC_c-Hex)₃Au₃O⁺ was obtained as an off-white solid. Then, this oxidized complex was reduced by CO, which yielded pale-yellow solid (CAAC_c-Hex)₃Au₃⁺, abbreviated as [Au]₃⁺ (Scheme 6). The melting point of [Au]₃⁺ was 240 °C (dec.), which was an indication of good stability. Indeed, the [Au]₃⁺ cluster was tested to be robust to oxygen or water, which was an important advantage for further catalytic utilizations.

The phosphine/NHC ligand exchange possibilities were also investigated using the phosphine-supported μ³-oxo cluster (PPh₃)₃Au₃O⁺ as starting material (Scheme 7). Adding excess equivalents of CAAC_diEt ligand to (PPh₃)₃Au₃O⁺ gave the μ³-oxo complex (CAAC_diEt)₃Au₃O⁺ with good isolated yield. The targeted trinuclear gold ring (CAAC_diEt)₃Au₃⁺ was obtained in nearly quantitative yield after a routine reduction process. When fewer equivalents of CAAC_diEt were used, the mixed (CAAC)(phosphine)-ligated complex (CAAC_diEt)₂PPh₃Au₃⁺ was formed with a moderate yield.

According to the crystallographic analyses, the Au–Au distances in CAAC_diEt)₃Au₃⁺ ranged from 2.6324(8) to 2.6706(8) Å; however, the value was 3.1669(3)–3.3149(3) Å for its oxidized analogue. Comparing the Au–Au bond lengths for these two crystal structures, apparent bond length differences were found. Normally, true covalent bonds are hinted in CAAC_diEt)₃Au₃⁺ due to their unbelievably short metal–metal bond distances. What is worth mentioning is that the CAACs ligand showed stronger σ-donor properties than triphenyl phosphine in the mixed-ligand coordinated structure (CAAC_diEt)₂PPh₃Au₃⁺. This improved synthetic route provided possibilities for preparation of a variety of polynuclear gold (1/3 valence) clusters through ligand exchanges. Similar to Sadighi’s structure, these mixed ligated clusters [Au]₃⁺ should also be the analogues of σ-aromatic [H₃]⁺.

2.4. Triangular σ-Aromatic [Zn₃]⁺, [Zn₂Cu]

In 2015, the triangular all-metal aromaticity was extended to Zinc-involved systems by Frenking and Fischer’s group. They synthesized and fully characterized the neutral triangular cluster [Zn₂CuCp*₃], abbreviated as [Zn₂Cu] and the cationic cluster [Zn₃Cp*₃][BAr₄^F] (Cp* = η-C₅Me₅), abbreviated as [Zn₃]⁺ (Figure 6) [68]. The synthetic procedures were convenient by simply adding [ZnCp*]⁺ and [CuCp*] to [Cp*Zn-ZnCp*] (Carmona’s compound) and stirring. The Zn–Zn bond in [Cp*Zn-ZnCp*] remained after the transformation to the final trimers. The authors claimed that using fluoroaromatic solvents was crucial for their efficient transformation. In addition, Fischer further reported the complementary experimental and theoretical investigations for the pseudo two-electron Cu/Zn clusters [69].

They crystalized both the [Zn₃]⁺ and [Zn₂Cu] clusters even sensitive properties were observed for these two ligated metallic trimers. Tri-zinc complex [Zn₃]⁺ crystallized in the ordinary monoclinic space group P2₁/c with two different formula units arranged together (Figure 6A). The neutral compound [Zn₂Cu] crystallized in the typical triclinic space group P₁⁻ (Figure 6B). The bond angles for [Zn₃]⁺ were very similar as 59.22(1)°, 59.71(2)° and 61.06(2)º, and for [Zn₂Cu], the angles were measured as 59.33(2)°, 60.34(3)° and 60.34(3)°. So, the metal atoms in both three-membered core structures presented almost equilateral triangular arrangements. The average distance of Zn–Zn bond in [Zn₃]⁺ was 2.430(1) Å. Similarly, the two Zn–Cu bonds in isosceles [Zn₂Cu] were measured to be the same and was equal to 2.381(1) Å. The only Zn1–Zn2 bond in [Zn₂Cu] was slightly shorter than 2.357(1) Å. All metal–metal bonds in these two complexes are slightly longer compared with their zinc-contained starting material decamethyldizincocene [70].

In order to better understand the bonding interactions of these two triangular clusters, the cationic [Zn₃]⁺ and its neutral homologue [Zn₂Cu] were calculated using the meta-GGA functional M06L method [71,72]. The nature of the bonding was analyzed by the energy decomposition analysis (EDA) method. They found that the σ donation of [Zn₂] to [Zn]⁺ into its empty 4s valence orbital was the main constitution for the overall orbital interactions (Figure 6C(I)). In addition, there are two types of leading orbital interactions for the neutral complex [Zn₂Cu] including the σ-donation from [Zn₂] to [Cu] and the π back-donation from [Cu] to [Zn₂] (Figure 6C(II,III)). Both the triangular cores [Zn₃]⁺ and [Zn₂Cu] showed valence electrons (cve) delocalization in one cyclic plane and presented apparent implications for σ-aromaticity which was further confirmed by quantum chemical calculations. Similar to the triatomic hydrogen ion [H₃]⁺, both the [Zn₃]⁺ and [Zn₂Cu] clusters revealed a high degree of σ-aromaticity.

2.5. Triangular Homoleptic [Hg₃]⁴⁺

The homoleptic triangular cluster [Hg₃(μ-dmpm)₄][O₃SCF₃]₄ [dmpm = bis(dimethylphosphino)methane], abbreviated as [Hg₃]⁴⁺ [73], was synthesized and crystalized by Peringer’s group many years after their other brief report on a similar cluster, [Hg₃(μ-dppm)₃][O₃SCF₃]₄ [dppm = bis(diphenylphosphino)methane] [74]. Each of the mercury atom in the cluster [Hg₃]⁴⁺ represented the 4/3 oxidation state. It was simply synthesized by mixing [Hg₂]²⁺ with excess dmpm (Equation (2)) or by the reduction of [Hg(Me₂SO)₆][O₃SCF₃]₂ in the presence of dmpm and mercury (Equation (3)).

2[Hg₂][O₃SCF₃]₂ + 4 dmpm →[Hg₃(μ-dmpm)₄][O₃SCF₃]₄ + Hg

(2)

2[Hg(Me₂SO)₆][O₃SCF₃]₂ + 4 dmpm + Hg → [Hg₃(μ-dmpm)₄][O₃SCF₃]₄ + 12 Me₂SO

(3)

According to the single-crystal X-ray analysis for complex [Hg₃]⁴⁺ (Figure 7), this molecule was composed of four CF₃SO₃⁻ anions for each [Hg₃]⁴⁺ cation, so the oxidation state for each Hg was distributed as +4/3 correspondingly. The cationic [Hg₃]⁴⁺ core was presented as a triangular mercury ring with bond lengths measured as Hg–Hg 276.68(14), 295.53(14) and 280.99(14) pm. There are four dmpm ligands involved for three mercury atoms, and the mercury edges are linked to either one or two phosphorous bridges. The three Hg atoms are nearly in a plane. Because of the intramolecular exchange between the two types of dmpm ligands, the complex showed fluxional ³¹P NMR signals at 81 MHz at room temperature. According to the experimental parameters for homoleptic [Hg₃]⁴⁺ cation, it was assumed to be another close analogue of the [Au₃]⁺ and the [Zn₃]⁺ systems, which are isolobal to σ-aromatic [H₃]⁺.

2.6. Triangular δ-Aromatic [Pd₃]⁺, [Pt₃]⁺, [Pd₂Pt]⁺, [PdPt₂]⁺

In 2014, Malacria’s group reported the mixed-ligand stabilized 44-cve tri-palladium cluster [Pd₃(μ₂-SPh)₃(PPh₃)]⁺, abbreviated as [Pd₃]⁺, as the first isolated d-block analogue of the π-aromatic cyclopropenyl cation [C₃H₃]⁺ [75]. These [Pd₃]⁺ clusters were firstly obtained by using S-aryl isothioureas as the sulfur source and have further been confirmed to be δ-aromatic by theoretical calculations (Scheme 8A). Then, a much-improved method was presented to access the identical tri-palladium cluster [Pd₃]⁺ and its homoaromatics tri-platinum cluster [Pt₃]⁺ and mixed-metallic heteroaromatics [Pd₂Pt]⁺ and [PdPt₂]⁺ (Scheme 8B) [76]. This new synthetic route was more efficient with almost quantitative conversions without preparing the S-aryl isothioureas. Using this updated method, aliphatic S-bridges and aliphatic phosphorous ligands are compatible and can be installed with high yields. The anionic part for the targeted [Pd₃]⁺ cluster could be changed simply by hiring different Ag^I salts. By replacing disulfide with diselenides, the Se-bridged [Pd₃]⁺ complex could also be obtained with much lower yields due to its intrinsic lower stabilities. The triangular triplatinum cluster [Pt₃]⁺ could also be prepared using Pt⁰ precursor Pt(dba)₃ instead of Pt^II species in order to achieve more efficient and more selective transformations [77]. Most importantly, when Pd(0) and Pt(0) precursors were mixed in 2:1 or 1:2 ratios, the peculiar heterobimetallic clusters [Pd₂Pt]⁺ and [PdPt₂]⁺ were formed, respectively (Figure 8). The isolated [Pd₂Pt]⁺ and [PdPt₂]⁺ clusters were considered the first confirmed triangular all-metal heteroaromatics among all the published Pd/Pt mixed complexes [78]. Both the heteroaromatic structures [Pd₂Pt]⁺ and [PdPt₂]⁺ were fully characterized including conclusive High Resolution Mass Spectroscopy (HRMS) and single-crystal X-ray diffraction (XRD).

For better illustrations and comparisons of the four analogue clusters [Pd₃]⁺, [Pt₃]⁺, [Pd₂Pt]⁺ and [PdPt₂]⁺ in the X-ray analyses, identical substituent groups were chosen with anion = SbF₆, R = 4-FC₆H₄ and R’ = 4-ClC₆H₄, as presented in Figure 9A. After optimization, these four crystalline [Pd₃]⁺, [Pt₃]⁺, [Pd₂Pt]⁺ and [PdPt₂]⁺ remained triangular, quasi-symmetric structures without any symmetry constraints. Their triangular all-metal core was found to be nearly D_3h-symmetric and their whole structure was almost C₃ symmetry for all these four complexes. The homometallic [Pd₃]⁺ and [Pt₃]⁺ triangles were always found to be perfectly equilateral with identical bond angles (60.0°) and exactly the same metal–metal bond lengths no matter the various substituted groups and anions. The Pd–Pd bond distance in [Pd₃]⁺ core equalled 2.8731(8) Å and the Pt–Pt length in [Pt₃]⁺ core was 2.8690(4) Å. These data were generally below the sum of Van der Waals radii (3.26 Å).

For the heteroaromatic complexes [Pd₂Pt]⁺ and [PdPt₂]⁺, 35% Pd and 65% Pt in [PdPt₂]⁺ and 66% Pd and 34% Pt in [Pd₂Pt]⁺ were presented in proportions. The bond angles in these heteroaromatic analogues deviated from 59.5(5) to 60.5(6)°. The averaged bond distance is 2.86(3) Å in [PdPt₂]⁺ and 2.87(4) Å in [Pd₂Pt]⁺. These parameters were very close to their homonuclear complexes [Pd₃]⁺ and [Pt₃]⁺. Because of the lanthanide contraction, the atomic radii for Pt and Pd are nearly equalized. So, the metal–metal distances are alike for the heterobimetallic clusters [Pd₂Pt]⁺ and [PdPt₂]⁺. However, the situation would be much different when two main-group elements were combined.

The most characteristic molecular orbitals for complexes [Pd₃]⁺, [Pt₃]⁺, [Pd₂Pt]⁺ and [PdPt₂]⁺ were analyzed and compared (Figure 9B). Interestingly, identical bonding HOMOs were found for these four analogues complexes. Not surprisingly, the two bimetallic complexes [Pd₂Pt]⁺ and [PdPt₂]⁺ presented reasonable sigmoidal symmetry properties. There is clearly a three-centre-two-electron bond delocalized in their HOMOs. In addition, not only the HOMO but also HOMO-6 and HOMO-36 showed many similarities with the other early transition metal complexes in which d-orbital and δ-aromatic properties were confirmed through molecular orbitals calculations.

Indeed, the presence of d-orbital aromaticity with a typical delocalized three-centre-two-electron bond among the three-membered metallic ring cores of [Pd₃]⁺, [Pt₃]⁺, [Pd₂Pt]⁺ and [PdPt₂]⁺ was further confirmed by natural bond orbital (NBO) analyses [79] and adaptive natural density partitioning (AdNDP) results [80]. The chemical shifts for these four clusters were always negative (5 Å) in the NICS analyses, which was also a characteristic phenomenon for aromatic structures. The delocalization state of the cyclic electron was also proved by magnetic susceptibilities obtained by SQUID measurements [81,82]. Normally, the χT for a non-aromatic, quasi-symmetric, triangular cluster, [Pd(CO)PPh₃]₃ equals 0.04 emuKmol⁻¹ at 300 K. This value was 0.44 emuKmol⁻¹ for [Pd₃]⁺ and 0.48 emuKmol⁻¹ for [Pt₃]⁺, respectively, under identical conditions. However, the susceptibilities for bimetallic complexes [Pd₂Pt]⁺ and [PdPt₂]⁺ were somehow as low as 0.22 and 0.26 emuKmol⁻¹, respectively, which was reasonable, because heteroarenes normally show much lower diamagnetic susceptibilities than their analogues homoarenes, even for the molecules made of main group elements [83]. In any case, all four complexes showed much higher values than those of ordinary non-aromatic clusters. The above experimental and theoretical proofs are fully consistent with the presence of δ-aromaticity for this group of triangular all-metal aromatics.

2.7. Triangular Heteroaromatic [Pd₂Ru]⁺

More recently, with the development and increasing interest in the fields of all-metal triangular aromatics, A. I. Boldyrev and coauthors presented another ligand-stabilised heterobimetallic triangular compound [Pd₂Ru]⁺ (Figure 10) [84]. This complex was air-stable and definitely was the first all-metal aromatic structure composed of noble metals Pd and Ru. Through structural and theoretical analyses, the authors demonstrated both similarities and differences for [Pd₂Ru]⁺ compared to its parental [Pd₃]⁺. Indeed, AdNDP and magnetic criteria proved that the [Pd₂Ru]⁺ core presented obvious overlap between the d_z² AO of ruthenium and the d_xy or d_x²—y₂ AOs of the other two palladium atoms. So, the comprehensive characterisations proved that there is σ-aromaticity among the metal triangular core [Pd₂Ru]⁺.

However, this structure was actually a byproduct of the reaction to obtain the expected metal triangle [PdRu₂]⁺ according to the ratio for the two metallic starting reagents. Notably, the two Pd atoms were coordinated with triaryl phosphines, the unique Ru atom was coordinated with p-cymene, and the employment of mixed ligands was quite rare compared to the other homo or hetero aromatics. By now, the application for this brilliant complex has not yet implemented, perhaps due to its comparatively low yields. In addition, Boldyrev further summarised the in silico advances in aromaticity and antiaromaticity for bridging transition-metal clusters by bonding and magnetic analyses [10].

2.8. Triangular σ-Aromatic [Th₃]²⁻

N. Kaltsoyannis and S. T. Liddle synthesized the first σ-aromatic radioactive metal composed tri-thorium cluster {Th₃(C₈H₈)₃Cl₆}²⁻, abbreviated as [Th₃]²⁻ (Figure 11) [85,86]. [Th₃]²⁻ was obtained through the interaction of [K₂{C₄(SiMe₃)₄}] with [Th(η⁸-C₈H₈)(Cl)₂(THF)₂]. Single-crystal X-ray diffraction showed that [Th₃]²⁻ was equilateral, and the average Th-Th bonds distances (3.99 Å) for the triangle was much lower than the sum of Van der Waals radii for thorium (4.74 Å). Indeed, various DFT and NICS analyses showed that the [Th₃]²⁻ core exhibited a valence-delocalized three-centre-two-electron σ-aromatic character. A short time later, Cuyacot and Foroutan-Nejad made more theoretical comparisons for [Th₃]²⁻, and questioned the employment of the NICS method and opposed the opinion that the Th–Th bond in the [Th₃]²⁻ was aromatic. I believe that further studies of this peculiar triangular [Th₃]²⁻ and the developments for the other predicted triangular metallic aromatics will give a more unified or developed view of triangular all-metal aromaticity.

3. Theoretical Predictions of Aromaticity for Triangular All-Metal Clusters

Aside from the above-discussed isolated all-metal aromatics as robust trinuclear crystals or solids, plenty of proposed triangular all-metal structures composed of various transition metals, semimetallics and other metals have been investigated by theoretical calculations in the past decade to predict their potentially possessed all-metal σ, π or δ-aromaticity [87,88,89,90,91,92]. Some typical and representative examples are discussed in the following for their peculiar structures, predicted triangular all-metal aromaticities and potential powerful applications.

3.1. d-Orbital Aromaticity in Triangular [M₃O₉]⁻, [M₃O₉]²⁻(M = W, Mo)

In 2005, Wang and co-workers reported the experimental and theoretical proofs of d-orbital aromaticity for early transition metal oxide clusters, [M₃O₉]⁻ and [M₃O₉]²⁻ (M = W, Mo) [28]. Both theoretical calculations and photoelectron spectroscopy showed that [W₃O₉] and [Mo₃O₉] are aromatic anions and presented D_3h symmetric property with a low-lying unoccupied molecular orbital occupied by one or two electrons. The single-occupied molecular orbital (SOMO) for [W₃O₉]⁻ and the highest occupied molecular orbital (HOMO) for [W₃O₉]²⁻ indicated M–M σ-bonding interactions which were inconsistent with their d-orbital aromaticity (Figure 12). The D_3h symmetric [M₃O₉]²⁻ dianions were given by adding another electron to this orbital and the three-centre-two-electron (3c–2e) bond was presented.

According to photoelectron spectroscopy (PES) analysis, the calculated resonance energy (about 1.05 eV) for the [M₃O₉]²⁻ cluster was comparable to the value estimated for highly aromatic benzene. In addition, the NICS values [93] among the ring centre of [W₃O₉]²⁻ and [Mo₃O₉] ²⁻ were obtained as −21.5 and −20.5, respectively. So, the [M₃O₉]⁻ and [M₃O₉]²⁻ clusters described by Wang’s group are peculiar with equal M−M bond lengths, large resonance energies and large negative NICS values, which all suggest that they are highly aromatic species involving fully delocalized metal–metal bonds. This new class of d-orbital aromatic species exhibited interesting chemical, electrochemical, and catalytic properties even without isolation and full characterization.

3.2. δ-Aromaticity in Triangular [Ta₃O₃]⁻

Many groups have discovered δ-aromaticity in metallic systems; the presence of δ bonds between identical or different transition-metal atoms indicated that delocalized cyclic δ bonds may exist in cyclic transition-metal structures [94]. However, different from valent s or p orbitals, d orbitals are more sterically restricted. As we know, d-orbital aromaticity needs substantial d–d bonding interactions, so the possibility for the transition metal to participate in forming chemical bonding mainly depends on the type of the transition metals and their spatial coordination circumstances [95,96,97]. In 2007, Wang and Boldyrev confirmed experimental and theoretical evidence for the presence of δ-aromaticity in the planar D_3h symmetric triangular cluster [Ta₃O₃]⁻ which possessed three delocalized electrons among Ta–Ta bonding [98]. The authors found the presence of not only π-aromaticity but also δ-aromaticity contributed by the d-bonding interactions in this proposed [Ta₃O₃]⁻ cluster according to the photoelectron spectroscopy (PES) [99] analysis under 193 nm (Figure 13a, 6.424 eV) and 157 nm (Figure 13b, 7.866 eV), respectively.

In addition, compared with the common ab initio calculations, both the optimized structures and the most representative electronic transition states for [Ta₃O₃]⁻ indicated that this cluster was arranged in a planar D_3h triangle (Figure 13c). Based on the chemical-bonding analysis, the five most representative valence molecular orbitals participated in the Ta–Ta bonding interactions. Concerning the contributions of the 5d atomic orbitals, the close bonding interactions between the δ and π orbitals of Ta atoms were apparent (Figure 13d). So, the most representative five valence molecular orbitals demonstrated the delocalization of the δ bonds and proved that the [Ta₃O₃]⁻ cluster is another example with δ-aromaticity. The ligated [Ta₃O₃]⁻ cluster has not yet been isolated experimentally.

3.3. d-Orbital Aromaticity in Triangular [Tc₃X₉]²⁻

Decades ago, scientists started the investigations for trinuclear complexes Re₃X₉ (X = Cl, Br, I) and found the electron delocalization character for these potential three-membered metallic cores [30,32,33]. However, the technetium analogues (Figure 14) gained less attention until Weck and coworkers reported their study for the technetium halide complexes [Tc₃(μ-X)₃X₆]^0/1−/2− (X = F, Cl, Br, I), which were further confirmed as isomorphous with their rhenium congeners by density functional theory calculations [88]. Even without isolation, the theoretical calculation results via adaptive natural density partitioning demonstrated that the [Tc₃X₉]²⁻ complex could exhibit aromaticity contributed by the d-orbital-based π electron delocalization over the tri-technetium centre.

3.4. Aromaticity/Antiaromaticity of Triangular [Sc₃]⁻

Rashidi-Ranjbar, Ruud, Foroutan-Nejad and coworkers reported new insights on the determination of aromaticity in triangular all-metal clusters [Sc₃]⁻ by giving computational arguments with the help of the ring-current model over local indices [100]. With their efforts, the evidence of the magnetic aromaticity was confirmed by nucleus-independent chemical shifts. Two typical methods for estimating magnetically induced ring currents were employed; the first approach was based on the quantum theory of atoms in molecules (QTAIM) and the second approach was explicit calculation using the magnetically induced current densities. Using the QTAIM-based magnetizabilities principle, they explained the presence of the two-zone aromaticity/antiaromaticity of several 3d all-metallic triangular clusters including Sc₃⁻, Cu₃⁺ and Cu₄²⁻ (Figure 15). According to the classical electromagnetic theory, their theoretical investigations indicated a comprehensive explanation for the anomalous magnetic shielding in different transition metal cluster. The authors also appealed that the nature of magnetic aromaticity/antiaromaticity for all types of triangular transitional all-metal complexes should be determined more comprehensively based on more reliable global indices.

3.5. Aromaticity of Triangular [Al₃]⁻

With strong interest, a multiconfigurational high-level electronic structure calculation method was employed to investigate the aromaticity property for the triangular all-metal cluster [Al₃]⁻ by Ugalde’s group (Figure 16a). Explicit theoretical calculation values indicated that the proposed three-membered-ring-like cluster anion Al₃⁻ presented the three most representative close low-lying electronic states with clearly different spins, and all these states presented strong multiconfigurational properties. Ugalde and coworkers further studied the aromaticity character of the all-metallic cluster [Al₃]⁻ using the total electron delocalization method. The related evaluations of the normalized multicentre electron delocalization indices for each state were calculated according to the multiconfigurational wave functions [101]. Combining the multifaceted results, the authors concluded that the lowest-lying singlet and triplet states of Al₃⁻ were proved to be highly aromatic. In contrast, the closest lowest-lying state (the quintet state) presented much fewer aromatic properties (Figure 16b). Later, theoretical investigations for the d-p hybridized aromaticity, neutral salts and photoelectron spectroscopy of the proposed LaX₂⁻ (X = Al, Ga, In) clusters were calculated [102]. Near-degenerate molecular-orbital-coupling-induced dual aromaticity in stable open-shell metal clusters was also observed [58]. Furthermore, the possibilities and behaviors of the aluminium trimer during the combination with different superatom clusters were illustrated [103].

3.6. Aromatic in Triangular [M₃]⁴⁺ (M = Ni, Pd, Pt)

Xiao and Li reported the structural, electronic and bonding properties of a system of triangular trinuclear clusters [M₃X₃]⁺ (M = Ni, Pd, Pt; X = F, Cl, Br, I) and explored the electronic and steric effects on the adsorptions of H₂ and other small molecules (Figure 17A) [104]. The authors presented the chemical bonding models of M₃ in detail and proposed that the oxidation state of each metal element in the M₃ cluster was +4/3. In fact, combined with the experimental successes on the isolations of [M₃]⁺ (M = Pd, Pt) crystals by Malacria’s group, all-metallic σ-aromaticity was also confirmed in these triangular all-metal clusters by theoretical calculations. It was suggested that by changing the energies and compositions of M–M and M–L chemical bonding orbitals, the stability and the catalytic activity of these complexes could be adjusted. The orbital interactions and relativistic effects could be responsible for the active H₂ dissociative adsorption on these complexes. This finding offered another example of adjusting catalytic properties by tuning different chemical bonding in the all-metal clusters. Actually, similar conclusions have been confirmed experimentally by the highly efficient triangular all-metal aromatic cluster [Pd₃]⁺-catalyzed internal alkyne reduction reactions. Later, the effects of ligands on bonding stability and σ-aromaticity for Pt nanoclusters were also examined [105]. Modulation of the backside-ligand effect of main-group aromatic ligands (benzene or naphthalene) binding to Pd was also investigated [106].

According to the earlier report of Fenske, treating transition metallic complex NiBr₂(PEt₃)₂ with Li₂NPh in THF/n-Heptane gave nitrogen-bridged triangular tri-nickel cationic cluster [Ni₃(μ₂-NHPh)₃(PEt₃)₂]Br, abbreviated as [Ni₃]⁺ [107]. This [Ni₃]⁺ cluster was investigated using single-crystal X-ray diffraction (Figure 17B). The crystallographic data proved that [Ni₃]⁺ was C₃ symmetric and composed of a Ni₃-ring with Ni–Ni bond distances ranging from 247.9(1) to 250.6(1) pm, which were the shortest as known and connected with three μ₂-bridging phenylamido ligands. Besides that, similar to their Pd/Pt analogues, each Ni atom was coordinated with one PEt₃ ligand. All Ni atoms lay in a plane, resulting in a trigonal-planar structure. Even the authors claimed that this [Ni₃]⁺ cluster was closer to the non-aromatic palladium compound discussed by Lee and Trogler [108]. However, according to the comprehensive investigations for their Pd/Pt analogues, all geometric parameters for [Ni₃]⁺ cluster hinted at the possibility of aromaticity in the triangular all-metal core. Taken together with the theoretical calculations, the triangular tri-nickel cationic cluster [Ni₃]⁺ may be another close analogue to the aromatic systems [Pd₃]⁺ and [Pt₃]⁺, which are isolobal to π-aromatic [C₃H₃]⁺, whereas further explicit theoretical calculations would be required to give a definitive conclusion on its proposed all-metal π-aromaticity.

3.7. Triangular Aromatic Heterobimetallic [Hf₃]⁺

Boldyrev presented that the lowest singlet D_3h¹A₁ structure of [Hf₃]⁺ is another example of a triple (σ-, π-, and δ-) aromatic system (Figure 18) [109]. An extensive search for the structure of Hf₃ using the B3LYP/LANL2DZ theory revealed that the lowest triplet and singlet states are D_3h³A₂‘ and D_3h¹A₁‘, respectively. Notably, the triplet state for [Hf₃]⁺ was the lowest one. However, the authors discovered that these two states are degenerate at the CASSCF/Stuttgart+2f1g level. These results indicated that the singlet state could be the global minimum structure at the higher level of theory. They concluded that the triplet D_3h³A₂‘ structure was doubly (σ- and π-) aromatic and the singlet D_3h¹A₁‘ structure was the first reported triple (σ-, π-, and δ-) aromatic system.

3.8. Triangular Aromatic Os₃N₃^+/−

Jin’s group investigated the stability, electronic character, and aromaticity of osmium-nitride clusters Os₃N₃^+/− using DFT (B3LYP methods) (Figure 19) [110]. The theoretical calculations revealed that five types of interactions existed in the hexagonal Os₃N₃⁺ (D_3h, ⁷A₁′) cation and the regular planar Os₃N₃⁻ (D_3h, ⁵A₁′) anion. Through detailed molecular orbitals (MOs) analysis, the authors concluded that the cationic Os₃N₃⁺ (D_3h, ⁷A₁′) and anionic Os₃N₃⁻ (D_3h, ⁵A₁′) possessed a triple-aromatic character (σ-, π- and δ-aromaticity), simultaneously. It was convinced that the solid structural stability was contributed by the five types of powerful d-orbital bonding interactions and triple aromaticity. Based on the aromaticity of the bare osmium trimers, the authors further calculated their binding interactions to group IA/IIA of all-metal series [111].

3.9. Aromaticity in Triangular Ir₃N₃^+/−

In 2017, Ding and coworkers fully investigated the chemical bonding, aromaticity and stabilities of the ground state of iridium low-nitride Ir₃N₃^2+/0/2− clusters (Figure 20), their metal complexes Ir₃N₃M^−/0 (M = Li, Na, K and Be, Mg, Ca) and bimetallic complexes Ir₃N₃M₂ (M = Li, Na, and K) using DFT calculations [112]. Through detailed molecular orbitals (MOs) analysis, they reveal that the neutral Ir₃N₃ (D_3h, ³A₂’’) cluster possesses σ-, π-, and partial δ- aromaticity, and the Ir₃N₃²⁻ (D_3h, ¹A₁’) dianion possesses both σ- and π- aromaticity. They further proved that the pyramidal Ir₃N₃M^−/0 and bipyramidal Ir₃N₃M₂ complexes also possess aromatic characters and preserved the Ir₃N₃²⁻ (D _3h, ¹A₁‘) motif.

3.10. Aromaticity in Triangular PrB₂⁻

Due to their highly contracted 4f atomic orbitals, lanthanide metals were rarely found in multiple types of aromatic systems. The first example was the lanthanide-boron PrB₂⁻, which was confirmed as a double-aromatic triatomic molecule based on joint investigations in photoelectron spectroscopy and quantum calculations (Figure 21) [113]. Global minimum structural studies revealed that PrB₂⁻ possessed a C_2v triangular structure. Their paramagnetic triplet ³B₂ electronic ground state could be viewed as featuring a B₂⁴⁻ and one trivalent Pr(III, f²). Chemical bonding analyses concluded that cyclo-PrB₂⁻ showed multiple Pr–B bonding characters and exhibited σ and π double aromaticity. This smallest 4f-metalla-aromatic system PrB₂⁻ also highlighted the formation of the rare B₂⁴⁻ tetra-anion, completing the isoelectronic C₂²⁻, N₂, and O₂²⁺ series.

According to the Geerlings findings, the metal–metal bonding interactions were much more complicated in the [Li₃]⁺ system [114]. Using the ring-current map computation, the assumed σ-aromaticity for the triangular [H₃]⁺ and [Li₃]⁺ was investigated. The [H₃]⁺ showed a characteristic diatropic ring current and could be considered as σ-aromatic on the magnetic criterion. However, even with a negative NICS value, [Li₃]⁺ was concluded to be non-aromatic according to the same criterion because it did not show any global current.

4. Catalytic Applications of Triangular All-Metal Aromatic Clusters

For trimetallic systems, scientists are becoming more and more interested in the investigations of their applications and synergistic effects. Clearly, triangular all-metal aromatic clusters contained characteristic delocalized cyclic electrons, metal–metal bonding interactions which determined their good stabilities, and aromaticity, making them become the metallic counterparts of the main-group aromatics. Different from biased opinions on all-metal aromatics for weak stability and rare utilizations, triangular all-metal aromatics were found to show great potential in applications like catalysis, coordination, and material science. In the following part, the most representative examples were summarized.

4.1. σ-Aromatic Tri-Gold Cation [Au₃]⁺ Catalyzed Amine Carbonylation

The first example referring to the applications of all-metal aromatics in catalytic reactions was reported by Bertrand and coworkers. They presented the syntheses and stabilities for the NHC stabilized aromatic tri-gold complex [Au₃]⁺ (Figure 22B), and further investigated their catalytic behaviors and plausible mechanisms in the carbonylation reaction [66,67]. Several types of transition metal catalysts presented efficiencies for the reaction of primary amines and carbon monoxide to prepare urea derivatives [115,116,117]. The tri-gold cluster was employed to catalyze this reaction and proved to be a green and selective catalyst (Figure 22A). The triangular all-metal aromatic [Au₃]⁺ catalyzed carbonylation reaction offered various corresponding substituted urea with mild yields. In order to achieve higher conversions, the catalyst loading reached 2.5 mol% in this reaction. Notably, the isolated yield for the corresponding substituted urea was much lower when the use of the catalyst was decreased. In addition, the substrate scope of this catalytic carbonylation reaction was screened and it was found that fewer spatially hindered amines could be transformed into urea derivatives with higher yields. The authors claimed a peculiar reaction mechanism for this aromatic tri-gold complex [Au₃]⁺ catalyzed carbonylation of amines. This proposed mechanism involved an oxidation of the valence state of the gold from Au₂⁰Au^I to Au₃^I, whereas this proposal demonstrated that these [Au₃]⁺ clusters lost their aromatic nature as an intermediate. Meanwhile, the more plausible catalytic mechanism is still underway. This reaction also had previously been screened by using the other types of gold-containing catalysts like mono-gold complexes and gold nanoparticles [118,119]. Generally, simple gold complexes presented good activity in this model reaction; for example, Corma et al. showed that small metallic gold clusters could display higher catalytic activities even than small-sized gold nanoparticles and many mono-gold catalysts [120,121,122,123]. Later, the behaviors of σ-aromatic cyclic [M₃]⁺ (M = Cu, Ag, Au) complexes were investigated when dimethyl imidazol-2-ylidene, isoxazole, pyridine, furan, carbon monoxide and noble gases were introduced for complexation [124]. In addition, the cyclic (R₂SnAu)₃ anion was proposed to be stable and showed in-plane σ-Möbius aromaticity [125].

Compared to the usual alkali and alkaline earth metals, analogues super alkalis showed lower ionization energy and possessed peculiar properties. According to the First Principles calculations, predicted superalkali clusters containing triangular all-metal aromatic Au₃ coordinated with either three imidazole (IMD) or three pyridine (Py) ligands are very possibly to be stable (Figure 22C) [126]. Similar to the properties of superalkali clusters, the calculated ionization energies (IE) for the organometallic complexes, Au₃(Py)₃ or Au₃(IMD)₃, are quite low. Additionally, the first-order hyperpolarizability calculations proved that some of these coordinated complexes presented non-linear optical properties, which were also similar to the properties of normal super alkali.

Based on the above work on gold clusters, our group synthesized the first triphenylphosphine stabilized triangular tri-gold complex [(PPh₃Au)₃]⁺BF₄⁻ through the facile reduction of μ³-oxo complex [μ₃-O(PPh₃Au)₃]⁺BF₄⁻ (Figure 23) [127]. The cyclic voltammetry of [(PPh₃Au)₃]⁺BF₄⁻ showed one single oxidation process. These tri-gold complexes were further employed in the catalytic hydroamination of phenylacetylenes with anilines. When the loading was decreased to 1.0 mol%, 95% of conversions were obtained. When the catalyst loading further decreased to 0.5 mol%, 83% of isolated yields were achieved in the preparative scale. This application presented great significance in the synthesis of imine-containing pharmaceutical intermediates. Moreover, further investigations of the catalytic mechanisms indicated that these tri-gold complexes played roles as pre-catalysts in the hydroamination.

4.2. Aromatic [Pd₃]⁺ Catalyzed Semi-Reduction of Internal Alkynes

Trinuclear all-metal aromatic [Pd₃]⁺ should be a good choice for catalytic applications in many types of reactions due to their essential palladium-containing nature and delocalized metal–metal bonding forms. In order to investigate the catalytic properties of all-metal aromaticity in [Pd₃]⁺, Malacria’s group investigated the catalytic properties of these tripalladium complexes in the classical semihydrogenation reaction of internal alkynes which were used as a model reaction for many types of mono-palladium complexes [128,129,130,131,132]. They found that when palladium clusters [Pd₃]⁺ were employed to reduce internal alkynes in a hydrogen transfer condition, quantitative (Z)-alkenes were obtained without forming any trace of alkane products [133]. The activity and selectivity for these [Pd₃]⁺ clusters are comparable with the best mono-palladium peers. At the end of the reaction, the HRMS still detected a trace of the complement catalyst.

In contrast to the excellent catalytic behavior of the [Pd₃]⁺ complex, the aromatic triangular tri-platinum complex [Pt₃]⁺ showed inert behavior under identical conditions. In order to explain their vastly different catalytic behaviors, the most representative frontier molecular orbitals of the analogues structures [Pd₃]⁺ and [Pt₃]⁺ were compared and analyzed (Figure 24A). The LUMO of the active [Pd₃]⁺ triangle is obviously lower than that of the heavier [Pt₃]⁺ cluster with the ΔE_gap reached to 14.9 kcalmol⁻¹. On the contrary, according to the literature research on palladium cluster-catalyzed reactions [134,135,136], monoatomic π-acidic Pd-containing clusters are much less electrophilic than their Pt-containing analogues.

The authors also carried out the deuterium-labelling experiment to figure out the possible catalytic pathways (Figure 24B). They found that deuterium atoms incorporated on the terminal carbon of 1-phenylpropyne with [Pd₃]⁺ complex as the catalyst and the intermediate phenylallene was formed during the catalytic reaction [137]. These findings indicated that the hydrogen transfer step might be rate-determining. In addition, the kinetic studies presented that the hydride-transfer process displayed obvious isotopic effects [138].

Later on, Malacria’s group further discovered that the semi-hydrogenation of internal alkynes could also be catalyzed by the aromatic [Pd₃]⁺ clusters under neat conditions and gave cis-alkenes with quantitative conversions (Figure 25A) [139]. The authors claimed that when 0.03 mol% amount of catalyst was loaded, a large gram scale of the catalytic semi-hydrogenation could be carried out successfully. This group further developed the zwitterionic [Pd₃]⁺ clusters and employed them in the semi-reduction reaction of internal alkynes as pre-catalysts or as a whole complex. These catalysts propelled the reaction which could be carried out under even neat conditions (Figure 25B) [140]. Generally, triangular all-metal aromatic [Pd₃]⁺ complexes presented excellent catalytic activities and outstanding selectivity in the semi-reduction of internal alkynes mainly due to their intrinsic aromaticity-induced stability. Aromatic tripalladium complexes could also catalyze inter/intramolecular cascade reactions of 1,6-enynes and carboxylic acids, giving excellent catalytic activities and selectivities [141].

Through a regular ion exchange process, anionic MOFs could become frameworks or templates with interesting properties. For example, crystalline MOFs could be further functionalized due to their modular character and novel properties could occur after elaborations. Ma and coworkers presented that by a simple ion exchange process, the trinuclear palladium complex [Pd₃]⁺ could be introduced to the anionic metal–organic framework (MOFs). In addition, the combined catalytic system was heterogeneous and showed excellent catalytic activity in the classical semi-reduction reaction of internal alkynes (Figure 26) [142]. The authors claimed that the thiol groups from the tri-palladium cluster and the delicate pore sizes of the MOF material significantly minimized the aggregations of palladium clusters during the recyclable catalytic process.

4.3. Aromatic [Pd₃]⁺ Catalyzed Cycloisomerization of 1,6-Enynes and Dienynes

Besides the semi-reduction reaction of internal alkynes, the triangular all-metal aromatic [Pd₃]⁺ complexes were further applied in the cycloisomerization reactions of substituted internal dienynes and terminal 1,6-enynes and also presented excellent catalytic activity and great selectivity under mild conditions (Figure 27A) [143]. The authors prepared various polycyclic frameworks by modification of substrates and assumed the mechanism of the reaction as a complex cascade process. T Interestingly, the authors later found that the substituted 1,6-enynes showed much different reactive properties when the reaction used carboxylic acid as an additive in the presence of an identical catalyst (Figure 27B) [144]. The catalytic activities of [Pd₃]⁺ complexes in this cyclisation reaction were comparable to that of much-reported mononuclear Pd(I) and Pd(II) clusters [145,146]. Different from the inertness in the semi-reduction reaction, the analogues tri-platinum catalyst [Pt₃]⁺ showed good reactivity to cyclize the substrate. However, the authors found that when the analogues tri-palladium complex [Pd₃]⁺ was used under identical conditions, successive cyclization/double bond shifts happened. So, the tri-platinum catalyst [Pt₃]⁺ showed a completely different reactive orientation compared with the analogues tri-palladium complex [Pd₃]⁺ mainly due to their different metallic properties.

4.4. Aromatic [Pd₃]⁺ Catalyzed Coupling Reactions

Besides the above applications in semi-reduction of internal alkynes and cyclisation, [Pd₃]⁺ also has been successfully applied to various types of carbon–carbon coupling reactions by several groups. Zhu and coworkers reported another type of tri-palladium complex [Pd₃Cl(PPh₂)₂(PPh₃)₃]⁺[SbF₆]⁻, abbreviated as Pd₃Cl, which was not C₃ symmetry and stabilized by one chlorine atom and two types of aryl phosphine ligands (Figure 28A) [147]. Compared with the other palladium-complex-catalyzed coupling reaction process [148,149], Pd₃ clusters presented precious C–X (X = halogens) selectivity and excellent efficiency. Pd₃Cl was proved to be a robust and air-stable cluster during characterization and application. The authors pointed out the σ-aromaticity for the Pd₃Cl cluster and also claimed that their outstanding stability was mainly contributed by their delocalized 3c–2e Pd–Pd–Pd bonds among the tri-palladium core. One of the most persuasive proofs for the existence of σ-aromaticity was the presence of delocalized 3c–2e σ-bonds.

Later, Pd₃Cl was further used in homogeneous catalysis with the Suzuki–Miyaura coupling as the model reaction under mild conditions. Among palladium complexes or nanoparticle-catalyzed aryl–aryl bond formation reactions, the Suzuki–Miyaura coupling is one of the most powerful pathways to construct versatile unsymmetrical. The authors investigated the catalytic mechanism of the reaction by monitoring the reaction of simple aryl bromides and arylboronic acids using HRMS and confirmed the presence of possible intermediates in the catalytic circle (Figure 28B). Interestingly, they found the mass for the Ar inserted intermediate through the simple substitution of the Cl atom between two palladium atoms and gave a new understanding of the palladium clusters involved mechanisms of C–C cross-coupling. Additionally, besides aryl bromides, the authors omitted the cases of the aryl chlorides and aryl iodides in this work and did not figure out the halogen selectivity in this case. Even though the mass of the Arinserted intermediate was detected, the proposed catalytic circle was not compliant with the widely recognized palladium-catalyzed coupling reactions.

Later, Schoenebeck and coworkers presented another robust palladium trimer using a facile synthetic method [150]. This palladium triangle possessed C₃ symmetric property and was stabilized by only one type of phosphine ligand. Different from the Pd₃Cl analogue, the Schoenebeck [Pd₃]⁺ presented unambiguous reactivities and privileged capabilities to C–I bond over C–Br and C–Cl bonds in aryl–aryl bond formations from polyhalogenated arenes and Grignard reagents (Figure 29). The catalytic procedure was classical, using the preformed palladium trimer allowed for quantitative preparation for the arylations and alkylations even for the coupling of steric substrates. Experimental data combined with computational investigations presented the feasibility and capability of the palladium trimer. Even the authors did not mention it, we believe this palladium core [Pd₃]⁺ presented unique C–I selectivity due to its aromaticity and stability. The cross-coupling catalyst Pd(OAc)₂/2PPh₃ could form a dinuclear Pd^I complex and further cyclic Pd₃ clusters during the catalytic process [151].

Most importantly, another aromatic tri-palladium complex [Pd₃]⁺ which was simultaneously stabilized by the sulfur and triaryl phosphine ligands showed outstanding photoelectric properties and was subsequently employed in the Suzuki–Miyaura coupling reaction by our group (Figure 30A) [152]. These C₃-symmetric triangular palladium complexes [Pd₃]⁺ showed great catalytic activity and exclusive C–X selectivity to aryl iodine thanks to the stability of the catalyst and the mild dissociation energy of the C–I bond. The substrate scope could be extended to thiophene, pyridine, pyrazine, and other common heterocyclic aromatic hydrocarbons. Research on the catalytic mechanism demonstrated that the reaction may involve the intermediate that one sulfur ligand was substituted by the iodine atom monitored by HRMS. A small amount of this tri-palladium complex [Pd₃]⁺ (0.06 mol%) showed powerful catalytic activity in the gram scale reaction and gave a 93% yield. For the Sonogashira and Heck coupling reactions, this triangular palladium complex [Pd₃]⁺ also presented unique selectivity for C–I bonds over the other halogen analogues (Figure 30B) [153]. HRMS monitoring for these two coupling reactions showed that the robust [Pd₃]⁺ remained as a whole at the end of the catalytic process, which proved that the [Pd₃]⁺ complex was extremely stable.

5. Cation–π Interactions of Triangular All-Metal Aromatics

5.1. Coordination of Aromatic [Pd₃]⁺ to Lewis Acids

Normally, delocalized bonds in main-group aromatics could play the roles as electron donors and form bonding interactions with Lewis-acid species. Indeed, the cation–π interactions are well documented for the main-group aromatics and they play a crucial role in chemistry and biology [154,155,156]. The sulfur atom and aryl phosphine ligand simultaneously stabilized tri-palladium complex [Pd₃]⁺ has been proven to be aromatic using many types of physical measurements and theoretical calculations. For further extending the applications of [Pd₃]⁺ in coordination, our group investigated the interactions of the aromatic [Pd₃]⁺ molecule with a variety of Lewis acids, such as Li(I), Ag(I), Au(I) and Cu(I). With a similar interest, by means of DFT methods, Tsipis predicted the tendency and possibilities of interactions between the ligand-stabilized aromatic triangular gold rings and Lewis acids years earlier [157]. As we know, Lewis acids are possible to form bonding interactions with main group arenes that possess a negative quadrupolar moment perpendicular to their plane (Q_zz) [158]. For comparison, the calculated Q_zz for cationic [Pd₃]⁺ is −52 Buckingham, and the Q_zz of benzene is −8.5 Buckingham, and for hexafluorobenzene is +9.5 Buckingham (Figure 31). By the same calculation model, the calculated Q_zz values for the [Pd₃]⁺ cations were seven times that of benzene. The largely negative values of aromatic [Pd₃]⁺ suggested that these cationic clusters hold a strong possibility of forming bonding interactions with Lewis acids. However, to the best of our knowledge on heterometallics [159,160,161,162], there were no experimentally isolated complexes to confirm this conjecture.

In 2017, we reported the first “pyramid” structure of [Pd₃Ag]²⁺ formed by the coordination of the triangular all-metal aromatic [Pd₃]⁺ and Lewis acid Ag⁺ (Figure 32) [163]. Cationic [Pd₃]⁺ can act as donor ligands thanks to their three-centre-two-electron metallic structure with delocalized metal–metal bonds. Driven by the cation–π interaction, the all-metal aromatic ligand [Pd₃]⁺ and Lewis acid Ag⁺ overcame the unavoidable electrostatic repulsions and formed these “peculiar” cation–cation-coordinated “pyramid” complexes. This result proved that parent all-metal aromatic complexes [Pd₃]⁺ have the properties of aromatic rings that are crafted with main group elements. Echoed with their main group element counterparts and overcoming the electrostatic repulsions, these noble-metallic rings formed stable bonding interactions with several Lewis acids, such as Ag⁺, Au⁺, Cu⁺, to deliver the corresponding tetranuclear bimetallic complexes with quantitative yields. These novel pyramid tetra-metal complexes [Pd₃M]²⁺ were fully characterized including HRMS and single-crystal X-ray diffractions. Through comprehensive modelling and experimental techniques for these bimetallic clusters, it was concluded that this bonding mode in [Pd₃Ag]²⁺ is an original coordination-like one rather than a four-centre-two-electron bond.

5.2. Coordination of π-Aromatic B₃⁺ to Transition Metals

The triangular B₃ unit is a fundamental bonding template among all boron compounds. Previously isolated B₃⁻ cluster possessed a D_3h structure and had double (σ and π) aromaticity. Using computational chemistry and high-resolution photoelectron imaging, several chemists investigated the bonding modes between a B₃ cluster and metallic Lewis acids. For example, Boldyrev and Wang studied the IrB₃⁻ cluster via high-resolution photoelectron imaging and theoretical calculations. They experimentally observed two isomers with different electron affinities and both structures had a B₃ ring coordinated with one Ir atom (Figure 33) [164]. The isomer with the higher EA consisted of one B₃ ring with a bridge-bonded Ir atom (C_s), and the isomer with the lower EA featured a tetrahedral structure (C_3v). Chemical bonding analyses showed that the neutral C_3v isomer could be viewed as an Ir–(η³-B₃⁺) complex, in which significant covalent bonding (Ir–B) and weak charge transfer from B₃ to Ir (ionic bonding) existed. The neutral tetrahedral structure was proved to be very stable. This study provided the experimental evidence and theoretical support for a π-aromatic B₃⁺ ring coordinating with a Lewis acid or transition metal.

In addition, Wang and colleagues investigated the high-resolution photoelectron imaging of MnB₃⁻ [165]. Theoretical calculation found that this is a C_2v planar structure in which the Mn is coordinated to only one side of the B₃ unit. The Mn atom showed weak interactions with the B₃ unit which still maintained the double aromaticity with small structural changes compared to the bare B₃ cluster. The present results thus pave the way for the use of suitable triangular all-metal rings as aromatic ligands for a variety of Lewis acids for cation–π interactions. It is anticipated that the introduction of aromatic all-metal cationic metallic rings as a new class of donor ligands is promising for vast innovation in coordination chemistry, materials, and catalysis [166,167,168,169].

6. Experimental and Theoretical Developments of Triangular All-Metal Aromatic Sandwiches

A sandwich structure usually means that one metal atom is bounded by two aromatic ligands like ferrocene, which was discovered at the beginning of metallocene chemistry. Now, it has been extended to include aromatic metallic ligands which could also be carbon-free as all-metal-containing layers [170]. In 2010, Fujita and coworkers reported the first three-dimensional all-metal aromatic complex in which metal ions were arrayed in a sandwich shape [171]. The planar polymetallic complexes could be coordinated and assembled as aromatic stacks. For the first time, a crystallographic X-ray of a trigonal prismatic cluster made of three cyclic triangular Au(I) layers was obtained (Figure 34a). By simply mixing the hexa-metallic Au₃-Au₃ complex with Ag(I) ion, the Ag-sandwiched complex Au₃-Ag-Au₃ could also be obtained. The composition of the initial polynuclear complexes and the shape of the yielded sandwiches determined the exact method of arrangement in the caged clusters (Figure 34b). In fact, based on triangular all-metal mono-layers, the preparation of all-metal aromatic sandwiches was an excellent breakthrough and opened the possibilities for their potential applications in material science as semiconductors.

Besides experimental findings, there are also theoretical proofs for the existence of aromaticity in peculiar all-metal sandwiches. For example, some metallic sandwich complexes made of arraying aromatic triangular layer Al₃R₃ (perfluorinated cyclotrialane) were calculated by Mercero and coworkers. For better understanding, their structural and electronic properties were also investigated using density functional theory (Figure 34c) [172]. As we know, the perfluorinated cyclotrialane ring possesses both strong σ and π aromaticities, and it has been confirmed to be a very stable ligand for metal–ligand coordination during the sandwich formation process. In addition, with thorough theoretical calculations, the authors claimed that when the perfluorocyclotrialane ligands were employed to complex the other single metals or metal dimers, the aromaticity of the all-metal triangular ligand remained.

In 2015, Sun, Zhai and coworkers reported the synthesis and isolation of another all-metal aromatic sandwich complex, [Sb₃Au₃Sb₃]³⁻, in which the structure and composition were confirmed by single-crystal X-ray diffraction (Figure 35a) [173]. Quantum chemical calculations demonstrated that there were obvious intramolecular electron transfers among the three all-metal layers. Their valence electrons were clearly rearranged from the cyclo-Sb₃ layers and Au₃ interlayers to the Au–Sb bonds. The whole triangular sandwich complex was combined via σ bonding. According to the reversed Hückel rule for aromaticity, the delocalization of the 3c–3e π bonds for each cyclo-Sb₃ layer proved the presence of aromaticity (Figure 35b,c). Theoretical studies on the electron structures and bonding forms of [Au₃Sb₆]³⁻ were also carried out [174]. This new type of all-metal, aromatic sandwich complex was prospected to hold great potential in applications as semiconducting materials.

A short time later, Li and coworkers reported one similar D_3h symmetric all-metal aromatic sandwich structure, [Bi₃Au₃Bi₃]³⁻; this species was investigated systematically via density functional theory (Figure 35d) [175]. Both molecular orbital analysis and nucleus-independent chemical shift data confirmed that similar to the cyclo-Sb₃ [Sb₃Au₃Sb₃]³⁻ complex, cyclo-Bi₃ in [Bi₃Au₃Bi₃]³⁻ possessed both σ and π aromaticities. However, the cyclo-Au₃ possessed σ and δ aromaticity and even weak π antiaromaticity. The authors analyzed the bonding nature of each type of bond in this sandwich complex and suggested that the Bi–Bi bond was a nonpolar σ covalent bond; the Au–Bi bond is a polar σ covalent bond; and the Au–Au bond was attributed to the typical aurophilic interaction. In addition, the intermolecular electron transfer phenomenon between cyclo-Au₃ and cyclo-Bi₃ was observed by charge decomposition analysis. Besides that, the robust stability of these complexes could be explained by their large E_gap and small Δ e^avg_R which also indicated their potential application as one new type of semiconductor material.

Additionally, the φ-aromaticity for prismatic {Bi₆}-based clusters was also demonstrated [176]. The aromaticity, electronic structures, and interactions with hydrogens for all-metal aromatic binuclear sandwich complexes were further illustrated [177]. The all-metal binuclear sandwich clusters Al₄Ti₂Al₄ were proved to be high-capacity hydrogen storage structures through multicentre bonds [178]. The four-fold π/σ aromatic sandwich-type Na₆B₇⁻ and Na₈B₇⁺ complexes featured charge-transfer complexes [179]. Trinuclear mixed-metal sandwich complexes could also construct axially chiral metal skeletons by selective dimerization [180]. The coordination modes and ligand influences for cyclooctatetraene ligated trinuclear Pd3 sandwich complexes were theoretically investigated [181]. Trinuclear Pd metal sheet sandwich complexes have peculiar electronic structures and hold great potential for further functionalization [182].

7. Conclusions

Triangular all-metal aromatics are a group of organometallic complexes that possess the smallest ring and conform to the Hückel (4n + 2) rule with all-metal σ, π or δ-aromaticity among transition metals, semimetallics and other metals. These metallic clusters proved to be either analogues of the σ-aromatic molecules [H₃]⁺ ion or analogues of the π-aromatic molecule’s [C₃H₃]⁺ ion. As summarized in the review, the syntheses, characterizations and theoretical calculations for isolated robust triangular all-metal aromatic clusters like [Ga₃]²⁻, [Al₃]²⁻, [B₃]²⁻ [Si]₃⁺, [Ge]₃⁺, [Au]₃⁺, [Zn₃]⁺, [Pd₃]⁺, [Pt₃]⁺, [Th₃]⁺, [Pd₂Ru]⁺, [Pd₂Pt]⁺ and [PdPt₂]⁺ have been demonstrated. Theoretical calculations have predicted the existence of multiple types of aromaticity in proposed structures like [M₃O₉]⁻ and [M₃O₉]²⁻ (M = W, Mo), [Ta₃O₃]⁻, [Tc₃X₉]²⁻, [Sc₃]⁻, [Al₃]⁻, [M₃]⁴⁺ (M = Ni, Pd, Pt), [Hf₃]⁺, Os₃N₃^+/−, Ir₃N₃^+/−, and PrB₂⁻. Experimental and theoretical developments of all-metal aromatics involved in sandwiches are well developed. The aromaticity of the all-metal triangular ligand in the sandwiches could also remain. The isolation of complex aromatic sandwiches will pave the way for the synthesis of entirely new classes of complexes and hold potential technological applications as semiconductors.

The first σ-aromatic cation [Au₃]⁺ showed clear activity in the amine carbonylation reaction and provided modest yields. Comparatively, trinuclear all-metal aromatic [Pd₃]⁺ proved to be a strong candidate for catalytic applications in several optimized catalytic systems like semi-reduction of internal alkynes to deliver cis-alkenes with almost complete selectivity on a gram scale with very low catalyst loadings. Powerful activity was also obtained in the cycloisomerization of terminal 1,6-enynes and internal dienynes under mild conditions and Suzuki–Miyaura, Sonogashira, and Heck C–C cross-coupling of aryl halides and arylboronic acids under mild conditions. By now, the all-metal aromatic cationic metal rings were applied in coordination chemistry successfully and used as a new class of donor ligands for a variety of Lewis acids through cation–π interactions. This summarised review will propagate the importance and developments of triangular and sandwich-typed all-metal aromatics and the prospect that this group of peculiar structures could participate in more types of in-depth applications in coordination chemistry, catalysis, and material science.

Above all, many types of triangular and sandwich-typed all-metal aromatics were synthetically isolated or theoretically predicted, and the types of metals in three-membered circles possessing aromaticity were still very limited compared to the integral metal species on the periodical table. By now, besides the three-membered circles, larger aromatic circles (four-, five-, six-membered) composed of pure metals were rarely discovered. In fact, most of the successful syntheses originated from unexpected discoveries rather than purposeful preparations and most of the designed syntheses routes cannot be achieved due to their exclusive metallic properties and subtle coordination relationships with ligands. Besides the enthusiasm for their peculiar structures of triangular and sandwich-typed all-metal aromatics, their applications in coordination chemistry, catalysis, and material science could be developed and carried forward in the future decades.