The potential of a cyclo-As₅ ligand complex in coordination chemistry

Abstract

The reaction of [Cp*Fe(η⁵-As₅)] (1) with Cu^I halides leads to the formation of the 1D-polymeric compounds [{Cu(μ-X)}₃(CH₃CN){Cp*Fe(η⁵:η²:η²:η²-As₅)}]_n (X = Cl (2), Br (3)), [{Cu(μ₃-I)}₂{Cp*Fe(η⁵:η²:η²-As₅)}]_n (4) and [{Cu(μ-I)}₃{CuI}{Cp*Fe(η⁵:η²:η²:η¹:η¹-As₅)}{Cp*Fe(η⁵:η⁵:η²-As₅)}]_n (5). The polymers are built up by the π-coordination of the cyclo-As₅ ring to (CuX)_n moieties forming discrete units, which are additionally linked by weak intermolecular As⋯Cu σ-interactions. Only polymer 4 is an exception revealing a (CuI)_n ladder which is alternately coordinated by molecules of 1. In a side arm of polymer 5 a novel η⁵-coordination of a Cu atom below the cyclo-As₅ ring is found, showing an unprecedented heteroleptic triple-decker sandwich complex with a polyarsenic middle-deck. All products are characterised by single crystal X-ray structure analysis. To define the differences in the coordination behaviour of 1 and its phosphorus analogue [Cp*Fe(η⁵-P₅)] (1a), DFT calculations are described.

---

Introduction

The self-organisation of discrete units to form supramolecular aggregates and networks has become one of the foremost areas of current research.¹ Contrary to established principles in this field which make almost exclusively use of N- and O-donor-containing ligands to connect different metal centres, we developed an alternative concept using P_n ligand complexes as linking moieties. The concept is based on the use of different organometallic P_n ligand moieties in the reaction with Lewis acidic Cu^I and Ag^I complex units to obtain inorganic polymers and large spherical oligomers.² Note that the first attempts in this direction were carried out by the groups of Stoppioni and Peruzzini using [(triphos)M(η³-P₃)] (M = Co, Rh, or Ir; triphos = 1,1,1-tris(diphenylphosphinomethyl)ethane) as starting material.^3,4 We have already shown that the reaction of the tetrahedral complex [{CpMo(CO)₂}₂(μ,η²-P₂)] with Cu^I and Ag^I salts yields dimeric and polymeric coordination compounds.⁵ Also the use of the cyclo-P₃ ligand complex [CpMo(CO)₂(η³-P₃)] results in the formation of novel 1D polymers.⁶

However, the most progress is obtained using the pentafold-symmetric molecule [Cp^RFe(η⁵-P₅)] (Cp^R = Cp*, C₅Me₄Et) containing a cyclo-P₅ ligand as a building block. Thus, in the reaction of [Cp^RFe(η⁵-P₅)] with Cu^I,⁷ Ag^I8 and Tl^I9 starting materials, one dimensional (1D) and two dimensional (2D) polymers are formed, in which the 1,2- (type A), η²:η¹- (type B) and η⁵:η¹-coordination modes (type C) are found. By applying special reaction conditions and stoichiometries, the reaction with Cu^I halides leads to the formation of soluble spherical aggregates with fullerene-like topology,¹⁰ in which all P atoms coordinate with their lone pairs (type D). Furthermore, these systems react in a template controlled reaction to form large spherical clusters by encapsulating guest molecules such as C₆₀ or o-carborane.¹¹ Besides pentaphosphaferrocene, so far, only the cyclo-P₄ ligand complex [Cp′′Ta(CO)₂(η⁴-P₄)] (Cp′′ = η⁵-C₅H₃tBu₂) forms a fullerene-like spherical cluster revealing a C₃₂ topology.¹²

In view of the versatility of P_n ligand complexes as linking moieties in coordination and supramolecular chemistry we are interested in exploring the potential of As_n ligand complexes.¹³ Recently, novel results have been reported based e.g. on the generation of AsP₃.¹⁴ Although the reactivity of As_n ligand complexes has been investigated in organometallic chemistry to some extent,¹⁵ no examples of supramolecular aggregates featuring these compounds as connecting moieties are known. The first attempts with [Cp*Mo(CO)₂(η³-As₃)]¹⁶ as a starting material in coordination chemistry towards Cu^I halides have demonstrated the formation of discrete dimers which are connected by very weak Cu⋯As interactions to incoherent polymeric assemblies.¹⁷ However, the coordination behaviour of [Cp*Fe(η⁵-As₅)] (1)¹⁸ is of particular interest. Here the question is raised as to whether similar coordination modes A–D of the [Cp*Fe(η⁵-P₅)] (1a) are to be designed for the cyclo-As₅ ligand complex 1 by using mainly the lone pairs of the group 15 element atoms. Furthermore, the ability of 1 to form spherical aggregates with fullerene-like topology is of special interest. We report herein on the first results using the cyclo-As₅ ligand complex 1 in coordination chemistry, which differs significantly from the coordination behaviour of the cyclo-P₅ homologue.

Results and discussion

DFT calculations

Based on the relatively similar behaviour of organophosphorus and organoarsenic compounds in general one would expect a similar reactivity of As_n ligand complexes as is observed for P_n ligand complexes. To gain insight into the potential reactivity of 1 and [Cp*Fe(η⁵-P₅)] (1a) towards Lewis acids, their electronic structures were modelled by DFT calculations using the hybrid B3LYP functional.†¹⁹ The geometries of 1 and 1a were optimised in the staggered conformation of the C_5v symmetry. The isosurfaces of the frontier molecular orbitals of 1 and 1a are depicted in Fig. 1.

Fig. 1 Isosurfaces of the frontier molecular orbitals in 1 and 1a.

The degenerate e₁ orbital pairs represent the highest occupied molecular orbital (HOMO) and are mainly localised on the cyclo-As₅ and cyclo-P₅ ligands. The a₁ symmetric HOMO−1 orbital is composed of the iron d_z² orbital with almost no ligand orbital contribution. The energies of the HOMO and HOMO−1 in 1 lie 0.33 eV and 0.27 eV higher than in 1a. Furthermore, differences in the electronic structures of 1 and 1a can also be observed in the lowest unoccupied molecular orbital (LUMO) region. The LUMO in 1 is doubly degenerate (e₂ symmetry) involving mainly π orbital contribution from the cyclo-As₅ ligand, and the LUMO+1 (e₁ symmetry) contains metal and Cp* as well as cyclo-As₅ ligand contributions. In 1a the relative energies of the unoccupied e₂ and e₁ orbitals are reversed.† However, it has to be noted that the energy difference between the unoccupied orbitals e₁ and e₂ is very narrow (1: 0.07 eV and 1a: 0.11 eV). The HOMO–LUMO gap in 1 (4.16 eV) is slightly smaller than that in 1a (4.48 eV) which is caused by the higher energy of the occupied frontier molecular orbitals in 1 in comparison to 1a. The natural charge distribution shows a negative charge accumulation on the iron atom (−1.33 in 1 and −1.11 in 1a) and a positive charge accumulation on the arsenic and phosphorus atoms (0.19 in 1 and 0.14 in 1a). This and the relative energy of the HOMO in 1 in comparison to 1a suggest a somewhat more electron releasing character of the cyclo-As₅ unit in 1 compared to the cyclo-P₅ moiety in 1a. Accordingly, one would expect that 1 should be a weaker ligand than 1a. Due to the small differences in the ground state electronic structure, a similar coordination behaviour might be expected. Surprisingly, the experimental results show a considerable difference.

Synthesis and characterisation of the products

The reactions of complex 1 with copper halides CuX were carried out in CH₂Cl₂/CH₃CN solutions using layering techniques (Scheme 1). In the cases of copper chloride and bromide the isostructural compounds [{Cu(μ-X)}₃(CH₃CN){Cp*Fe(η⁵:η²:η²:η²-As₅)}]_n (X = Cl (2), Br(3)) were obtained in good yields. In contrast, the reaction with copper iodide yields the products [{Cu(μ₃-I)}₂{Cp*Fe(η⁵:η²:η²-As₅)}]_n (4) and [{Cu(μ-I)}₃{CuI}{Cp*Fe(η⁵:η²:η²:η¹:η¹-As₅)}{Cp*Fe(η⁵:η⁵:η²-As₅)}]_n (5), comprising the same stoichiometric composition but showing different structural arrangements.† By varying the stoichiometry of the reactions no change in the nature of the products is observed. As mentioned before, the reaction of the phosphorus analogue 1a with copper halides yields 1D and 2D polymeric compounds,¹⁰ whereas under diluted reaction conditions or in the presence of templates this reaction proceeds to form spherical aggregates.¹¹ Surprisingly, by applying the latter reaction conditions to the arsenic derivative 1 (reactions in diluted solutions or in the presence of e.g. C₆₀ as a template), the polymeric compounds 2–5 are exclusively formed. This behaviour of the pentaarsaferrocene was unexpected at a first glance; however, it demonstrates the extended π coordination behaviour of 1 in contrast to a more σ donation reactivity pattern of the pentaphosphaferrocene.

Scheme 1 Reaction behaviour of [Cp*Fe(η⁵-As₅)] (1) with different Cu(I) halides. Depicted are the repeating units of the products 2–5.

The products are stable in the solid state under an atmosphere of nitrogen and are insoluble in common solvents without decomposition. This is also reflected in the ESI mass spectra.† In CH₃CN solutions no peaks of molecular units of the polymers are detected. Peaks of low intensity could be assigned to aggregates formed by two molecules of 1 and one Cu⁺ cation or a fragment of (Cu₂X)⁺ (X = Cl, Br, I). Furthermore, a peak of 1 in combination with one Cu⁺ cation and one solvent molecule CH₃CN is identified. This indicates that small amounts of depolymerised material are present in the solutions of the solids at ambient temperatures.

The molecular structures of compounds 2–5 were determined by single crystal X-ray diffraction.‡ The 1D-polymeric compounds 2 and 3 are isostructural; their molecular structure is shown in Fig. 2, and the interaction between the monomeric building blocks forming the 1D polymeric chain is depicted in Fig. 3. Table 1 contains selected bond lengths and angles.

Fig. 2 Section of the 1D polymeric structure of 2 (X = Cl) and 3 (X = Br) revealing the repeating units. Hydrogen atoms are omitted for clarity.

Fig. 3 Orientation of the repeating units of 2 and 3 along the crystallographic a-axis. Hydrogen atoms and Cp*Fe fragments are omitted for clarity.

Table 1 Comparison of selected geometrical parameters (distances/Å) for 2 and 3.

	2 (X = Cl)	3 (X = Br)
As(1)–As(2)	2.3911(10)	2.3880(9)
As(1)–As(5)	2.3613(11)	2.3671(9)
As(2)–As(3)	2.3751(10)	2.3793(8)
As(3)–As(4)	2.4107(10)	2.4100(9)
As(1)–Cu(1)	2.4733(10)	2.4698(11)
As(2)–Cu(1)	2.4692(10)	2.4809(11)
As(3)–Cu(2)	2.4534(10)	2.4577(11)
As(4)–Cu(2)	2.5647(11)	2.5513(11)
As(4)–Cu(3)	2.5677(11)	2.5783(11)
As(3)–Cu(2′)	2.8372(13)	2.9653(11)
As(1)–Cu(1′)	2.9075(13)	2.9608(11)
Fe(1)–As(3)	2.4660(11)	2.4664(10)
Fe(1)–As(4)	2.5538(11)	2.5621(11)
Cu(1)–X(1)	2.3105(17)	2.4260(11)
Cu(2)–X(3)	2.2756(16)	2.3944(11)
Cu(3)–X(3)	2.4005(17)	2.5283(10)

---

The main structural motif of 2 and 3 is a monomer-like building block, consisting of the pentaarsaferrocene moiety coordinated by all five arsenic atoms to the three copper atoms of a six-membered (CuX)₃ ring in a novel η²:η²:η² coordination mode. The (CuX)₃ ring is centred below the cyclo-As₅ ring and shows a distorted chair conformation. The formation of such copper halide rings is known in the literature,^20,21 but usually, they are sections of a polymeric (CuX)_n network, in contrast to the discrete (CuX)₃ rings found in 2 and 3. The Cu–X (X = Cl, Br) distances are comparable to those found in the literature^21–23 ranging from 2.276(2) Å to 2.401(2) Å in 2 and 2.394(1) to 2.528(1) Å in 3, respectively. Interestingly, the distance between Cu(2) and Cu(3) is remarkably short (2.669(1) Å (2); 2.667(1) Å (3), respectively), indicating the presence of Cu⋯Cu interactions.²⁴ Interactions of this kind have been well documented and often cause photoluminescence properties.^25,26 However, preliminary investigations for 2 and 3 show no luminescence.

Due to the π-coordination of the As–As edges towards the copper atoms, all As–As distances in the cyclo-As₅ unit are elongated in comparison to 1 (2.312(2)–2.319(2) Å)¹⁸ and vary in a bigger range. The shortest bond is found between As(1)–As(5) (2.361(1) Å (2); 2.367(1) Å (3)) and the longest is found between the doubly coordinating atom As(4) and its neighbours (e.g. As(3)–As(4): 2.411(1) Å (2); 2.410(1) Å (3)). A further consequence of the double As⋯Cu interaction of As(4) is the distortion of the planar cyclo-As₅ unit to form a more envelope-like shape, dragging As(4) about 0.26 Å out of the cyclo-As₅ plane towards the copper atoms and elongating the Fe(1)–As(4) bond to 2.554(1) Å (2) and 2.562(1) Å (3) when compared e.g. to the Fe(1)–As(3) bond (2.466(1) Å (2) and 2.466(1) Å (3)). The Cu–As distances (2.453(1)–2.568(1) (2); 2.458(1)–2.578(1) (3)) of the arsenic atoms (As(1) and As(3), respectively) of the cyclo-As₅ ring and the copper atoms of the (CuX)₃ ring are in the range of those reported.^17,27 Besides these interactions, one copper atom is additionally coordinated by a solvent molecule of acetonitrile. The most intriguing feature of 2 and 3 is the additional end-on coordination of two copper atoms by arsenic atoms of neighbouring molecules of 1, leading to an infinite aggregation of the monomeric building blocks along the crystallographic a axis. In this way, a sinusoidal undulated 1D polymer is formed (Fig. 3). The distances between these atoms (2.837(1) to 2.908(1) Å for 2 and 2.961(1) to 2.965(1) Å for 3) are shorter than the sum of the van der Waals radii of As and Cu (3.25 Å)²⁸ and indicate a weak interaction. However, they are responsible for the low solubility of these 1D polymers. Similar weak interactions are found in [{Cu(μ-X)[Cp*(CO)₂Mo(μ,η³:η²-As)]}₂] (X = Cl, Br, I)¹⁷ and demonstrate the capability of arsenic to realize higher coordination numbers when compared to its lighter homologue phosphorus. Bearing in mind the π-coordination mode of the As–As edges, one could propose a distorted trigonal-planar environment for the copper atoms, with one vertex in the middle of the As–As bond and two vertices occupied by the halide atoms (alternatively, a square pyramidal environment regarding both As atoms). The copper atoms are slightly pulled out of this plane by the additional coordination from the acetonitrile molecule or the σ-coordination of the neighbouring arsenic atom of the next molecule of 1 in the polymeric strand.

A section of the 1D polymeric structure of 4 is depicted in Fig. 4; selected bond lengths and angles are summarized in Table 2.‡ Compound 4 shows a (CuI)_n ladder as the main structural feature, of which only a few examples are known in the literature.²⁰ However, in 4 the copper atoms of this stair-like strand are coordinated by molecules of 1, alternately placed on both sides of the CuI chain. The arsenic atoms of the cyclo-As₅ moiety adopt a η²:η² coordination mode which always has two arsenic atoms coordinated side-on to one copper atom, similar to 2 and 3. Also, all of the As–As bonds are elongated relative to 1 due to the additional coordination revealing longer bonds between As(2)–As(3) (2.392(2) Å) and shorter bonds between As(3)–As(4) (2.346(2) Å). In contrast to 2 and 3 a unique arsenic atom shows no further coordination and also the planarity of the cyclo-As₅ ring is better maintained. The Cu–As distances are between 2.474(2) and 2.521(2) Å and are comparable to those found in the literature.^17,27 Taking into account the π-coordination mode of the As–As edges, the Cu atoms are in a tetrahedral environment, spanned by the three halogen atoms and the centre of the As–As bond. The 1D polymeric (CuI)_n strand shows a slightly distorted stair-like structure, which is similar to the coordination polymer [(CuI)₂(Quin)]_n.^29a In contrast, in 4 one molecule of 1 bridges one “step” of the (CuI)_n chain, causing a narrow Cu(1)–I(2)–Cu(2′′) angle of 88.44(5)° compared to the unbridged “step” with a wider Cu(2)–I(1)–Cu(1′′) angle of 120.51(6)°. The Cu–I distances (2.671(2)–2.715(2) Å) are within the expected range.^21,22,29

Fig. 4 Section of the polymeric structure of 4. Hydrogen atoms are omitted for clarity.

Table 2 Selected geometrical parameters (distances/Å and angles/°) for 4

As(2)–As(3)	2.392(2)	As(2)–Cu(1)–As(3)	57.29(6)
As(3)–As(4)	2.346(2)	I(1)–Cu(1)–I(1′)	92.25(5)
As(3)–Cu(1)	2.512(2)	Cu(1)–I(2)–Cu(2′′)	88.44(5)
Fe(1)–As(3)	2.519(2)	Cu(2)–I(1)–Cu(1′)	120.51(6)
I(2)–Cu(2′′)	2.671(2)
Cu(1)–I(1)	2.715(2)

---

The solid state structure of 5 is shown in Fig. 5; selected bond lengths and angles are summarized in Table 3.‡ Despite the fact that 4 and 5 contain pentaarsaferrocene and CuI in the same stoichiometric ratio of 1 : 2, the latter compound shows a strikingly different structure in the solid state, combining structural motifs of 2, 3, and 4 along with new coordination modes. At a first glance, the central pentaarsaferrocene moiety reveals the same coordination mode as in 4, adopting a η²:η² coordination towards two copper atoms. The coordinating As–As edges are slightly elongated (As(2)–As(3): 2.385(1) Å) in contrast to the non-coordinating bonds (As(1)–As(2): 2.336(1) Å) similar to 4. However, the 1D polymer is not formed by a (CuI)_n chain as in 4. Instead a further end-on coordination of two arsenic atoms towards copper atoms is observed, as is found in 2 and 3. However in 5, the σ-coordination is much stronger, giving a short Cu–As distance (As(2)–Cu(1′′): 2.569(1) Å) which is even shorter than the distance to the π-coordinated Cu atom (As(2)–Cu(1): 2.601(1) Å). Finally, this leads to a novel η²:η²:η¹:η¹ coordination mode of the cyclo-As₅ ring. The atom Cu(1) is situated in a tetrahedral environment and coordinated by two iodide atoms, one σ-coordinating As atom and one π-coordinating As–As edge. The two copper atoms below the cyclo-As₅ ring belong to a (CuI)₃ ring, revealing a distorted chair structure similar to compounds 2 and 3. The Cu–I distances are within the expected range (2.591(1)–2.622(1) Å) and are comparable to those found in the literature.^21,29 An interesting structural feature is introduced by the additional coordination of the third copper atom of the six-membered CuI ring. This copper atom is side-on coordinated by two arsenic atoms of a second molecule of 1, revealing a slightly elongated As–Cu bond (As(4)–Cu(2): 2.663(2) Å). Furthermore, this cyclo-As₅ ring coordinates in an η⁵ fashion to another Cu atom of a dumbbell shaped CuI located nearly perpendicular above the cyclo-As₅ plane (angle between the cyclo-As₅ plane and the CuI-dumbbell ≈ 79°). This structural motif is reminiscent of a triple-decker sandwich complex which exists for cyclo-As₅ complexes only in a few examples in a pure homoleptic environment.^15a However, here the first example of a heteroleptic arrangement is found. The copper atom is not ideally centred above the cyclo-As₅ ring; this is deduced from the differences in the As–Cu distances, decreasing from 2.696(2) Å (As(4)–Cu(3)) to 2.479(2) (As(6)–Cu(3)). Within the cyclo-As₅ ring, similar As–As distances are found (As(4)–As(5): 2.370(1) Å; As(5)–As(6): 2.368(1) Å) with a slightly elongated bond (As(4)–As(4′): 2.389(1) Å) due to the additional π-coordination. The CuI dumbbell is attached to the (CuI)₃ ring through a Cu(2)⋯I(3) interaction of 2.740(2) Å and a very short Cu⋯Cu interaction of 2.563(2) Å; this is in the range of distances usually found in copper metal (2.56 Å).^24,26

Fig. 5 Section of the polymeric structure of 5. Hydrogen atoms are omitted for clarity.

Table 3 Selected geometrical parameters (distances/Å and angles/°) for 5

As(1)–As(2)	2.3358(10)	Cu(1)–I(1)	2.5908(12)
As(2)–As(3)	2.3852(9)	Cu(1)–I(2)	2.6222(11)
As(4)–As(5)	2.3696(12)	Cu(2)⋯I(3)	2.7404(16)
As(5)–As(6)	2.3675(12)	Cu(3)–I(3)	2.5237(17)
As(4)–As(4′)	2.3890(13)	Cu(2)–Cu(3)	2.563(2)
As2–Cu(1)	2.6007(12)	As(2)–Cu(1)–As(3)	55.72(3)
As(2)–Cu(1′′)	2.5694(14)	As(4)–Cu(2)–As(4′)	53.30(4)
As(4)–Cu(2)	2.6631(15)	As(4)–Cu(3)–As(4′)	52.61(4)
As(4)–Cu(3)	2.6956(18)	Cu(1)–I(2)–Cu(2)	106.80(4)
As(5)–Cu(3)	2.5968(14)	I(2)–Cu(2)–I(2′)	112.15(5)
As(6)–Cu(3)	2.4794(18)	Cu(2)–I(3)–Cu(3)	58.10(5)

---

Conclusions

The results show that the cyclo-As₅ ligand complex 1, surprisingly, reveals a preferred edge-bridging coordination mode to form π-coordination as found in 4 (type E), 5 (type F and G) and in a maximum manner in 2 and 3 (type H).

This is in contrast to the preferred σ-coordination behaviour of the cyclo-P₅ complex homologue [Cp*Fe(η⁵-P₅)]. In addition, σ-coordination modes of two As atoms of the ring are found, revealing intermolecular interactions responsible for forming 1D polymeric structures. Whereas this feature is present in H (compounds 2 and 3) to give rather weak interactions, this behaviour is more pronounced in G (compound 5) to result in short As–Cu bonds. However, the most striking observation is the formation of a heteroleptic triple-decker-like sandwich complex motif in 5 in which Fe and Cu are incorporated into the multiple-decker arrangement. These surprising results prove the intriguing new coordination features of the arsenic complexes and extend our knowledge on the variety of the palette of coordination modes and compounds formed by ligand complexes of the group 15 elements.

Notes and references

1. (a) J.-M. Lehn, Supramolecular Chemistry: Concepts and Perspectives, VCH, Weinheim, 1995; (b) R. W. Saalfrank and B. Demleitner, in Transition Metals in Supramolecular Chemistry, ed. J.-P. Sauvage, Wiley-VCH, Weinheim, 1999, vol. 5, pp. 1–51; (c) M. M. Conn and J. Rebek, Jr., Chem. Rev., 1997, 97, 1647–1668; (d) C. J. Jones, Chem. Soc. Rev., 1998, 27, 289–299; (e) G. F. Swiegers and T. J. Malefetse, Chem. Rev., 2000, 100, 3483–3537; (f) S. Leininger, B. Olenyuk and P. J. Stang, Chem. Rev., 2000, 100, 853–908; (g) M. Ruben, J. Rojo, F. J. Romero-Salguero, L. H. Uppadine and J.-M. Lehn, Angew. Chem., 2004, 116, 3728–3747 ( Angew. Chem., Int. Ed. , 2004 , 43 , 3644–3662 ); (h) M. Fujita, M. Tominaga, A. Hori and B. Therrien, Acc. Chem. Res., 2005, 38, 369–378; (i) T. Kreickmann and F. E. Hahn, Chem. Commun., 2007, 1111–1120.
2. M. Scheer, Dalton Trans., 2008, 4372.
3. (a) M. Di Vaira, M. P. Ehses, M. Peruzzini and P. Stoppioni, Polyhedron, 1999, 18, 2331–2336; (b) M. Di Vaira, P. Stoppioni and M. Peruzzini, J. Chem. Soc., Dalton Trans., 1990, 109–113.
4. (a) M. F. Cecconi, C. A. Ghilardi, S. Midollini and A. Orlandini, J. Chem. Soc., Chem. Commun., 1982, 229–230; (b) F. Cecconi, C. A. Ghilardi, S. Midollini and A. Orlandini, Angew. Chem., 1983, 95, 554–555 ( Angew. Chem., Int. Ed. Engl. , 1983 , 22 , 554–555 ).
5. (a) J. Bai, E. Leiner and M. Scheer, Angew. Chem., 2002, 114, 820–823 ( Angew. Chem., Int. Ed. , 2002 , 41 , 783–786 ); (b) M. Scheer, L. Gregoriades, J. Bai, M. Sierka, G. Brunklaus and H. Eckert, Chem.–Eur. J., 2005, 11, 2163–2169; (c) M. Scheer, L. J. Gregoriades, M. Zabel, M. Sierka, L. Zhang and H. Eckert, Eur. J. Inorg. Chem., 2007, 2775; (d) L. J. Gregoriades, G. Balázs, E. Brunner, C. Gröger, J. Wachter, M. Zabel and M. Scheer, Angew. Chem., 2007, 119, 6070 ( Angew. Chem., Int. Ed. , 2007 , 46 , 5966 ); (e) M. Scheer, L. J. Gregoriades, M. Zabel, J. Bai, I. Krossing, G. Brunklaus and H. Eckert, Chem.–Eur. J., 2008, 14, 282.
6. L. J. Gregoriades, B. K. Wegley, M. Sierka, E. Brunner, C. Gröger, E. V. Peresypkina, A. V. Virovets, M. Zabel and M. Scheer, Chem.–Asian J., 2009, 4, 1578–1587.
7. J. Bai, A. V. Virovets and M. Scheer, Angew. Chem., 2002, 114, 1808–1811 ( Angew. Chem., Int. Ed. , 2002 , 41 , 1737–1740 ).
8. M. Scheer, L. J. Gregoriades, A. V. Virovets, W. Kunz, R. Neueder and I. Krossing, Angew. Chem., 2006, 118, 5818 ( Angew. Chem., Int. Ed. , 2006 , 45 , 5689 ).
9. S. Welsch, L. J. Gregoriades, M. Sierka, M. Zabel, A. V. Virovets and M. Scheer, Angew. Chem., 2007, 119, 9483 ( Angew. Chem., Int. Ed. , 2007 , 46 , 9323 ).
10. (a) J. Bai, A. V. Virovets and M. Scheer, Science, 2003, 300, 781–783; (b) M. Scheer, J. Bai, B. P. Johnson, R. Merkle, A. V. Virovets and C. E. Anson, Eur. J. Inorg. Chem., 2005, 4023–4026; (c) M. Scheer, A. Schindler, J. Bai, B. P. Johnson, R. Merkle, R. Winter, A. V. Virovets, E. V. Peresypkina, V. A. Blatov, M. Sierka and H. Eckert, Chem.–Eur. J., 2010, 16, 2092–2107.
11. (a) M. Scheer, A. Schindler, C. Gröger, A. V. Virovets and E. V. Peresypkina, Angew. Chem., 2009, 121, 5148; (b) M. Scheer, A. Schindler, R. Merkle, B. P. Johnson, M. Linseis, R. Winter, C. E. Anson and A. V. Virovets, J. Am. Chem. Soc., 2007, 129, 13386.
12. B. P. Johnson, F. Dielmann, G. Balázs, M. Sierka and M. Scheer, Angew. Chem., 2006, 118, 2533 ( Angew. Chem., Int. Ed. , 2006 , 45 , 2473 ).
13. For examples of supramolecular aggregates based on [{CpMo(CO)₂}₂(μ,η²-Sb₂)], see: H. V. Ly, M. Parvez and R. Roesler, Inorg. Chem., 2006, 45, 345–351.
14. (a) B. M. Cossairt, M.-C. Diawara and C. C. Cummins, Science, 2009, 323, 602; (b) H. A. Spinney, N. A. Piro and C. C. Cummins, J. Am. Chem. Soc., 2009, 131, 16233–16243; (c) B. M. Cossairt and C. C. Cummins, J. Am. Chem. Soc., 2009, 131, 15501–15511; (d) J. J. Curley, N. A. Piro and C. C. Cummins, Inorg. Chem., 2009, 48, 9599–9601.
15. Review articles: (a) O. J. Scherer, Acc. Chem. Res., 1999, 32, 751–762; (b) K. H. Whitmire, Adv. Organomet. Chem., 1998, 42, 1–145; (c) O. J. Scherer, Angew. Chem., 1990, 102, 1137–1155 ( Angew. Chem., Int. Ed. Engl. , 1990 , 29 , 1104–1122 ); (d) M. Scheer and E. Herrmann, Z. Chem., 1990, 30, 41–55.
16. I. Bernal, H. Brunner, W. Meier, H. Pfisterer, J. Wachter and M. L. Ziegler, Angew. Chem., 1984, 96, 428–429 ( Angew. Chem., Int. Ed. Engl. , 1984 , 23 , 438–439 ).
17. L. J. Gregoriades, H. Krauss, J. Wachter, A. V. Virovets, M. Sierka and M. Scheer, Angew. Chem., 2006, 118, 4295–4298 ( Angew. Chem., Int. Ed. , 2006 , 45 , 4189–4192 ).
18. O. J. Scherer, C. Blath and G. Wolmershäuser, J. Organomet. Chem., 1990, 387, C21–C24.
19. (a) A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652; (b) C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter, 1988, 37, 785–789.
20. Review article: R. Peng, M. Li and D. Li, Coord. Chem. Rev., 2010, 254, 1–18.
21. A. Biegerl, E. Brunner, C. Gröger, M. Scheer, J. Wachter and M. Zabel, Chem.–Eur. J., 2007, 13, 9270–9276.
22. A. Schindler, M. Zabel, J. F. Nixon and M. Scheer, Z. Naturforsch., B: Chem. Sci., 2009, 64, 1429–1437.
23. W. S. Sheldrick and T. Häusler, Z. Anorg. Allg. Chem., 1994, 620, 334–342.
24. P. Pyykkö, Chem. Rev., 1997, 97, 597–636; H. L. Hermann, G. Boche and P. Schwerdtfeger, Chem.–Eur. J., 2001, 7, 5333–5342.
25. P. C. Ford, E. Cariati and J. Bourassa, Chem. Rev., 1999, 99, 3625–3647.
26. N. P. Rath, E. M. Holt and K. Tanimura, Inorg. Chem., 1985, 24, 3934–3938.
27. J. Besinger, J. Treptow and D. Fenske, Z. Anorg. Allg. Chem., 2002, 628, 512–515.
28. A. Bondi, J. Phys. Chem., 1964, 68, 441–451.
29. (a) P. M. Graham, R. D. Pike, M. Sabat, R. D. Bailey and W. T. Pennington, Inorg. Chem., 2000, 39, 5121–5132; (b) D. M. Haddleton, D. J. Duncalf, A. J. Clark, M. C. Crossman and D. Kukulj, New J. Chem., 1998, 22, 315–317.

---

Footnotes

† Electronic supplementary information (ESI) available: Experimental, analytical, crystallographic and computational details. CCDC reference numbers 772393–772396. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c0sc00254b

‡ Crystal data for 2: C₁₂H₁₈As₅Cl₃Cu₃FeN, M = 903.72, monoclinic, space group P2₁/c (no. 14), a = 11.660(3) Å, b = 21.626(6) Å, c = 8.490(2) Å, β = 90.00(2)°, V = 2140.8(10) ų, Z = 4, μ = 11.647 mm⁻¹, F(000) = 1704, T = 123(1) K, 24053 reflections measured, 6531 unique (R_int = 0.0929), R₁ = 0.0427, wR₂ = 0.1015 for I > 2σ(I). CCDC 772393. Crystal data for 3: C₁₂H₁₈As₅Br₃Cu₃FeN, M = 1037.07, monoclinic, space group P2₁/c (no. 14), a = 11.8229(4) Å, b = 21.7908(8) Å, c = 8.6497(2) Å, β = 90.118(2)°, V = 2228.42(12) ų, Z = 4, μ = 22.890 mm⁻¹, F(000) = 1920, T = 123(1) K, 15185 reflections measured, 3774 unique (R_int = 0.0816), R₁ = 0.0381, wR₂ = 0.0840 for I > 2σ(I). CCDC 772394. Crystal data for 4: C₁₀H₁₅As₅Cu₂FeI₂, M = 946.57, triclinic, space group P1̄ (no. 2), a = 8.4256(6) Å, b = 10.9119(8) Å, c = 11.4820(9) Å, α = 71.256(7)°, β = 81.825(6)°, γ = 78.273(6)°, V = 975.42(13) ų, Z = 2, μ = 42.587 mm⁻¹, F(000) = 860, T = 123(1) K, 4805 reflections measured, 2111 unique (R_int = 0.0375), R₁ = 0.0337, wR₂ = 0.0711 for I > 2σ(I). CCDC 772395. Crystal data for 5·nCH₂Cl₂: C₁₁H₁₇As₅Cl₂Cu₂FeI₂, M = 1031.50, orthorhombic, space group Pnma (no. 62), a = 15.6513(1) Å, b = 11.7371(3) Å, c = 24.5300(5) Å, V = 4506.18(16) ų, Z = 8, μ = 39.105 mm⁻¹, F(000) = 3776, T = 123(1) K, 10949 reflections measured, 3650 unique (R_int = 0.0367), R₁ = 0.0322, wR₂ = 0.0741 for I > 2σ(I). CCDC 772396.

---