The tetrahedral clusters Fe3(CO)10(mu3-PR) (1) react on thermal as well as photochemical activation with Ph2PC=CPh to give Fe3(CO)8(mu3-PR)[mu3-eta4-Ph2P-CC-Ph(CO)] (3). In 3 an edge-opened iron triangle bears a mu3-PR group on one side and the alkynylphosphane on the other side as bridging moieties. One of the terminal CO groups present in 1 in addition to being coordinated to the appropriate iron atom, is bonded to the phenylsubstituted carbon atom of the alkynylphosphane.Irradiation of the square-pyramidal cluster Fe3(CO)9(mu3-P(t)Bu)(mu3-Se) (2) in the presence of Ph2PC=CPh leads to the compounds Fe3(CO)7(Ph2PC=CPh)2(mu3-P(t)Bu)(mu3-Se) (4) and Fe3(CO)6(Ph2PC=CPh)3(mu3-P(t)Bu)(mu3-Se) (5). In these clusters the square-pyramidal framework of 2 is preserved. In each case the alkynylphosphane ligands are coordinated via phosphorus only. In 4 one of these ligands is bonded to each of the two equatorial iron centres; in 5 each of the three iron atoms bears an alkynylphosphane ligand.The results are supported by conventional spectroscopic and analytical data as well as by single-crystal analyses.

The clusters (mu-3-PR)Fe3(CO)10 produce adducts with chalcogenocyanates XCN- (X = S, Se, Te) which, when alkylated with Et3O+, lead to the carbonyl clusters (mu-3-PR)(mu-3-X)Fe3(CO)9 (1) or their isonitrile substitution products (mu-3-PR)(mu-3-X)Fe3(CO)8(CNEt) (2), respectively. One or two CO groups of the clusters 1 can be substituted by trimethylphosphite (TMP) by means of thermal activation. In the case of the monosubstituted products (mu-3-PR)(mu-3-X)Fe3(CO)8(TMP) (3), the isomer which crystallizes is that in which one of the CO-groups at the apical iron of the square-pyramidal cluster 1 is substituted. In solution a slow isomerisation (on the NMR timescale) occurs to give substitution products in which the TMP ligand is bound to an equatorial iron. Only two of the three geometrically different ligand positions at the equatorial iron centre are markedly populated. The rapid exchange of the TMP ligand between these two positions is frozen out at 210 K. Activation barriers of about 50 kJ mol-1 for this isomerisation process can be calculated from the coalescence temperature of the P-31 NMR resonances.For the disubstituted products 4 a solid state structure is observed in which one TMP ligand is bound to the apical iron, and the second occupies an axial position at an equatorial iron. The P-31 NMR spectra of solutions of 4 demonstrate a dynamic equilibrium between all substitution patterns, in which both TMP ligands are bound separately to an iron centre of the cluster. Exchange between the positions at the apical iron and those at the equatorial iron centres appears to be slow on the NMR timescale. The exchange process between positions at the equatorial iron centres takes place rapidly and is frozen out at 200 K. Activation barriers of about 60 kJ mol-1 for the corresponding exchange phenomena can be deduced from the coalescence temperatures.The compounds 1 and all their substitution derivatives 2-4 show two reversible reductions in the cyclovoltammogram. The observed sequence of the reduction potentials corresponds to the nature of the substituents. The compounds 4 which are most difficult to reduce also show a reversible oxidation peak.The results reported herein are documented by the conventional spectroscopic and analytical data as well as by seven X-ray diffraction studies.

In clusters, in which an alkyne RCCR' is added to an (mu-3-RP)Fe3(CO)9 fragment, normally both carbon atoms of the alkyne group are engaged in interaction with the metal constituents. We have now found that with aminoalkynes RCCN(Et)2 the amino-substituted carbon atom of the alkyne fragment normally eludes this complete integration into the cluster by formation of a C = N double bond. These observations may be reduced to the statement that an amino group connected to a skeletal carbon atom of a cluster increases the effective number of skeletal electrons in such a way, that the observed structures correspond to the ones, which, on the basis of Wade's rules, should ensue from an increased number of skeletal electrons. Exceptions from this statement are found if potentially conjugative groups, i.e. CO, are connected to the other carbon atom of the aminoalkyne constituent.These general propositions are documented here for two isomeric clusters (iPrP)Fe3(CO)9(MeCCNEt2) (2 and 3) and one cluster with the composition (iPrP)Fe3(CO)10(MeCCNEt2), (4) by spectroscopic and X-ray structural data.

The two isostructural compounds RPFe3(CO)9(R′CCN(Et)2), 2 and 3, as well as compound 4, RPFe3(CO)10(R′CCN(Et2), are obtained by reaction of (μ3-RP)Fe3(CO)10, 1, with aminoalkynes. The N(Et)2-substituents donate their lone pair of electrons into the cluster framework of 2-4. The observed cluster geometry of 2-4 reflects this increased number of skeletal electrons. The structures of 2a, 3a and 4b have been determined by X-ray diffraction studies. All compounds have been characterized by the usual spectroscopic and analytical methods.

The clusters (mu-3-RP)Fe3(CO)10, 1, behave like coordinatively unsaturated compounds in many reactions; they add two electron donor ligands L via opening of one Fe-Fe edge. This reaction pattern presumably represents the initiation of a process, in which these clusters split thiocompounds R2C = S into sulfur and carbene ligands, both of which are integrated into the resulting product cluster. With thiourea, N,N'-dimethylthiourea and thioacetamide the compounds 1 react to give the clusters eta-4-[R activated P...Fe(CO)3...S...Fe(CO)2(CR' R")]Fe(CO)3, 2 and 3, in fair yields. The structure of the carbene-substituted clusters 2 and 3, the square pyramidal framework of which may be regarded as an eta-4-coordinated 4-pi-metallaheterocycle, is documented by two X-ray analyses. The dynamic behaviour of the clusters 2 and 3 is analysed by NMR-spectroscopy: the strongly hindered rotation of the amino groups around the C(carbene)-N-bond is characteristic in all cases.

While attempts to stack the nido clusters (mu-3-RP)(mu-3-S)Fe3(CO)9, 2, in a direct way to the closo compounds (mu-4-RP)(mu-4-S)Fe4(CO)11, 4, have been unsuccessful so far, the carbene substituted derivatives of 2 (mu-3-RP)(mu-3-S)Fe3(CO)8(CRR'), 1, react with Fe2(CO)9 to give the closo clusters 4 and their carbene derivatives (mu-4-RP)(mu-4-S)Fe4(CO)10(CRR'), 3, in fair yields. The carbene ligand of 1 obviously makes it considerably easier to introduce an Fe(CO)2 capping group.Cyclovoltammetric analysis of the nido/closo pairs of compounds 1/3 and 2/4 respectively demonstrates that one electron reduction of the nido complexes requests a potential which is about 0.5 V more negative than the potential necessary for the corresponding reduction of the respective closo compounds. The observed sequence of reduction potentials corresponds to the predictions which may be derived from a model in which nido and closo compounds are considered as derivatives of organometallic pi-systems.All results are documented by the conventional analytical and spectroscopic techniques and in addition by two X-ray structure analyses.

The halfsandwich-type compounds eta-4-[RPactivated...Fe(CO)3...X...Fe(CO)3]Fe(CO)3, (X = Se, Te) 1a, 1b (type A), may be stacked to the tripeldecker-type compounds mu-2-eta-4-[RPactivated...Fe4-(CO)3...X...Fe(CO)3]Fe2(CO)5, 2a (X = Se), 2b (X = Te) (X-ray analysis) (type B). The relation between the compounds 1 and 2 is the same as the one between the nido pentagonal pyramidal clusters eta-5-[RPactivated...Fe(CO)3...R'C...R''C...Fe(CO)3]Fe(CO)3, 3 (type C) and their closo derivatives mu-2-eta-5-[RPactivated...Fe-(CO)3...R'C...R''C...Fe(CO)3]Fe2(CO)5, 4 (type D). The analysis of a specific dynamic behaviour of compounds 3 is reported. In addition a model is proposed, which explains the observed pairwise relation in the redox potentials of structurally related nido and closo compounds in a simple manner.

The compound (1a) may be described as an Fe(CO)3 derivative of a 1,2-ferrathiacyclobutadiene ligand. The CO groups of 1a, which are very difficult to substitute by thermal activation, are readily exchanged when compared with P(OMe)3, by reductive electrocatalysis. Reductive initiation of substitution is also possible by the reduction of 1a with Na/Hg or t-BuLi. The mechanism of the reductively-initiated formation of the products 1b and 1c, in which one or two carbonyl groups, respectively, of 1a are replaced by P(OMe)3 as well as the formation of a monosubstituted derivative of 1a, in which an axial CO group of the heterometallacyclobutadiene ligand is replaced by tBuP(OMe)2 (1d) is discussed with respect to electrochemical, spectroscopic and structural data. tBuLi can activate 1a not only by single electron transfer: the addition of tBu− at the sulfur center of 1a leads to an anionic CO-insertion product, which, by subsequent alkylation, may be trapped as (2b). Insertion of CO in 1a is also initiated by Na/Hg reduction; oxidative work up with I2 gives (3). Thermal activation readily decarbonylates 3 to give 1a. The individual results are consistently described in a common reaction scheme. The structures of 1d, 2 and 3 have been determined by X-ray diffraction studies.