C,C- and C,N-Chelated Organocopper Compounds

Abstract

Copper-catalyzed and organocopper-involved reactions are of great significance in organic synthesis. To have a deep understanding of the reaction mechanisms, the structural characterizations of organocopper intermediates become indispensable. Meanwhile, the structure-function relationship of organocopper compounds could advance the rational design and development of new Cu-based reactions and organocopper reagents. Compared to the mono-carbonic ligand, the C,N- and C,C-bidentate ligands better stabilize unstable organocopper compounds. Bidentate ligands can chelate to the same copper atom via η²-mode, forming a mono-cupra-cyclic compounds with at least one acute C-Cu-C angle. When the bidentate ligands bind to two copper atoms via η¹-mode at each coordinating site, the bimetallic macrocyclic compounds will form nearly linear C-Cu-C angles. The anionic coordinating sites of the bidentate ligand can also bridge two metals via μ₂-mode, forming organocopper aggregates with Cu-Cu interactions and organocuprates with contact ion pair structures. The reaction chemistry of some selected organocopper compounds is highlighted, showing their unique structure–reactivity relationships.

1. Introduction

Copper-mediated or -catalyzed reactions have been upgrading the toolbox of organic synthesis with a variety of cheap, efficient transformations, which benefit scientific research in a broad range of areas, such as in polymer chemistry and biochemistry [1,2,3,4,5,6,7,8,9,10,11,12]. To have a deep understanding of the reaction mechanisms, organocopper species, as proposed intermediates, are sought to be isolated. On the other hand, organocopper compounds can be easily prepared in situ and utilized as organometallic synthons to construct various small organic molecules [13,14,15]. A well-recognized example is that “hard” nucleophiles, e.g., organolithium or Grignard reagents, react with α,β-unsaturated carbonyl compounds via 1,2-addition, whereas the “soft” organocuprates will end up with 1,4-addition products with the same substrates [16,17]. In a catalytic version, in the presence of copper salts, Grignard reagents can behave similarly to organocuprates [18]. It is known that reactivities are closely connected to structural configurations. Organocopper chemistry focuses on establishing the structure–reactivity relationship of well-defined organocopper compounds, rationalizing the dissimilar reaction patterns of different organometallic copper reagents.

Transmetalation provides the most straightforward and reliable method f preparing organocopper compounds from readily available organometallic reagents [19,20], though many other synthetic strategies are also known, such as copper-halide exchange reactions [21], direct cupration [22], and the oxidative addition of reactive Rieke Cu* [23]. It should be emphasized that the identity of forming organocopper compounds is very much dependent on the ratio of starting materials. For example, in the early 19th century, Gilman et al. found that the reaction of methyl lithium and 1.0 equiv. of copper(I) salts formed insoluble methylcopper polymers as yellow solids in diethyl ether [24]. When 0.5 equiv. of copper(I) salts was used in the above condition, a clear solution of dimethyl cuprate (Gilman reagent) was created. The initial isolation and characterizations of organocopper compounds were very slow since the reactive species were subjected to thermally homolytic, oxidative, and hydrolytic decompositions [25]. A general trend is that the stability increases from alkyl, alkenyl, and aryl to alkynyl copper compounds. For example, methylcopper [24] and phenylcopper [26] clusters display rapid and slow decompositions at room temperature, respectively, while (phenylethynyl)copper [27] looks stable under ambient conditions. This stability trend could also be reflected by the numbers of single-crystal structures in the Cambridge Structural Database (CSD 5.42, 2020 November), in which roughly 185 alkyl, 288 aryl, and 326 alkynyl copper compounds have been collected. Moreover, the aggregate state, which is favorable in charge-neutral organocopper compounds, makes these polymeric compounds poorly soluble in common organic solvents. To solve the above inherent problems of organocopper compounds, coordinating heteroatom-containing moieties, such as dimethyl amino groups, were rationally introduced into the ligand skeleton by van Koten and other researches to enhance stability and solubility [28,29]. The hybrid C,N-bidentate ligands can bind to copper atoms with an enhanced chelating effect. Following a different path, Xi’s group discovered that 1,4-dilithio 1,3-butadienes (C,C-bidentate ligand) can stabilize reactive organocopper species to a great extent via a cooperative effect [30,31]. In addition, this C,C-bidentate ligand can also be regarded as an analogue of 1,2-diketimine, which turns out to be a special Z-type non-innocent ligand, producing an unprecedented aromatic dicupro-annulenes [32].

Here, in this review, we are interested in summarizing organocopper compounds with at least one Cu-C σ-bond in which the organic fragments behave as actor ligands rather than spectator ligands (e.g., Cp, NHC). For mono-dentate carbonic ligands, the binding between copper and carbon atoms normally consists of the regular two-center two-electron (2c-2e) σ-bonds and three-center two-electron (3c-2e) bonds [33] when the carbon atom bridges two copper atoms. In cases where the bridging carbon atom is linked with a copper and a lithium (or magnesium) atom, the binding is more likely between the 2c-2e and 3c-2e bonds. For neutral nitrogen-containing groups, the binding of Cu-N is a traditional dative bond. When extending from one coordinating site to C,N- or C,C-chelating ligands, the binding in the corresponding organocopper compounds becomes more diverse. Based on the characteristic connectivity of the fragments in these structures, the C,N- and C,C-chelated organocopper compounds can be put into the following four categories:

(1)

Mononuclear organocopper complexes:

The two coordinating sites of the C,C- and C,N-bidentate ligand bind to the same Cu center together, forming a 5-membered chelate ring, as shown in (I) of Scheme 1. The two chelate rings are connected with a copper atom and will generate a spiro complex.

(2)

Organocuprates with contact ion pair structures:

The two coordinating sites can be attached with different metals, with one site bound to the Cu/M atoms and the other one bound to the main group metal atoms M only. The resulting product can be regarded as being organocuprates with contact ion pair structures (II).

(3)

Binuclear macrocyclic copper complexes:

Each coordinating site of the C,C- and C,N-bidentate ligand can also bind an individual Cu atom with 2c-2e bonds, forming a macrocyclic dinuclear complex. As a result, each Cu atom forms a linear geometry with two units of bidentate ligands that provide only one coordinating site to each Cu atom (III).

(4)

Polynuclear organocopper aggregates:

If the carbanion bridges two copper atoms, the structural feature will likely be polynuclear copper aggregates with Cu-Cu interactions (IV).

2. Mononuclear Organocopper Complexes

Although neutral N,N-chelated spiro copper complexes are relatively common, the anionic C,N- or C,C-cheated spiro counterparts are sporadic. As mentioned above, Xi’s group found that the C,C-bidentate ligand is an excellent platform to build up diversified coordination complexes across the periodic table [32,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60]. The chelating effect increases the thermal stability of these organometallic species, facilitating isolation and following characterization. Additionally, their metal–carbon bonds remain reactive toward electrophiles. As a result, a good balance between stability and reactivity is achieved.

Starting from the 1,4-dilithio 1,3-butadienes 1 (dilithio reagent for short) and 0.5 equiv. of copper(I) salts, Xi’s group isolated the first series of spiro organocopper(I) compounds 2–4 (Scheme 2) [59,60]. It seems that the terminal alkenyl carbon atom is favorable for LiBr coordination, probably because it is more electron-rich than the phenyl ones. The butadienyl spiro copper complex 2 forms one-dimensional polymers with repeating units linked by Li—Br. When one vinyl moiety is replaced by a phenyl ring, LiBr forms a Li₂Br fragment between the two vinyl carbon atoms of 3. No LiBr salt can be found in the biphenyl spiro copper complex 4 (Figure 1). The copper atoms are all coordinated with two dianionic ligands, forming a distorted tetrahedral geometry (dihedral angels between two five-membered chelate rings is about 63–84°). The averaged Cu(I)-C bond lengths of 2–4 are from 2.01 to 2.06 Å. Upon oxidation, either the dianionic ligand or the metal center can be oxidized depending on the electron-richness of the ligands. Compound 1 is transformed to a hexanuclear tetraenyl copper cluster 5 with a Cu₆ hexagon. In this process, the oxidative dimerization of the original butadienyl ligand occurs. It might undergo a spiro Cu(III) intermediate, which is too unstable to be isolated. When 3 is treated with oxidizing agents, a spiro Cu(III) complex 6 is afforded and isolated successfully. Compound 6 is thermally unstable and is difficult to handle at room temperature. With an extended conjugated backbone, 4 is oxidized to form a very stable spiro Cu(III) complex 7. In addition, 7 can be reduced by metal lithium to regenerate 4 quantitatively. Both 6 and 7 have copper atoms with a distorted square planar geometry. The averaged Cu(III)-C bonds of 6 and 7 (1.96–1.97 Å) are noticeably shorter than the Cu(I)-C bonds of 2–4. It is noteworthy that 6 and 7 represent a novel type of unprecedented organocopper(III) complex. By virtue of Cu L_2,3-edge X-ray absorption spectroscopy and experimentally calibrated electronic structure calculations, Lancaster and co-workers found that Cu(III) complex 7 has a LUMO (lowest unoccupied molecular orbital) with only ∼25% Cu 3d character, which suggests that the formal Cu(III) is not a physical d⁸ configuration [61].

As revealed by the crystal structure of 7, all four aryl rings are at cis-positions, which, in principle, provides a favorable arrangement for reductive elimination. However, no C-C bond formation occurred when the anionic compound 7 itself was refluxed overnight. However, when treated with electrophiles, e.g., aqueous HCl solution, iodine, and alkylation reagents, 7 afforded the symmetrical quaterphenyl derivatives 8–10 facilely in almost all quantitative yields via an electrophilic attack followed by a reduction elimination (Scheme 3). A stepwise experiment showed that 7 reacted with 1.0 equiv. of methylation agent, producing a quaterphenylcopper(I) species 11, which can couple with other electrophiles to form the di-substituted quaterphenyl compounds 10, 12, and 13 in high yields. The reductive elimination of high-valent copper compounds has often been proposed as the final step to release the product, but there has been nearly no concrete experimental evidence of that until this work.

In a similar vein, López-Ortiz and co-workers reported the reaction between dilithiated phosphazene 14 with 2.2 equiv. of CuBr followed by an aqueous workup to afford two diastereotopic isomers of mixed-valent Cu(I)/Cu(III) complexes 15 and 16 in 53% combined yield, in which the Cu(III) center was chelated with two C,C-bidentate aryl-alkyl ligands (Scheme 4) [62]. A byproduct with a four-membered ring 17 was isolated as well. Both 15 and 16 can be purified by column chromatography. Accordingly, they are stable in air and moisture. Cu(III) and Cu(I) adopt a distorted square planar and linear geometry, respectively (Figure 2). The averaged Cu(III)-C bond (1.98 Å) in 15 and 16 is similar to that in 6 and 7. The oxidation states of the Cu ions in 15 were further confirmed by extended X-ray absorption fine structure (EXAFS). Furthermore, all of the pieces of evidence, including the short Cu-Cu distance (2.74 Å) in the single-crystal structure, a critical bond point in the topological analysis of the X-ray experimental Fourier map as well as a diagnostic absorption band at 343 nm, strongly supported a weak metal–metal d⁸-d¹⁰ bonding between two Cu ions. Compound 15 can be used as a catalyst in catalytic cyclopropanation and aerobic oxidations. The characteristic signal of 15 has been observed and suggests that 15 remains unaffected during these transformations, probably due to its rigid structural configuration and extremely high stability.

The compound 2-phenylpyridine and its derivatives are among the most popular scaffolds for photoactive transition metal complexes. Partial fluorination of the aryl ring makes the C-H bond at the 2-position very acidic. Additionally, at the same time, it increases the stability of the resulting Cu-C bond. Steffen and co-workers reported that the C,N-chelated organocopper compound 18 was obtained from a direct metalation from Cu(OH)(NHC), in which 4A molecular sieves play an essential role in removing water (Scheme 5) [63]. In the presence of water, compound 18 will inevitably undergo a hydrolysis process, regenerating the C-H bond and forming an ionic species 20. The counteranion of hydroxide in 20 will replace a fluoride at the 4-position via a nucleophilic aromatic substitution in THF to 4-hydroxyl-2-arylpyridine and copper fluoride species. When MeCN is used as the solvent, hydroxide will attack MeCN, forming a copper acetamide species 21. Similarly, a spiro complex 19 can also be prepared from lithiation and subsequent transmetalation with CuBr·SMe₂ in the presence of diphosphine ligands. It has been shown in single crystal structures that 18 and 19 adopt distorted trigonal and tetrahedral coordination geometries, respectively. In solution, 18 is weakly emissive. In the solid state, the C,N-chelated organocopper compounds exhibit intense orange-red luminescence (λ_max = 610 (18), 607 (19) nm) at room temperature and have phosphorescence lifetimes of τ = 8.6 (18) and 9.5 (19) μs.

3. Organocuprates with Contact Ion Pair Structures

As mentioned in the Introduction, organocuprates have been extensively used in organic synthesis. Although numerous lithium organocuprates have been structurally characterized, the equally important magnesium organocuprates remain less unexplored, especially in terms of their well-defined structures. Only a few examples of aryl magnesium organocuprates have been previously reported [64]. As reported in the previous work of Xi and co-workers [35], magnesiacyclopentadienes 22, the first alkaline-earth metallocyclopentadienes, have a small C-Mg-C bite angle and a coplanar 5-membered chelate ring. The bridging character of anionic carbon atoms (Lewis base) and the coordinative unsaturation of the Mg atom (Lewis acid) make 22 available to bind other metal salts. Two equiv. of 22 reacted with 1.0 equiv. of copper halides or (phenylethynyl)copper to create magnesium organocuprates 23a–c and 24a with a Cu:Mg ratio of 1:2 (Scheme 6) [58]. In addition, inert silver halides as well as alkynyl silver compounds can apply to the above transformations, producing magnesium organoargenates 23d and 24b. The structures of forming organocuprates are dependent on the substituents at the 2,3-positions for some reason. When the substituents are changed from Ph to Me groups, 25, with a Cu:Mg ratio of 2:3, was obtained as the major product. Structural analysis and DFT calculation (AIM) revealed that as the Cu-C bonds form, one Mg-C bond becomes very weak (2.34 Å), while the other one remains unchanged (2.11 Å) in 23a. It seems that two Mg atoms provide a pre-organized position to anchor the halides. The C-Cu-C bond angle (157°) as well as the Cu-C bond lengths (1.95 Å) indicates a Cu-C-Mg 3c-2e bonding character (Figure 3). As supported by the molecular structure and DFT calculations, no bonding interaction was found between Cu-Br in 23a. On the whole, the binding of the anionic C (sp²) carbon atoms to the Cu atom and the bridging of the Mg atoms by the halides leads to a formal heterolytic cleavage of the M-X bond (M = Cu, Ag; X = halide, alkynyl) in a cooperative way along with the transfer of anionic X units from M centers to Mg centers. Accordingly, 23–25 can be considered as a resting-state intermediate of the transmetalation reaction between organomagnesium reagents and coinage-metal salts. The bridging ligand can be transformed quantitatively, e.g., 23c → 23a or 24a. The preliminary reactivity of these rigid magnesium organocuprates was provided by the reactions of 23c with various electrophiles, producing small organic molecules 26–28.

When a built-in coordinating site is added to the monoanionic phenyl skeleton at the ortho-position, it will generate a C,N-bidentate ligand, as shown in Scheme 7. With the (dimethylamino)methylphenyl (DMMP) lithium reagent 29 in hand, van Koten’s group developed two methods to synthesize the dicopper-dilithium organocuprate 30: (1) a direct transmetalation of 29 with 0.5 equiv. of CuBr and (2) an interaggregate exchange reaction between DMMP copper cluster 31 (vide infra) and 29 [65,66]. The short averaged Cu-C bonds (1.94 Å) as well as the weak averaged Li-C bonds (2.39 Å) in 30 (Figure 4) suggest that the 2c-2e bond of the Cu-C bonds has a higher character [67]. However, the split coupling of ¹³C-⁷Li (7 Hz) was still observed in solution, which indicates that the s electron density is present between the Li and C (bridge) nuclei [68]. Meanwhile, J (¹³C-⁷Li) did not change during variable-temperature NMR studies, allowing it to exclude the intraaggregate exchange processes. By monitoring the characteristic signals of methylene protons (AB pattern) and NMe₂ protons (two singlets) using dynamic ¹H and ¹³C NMR spectroscopies, van Koten and co-workers found the rotation of the 3c-2e bonded aryl groups around the C(1)-C(4) axis for the first time [69]. The C₄Cu₂Li₂ core fragment appears to be a persistent structural feature in neutral aryl lithium organocuprate (contact ion structure [64]).

4. Binuclear Macrocyclic Copper Complexes

It seems that trimethylmethylenephosphorane (Me₃P=CH₂) shows a better stabilization effect on the Cu-C bonds than its isoelectronic analogue Me₃Si-CH₂. Buchner and Zangrando found that the treatment of Me₃P=CH₂ with CuCl afforded a binuclear cuprate complex 32 in which negative charges are formally located at the Cu atoms while positive charges are located at the phosphine atoms (Scheme 8) [70,71]. In this process, the hybridization of CH₂ was changed from sp² to sp³. The core structure of P₂(CH₂)₄Cu₂ forms a coplanar 8-membered macrocycle. The averaged Cu-C bond (1.96 Å) and C-Cu-C bond angel (176°) are in the typical ranges for organocuprates.

Recently, there has an increasing interest in (CF₂)_n-containing materials in industry. However, the methods of incorporating difluoromethylene into molecules are lagging behind other perfluroalkyl counterparts. Vicic and co-workers reported a new route to access perfluoroalkyl-based metallacycles with varied ring sizes (Scheme 9) [72]. Tetra-perfluoroalkyl-dizinc compounds 33 were obtained from the reaction between diiodoperfluoroalkanes and 2.0 equiv. of ZnEt₂. The treatment of 33b with 2.0 equiv. of CuCl in DMF afforded a 10-membered dicopper compound 34. The averaged Cu-C bond (1.93 Å) is relatively shorter than that of other non-fluorinated alkyl cuprates. Comound 34 was inactive towards electrophiles, which is consistent with its strong Cu-C bond and high stability.

Metal-containing aromatic molecules have attracted much attention recently. Compared to six-membered metallobezenes and metallabenzynes as well five-membered metallocylcopentadienes, macrocyclic metalla-aromatics are less explored. Trans,cis,trans,cis,cis-[10]annulene is a conjugate molecule with 10-π electrons. However, it is non-aromatic due to the steric repulsion between the two internal hydrogen atoms. It was previously envisioned that the replacement of two internal CH with a transition metal would not only release the steric hindrance but would also provide electrons to the delocalized π-system. Aromatic dicupra[10]annulenes 36 with four lithium atoms were first synthesized and isolated from the reaction of dilithio reagents 1 and a CuCl complex by Xi’s group (Scheme 10) [32]. In the above reactions, 35 was isolated with two lithium atoms as a small amount of byproduct that is also able to be efficiently synthesized from dilithio reagent and mesitylcopper. More importantly, reversible two-electron transformations between 35 and 36 were observed [47]. Compared to the alternating bonds of the butadienyl skeleton in 35a (1.52, 1.37 Å), the C-C bonds of annulene moiety in 36a are remarkably averaged (1.47, 1.42 Å), suggesting a considerable π-delocalization (Figure 5). The Cu-C bonds of 36a (1.92 Å) were notably shorter than those of 35a (1.96 Å), which might be attributed to the extra two electrons posed to 36a. The Cu 2p_3/2 binding energy (932.9 eV) and Cu LMM kinetic energy (915.8 eV) of 36b, as displayed by XPS data, show that the Cu atoms in 36 are more likely to be Cu(I). With the assembled experimental results (the above crystallographic data; low-frequency resonance signals of ⁷Li NMR: −5.1 to −6.2 ppm) and DFT calculations (negative values of ISE: −11.7 and −21.5; large negative NICS values: −9.0 to −12.8), it is safe to conclude that dicupra[10]annulenes are aromatic. AdNDP analysis has previously suggested dicupra[10]annulenes could be regarded as a 10-π aromatic system. Based on independent theoretical calculations, Grande-Aztatzi and co-workers consider dicupra[10]annulenes to be a metalla-naphthalene [73]. Zhu and co-workers suggest that dicupra[10]annulenes have 16e Craig-type Möbius aromaticity [74]. Additionally, both Zhu’s and Sundholm’s calculations reveal that lithium counterions play an important role in these organocopper complexes [74,75].

As shown in Scheme 11, the C,N-chelated dicopper compounds 37,38 and 40–42 documented in the literature are generally macrocyclic neutral species since the two negative ligand charges are exactly balanced with two Cu(I) atoms [76,77,78,79,80,81,82]. An early example of structurally well-defined alkylcopper compounds was [Me₃SiCH₂Cu]₄, in which trimethylsilyl groups not only pose steric hindrance to increase the stability but also improve the solubility by virtue of the lipophilicity of the silicon groups [83,84]. White and others found that incorporating an ortho-bulky bis-trimethylsilyl-alkyl ligand at a pyridine can stabilize t dicopper compound 37 to a great extent [76,80]. Compound 37 can be sublimated at 160 °C without decomposition. Compund 37 features an 8-membered core structure (C₄N₂Cu₂). When increasing the distance between the alkyl moiety and the pyridine, a dicopper compound 38 with a 10-membered ring can also be generated, as reported by Eaborn and Smith [81]. However, with a smaller steric effect, the alkyl-pyridine ligands will bind four copper atoms, generating a tetracopper complex 39 [77]. Similarly, three dicopper compounds 40–42 based on 1-azaallyl, aryl-phosphanimine, and aryl-imine ligands were synthesized from the corresponding lithiated starting materials and copper salts by Lappert [78], Stalke [79], and Schmidt [82], respectively. The Cu-C bonds of 37–40 are in the range of 1.93–1.95 Å, which suggests a mostly 2c-2e Cu-C bond, while the Cu-C bonds (1.91 Å) in 41 and 42 fall in the short end of the range observed for Cu-C 2c-2e bonds. Cyclic voltammetry and bulk electrolysis of 37 show very promising evidence of the existence of formal copper(II) alkyl species with relatively high stability.

5. Polynuclear Organocopper Clusters

Fifteen years ago, Xi and co-workers reported that the cis-1,3-butadienyl 1,4-dicopper species, obtained from the reaction beteen dilithio reagent and 2.0 equiv. of CuCl, efficiently produced multi-substituted semibullvalenes after thermolysis at 50 ℃ in non-coordinating toluene solvent [85]. Later on, it was found that the same copper species underwent lithium iodide-assisted linear dimerization in Et₂O at 0 °C to afford an all-cis octatetraenyl dicopper intermediate that could be subject to subsequent Pd-catalyzed cross-coupling reactions with halides [86]. In order to unveil the nature of butadienyl copper species, Xi and co-workers carefully conducted transmetalation reactions at a low temperature (−78 ℃) in order tto avoid uncontrollable thermal decompositions. The reaction of the dilithio reagents 4Pr-1 with 2.0 equiv. of CuCl in Et₂O afforded 1,4-dicopper tetramer 43 at a 75% yield (Scheme 12) [56]. When 2.5 equiv. of CuCl was used in the presence of 2.0 equiv. of TMEDA, a different copper aggregate 44 was obtained at a 63% yield. Compound 44 could be transformed to 43 through treatment with 1.0 equiv. of dilithio reagent. Additionally, the products were also dependent on the solvents and additives, e.g., lithium salts. When THF was used as a solvent, the reaction between dilithio reagent and 2.5 equiv. of CuCl produced 45, which was linked by a 1,3-butadienyl and 1,3,5,7-octatetraenyl units at a 43% yield. In the presence of extra LiI, the above reaction produced the 1,3,5,7-octatetraenyl-1,8-dicopper compound 46 at a 63% yield. The unit of 1,3,5,7-octatetraenyl was formed from the dimerization of butadienyl copper complexes, as observed in previous work [86]. Compounds 46 and 43 can also be generated from the reaction between 45 and LiI. Structural analysis of 43–46 reveals that the Cu-C bonds and C-Cu-C bond angles are in the range of 1.96–2.08 Å and 73–84°, respectively. The close Cu-Cu distances (2.37–3.01 Å) suggests weak d¹⁰-d¹⁰ interactions in all of the copper aggregates above (Figure 6).

The reactivity of the above bis-enyl copper compounds was displayed by the reactions of 43 and 46 (Scheme 13). Compound 43 produced tricycle-octa-diene 47 and cyclopenta-2,4-dien-1-imine 48 via thermolysis and the insertion of isocyanide, respectively. When heated in THF, 46 afforded multi-substituted cyclooctatetraene 49. In toluene, the thermolysis of 46 produced semibullvalene 50 as the major product, which was consistent with earlier work [85].

Copper-mediated reactions of zirconacyclopentadienes (or zirconaindenes) and electrophiles developed by Takahashi and others are extensively used to construct a variety of functional organic molecules [87,88]. However, organocopper intermediates remain elusive though the methodologies were reported about twenty years ago. Inspired by this work, Xi’s group revisited the transmetalation reaction between zirconium reagents and copper salts. Salt-free zirconaindenes 51 and zirconacyclopentadienes 52 were readily prepared according to the reported procedure [87]. A 1:3 reaction of 51 and CuCl in THF/ TMEDA afforded a tetrameric styrenyl copper aggregate 53 at a 53% isolated yield (Scheme 14) [57]. When 51 was treated with 2.5 equiv. of CuCl, a hexameric styrenyl copper aggregate 54 was obtained at a 27% isolated yield after crystallization in Et₂O and tetrahydrothiophene (THT). Zirconacyclopentadienes 52a,b reacted with 3.4 equiv. of CuCl in THF, affording trimeric butadienyl copper aggregates 55 and 56 at 72% and 51% yields, respectively. The additional ancillary ligand (TMEDA vs. THT) could change the aggregate state of the styrenyl clusters. In terms of molecular structures, 53 consists of ten copper atoms, while 54 is a hexanuclear cluster with twelve copper atoms. The lengths of the Cu-C(alkenyl) bonds (1.98–2.01 Å) are comparable to those of the Cu-C(aryl) bonds (1.98–2.02 Å), as shown in 53 (Figure 7). The binding mode of 54 looks more complicated and also more unique. Compound 54 comprises two unusual μ₃-alkenyl carbon atoms (black dots: C17, C17′) and the corresponding two tri-coordinated copper atoms (red: Cu2, Cu2′). Compounds 55 and 56 share identical core copper structures. In addition, 55 and 56 can be regarded as variants of 44 by replacing the two lateral lithium atoms with copper atoms. The substituents of the butadienyl skeleton seem to have an influence on the stability of their copper clusters. In Et₂O, 55 is relatively stable below 10 ℃, whereas 56 is would thermally decompose and release the copper mirror at 10 ℃ within 0.5 h. The reactivity of 55 was provided by its coupling reaction with diiodobenzene, producing multi-substituted naphthalene.

The intramolecular stabilization effect developed by van Koten was initially used to synthesize organocopper clusters. The reaction of 29 with 1.0 equiv. of CuBr afforded a LiBr-free tetranuclear copper cluster 31, which is thermally stable up to 170–185 ℃ (phenylcopper: slow decomposition at room temperature) [89,90]. The molecular structure of 31 comprises four copper atoms in a butterfly arrangement (Scheme 15) [91,92]. Each C_ipso bridges two copper atoms via the 3c-2e Cu-C bonds (1.97–2.16 Å). Each copper atom has a distorted trigonal planar geometry (Figure 8). Variable temperature ¹H NMR spectroscopy shows that the methylene and NMe₂ signals are singlets that are independent of the temperature, which suggest a weak Me₂N→Cu interaction in solution. Whereas the reaction between 31 and 1.0 equiv. of dppe afforded a 1:1 donor–acceptor copper complex 57, the addition of excess dppe resulted in an unusual C-P bond cleavage with the formation of a phosphidocopper complex 58, Ph₂PCH=CH₂, and Ar-H [93,94]. This process can be explained by a concerted mechanism, as shown in the box. A 1:1 reaction of 31 with CuBr affords a polymeric complex 59. When cupric halides were used instead, the oxidative coupling products Ar-Ar, Ar-X, and Ar-H were obtained [95]. In addition, van Koten and co-workers reported a serial of organocopper aggregates with the N-C-C-C-Cu connectivity supported on other C,N-bidentate ligands, such as 8-(dimethylamino)naphthyl [96], 2-oxazoline-4-methylphenyl [97,98], and (2-(dimethylamino)phenyl)-1-vinyl [99,100,101,102].

When a built-in heteroatom moiety is directly attached to the phenyl skeleton, the C,N-bidentate ligand displays different features when binding to the copper atoms (N-C-C-Cu connectivity). Van Koten reported that the treatment of 60 with copper bromide affords a hexanuclear copper cluster 61 that includes six copper atoms in a distorted octahedral arrangement (Scheme 16) [103,104,105,106]. Each bromide atom bridges the trans-equatorial edges. Each C,N-bidentate ligand bridges three copper atoms on the face of an octahedron. Molecular weight determinations and ¹H NMR spectroscopy supported that 61 retained its aggregate structure in solution. Me₂N was observed as two broadened singlets (1.84, 2.92 ppm) at room temperature, suggesting a relatively strong Me₂N→Cu bonding interaction in contrast to the one seen in 31. When the temperature increases to 90 ℃, the two singlets coalesced to a single singlet (2.42 ppm). In 31, the C_ipso atom and Me₂N bind to the same copper atom, leading to a 5-membered chelate ring (A), whereas in 61, Me₂N→Cu coordination takes place with a different copper atom to which the aryl group is not bonded (B), which is due to the steric effects. When 61 was treated with CuOTf, a selective, quantitative coupling of biaryls Ar’-Ar’ occurred, which was unexpected [107,108]. Only a trace amount of Ar’-H was formed, which excluded the pathway involving free radicals. The charge transfer from the core coppers to the strongly electron accepting OTf groups, which reduces the electron density of the core skeleton, accounts for the C-C bond formation. The proposed mechanism (3), including the valence disproportionation inside the metal core followed by the reductive elimination of Ar’-Ar’, rationalized the large influence of the identity of the anions.

Based on the earlier work shown in Scheme 16, van Koten reported that 62 could be synthesized via three methods (Scheme 17) [109,110,111,112]. The thermolysis of Ar’₄(CCPh)₂Cu₆ 62 results in the exclusive formation of the mixed coupling product Ar’-CCPh. The absence of Ar’-H or PhCC-H from hydrogen abstraction indicates that the reaction undergoes a concerted pathway rather than proceed via free radicals. The molecular structure of 62 comprises exclusively triangular copper faces occupied by one Ar’ and one PhCC ligand, whereas Cu₃Ar’₂ or Cu₃(CCPh)₂ faces are absent, explaining the high selectivity of the asymmetric C-C bond formation (Figure 9).

6. Conclusions and Perspective

In this review, we have summarized the structurally well-defined organocopper compounds based on C,N- and C,C-bidentate ligands, from mononuclear copper spiro compounds to polynuclear copper clusters. It was shown that the intramolecular stabilization effect of the C,N-bidentate ligands and the cooperative effect of the C,C-bidentate ligands (cis-1,3-butadienyl) can both enhance the solubility and stability of the resulting organocopper compounds. In the molecular structures, copper(I) atoms adopt linear, trigonal, or tetrahedral geometries, depending on the coordination numbers. In charge-neutral copper aggregates, the electron-deficient 3c-2e Cu-C bonds are predominant, whereas in more electron-rich organocuprates, the Cu-C bonds have a much higher degree of 2c-2e bond character.

The strategy for enhancing the intramolecular coordination was further developed by adding two build-in heteroatom moieties to the aryl ring. Consequently, the original ligands turn into tridentate pincer ligands, e.g., C,N,N and N,C,N-ligands [29], which are now widely used in organic synthesis and small molecules activations. Another point of interest in organocopper chemistry is high-valent organocopper compounds [113]. It was found that the macrocyclic ligands with three intramolecular heteroatom moieties, e.g., C,N,N,N-cyclic ligands, were excellent platforms to support organocopper(III) compounds. Moreover, there has been increasing interest in more challenging organocopper(II) compounds in recent years since many Cu-catalyzed reactions may have Cu(II)-C containing species as their intermediates.