摘要:

DALTON FULL PAPER J. Chem. Soc. Dalton Trans. 1999 1213–1220 1213 The mechanism of displacement of dihydrogen and dinitrogen from iron ruthenium and osmium hydrides and implications for models of nitrogenase action Caroline A. Helleren,a Richard A. Henderson b and G. JeVery Leigh *a a School of Chemistry Physics and Environmental Science University of Sussex Brighton UK BN1 9QJ b Nitrogen Fixation Laboratory John Innes Centre Norwich Research Park Colney UK NR4 7UH Received 24th December 1998 Accepted 1st March 1999 The substitution of dihydrogen in complexes [FeH(H2)(phosphine)x]1 [phosphine = R2PCH2CH2PR2 (R = Et or Me) or P(CH2CH2PR92)3 (R9 = Me or Ph)] by ligands L (MeCN PhCN or Cl2) has been shown to be first order in the concentration of complex and zero order in the concentration of L in both acetone and thf.Activation parameters have been determined and the mechanism of substitution is proposed to involve rate-determining loss of H2 from the parent complexes and subsequent rapid co-ordination of L. This mechanism diVers from that recently proposed for an analogous complex of Ph2PCH2CH2PPh2 and the reasons for this are discussed. Less thorough studies of some related dinitrogen complexes and of some homologous complexes of Ru and Os are consistent with a similar loss of dinitrogen or dihydrogen being rate determining. Introduction There is considerable circumstantial evidence that the dinitrogen-binding site of the molybdenum–iron nitrogenases is hydridic during at least part of the catalytic cycle.1 In particular the fixation of dinitrogen involves the obligatory evolution of at least one molecule of dihydrogen for each molecule of dinitrogen fixed and dihydrogen is a competitive inhibitor of nitrogen fixation.Both features can be explained by dinitrogen binding involving displacement of dihydrogen.2 In a chemical context of models for dinitrogen binding it has been recognised for some time that iron trihydrides such as trans-[FeH(H2)(dmpe)2]1 (dmpe = Me2PCH2CH2PMe2) react with several donors (L) such as CO N2 and MeCN losing H2 and forming trans-[FeH(L)(dmpe)2]1.3 The mechanism of these displacements has not been definitively determined. In order to extend this type of reaction we originally attempted 4 to prepare complexes containing a dinitrogen molecule bridging between iron and molybdenum by a reaction of trans-[FeH(H2)(dmpe)2]1 with trans-[Mo(N2)2(dppe)2](dppe = Ph2PCH2CH2PPh2) but all we were able to isolate was a mixture containing trans-[MoH4(dppe)2] and trans-[FeH(N2)- (dmpe)2]1 although it is now evident 5 that there are also further products.A preliminary study 6 suggested that the rate of this surprising exchange exhibited a first-order dependence on the concentration of the iron complex and this prompted us to pursue the work reported here on the kinetics of substitution reactions of a series of metal hydrides as generalised in the following equations (M = Fe Ru or Os; pp = dmpe or Et2PCH2CH2PEt2; depe or less often tetraphosphines; x = 1 or 2; n = 1 or 0; and L = MeCN PhCN or Cl2). trans-[MH(X2)(pp)x]1 1 L æÆ trans-[MH(L)(pp)x]n1 1 X2 Such studies should enable us to determine the influences of substituent stereochemistry incoming nucleophile and solvent on the reaction.When this work had been completed some parallel studies on the reactions of trans-[FeH(H2)(dppe)2]1 with nitriles were described.7 We shall include these published data in our discussions and show how a spectrum of mechanisms can operate within the family of complexes of general formula trans-[FeH(H)2(pp)2]1 (pp = dmpe depe or dppe). Preliminary work in our group4,6 has indicated that the reactions we are interested in are first order in the concentration of complex. Thus in the substitution of dihydrogen by isocyanides and nitriles in trans-[FeH(H2)(pp)2][BPh4] (pp = dmpe or depe) 3,4,6 the rate of reaction exhibits a first-order dependence on the concentration of complex and is independent of the concentration of nitrile.The reaction of trans-[FeH(H2)(dmpe)2]- [BPh4] with MeCN in thf under argon was studied by following the intensity of the band assigned to the n(N]] ] C) of the product in the IR spectrum and the first-order rate constant kobs was found to be (1.6 ± 0.2) × 1023 s21 essentially identical to that for the formation of trans-[FeH(N2)(dmpe)2][BPh4] from trans-[FeH(H2)(dmpe)2]1 (1.2 ± 0.1) × 1023 s21 determined previously.3,6 This measurement was performed at only one temperature that of the IR beam estimated to be 325 K. The reaction of trans-[FeH(H2)(depe)2][BPh4] with MeCN eqn. (3) trans-[FeH(H2)(depe)2]1 1 MeCN thf trans-[FeH(MeCN)(depe)2]1 1 H2 (3) was studied 6 in thf by following the intensities of the 1214 J. Chem. Soc. Dalton Trans. 1999 1213–1220 resonances of the starting material and product in their 31P- {1H} NMR spectra.Under pseudo-first-order conditions a rate constant kobs = (3.2 ± 0.2) × 1024 s21 at 296 K was determined. In some related work8 the kinetics of the reactions of trans- [FeH(N2)(depe)2]1 with several nucleophiles at 298 K has been studied. The reactions exhibit a first-order dependence on the concentration of trans-[FeH(N2)(depe)2]1 with kav = 1.0 × 1023 s21 for the reactions with CO MeCN and PhCN. The present study is one of the first comprehensive quantitative studies to show the eVect of changing the metal ancillary ligands and nucleophile on the kinetics and mechanisms for the substitution of H2 or N2 in octahedral Group 8 phosphine complexes. We have also extended the studies to a limited number of homologous hydrides of ruthenium and osmium (Table 1).However we were unable to complete this study to include all complexes trans- [MH(X2)(pp)2][BPh4] and cis-[MH(X2){P(CH2CH2PR2)3}]- [BPh4] (M = Fe Ru or Os; X = H or N; pp = dmpe or depe; R = Me [pp3Me] or Ph [pp3]) because some of them are too kinetically stable or unstable in solution whilst others have yet to be reported. Experimental All manipulations were carried out using standard Schlenk techniques under argon unless otherwise stated. House dinitrogen was dried with potassium hydroxide and silica gel before use. Otherwise pure dinitrogen (Air Products) was used directly from the cylinder. Solvents were dried by heating them to reflux over an appropriate drying agent under dinitrogen. Absolute ethanol was used as supplied and analytical grade acetone was dried over a succession of molecular sieves.The volatile solvents were degassed by freeze-thawing. Non-volatile solvents were purged of dioxygen by bubbling argon or dinitrogen through them. The preparations of complexes [FeH(H2)(dmpe)2][BPh4],3,5,9 [FeH(N2)(dmpe)2][BPh4],3,9,10 [FeH(H2)(depe)2][BPh4],3,5,9 [FeH- (N2)(depe)2][BPh4],3,9,10 [FeH(H2)(pp3)][BPh4],11 [FeH(N2)- (pp3)][BPh4],11 [FeH(H2)(pp3Me)][BPh4],12 [FeH(N2)(pp3Me)]- [BPh4],12 [RuH(H2)(depe)2][BPh4],9a,10,13 [RuH(N2)(depe)2]- [BPh4],9a,10,13 and their ruthenium dmpe homologues [OsH- (H2)(depe)2][BPh4],3,5,6,9,14,15 [OsH(N2)(depe)2][BPh4],3,5,6,9,14,15 and the osmium dmpe homologues and their derivatives [MH(L)(phosphine)n][BPh4] (M = Fe Ru or Os; L = MeCN Table 1 Summary of systems studied Group 8 complex trans-[FeH(N2)(dmpe)2][BPh4] trans-[FeH(X2)(depe)2][BPh4] trans-[RuH(H2)(depe)2][BPh4] trans-[OsH(H2)(dmpe)2][BPh4] Nucleophile MeCN PhCN and chloride MeCN PhCN and chloride MeCN MeCN and PhCN PhCN or Cl as detailed in the text; n = 1 or 2 as required by the stoichiometry; phosphine = dmpe depe pp3 or pp3Me)3,5,6,10 were attempted and most were characterised as described in the literature.Not all our attempts were successful and only some of these complexes (see text) were found to be amenable to our kinetic experimental techniques. Nuclear magnetic resonance (NMR) spectra were measured using either a Bruker ACP-250 (operating frequencies 1H 250.2 MHz 31P 101.3 MHz) or a Bruker DPX-300 spectrometer (operating frequencies 1H 300.1 MHz and 31P 121.5 MHz) with 1H referenced against the deuteriated solvent and 85% H3PO4 in D2O (d 0) used as the external 31P reference.The NMR solvents were used as supplied from Cambridge Isotope Laboratory after they were transferred into dried Schlenk tubes and kept under argon. The decoupler frequency in 31P-{1H} NMR spectroscopy was centred on the metal hydride resonances which exhibit the strongest P–H coupling as determined from the 1H NMR spectra. The UV-vis spectra were obtained on a UV-2101 PC scanning spectrophotometer and stopped-flow measurements were carried out using a Hi-Tech SF-51 spectrophotometer modified to enable manipulation of air-sensitive solutions. The temperature was maintained at 25 8C using a Grant LE8 thermostat bath and the spectrophotometer was interfaced to a computer via an A/D converter.Data were transferred directly to the computer and analysed by a computer program which fitted the exponential absorbance–time curves by use of single exponential functions. Methods of kinetic measurement (a) Iron and osmium dihydrogen and/or dinitrogen complexes. The intensity of the phosphorus resonances for the starting material and product was monitored by 31P-{1H} NMR spectroscopy at 101.3 or 121.5 MHz on spectrometers with variabletemperature facilities. An appropriate weight of trans-[MH(X2)(pp)2][BPh4] (M = Fe or Os; X = N or H) (ca. 0.1 mmol) was dissolved in either 0.5 cm3 of a mixture of deuteriated and undeuteriated solvent (1 9 v/v) when the resonances were measured with a base frequency Table 3 Rate constants for the reactions of trans-[FeH(X2)(pp)2]- [BPh4] in the concentration range 100–200 mmol dm23 with [nBu4N]Cl (7–10-fold excess) in acetone at 298.2 K kobs/s21 for Complex [FeH(H2)(depe)2]1 [FeH(N2)(depe)2]1 [FeH(N2)(dmpe)2]1 Loss of X2 from [FeH(X2)(pp)2]1 (3.5 ± 0.1) × 1024 (9.7 ± 0.1) × 1024 (2.1 ± 0.1) × 1024 Formation of [FeH(Cl)(pp)2] (5.8 ± 0.5) × 1024 (11.1 ± 0.7) × 1024 Table 2 Rate constants for the reactions of trans-[MH(X2)(pp)2][BPh4] (in the concentration ranges Fe 100–200 Ru ca.0.2 Os 150–200 mmol dm23) with nitriles at 298.2 K kobs/s21 for Complex [FeH(H2)(depe)2]1 [FeH(N2)(depe)2]1 [FeH(N2)(dmpe)2]1 [RuH(H2)(depe)2]1 [OsH(H2)(depe)2]1 Solvent Acetone thf Acetone thf Acetone thf Acetone thf Loss of X2 from [MH(X2)(pp)2]1 (4.0 ± 0.3) × 1024 (2.7 ± 0.2) × 1024 (12.5 ± 0.8) × 1024 (10.6 ± 0.6) × 1024 (2.5 ± 0.2) × 1024 (1.7 ± 0.3) × 1024 (2.8 ± 0.2) × 1021 (3.5 ± 0.2) × 1025 a Formation of [MH(L)(pp)2]1 (4.5 ± 1.6) × 1024 (2.5 ± 0.3) × 1024 (12.6 ± 1.8) × 1024 (9.5 ± 1.6) × 1024 (2.7 ± 0.5) × 1024 (2.1 ± 0.8) × 1024 a Taken at 323.2 K.J. Chem. Soc. Dalton Trans. 1999 1213–1220 1215 of 121.5 MHz (5 mm NMR tube) or 2 cm3 of undeuteriated solvent when the spectra were measured with a base frequency of 101.3 MHz (10 mm tube with an insert containing D2O as the lock solvent). The predried solvent was degassed prior to use and the reactions were carried out under argon to prevent side reactions of the dihydrogen complexes with dinitrogen. It was assumed (reasonably) that the small percentage of deuteriated solvent would not significantly aVect the rate constant of the reaction that would otherwise occur in the undeuteriated solvent.The solution of starting material was transferred to an NMR tube and capped with an appropriately sized Subaseal. We assumed that the production of dihydrogen or dinitrogen would not exert undue pressure on the Subaseal of the argon-filled NMR tube. The sample was equilibrated in the NMR probe to the required temperature for 20 to 30 min prior to the addition of nucleophile. The initial spectrum of starting material was taken with a number p of scans p depending on the concentration of the complex. The tube was removed from the probe an excess of nucleophile L was quickly added to the solution via a syringe and the tube was then shaken and replaced in the probe. The same number p of scans was taken after set periods of time until the reaction was more than 75% complete.Solutions of the starting materials trans-[MH(X2)(pp)2]- [BPh4] had to be essentially stable in solution under argon to irreversible loss of X2 in the absence of a nucleophile for the period of the experiment. They were placed in NMR tubes in a thermostatically controlled probe and spectra taken periodically. No concentration of any starting complex changed detectably over a period of up to 12 h. (b) Ruthenium dihydrogen complex. The reactions of trans- [RuH(H2)(depe)2][BPh4] with nucleophiles are too rapid to be followed by the NMR spectroscopic methods outlined above. Therefore the reaction with MeCN was followed using stopped- flow spectrophotometry. At l = 420 nm the reaction is characterised by a single exponential absorbance–time curve with an initial absorbance corresponding to that of trans-[RuH(H2)- (depe)2]1 and a final absorbance which is that of trans- [RuH(NCMe)(depe)2]1.Results Determination of rate constants The kinetics of all the reactions studied was determined in the presence of a suYcient excess of nucleophile to ensure pseudo- first-order conditions. In the first instance (see later) the data were analysed in the usual way by plotting loge{[M(X2)- Ln5]t 2 [M(X2)Ln5]0} against t where [M(X2)Ln5]t is the concentration of complex at time t [M(X2)Ln5]0 is its concentration at the beginning of the reaction and Ln5 represents the remaining non-reactive ligands in the metal coordination sphere. The observed rate constant kobs is the slope of this straight line. The kinetic data for the reaction between trans-[RuH- (H2)(depe)2]1 and MeCN determined on the stopped-flow apparatus were analysed by a computer curve-fitting program.The curve was a good fit to a single exponential for at least three half-lives. Systematic variation of the concentration of the nucleophile (in the range 1.1- to 15-fold excess) led to no appreciable change in kobs demonstrating that the rate is independent of the concentration of the nucleophile. The average rate constants for the various complexes reacting in diVerent solvents with MeCN or PhCN are summarised in Tables 2 and 3. The data for the iron and osmium complexes derived from the NMR spectroscopic studies show rate constants obtained by analysing both the disappearance of the starting material and the appearance of the product. The latter type of study presented some problems because the value of kobs determined in this way was sometimes appreciably diVerent from that determined by monitoring the disappearance of reactant.This is because the semi-logarithmic plot will only give accurate values reliably if the concentration of product at the end of the reaction is known reliably. We estimated [M(X2)Ln5]e the concentration at the end of the reaction by extrapolating the data on the basis of an exponential increase in the intensity of the resonance in the 31P-{1H} NMR spectrum. However even small changes in the estimated value resulted in significantly diVerent values of kobs. To avoid this problem we used both the Guggenheim16 and Kezdy–Swinbourne17 methods of analysis that do not require the value of [M(X2)Ln5]e. Unless data analysis was performed in this way we could not obtain accurate values of kobs.It has been claimed that both the Guggenheim 16 and Kezdy– Swinbourne17 methods of analysis have similar orders of accuracy. Both give linear plots for first-order reactions. There is an evident advantage in the Guggenheim method when the order of reaction is not known since this method relies principally on data obtained at the end of the reaction where there is a clear distinction between exponential (for reactions exhibiting a first-order dependence upon complex concentration) and hyperbolic (for reactions exhibiting a second-order dependence upon complex concentration) curves. However in our case this distinction is not essential because the first-order dependence on the concentration of complex has been established by analysis of other data (see below).The average values of kobs obtained by the Kezdy–Swinbourne17 method of analysis are presented in Tables 2–4. Order of reaction in complex The linearity of the semi-logarithmic plots described above is consistent with the reactions exhibiting a first-order dependence on the concentration of complex. This was confirmed by experiments in which the concentration of the complexes was varied whilst maintaining a constant concentration of nucleophile. Under these conditions kobs did not change. Specimen data are shown in Tables 4 and 5. Temperature dependence of rate constants The reactions of the iron complexes were investigated at various temperatures to determine activation parameters for the Table 4 Rate constants for the reaction of trans-[FeH(H2)(depe)2]- [BPh4] in the concentration range 40–150 mmol dm23 with MeCN in acetone at 298.2 K 104 kobs/s21 for Ratio Fe :MeCN or weight of iron complex (g) 1 1.1 0.015 0.029 0.050 Loss of H2 from [FeH(H2)(depe)2]1 3.9 ± 0.1 4.3 ± 0.1 4.3 ± 0.1 4.2 ± 0.1 Formation of [FeH(MeCN)(depe)2]1 4.0 ± 0.3 4.2 ± 0.3 4.0 ± 0.3 4.1 ± 0.3 Table 5 Rate constants for the reaction of trans-[RuH(H2)(depe)2]- [BPh4] at the concentration ca.0.2 mmol dm23 with MeCN in excess in acetone at 298.2 K Ratio of Ru to MeCN 1:50 1:25 1 12.5 1 6.3 1 6.3 kobs/s21 for loss of H2 from [RuH(H2)(depe)2]1 0.260 0.285 0.285 0.285 0.275 1216 J. Chem. Soc. Dalton Trans. 1999 1213–1220 Table 6 Activation parameters for the reactions of trans-[FeH(X2)(pp)2][BPh4] with nitriles in the temperature range 291–313 K Complex [FeH(H2)(depe)2]1 [FeH(N2)(depe)2]1 [FeH(N2)(dmpe)2]1 Solvent Acetone thf Acetone thf Acetone thf Eact/kJ mol21 115.1 ± 4.7 124.3 ± 4.2 109.3 ± 4.7 118.6 ± 3.4 123.5 ± 3.7 125.0 ± 1.4 DH‡/kJ mol21 112.4 ± 4.7 121.7 ± 4.1 107.2 ± 4.1 115.9 ± 3.4 120.8 ± 3.7 122.3 ± 1.4 DS‡/J K 21 mol21 48 ± 15 77 ± 14 40 ± 14 68 ± 12 73 ± 13 76 ± 5 DG‡ 298.2 K/kJ mol21 98.1 ± 0.2 98.9 ± 0.1 95.2 ± 0.1 95.6 ± 0.1 99.2 ± 0.2 99.8 ± 0.1 Table 7 Summary of the trends in rate constants and activation parameters for the reactions of trans-[FeH(X2)(pp)2][BPh4] with nucleophiles Factor Metal X2 Ancillary ligand Nucleophile Solvent Stereochemistry Temperature kobs Ru > Fe@Os N2 > H2 depe > dmpe @ pp3 Independent of nucleophile Acetone > thf trans @ cis Increases with increasing temperature Eact DH‡ DS‡ N2 < H2 dmpe > depe Independent of nucleophile thf > acetone DG‡ N2 < H2 N2 < H2 Independent of nucleophile thf > acetone dissociation of X2 (X = H or N) from trans-[FeH(X2)(depe)2]1.The complexes trans-[RuH(H2)(depe)2]1 and trans-[OsH(H2)- (depe)2]1 were studied only at 298.2 and 323.2 K respectively. Other structurally similar diphosphine or tetradentate tetraphosphine iron osmium and ruthenium compounds were not amenable to study since they react either too rapidly or too slowly for an extensive temperature-dependence study to be possible. Plots of loge(kobs) against 1000/T give slopes of 2Eact/R from which the activation energies Eact were calculated (Table 6). The data are independent of the identity of the nitrile and are averaged. The errors reported are those for the data points about the line as given by the least squares analysis.The temperature range employed (291–313 K) in these studies depended on the rate of reaction the fastest reaction studied being complete within ca. 10 min. The activation parameters DH‡ and DS‡ were calculated using the Eyring equation and the usual thermodynamic relationships.17,18 A summary showing the main trends in the rate constants and activation parameters is given in Table 7. Discussion The mechanism of substitution of trans-[MH(X2)(pp)2][BPh4] (X 5 H or N) In the studies with dmpe and depe complexes reported in this paper we have shown that (i) the rate constants for the loss of trans-[MH(X2)(pp)2][BPh4] are equal to those for formation of trans-[MH(L)(pp)2][BPh4] within experimental error and (ii) that the rates of reaction are strictly independent of the nature and concentration of nucleophile in both acetone and thf.Point (i) is consistent with the simple stoichiometries observed for these reactions and point (ii) dictates a mechanism involving a rate-limiting unimolecular reaction of trans-[MH(X2)(pp)2]- [BPh4]. The possible mechanisms are discussed below after an account of some solution properties of complexes trans- [MH(H2)(pp)2]1. At low temperatures in solution complexes trans-[MH(H2)- (pp)2]1 show separate resonances for the hydride and dihydrogen ligands in the 1H NMR spectra. As the temperature increases the hydride and dihydrogen atoms undergo intramolecular exchange,3,9a the rate of which increases with temperature. At the fast-exchange limit only a single broad resonance is observed in the 1H NMR spectra of trans- [MH(H2)(pp)2]1.This eVectively scrambles all the hydrogen atoms. The question then arises as to which of the tautomers is the reactive species. We studied the reactions of trans-[FeH(H2)(pp)2][BPh4] with nucleophiles at temperatures where this intramolecular exchange is rapid and we assumed that the loss of dihydrogen occurred via [MH(h2-H2)(pp)2]1 derived from [MH3(pp)2]1. However intramolecular hydrogen-atom exchange in trans- [RuH(H2)(depe)2][BPh4] at 298.2 K is slow and the loss of H2 upon reaction with a nitrile is almost certainly from the hydrido(dihydrogen) tautomer. Finally [OsH(H2)(depe)2][BPh4] exists as two tautomers in rapid temperature-dependent equilibrium. 9a,14,19,20 We again assumed that at 323.2 K the loss of dihydrogen was via the hydrido(dihydrogen) tautomer.Indeed it is diYcult to conceive of any reasonable alternative. The simplest (and most likely) mechanisms consistent with our data on the depe and dmpe complexes are the dissociative mechanism and the dissociative interchange mechanism.17,18,21–23 For a dissociative mechanism as shown in eqn. (4) rate-limiting [MH(X2)Lnm] k1 k21 [MHLnm] 1 X2 k2 L [MH(L)Lnm] (4) dissociation of X2 generates the co-ordinatively unsaturated intermediate [MHLnm]1 which is rapidly attacked by nucleophile (L) or solvent (solv) to form the product. Other work† has shown that in the case of osmium a five-co-ordinate species may actually be more stable than the dinitrogen adduct. The rate law associated with this mechanism is readily derived by treating [MHLnm] as a steady-state intermediate.The resulting expression is (5). If k2[L] > k-1[X2] (a condition that can be 2d[MH(X2)Lnm] dt = k1k2[L][MH(X2)Lnm] k21[X2] 1 k2[L] (5) fulfilled when there is a large excess of nucleophile present) eqn. (5) simplifies to (6). Eqn. (6) is consistent with our kinetic 2d[MH(X2)Lnm]/dt = k1[MH(X2)Lnm] (6) data since it dictates that the rate of reaction is independent of the concentration of nucleophile and that k1 is independent of the nature of the nucleophile. However in the dissociative interchange mechanism eqn. (7) [MH(X2)Lnm]1 1 L k3 k23 [MH(X2)Lnm]1?L k4 [MH(L)Lnm]1?X2 fast [MH(L)Lnm]1 1 X2 (7) † As reported in ref. 10 for example trans-[OsH(N2)(depe)2]1 loses N2 rapidly even in the solid state. J. Chem. Soc.Dalton Trans. 1999 1213–1220 1217 Fig. 1 The correlation of DS‡ and DH‡ for reactions of trans-[FeH(X2)(pp)2]1 with nucleophiles. outer-sphere association of the nucleophile with the complex occurs prior to dissociation of the leaving group. Upon dissociation of X2 the nucleophile present in the first solvation sphere is advantageously positioned to bind to the vacant site. Assuming that association of L is a rapidly established equilibrium (K3) and that k4 represents the rate-limiting dissociation of M–X2 the dissociative interchange mechanism gives the rate law (8). If K3[L] @ 1 eqn. (8) simplifies to (9) which is also 2d[MH(X2)Lnm] dt = k4K3[MH(X2)Lnm][L] 1 1 K3[L] (8) 2d[MH(X2)Lnm]/dt = k4[MH(X2)Lnm] (9) consistent with the observed kinetics. Although the form of this rate law is identical to that observed experimentally we consider this mechanism to be less likely than the dissociative mechanism for the following reasons.First k4 should be dependent to some degree on the nature of the nucleophile because of the presence of the nucleophile within the solvation sphere during this elementary reaction step but we see no appreciable variation of the value of kobs with a variety of nucleophiles. Secondly in order for the limiting rate law to operate in the form of eqn. (9) K3[L] @ 1 even at the lowest concentration of nucleophile employed (1.0 mmol dm23). This allows us to calculate a limit of K3 @ 1000 dm3 mol21. This is very tight binding for a neutral molecule such as MeCN to a monocation such as trans-[FeH(X2)(pp)2]1 and seems unlikely. Consistent with our proposed dissociative mechanism is the independence of DH‡ and DS‡ of the nature of L.Particularly significant are the large and positive values of DS‡ that are entirely consistent with the dissociative mechanism. We discuss further aspects of the DH‡ and DS‡ parameters in the next section. Comparison with the reaction mechanism proposed for [FeH(H2)- (dppe)2]1 Recently the kinetics of the reactions of trans-[FeH(H2)- (dppe)2][BPh4] with MeCN PhCN and Me2SO has been studied in acetone and thf.7 The results obtained diVer from ours in two key respects. First the rate of the reaction depends on the concentration and nature of the nucleophile. Secondly DH‡ = ca. 80 kJ mol21 and DS‡ = ca. 220 J K21 mol21. This has been interpreted as indicating the mechanism shown in eqns. (10)–(12) in which one arm of a dppe ligand dissociates [FeH(H2)(dppe)2]1 [FeH(H2)(dppe-k2P)(dppe-kP)]1 (10) [FeH(H2)(dppe-k2P)(dppe-kP)]1 1 L [FeH(H2)(L)(dppe-k2P)(dppe-kP)]1 (11) [FeH(H2)(L)(dppe-k2P)(dppe-kP)]1 æÆ [FeH(L)(dppe)2]1 1 H2 (12) and then solvent binds weakly to the vacant site.The ratelimiting step is proposed to be associative attack of the nucleophile on this intermediate. The activation volumes DV‡ for the reaction were found to be ca. 220 cm3 mol21 consistent with an associative mechanism. Analogous lability of diphosphine ligands on iron(II) sites has been noted before 8,24 and suggested as the pathway for substitution reactions of other iron complexes. Clearly our rate data on the dmpe and depe analogues are not consistent with this mechanism. This conclusion is supported by the temperature-dependence data for these reactions.Fig. 1 shows the correlation between DH‡ and DS‡ for the substitution reactions of the complexes trans-[FeH(X2)(pp)2]1 where pp = dmpe depe and dppe. The data points corresponding to the dmpe and depe complexes studied in this work cluster in the top right hand corner. The line presented is that defined by a least squares analysis of our data in Table 6 alone together with the additional restriction that the intercept at DS‡ = 0 is DH‡ = 98 kJ mol21 (the mean value of DG‡ observed for all the depe and dmpe complexes Table 6). Clearly our data on the dmpe and depe analogues are not consistent with the mechanism proposed for the dppe complexes. This conclusion is supported by the temperature-dependence studies. The enthalpies of activation and entropies of activation for the reactions of the iron complexes apparently compensate to give an almost invariant DG‡.When a reaction has a large value of DH‡ then it is generally assumed that the M–X2 bond dissociation energy is large. The tighter binding of X2 may 1218 J. Chem. Soc. Dalton Trans. 1999 1213–1220 mean that its mobility in the ground and excited states is restricted and therefore the associated value of DS‡ is small. A weak M–X2 bond should give rise to a larger value of DS‡. Since DG‡ = DH‡ 2 TDS‡ DH‡ and DS‡ then balance to give similar DG‡ values at the same temperature. There are no published data comparable to those for our iron complexes nor for the corresponding ruthenium or osmium compounds [MH(X2)(pp)2]1 so we cannot assess whether changing metal would aVect this generalisation.Observations of such compensatory eVects are common in many areas of chemistry both for kinetic and equilibrium data.25 The points in the bottom left hand corner of Fig. 1 correspond to the published data for trans-[FeH(H2)(dppe)2]1. Clearly these points do not lie on the line defined by the dmpe and depe complexes. This is a further indication that a diVerent mechanism operates for these analogues. The distinctly diVerent behaviour between the dmpe or depe and the dppe analogues immediately poses the question of which factors control the mechanisms of substitution reactions in this family of compounds. In reactions operating by a dissociative mechanism the primary controlling factor must be the metal– ligand bond strengths. We infer that the Fe–P Fe–H2 and Fe–N2 bond strengths are very similar in the family of complexes [FeH(X2)(pp)2]1 and that depending on the diphosphine either Fe–P (pp = dppe) or Fe–X2 cleavage (pp = depe or dmpe) can occur.The dppe is sterically quite demanding because of the bulky phenyl groups and this would favour dissociation of one arm of this ligand. In contrast dmpe and depe are sterically less demanding and better electron donors due to their alkyl groups. This would make their dissociation less likely. In addition the increased electron density on Fe in the dmpe and depe complexes would make the site more electron-rich facilitating Fe-to-X2 back bonding. Clearly changing the electronic and steric properties of the phosphine co-ligands by varying the substituents on the phosphorus atoms will perturb the labilities of both the H2 and phosphine ligands with the result that in the family of complexes trans-[FeH(X2)(pp)2]1 a change in mechanism of substitution may occur as the diphosphine ligand is changed.Factors aVecting the rate constants In this and the remaining sections of the discussion we shall restrict consideration to the dmpe and depe complexes that undergo substitution by the dissociative mechanism depicted in eqn. (4). The largest change in rate constants (Table 7) occurs when the metal is changed. For example trans-[RuH(H2)- (depe)2]1 reacts 700 times faster than trans-[FeH(H2)(depe)2]1 at 298.2 K. As trans-[OsH(H2)(depe)2]1 is unreactive at 298.2 K the reactions with nitriles were carried out at 323.2 K. These results show a very large decrease in rate constant when iron or ruthenium is replaced by osmium in structurally analogous complexes.The reactions involving trans-[OsH(H2)(depe)2]1 were studied only in thf because 323.2 K is too close to the boiling point of acetone for measurements to be made in that solvent. Our conclusion that the rate constants for the loss of H2 decrease in the sequence 4d @ 3d @ 5d is consistent with other studies of Group 8 complexes by Halpern et al. 26 (quantitative) by Amendola et al.27 and by Morris and coworkers 9a,13,20 (qualitative). For example Jessop and Morris 20 concluded that 5d dihydrogen complexes are always more stable to dihydrogen loss than the analogous 3d or 4d complexes. However the relative lability of H2 in 3d and 4d metal complexes (3d < 4d or 3d < 4d) depends on the ancillary ligands.20 We have now shown that compounds trans-[MH(X2)- (depe)2]1 are more labile to dissociation of X2 when X = N than when X = H but only by a factor of ca.3.5 (see rate constants in Table 2). This trend in the lability is independent of temperature. The rate constants are also dependent upon the alkyl group of the diphosphine increasing ca. 5-fold from dmpe to depe in trans-[FeH(N2)(pp)2]1 (pp = dmpe or depe) in all solvents and for all nucleophiles. It has been reported 9a that there is not much diVerence in the “stabilities” of trans-[MH(H2)(pp)2]1 (M = Fe Ru or Os; pp = depe or dppe) though whether they are kinetic or thermodynamic stabilities was not clarified. We rationalise the diVerence between the dmpe and depe complexes in the following terms. Kubas et al.28 have stated that steric interactions are of much less consequence than electronic eVects in stabilising H2 (and presumably also N2) bound to a metal.The ethyl group in depe makes it a better s donor than the methyl group in dmpe. Consequently less s donation to the metal from dinitrogen may occur although there may be more p acceptance into dinitrogen. This could weaken Fe-to-N2 bonds more in the depe complex than in the dmpe complex. Arguments based upon steric interactions between the phosphine substituents and the dinitrogen in the ground state and in the transition state would lead to the opposite conclusion. We conclude that electronic eVects appear to be more important than steric for our complexes. Although there are no literature data available for a direct comparison with ours it is instructive to extend our discussion to closely related compounds for which some information is available.For example it has been shown qualitatively that trans ligands which compete eVectively with dihydrogen for p-electron density weaken the back donation (MÆH2).20 The trans eVect influences both the kinetic and thermodynamic stability of dihydrogen complexes and is shown clearly by Group 8 trans-[MY(X2)(pp)2]1 complexes (Y = H or Cl; X = H or N; pp = diphosphine) where the complexes trans-[MCl(H2)- (pp)2]1 are generally more reactive than the analogous hydrido- (dihydrogen) complexes. The strength of the metal-to-H2 bond follows a diVerent trend (3d < 4d < 5d) in Group 8 for a chloride trans to dihydrogen than for a hydride trans to dihydrogen (4d < 3d< 5d).15,20,29,30 Ruthenium complexes containing a chloride trans to dihydrogen are relatively more stable than for the trans hydride series in trans-[RuY(H2)(pp)2]1 (Y = H or Cl; pp = depe or dppe).For example trans-[RuCl(H2)(depe)2]1 is stable to H2 loss (although it can lose HCl through reductive elimination) but dihydrogen in trans-[RuH(H2)(depe)2]1 is labile.31 However when pp = Cy2PCH2CH2PCy2 (Cy = cyclohexyl) the converse is true.30 Therefore the eVect of the trans ligand on the lability of the dihydrogen also depends on the ancillary ligands. Dinitrogen in trans-[FeCl(N2)(depe)2]1 is very labile,32 dissociating from the complex both in solution and in the solid state whereas we have shown that the corresponding transhydride complexes are more stable to dinitrogen loss under the same conditions.We conclude that the rate constant for the loss of X2 depends subtly on the metal and on all the ancillary ligands. Our results (Table 6) show that changing the solvent from acetone to thf reduces the rate constants by a very small amount ca. 15–35%. Similar small changes were observed in the reactions of trans-[FeH(H2)(dppe)2]1. In these substitution reactions a similar degree of solvation is to be expected for the transition state and reactants since the solvents are of similar polarity. The observation that the rate constants for the substitution of X2 in our complexes increase slightly from thf to acetone may be more to do with the donor power of the individual solvents than with any changes in polarity of species during the reaction. Factors aVecting the activation parameters The values of Eact DH‡ and DS‡ for trans-[FeH(H2)(pp)2]1 are all larger than for trans-[FeH(N2)(pp)2]1 although the values of DG‡ are very similar.This implies that dihydrogen is more strongly bound to iron than dinitrogen in the same ancillary ligand environments. Dihydrogen complexes are there- J. Chem. Soc. Dalton Trans. 1999 1213–1220 1219 fore more thermodynamically stable than dinitrogen complexes as well as being more kinetically stable. Comparable values also increase from depe to dmpe although again DG‡ values are almost invariant. The increase of these activation parameters is associated with the ca. 5-fold decrease in rate constant when depe is replaced by dmpe. Changing the solvent has little eVect on the activation parameters. The mechanism of substitution reactions of [MH(X2)- (tetraphos)]1 The tetraphosphines pp3 and pp3Me force a cis configuration on the associated X2 and hydride ligands.The kinetics and thermodynamics for the loss of X2 from cis-[MH(X2)- {P(CH2CH2PR2)3}]1 were not studied because of the lack of reactivity or else the instability of these complexes in solution under argon. Several of the homologues of this kind have yet to be synthesized. Consequently we can make only qualitative comparisons between the complexes with two diphosphines and those with one tetraphosphine. The diVerences in reactivity between cis and trans (hydride and X2) complexes are pronounced for iron. The complex cis-[FeH(H2)(pp3)]1 is kinetically stable to nucleophiles in solution 11a,b under argon at room temperature whereas trans- [FeH(H2)(pp)2]1 react within a couple of hours under the same conditions.3,9a The faster rates of dissociation of N2 from cis- [FeH(N2)(pp3)]1 compared to that from trans-[FeH(N2)(pp)2]1 are highlighted by the fact that the cis complex is unstable in potentially ligating solvents even under dinitrogen.This made measurement of NMR spectra diYcult although no solvolysis products were detected spectroscopically. The complex cis-[FeH(H2)(pp3Me)]133 is more labile than cis- [FeH(H2)(pp3)]1.11a,b It will react with nucleophiles to completion after several hours at room temperature or after 30 min at 333 K whereas cis-[FeH(H2)(pp3)]1 takes several days to react at room temperature. The rates of reaction of ruthenium or osmium tetraphosphine complexes are not as well documented as of iron.Nevertheless the trend in reactivity Ru @ Fe @ Os was found in cis-[MH(X2)(pp3)]1 just as in trans-[MH(X2)- (pp)2]1. Bianchini et al. 34 stated that cis-[MH(H2)(pp3)]1 is more stable than trans-[MH(H2)(pp)2]1 because of the “attractive cis-eVect” in the former. This eVect is supposed to arise by incipient formation of an H3 ligand due to the enforced cis arrangement of H and H2 and it is reported to be significant for X2 = H2. There is no comparable eVect for X2 = N2 or CO.35 A cis interaction is obviously impossible in a trans arrangement of H and H2.11b,36 There is qualitative evidence that cis-[RuH- (H2)(pp3)]137 reacts with nucleophiles much faster than cis- [FeH(H2)(pp3)]1,11a,b and that cis-[OsH(H2)(pp3)]1 reacts much slower than the iron complex.34 The hydride complex cis- [RuH(H2)(pp3)]1 is unstable in solution unless under dihydrogen and it is therefore less kinetically stable than trans- [RuH(H2)(pp)2]1.9a,37 The complex cis-[OsH(H2)(pp3)]135 reacts with nitriles only at high temperatures which is similar to our observations for trans-[OsH(H2)(depe)2]1.The compound cis-[OsH(N2)(pp3)]134 reacts rapidly with nitriles which is also consistent with our data because we had diYculty even isolating trans-[OsH(N2)(depe)2]1 due to the lability of dinitrogen.10 The complex cis-[MH(X2)(pp3Me)]1 (M = Ru or Os) have not been described in the literature. The strengths of the bonds between metals and the dihydrogen ligand,20 as indicated by the magnitude of the thermodynamic parameter (H8 and by IR spectroscopy increase in the order 4d < 3d < 5d. This order is common for many isostructural complexes of metals of Groups 6 and 8,20 for example for trans-[MH(H2)(pp)2]1.9a A less common order 3d < 4d < 5d has been noted 20 for trans-[MCl(H2)(depe)2] and the order 3d 4d < 5d has been reported for [MH(H2){PPh(OEt)2}4]1 and for the restricted series [MH4(PPh3)4] (M = Ru or Os).20 The steric crowding at the metal in the Group 8 complexes cis-[MH(H2)- (pp3)][BPh4] decreases with increasing metal radius,38 but the catalytic activity has been shown to relate to the strengths of the metal–dihydrogen bond which changes in the order Os � Fe @ Ru.Our data are clearly not inconsistent with these trends. Possible implications for nitrogenase mechanism The production of H2 and the fixation of N2 are two reactions which are intimately associated with one another in the action of the nitrogenases.In the absence of any other reducible substrate nitrogenases will reduce H1 to H2. Progressively introducing more N2 results in a decrease in the amount of H2 formed and a concomitant increase in the amount of NH3 produced. However even at high pressures of N2 the production oH2 cannot be suppressed entirely. Thus the limiting stoichiometry for the action of the molybdenum-containing nitrogenases is that shown in eqn. (13) (Pi represents inorganic N2 1 8H1 1 16 ATP 1 8 e æÆ 2 NH3 1 H2 1 16 ADP 1 16 Pi (13) phosphate) in which approximately one mole of H2 is produced for every mole of N2 reduced. With the vanadium-based nitrogenases the limiting stoichiometry involves proportionately even more H2 ca. 3 H2 per N2 fixed.2 A variety of mononuclear chemical systems and the work presented in this paper clearly show that displacement of H2 at a metal site by N2 can occur.However most studies (including our own) show that the mechanism is dissociative and hence that the H2 dissociation is not facilitated by the attacking nucleophile dinitrogen. It has been proposed that these complexes represent models for the N2 binding in the enzyme.39 However when the reactions of nitrogenases with D2 are studied it becomes clear that the enzyme is performing much more elaborate chemistry which these simple complexes are not mimicking in any sense. The formation of HD by conventional molybdenum–iron nitrogenases when they reduce N2 in the presence of D2 is one of the most intriguing phenomena associated with biological nitrogen fixation.2 The stoichiometry is represented by eqn.(14). The most striking features of the HD formation are as D2 1 2 H1 1 2 e æÆ 2HD (14) follows. It is associated with no reducible substrate other than N2 there is no indication that any of the deuterium ever passes into solution and no D2 is formed when fixing N2 in the presence of HD. The clear implications are that the H and D which are eventually combined in HD come from diVerent sources that do not mix their hydrogen atoms and most important that the reaction is facilitated only when N2 is bound. This implies that the displacement of H2 by N2 at a single active site must be an associative process. The most complete explanation put forward to explain this phenomenon in nitrogenase is that N2 binds to a trihydride species MH3 with displacement of H2.Subsequent loss of N2 (by reaction with protons towards ammonia or by simple dissociation) followed by binding of D2 would generate MHD2 and this last species is a plausible source of HD.40 This scheme is made more attractive by the Lowe–Thorneley 41 model of nitrogenase mechanism that has been interpreted to mean that a trihydride species is indeed generated before dinitrogen is bound and that some dihydrogen is released when that process occurs. Further there are model chemical compounds such as [CoH3(PPh3)3],42 that apparently exhibit similar N2/H2 displacement reactivity. However this model does not explain why in the comparable experiment performed under HD no D2 is ever formed. Nor does it explain why substrates other than dinitrogen do not also stimulate HD formation.After all those substrates may also 1220 J. Chem. Soc. Dalton Trans. 1999 1213–1220 be imagined to bind at the nitrogenase active site. Finally if dihydrogen is also able to interact with the active site why is any substrate at all necessary to promote HD formation? Chemical models and the work presented above clearly show that displacement of dihydrogen is not a necessity for binding dinitrogen. The formation of a considerable proportion of the dihydrogen generated by nitrogenases during turnover does not require the presence of an incoming group to provoke it. Consequently why is HD formation observed only when dinitrogen is being reduced? Chemical systems have been developed in which binding of N2 can occur before the release of H2. These involve dissociation of a carboxylate group from the co-ordination sphere of molybdenum hydrido species.43 This also has the additional merit of being consistent with the sequence of events presented by the Lowe–Thorneley model.41 However the specific reactivity of the enzyme as described above has yet to be successfully mimicked in a chemical system.The simplest rationalisation is that HD formation and dinitrogen binding (and perhaps by extension ordinary dihydrogen evolution) occur at diVerent places. The active site of these nitrogenases is the iron–molybdenum cofactor an Fe– S-based cluster with the stoichiometry MoFe7S9.2 It is not unreasonable to assume that diVerent substrates bind and are transformed at diVerent parts of this large cluster. More data are required. For example dinitrogen is a ligand that like CO should stabilise low oxidation states of metals in complexes.Carbon monoxide inhibits nitrogen fixation in nitrogenases but not dihydrogen evolution. It should also be able to facilitate HD formation though we know of no attempts to check this. However evidence is now accumulating that the nitrogenase cluster promotes multi-site processes and chemical models for nitrogenase function must begin to take account of this. Acknowledgements We acknowledge support from EPSRC through a CASE award to C. A. H. and from BBSRC and from Johnson Matthey plc for the loan of ruthenium and osmium compounds. We are exceedingly grateful to Professor L. D. Field and Dr R. J. Smernik (University of Sydney) for much helpful information concerning syntheses and for many kindnesses.References 1 See for example G. J. Leigh in Comprehensive Biological Catalysis ed. M. J. Sinnott Academic Press Limited Chichester 1998 ch. 35. 2 For reviews see R. R. Eady and G. J . Leigh J. Chem. Soc. Dalton Trans. 1994 2739; G. J. Leigh Eur. J. Biochem. 1995 229 14. 3 A. Hills D. L. Hughes M. Jimenez Tenorio G. J. Leigh and A. T. Rowley J. Chem. Soc. Dalton Trans. 1993 3041 and refs. therein. 4 M. Jimenez Tenorio D. Phil. Thesis University of Sussex 1990; M. Jimenez Tenorio and G. J. Leigh J. Am. Chem. Soc. 1991 113 5802. 5 C. A. Helleren D. Phil. Thesis University of Sussex 1998. 6 D. A. Hall D. Phil. Thesis University of Sussex 1994. 7 M. G. Basallotte J. Duran M. J. Fernandez-Trujillo G. Gonzales M. A. Manez and M. Martinez Inorg. Chem. 1998 37 1623. 8 R. A. Henderson J.Chem. Soc. Dalton Trans. 1988 509 515. 9 (a) M. T. Bautista E. P. Cappellani S. D. Drouin R. H. Morris C. T. Schweitzer A. Sella and J. Zubkowski J. Am. Chem. Soc. 1991 113 4876 and refs. therein; (b) A. Hills D. L. Hughes M. Jimenez Tenorio and G. J. Leigh J. Organomet. Chem. 1990 391 C41; (c) C. N. Mc Mahon D. Phil. Thesis University of Sussex 1995. 10 G. M. Bancroft M. J. Mays B. E. Prater and F. P. Stefanini J. Chem. Soc. A 1970 2146. 11 (a) C. Bianchini M. Peruzzini and F. Zanobini J. Organomet. Chem. 1988 354 C19; (b) C. Bianchini M. Peruzzini A. Polo A. Vacca and F. Zanobini Gazz. Chim. Ital. 1991 121 543; (c) P. Stoppioni F. Mani and L. Sacconi Inorg. Chim. Acta 1974 11 227. 12 L. D. Field B. A. Messerle and R. J. Smernik Inorg. Chem. 1997 36 5984; R. J. Smernik University of Sydney personal communication.13 E. P. Cappellani S. D. Drouin G. Jia P. A. Maltby R. H. Morris and C. T. Schweitzer J. Am. Chem. Soc. 1994 116 3375. 14 E. P. Cappellani P. A. Maltby R. H. Morris C. T. Schweitzer and M. R. Steele Inorg. Chem. 1989 28 4437. 15 M. Bautista K. A. Earl R. H. Morris and A. Sella J. Am. Chem. Soc. 1987 109 3780. 16 E. A. Guggenheim Philos. Mag. 1926 2 538. 17 J. H. Espenson Chemical Kinetics and Reaction Mechanisms McGraw-Hill New York 1981. 18 S. R. Logan Fundamentals of Chemical Kinetics Longman London 1996; P. W. Atkins Physical Chemistry 4th edn Oxford University Press Oxford 1990; E. L. Tepper and E. Pollak Chem. Br. 1997 33 22. 19 K. A. Earl G. Jia P. A. Maltby and R. H. Morris J. Am. Chem. Soc. 1991 113 3027. 20 P. G. Jessop and R. H. Morris Coord.Chem. Rev. 1992 121 155 and refs. therein. 21 D. J. Darensbourg Adv. Organomet. Chem. 1982 21 113. 22 F. A. Cotton and G. Wilkinson Advanced Inorganic Chemistry 5th edn. Wiley Interscience Chichester 1988. 23 H. M. Marques J. C. Barclay and L. A. Campbell J. Chem. Soc. Dalton Trans. 1992 2019. 24 C. Bianchini A. Meli M. Peruzzini P. Frediani C. Bohanna M. A. Esteruelas and L. A. Oro Organometallics 1992 11 138. 25 W. Linert and R. F. Jameson Chem. Soc. Rev. 1989 18 477; W. Linert Chem. Soc. Rev. 1994 23 ,429. 26 J. Halpern L. S. Cai P. J. Desrosiers and Z. R. Lin J. Chem. Soc. Dalton Trans. 1991 717. 27 P. Amendola S. Antoniutti G. Albertin and E. Bordignon Inorg. Chem. 1990 29 318. 28 G. J. Kubas R. R. Ryan and C. J. Unkefer J. Am. Chem. Soc. 1987 109 8113. 29 R. H. Crabtree and D.G. Hamilton J. Am. Chem. Soc. 1986 108 3124. 30 A. Mezzetti A. Del Zotto P. Rigo and E. Farnetti J. Chem. Soc. Dalton Trans. 1991 1525. 31 A. C. Albinez M. Heinekey and R. H. Crabtree Inorg. Chem. 1991 30 3632. 32 B. E. Wiesler N. Lehnert F. Tuczek J. Neuhausen and W. Tremel Angew. Chem. Int. Ed. Engl. 1998 37 815. 33 L. D. Field University of Sydney personal communication. 34 C. Bianchini K. Linn D. Masi M. Peruzzini A. Polo A. Vacca and F. Zanobini Inorg. Chem. 1993 32 2366. 35 C. Bianchini D. Masi M. Peruzzini M. Casarin C. Maccato and G. A. Rizzi Inorg. Chem. 1997 36 1061. 36 L. S. Van der Sluys J. Eckert O. Eisenstein J. H. Hall J. C. HuVman S. A. Jackson T. F. Koetzle G. J. Kubas P. J. Vergamini and K. G. Caulton J. Am. Chem. Soc. 1990 112 4831. 37 C. Bianchini P. J. Perez M. Peruzzini F. Zanobini and A. Vacca Inorg. Chem. 1991 30 279. 38 C. Bianchini E. Farnetti M. Graziani M. Peruzzini and A. Polo Organometallics 1993 12 3753. 39 R. A. Henderson G. J. Leigh and C. J. Pickett Adv. Inorg. Chem. Radiochem. 1983 27 197. 40 J. Chatt Proc. Phytochem. Soc. Eur. Symp. 1980 18 1. 41 D. J. Lowe and R. N. F. Thorneley Biochem. J. 1984 224 887. 42 A. Yamamoto L. S. Pu S. Kitazume and S. Ikeda J. Am. Chem. Soc. 1967 ,89 3071. 43 C. J. Pickett J. Bioinorg. Chem. 1996 1 601 and refs. therein. Paper 8/10001B

ISSN:1477-9226    

DOI:10.1039/a810001b

出版商:RSC    

年代:1999

数据来源: RSC