摘要:

DALTON PERSPECTIVE J. Chem. Soc., Dalton Trans., 1998, 3529–3538 3529 Transition metal catalysed reactions using glass bead technology Michael S. Anson,a Mathew P. Leese,b Louise Tonks b and Jonathan M. J. Williams *b a GlaxoWellcome Research and Development, Medicines Research Centre, Gunnels Wood Road, Stevenage, Herts, UK SG1 2NY b Department of Chemistry, University of Bath, Claverton Down, Bath, UK BA2 7AY Received 6th April 1998, Accepted 28th July 1998 A successful method for rendering homogeneous transition metal catalysts heterogeneous has emerged from glass bead technology.The process generally combines high activity and selectivity with facile recyclability. Glass bead technology, and related approaches, have been applied to a wide range of reactions with results paralleling, and sometimes exceeding, those obtained from their homogeneous counterparts. Outlined is the current scope of this research, with an emphasis on work on supported palladium reagents.A novel approach for the retrieval of precious metals from reaction mixtures stemming from these studies is also outlined. 1 Introduction It is desirable that a catalyst should possess high activity/ catalytic turnover combined with high regio- and stereoselectivity, indeed these desirable properties have emerged from the development of homogeneous transition metal catalysis over several decades.1 Yet despite these developments and the inherent advantages over their heterogeneous counterparts, homogeneous catalysts have been underused by the fine chemical industry partly due to the problematic separation of catalyst and product.Industrial interest spurred the introduction of biphasic counterparts of the rhodium catalysed hydroformylation in the early 1980s by development and application of water soluble ligands (notably TPPTS 1, a sulfonated triphenylphosphine analogue) and also served to reinvigorate academic interest in overcoming the technological limitations of the powerful catalysts available.2,3 Representative anionic and cationic ligands are 1 to 6.Although biphasic reactions have proved to be a successful solution in some cases, these reactions are impeded by solubility of substrates in the aqueous layer in which the catalyst resides and a limited interface between the reactants, although this technique does aVord the opportunity for rational catalyst design and optimisation.4 Many imaginative strategies have been developed for harness- Mike Anson received his BSc and PhD from University College London, where he worked with Professor Chris McGuigan until 1988.After 2 years as a post-doc with Professor Mike Blackburn at Sheffield he joined Process Research at Glaxo in 1990. He is currently a team leader at GlaxoWellcome (Stevenage) where he is involved in evaluating new technologies for rapid scale-up. Louise Tonks was born in 1973 in Rochdale, Lancashire. She received her BSc in 1995 from Sheffield Hallam University before moving on to PhD studies with Professor J.M. J. Williams at Loughborough and the University of Bath. Her research interests lie in transition metal catalysis and, in particular, C–C coupling reactions using palladium catalysts supported on glass beads. Mathew Leese was born in Barnsley in 1973. He first studied chemistry at the University of Nottingham, obtaining his BSc in 1994 before joining Professor D. W. Knight to study for a PhD, consequent relocation to the University of Wales, Cardiff followed in late 1995.He also received a Monbusho fellowship to work with Professor T. Takahashi at the Institute of Molecular Sciences, Okazaki, Japan during 1995. After completing his studies he moved to Bath to join Jon Williams group in November 1997 to work on novel catalyst systems. Jonathan Williams obtained his BSc from the University of York in 1986, and then did a D.Phil with Professor Steve Davies at Oxford.After a post-doctoral stay with Professor David Evans at Harvard, and a lectureship at Loughborough University, he moved to the Chair of Organic Chemistry at the University of Bath in 1996. His main research interests are in asymmetric catalytic methods. Michael Anson Louise Tonks Mathew Leese Jonathan Williams3530 J. Chem. Soc., Dalton Trans., 1998, 3529–3538 ing the power of transition metal catalysis by anchoring the ligands to various solid supports, but the eYcacy of modified catalysts is often hampered by the suppression of catalyst mobility stemming from the anchoring process.5 In our eyes, one innovative solution to this problem which appears to oVer the ideal union of the chemical advantages of homogeneous catalysis with the practicability of a solid, heterogeneous system is supported aqueous phase (SAP) catalysis, first demonstrated by Arhancet et al.6 in 1989.This technique uses the hydrophilic nature of the surface of controlled pore glasses (CPGs) to adhere a layer of water (or another suitable polar solvent) in which a hydrophilic catalyst can be anchored.This allows a large surface area of catalyst (surface area of 1 g of CPG-240 = 77.5 m2), whilst the catalyst retains mobility within the polar solvent on the bead. In contrast to many other supports the beads are stable to mechanical and thermal fatigue. The essential features of glass bead technology are illustrated in Fig. 1. The philosophy underlying this approach is long established by supported liquid phase catalysis (SLPC),7 a technique in which a catalyst supported in a liquid film interacts with substrates in the gaseous phase. This approach is limited to gaseous reactants. Glass bead technology and related approaches have been successfully applied to many transition metal catalysed reactions, and this information is summarised in Table 1. Subsequent sections of this article describe the chemistry in further detail. 2 Rhodium hydroformylation catalysts The catalyst system initially studied by Davis and co-workers 6 was the rhodium hydroformylation catalyst [RhH(CO)- (TPPTS)3] used in the biphasic reaction.2,26 Application to glass bead technology is eVected by pouring an aqueous solution of the catalyst and an excess of TPPTS onto degassed CPG-240. After further degassing and stirring, the water was removed by evaporation to yield the “dry” catalyst with a water content of 2.9%.The resultant “glass bead catalyst” was tested in the hydroformylation of oleyl alcohol 7 (a water insoluble substrate which is not hydroformylated under biphasic conditions),8 which provided a good conversion into the hydroformylated product 8 (Scheme 1). An alternative, self assembly method of catalyst preparation involves loading [Rh(acac)(CO)2] on to the bead and then treating with a solution of water soluble phosphine. 9 Ligand exchange in this case proves to be rapid as indicated by the change from white to yellow; and this approach is advantageous in that oxidative catalyst degradation is minimised.In order to maximise catalyst stability it is necessary to use an excess of phosphine (in this case P :Rh < 3 : 1); it has also been found that the dry glass bead catalyst is more stable to storage. The water content of the glass bead catalyst has been shown greatly to aVect the activity of the catalyst and the selectivity of the reaction.Davis and co-workers 8 reported optimum activity for the rhodium catalysed hydroformylation reaction at a water content of ª8%. At this level the catalyst has a high degree of mobility and yet is still available at the interface, whereas lower water content suppresses mobility as reflected by lower activity, and higher water content renders the catalyst less available to the water insoluble substrate. Varying the degree of hydration is Fig. 1 A schematic of the supported aqueous phase catalyst.The reaction is thought to proceed at the interface. Scheme 1 (i) [RhH(CO)(TPPTS)3] on CPG-240, cyclohexane, H2–CO (1 : 1), 50 atm, 5.5 h, 100 8C, 96.6% conversion.J. Chem. Soc., Dalton Trans., 1998, 3529–3538 3531 Table 1 Current scope of glass bead technology Reaction Hydroformylation Hydrogenation Wacker oxidation Heck reaction Suzuki reaction Allylic substitution Metal Rh Co Pt Pt/Sn Ru Rh Pd/Cu Pd Pd Pd Ligand L TPPTS TPPTS HexDPPS TPPTS TPrPTS TPPTS TPPTS TPrPTS (S,S)-BDPP* BINAP* TPPTS (S,S)-BDPP* Chiraphos* TPPTS TPPMS TPPTS TPPTS Catalyst [RhH(CO)L3] [RhH(CO)L3] [RhH(CO)L3] [Co2(CO)6L2] Na5[Co(CO)3L2] [PtCl2L2] [PtCl(SnCl3)L2] [PtCl(SnCl3)L2] [PtCl(SnCl3)L] [Ru(C6H6)L(Cl)]Cl [RuCl2L3] [RuH2L3] [Rh(COD)L]BF4 [Rh(COD)L]BF4 PdCl2–CuCl2 [PdCl2L2] [PdCl2L2] [PdCl2L2] [PdCl2L2] Substrate(s) Alkenes a,b-Unsaturated esters Alkenes Alkenes Alkenes Alkenes Alkenes Alkenes Styrene Alkenes a,b-Unsaturated aldehydes a,b-Unsaturated aldehydes Alkenes Alkenes Alkenes Aryl halides Aryl halides Aryl halides Allyl acetates Ref. 8–13,16 14,15 13 17 13 18 18 13 13 19–21 22 22 13 13 23 24 24 — 24,25 * Water soluble derivative of this ligand as illustrated. achieved by exposure of the catalyst to a known vapour pressure of water or by water transfer from hydrated controlled pore glass beads. The activity diVerence between the hydrated and “dry” glass bead catalysts has been reported to be typically two or more orders of magnitude.10 The water content is also seen to influence the regioselectivity of the reaction which can be ascribed to a variation in catalytic pathway.The hydroformylation of hept-1-ene yields similar results in terms of conversion and n:b ratio as that of tetradec-1-ene and heptadec-1-ene (Scheme 2). The dependence on water content was also reported by Horváth;11 starting from high water content the activity of the catalyst in a trickle bed reaction was seen to increase as water leached from the bead into the organic layer until the water content was only suYcient to supply two monolayers to the surface of the controlled pore glass bead.This corresponds to greater presence of the active site of the mobile catalyst at the interface, instead of a catalyst immersed in the supported solvent. Once the catalyst is placed on the bead structural elucidation becomes more complex, and inferences on the actual catalytic species can only be made with respect to the observed activity of the catalyst compared with the homogeneous and biphasic counterparts.Infrared spectroscopy can be used on the reaction solution and indicates the absence of rhodium carbonyls in the reaction, but low catalyst loading, combined with the broad silica band at ª1850 cm21 render both diVuse reflective and transmission IR ineVectual for the study of the glass bead systems. In the solution phase the catalyst, [RhH(CO)(TPPTS)3], loaded onto the bead exists as a trigonal bipyramid with the phosphines occupying the equatorial positions.Solid phase 1H NMR techniques on the dry catalyst suYce only to confirm the presence of the phosphine ligands, whilst CP MAS 31P NMR indicates that the phosphines are stationary in the dehydrated Scheme 2 (i) [RhH(CO)(TPPTS)3] on CPG-240, cyclohexane, H2–CO (1 : 1), 6.8 atm. catalyst. Fortunately, once the catalyst is partially hydrated then normal solution phase NMR techniques can be utilised and indicate mobility of the phosphine ligands in the supported solvent layer.27 Electron microprobe analysis has been utilised to reveal a uniform Rh:P ratio across the surface of the catalyst.8 Horváth11 conducted further studies on the rhodium catalysed hydroformylation in order to investigate the exact nature of the catalyst itself.The ratio of straight to branched products (n :b ratio) of the reaction varies greatly between the supported aqueous phase catalyst and the biphasic reaction, indicating the mechanistic diVerence between the two cases.On comparison of the supported aqueous phase reaction and its homogeneous and biphasic analogues the nature of the rhodium centre was proposed from a consideration of the branched to straight chain ratio of the aldehyde products; in the case of the biphasic reaction the n :b ratio is high suggesting the catalytic rhodium species is [RhH(CO){P(C6H4SO3Na-m)3}2] with the two phosphines in the trans position.The glass bead catalyst system produces an n:b ratio, similar, yet slightly higher than the homogeneous system in which the key intermediates are either [RhH(CO)(PPh3)2] or [RhH(CO)2(PPh3)],28 leading Horváth to propose that the catalytic species in the glass bead catalyst system is [RhH(CO)2{P(C6H4SO3Na-m)3}], although the possibility that a complex with two phosphines with a cis configuration could produce the same result was acknowledged. 29,30 Davis proposed this eVect may stem either from the higher eVective phosphine concentration in the glass bead or a diVerent catalytic species for the glass bead catalyst.19,20 Yuan et al.12 investigated the eVect of codeposition of alkali metal salts on the surface of the beads prior to hydration.The resultant glass bead catalysts were reported to benefit in terms of both activity and n :b ratios in the hydroformylation of terminal alkenes when the ratio KCl :Rh was in the range 2–10 : 1. Addition of salt can be expected to assist CO insertion in alkyl ligands by stabilising the polar transition state thus increasing the steady state concentration of the acyl intermediate, leading to an increase in turnover frequency.Tóth et al.13 synthesized HexDPPS 2 in order to attempt to make the ligand more surface active, with the reasoning that the lipophilic chain should bring the metal closer to the interface. In fact the supported aqueous phase catalyst formed from this ligand proved to be less active, yet showed similar n :b ratios, no sign of leaching and was recyclable.The analogies were suggested to reflect the similar action of the ligand, whilst reinforcing the previous conclusions that the TPPTS analogues act at the interface.3532 J. Chem. Soc., Dalton Trans., 1998, 3529–3538 Frémy et al.14,15 reported that hydroformylation using supported aqueous phase catalysis can proceed with higher turnover frequencies than the analogous homogeneous reaction.Methyl acrylate, a polar substrate, was subjected to the hydroformylation reaction introduced by Davis (see above) with silica as the support, and proved to yield average turnover frequencies over ten times greater than those observed for the analogous biphasic and homogeneous reactions, this eVect being ascribed to the beneficial interactions between methyl acrylate, supported solvent and surface hydroxyl groups. Once again a great dependence of activity on water content was observed, which is in contrast to Davis’s results for nonpolar substrates, thus reflecting the high degree of dependence on substrate functionality.Optimum conditions were realised with total pore filling; a steady rise in activity is observed as pore filling increases to unity, once water content exceeds this level however the activity drops rapidly and approaches that of the biphasic system. This result was initially obtained using silica gel (60 Å) as a support; subsequent studies show diVerent silicas gave identical activity as a function of the degree of pore filling. Interestingly for these supports catalyst performance is independent of mean pore diameter and surface area.Supported aqueous phase catalysts were also prepared from CPG-240 and -350 (controlled pore glasses with a mean pore diameter of 240 and 350 Å respectively) in order to compare the eVect of these more defined particles; surprisingly they were shown to exhibit lower activities than the far cheaper silica analogues, an eVect that may be due to the higher surface area of the silica as, in this reaction, the solvent content is such that the pores are full.Reasonable catalytic activity was also observed when molecular sieves and alumina were used as the support. Substrates which were less polar than methyl acrylate reacted with optimum activity at low pore filling, where a large surface area contact was required. The use of alternative supported hydrophilic layers in the hydroformylation reaction has been investigated by Naughton and Drago,16 who, after studying the catalytic properties of their system proposed the term supported homogeneous film catalysts (SHFCS).High boiling point liquid or liquid polymer films containing the hydrophilic rhodium catalyst [RhH(CO)- (TPPTS)3] on the surface of silica were synthesized in an analogous manner to that described previously by Davis and co-workers.9,10 Polyethylene glycol was selected for its ability to form films, its non-volatile nature, its insolubility in substrate and product, and its ability to dissolve and retain the catalyst.The activity of the bare bead (in eVect Davis’s dry catalyst) showed low activity, addition of suYcient polyethylene glycol (PEG 600, average molecular weight 600) to give a pore filling (d) of 0.5 doubled the activity (TOF 2.6 min21), whilst use of water–PEG 600 (1 : 1) gave a further increase in activity (TOF 8.3 min21, d = 1.4).It was found that, in the hydroformylation of hex-1- ene, activities of 24.7 min21 and n:b ratios of 6 : 1 could be achieved by using a pore filling of 1.4 (100 8C and 85 psig; psi ª6895 Pa) which compares well with the homogeneous analogue under identical conditions which gives TOF 68 min21 and n :b 3. In contrast to the homogeneous reaction it was found desirable to carry out the reaction without the addition of bulk solvent. In the hydroformylation of oct-1-ene a reduced activity (when compared with hex-1-ene) is observed which is thought to reflect the lower solubility of the substrate in the film.Interestingly the activity of the catalyst in this case increases as the reaction proceeds; this is ascribed to retention of the product in the film (as evidenced by IR spectroscopy) which can then function as a surfactant and thus increases the rate of transfer of the substrate into the film. Here the observed dependence of TOF on the solubility of the substrate contrasts with that for the supported aqueous phase catalysts reported previously, and leads the authors to suggest that the substrate is dissolved in and then reacts in the film.A number of results support this conclusion. (i) Use of a more viscous PEG film (PEG 8000) serves to impede the reaction; if the catalyst were solely active at the interface then the detrimental eVect of increased viscosity would not be expected.Indeed modification of this catalyst by addition of water was responsible for an increase in activity reflecting increased reaction rate from reduced viscosity of the film. (ii) Addition of a surfactant to the catalyst for the octene hydroformylation is accompanied by an increase in initial TOF which parallels that of the pure PEG film after longer reaction times. (iii) The optimum pore filling is 1.4 (compared with 0.06 for the normal supported aqueous phase catalyst) which corresponds to full pores in the support and is approaching the level at which droplets form on the bead.The film has some solubility in concentrated solutions of the aldehyde product, which is reflected by some leaching of the PEG into the bulk, yet no leaching of the metal is observed. Some suppression of PEG leaching could be achieved by addition of sodium sulfate to the film. Alternative films of polyvinylpyrrolidine, polyethylene oxide and polyvinyl alcohol showed drastically reduced activity or no activity. 3 Cobalt hydroformylation catalysts Glass bead technology has been applied to cobalt reagents for the hydroformylation reaction of alkenes.17 The cobalt complex [Co2(CO)6(TPPTS)2] has been investigated as a cheap alternative to rhodium catalysts in biphasic hydroformylations but suVers from much lower activity.31 Another drawback of the biphasic reaction is the high leaching levels stemming from the equilibria in which the catalyst likely exists with the organic soluble dicobalt octacarbonyl [Co2(CO)8] and hydridocobalt tetracarbonyl [CoH(CO)4] species under the high pressure of CO and H2 in the reaction.The glass bead cobalt catalyst displays greatly reduced leaching (a tenth of the biphasic counterpart); a twofold excess of phosphine complements the relative increase in phosphine concentration already apparent on the bead surface to yield this desirable eVect. It is notable that in contrast with the rhodium system little dependency of activity of the catalyst is observed with a variation in water content of the support; this may imply little relationship of catalyst mobility and activity, alternatively the substrate phase may be homogeneous at the high temperature at which the reaction takes place.Significantly the activity of the supported catalyst does not correlate with cobalt leached to the organic phase, indicating the catalytic species to be cobalt supported on the glass. This conclusion is supported by the higher n :b ratio obtained in the reaction, which is similar to the equivalent homogeneous reaction (Scheme 3).The n :b ratio is double that observed for the biphasic reaction implying an alternative reaction path whilst lower loading levels are accompanied by higher activity per mole cobalt suggesting that much of the cobalt in the Scheme 3 (i) [Co2(CO)6(TPPTS)2] on CPG-340 (3 g), toluene (50 ml), H2–CO (1: 1), 54 atm, 190 8C, 8 h.J. Chem.Soc., Dalton Trans., 1998, 3529–3538 3533 supported aqueous phase catalyst is unavailable for catalytic activity. In an attempt to suppress displacement of the phosphine ligands in the desired catalyst [Co2(CO)6(PR3)2] by CO to yield a less selective catalytic species, the same research group synthesized alkylphosphines with sulfonated aromatic groups to serve as anchors.13 4 Platinum hydroformylation catalysts Hanson and co-workers 18 have explored hydroformylation with [Pt(TPPTS)2Cl2]. The non-sulfonated phosphine complex [Pt(PPh3)2Cl2] is not active in homogeneous hydroformylation reactions, yet for both biphasic and glass bead catalyst conditions the TPPTS complex does serve as a catalyst.This activation by the modified ligand is attributed to the sulfonate group functioning as a ligand in analogy with the activation of other platinum complexes by the addition of methanesulfonic acid.32 Platinum hydroformylation catalysts are commonly activated by tin(II) chloride as cocatalyst, however Pt–Sn bonds are subject to facile hydrolysis and for this reason were not considered likely candidates for water soluble catalysts.However, it is possible to use only a minimum amount of water on the surface of the glass bead to anchor the sulfonate groups, therefore allowing the possibility of using supported aqueous phase conditions to support a catalyst which is inherently unstable in water. Catalyst preparation in this case was achieved by deposition of complex 15 (cis isomer) on the bead as described above; the tin analogue complex 16 was then synthesized by treating the dried beads with 2 equivalents of tin(II) chloride in dichloromethane. Evaporation of solvent gives a system with a water content of just 1.4%.A colour change (yellow to orange) is observed but no direct analytical evidence of the formation of the platinum–tin complex could be obtained so that inferences on structure were made from the activity of the derived catalyst.Both catalysts showed similar activity in the hydroformylation reaction yet exhibit a drastic diVerence in regioselectivity. Complex 15 under identical reaction conditions gives an n :b ratio of 3.0 : 1 whilst the tin modified catalyst 16 gives an n:b ratio of 11.5 : 1 (consistent with the results obtained in the homogeneous analogue of the reaction), Scheme 4. In order to try to make the catalyst more remote from the anchoring aqueous layer ligands with varying alkyl chains separating the phosphine from the anchoring hydrophilic groups were employed.13 In contrast to TPPTS, the platinum complex 17 formed from the sulfonate anchored alkyl ligands is formed as the trans complex as is usual for monodentate ligands, although this detail may be insignificant due to the equilibrium in which the catalytic intermediates exist.The Pt–Sn hydroformylation catalyst was prepared as before and showed similar selectivities to that of the analogous homogeneous reaction, although in contrast to the reactions above it shows greatly reduced activity.Again, no leaching is observed from this catalyst. Asymmetric hydroformylation with an enantiomerically pure platinum catalyst 18 has been demonstrated by Tóth et al.13 although only with modest enantioselectivity. 5 Asymmetric hydrogenation Wan and Davis 19 then went on to explore the possibility of applying glass bead catalysts to asymmetric hydrogenation reactions. The polar ruthenium complex [Ru(C6H6)(BINAP- 4SO3Na)Cl]Cl was examined in homogeneous, heterogeneous and glass bead cases.The chosen substrate for the studies was 2- (6-methoxy-2-naphthyl)acrylic acid 19, whose hydrogenation product is the important non-steroidal anti-inflammatory drug, naproxen 20 (Scheme 5). As expected the homogeneous case proved to be the most active and gave high enantioselectivities, reflecting the intimacy of mixing of catalyst and substrate in this case.The homogeneous reaction carried out in methanol–water (1 : 1) showed high activity but diminished enantioselectivity when compared to the same reaction in methanol (77.6 vs. 87.5% enantiomeric excess, e.e.) and is thought to reflect the crucial role of the easily hydrolysed chloro ligand in maximising enantioselectivity. Addition of triethylamine was also observed to have a beneficial eVect on enantioselectivity. The dry supported phase catalyst is inactive in the reaction, yet when water is added to the reaction to hydrate the catalyst, activities fifty times greater than those of the biphasic reaction and only seven times slower than the homogeneous system are observed. The supported aqueous phase reaction was carried out with ethyl acetate as the bulk solvent and the extent of catalyst hydration is limited by the solubility of water in this solvent.Increased loadings of water in the system are accompanied by increased activity and enantioselectivity though this still falls short of the selectivity of the heterogeneous system due to aquation of the Ru–Cl bond.Improvements in enantioselectivity are possible by decreasing the temperature of reaction but at severe cost to the turnover frequency. This report demonstrates the practical approach of glass bead technology to enantioselective hydrogenation limited only by the intrinsic selectivity of the reaction in water. Further development of this reaction came in the form of substitution of water as the supported solvent by ethylene Scheme 4 (i) Platinum catalyst on CPG-350 (3 g), toluene, H2–CO (1 : 1), 69 atm, 100 8C, 120 h.Scheme 5 (i) [Ru(C6H6)(BINAP-4SO3Na)Cl]Cl on CPG-240, ethyl acetate, 95 atm, H2, 3 8C.3534 J. Chem. Soc., Dalton Trans., 1998, 3529–3538 glycol,20,21 a polar solvent immiscible with most organic solvents. The 31P NMR spectrum of a solution of the catalyst in ethylene glycol–methanol (1 : 1) is identical to that in methanol (a double doublet) indicating the integrity of the catalyst, [Ru(C6H6)(BINAP-4SO3Na)Cl]Cl, whereas addition of water leads to production of the singlet due to scission of the Ru–Cl bond.By taking the dry catalyst prepared by standard methods (see below) and stirring in a solution of ethylene glycol in ethyl acetate the solvated supported catalyst is formed. Once again the catalytic activity is seen to be dependent on the amount of solvent on the bead, crucially in this case the enantioselectivity matches that of the homogeneous case whilst the activity observed is three times greater than those of the previously tested supported aqueous phase catalysts (as is also reflected by a 50 fold increase in activity of the biphasic case between ethylene glycol and water). However this method of in situ preparation is also accompanied by leaching of the ruthenium into the bulk organic probably stemming from the greater relative solubility of ethylene glycol in the bulk when compared to water.An alternative preparation of the active catalyst which overcomes this problem involves separation of the ethylene glycol impregnated beads from the bulk by filtration and vacuum drying in two cycles. The bulk organic solvent is then saturated with ethylene glycol, mass transfer between the two phases is thus minimised and no leaching is observed. Thus Davis realised the practical preparation of a recyclable heterogeneous catalyst for asymmetric hydrogenation displaying high activity and enantioselectivity directly comparable to that of its homogeneous counterpart (88.4 vs. 88.2% at r.t., increasing to 95.7% at 3 8C). Additionally, Tóth et al.13,33 synthesized and utilised rhodium complexes of water soluble derivatives of enantiomerically pure bidentate phosphine ligands for asymmetric hydrogenation. 6 Hydrogenation of ·,‚-unsaturated aldehydes A Rhône-Poulenc group investigated the hydrogenation of a,b-unsaturated aldehydes to allylic alcohol with supported aqueous phase catalysts formed from silica (Merck 60H), [RuCl2(TPPTS)3] and [RuH2(TPPTS)4].22 The non-supported solid catalyst showed little activity in the reaction in contrast to the supported catalysts for which conversions of 100% together with selectivity of 88.5% were obtained in the hydrogenation of 3-methylbut-2-enal 21 to give the product 3-methylbut-2-enol 22 (Scheme 6).It was possible to recycle the catalyst and successfully maintain the selectivity of reaction but yields are seen gradually to diminish. The reduced catalytic activity was attributed to poisoning of the catalyst by reactant and product whose presence on the beads after use was evidenced by IR spectroscopy and microanalysis; no mention was made of phosphine oxidation in this report although it has been cited as the limiting factor in recycling in other supported aqueous phase catalysts (see above).Use of a polar solvent in the reaction (methanol) led to reduced activity when compared to hexane and leaching of catalyst whilst an attempt to use the substrate as reaction solvent gave good selectivities but low yields. 7 Wacker oxidation The Wacker oxidation is a palladium/copper catalysed partial Scheme 6 (i) [RuH2(TPPTS)2] on Merck 60H silica, hexane, 99 atm H2, 50 8C, 3 h. oxidation of an alkene.34 One feature of this reaction is that although the oxidant is molecular oxygen the incorporated oxygen comes from the water,35 hence the reaction is usually carried out in aqueous solvent which retards the oxidation of sparingly soluble higher alkenes, a limitation which may be overcome by the application of supported aqueous phase catalysis.A supported aqueous phase catalyst was prepared by deposition of a 1 : 1 ratio of the metals as their respective dichlorides from aqueous solution.23 In the oxidation reaction the catalyst is first hydrated by either direct treatment with water or water transfer from controlled pore glass beads of known water content; this latter method proves superior than the hydration process because it appears to be less disruptive to the catalyst surface.The oxidation of hept-1-ene 23 into heptanone 24 (Scheme 7) was chosen for the study, with temperatures of 100–130 8C and varying pressure of oxygen. Optimum activities were obtained with much greater water content than the hydroformylations above, probably reflecting the fact that there are no hydrophilic ligands employed in the Wacker chemistry.Conversions of hept-1-ene under optimum conditions were found to be of the order of 25% with significant isomerisation to hept-2- and -3-ene taking place during the course of the reaction. A high dependence on oxygen pressure was also noted, which is in contrast to the homogeneous system and was proposed to reflect the increased diYculty of palladium reoxidation by the Cu21/Cu1/O2 couple in the supported phase case. At temperatures above 100 8C the conversion and selectivity rapidly fall oV reflecting the domination of isomerisation at higher temperatures; again this problem is much more pronounced for the supported aqueous phase system when compared to the homogeneous analogue.Importantly an increase in palladium content of the catalyst is accompanied by increased conversion of heptene into product. No leaching of copper or palladium was observed down to the detection limits of 1 and 2 ppm respectively. 8 Palladium catalysed Heck reactions Our own research eVorts at the University of Bath have centred around applications of palladium catalysed reactions using glass bead technology. The Heck reaction is synthetically important, and has previously attracted interest in the development of a heterogeneous variant. One such example uses modified silica supports impregnated with various palladium catalysts to carry out the Heck reaction of iodobenzene 25 with methyl acrylate 26 to aVord the coupled product methyl cinnamate 27.36 Another method uses palladium supported on porous glass tubing as a catalyst (without ligand) for Heck coupling reactions.37 Additionally, there is a significant body of knowledge on the use of polar catalysts for aqueous phase Heck reactions.38 Palladium catalysed coupling reactions utilising water-soluble ligands in aqueous media have been reviewed.39,40 Dibowski and Schmidtchen 41 investigated the synthesis and use of various guanidinium phosphines in aqueous Heck reactions.Other Heck reactions have involved the coupling of halogenoarenes to ethylene.42 Simple palladium coupling reactions catalysed by mono- and tri-sulfonated triphenylphosphine palladium complexes have been carried out at the University of Bath using glass bead Scheme 7 (i) PdCl2–CuCl2 on CPG-240, hexane, 4 atm O2, 3 h, 100 8C.J. Chem. Soc., Dalton Trans., 1998, 3529–3538 3535 Table 2 Preparation of methyl cinnamate 27 using the Heck reaction and glass bead technology Beads CPG 239 Å CPG 239 Å CPG 239 Å CPG 239 Åa CPG 290 Åb CPG 120 Å CPG 500 Å Davisil 300 Å Ligand TPPMS TPPMS TPPTS TPPTS TPPMS TPPMS TPPMS TPPMS Catalyst PdCl2 Pd(OAc)2 PdCl2 PdCl2 PdCl2 PdCl2 PdCl2 PdCl2 t/h 25 25 25 25 48 23 24 21 Yield (%) 71 63 75 27 61 74 c 59 c 73 c Palladium leaching (ppm) 0.5 0.5 0.2 0.4 0.4 2.5 8.4 2.5 a The fourth run using the same prepared catalyst.b Reaction performed at room temperature. c For these reactions only, the quoted figures are conversions rather than isolated yields. These reactions were performed using 5 mol % Pd, at a concentration of 1.9 mmol in 5 ml. technology. The constitution of the glass bead/palladium catalyst was found to have a highly significant eVect on reaction rates, as well as on the amount of palladium leaching (see below). A general purpose catalyst could be prepared as follows. 24 The palladium catalyst (based on chloride or acetate) was treated with 2.2 equivalents of the polar ligand TPPTS in the minimum amount of ethylene glycol.43 After heating to ensure complexation, controlled pore glass beads were added as well as additional ethylene glycol.44 After stirring, to facilitate an even coating of the palladium complex onto the beads, they were cooled and freeze dried.45 The freeze dried catalyst is stable and can be resolvated by dosing with ethylene glycol prior to use in reactions. Addition of 10% w/w ethylene glycol to glass beads provided a robust and active catalyst (Scheme 8).The prepared catalysts derived from either palladium chloride or acetate were examined for their ability to catalyse the Heck reaction. Thus, iodobenzene and methyl acrylate were treated with the prepared palladium catalyst (1 mol % based on palladium) in the presence of triethylamine in hexane– diethyl ether (4 : 1) as solvent (Scheme 9). At the end of the reaction the solution was decanted from the glass beads and analysed for leaching of palladium.46 The basic palladium catalysed Heck reaction using glass bead technology was optimised using diVerent glass beads, ligands, and the results are summarised in Table 2.It was observed that the use of TPPTS rather than TPPMS (which is the analogous monosulfonated triphenylphosphine ligand) in Scheme 8 Scheme 9 (i) mol % prepared catalyst, Et3N, hexane–diethylether, (4 : 1), reflux.the catalyst complex resulted in less leaching of the palladium complex from the beads after performing the simple Heck reaction as described above and a greater yield. This is probably attributed to the stronger hydrophilic nature of the TPPTS palladium complex; TPPTS reactivates the catalyst and an excess of ligand stabilises the system against decomposition.39 The glass bead catalysts do work in subsequent reactions but a reduction in yield is observed, which is thought to be due to oxidation of phosphines rendering a less competent catalyst.This was proved by disintegration of the complex by dissolving it in methanol after the third reaction. A phosphorus NMR analysis was performed, and the results showed a greater level of phosphine oxide compared with the original ligand used in the initial experiment. After catalyst optimisation, the prepared catalysts derived from either palladium chloride or acetate were tested to see if they could catalyse other Heck reactions.24 Thus, bromobenzene 28 and bromoiodobenzene 29 also underwent coupling reactions with methyl acrylate to produce the cinnamate products 27 and 30 in moderate yields with low levels of palladium leaching (Scheme 10).The coupling of iodobenzene 25 with the allylic alcohol 31 aVorded the ketone 32.47 In all cases above the catalyst was shown eVectively to catalyse the Heck reactions and consistently low levels of leaching into the product were observed.However, this catalyst system was not deemed suitable for use with very polar substrates (such as acrylic acid), Scheme 10 (i) 1 mol % prepared catalyst, PdCl2, Et3N, hexane– diethylether (4 : 1), reflux, 46 h, 54%; (ii) 1 mol % prepared catalyst, Pd(OAc)2, Et3N, hexane–diethyl ether (4 : 1), reflux, 47 h, 59%; (iii) 1 mol % prepared catalyst, Pd(OAc)2, Et3N, hexane–diethyl ether, (4 : 1), reflux, 48 h, 88%.3536 J. Chem. Soc., Dalton Trans., 1998, 3529–3538 since they aVorded low yields of substitution and high levels of leaching, which we attributed to their aYnity for the more polar supported phase.Several procedures for the preparation of the glass bead catalyst were adopted with the aim of ensuring good assembly, distribution and immobilisation of the catalyst complex upon the surface of the beads. The eVect on solvation of the supported liquid-phase catalyst upon activity proved to be very interesting. Maximum activity was attained at a loading of around 10% wt.(= 0.1 d, theoretical film thickness = 16 Å) ethylene glycol to glass beads. At low loading (6.5 Å theoretical film thickness of ethylene glycol) no activity was observed indicating the lack of mobility of the catalyst complex within such a thin film. Activity decreases at higher levels of ethylene glycol, presumably due to diminishing interfacial area as the pores become filled. Hence our results compared favourably with those observed by Horváth.11 An alternative catalyst preparation was carried out to see if ethylene glycol (or a polar film) was an essential component of the catalyst preparation for successful catalytic activity.The ethylene glycol (polar film) is thought to bind to the hydrophilic sulfonate groups in the ligand structure and adhere to the controlled pore glass beads. An alternative glass bead catalyst was prepared as follows: palladium chloride was treated with 2.2 equivalents of the polar ligand (TPPTS) and heated in methanol to dissolve.After 1 h the solution had changed from a brown to a green coloration. After complexation, Davisil beads were introduced and stirred at room temperature to ensure an even coating of beads onto the organometallic complex.48 The solvent was removed in vacuo and the isolated green powder dried in an oven (Scheme 11). The catalyst must contain only the minimum amount of solvent, and yet is still active for Heck reactions, albeit with a slightly higher level of palladium leaching. In the standard conversion of iodobenzene 25 and methyl acrylate 26 to give the Heck product 27 the ethylene glycol-free beads aVorded 71% yield and 2.3 ppm palladium leaching.Preliminary results suggest that this new catalyst is more eVective when polar substrates are being employed, since the Heck chemistry works well with acrylic acid as one of the coupling partners. 9 Palladium catalysed allylic substitution We also investigated the use of glass bead technology for palladium catalysed allylic substitution reactions.24,25 Allylic substitution reactions have previously been reported using a two phase system with a palladium catalyst and TPPTS as ligand.49,50 We opted for a phosphazene base 33 to deprotonate dimethyl malonate 35, since we reasoned that the resultant active nucleophile would still prefer to reside in the bulk organic layer. In all cases palladium acetate and TPPTS were employed in the construction of the catalyst, which contained the same Scheme 11 level of ethylene glycol as used in the Heck reactions (d = 0.1).The allyl acetates 34, 37, and 39 were successfully converted into the corresponding allylic substitution products 36, 38, and 40 with reasonable yields and with consistently low levels of palladium leaching (Scheme 12). 10 Palladium catalysed Suzuki couplings Palladium catalysed cross-coupling reactions of organoboron compounds has been widely published.51,52 Such coupling reactions have been carried out in aqueous media too.39 Genét et al.53 have reported the use of the Suzuki coupling reaction to synthesize functionalised dienes using a palladium water soluble catalyst.We decided to test the new catalyst in several Suzuki coupling reactions to show the reactions work in the absence of a polar film.54,55 Aryl bromides 28, 43, and 45 were coupled with phenylboronic acid 41 to give the corresponding biphenyls 42, 44, and 46 in good yields and with low levels of palladium leaching into the organic layer (Scheme 13).Surprisingly, the presence of aqueous sodium carbonate does not increase leaching significantly, although it is reasonable to assume that the beads have been hydrated. Further investigations into the exact nature of this catalytic system are currently underway. 11 Glass beads as sponges for transition metals The ability to restrict the movement of a transition metal catalyst in a given reaction is undoubtedly of great importance.However, we felt that the utility of glass bead technology would be further enhanced if we could remove transition metals from solutions at the end of a reaction using hydrophilic ligands supported in a polar phase on the glass bead. This work originated when trying to carry out palladium catalysed [3,3] sigmatropic allylic rearrangement reactions, such as the conversion of the acetate 47 into the rearranged product 48 (Scheme 14).56,57 Scheme 12 (i) mol % prepared catalyst, phosphazene base, toluene, reflux, 6 h, 59%; (ii) 1 mol % prepared catalyst, phosphazene base, toluene, reflux, 5 d, 51%; (iii) 1 mol % prepared catalyst, phosphazene base, toluene, reflux, 48 h, 88%.J.Chem. Soc., Dalton Trans., 1998, 3529–3538 3537 We were having little success in applying the originally prepared glass bead catalysts to such reactions. The presence of a phosphine ligand dramatically inhibits the reaction, presumably by blocking the co-ordination sites for the incoming alkene.We had also tried various other ways around the problem such as making polar equivalents of the nitrile ligand. However, with little achievement the idea that we might be able to introduce the glass beads as a sponge to “mop-up” any palladium in solution became more appealing. The sponge beads were synthesized simply by mixing the polar ligand and Davisil beads (500 Å) in the minimum amount of ethylene glycol to form the palladium sponge (Scheme 15).The resulting glass bead sponge is a free-flowing powder which is easy to add to reaction mixtures and subsequently filter. Very little work has been carried out on removal of palladium from reaction solvents. Palladium on charcoal is a common reagent used in removal of tetrakis(triphenylphosphine) palladium, however a substantial amount is usually required. Feng et al.58 used mesoporous silica materials containing functionalised organic monolayers (of propylsilane) eY- Scheme 13 (i) mol % prepared catalyst, 2 M Na2CO3, toluene, reflux, 5 h, 87%; (ii) 1 mol % prepared catalyst, 2 M Na2CO3, toluene, reflux, 7 h, 95%; (iii) 1 mol % prepared catalyst, 2 M Na2CO3, toluene, reflux, 4 h, 86%.Scheme 14 (i) 1 mol % [PdCl2(CH3CN)2], THF, r.t, 24 h, 99.5% conversion. Scheme 15 ciently to remove mercury and other heavy metals from both aqueous and non-aqueous waste streams. Degussa 59 has synthesized a metal-absorbing resin called Deloxan THP II which contains organofunctional polysiloxanes bearing thiourea, sulfanyl) or thioether groups to recover Rh, Pd, Pt, Ir and Ru from highly diluted product or waste streams.Using the allylic rearrangement reaction as a model example, we took the reaction product, removed solvent, redissolved in a less polar organic solvent, and added 1 equivalent of beads with respect to substrate. The results were significant as the palladium level dropped to 0.05 ppm, and visually the yellow solution rapidly became colourless.The sponge beads were also added to the reaction solution of a Heck product, and despite the presence of a non-polar phosphine ligand (triphenylphosphine) the palladium levels very significantly lowered, such that using 100% w/w equivalents of beads with respect to substrate no palladium was detected in the product (Schemes 16 and 17). Furthermore, we have filtered a solution of palladium acetate in toluene through a filter pad of the sponge beads. The glass bead sponges were able to remove nearly all of the palladium from such a solution in just one pass of it over the beads.Thus a solution of palladium acetate (50 mg) in toluene (10 ml) was passed over the glass bead sponges. Using 1 g of the glass bead sponge (which contains 5 equivalents of TPPTS) only 0.4% of the original palladium was present in the filtrate. Using 4 g of the glass bead sponge the level was reduced to just 0.1% of the original palladium present in the filtrate.This methodology may have potential for use in extracting other metal contaminants after the metal catalysed reaction has gone to completion. 12 Outlook Glass bead technology has been successfully applied to many transition metal catalysed processes. The reactions retain much of the selectivity of their homogeneous counterparts, but retain the catalyst in a separate phase from the bulk reaction. Low levels of transition metal leaching are generally observed, which represents an economically sound and environmentally friendly approach to transition metal catalysed reactions.Many transition metal catalysed reactions have still to be examined using glass bead technology, but the outlook is promising. Perhaps more “academic” reactions can now be explored in scale-up and process chemistry using these methods. 13 References 1 W. A. Herrmann and B. Cornils, Angew. Chem., Int.Ed. Engl., 1997, 36, 1048. 2 E. Kuntz, Chemtech, 1987, 17, 570. Scheme 16 (i) 1 mol % PdCl2(CH3CN)2, THF, r.t, 24 h, 99.5%, conversion; (ii) removal of THF, hexane–diethyl ether (2 : 1); (iii) palladium sponge beads, stirring, r.t. 10 min. Scheme 17 (i) 1 mol % PdCl2, 2 mol % PPh3, Et3N, toluene, reflux, 3 h, 71%; (ii) palladium sponge beads, stirring, r.t, 10 min.3538 J. Chem. Soc., Dalton Trans., 1998, 3529–3538 3 W. A. Herrmann and C. W. Kohlpaintner, Angew. Chem., Int.Ed. Engl., 1993, 32, 1524. 4 M. E. Davis, Chemtech, 1992, 498. 5 F. R. Hartley, in Supported Metal Complexes, eds. R. Ugo and B. R. James, Reidel, New York, 1985. 6 J. P. Arhancet, M. E. Davis, J. S. Merola and B. E. Hanson, Nature (London), 1989, 339, 454. 7 R. Z. Moravec, W. T. Schelling and C. F. Oldershaw, Br. Pat., 511556, 1939. 8 J. P. Arhancet, M. E. Davis, J. S. Merola and B. E. Hanson, J. Catal., 1990, 121, 327. 9 J. P. Arhancet, M. E. Davis and B. E. Hanson, J.Catal., 1991, 129, 94. 10 J. P. Arhancet, M. E. Davis and B. E. Hanson, J. Catal., 1991, 129, 100. 11 I. T. Horváth, Catal. Lett., 1990, 6, 43. 12 Y. Yuan, J. Xu, H. Zhang and K. Tsai, Catal. Lett., 1994, 29, 387. 13 I. Tóth, I. Guo and B. E. Hanson, J. Mol. Catal. A: Chemical, 1997, 116, 217. 14 G. Frémy, E. Monflier, J.-F. Carpentier, Y. Castanet and A. Mortreux, J. Catal., 1996, 162, 339. 15 G. Frémy, E. Monflier, J.-F. Carpentier, Y. Castanet and A. Mortreux, Angew. Chem., Int.Ed. Engl., 1995, 34, 1474. 16 M. J. Naughton and R. S. Drago, J. Catal., 1995, 155, 383. 17 I. Guo, B. E. Hanson, I. Tóth and M. E. Davis, J. Organomet. Chem., 1991, 403, 221. 18 I. Guo, B. E. Hanson, I. Tóth and M. E. Davis, J. Mol. Catal., 1991, 70, 363. 19 K. T. Wan and M. E. Davis, J. Catal., 1994, 148, 1. 20 K. T. Wan and M. E. Davis, Nature (London), 1994, 370, 449. 21 K. T. Wan and M. E. Davis, J. Catal., 1995, 152, 25. 22 E. Fache, C. Mercier, N. Pagnier, B. Despeyroux and P.Panster, J. Mol. Catal., 1993, 79, 117. 23 J. P. Arhancet, M. E. Davis and B. E. Hanson, Catal. Lett., 1991, 11, 129. 24 L. Tonks, M. S. Anson, K. Hellgardt, A. R. Mirza, D. F. Thompson and J. M. J. Williams, Tetrahedron Lett., 1997, 38, 4319. 25 During the course of this work, the use of silica-supported aqueous phase palladium catalysed allylic substitution has been reported: P. Schneider, F. Quignard, A. Choplin and D. Sinou, New. J. Chem., 1996, 20, 545; S.dos Santos, Y. Tong, F. Quignard, A. Choplin, D. Sinou and J. P. Dutasta, Organometallics, 1998, 17, 78. 26 E. Kuntz, US Pat., 4 248 802, 1981. 27 B. B. Bunn, T. Bartik, B. Bartik, W. R. Bedout, T. E. Glass and B. E. Hanson, J. Mol. Catal., 1994, 94, 157. 28 I. T. Horváth, R. V. Kastrup, A. A. Oswald and E. J. Mozeleski, Catal. Lett., 1989, 2, 85. 29 A. R. Sanger, J. Mol. Catal., 1977, 3, 221. 30 C. V. Pittman, J. R. Hiraio and A. Hiraio, J. Org. Chem., 1978, 43, 640. 31 L. H.Slaugh and R. D. Mullineaux, J. Organomet. Chem., 1968, 13, 469. 32 C. Botteghi, S. Paganelli, U. Matteoli, A. Scrivanti, R. Ciorcaro and L. G. Venanzi, Helv. Chim. Acta, 1990, 73, 284. 33 I. Tóth, B. E. Hanson, I. Guo and M. E. Davis, Organometallics, 1993, 12, 848. 34 G. Smidt, W. Hafner, R. Jira, J. Sedlmeier, R. Sieber, Rultinger and H. Koger, Angew. Chem., 1959, 71, 176. 35 J. E. Bäckvall, B. Åkermark and S. O. Ljunggren, J. Am. Chem. Soc., 1979, 107, 2411. 36 J. Kiviaho, T. Hanaoka, Y. Kubota and Y. Sugi, J. Mol. Catal., 1995, 101, 25. 37 J. Li, A. W.-H. Mau and C. R. Strauss, Chem. Commun., 1997, 1275. 38 B. Cornils and E. Hintz, J. Organomet. Chem., 1995, 502, 177. 39 J. P. Genêt, E. Blart and M. Savignac, Synlett., 1992, 715. 40 N. A. Bumagin, P. G. More and I. P. Beletskaya, J. Organomet. Chem., 1989, 371, 397. 41 H. Dibowski and F. P. Schmidtchen, Tetrahedron, 1995, 8, 2325. 42 J. Kiji, T. Okano and T. Hasegawa, J. Mol. Catal., 1995, 97, 73. 43 The use of fewer equivalents of ligand aVorded higher levels of leaching. 44 CPG beads are commercially available: Cambio, 34 Newnham Road, Cambridge, UK CB3 9ET (postmaster@cambio.demon. co.uk). We are grateful to GlaxoWellcome for purchasing these beads on our behalf. 45 Freeze drying was performed on an Edwards Modulyo Pirani 10 freezedrier. 46 Leaching of palladium was determined by atomic absorption or ICP analysis. For atomic absorption analysis the decanted samples were filtered and furnaced at 400 8C, extracted into 3 drops of aqua regia and 3 ml deionised water added. The samples were analysed on a Perkin-Elmer 1100B atomic absorption instrument in conjunction with a series of palladium standards. For inductively coupled plasma (ICP) analysis the decanted samples were filtered and furnaced at 400 8C, and extracted into 3 ml dmso–2% HCl solution. The samples were analysed on a Varian Liberty 200 ICP-AES instrument with a series of palladium standards. 47 J. B. Melpolder and R. F. Heck, J. Org. Chem., 1976, 41, 265. 48 Davisil 500 Å beads were employed rather than CPG beads as both sets of beads gave comparable yields and palladium leaching. The Davisil beads are more economically viable, and can be purchased from Aldrich Chemicals. 49 M. Safi and D. Sinou, Tetrhedron Lett., 1991, 32, 2025. 50 E. Blart, J. P. Genêt, M. Safi, M. Savignac and D. Sinou, Tetrahedron, 1994, 50, 505. 51 N. Miyaura and A. Suzuki, Chem. Rev., 1995, 95, 2452. 52 M. Moreno-Manas, M. Perez and R. Pleixats, J. Org. Chem., 1996, 61, 2346. 53 J. P. Genét, A. Linquist, E. Blart, V. Mouriès, M. Savignac and M. Vaultier, Tetrahedron Lett., 1995, 36, 1443. 54 N. Miyuara, T. Yanagi and A. Suzuki, Synth. Commun., 1981, 513. 55 A. R. Martin and Y. Yang, Acta Chem. Scand., 1993, 47, 221. 56 L. E. Overman, Angew. Chem., Int. Ed. Engl., 1984, 23, 579. 57 L. E. Overman and F. M. Knoll, Tetrahedron Lett., 1979, 321. 58 X. Feng, G. E. Fryxell, L. Q. Wang, A. Y. Kim, J. Liu, K. M. Kemner, Science, 1997, 276, 923. 59 Available from Degussa Limited, Winterton House, Winterton Way, Macclesfield, Cheshire, UK SK11 0LP. Paper 8/02593B

ISSN:1477-9226    

DOI:10.1039/a802593b

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, 3539–3541 3539 The metal directed assembly of a trinuclear macrocyclic copper(II) complex Paul V. Bernhardt * and Elizabeth J. Hayes Department of Chemistry, University of Queensland, Brisbane 4072, Australia Received 21st August 1998, Accepted 17th September 1998 A trinuclear macrocyclic complex is reported from the metal directed condensation between melamine, formaldehyde and the CuII complex of a linear tetraamine. Metal directed condensation reactions of coordinated primary amines with aldehydes and dibasic acids such as nitroalkanes or primary amines oVer eVective and facile routes to pendentarmed macrocyclic ligands and their complexes.1 In each case, a new six-membered chelate ring is formed across a pair of cis disposed primary amines, and the pendent substituents introduced via the acid ‘locking fragment’ reside at the apex of this chelate ring.The majority of examples of this type of chemistry involve the reaction of CuII polyamines with nitroethane and formaldehyde,2 but other metal ions,3 aldehydes 4 and acids 5 have been employed successfully.Mononuclear complexes have almost invariably been the targets. There is continual interest in viable synthetic routes to oligonuclear complexes by virtue of the unusual cooperative magnetic and electronic interactions that arise when metal ions are constrained to be in proximity of one other. If a difunctional acid, bearing two sets of acidic methylene or primary amino groups is employed, then one may close two diVerent ring systems connected through the locking group.However, this approach has rarely been successful, these examples involving the locking groups propane-1,3-diamine,6 1,3,5-trinitropentane,7 1,2-dinitroethane, 1,3-dinitropropane and 1,4-dinitrobutane.8 Unsuccessful attempts to bridge two metal centres using this chemistry are known9 where mononuclear complexes have been formed via intramolecular cyclisation.Herein, we describe the first example where a trifunctional locking group has been employed in a metal directed aldehyde/amine condensation reaction to give a trinuclear complex. The aromatic triazine melamine possesses primary amino groups at the 2-, 4- and 6-positions and is ideally suited to act as a trifunctional locking group. In particular, the rigidity of melamine forbids intramolecular reactions that have been N N N N N NH2 NH2 N N N Cu N NH2 NH2 N Cu N N N N N N N N N N N N N N N N N N Cu Cu Cu H H H H H H H H H H H H H H H H H H [CuL2]2+ [CuL1]2+ [Cu3L3]6+ observed to compete with oligonuclear complex formation in more flexible difunctional analogues.Recently we reported 10 the metal-directed assembly of the mononuclear, pendentarmed macrocyclic complex [CuL1]21 using melamine as a locking fragment in conjunction with formaldehyde and the CuII complex of the linear tetraamine bis-N,N9-(2-aminoethyl)- propane-1,3-diamine (L2).We have shown that the melamine pendent group exhibits a very strong propensity for H-bonding with other molecules and ions in the solid state, which is a feature of other crystal structures containing the melamine group or fragment.11 However, it became apparent to us that the remaining two primary amino groups in [CuL1]21 might also be employed as locking fragments for other macrocyclic rings. To this end, by varying the stoichiometry of the reaction, we have successfully condensed three molecules of [CuL2]21 with melamine and formaldehyde to produce the new trinuclear macrocyclic complex [Cu3L3][ClO4]6?5H2O.† Single crystals of the complex were grown from a saturated aqueous solution of the complex, and we have determined the crystal structure of this compound.‡ A view of the complex ion [Cu3L]61 is shown in Fig. 1.The three fourteen-membered macrocyclic CuII units are linked to the triazine ‘hub’ via the aromatic amino groups.Water molecules and/or perchlorate anions (not shown in Fig. 1) occupy the two axial sites of each metal centres at Cu–O distances of ca. 2.5 Å. The configuration of each set of four N-donors is RSSR (trans-III), which is the most commonly observed N-based isomer in fourteen-membered macrocyclic complexes. The corresponding bond lengths and angles in each macrocyclic sub-unit are the same within experimental error. However, the relative dispositions of the rings are diVerent.Two of the macrocyclic sub-units are found on the same side of the triazine hub, while the other is on the opposite side (syn,syn,anti). Therefore, the overall molecular (but not crystallographic) symmetry is Cs. There are no significant intermolecular contacts between trinuclear complex units in the Fig. 1 View of the [Cu3L3]61 cation showing 30% probability ellipsoids. Selected bond lengths: Cu–N 1.98(2)–2.03(2) Å, N(1)–C(1), C(3) 1.33(2) Å.3540 J. Chem.Soc., Dalton Trans., 1998, 3539–3541 present crystal structure. The free primary amines of the protonated analogue [Cu(HL1)]31 are very eVective H-bond donors, and H2O-linked macrocyclic chains result through intermolecular interactions in the solid state. By comparison, these amino groups are no longer available for H-bonding in the trinuclear analogue [Cu3L]61, as they have each been incorporated into one of the macrocyclic rings. The amino N-atoms in melamine are trigonal planar, as a result of conjugation of their lone pairs with the aromatic ring, so the N(5n)–C(n) (n = 1, 2, 3) torsional angles are close to zero degrees.The conformation of the triazine-substituted sixmembered chelate ring is somewhat distorted by the inclusion of a trigonal atom at the apex, which results in the melamine ring being tilted by ca. 408 relative to each CuN4 plane. It is apparent that all three macrocyclic sub-units may be on the Fig. 2 EPR spectra of [Cu3L3][ClO4]6 (top) and [CuL1][ClO4]2 (bottom).Experimental conditions: 1 mmol dm23 solutions in DMF– H2O (1 : 2), T = 77 K, v = 9.272 GHz. Fig. 3 Square wave voltammogram of [Cu3L3][ClO4]6 (top); cyclic voltammogram of [Cu3L3][ClO4]6 (centre) and cyclic voltammogram of [CuL1][ClO4]2 (bottom). Experimental conditions: 5 mmol dm23 solutions in MeCN, 0.1 mol dm23 n-Bu4NClO4, glassy carbon working, Pt auxiliary and Au reference electrodes (potentials referenced vs. ferrocene/ferrocenium couple).same side (syn,syn,syn) of the triazine ring or one may be on the opposite side to the other two (syn,syn,anti). Molecular mechanics modeling found that the minimised strain energies of the two conformers diVered by less than 2 kJ mol21, so steric eVects evidently do not play an important role in determining the preferred conformation. Three-way dipole–dipole coupling between the metal centres in [Cu3L]61 was observed in the electron paramagnetic resonance spectrum of the complex in a DMF–H2O (1 : 2) glass at 77 K (Fig. 2). For comparison, the EPR spectrum of the mononuclear analogue [CuL1]21 is also shown. From the crystal structure analysis, the Cu ? ? ? Cu distances in the triangular array of metal centres in [Cu3L3]61 are 7.97 Å (Cu(1) ? ? ? Cu(3)), 9.54 Å (Cu(1) ? ? ? Cu(2)) and 9.36 Å (Cu(2) ? ? ? Cu(3)). These internuclear distances are within the range over which we have previously observed dipole–dipole coupling in the EPR spectra of dinuclear CuII complexes.7,12 Nevertheless, the EPR spectra of [Cu3L3]61 and [CuL1]21 are qualitatively similar; with the former exhibiting much broader peaks.This shows that metal– metal interactions in the trinuclear complex are only a minor perturbation on the basic mononuclear spectrum. Cyclic voltammetry of [Cu3L3]61 in MeCN solution revealed overlapping waves in the region 21.2 to 21.4 V vs. ferrocene/ ferrocenium (Fig. 3). Square wave voltammetry resolved these processes into one-electron and two-electron responses at 21.14 and 21.35 V vs.Fc/Fc1 respectively. For comparison, the cyclic voltammogram of [CuL1]21 is also shown, where a quasi-reversible CuII/I couple at 21.40 V vs. Fc/Fc1 is found. The positive shift in the CuII/I waves of [Cu3L3]61 relative to [CuL1]21 reflect the electrostatic metal–metal interactions that facilitate reduction to the monovalent state. In aqueous solution, all of these electrochemical responses become totally irreversible due to the instability of CuI amines in water.Magnetic coupling between the three S = ��� centres in [Cu3L3]61 should be significant on the basis of the EPR spectroscopic results, and we are currently investigating this through low temperature magnetic moment measurements. Acknowledgements We gratefully acknowledge financial support by the University of Queensland and the Australian Research Council. Mr D. Hunter is also thanked for technical assistance with the EPR measurements.Notes and references † To a refluxing solution of Cu(NO3)2?3H2O (7.24 g), L2 (5.40 g), Et3N (6.07 g), formaldehyde (6.0 cm3, 32%) was added a solution of melamine (1.53 g) in MeOH–water (1 : 1, 100 cm3). The reaction was allowed to proceed for 5 d. Column chromatography (Sephadex C-25, 0.4 mol dm23 NaClO4) separated unreacted [CuL2]21, [CuL1]21 (major) then the desired product [Cu3L3]61 well behind the mononuclear complexes. The complex precipitated on concentration of the eluate to ca. 100 cm3 (yield 5%) (Found: C, 22.3; H, 4.5; N, 16.1. [Cu3L3]- [ClO4]6?6H2O, C30H78Cl6Cu3N18O30, requires C, 22.88; H, 4.99; N, 16.02%). ‡ Crystal data: [Cu3L3][ClO4]6?5H2O, C30H76Cl6Cu3N18O29, M = 1556.41, monoclinic, space group P21 (no. 4), a = 8.643(3), b = 22.580(4), c = 15.759(5) Å, b = 95.78(2)8, U = 3057(1) Å3, Z = 2, m(Mo- Ka) = 13.93 cm21, T = 293 K, final R1 = 0.0644 for 2233 observed reflections [|Fo| > 2s|Fo|, 2q < 508], wR2 = 0.2191 for 5515 unique reflections (Rint 0.0813). The structure was solved by Patterson methods with SHELXS-8613 and refined by full matrix least squares with SHELXL- 93.14 CCDC reference number 186/1168. See http://www.rsc.org/ suppdata/dt/1998/3539/ for crystallographic files in .cif format. 1 P. V. Bernhardt and G. A. Lawrance, Coord. Chem. Rev., 1990, 104, 297. 2 P. Comba, N. F. Curtis, G. A. Lawrance, M. A. O’Leary, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1988, 2145; P.Comba, N. F. Curtis, G. A. Lawrance, A. M. Sargeson, B. W. Skelton and A. H. White, Inorg. Chem., 1986, 25, 4260. 3 N. F. Curtis, G. J. Gainsford, A. Siriwardena and D. C. Weatherburn, Aust. J. Chem., 1993, 46, 755; M. Rossignoli, P. V. Bernhardt, G. A. Lawrence and M. Maeder, J. Chem. Soc., DaltonJ. Chem. Soc., Dalton Trans., 1998, 3539–3541 3541 Trans., 1997, 323; M. Rossignoli, C. C. Allen, T. W. Hambley, G. A. Lawrence and M. Maeder, Inorg. Chem., 1996, 35, 4961. 4 P. V. Bernhardt and P. C. Sharpe, Inorg. Chem., 1998, 37, 1629; L. Fabbrizzi, M. Licchelli, A. M. Lanfredi, O. Vassalli and F. Ugozzoli, Inorg. Chem., 1996, 35, 1582. 5 Y. D. Lampeka, A. I. Prikhod’ko, A. Y. Nazarenko and E. B. Rusanov, J. Chem. Soc., Dalton Trans., 1996, 2017; P. V. Bernhardt, K. A. Byriel, C. H. L. Kennard and P. C. Sharpe, Inorg. Chem., 1996, 35, 2045. 6 S.-G. Kang, S.-K. Jung, J. K. Kweon and M.-S. Kim, Polyhedron, 1993, 12, 353. 7 P. V. Bernhardt and L. A. Jones, Chem. Commun., 1997, 655. 8 P. Comba and P. Hilfenhaus, J. Chem. Soc., Dalton Trans., 1995, 3269. 9 M. P. Suh, S.-G. Kang, V. L. Goedken and S.-H. Park, Inorg. Chem., 1991, 30, 365. 10 P. V. Bernhardt and E. J. Hayes, Inorg. Chem., 1998, 37, 4214. 11 G. M. Whitesides, E. E. Simanek, J. P. Mathias, C. T. Seto, D. N. Chin, M. Mammen and D. M. Gordon, Acc. Chem. Res., 1995, 28, 37. 12 P. V. Bernhardt, P. Comba, T. W. Hambley, S. S. Massoud and S. Stebler, Inorg. Chem., 1992, 31, 2644. 13 G. M. Sheldrick, Acta. Crystallogr., Sect. A, 1990, 46, 467. 14 G. M. Sheldrick, SHELXL-93, Program for Crystal Structure Determination, University of Göttingen, 1993. Communication 8/

ISSN:1477-9226    

DOI:10.1039/a806586a

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3543–3548 3543 The interaction of indium(III) iodide species with substituted ortho- and para-quinones Martyn A. Brown, Bruce R. McGarvey and Dennis G. Tuck Department of Chemistry and Biochemistry, University of Windsor, Windsor, Ontario, Canada N9B 3P4 Received 2nd June 1998, Accepted 25th August 1998 The interactions of substituted ortho- and para-quinones with indium(III) halides and the InI4 2 anion have been studied in non-aqueous solution.para-Quinones and InI3 give rise to stable 1 : 1 adducts, which are diamagnetic in the solid state, but which decompose in solution to form (p-sq)InI2 derivatives, where p-sq~2 is the corresponding semiquinonate. With ortho-quinones, the reaction products are (o-sq)InI2 which react with 4-methylpyridine (pic) to form (o-sq)InI2pic2. The electron spin resonance spectra of these products, and their solution chemistry, are discussed. The reactions involve intramolecular one-electron transfer, resulting in oxidation of the iodide ligand.In contrast, the reaction of 3,5-di-tert-butyl-1,2-benzoquinone with InI4 2 apparently involves intermolecular electron transfer; in this case, the products are I3 2 and the corresponding catecholate (dbc), isolated as the solid InI(dbc)pic2. The mechanisms of these various processes are discussed. The oxidation of Main Group elements, and of the lower oxidation state derivatives of such elements, has been the subject of a number of papers from this laboratory.The most significant conclusions include the identification of successive one-electron transfer processes in such redox reactions, and the confirmation of a mechanism involving nucleophilic attack by a quinone at the appropriate metal or non-metal centre. Both electron spin resonance spectroscopy and X-ray crystallography have been important experimental techniques in these investigations, and the use of these and other methods has been discussed elsewhere.1,2 We have recently carried out a series of experiments in which both ortho- and para-quinones interact with the halide of a Main Group metal in its highest oxidation state.In some cases, it has been possible to characterize crystalline adducts of such systems, and in others there is clear evidence of intramolecular electron transfer within the complex. The present paper reports studies of the interaction of a range of ortho- and paraquinones with neutral and anionic halide derivatives of indium(III).With para-quinones, we have isolated 1 : 1 adducts of InI3, which are stable in the solid state, but which decompose in non-aqueous solution, forming paramagnetic species. ortho- Quinones react directly with InI3, to give compounds in which indium(III) is coordinated by two halides and a semiquinonate ligand. These results appear to involve a hitherto unrecognized type of intramolecular electron transfer reaction, with important implications in Main Group chemistry.We have also observed a reaction between an ortho-quinone and the InI4 2 anion, and here the course of the electron transfer is apparently through an intermolecular attack of the quinone on the ligand, a process which is similar to that found in reactions of quinones with organometallic substrates.3 The quinones studied in the present work, and the abbreviations used, are shown right. Experimental Indium halides were prepared by the thermal reaction of the elements in refluxing xylene, and the salt Et4N[InI4] was synthesized from the reaction of Et4NI and InI3 in ethyl acetate (Found: C, 13.2; H, 2.65.Calc. for C8H20InI4: C, 12.8; H, 2.65%). Substituted quinones, and organic bases, were dried over sodium hydroxide and recrystallised under nitrogen. Solvents were distilled from, and stored over, drying agents, and degassed before use. All reactions were carried out in an atmosphere of dry nitrogen, and reaction products were handled by the normal methods for air-sensitive materials.Electron spin resonance (ESR) spectra were recorded on a Bruker ESP-300E instrument operating in the X-band region, using the calibration and other techniques outlined elsewhere.4 O O O O But But O But But O O X4 X = Cl, Br p-dbbq phenanthrenequinone (pq) O – p-dbbsq• – But O o-dbbq O But But O – O – But But O O But o-dbbsq• – dbc2– •– O O O O X4 X = Cl, Br (Cl4q, Br4q) 1,2-naphthoquinone (nq)3544 J.Chem. Soc., Dalton Trans., 1998, 3543–3548 Table 1 Characterization of derivatives of InI3 and ortho-quinones Identifying Analysis (%) a o-Quinone 3,5-di-tert-butylbenzo- Tetrachlorobenzo- Tetrabromobenzo- 1,2-naphtho- Phenanthrene- Compound (sq)InI2 (sq)InI2pic2 (sq)InI2 (sq)InI2pic2 (sq)InI2 (sq)InI2pic2 (sq)InI2 b (sq)InI2pic2 (sq)InI2 c (sq)InI2pic2 number 1 1A 2 2A 3 3A 4 4A 5 5A Yield (%) 75 65 55 45 38 26 53 42 65 58 Colour Purple Brown Red-orange Red Red Orange Dark blue Dark green Green Brown C 28.5 (28.5) 40.0 (40.3) 12.2 (11.7) 27.6 (27.0) 9.50 (9.10) 22.2 (22.1) 22.9 (22.8) 37.2 (37.0) 28.1 (29.1) 41.6 (40.9) H 3.35 (3.40) 4.30 (4.40) 0.14 (0) 1.80 (1.75) 0.02 (0) 1.80 (1.45) 1.35 (1.15) 2.90 (2.80) 1.70 (1.40) 3.00 (2.90) a Calculated values in parentheses.b Mass spectrum; M1 = 527 observed. c Mass spectrum; M1 = 557 observed. Microanalysis was by Canadian Analytical Services. X-Ray crystallographic studies were by the methods previously described.4,5 Mass spectra were recorded on a Shimadzu 14-B instrument operating in the EI mode, with Sun Sparc software.In this paper, we follow the previous practice of identifying the triad of quinone, semiquinone and catecholate by the abbreviations q, sq~2, and cat22, with appropriate prefixes. Preparative studies In a typical experiment with ortho-quinones, a solution of the quinone in tetrahydrofuran (5 mmol in 20 cm3) was added to a stirred solution of InI3 (5 mmol) in the same solvent.In every case, a colour change was observed, with the fastest reaction being with di-tert-butyl-o-benzoquinone. Samples were removed for ESR analysis, and approximately half of the remaining mixture cooled to 0 8C; this led to the precipitation of coloured solids which were shown by analysis to be the corresponding (sq)InI2 derivatives (see Table 1 for analytical results and related experimental data). A six-fold excess of picoline (pic; 4-methylpyridine) was added to the remaining reaction mixture, the volume of the solution reduced by 50%, and the residue cooled to 0 8C; the crystals which formed were identified analytically as the bis(picoline) adducts (sq)InI2pic2.In the case where sq~2 = 3,5- di-tert-butyl-o-benzosemiquinonate (dbbsq) an X-ray crystallographic study showed that this compound was structurally identical with the material formed in the reaction between indium(II) iodide and the corresponding ortho-quinone.5 (Unit cell dimensions, a = 13.0097(5), b = 13.302(4), c = 10.811(5) Å, a = 97.677(4), b = 107.992(3), g = 104.003(4)8, U = 1681.9(6) Å3.Found5 for (dbbsq)InI2(pic)2, a = 13.013(3), b = 13.317(3), c = 10.828(5) Å, a = 97.71(3), b = 107.98(3), g = 103.92(3)8, U = 1684.8(1.2) Å3). We also attempted to prepare adducts with pyridine as the neutral donor, but in each case the analytical results were less than satisfactory, except in the case of phenanthrenequinone (Found: C, 39.4; H, 2.93. Calc.for (psq)InI2py2, C24H18O2N2InI2: C, 39.2; H, 2.47%), and this aspect of the work was not continued. A similar sequence of reactions starting with InCl3 gave the analogous (dbbsq)InCl2pic2 (Found: C, 52.7; H, 6.09. Calc. for C26H24O2N2Cl2In: C, 52.7; H, 5.74%) In each case, the infrared spectra of both the initial products and the picoline adducts showed that n(C]] O) of the o-quinones, in the region 1650–1695 cm21, had disappeared and was replaced by n(C–O) modes at 1440–1490 cm21 (cf.ref. 1). When para-quinones were used in essentially identical experiments, the products were the 1 : 1 adducts, q?InI3, which were isolated and analyzed (Table 2). It was not possible to obtain material of suitable quality for X-ray crystallography from these experiments. Studies with InI4 2 A rapid reaction, identified by a colour change, occurred when millimolar quantities of Et4N[InI4] and dbbq in tetrahydrofuran were mixed together at room temperature.The resultant brown solution was divided into two equal portions. Evaporation of one of these to 50%, followed by cooling, gave a dark brown powder, which was identified analytically as Et4NI3 (Found: C, 18.5; H, 3.82. Calc. for C8H20NI3: C, 18.8; H 3.91%). 13C NMR (Me4Si = 0): d 7.995 CH3, 32.286 CH2. IR n(C–H) 3009–2908, n(C–N) 1585, 1443, 1416 cm21. Yield 100%, based on the initial quantity of cation. Samples of the remaining portion were used for ESR studies (see below).Addition of picoline to this solution gave a colourless precipitate, identified as InI(dbc)(pic)2 (Found: C, 49.5; H, 5.25. Calc. for C26H34N2O2InI: C, 48.2; H, 5.25%). The infrared spectrum confirmed the presence of catecholate, with n(C–O) at 1463, 1433 and 1415 cm21. Similar experiments were attempted using toluene as the reaction medium, but these were unsatisfactory because of the insolubility of the InI4 2 salt, although the general course of the reaction appeared to be similar to that described above.Results and discussion Reactions with ortho-quinones The results in Table 1 for compounds numbered 1–5 show that the reaction of ortho-quinones with InI3 produces (sq)InI2 species, which can be obtained as insoluble solids at 0 8C. The molecularity of these compounds is not known, but treatment with excess picoline yields the bis-adduct, which in the case of (dbbsq)InI2pic2 was shown to be crystallographically identical with the mononuclear six-coordinate compound obtained from the reaction the reaction of dbbq and In2I4.5 The initial overall reaction is dbbq 1 InI3 æÆ (dbbsq)InI2 1 I? (1) followed by 2I? æÆ I2 (2) Table 2 Characterization of adducts of InI3 and para-quinones Yield Analysis (%) a p-Quinone Benzo 2,6-Di-tert-butylbenzo b,c Tetrachlorobenzo Tetrabromobenzo 1,4-Naphtho (%) 80 73 26 35 10 Colour Grey Red Yellow Orange Brown C — (11.9) 23.7 (23.5) 9.70 (10.1) 7.94 (7.83) 15.5 (15.2) H — (0.67) 2.95 (2.80) 0.04 (0) 0.03 (0) 1.08 (0.95) a Calculated values in parentheses.b For 1 : 1 adduct with InCl3, orange, C 38.0 (38.1), H 4.10 (4.53), 62% yield. c Mass spectral peaks include m/z = 496 (InI3).J. Chem. Soc., Dalton Trans., 1998, 3543–3548 3545 and (dbbsq)InI2 1 excess pic æÆ (dbbsq)InI2pic2 (3) Reactions (1) and (2) are complete within about 15 min, with the relative rates, as judged by the colour change, in the order dbbq > Br4q > CI4q > nq > pq, and InI3 > InCl3.We return to the mechanistic details of eqn. (1) below. In each case, the reaction of InI3 and o-quinone gave rise to solutions which were strongly ESR active, as would be expected if the product is a semiquinone derivative. The generation of radical species by the interaction of dbbq with indium and gallium trihalides was reported some years ago, but the reaction pathway was not apparently explored.6,7 In the present studies, the system most thoroughly explored involved InI3 1 dbbq, and Fig. 1a shows the ESR spectrum of the (diluted) solution resulting from this reaction. The addition of excess picoline caused marked changes, giving rise to a spectrum (Fig. 1b) essentially identical to that assigned previously5 to the sixcoordinate indium(III) complex (dbbsq)InI2pic2. The parameters found for Fig. 1b by simulation are g = 2.0038, AIn = 4.9 G, AH = 3.6 G (1 proton) (1 G = 0.1 mT), compared with the earlier values g = 2.00391, AIn = 4.86 G, AH = 3.42 G (1H), 0.36 G (9H): in this spectrum, and those in Fig. 1a, we did not observe any splitting by the proton on C3, as is commonly the case in derivatives of dbbsq.8 Fig. 1a can be analysed as the ESR spectrum of a mixture of two closely related indium(III) species; for one, g = 2.0032, AIn = 8.93 G, AH = 3.65 G (1H), and for the other g = 2.0032, AIn = 8.50 G, AH = 3.65 G (1H). The simulation in Fig. 1a assumes that these two species are present in equimolar proportions, and we therefore suggest that eqn.(1) is followed by 2 (dbbsq)InI2 (dbbsq)IInI2InI(dbbsq) (4) in which the dimerisation is presumed to involve iodide bridging, In–m-I2–In, similar to that reported for In2I6 in nonaqueous solution.9 Such a dimer can exist as cis and trans isomers, which explains the identification of two similar AIn Fig. 1 a, ESR spectrum of the (diluted) solution arising from the reaction of InI3 and o-dbbq, at room temperature.The upper trace is the experimental result, and the lower is the simulated spectrum, using the parameters discussed in the text. b, The same solution, after addition of a six-fold excess of picoline. values. These structures are preferred to the form (dbbsq)2- InI2InI2, which does not give two stereoisomers. The stereochemistry of these various species is an important factor in these arguments. The (dbbsq)InI2 monomer must be highly strained, since it is diYcult for a pseudo-tetrahedral molecule to accommodate the bidentate semiquinonate, given that the bite angle for this ligand is ca. 758 (cf. ref. 5). In a dimeric molecule, indium has pentagonal bipyramidal stereochemistry, and the consequent lowering of strain therefore serves to move eqn. (4) to the right. Both cis and trans dimers can be readily converted to the six-coordinate (dbbsq)InI2pic2 by excess picoline. Coordination by an electron-donating neutral ligand lessens the eVective positive charge at the metal centre, thereby weakening the interaction of the unpaired electron and lowering the hyperfine constant AIn from the relatively high value of ca. 8.6 G in the dimers to one more typical of six-coordinate indium(III) complexes. In an attempt to further characterize these dimeric species, we recorded the ESR spectra of the frozen solution (Fig. 2a), and the associated half-field resonance (Fig. 2b). Surprisingly, the simulation of these spectra identifies the low temperature species as being an S = ��� state, rather than a biradical, and the predominant complex under these conditions is therefore In(dbdsq)3. This molecule, which has been prepared independently 10 by the metathesis of InI3 and 3Na1dbbsq2, is an analogue of Ga(dbbsq)3, whose preparation and crystallographic structure were reported earlier.11 The simulation for S = ��� state assumed a spin-Hamiltonian of the form H = gBeS·H 1 AIn S· I 1 D[3 S2z 2 1 3 – S (S 1 1)] (5) with the parameters g = 2.003, AIn = 8 ± 0.5 G, and D = 103 ± 5 Fig. 2 a, ESR spectrum of frozen solution at 100 K from Fig. 1a. b, The half-field transition. In I In I O O O I O I In I In I O O O I I O • • • • cis trans3546 J. Chem. Soc., Dalton Trans., 1998, 3543–3548 G. The corresponding values for the gallium complex are g = 2.003, D = 108; the slightly lower value of D for the indium species is in keeping with the larger ionic radius of the latter element. In addition to the dimerisation processes discussed above, the solution chemistry of the (dbbsq)InI2 product of eqn.(1) must therefore also involve the equilibria InX3 InX2Y InXY2 InY3 (6) where X = I, Y = dbbsq. Such facile redistribution reactions are a known feature of the chemistry of indium(III) complexes in non-aqueous solution,9 and in the present context explain the presence of di- and tri-radical species. The relative quantity of each species will be a function of solvent, temperature, and the nature of the semiquinonate anion, and it is therefore understandable that the predominant trimer seen at 100K is not observed at room temperature.The ESR spectra of solutions produced by the reaction of o-Cl4q and o-Br4q with InI3 were weak, due to the poor solubility of the (X4sq)InI2 products, and of their picoline derivatives (2–3A, Table 1). There are also some diVerences in the case of the derivatives of 1,2-naphthoquinone; the AIn value in the presence of excess picoline is 3.6 G, in reasonable agreement with that for the dbbsq analogue, but the spectrum of the initial reaction solution shows no indium hyperfine coupling, suggesting that the species in solution are the free napththosemiquinonatInI3.In the case of phenanthrenequinone, the reaction solution gave g = 2.0028, AIn = 1–2 G, but the simulation of the picoline adduct spectrum was clearer, with AIn = 2.6 G, AH = 2.25 (1H) and 1.35 G (1H), and g = 2.0036. These results in general are in keeping with those for dbbq 1 InI3, although the naphthoquinone results emphasize the significance of the solution equilibria, and their dependence on the properties of the o-quinone involved.The main feature of these reactions is the oxidation of a ligand bonded to a metal which is in its highest oxidation state. In previous studies 5,8,12 of the reactions of substituted orthoquinones with indium-(I) and-(II) halides, we established that oxidation occurred at the metal centre, leading to the eventual production of sq~2 or cat22 derivatives of indium(III), depending on the detailed behaviour of the system.The present results show that a diVerent reaction q 1 InX3 æÆ (sq?)InX2 1 X? (7) can also occur, and we note that this parallels a known fast solution-phase reaction in the case of iodide 13 q 1 I2 æÆ sq~2 1 ��� I2 (8) While this readily explains the reactions of InI3, the reaction dbbq 1 InCl3 æÆ (dbbsq?)InCl2 1 Cl? (9) is more surprising, since Cl2 is not oxidized by ortho-quinones in aqueous solution, but it must be emphasized in this context that we are dealing here with a complex of a metal in its highest oxidation state, and that the redox behaviour of a bonded halide ligand in such a molecule will be quantitatively diVerent from that of the free anion in aqueous solution. The first step in the reaction is the coordination of the quinone to the metal centre, as is suggested by the chemistry of the para-quinone systems (see below), and confirmed by studies of complex formation between ortho-quinones and AlCl3 or SnX4 (X = Cl, Br),14 and this must be followed by intramolecular electron transfer within the bonded system.q In X e One important diVerence between this process and the solution phase reaction of eqn. (7) lies in the entropy of activation. In a series of reactions involving tetrahalogeno-p-quinones and MI (M = Na1, K) in acetone,13 DS‡ was found to be in the order of 280 J K21 mol21, but this parameter is probably close to zero for an intramolecular electron transfer in a relatively large molecule in non-aqueous solution, and to this extent, the energetics of the latter process are more favorable than for the solution reaction.The detailed implication of these principles will be the subject of future work. Reactions with para-quinones The analytical results in Table 2 show that the products of the reaction between indium(III) iodide and a series of substituted para-quinones are the 1 : 1 adducts, which are stable at room temperature in the absence of moisture.An analogous compound was also obtained with indium(III) chloride. This adduct formation is not surprising, given the known properties of the indium(III) halides,9 and it seems reasonable to assume that these are four-coordinate monomeric species in the solid state. As noted above, crystallographic investigations were not possible, but we have recorded the infrared spectra with particular reference to the carbonyl stretching mode, which is typically in the region of 1650–1690 cm21 for the parent quinones, and is sometimes observed as a doublet due to Fermi resonance.15,16 For p-dbbq, the complex has a vibration at 1655 (cf.n(C]] O) at 1655 cm21 in the parent p-quinone), and similar features are seen in the adducts of p-Br4q, p-Cl4q and 1,4-naphthoquinone. We discuss the p-C6H4O2 system below. These spectra are compatible with weak bonding of p-q to InI3.Similar results have been obtained with adducts of InI3 and cyclic ketones, for which both X-ray and infared results show only small changes in the properties of the C]] O group on coordination.14 The solid complexes show no significant ESR activity, in keeping with the above structure which implies no singleelectron transfer, but significant spectra are observed on dissolution in non-aqueous solvents. The solution spectra were recorded under a variety of conditions, and q : InI3 ratios, and all show clear evidence of a free radical coupled to indium. Most of the spectra are apparently of a mixture of species, but that shown in Fig. 3 analyses as being from an essentially single species and the spectrum was well simulated with the parameters g = 2.0065, AIn = 1.88 G, AH = 3.45 G (4H). The identifi- cation of four equivalent hydrogen nuclei requires the molecule in question to contain two p-dbbsq~2 groups, and two possible structures are shown below.The simulation results eliminate structures analogous to those proposed for the ortho-quinone system, since these would require contributions from two indium and four hydrogen atoms. Although we observed a half-field transition in the frozen solution of Fig. 3, it was not possible to deduce the D value, which might have cast light on this matter. We conclude that the solution chemistry of this 1 : 1 adduct involves as the first step an intramolecular electron transfer which is facilitated by the elimination of a halogen atom. This latter process may require solvation O O In I I I I I In I I In solv I In p-sq p-sq p-sq p-sqJ.Chem. Soc., Dalton Trans., 1998, 3543–3548 3547 (p-q)InI3 æÆ (p-sq?)InI2 1 I?(solv) (10) 2 I?(solv) æÆ I2 (11) and/or I?(solv) æÆ decomposition products (12) Eqn. (10)–(12) are obviously reactions whose rates will depend on solvent, the quinone, and temperature. The semiquinonate derivative (p-sq?)InI2 may be the starting point for redistribution processes, following eqn.(6). Dimerisation is also possible, but the presumably low steady state concentrations of (p-sq?)- InI2 argue against this, in contrast to the ortho-quinone system in which (o-sq?)InI2 is initially the predominant solute species. A reaction which is more probable than dimerisation is association with unreacted (p-q)InI3 to give (p-sq)InI–m-I–InI2(p-q), which can then be the starting point for redistribution processes.Given these possibilities, it is not surprising that the ESR spectra demonstrate the presence of a mixture of radicalcontaining complexes. The most important conclusion is that the para-quinone adducts studied are stable four-coordinate diamagnetic monomers in the solid state, decomposing in non-aqueous solution to give semiquinonate derivatives. The conclusion that intramolecular electron transfer is aVected by the phase seems counter-intuitive, but the critical factor here is the removal of halogen by eqn.(10). The diVerent solution behavior of the ortho- and para-quinone systems reflects both the redox properties of the quinones, and their diVerent coordinating properties. In particular the larger hyperfine constants for indium in the o-quinone complexes (5–8 G), compared to the p-quinones (ª2 G) shows that bidentate coordination produces a stronger interaction between ligand and metal centre. The relative strengths of the ESR signals for the para-quinone systems suggests that eqn.(10) goes to the right in the order p-dbbq > p-Cl4q > p-Br4q ª p-nq, with only very weak activity being observed in the last two systems. The absence of biradical activity in all these systems may be evidence of the stability of (p-sq)InI2 species relative to the redistribution processes. The p-C6H4O2 system diVers substantially from the others just discussed. The solid is strongly ESR-active, and the solution spectrum shows strong coupling to indium(III), but no evidence was obtained for biradical species.The infrared spectrum had no features in the n(C]] O) region. The solid slowly turns grey on standing, and we conclude that the adduct InI3(p-C6H4O2) is sensitive to both air and moisture, and that the strong ESR activity is evidence of decomposition rather than of intramolecular electron transfer, but we did not investigate these eVects in any detail. Fig. 3 ESR spectrum of a dilute solution of p-dbbq 1 InI3 (mole ratio 1 : 2) in tetrahydrofuran at room temperature.Upper trace, experimental result: lowce, simulated spectrum, using the parameters discussed in the text. Reaction with InI4 2 The previous systems, involving InI3 with o- or p-quinones, have been discussed in terms of coordination followed by intramolecular electron transfer. In the course of this work, we also found an unexpected reaction between dbbq and the InI4 2 anion.An important experimental point is that this reaction can only be studied spectroscopically by using salts of Et4N1 or some similar cation which resists oxidation by the o-quinone; initial studies with tetraphenylphosphonium salts were hindered by a reaction apparently involving this cation. The production of Et4NI3 in quantitative yield, and of (dbc)InIpic2 in the presence of picoline, shows that the overall reactions are Et4NInI4 1 dbbq æÆ Et4NI3 1 (dbc)InI (13) (dbc)InI 1 excess pic æÆ (dbc)InIpic2 (14) Eqn.(13) begs the question of mechanism, and we propose a process based on the work of Davies et al.17 for reactions such as q 1 Ph4Sn æÆ sq?(Ph?)SnPh3 æÆ (sq?)SnPh3 1 Ph? (15) which was substantiated for a range of substituted quinones. The same mechanism has been invoked for the reaction between o-quinones and Sn2Ph6,3 and for the oxidation of phenyl, as in LiPh, by o- and p-quinones to give Li1(sq~2).18,19 Equally important are that Ph4Sn has no acceptor properties,20 which eliminates the possibility of nucleophilic attack at the metal centre, and that the solution chemistry of (dbbsq)SnPh3 shows the importance of redistribution reactions subsequent to eqn.(15).3 The detailed discussion of eqn. (13) and (14) also begins with the lack of evidence 9 for any acceptor properties for InI4 2, in contrast to InCI4 2 and InBr4 2, so that the primary process is assumed to be analogous to eqn. (15) dbbq 1 InI4 2 æÆ [(dbbsq?)(I?)InI3]2 æÆ dbbsq~2 1 InI? 4 (16) followed by InI4 ? æÆ InI31I? (17) dbbsq~2 1 InI3 (dbbsq?)InI2 1 I2 (18) (dbbsq?)InI2 æÆ (dbc)InI 1 I? (19) 2I? 1 I2 æÆ I3 2 (20) This sequence explains the products identified (see Experimental section), and is compatible with the ESR spectra of the reaction solution.The results for the dbbq/InI4 2 reaction raise an ambiguity in the matter of the initial step in the reaction of o-quinones with InI3. As noted above, p-quinones form stable 1 : 1 adducts with InI3, implying that nucleophilic attack is probable in the o-quinone/InI3 system. Earlier evidence from NMR studies of InI3dppe mixtures 21 (dppe = 1,2-bis(diphenylphosphino)- ethane) show that the formation of the 1 : 1 complex dppe 1 InI3 æÆ InI3(dppe) (21) lies strongly to the right, which supports the coordinative mechanism of eqn.(1). On the other hand, InI3 in toluene is present as the dimer In2I6, in which indium is psuedotetrahedrally coordinated, as it is in InI4 2, so that a reaction sequence analogous to that in eqn.(14)–(16) can also be3548 J. Chem. Soc., Dalton Trans., 1998, 3543–3548 constructed. The presence of the dimer is not itself an argument against the coordinative mechanism, as has been shown for the ddpe case,21 and we therefore favour the model proposed, earlier. Further work on this is planned. Acknowledgements This work was supported in part by Research Grants (to B. R. M. and D. G. T.) from the Natural Sciences and Engineering Research Council of Canada.References 1 D. G. Tuck, Coord. Chem. Rev., 1993, 112, 215. 2 B. R. McGarvey, A. Ozarowski and D. G. Tuck, Inorg. Chem., 1993, 32, 4474. 3 M. A. Brown, B. R. McGarvey, A. Ozarowski and D. G. Tuck, J. Organomet. Chem., 1998, 550, 165. 4 T. A. Annan, M. A. Brown, A. A. El-Hadad, B. R. McGarvey, A. Ozarowski and D. G. Tuck, Inorg. Chim. Acta, 1994, 225, 207. 5 M. A. Brown, B. R. McGarvey, A. Ozarowski and D. G. Tuck, Inorg. Chem., 1996, 35, 1560. 6 G. A. Razuvaev, G. A. Abakumov and E. S. Klimov, Dokl. Akad. Nauk SSSR, 1971, 201, 624. 7 G. A. Abakumor and E. S. Klimov, Dokl. Akad. Nauk SSSR, 1972, 202, 827. 8 T. A. Annan, R. K. Chadha, P. Doan, D. H. McConville, B. R. McGarvey, A. Ozarowski and D. G. Tuck, Inorg. Chem., 1990, 29, 3936. 9 D. G. Tuck, Comprehensive Coordination Chemistry, eds. G. Wilkinson, R. D. Gillard and J. A. McCleverty, Pergamon Press, Oxford, 1987, vol. 3, ch. 25.2, p. 165. 10 A. A. El-Hadad, M.Sc. Thesis, University of Windsor, 1996. 11 A. Ozarowski, B. R. McGarvey, A. A. El-Hadad, Z. Tian, D. G. Tuck, D. J. Krovich and G. C. DeFotis, Inorg. Chem., 1993, 32, 841. 12 T. A. Annan and D. G. Tuck, Can. J. Chem., 1988, 66, 2935. 13 M. Sasaki, Rev. Phys. Chem. Jpn., 1996, 39, 27. 14 T. L. Brown, Spectrochim Acta, 1963, 19,1065. 15 T. Anno and A. Sado, Bull. Chem. Soc. Jpn., 1958, 31, 734. 16 M. A. Brown and D. G. Tuck, unpublished work. 17 A. G. Davies and J. A. A. Hawairi, J. Organomet. Chem., 1983, 251, 53. 18 M. A. Brown, B. R. McGarvey, H. Ozarowski and D. G. Tuck, J. Am. Chem. Soc., 1996, 118, 9691. 19 M. A. Brown, B. R. McGarvey and D. G. Tuck, J. Chem. Soc., Dalton Trans., 1998, 1371. 20 A. G. Davies and P. J. Smith, Comprehensive Organometallic Chemistry, eds. G. Wilkinson, F. G. A. Stone and E. W. Abel, Pergamon Press, Oxford, 1982, vol. 2, ch. 11, p. 548. 21 M. A. Brown, D. G. Tuck and E. J. Wells, Can. J. Chem. 1996, 74, 1535. Paper 8/04124E

ISSN:1477-9226    

DOI:10.1039/a804124e

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3549–3558 3549 Preparation, molecular structure and reactivity of mono- and di-nuclear sulfonato rhodium(I) complexes Helmut Werner, Marco Bosch, Michael E. Schneider, Christine Hahn, Frank Kukla, Matthias Manger, Bettina Windmüller, Birgit Weberndörfer and Matthias Laubender Institut für Anorganische Chemie der Universität Würzburg, Am Hubland, D-97074 Würzburg, Germany. E-mail: helmut.werner@mail.uni-wuerzburg.de Received 28th July, Accepted 7th September 1998 The reaction of [Rh(h3-C3H5)(PPri 3)2] 1 or [Rh(h3-CH2Ph)(PPri 3)2] 2 with an equimolar amount of RSO3H (R = Me, p-tolyl, CF3, F, Camph) led to the formation of the monomeric sulfonatorhodium(I) complexes [Rh{h2-O2S(O)R}- (PPri 3)2] 3–7 in excellent yield.An alternative route for the preparation of 4 (R = p-tolyl) and 5 (R = CF3) is based on the reaction of PPri 3 with the dinuclear compounds [{Rh(C8H14)2[m-O2S(O)R]}2], which were obtained either from [{Rh(C8H14)2(m-Cl)}2] 8 or [{Rh(C8H14)2(m-OH)}2] 9 as starting materials.Compounds 3–7 react smoothly with hydrogen by oxidative addition to give the dihydridorhodium(III) complexes [RhH2{h2-O2S(O)R}(PPri 3)2]. Moreover, on treatment of 3–6 with CO and C2H4 the chelating bond of the sulfonate ligand is partially opened and the carbonyl and ethene complexes trans-[Rh{h1-OS(O)2R}(L)(PPri 3)2] (L = CO or C2H4) are formed. The bis(stibine)- rhodium(I) derivative trans-[Rh{h1-OS(O)2CF3}(C2H4)(SbPri 3)2] was obtained from [{Rh(C2H4)2[m-O2S(O)CF3]}2] and SbPri 3.Reaction of the compounds [Rh{h2-O2S(O)CF3}(olefin)(PPri 3)] (olefin = C8H14 or C2H4) with benzene led to the displacement of the sulfonate ligand and to the formation of the half-sandwich-type complexes [Rh{h6-C6H6)- (olefin)(PPri 3)][CF3SO3] containing a rather labile benzene–rhodium bond. The preparation of the vinylidene complex trans-[Rh{h1-OS(O)C6H4Me-p}(]] C]] CHPh)(PPri 3)2] is also described and the crystal and molecular structures of three compounds have been determined. The four-co-ordinate sulfonato complexes 3–6 are active catalysts in the C–C coupling reaction of ethene and diphenyldiazomethane.Besides the three isomeric 1 : 1 adducts of C2H4 and CPh2, quite unexpectedly also the 2 : 1 adduct 3,3-diphenylpent-1-ene is formed. One of the most noteworthy discoveries, which we made in recent years, was that in the presence of catalytic amounts of chlororhodium(I) complexes such as [{RhCl(PPri 3)2}2] or [{RhCl(C2H4)2}2], ethene and diphenyldiazomethane react to give almost selectively 1,1-diphenylprop-1-ene Ph2C]] CHMe.1 This trisubstituted olefin is formally built up by the coupling of two carbene fragments ?? CPh2 and ?? CHMe, of which the latter is generated from the isomeric ethene.It was previously known that dinuclear rhodium(II) compounds such as [Rh2(m-O2- CMe)4] and derivatives thereof are active catalysts for the synthesis of cyclopropanes from olefins and diazoalkanes,2 but the formation of Ph2C]] CHMe from Ph2CN2 and C2H4 was without precedent.Following our initial studies, we also found that, if the acetylacetonato complex [Rh(acac)(C2H4)2] was used instead of [{RhCl(C2H4)2}2] as the catalyst, 1,1-diphenylcyclopropane and not the isomeric 1,1-diphenylprop-1-ene was formed from ethene and diphenyldiazomethane. In contrast, the hexafluoroacetylacetonato derivative [Rh(acac-F6)- (C2H4)2] behaves similarly to the chloro complex [{RhCl- (C2H4)2}2] and with C2H4–Ph2CN2 generates catalytically (although with low turnover numbers) 1,1-diphenylprop-1- ene.3,4 It was this apparent influence of the anionic ligand of the rhodium(I) complexes on both the reactivity and selectivity of the C–C coupling reaction that prompted us to prepare a series of compounds of the general composition [Rh{h2-O2S(O)R}- (PPri 3)2].The main reason why we chose the sulfonate derivatives was that the low nucleophilicity of the RSO3 2 anions makes this ligand an excellent leaving group, and we assumed that this could be an important aspect for the catalytic activity.Moreover, the sulfonate ligand may adopt either a bridging, a h1- or a h2-bonding mode upon co-ordination to a metal centre as was recently reported for carboxylato rhodium(I) compounds of a similar type.5,6 The present paper describes the preparation of mono- and di-nuclear sulfonatorhodium(I) complexes, the reactivity of these species toward H2, CO, C2H4, phenylacetylene and benzene, and the use of the bis(phosphine) complexes [Rh{h2- O2S(O)R}(PPri 3)2] as catalysts for the reaction of ethene and diphenyldiazomethane.Some preliminary results of these studies have been communicated.7 Results and discussion Preparation of mono- and di-nuclear sulfonate complexes The most convenient synthetic routes leading to the mononuclear sulfonatorhodium(I) compounds 3–7 are shown in Scheme 1.The reactions of the starting materials 1 and 2 with sulfonic acids were carried out in diethyl ether at 278 8C and gave propene and toluene, respectively, as by-products. Similarly, Stuhl and Muetterties 8 prepared the sulfonatomanganese( I) derivative [Mn{h2-O2S(O)CF3}(CO)2{P(OPri)3}2] by protonation of [Mn(h3-C3H5)(CO)2{P(OPri)3}2] with CF3- SO3H. The rhodium complexes 3–7 are red or violet airsensitive solids which have been characterized by elemental analysis and spectroscopic techniques.While the IR spectra of Scheme 1 L = PPri 3.3550 J. Chem. Soc., Dalton Trans., 1998, 3549–3558 the related carboxylato compounds [Rh(h2-O2CR)(PPri 3)2] clearly support the chelate structure,5,6 the IR data of 3–7 are less informative and do not distinguish between a h1- and h2- bonding mode of the sulfonato unit.9 As far as compounds 3–6 are concerned, the cis disposition of the two phosphine ligands is indicated by the appearance of one signal for the CH3 protons of the isopropyl groups in the 1H NMR spectrum which due to P–H and H–H coupling is split into a doublet of doublets in the case of 3, 5 and 6.For 7, which contains a chiral substituent at the sulfur atom, two doublets of doublets are observed. The 31P NMR spectra of 3–7 display one doublet, the Rh–P coupling of which (210–220 Hz) is also consistent with a cis disposition of the PPri 3 ligands.5,10 To confirm the structural proposal for the complexes 3–7, a single-crystal X-ray diVraction study of 4 was carried out.The ORTEP11 plot (Fig. 1) reveals that the ligand sphere around the metal centre is distorted square planar with the two phosphorus and the two oxygen atoms O(1) and O(2) lying exactly in the same plane as rhodium. The symmetrical arrangement of the ligands is illustrated by almost identical Rh–P and Rh–O distances (see Table 1), the latter [Rh–O(1) and Rh–O(2)] being about 0.05 Å longer than in the analogous acetato compound [Rh(h2-O2CMe)(PPri 3)2]. However, the bite angle O–Rh–O in 4 [64.21(7)8] and in [Rh(h2-O2CMe)(PPri 3)2] [60.2(1)8] is quite similar which could be due to the steric requirements of the bulky phosphine groups.The angle O(1)–S–O(2) [106.3(1)8] is only slightly smaller than would be anticipated for an ideal tetrahedral geometry. An alternative synthetic pathway to compounds 4 and 5 is outlined in Scheme 2. Treatment of the well known cyclooctene complex 8 with 2 equivalents of RSO3Ag (R = p-Tol or CF3) in CH2Cl2 or CH2Cl2–Et2O led to a displacement of the bridging chlorides by p-toluene- or trifluoromethane-sulfonates and gave compounds 11 and 12 as yellow solids in excellent yield.Instead of 8, the corresponding dimeric m-hydroxo complex 9 could also be used as starting material for the preparation Fig. 1 An ORTEP plot of complex 4. Table 1 Selected bond lengths (Å) and angles (8) for complex 4 Rh–O(1) Rh–O(2) Rh–P(1) Rh–P(2) P(1)–Rh–P(2) P(1)–Rh–O(1) P(1)–Rh–O(2) P(2)–Rh–O(1) P(2)–Rh–O(2) O(1)–Rh–O(2) Rh–O(1)–S 2.217(2) 2.227(2) 2.206(1) 2.218(1) 106.31(3) 93.85(6) 157.86(6) 159.79(6) 95.70(6) 64.21(7) 94.8(1) S–O(1) S–O(2) S–O(3) S–C(19) Rh–O(2)–S O(1)–S–O(2) O(1)–S–O(3) O(2)–S–O(3) O(1)–S–C(19) O(2)–S–C(19) O(3)–S–C(19) 1.476(2) 1.475(2) 1.429(2) 1.761(3) 94.5(1) 106.3(1) 114.8(1) 114.4(2) 107.7(1) 105.9(1) 107.2(2) of 10 and 11.It reacted with 2 equivalents of MeSO3H or p- MeC6H4SO3H?H2O in CH2Cl2 to aVord the m-sulfonato derivatives almost quantitatively.Compound 9 was prepared by reaction of 8 with an excess of NaOH in a two-phase system of C6H6–water following the procedure described by Alper and coworkers 12 for the synthesis of [{Rh(PPh3)2(m-OH)}2]. Along a similar route, the related triisopropylphosphine complex [{Rh(PPri 3)2(m-OH)}2] was prepared in our laboratory and characterized by crystal structure analysis.13 The m-sulfonato compounds 10–12 are yellow or orange-yellow, only moderately air-sensitive solids which are soluble in CH2Cl2, thf or ether and in the case of 12 even in saturated hydrocarbons. The reactions of 11 and 12 with triisopropylphosphine in ether at 278 8C proceed rather quickly and aVord 4 and 5 in 75–90% yield.Addition reactions of Á2-sulfonato and Ï-sulfonato rhodium(I) complexes The chelate complexes 3–7 are quite labile and react smoothly at room temperature with H2 as well as with CO and C2H4.On treatment with hydrogen, the dihydridorhodium(III) compounds 13–17 (Scheme 3) are obtained as white, nearly air-stable solids in excellent yield. The hydrido complexes are soluble in most organic solvents and can be stored under H2 at 278 8C for weeks. In vacuo they slowly lose hydrogen and regenerate the starting materials 3–7. The 31P NMR spectra of 13–17 display at room temperature only one resonance (doublet in 31P-{1H} and doublet of triplets in oV-resonance) with a 103Rh–31P coupling which is typical for trans disposed triisopropylphosphine ligands.14 In the 1H NMR spectra there Scheme 2 = C8H14, L = PPri 3.Scheme 3 L = PPri 3.J. Chem. Soc., Dalton Trans., 1998, 3549–3558 3551 is also only one set of signals at d 225.3 to 225.8 for the hydride ligands. Since for a rigid six-co-ordinate structure as shown in Scheme 3 two stereoisomers should exist, we assume that at 25 8C these isomers rapidly interconvert. If the 31P NMR spectra are measured in CD2Cl2 at 290 8C or in [2H8]toluene at 280 8C, a slight broadening of the single resonance is observed indicating that even under these conditions the interconversion of the two isomers is very fast on the NMR timescale.The same seems to be true in the case of the four-co-ordinate complex 7 for which the NMR data equally suggest an eVective C2 symmetry. It should be mentioned that compound 4 catalyses the hydrogenation of phenylactylene to styrene and we assume that the dihydrido derivative 14 is involved in this process.15 Like 13–17, the carbonyl complexes 18–21 are formed almost quantitatively by passing a slow stream of CO through a solution of 3–6 in hexane at room temperature.As far as the properties of 18–21 are concerned, a special feature is that while 18 and 19 are soluble in benzene or ether, the related species 20 and 21 are not. Conductivity measurements in nitromethane indicate that in this solvent the neutral compounds 20 and 21 are in equilibrium with the ionic species 20a and 21a (see Scheme 3).This can be understood by the general behaviour of CF3SO3 2 and FSO3 2 as good leaving groups. Owing to these results we assume that the NMR data measured for the carbonyl derivatives of the trifluoromethanesulfonato and the fluorosulfato rhodium complexes in nitromethane correspond to 20a and 21a and not to 20 and 21, respectively. Treatment of complexes 3–6 with C2H4 led to the formation of 22–25 the structure of which is probably quite similar to that of the carbonyl compounds 18–21 (see Scheme 4).Since the ethene complexes in analogy to the dihydrido derivatives are also unstable in vacuo, undergoing loss of C2H4, the NMR spectra of 22–25 were recorded in C6D6 which was saturated with ethene. The mass spectrum of 23 confirmed that the compound is monomeric in the solid state. The ethene complexes 23 and 24 are not only formed from 4 and 5 but also from the sulfonato-bridged compounds 11 and 12 as starting materials.In the initial step the cyclooctene ligands of 11 and 12 are displaced by ethene to aVord the corresponding dinuclear intermediates 26 and 27 both of which were isolated as yellow solids in excellent yield. While compound 26 is stable (and thus could be characterized by elemental analysis), the triflato derivative 27 is rather labile and decomposes rapidly in solution. We note that Aresta et al.16 reported the preparation of monomeric [Rh(O3SCF3)(C2H4)2] from [{RhCl(C2H4)2}2] and CF3SO3Ag which was found to be Scheme 4 L = PPri 3.unstable under nitrogen and even under ethene. In contrast, the mass spectrum of 27 revealed that this compound (obtained from 12) is a dimer and not a monomer in the solid state. Treatment of 26 and 27 with triisopropylphosphine led both to bridge cleavage and partial displacement of ethene to give the complexes 23 and 24 almost quantitatively. The dinuclear triflato derivative 12 reacts with triisopropylstibine in pentane even at 240 8C yielding a labile species that presumably contains both SbPri 3 and cyclooctene as ligands.17 Treatment of this intermediate with ethene aVords compound 28 which was isolated as an analytically pure solid in 86% yield.The same product is also obtained from 27 and an equimolar amount of SbPri 3. We would like to point out that, recently, we described the synthesis of the corresponding chlorobis(stibine) complex trans-[RhCl(C2H4)(SbPri 3)2] which is an excellent starting material for the preparation of a whole series of carbene rhodium(I) derivatives.18,19 The mixed cyclooctene–triisopropylphosphine rhodium(I) complex 29 containing a chelating triflate ligand is accessible from compound 12 and 2 equivalents of PPri 3 (Scheme 5).If this reaction is monitored at room temperature a change from orange to violet-brown initially occurs which is smoothly reversed after ca. 2 h. The 31P NMR measurements revealed that in the first stage of the process the bis(phosphine) complex 5 is formed which in the presence of unchanged starting material 12 (and cyclooctene) aVords the mixed olefin– phosphine derivative 29.The same product is obtained from 12 and 5 in the molar ratio of 1 : 2 in pentane. Treatment of 29 with C2H4 led to the formation of the ethene–phosphine complex 30 which was isolated in ca. 70% yield as an analytically pure yellow solid. The 1H NMR spectrum of 30 in CD2Cl2 at room temperature displays a broad singlet at d 2.75 for the C2H4 protons indicating that under these conditions rotation of the olefinic ligand around the Rh–C2H4 bond is only slightly hindered.In the 13C NMR spectrum of 30 the resonance for the C2H4 carbon atoms appears at d 43.7 as a doublet with a 103Rh–13C coupling constant of 15.3 Hz. The molecular structure of compound 29 was determined by X-ray crystallography. There are two independent molecules A and B in the unit cell, of which A is shown in Fig. 2. As the ORTEP plot reveals, the configuration around rhodium is slightly distorted square planar and therefore to some extent analogous to that of the tosylate complex 4. However, the bond angle between phosphorus, rhodium and the centre of the C]] C double bond for 29 (molecule A) is 97.158 (94.598 for B) and thus somewhat smaller than the bond angle P(1)–Rh–P(2) in compound 4. The atoms S, O(1), O(2), Rh, P and the centre of the C]] C bond lie almost in the same plane, the dihedral angle between the planes [O(1), S, O(2)] and [O(1), Rh, O(2)] for molecule A being 11.1(1)8 and for molecule B 10.7(1)8.The corresponding dihedral angles between [O(1), Rh, O(2)] and (P, Scheme 5 L = PPri 3.3552 J. Chem. Soc., Dalton Trans., 1998, 3549–3558 Rh, centre of C]] C) are 7.9(1)8 for A and 8.0(1)8 for B, respectively. While the Rh–O bond lengths in 4 are nearly identical, those in 29 are not (see Table 2); due to the diVerent ligands (C8H14 and PPri 3) in trans position they diVer by ca. 0.13 Å. In contrast, the distances Rh–P(1) and Rh–P(2) in 4 and Rh–P in 29 are almost the same. Dissolving either compound 29 or 30 in benzene leads to the displacement of the triflate ligand and to the formation of the cationic half-sandwich-type complexes 31 and 32 in virtually quantitative yield. Both 31 and 32 are yellow, only moderately air-sensitive solids which have been characterized by elemental analysis, IR, NMR and (in the case of 31) FAB mass spectroscopy.In solution in the absence of benzene they are quite labile and regenerate the starting materials 29 and 30, respectively. Despite this lability, a crystal structure analysis of 31 could be carried out, and the result of this study is summarized in Fig. 3 and Table 3. Similar to compound 29, there are two independent molecules A and B in the unit cell which diVer in the conformation of the cyclooctene ligand.The bond length Rh–P as well as the distances Rh–C(30) and Rh–C(31) are somewhat longer than in the square-planar complex 29 Fig. 2 An ORTEP plot of complex 29. which could be due to the cationic nature and the closed-shell electron configuration of the half-sandwich-type molecule. The angle between P, Rh and the centre of the C]] C bond in 31 is 92.698 for molecule A and 93.278 for molecule B, respectively. The distances between rhodium and the carbon atoms of the benzene ligand lie between 2.313(3) and 2.361(4) Å for A and between 2.296(4) and 2.365(4) Å for B, and this range is typical also for other arenerhodium(I) compounds.20 The reactivity of complex 4 as a representative of fourco- ordinate bis(triisopropylphosphine)rhodium(I) sulfonato compounds toward a terminal alkyne such as phenylacetylene is illustrated in Scheme 6.In toluene solution at room temperature a smooth reaction between 4 and PhC]] ] CH takes place which gives the vinylidene complex 33 as a violet crystalline solid in 91% isolated yield.The most typical spectroscopic Scheme 6 L = PPri 3. Fig. 3 An ORTEP plot of the cation of complex 31. Table 2 Selected bond lengths (Å) and angles (8) for complex 29 (there are two independent molecules A and B in the unit cell) Rh–P Rh–O(1) Rh–O(2) Rh–C(20) Rh–C(21) P–Rh–O(1) P–Rh–O(2) P–Rh–C(20) P–Rh–C(21) Rh–O(1)–S Rh–O(2)–S A 2.192(1) 2.219(3) 2.351(2) 2.091(4) 2.074(4) 98.41(7) 160.55(7) 94.9(1) 96.9(1) 95.8(1) 90.9(1) B 2.194(1) 2.205(3) 2.339(2) 2.091(4) 2.073(4) 97.64(7) 160.16(7) 95.3(1) 96.3(1) 96.2(1) 91.1(1) C(20)–C(21) S–O(1) S–O(2) S–O(3) S–C(1) O(1)–Rh–O(2) O(1)–S–O(2) O(1)–S–O(3) O(2)–S–O(3) O(1)–S–C(1) O(2)–S–C(1) A 1.413(5) 1.477(3) 1.458(3) 1.422(3) 1.820(5) 63.02(9) 109.1(2) 116.5(2) 117.3(2) 104.0(2) 103.6(2) B 1.404(5) 1.471(3) 1.456(3) 1.424(3) 1.820(4) 63.00(9) 108.6(2) 116.5(2) 117.6(2) 103.9(2) 103.9(2) Table 3 Selected bond lengths (Å) and angles (8) for cationic complex 31 (there are two independent molecules A and B in the unit cell) Rh–P Rh–C(30) Rh–C(31) Rh–C(50) Rh–C(51) P–Rh–C(30) P–Rh–C(31) Rh–C(30)–C(31) A 2.294(1) 2.141(3) 2.138(3) 2.330(3) 2.313(3) 94.43(9) 90.43(8) 70.7(2) B 2.290(1) 2.141(3) 2.143(3) 2.332(4) 2.346(4) 93.92(9) 91.35(9) 71.0(2) Rh–C(52) Rh–C(53) Rh–C(54) Rh–C(55) C(30)–C(31) Rh–C(31)–C(30) C(30)–C(31)–C(32) C(31)–C(30)–C(37) A 2.338(4) 2.361(4) 2.320(3) 2.328(3) 1.403(4) 71.0(2) 122.8(3) 123.0(3) B 2.324(3) 2.365(4) 2.362(4) 2.296(4) 1.396(5) 70.9(2) 124.8(3) 122.6(3)J.Chem. Soc., Dalton Trans., 1998, 3549–3558 3553 features of 33 are the doublet of triplets at d 1.51 for the ]] CHPh proton in the 1H NMR and the low-field resonance (also a doublet of triplets) at d 301.1 for the a-carbon atom of the vinylidene unit in the 13C NMR spectrum. These data are in good agreement with those of the corresponding chloro and acetato derivatives trans-[RhX(]] C]] CHPh)(PPri 3)2] (X = Cl or MeCO2) which have recently been prepared in our laboratory.5,14 Catalytic studies In the same way as the dimeric chloro derivative [{RhCl- (PPri 3)2}2], the new monomeric sulfonatorhodium(I) complexes 3–6 are also active catalysts in the C–C coupling reaction of ethene and diphenyldiazomethane. The most noteworthy feature is that the selectivity depends significantly on the substituent R of the sulfonate ligand.While in the reaction with either 3 or 4 as catalyst 1,1-diphenylprop-1-ene Ib (Scheme 7) is the major product, in the presence of 5 and 6 3,3-diphenylpent- 1-ene Id is the dominating species (see Table 4).This 2 : 1 adduct of C2H4 and Ph2CN2 is formed in only minor quantities if 3, 4 or the chlororhodium(I) dimer [{RhCl(PPri 3)2}2] is used as catalyst. In all C–C coupling reactions of ethene and diphenyldiazomethane, which are catalysed by bis(triisopropylphosphine) rhodium(I) compounds of the general type [{RhX(PPri 3)2}n] (n = 1 or 2), only trace amounts of 3,3- diphenylprop-1-ene Ic (which is generated selectively in the stoichiometric reaction of trans-[{RhCl(]] CPh2)(SbPri 3)2}2] with ethene) 19 are obtained.The mechanism for the formation of compound Id is not clear as yet. We assume that in analogy to the reaction of ethene and diphenyldiazomethane with [{RhCl(PPri 3)2}2] as catalyst,1 in the initial stage of the catalytic process both C2H4 and CPh2 are co-ordinated to rhodium, and that subsequently a rhodacyclobutane is formed.The next step could be either an insertion of ethene into one of the Rh–C bonds of the metallacyclobutane to give a six-membered RhC5 cycle or a b-H shift to generate a RhH(CH2CHCPh2) intermediate. Addition of C2H4 to this intermediate followed by the insertion of the olefin into the Rh–H bond could aVord a Rh(C2H5)(CH2CHCPh2) species which by metal-mediated C–C coupling yields Id. If the catalytic cycle involves the above mentioned RhC5 ring system, then upon a b-H shift a RhH(CH2CH2CPh2CH]] CH2) intermediate could be formed which by reductive elimination generates Id.Precedence for the postulated insertion of ethene into the metal–carbon bond of a metallacyclobutane can be found in the work of Binger and Schuchardt,21 who argue that the palladium(0)-catalysed cycloaddition of methylenecyclopropane and olefins presumably proceeds through a PdC5 intermediate. We cannot rule out that one of the initial steps in the formation of Id (following the co-ordination of ethene to Scheme 7 Table 4 Catalytic cycles and composition of the mixture of products obtained from ethene and diphenyldiazomethane in methylcyclohexane according to Scheme 7 Product (%) [Rh]-cat 345 b 6 b Cycles a 24 11 47 37 Ia 2275 Ib 86 87 37 29 Ic 4210 Id 89 55 66 a Cycle = mmol product/mmol catalyst.b In toluene. the metal centre) is an intramolecular C–H activation to give a RhH(CH]] CH2) intermediate which could then react with a second molecule of C2H4 to aVord a Rh(C2H5)(CH]] CH2) species.However, the reason why we consider this mechanistic route as less likely is that we failed to observe the generation of a hydrido(vinyl)rhodium(III) compound on photolysis of 24 or 25, respectively. With regard to the conversion of two ethene molecules into two s-bonded ligands in the co-ordination sphere of a d8 metal centre it should be pointed out that Carmona and co-workers 22 recently showed that the iridium complex [IrTp*(C2H4)2] [Tp* = tris(3,5-dimethylpyrazol-1-yl)hydroborate] rearranges thermally or photochemically to the isomer [IrTp*(H)(CH]] CH2)(C2H4)].On treatment with acetonitrile this compound yields the ethyl–vinyl derivative [IrTp*(CH]] CH2)(C2H5)- (NCMe)], which in the presence of catalytic amounts of water undergoes an intramolecular coupling of the vinyl and acetonitrile ligands to aVord a five-membered iridapyrrole ring.23 With [RhTp*(C2H4)2] as the starting material a related conversion into [RhTp*(CH]] CH2)(C2H5)(L)] (L = NCMe or py) takes place.24 In order to find out whether olefins other than ethene would also react with diphenyldiazomethane by C–C coupling, the reaction of Ph2CN2 with methyl acrylate and CH2]] CHCH2- CO2Me in the presence of the triflato complex 5 as the catalyst was also investigated.As is shown in Scheme 8, only cyclopropanation occurs and the corresponding esters IIa and IIb are formed.Whereas for IIa the number of cycles is 55, the yield of IIb is rather low and could not be increased by using an excess of the diazoalkane. Current work in our laboratory is aimed at the further exploration of the use of the sulfonatorhodium(I) as well as the related carboxylatorhodium(I) complexes in other types of C–C coupling reactions among which the cyclooligomerization of butadiene is of particular interest.7 Experimental All reactions were carried out under an atmosphere of argon by Schlenk-tube techniques.Solvents were dried by the usual procedures and distilled under argon prior to use. The starting materials 1, 25 and 8 25 were prepared by published methods. The NMR spectra were recorded on Bruker AC 200 and AMX 400 instruments and the IR spectra on a Perkin-Elmer 1420 spectrometer. Abbreviations used: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; vt, virtual triplet; N = 3J(PH) 1 5J(PH) or 1J(PC) 1 3J(PC), respectively.Conductivity data (L) in nitromethane. Preparations [Rh{Á2-O2S(O)Me}(PPri 3)2] 3. (a) A solution of compound 1 (0.219 g, 0.47 mmol) in ether (15 cm3) was treated dropwise with MeSO3H (0.031 cm3, 0.48 mmol) at 278 8C. After the addition a white suspension was formed which was warmed to room temperature to give an orange-red solution. The solution was stirred for 1 h, the solvent removed in vacuo and the residue extracted twice with pentane (40 cm3). The combined extracts were brought to dryness in vacuo, the remaining red solid was washed with small portions of pentane (220 8C) and dried: yield 0.153 g (63%).(b) A solution of compound 2 (0.241 g, 0.47 mmol) in ether (15 cm3) was treated dropwise with MeSO3H (0.031 cm3, 0.48 mmol) at 278 8C. A red solution formed which was warmed to Scheme 83554 J. Chem. Soc., Dalton Trans., 1998, 3549–3558 room temperature. After it was stirred for 30 min, the solvent was removed in vacuo and the residue washed twice with small portions of pentane (220 8C) to give a red solid: yield 0.130 g (54%); mp 110 8C (decomp.) (Found: C, 43.71; H, 8.94; S, 5.88.C19H45O3P2RhS requires C, 44.02; H, 8.75; S, 6.18%). IR (KBr): n(O3S) 1201, 1190 and 1049 cm21. NMR (C6D6): dH (200 MHz) 2.68 (3 H, s, SCH3), 1.80 (6 H, m, CHCH3) and 1.23 [36 H, dd, J(PH) 13.0, J(HH) 7.1 Hz, CHCH3]; dC (50.3 MHz) 40.0 (s, SCH3), 25.5 (vt, N 21.3 Hz, CHCH3) and 20.3 (s, CHCH3); dP (81.0 MHz) 70.2 [d, J(RhP) 212.2 Hz].[Rh{Á2-O2S(O)C6H4Me-p}(PPri 3)2] 4. A solution of compound 1 (0.116 g, 0.25 mmol) in toluene (2 cm3) was treated with p-MeC6H4SO3H (0.048 g, 0.25 mmol) and stirred for 1 h at room temperature. A change from yellow to red occurred. The solvent was removed in vacuo, the residue extracted with acetone (20 cm3) and the extract concentrated to ca. 2 cm3 in vacuo. Red crystals precipitated which were filtered oV, washed twice with small portions of acetone (0 8C) and dried: yield 0.131 g (88%); mp 80 8C (decomp.) (Found: C, 50.62; H, 8.63.C25H49O3P2RhS requires C, 50.50; H, 8.31%). NMR (C6D6): dH (200 MHz) 8.34, 6.86 (4 H, both m, C6H4), 1.90 (3 H, s, C6H4CH3), 1.83 (6 H, m, CHCH3) and 1.24 (36 H, m, CHCH3); dC (50.3 MHz) 141.3 [d, J(RhC) 15.8 Hz, ipso-C of C6H4], 129.5, 129.0, 127.2 (all s, C6H4), 25.7 (vt, N 21.0 Hz, CHCH3), 21.1 (s, C6H4CH3) and 20.4 (s, CHCH3); dP (81.0 MHz) 70.3 [d, J(RhP) 212.5 Hz]. Alternatively, compound 4 was prepared on treatment of a solution of 11 (0.040 g, 0.04 mmol) in ether (4 cm3) with PPri 3 (0.031 cm3, 0.16 mmol) at 278 8C.After the solution was stirred for 5 min the solvent was removed in vacuo. The oily residue was dissolved in pentane (1.5 cm3) and the solution stored for 12 h at 278 8C to give a red solid: yield 0.066 g (72%). [Rh{Á2-O2S(O)CF3}(PPri 3)2] 5. This compound was prepared as described for 3, using either 1 (0.088 g, 0.19 mmol) and CF3SO3H (0.017 cm3, 0.19 mmol) or 2 (0.160 g, 0.31 mmol) and CF3SO3H (0.028 cm3, 0.32 mmol) as starting materials.Violet solid: yield 0.102 g (95%) from 1 and 0.103 g (58%) from 2; mp 80 8C (decomp.) (Found: C, 39.60; H, 7.48; S, 5.36. C19H42F3O3P2RhS requires C, 39.87; H, 7.40; S, 5.60%). MS (70 eV): m/z 573 (M1). IR (KBr): n(O3S) 1263 and 1030, n(CF3) 1253 and 1161 cm21. NMR (C6D6): dH (200 MHz) 1.70 (6 H, m, CHCH3) and 1.13 [36 H, dd, J(PH) 13.5, J(HH) 7.3 Hz, CHCH3]; dC (50.3 MHz) 121.4 [q, J(FC) 316.9 Hz, CF3], 25.7 (vt, N 23.1 Hz, CHCH3) and 20.0 (s, CHCH3); dF (188.2 MHz) 277.1 (s); dP (81.0 MHz) 69.9 [d, J(RhP) 219.5 Hz].Alternatively, compound 5 was prepared on treatment of a solution of 12 (0.038 g, 0.04 mmol) in ether (4 cm3) with PPri 3 (0.031 cm3, 0.16 mmol) at 278 8C. After the solution was stirred for 5 min the solvent was removed in vacuo, the violet residue washed twice with 2 cm3 portions of pentane (240 8C) and dried: yield 0.041 g (89%). [Rh{Á2-O2S(O)F}(PPri 3)2] 6.This compound was prepared as described for 3, using either 1 (0.154 g, 0.33 mmol) and FSO3H (0.019 cm3, 0.33 mmol) or 2 (0.251 g, 0.49 mmol) and FSO3H (0.028 cm3, 0.49 mmol) as starting materials. Violet solid: yield 0.152 g (88%) from 1 and 0.256 g (81%) from 2; mp 64 8C (decomp.) (Found: C, 41.02; H, 8.02; S, 6.62. C18H42FO3P2RhS requires C, 41.38; H, 8.10; S, 6.14%). MS (70 eV): m/z 522 (M1). IR (KBr): n(O3S) 1312, 1240 and 1062, n(SF) 775 cm21. NMR (C6D6): dH (200 MHz) 1.69 (6 H, m, CHCH3) and 1.14 [36 H, dd, J(PH) 13.4, J(HH) 7.2 Hz, CHCH3]; dC (50.3 MHz) 25.8 (vt, N 22.2 Hz, CHCH3) and 20.1 (s, CHCH3); dF (188.2 MHz) 40.7 (s); dP (81.0 MHz) 70.5 [d, J(RhP) 220.1 Hz]. [Rh{Á2-O2S(O)C10H15O}(PPri 3)2] 7.This compound was prepared as described for 4, using 1 (0.116 g, 0.25 mmol) and (1S)-camphor-10-sulfonic acid (0.058 g, 0.25 mmol) as starting materials. Red solid: yield 0.134 g (82%); mp 45 8C (decomp.) (Found: C, 51.12; H, 8.67; S, 4.64.C28H57O4P2RhS requires C, 51.37; H, 8.78; S, 4.90%). NMR (C6D6): dH (400 MHz) 4.14, 3.18 [2 H, both d, J(HH) 15.1, CH2SO3], 2.10–1.45 (7 H, br m, 7 H of C10H15), 1.85 (6 H, m, CHCH3), 1.28 [18 H, dd, J(PH) 12.8, J(HH) 5.6, CHCH3], 1.27 [18 H, dd, J(PH) 12.8, J(HH) 5.5 Hz, CHCH3], 1.15, 0.59 (6 H, both s, CH3 of C10H15); dP (162.0 MHz) 69.8 [d, J(RhP) 213.0 Hz]. [{Rh(C8H14)2}2(Ï-OH)2] 9. A two phase system of complex 8 (0.315 g, 0.44 mmol) in C6H6 (15 cm3), NaOH (0.12 g, 3.0 mmol) and [PhCH2NEt3]Cl (0.050 g) in water (10 cm3) was stirred at room temperature for 4 h.The C6H6 layer was decanted and the aqueous phase extracted with 10 cm3 of C6H6. The combined benzene fractions were dried over Na2SO4 and then filtered. From the filtrate the solvent was removed in vacuo to give a yellow solid. The solid was washed three times with 5 cm3 portions of pentane and dried in vacuo. Pale yellow solid: yield 0.265 g (89%); mp 70 8C (decomp.) (Found: C, 56.66; H, 8.29. C32H58O2Rh2 requires C, 56.47; H, 8.58%).IR (KBr): n(OH) 3676 cm21. NMR (C6D6): dH (200 MHz) 2.44 (8 H, m, ]] CH of C8H14), 2.48–1.13 (48 H, br m, CH2 of C8H14) and 21.23 (2 H, br s, OH); dC 72.49 (br m, ]] CH of C8H14), 30.24, 28.64, 28.59 (all s, CH2 of C8H14). [{Rh(C8H14)2}2{Ï-O2S(O)Me}2] 10. A solution of complex 9 (0.455 g, 0.67 mmol) in CH2Cl2 (30 cm3) was treated at room temperature with MeSO3H (0.087 cm3, 1.34 mmol). The mixture was stirred for 2 h and then the solvent was removed in vacuo.The resulting yellow solid was washed three times with 5 cm3 portions of Et2O and dried: yield 0.416 g (74%); mp 124 8C (decomp.) (Found: C, 48.49; H, 7.27; S, 7.87. C34H62O6Rh2S2 requires C, 48.80; H, 7.47; S, 7.66%). IR (KBr): n(O3S) 1276, 1207, 1070 and 1012, n(CF3) 1260 and 1170 cm21. NMR (C6D6): dH (200 MHz) 2.47 (6 H, s, CH3), 2.14 (16 H, m, ]] CH and CH2 of C8H14), 1.78, 1.52, 1.38 (40 H, all br m, CH2 of C8H14); dC (50.3 MHz) 73.6 [d, J(RhC) 15.7 Hz, ]] CH of C8H14], 39.4 (s, CH3), 29.7, 28.2, 26.6 (all s, CH2 of C8H14).[{Rh(C8H14)2}2{Ï-O2S(O)C6H4Me-p}2] 11. A suspension of compound 8 (0.520 g, 0.72 mmol) and p-MeC6H4SO3Ag (0.420 g, 1.45 mmol) in CH2Cl2 (35 cm3) was stirred for 2 d at room temperature. A white solid precipitated and a change of the solution from orange-red to orange-yellow occurred. The solvent was removed in vacuo, the residue extracted with ether (40 cm3) and the extract brought to dryness in vacuo.A yellow solid was isolated which was repeatedly washed with 5 cm3 portions of pentane and dried: yield 0.570 g (80%); mp 168 8C (decomp.) (Found: C, 55.81; H, 6.95; Rh, 20.62; S, 6.54. C46H70O6Rh2S2 requires C, 55.87; H, 7.13; Rh, 20.81; S, 6.48%). IR (CH2Cl2): n(O3S) 1190, 1115, 1068 and 1022 cm21. NMR (CDCl3): dH (200 MHz) 8.03 (4 H, m, ortho-H of SC6H4), 7.27 (4 H, m, meta-H of SC6H4), 2.56 (8 H, m, ]] CH of C8H14), 2.39 (6 H, s, C6H4CH3), 2.32–1.17 (48 H, br m, CH2 of C8H14); dC (50.3 MHz) 142.6 (s, ipso-C of SC6H4), 137.4, 129.1, 126.4 (all s, C6H4), 73.9 [d, J(RhC) 15.7 Hz, ]] CH of C8H14], 29.2, 28.0, 26.1 (all s, CH2 of C8H14) and 21.4 (s, C6H4CH3).Alternatively, compound 11 was prepared on treatment of a solution of 9 (0.352 g, 0.519 mmol) in CH2Cl2 (30 cm3) with p-MeC6H4SO3H?H2O (0.247 g, 1.30 mmol). The mixture was stirred for 4 h at room temperature, the solvent removed in vacuo and the residue extracted twice with 15 cm3 portions of C6H6.The benzene layers were dried over Na2SO4 and filtered. Removal of the solvent in vacuo aVorded a yellow solid which was washed three times with 5 cm3 portions of ether and dried: yield 0.375 g (72%). [{Rh(C8H14)2}2{Ï-O2S(O)CF3}2] 12. A solution of compound 8 (0.340 g, 0.47 mmol) in CH2Cl2–ether (2 :1, 25 cm3) was treated with a solution of CF3SO3Ag (0.234 g, 0.91 mmol) inJ. Chem. Soc., Dalton Trans., 1998, 3549–3558 3555 ether (15 cm3) and stirred for 2 h at room temperature.A white solid precipitated and a gradual change of the solution from orange-yellow to yellow occurred. The solvent was removed in vacuo and the residue extracted with hexane (40 cm3). The extract was slowly concentrated in vacuo until an orange-yellow solid began to precipitate. The solution was then stored for 24 h at 278 8C, the precipitate separated from the mother-liquor, washed twice with 3 cm3 portions of pentane (240 8C) and dried: yield 0.318 g (71%); mp 73 8C (decomp.) (Found: C, 42.86; H, 5.95; S, 6.64.C34H56F6O6Rh2S2 requires C, 43.22; H, 5.97; S, 6.79%). IR (CH2Cl2): n(CF) 1245, n(O3S) 1195, 1135, 1045 and 1005 cm21. NMR (CDCl3 for 1H and 13C): dH (200 MHz) 2.63 (8 H, m, ]] CH of C8H14), 2.24–1.32 (48 H, br m, CH2 of C8H14); dC (50.3 MHz) 119.3 [q, J(FC) 318.8 Hz, CF3], 74.1 [d, J(RhC) 15.7 Hz, ]] CH of C8H14], 29.1, 27.4, 26.1 (all s, CH2 of C8H14); dF (188.3 MHz, CD2Cl2) 277.1 (s). [RhH2{Á2-O2S(O)Me}(PPri 3)2] 13.A slow stream of hydrogen was passed for ca. 5 s through a solution of complex 3 (0.064 g, 0.12 mmol) in ether (8 cm3). A rapid change from red to almost white occurred. After the solution was stirred for 15 min at room temperature the solvent was removed in vacuo. The remaining white solid was washed with small quantities of pentane (278 8C) and quickly dried: yield 0.057 g (88%): mp 41 8C (decomp.) (Found: C, 43.60; H, 9.18; S, 5.99. C19H47O3- P2RhS requires C, 43.84; H, 9.10; S, 6.16%).IR (KBr): n(RhH) 2165 and 2135, n(O3S) 1207, 1193 and 1043 cm21. NMR (C6D6): dH (200 MHz) 2.60 (3 H, s, SMe), 2.18 (6 H, m, CHCH3), 1.25 [36 H, dvt, N 14.3, J(HH) 7.1, CHCH3] and 225.30 [2 H, dt, J(RhH) 29.1, J(PH) 14.8 Hz, RhH]; dC (100.6 MHz) 39.5 (s, SCH3), 25.7 [vt, N 21.4 Hz, CHCH3] and 20.4 (s, CHCH3); dP (162.0 MHz) 60.9 [d, dt in oV-resonance, J(RhP) 115.0 Hz]. [RhH2{Á2-O2S(O)C6H4Me-p}(PPri 3)2] 14. This compound was prepared as described for 13, using 4 (0.059 g, 0.10 mmol) in CH2Cl2 (5 cm3) as starting material.After the white solid was washed with small quantities of pentane (278 8C) it was dried in vacuo for not more than 5 min: yield 0.057 g (95%); mp 112 8C (decomp.) (Found: C, 49.90; H, 8.79; S, 5.30. C25H51O3- P2RhS requires C, 50.33; H, 8.62; S, 5.37%). NMR (C6D6): dH (200 MHz) 8.01–6.77 (4 H, m, C6H4), 2.14 (6 H, m, CHCH3), 1.89 (3 H, s, C6H4CH3), 1.18 [36 H, dvt, N 13.1, J(HH) 7.3, CHCH3] and 225.26 [2 H, dt, J(RhH) 30.5, J(PH) 14.5 Hz, RhH]; dP (81.0 MHz) 61.8 [d, dt in oV-resonance, J(RhP) 114.8 Hz].[RhH2{Á2-O2S(O)CF3}(PPri 3)2] 15. A slow stream of hydrogen was passed for ca. 5 s through a solution of complex 5 (0.117 g, 0.21 mmol) in ether (5 cm3). A rapid change from violet to pale yellow occurred. After the solution was stirred for 15 min at room temperature the solvent was removed in vacuo. The residue was extracted with pentane (30 cm3), the extract filtered and the filtrate brought to dryness in vacuo.The remaining white solid was washed with small quantities of pentane (278 8C) and quickly dried: yield 0.085 g (68%); mp 52 8C (decomp.) (Found: C, 39.84; H, 7.64; S, 5.17. C19H44F3O3P2RhS requires C, 39.71; H, 7.72; S, 5.57%). MS (70 eV): m/z 574.8 (M1). IR (KBr): n(RhH) 2194 and 2135, n(CF3) 1258 and 1171, n(O3S) 1270 and 1032 cm21. NMR (C6D6): dH (200 MHz) 2.16 (6 H, m, CHCH3), 1.08 [36 H, dvt, N 13.9, J(HH) 6.9, CHCH3] and 225.80 [2 H, dt, J(RhH) 33.5, J(PH) 14.5 Hz, RhH]; dC (50.3 MHz) 121.0 [q, J(CF) 319.5 Hz, CF3], 24.9 (vt, N 21.9 Hz, CHCH3) and 20.2 (s, CHCH3); dF (188.2 MHz) 277.3 (s); dP (81.0 MHz) 60.9 [d, dt in oV-resonance, J(RhP) 114.1 Hz].[RhH2{Á2-O2S(O)F}(PPri 3)2] 16. This compound was prepared as described for 13, using 6 (0.136 g, 0.26 mmol) in ether (3 cm3) as starting material. White solid: yield 0.105 g (77%); mp 45 8C (decomp.) (Found: C, 40.98; H, 8.42; S, 6.11. C18H44FO3P2RhS requires C, 41.22; H, 8.46; S, 6.11%).IR (KBr): n(RhH) 2180 and 2158, n(O3S) 1285, 1252 and 1078, n(SF) 740 cm21. NMR (C6D6): dH (200 MHz) 2.07 (6 H, m, CHCH3), 1.10 [36 H, dvt, N 13.7, J(HH) 6.7, CHCH3] and 225.60 [2 H, dt, J(RhH) 33.0, J(PH) 14.8 Hz, RhH]; dC (50.3 MHz) 25.2 (vt, N 22.1 Hz, CHCH3) and 20.2 (s, CHCH3); dF (188.2 MHz) 41.6 (s, SF); dP (81.0 MHz) 60.6 [d, dt in oV-resonance, J(RhP) 114.5 Hz]. [RhH2{Á2-O2S(O)C10H15O}(PPri 3)2] 17. This compound was prepared as described for 13, using 7 (0.059 g, 0.09 mmol) in CH2Cl2 (5 cm3) as starting material.White solid: yield 0.054 g (92%); mp 42 8C (decomp.) (Found: C, 50.70; H, 9.05; S, 4.47. C28H59O3P2RhS requires C, 51.21; H, 9.06; S, 4.88%). IR (KBr): n(RhH) 2120 cm21. NMR (C6D6): dH (200 MHz) 3.86, 3.04 [2 H, both d, J(HH) 16.0 Hz, CH2SO3], 2.32 (6 H, m, CHCH3), 2.07– 1.40 (7 H, m, 7 H of C10H15), 1.25 [36 H, dvt, N 13.1, J(HH) 5.8, CHCH3], 1.21, 0.57 (6 H, both s, CH3 of C10H15) and 225.31 [2 H, dt, J(RhH) 30.5, J(PH) 14.4 Hz, RhH]; dP (81.0 MHz) 61.4 [d, dt in oV-resonance, J(RhP) 114.8 Hz].trans-[Rh{Á1-OS(O)2Me}(CO)(PPri 3)2] 18. A slow stream of CO was passed for ca. 5 s through a solution of complex 3 (0.067 g, 0.13 mmol) in hexane (2 cm3). The resulting pale yellow suspension was stirred for 15 min at room temperature. After the solvent was removed in vacuo, the remaining white solid was washed three times with pentane (2 cm3) and dried: yield 0.063 g (88%); mp 122 8C (decomp.) (Found: C, 43.69; H, 8.05; S, 5.90.C20H45O4P2RhS requires C, 43.94; H, 8.30; S, 5.87%). MS (70 eV): m/z 546 (M1). IR (KBr): n(CO) 1964, n(O3S) 1265 and 1033 cm21. NMR (C6D6): dH (200 MHz) 2.61 (3 H, s, SCH3), 2.54 (6 H, m, CHCH3) and 1.25 [36 H, dvt, N 14.1, J(HH) 7.2 Hz, CHCH3]; dC (50.3 MHz) 191.1 [dt, J(RhC) 76.7, J(PC) 16.4 Hz, CO], 40.1 (s, SMe), 24.9 (vt, N 20.3 Hz, CHCH3) and 20.3 (s, CHCH3); dP (81.0 MHz) 51.4 [d, J(RhP) 119.2 Hz].Alternatively, compound 18 was prepared on treatment of a solution of 13 (0.068 g, 0.13 mmol) in hexane (2 cm3). The resulting suspension was worked up as described above. White solid: yield 0.060 g (85%). trans-[Rh{Á1-OS(O)2C6H4Me-p}(CO)(PPri 3)2] 19. This compound was prepared as described for 18, using 4 (0.071 g, 0.12 mmol) as starting material. Light yellow solid: yield 0.066 g (89%); mp 117 8C (decomp.) (Found: C, 49.69; H, 8.00; S, 5.18. C26H49O4P2RhS requires C, 50.16; H, 7.93; S, 5.15%). IR (KBr): n(CO) 1945 cm21.NMR (C6D6): dH (200 MHz) 7.99–6.83 (4 H, m, C6H4), 2.46 (6 H, m, CHCH3), 1.95 (3 H, s, C6H4CH3) and 1.23 [36 H, dvt, N 13.9, J(HH) 7.3 Hz, CHCH3]; dP (81.0 MHz) 51.7 [d, J(RhP) 119.1 Hz]. trans-[Rh{Á1-OS(O)2CF3}(CO)(PPri 3)2] 20. This compound was prepared as described for 18, using either 5 (0.089 g, 0.16 mmol) or 15 (0.090 g, 0.16 mmol) as starting material. Light yellow solid: yield 0.087 g (93%) from 5 or 0.083 g (90%) from 15; mp 112 8C (decomp.) (Found: C, 39.69; H, 7.11; S, 5.39.C20H42F3O4P2RhS requires C, 40.01; H, 7.05; S, 5.34%). L 35 W21 cm2 mol21. MS (70 eV): m/z 600 (M1). IR (KBr): n(CO) 1964, n(O3S) 1273 and 1042, n(CF3) 1253 and 1171 cm21. NMR (CD3NO2): dH (400 MHz) 2.69 (6 H, m, CHCH3) and 1.44 [36 H, dvt, N 15.6, J(HH) 7.2 Hz, CHCH3]; dC (100.6 MHz) 191.6 [dt, J(RhC) 65.4, J(PC) 13.6, CO], 122.4 [q, J(CF) 320.9 Hz, CF3], 28.7 (vt, N 24.8 Hz, CHCH3) and 20.6 (s, CHCH3); dF (376.4 MHz) 278.4 (s); dP (162.0 MHz) 57.4 [d, J(RhP) 100.4 Hz].Since the NMR spectra were measured in CD3NO2 the signals probably correspond to the ionic species 20a (see Scheme 3). trans-[Rh{Á1-OS(O)2F}(CO)(PPri 3)2] 21. This compound was prepared as described for 18, using either 6 (0.108 g, 0.21 mmol)3556 J. Chem. Soc., Dalton Trans., 1998, 3549–3558 or 16 (0.116 g, 0.21 mmol) as starting material. Pale yellow solid: yield 0.110 g (95%) from 6 or 0.105 g (91%) from 16; mp 121 8C (decomp.) (Found: C, 41.21; H, 7.36; S, 5.91.C19H42FO4PRhS requires C, 41.46; H, 7.69; S, 5.82%). L 31 W21 cm2 mol21. MS (70 eV): m/z 550 (M1). IR (KBr): n(CO) 1946, n(O3S) 1286 and 1063, n(SF) 700 cm21. NMR (CD3NO2): dH (400 MHz) 2.68 (6 H, m, CHCH3) and 1.44 [36 H, dvt, N 15.2, J(HH) 7.2 Hz, CHCH3]; dC (100.6 MHz) 191.6 [dt, J(RhC) 65.4, J(PC) 13.6 Hz, CO], 28.7 (vt, N 24.8 Hz, CHCH3) and 20.5 (s, CHCH3); dF (376.4 MHz) 36.5 (s); dP (162.0 MHz) 57.4 [d, J(RhP) 100.3 Hz].Since the NMR spectra were measured in CD3NO2 the signals probably correspond to the ionic species 21a (see Scheme 3). trans-[Rh{Á1-OS(O)2Me}(C2H4)(PPri 3)2] 22. A slow stream of ethene was passed through a suspension of complex 3 (0.102 g, 0.20 mmol) in pentane (5 cm3). After the reaction mixture was stirred for 30 min at room temperature it was concentrated to ca. 1 cm3 by passing a stream of ethene through the solution. After the solution was stored for 12 h at 278 8C a yellow microcrystalline solid precipitated which was separated from the mother-liquor, washed twice with small portions of pentane (278 8C) and dried with a stream of ethene: yield 0.104 g (97%); mp 74 8C (decomp.) (Found: C, 45.71; H, 9.15; S, 5.56.C21H49O3P2RhS requires C, 46.14; H, 9.04; S, 5.85%). IR (KBr): n(O3S) 1252, 1162 and 1039 cm21. NMR (C6D6, saturated with C2H4): dH (400 MHz) 2.70 (3 H, s, SCH3), 2.47 (4 H, m, C2H4), 2.26 (6 H, m, CHCH3) and 1.21 [36 H, N 13.2, J(HH) 6.4 Hz, CHCH3]; dC (100.6 MHz) 40.2 (s, SCH3), 33.3 [d, J(RhC) 16.5 Hz, C2H4], 22.8 (vt, N 16.3 Hz, CHCH3) and 20.5 (s, CHCH3); dP (162.0 MHz) 35.2 [d, J(RhP) 119.1 Hz]. trans-[Rh{Á1-OS(O)2C6H4Me-p}(C2H4)(PPri 3)2] 23.This compound was prepared as described for 22, using 4 (0.075 g, 0.13 mmol) as starting material. Yellow microcrystalline solid: yield 0.067 g (85%); mp 58 8C (decomp.) (Found: C, 52.14; H, 8.86; S, 5.16. C27H53O3P2RhS requires C, 52.09; H, 8.58; S, 5.15%).MS (70 eV): m/z 622 (M1). IR (KBr): n(O3S) 1259 and 1032 cm21. NMR (C6D6, saturated with C2H4): dH (200 MHz) 8.02, 6.85 (4 H, both m, C6H4), 2.51 (4 H, m, C2H4), 2.17 (6 H, m, CHCH3), 1.99 (3 H, s, C6H4CH3) and 1.20 [36 H, N 13.2, J(HH) 6.6 Hz, CHCH3]; dC (50.3 MHz) 144.2 (s, ipso-C of SC6H4), 139.3, 128.3, 127.0 (all s, C6H4), 33.2 [d, J(RhC) 16.6 Hz, C2H4], 22.9 (vt, N 16.6 Hz, CHCH3), 21.1 (s, C6H4CH3) and 20.4 (s, CHCH3); dP (162.0 MHz) 35.6 [d, J(RhP) 119.2 Hz]. Alternatively, compound 23 was prepared on treatment of a suspension of 26 (0.028 g, 0.04 mmol) in C6D6 (0.5 cm3) with PPri 3 (0.030 cm3, 0.16 mmol).The 1H NMR spectrum displayed only the signals of 23 and of free ethene. trans-[Rh{Á1-OS(O)2CF3}(C2H4)(PPri 3)2] 24. This compound was prepared as described for 22, using 5 (0.077 g, 0.14 mmol) as starting material. Yellow microcrystalline solid: yield 0.065 g (80%); mp 76 8C (decomp.) (Found: C, 41.90; H, 7.53; S, 5.27. C21H46F3O3P2RhS requires C, 41.99; H, 7.72; S, 5.33%).IR (KBr): n(O3S) 1305 and 1026, n(CF3) 1230 and 1165 cm21. NMR (C6D6): dH (200 MHz) 2.54 (4 H, m, C2H4), 2.13 (6 H, m, CHCH3) and 1.15 [36 H, N 13.2, J(HH) 6.9 Hz, CHCH3]; dC (100.6 MHz) 120.8 [q, J(FC) 321.0 Hz, CF3], 33.7 (br s, C2H4), 22.9 (vt, N 17.3 Hz, CHCH3) and 20.3 (s, CHCH3); dF (188.2 MHz) 276.7 (s); dP (188.2 MHz) 34.6 [d, J(RhP) 117.5 Hz]. Alternatively, compound 23 was prepared on treatment of a suspension of 27 (0.040 g, 0.06 mmol) in C4D8O (0.5 cm3) at 220 8C with PPri 3 (0.045 cm3, 0.24 mmol).The 1H NMR spectrum displayed only the signals of 24 and of free ethene. trans-[Rh{Á1-OS(O)2F}(C2H4)(PPri 3)2] 25. This compound was prepared as described for 22, using 6 (0.098 g, 0.19 mmol) as starting material. Yellow microcrystalline solid: yield 0.102 g (98%); mp 52 8C (decomp.) (Found: C, 43.22; H, 8.23; S, 5.74. C20H46FO3P2RhS requires C, 43.64; H, 8.42; S, 5.82%). MS (70 eV): m/z 550 (M1).IR (KBr): n(O3S) 1332, 1231 and 1083, n(SF) 739 cm21. NMR (C6D6): dH (200 MHz) 2.38 [4 H, dd, J(RhH) 6.7, J(PH) 4.0, C2H4], 2.01 (6 H, m, CHCH3) and 1.15 [36 H, N 13.1, J(HH) 6.8 Hz, CHCH3]; dC (50.3 MHz) 34.8 [d, J(RhC) 17.7 Hz, C2H4], 21.6 (vt, N 17.6 Hz, CHCH3) and 19.6 (s, CHCH3); dF (188.2 MHz) 42.8 (s); dP (81.0 MHz) 35.8 [d, J(RhP) 117.3 Hz]. [{Rh(C2H4)2}2{Ï-O2S(O)C6H4Me-p}2] 26. A slow stream of ethene was passed through a suspension of complex 11 (0.110 g, 0.11 mmol) in hexane (4 cm3) for 1 min at room temperature.After the reaction mixture was stirred for 5 min it was stored until the pale yellow solid and the solution were separated. The mother-liquor was decanted, the pale yellow solid repeatedly washed with 5 cm3 portions of pentane and dried: yield 0.062 g (85%); mp 134 8C (decomp.) (Found: C, 39.75; H, 4.62; S, 9.28. C22H30O6Rh2S2 requires C, 40.01; H, 4.58; S, 9.71%). IR (CH2Cl2): n(O3S) 1195, 1132, 1048 and 1015 cm21. NMR (CDCl3, 60 8C): dH (200 MHz) 7.82 (4 H, m, ortho-H of SC6H4), 6.80 (4 H, m, meta-H of SC6H4), 3.00 (12 H, s, C2H4) and 1.93 (6 H, s, C6H4CH3); dC (50.3 MHz) 141.5 (s, ipso-C of SC6H4), 140.7, 129.1, 126.7 (all s, C6H4), 60.6 (br m, C2H4) and 21.0 (s, C6H4CH3).[{Rh(C2H4)2}2{Ï-O2S(O)CF3}2] 27. This compound was prepared as described for 26, using 12 (0.080 g, 0.08 mmol) as starting material. Yellow solid: yield 0.040 g (76%); mp 131 8C (decomp.). MS (70 eV): m/z (%) 616 (0.4) [M1], 5.88 (1.0) [M1 2 C2H4], 560 (2.3) [M1 2 2C2H4], 532 (0.4) [M1 2 3C2H4], 504 (2.8) [M1 2 4C2H4], 252 (12.3) [RhO3SCF3 1] and 103 (33) [Rh1].trans-[Rh{Á1-OS(O)2CF3}(C2H4)(SbPri 3)2] 28. A suspension of complex 12 (0.122 g, 0.12 mmol) in pentane (5 cm3) was treated at 240 8C with SbPri 3 (0.105 cm3, 0.49 mmol). A red solution was formed which was stirred for 10 min at 240 8C. A slow stream of ethene was then passed through the solution (ca. 1 min) and a yellow solid precipitated.The mother-liquor was decanted, the solid washed three times with 2 cm3 portions of pentane (240 8C) and dried: yield 0.171 g (86%); mp 47 8C (decomp.) (Found: C, 40.35; H, 6.41; S, 3.98. C27H53O3RhSSb2 requires C, 40.33; H, 6.64; S, 3.99%). IR (C6H6): n(O3S) 1263, 1153 and 1105 cm21. NMR: dH (200 MHz, C6D6) 8.05, 6.88 (4 H, both m, C6H4), 3.59 (4 H, br s, C2H4), 2.11 (6 H, m, CHCH3), 1.95 (3 H, s, C6H4CH3) and 1.30 [36 H, d, J(HH) 7.0 Hz, CHCH3]; dC (50.3 MHz, CDCl3) 143.4 (br s, ipso-C of SC6H4), 139.6, 128.7, 126.8 (all s, C6H4), 29.5 (m, C2H4), 22.1 (s, CHCH3) and 19.0 (br s, CHCH3).Alternatively, compound 28 was prepared on treatment of a suspension of 27 (0.098 g, 0.10 mmol) and SbPri 3 (0.098 cm3, 0.40 mmol) in pentane (15 cm3) at 240 8C. Yellow solid: yield 0.142 g (89%). [Rh{Á2-O2S(O)CF3}(C8H14)(PPri 3)] 29. A solution of complex 12 (0.529 g, 0.56 mmol) in pentane (25 cm3) was treated at 0 8C with PPri 3 (0.148 cm3, 1.12 mmol) to give a violet-brown reaction mixture which was stirred for 2 h at room temperature.The resulting orange solution was concentrated to ca. 2 cm2 in vacuo which led to the precipitation of an orange-red microcrystalline solid. The mother-liquor was decanted, the residue washed three times with 5 cm3 portions of pentane (278 8C) and dried in vacuo: yield 0.380 g (65%); mp 36 8C (decomp.) (Found: C, 41.78; H, 6.51; S, 6.01. C18H35F3O3PRhS requires C, 41.38; H, 6.75; S, 6.14%). MS (70 eV): m/z 522 (M1).IR (KBr): n(O3S) 1259, 1249 and 1021, n(CF3) 1244, 1179 and 1169 cm21. NMR (CD2Cl2): dH (400 MHz) 3.01 [2 H, m, J(HH) 9.8, ]] CH of C8H14], 2.13 [2 H, dd, J(HH) 12.3, 3.1, CH2 of C8H14], 1.75J. Chem. Soc., Dalton Trans., 1998, 3549–3558 3557 (3 H, m, CHCH3), 1.55 (2 H, m, CH2 of C8H14), 1.38 (8 H, br m, CH2 of C8H14) and 1.23 [18 H, dd, J(PH) 13.9, J(HH) 7.2 Hz, CHCH3]; dC (100.6 MHz) 120.0 [q, J(CF) 319.4, CF3], 61.1 [d, J(RhC) 18.3, ]] CH of C8H14], 29.7, 28.5, 26.8 (all s, CH2 of C8H14), 23.8 [d, J(PC) 26.4 Hz, CHCH3) and 19.7 (s, CHCH3); dF (376.4 MHz) 278.5 (s); dP (162 MHz) 78.7 [d, J(RhP) 212.8 Hz].[Rh{Á2-O2S(O)CF3}(C2H4)(PPri 3)] 30. A solution of complex 29 (0.197 g, 0.38 mmol) in pentane (10 cm3), prepared in situ from 12 (0.178 g, 0.19 mmol) and PPri 3 (0.074 cm3, 0.38 mmol), was treated at room temperature for 10 s with a stream of ethene which aVorded a yellow suspension. The solvent was decanted, the yellow microcrystalline residue washed three times with 5 cm3 portions of pentane and dried in vacuo: yield 0.114 g (69%); mp 50 8C (decomp.) (Found: C, 32.54; H, 5.72; S, 7.05.C12H25F3O3PRhS requires C, 32.74; H, 5.72; S, 7.28%). NMR (CD2Cl2): dH (400 MHz) 2.75 (4 H, br s, C2H4), 1.83 (3 H, m, CHCH3) and 1.33 [18 H, J(PH) 13.8, J(HH) 7.0 Hz, CHCH3]; dC (100.6 MHz) 119.0 [q, J(FC) 318.5, CF3], 43.7 [d, J(RhC) 15.3, C2H4], 23.1 [d, J(PC) 25.8 Hz, CHCH3] and 19.8 (s, CHCH3); dF (376.4 MHz) 278.4 (s); dP (162 MHz) 69.5 [d, J(RhP) 189.7 Hz].[Rh(Á6-C6H6)(C8H14)(PPri 3)][O3SCF3] 31. A solution of compound 29 (0.089 g, 0.17 mmol) in C6H6 (5 cm3) was stirred for 12 h at room temperature. A yellow solution was formed, which was filtered and the filtrate concentrated to ca. 1 cm3 in vacuo. Addition of pentane (5 cm3) aVorded a yellow suspension which was stored for 2 h. The solvent was then decanted, the yellow microcrystalline residue washed twice with 5 cm3 portions of pentane and dried in vacuo: yield 0.093 g (92%); mp 67 8C (decomp.) (Found: C, 48.06; H, 6.90; Rh, 17.48; S, 5.15.C24H41F3O3PRhS requires C, 47.99; H, 6.89; Rh, 17.15; S, 5.33%). MS-FAB: m/z (%) 451 (0.7) [M1 2 O3SCF3], 373 (1.0) [Rh(C8H14)(PPri 3)]1 and 263 (4.0) [Rh(PPri 3)]1. IR (KBr): n(O3S) 1276 and 1059, n(CF3) 1183 cm21. NMR (CD2Cl2): dH (400 MHz) 6.70 (6 H, s, C6H6), 3.09 [2 H, d, J(RhH) 9.4, ]] CH of C8H14], 2.38 [2 H, dd, J(HH) 9.7, 3.0, CH2 of C8H14], 1.87 (3 H, m, CHCH3), 1.49 (2 H, m, CH2 of C8H14), 1.39 (8 H, br m, CH2 of C8H14) and 1.32 [18 H, dd, J(PH) 14.1, J(HH) 7.4 Hz, CHCH3]; dC (100.6 MHz) 67.9 [d, J(RhC) 14.3 Hz, ]] CH of C8H14], 34.2, 32.4, 26.4 (all s, CH2 of C8H14), 25.5 [d, J(PC) 23.8 Hz, CHCH3] and 19.9 (s, CHCH3); dF (376.4 MHz) 278.0 (s); dP (162 MHz) 63.7 [d, J(RhP) 182.4 Hz].[Rh(Á6-C6H6)(C2H4)(PPri 3)][O3SCF3] 32. This compound was prepared as described for 31, using 30 (0.264 g, 0.60 mmol) as starting material.Pale yellow solid: yield 0.281 g (90%); mp 82 8C (decomp.) (Found: C, 40.81; H, 5.85; S, 6.02. C18H31F3O3PRhS requires C, 41.71; H, 6.03; S, 6.19%). IR (KBr): n(O3S) 1270 and 1028, n(CF3) 1156 cm21. NMR (CD2Cl2): dH (400 MHz) 6.77 (6 H, s, C6H6), 3.33, 2.23 (4 H, both m, C2H4), 1.84 (3 H, m, CHCH3) and 1.20 [18 H, dd, J(PH) 14.1, J(HH) 7.0 Hz, CHCH3]; dC (100.6 MHz) 121.2 [q, J(FC) 321.1 Hz, CF3], 104.5 (br s, C6H6), 40.6 [d, J(RhC) 13.2, C2H4], 25.3 [d, J(PC) 24.4 Hz, CHCH3] and 19.6 (s, CHCH3); dF (376.4 MHz) 278.5 (s); dP (162 MHz) 65.8 [d, J(RhP) 176.3 Hz]. [Rh{Á1-OS(O2)C6H4Me-p}(C]] CHPh)(PPri 3)2] 33.A solution of complex 4 (0.089 g, 0.15 mmol) in toluene (2 cm3) was treated with phenylacetylene (0.017 cm3, 0.15 mmol) and stirred for 6 h at room temperature. A smooth change from red to violet occurred. The solvent was removed, the residue extracted with ether (20 cm3) and the extract brought to dryness in vacuo. The remaining solid was dissolved in acetone (2 cm3) and the solution stored for 12 h at 278 8C.Violet crystals precipitated which were washed twice with 2 cm3 portions of acetone (0 8C) and dried: yield 0.095 g (91%); mp 74 8C (decomp.) (Found: C, 56.61; H, 8.01; S, 4.53. C33H55O3P2RhS requires C, 56.89; H, 7.96; S, 4.60%). MS (ES): m/z (%) 605 (2.7) [M1 2 C6H4CH3]. NMR (C6D6): dH (200 MHz) 7.99–6.85 (9 H, m, C6H5 and C6H4), 2.61 (6 H, m, CHCH3), 1.97 (3 H, s, C6H4CH3), 1.51 [1 H, dt, J(RhH) 1.5, J(PH) 2.9, ]] CHPh] and 1.24 [36 H, dvt, N 13.9, J(HH) 7.3 Hz, CHCH3]; dC (100.6 MHz) 301.1 [dt, J(RhC) 61.0, J(PC) 17.3 Hz, Rh]] C], 143.6, 139.7, 135.2, 128.6, 128.5, 126.7, 125.7, 125.5 (all s, C6H5 and C6H4), 112.2 [dt, J(RhC) 17.3, J(PC) 6.1, =CHPh], 24.1 [vt, N 19.6 Hz, CHCH3], 21.1 (s, C6H4CH3) and 20.3 (s, CHCH3); dP (81.0 MHz) 44.8 [d, J(RhP) 136.6 Hz].Catalytic studies Reactions of Ph2CN2 and C2H4 with complexes 3–6 as catalysts. A solution of a complex (ca. 20 mg, ca. 0.04 mmol) in methylcyclohexane (6 cm3) for 3, 4 or toluene (6 cm3) for 5, 6 was treated dropwise at 40 8C with a 0.5 mol dm23 solution of diphenyldiazomethane in methylcyclohexane while bubbling ethene through the solution. The catalytic reaction was finished when the violet colour of the diazoalkane solution did not disappear on further addition to the reaction mixture. The solvent was removed in vacuo, and the oily residue dissolved in 2–3 cm3 of hexane. In order to destroy the excess of Ph2CN2 and separate the catalyst, the mixture was filtered through Al2O3 (neutral, activity grade III, height of column 3 cm).After evaporation of the solvent, an oil containing a mixture of Ia–Id was isolated from the eluate. The ratio of the products was determined by integration of characteristic signals in the 1H NMR spectra and by GC-MS analysis. The results are summarized in Table 4. Reaction of Ph2CN2 and methyl acrylate with complex 5 as catalyst. In an analogous manner to the catalytic reaction of Ph2CN2 and ethene, a solution of complex 5 (17 mg, 0.03 mmol) and methyl acrylate (2.6 cm3, 3.0 mmol) in methylcyclohexane (6 cm3) was treated dropwise at 40 8C with a 0.1 mol dm23 solution of diphenyldiazomethane in toluene.After work-up, a clear oil was isolated and characterized by 1H NMR data and GC-MS analysis as IIa: yield 390 mg (1.55 mmol). If instead of methyl acrylate the corresponding ester CH2]] CHCH2CO2Me was used as the substrate, a small quantity (11 mg, 0.04 mmol) of an oV-white oil was isolated.It was characterized by 1H NMR data and GC-MS analysis as IIb. Crystallography Single crystals of complex 4 were grown from acetone (8 8C), those of 29 from pentane (20 8C) and those of 31 from benzene (20 8C). Crystal data collection parameters are summarized in Table 5. Intensity data were corrected for Lorentz-polarization eVects. Data reduction were performed for 4 and 31 with SDP 26 and for 29 with Stoe IPDS software.The structures were solved by direct methods (SHELXS 86).27 For 29 and 31 two independent molecules (A and B) were found in the asymmetric units with diVerent conformations of the cyclooctene ligand. In Figs. 1 and 2 only molecule A of 29 and 31, respectively, is shown. Table 5 contains the crystallographic data of each whole asymmetric unit (molecule A and B), the chemical formula and the formula weight, however, belong to one molecule only. Atomic coordinates and anisotropic displacement parameters of the non-hydrogen atoms were refined anisotropically by fullmatrix least squares on F2 (SHELXL 93).28 The positions of the hydrogen atoms [except of H(20), H(21), H(40) and H(41) in 29 and C(30), H(31), H(40) and H(41) in 31] were calculated according to ideal geometry using the riding method.CCDC reference number 186/1157. See http://www.rsc.org/suppdata/dt/1998/3549/ for crystallographic files in .cif format.3558 J.Chem. Soc., Dalton Trans., 1998, 3549–3558 Table 5 Crystal data for complexes 4, 29 and 31 Formula M Crystal system Space group a/Å b/Å c/Å a/8 b/8 g/8 U/Å3 T/K Z Dc/g cm23 l(Mo-Ka)/Å m/mm21 No. reflections measured No. unique reflections (Rint) R1 b wR2c 4 C25H49O3P2RhS 594.55 Monoclinic P21/c (no. 14) 10.498(4) 14.104(3) 20.276(7) — 92.89(2) — 2998(2) 293 4 1.317 0.71073 0.761 4422 4156 (0.0127) 0.0266 0.0717 29 C18H35F3O3PRhS 522.40 Monoclinic P21/c (no. 14) 14.177(3) 17.626(6) 19.038(6) — 103.61(3) — 4624(2) 173 8 1.501 0.71073 0.928 36543 8695 (0.0746) 0.0334 0.0683 31 a C24H41F3O3PRhS 600.51 Triclinic P1� (no. 2) 14.042(4) 15.347(4) 15.532(2) 63.12(2) 73.03(2) 70.98(2) 2781(1) 293 4 1.434 0.71073 0.782 9104 8703 (0.0096) 0.0294 0.0751 a For complex 31 an extinction parameter was refined to (5.98 ± 0.23) × 1023. b R = S|Fo 2 Fc|/SFo [for Fo > 2s(Fo)] for the number of observed reflections [I > 2s(I)], respectively. c wR2 = [Sw(Fo 2 2 Fc 2)2/Sw(Fo 2)2]� �� ; w21 = [s2(Fo 2) 1 (0.0377P)2 1 1.1100P] (4), [s2(Fo 2) 1 (0.0289P)2 1 0.0000P] (29), [s2(Fo 2) 1 (0.0347P)2 1 2.5370P] (31), where P = (Fo 2 1 2Fc 2)/3; for all data reflections, respectively.Acknowledgements We thank the Deutsche Forschungsgemeinschaft (SFB 347) and the Fonds der Chemischen Industrie for financial support, the latter in particular for Ph.D. grants to M. E. S. and M. M. We are also grateful to Mrs R. Schedl and Mr C. P. Kneis (DTA measurements and elemental analyses), to Dr G.Lange and Dr M. Herderich (mass spectra), to Degussa AG and BASF AG (chemicals), and to Dr J. Wolf for numerous valuable discussions. References 1 J. Wolf, L. Brandt, A. Fries and H. Werner, Angew. Chem., 1990, 102, 584; Angew. Chem., Int. Ed. Engl., 1990, 29, 510; H. Werner, J. Organomet. Chem., 1994, 475, 45. 2 A. J. Anciaux, A. J. Hubert, A. F. Noels, N. Petiniot and P. Teyssié, J. Org. Chem., 1980, 45, 695; M. P. Doyle, Acc. Chem. Res., 1986, 19, 348; M.P. Doyle, W. R. Winchester, J. A. A. Hoorn, V. Lynch, S. H. Simonsen and R. Ghosh, J. Am. Chem. Soc., 1993, 115, 9968. 3 L. Brandt, Ph.D. Thesis, Universität Würzburg, 1991. 4 L. Brandt, A. Fries, N. Mahr, H. Werner and J. Wolf, Selective Reactions of Metal-Activated Molecules, eds. H. Werner, A. G. Griesbeck, W. Adam, G. Bringmann and W. Kiefer, Vieweg Verlag, Braunschweig, 1992, p. 171. 5 M. Schäfer, J. Wolf, H. Werner, J. Chem. Soc., Chem. Commun., 1991, 1341; H. Werner, M. Schäfer, O. Nürnberg and J. Wolf, Chem. Ber., 1994, 127, 27; M. Schäfer, J. Wolf and H. Werner, J. Organomet. Chem., 1994, 476, 85. 6 H. Werner, S. Poelsma, M. E. Schneider, B. Windmüller and D. Barth, Chem. Ber., 1996, 129, 647. 7 M. E. Schneider and H. Werner, 10. International Symposium on Homogeneous Catalysis, Princeton, 1996. 8 L. S. Stuhl and E. L. Muetterties, Inorg. Chem., 1978, 17, 2148. 9 G. A. Lawrance, Chem. Rev., 1986, 86, 17. 10 K. Wang, G. P. Rosini, S. P. Nolan and A. S. Goldman, J. Am. Chem. Soc., 1995, 117, 5082. 11 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. 12 V. V. Grushin, V. F. Kuznetsov, C. Bensimon and H. Alper, Organometallics, 1995, 14, 3927. 13 O. Gevert, J. Wolf and H. Werner, Organometallics, 1996, 15, 2806. 14 H. Werner, F. J. Garcia Alonso, H. Otto and J. Wolf, Z. Naturforsch., Teil B, 1988, 43, 722; H. Werner and U. Brekau, Z. Naturforsch., Teil B, 1989, 44, 1438; T. Rappert, O. Nürnberg, N. Mahr, J. Wolf and H. Werner, Organometallics, 1992, 11, 4156. 15 U. Möhring, M. Schäfer, F. Kukla, M. Schlaf and H. Werner, J. Mol. Catal. A, 1995, 99, 55. 16 M. Aresta, E. Quaranta and A. Albinati, Organometallics, 1993, 12, 2032. 17 For a recent review on stibine transition-metal complexes see N. R. Champness and W. Levason, Coord. Chem. Rev., 1994, 133, 115. 18 P. Schwab, N. Mahr, J. Wolf and H. Werner, Angew. Chem., 1993, 105, 1498; Angew. Chem., Int. Ed. Engl., 1993, 32, 1480; H. Werner, P. Schwab, E. Bleuel, N. Mahr, P. Steinert and J. Wolf, Chem. Eur. J., 1997, 3, 1375. 19 H. Werner, J. Organomet. Chem., 1995, 500, 331. 20 J. Halpern, D. P. Riley, A. S. C. Chan and J. J. Pluth, J. Am. Chem. Soc., 1977, 99, 8055; E. T. Singewald, C. S. Slone, C. L. Stern, C. A. Mirkin, G. P. A. Yap, L. M. Liable-Sands and A. L. Rheingold, J. Am. Chem. Soc., 1997, 119, 3048; M. Manger, Ph.D. Thesis, Universität Würzburg, 1997. 21 P. Binger and U. Schuchardt, Angew. Chem., 1977, 89, 254; Angew. Chem., Int. Ed. Engl., 1977, 16, 249; Chem. Ber., 1981, 114, 3313. 22 Y. Alvarado, O. Boutry, E. Gutiérrez, A. Monge, C. M. Nicasio, M. L. Poveda, P. J. Pérez, C. Ruíz, C. Bianchini and E. Carmona, Chem. Eur. J., 1997, 3, 860; P. J. Perez, M. L. Poveda and E. Carmona, J. Chem. Soc., Chem. Commun., 1992, 8. 23 Y. Alvarado, P. J. DaV, P. J. Perez, M. L. Poveda, R. Sanchez- Delgado and E. Carmona, Organometallics, 1996, 15, 2192. 24 P. J. Perez, M. L. Poveda and E. Carmona, Angew. Chem., 1995, 107, 242; Angew. Chem., Int. Ed. Engl., 1995, 34, 231. 25 A. van der Ent and A. L. Onderdelinden, Inorg. Synth., 1973, 14, 92. 26 B. A. Frenz, The Enraf-Nonius CAD4 SDP, a real time system for concurrent X-ray data collection and structure determination, in Computing in Crystallography, Delft University Press, Delft, 1978, p. 64. 27 G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467. 28 G. M. Sheldrick, SHELXL 93, A program for crystal structure refinement, University of Göttingen, 1993. Paper

ISSN:1477-9226    

DOI:10.1039/a805900d

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3559–3564 3559 Stoichiometric enantioselective alkene epoxidation with a chiral dioxoruthenium(VI) D4-porphyrinato complex Tat-Shing Lai,a Hoi-Lun Kwong,b Rui Zhang a and Chi-Ming Che *a a Department of Chemistry, The University of Hong Kong, Pokfulam Road, Hong Kong, P. R. China b Department of Biology and Chemistry, City University of Hong Kong, Tat Chee Avenue, Kowloon Tong, Hong Kong, P. R. China Received 6th April 1998, Accepted 29th July 1998 A dioxoruthenium(VI) complex containing a D4-porphyrinato ligand por* {H2por* = 5,10,15,20-tetrakis- [(1S,4R,5R,8S)-1,2,3,4,5,6,7,8-octahydro-1,4 : 5,8-dimethanoanthracen-9-yl]porphyrin} has been prepared by oxidation of its ruthenium(II) carbonyl precursor with m-chloroperoxybenzoic acid and characterised by spectroscopic methods.The [RuVI(por*)O2] complex undergoes enantioselective epoxidation of alkenes and the highest enantiomeric excess (ee) attainable is 77%.In the presence of pyrazole the complex transforms to [RuIV(por*)(pz)2] when reacting with alkenes. The kinetics of the epoxidation of para-substituted styrenes has been studied. The experimental rate law is 2d[RuVI]/dt = k2[RuVI][alkene]. The second order rate constants k2 at 25 8C fall in a narrow range, 2.1 × 1023–9.7 × 1023 dm3 mol21 s21. Comparison of the Hammett plot (log krel vs. s?) with those for achiral analogues [RuVI(tpp)O2] (H2tpp = 5,10,15,20-tetraphenylporphyrin) and [RuVI(oep)O2] (H2oep = 2,3,7,8,12,13,17,18-octaethylporphyrin) suggests the formation of a radical intermediate for the alkene epoxidations.Both [RuII(por*)(CO)(EtOH)] and [RuVI(por*)O2] were examined for enantioselective catalysis. Enantioselectivities of the stoichiometric and catalytic reactions showed good correlation. There is no solvent dependence on enantioselectivity when changing the solvent from dichloromethane to benzene. Functionalization of hydrocarbons via epoxidation with transition metal catalysts has continued to receive great attention.1,2 In particular, there has been considerable interest in the oxidation chemistry of metalloporphyrins owing to their biological relevance to enzymatic reactions mediated by cytochrome P-450.Owing to the periodic relationship of ruthenium to iron, ruthenium porphyrins have been of recent interest. They have been shown to be active catalysts for epoxidation,3 aziridination 4 and cyclopropanation 5 of alkenes and hydroxylation of alkanes.6 More importantly, reactive oxo,7 imido8 and carbene 5a,b ruthenium porphyrinato complexes have been isolated and/or spectroscopically characterized.This renders direct measurement of rates of group and atom transfer reactions feasible. Although oxometal species are tacitly agreed to be the active intermediates in many discussions of asymmetric alkene epoxidations,1b their characterization is elusive. Indeed, in some of the reported metalloporphyrin catalysed organic oxidations, more than one reactive intermediate has been suggested.We anticipated that isolation of highly reactive and chiral oxometal porphyrin complexes and examination of their reactivities towards asymmetric alkene epoxidations should partly provide a clue to the factors aVecting enantioselectivity. In our laboratory, a chiral monooxoruthenium(IV) complex supported by a C2-chiral meridional ligand has been found to epoxidize styrene-type substrates stoichiometrically with enantioselectivity up to 57% enantiomeric excess (e.e.).9 Here, we report a detailed study of enantioselective epoxidation with a chiral dioxoruthenium(VI) complex, [RuVI(por*)O2], containing a D4- porphyrin ligand, H2por*{5,10,15,20-tetrakis[(1S,4R,5R,8S)- 1,2,3,4,5,6,7,8-octahydro-1,4 : 5,8-dimethanoanthracen-9-yl]- porphyrin}.Its use in stoichiometric asymmetric epoxidations and the kinetics of its reactions with para-substituted styrenes are also presented.The Hammett plots of the achiral [Ru- (tpp)O2] (H2tpp = 5,10,15,20-tetraphenylporphyrin) and [Ru- (oep)O2] (H2oep = 2,3,7,8,12,13,17,18-octaethylporphyrin) complexes are compared with that for this chiral analogue. Interpretation of these kinetic data provides a new picture for understanding the mechanism of asymmetric epoxidation by oxometal species. In this work the results of catalytic epoxidation of alkenes with PhIO as a terminal oxidant will be presented and compared with the stoichiometric ones.10 Experimental Preparation of compounds Ruthenium dodecacarbonyl [Ru3(CO)12] was purchased from Strem, meta-chloroperoxybenzoic acid (Merck, 85%) was used as received and H2por* was prepared according to the literature method.11 All solvents for the syntheses were of analytical grade.[RuII(por*)(CO)(EtOH)]. A mixture of H2por* (100 mg) and [Ru3(CO)12] (120 mg) in decalin (50 cm3) was refluxed under an inert atmosphere for 36 h.After cooling, the orange solution was chromatographed on a silica gel column. The product was3560 J. Chem. Soc., Dalton Trans., 1998, 3559–3564 eluted by dichloromethane as an orange band. Removal of the solvent and recrystallization of the crude product using CH2Cl2–EtOH gave a red crystalline solid. Yield 80%. 1H NMR (300 MHz, CDCl3): d 22.56 (s, 2 H), 1.05 (m, 8 H), 1.32 (m, 24 H), 1.85 (m, 8 H) 2.00 (d, 8 H, J = 8.0 Hz), 2.77 (s, 8 H), 3.56 (s, 8 H), 7.36 (s, 4 H) and 8.72 (s, 8 H).IR (KBr): 2969, 1942 and 1918 cm21. UV: 414 and 529 nm. FAB MS: m/z 1270 (M1) and 1242 (M1 2 CO). [RuVI(por*)O2]. A dichloromethane solution of [RuII(por*)- (CO)(EtOH)] (50 mg) was added to a well stirred solution of m-chloroperoxybenzoic acid in dichloromethane (100 mg in 15 cm3). After 3 min the solution was chromatographed on a short alumina column. The product was eluted by dichloromethane. The solution obtained was evaporated to dryness by rotatory evaporation.A dark purple residue (35 mg) was obtained. Yield 70%. IR (KBr): 2960vs, 2867s, 1684m, 1559s, 1292vs, 1106m, 1076m, 1019m, 965w, 948w, 822s, 797s, 754m and 705s cm21. 1H NMR (300 MHz, CDCl3): d 1.12 (m, 8 H), 1.35 (m, 24 H), 1.88 (m, 8 H), 2.05 (m, 8 H), 2.88 (s, 8 H), 3.62 (s, 8 H), 7.44 (s, 4 H) and 8.96 (s, 8 H). 13C NMR (300 MHz, CDCl3): d 27.183, 27.565, 42.417, 44.403, 49.339, 113.988, 118.923, 127.908, 130.780, 141.533, 144.272 and 148.062.UV/VIS: 424 nm (log e 5.38). FAB MS: m/z 1274 (M1), 1258 (M1 2 O) and 1242 (M1 2 2O). Instrumentation Ultraviolet-visible spectra were recorded on a HPUV 8452 spectrophotometer, infrared spectra as KBr discs on a Bio-Rad FTIR spectrophotometer. The GLC analyses were performed on a HP GC instrument equipped with a flame ionization detector. The NMR spectra were recorded on a DPX300 spectrophotometer. The chiral columns for separation of enantiomers were J & W cyclodex-B (30 m) and G-TA (30 m).The enantiomeric excesses of 3-nitrostyrene oxide and 4-methylstyrene oxide were determined by 1H NMR in the presence of the chiral shift reagent {[Eu(hfc)3] = tris[3-(heptafluoropropylhydroxymethylene)- D-camphorato]europium(III)} (Aldrich). The temperatures of kinetic measurements were stabilized with a thermostat (±1.0 8C). Stoichiometric oxidation of alkenes by [RuVI(por*)O2] and isolation of [RuIV(por*)(pz)2] Alkene (0.2 g) and pyrazole (0.05 g) were dissolved in dichloromethane (5 cm3). The [RuVI(por*)O2] complex (50 mg) was added with stirring and the resulting solution stirred for 12 h.Organic products were obtained through column chromatography with Et2O–light petroleum (1: 10) as eluent and [RuIV(por*)(pz)2] was eluted with dichloromethane. The organic products were analysed by 1H NMR and/or GC. The complex [RuIV(por*)(pz)2] was characterized by UV/VIS, IR, MS and magnetic moment measurements. Yield 72%. IR (KBr): 2957vs, 2865vs, 1639m, 1518w, 1469m, 1445m, 1293s, 1193m, 1106s, 1064m, 1009s, 948m, 862w, 796s, 755m and 710m cm21.FAB MS: m/z 1376 (M1) and 1242 (M1 2 2pz). UV/VIS (CH2Cl2): lmax 414 and 512 nm. meff(solid sample, r.t.): 2.9 mB. Alternatively, [RuVI(por*)O2] was generated in situ. A mixture of PhIO (0.1 g) and [RuII(por*)(CO)(EtOH)] (0.1 g) in CH2Cl2 was stirred for 10 min. The complete conversion of [RuII- (por*)(CO)(EtOH)] into [RuVI(por*)O2] was confirmed by electronic absorption spectroscopy. The solution was then injected into a solution of alkene containing 3 equivalents of pyrazole.The yield of organic product was calculated on the assumption of a RuVI æÆ RuIV transformation. Kinetic measurements Dichloromethane was distilled over CaH2. Alkenes were from commercial sources and purified by distillation or chromatography. The rates of reduction of [RuVI(por*)O2] by alkenes were measured by monitoring the decrease in absorbance of a 1,2-dichloroethane solution of the complex in the presence of 2% pyrazole at 424 nm. The reactions were carried out with [alkene] @ [RuVI] (at more than 1000 fold). Plots of ln |A• 2 At| vs.time were linear over at least three half-lives. The pseudo- first order rate constants (kobs) were determined on the basis of least squares fits using eqn. (1) where A• and At are the absorbln |A• 2 At | = 2kobst 2 ln |A• 2 A0 | (1) ance at the completion of reaction and at time t respectively; A• readings were obtained after at least four half-lives.Second order rate constants (k2) were determined from plots of kobs vs. [alkene]. Oxidation of alkenes by PhIO catalysed by [RuVI(por*)O2] or [RuII(por*)(CO)(EtOH)] A mixture of substrate (100 mg), PhIO (50 mg) and catalyst (2 mg) in dichloromethane (4 cm3) was stirred with strict exclusion of air. The reaction was completed when all PhIO solid dissolved. The GC analysis of the reaction mixture was carried out with halogenated aromatics such as 1-bromo-4-chlorobenzene as internal standard.The yields were calculated using PhIO as the limiting reactant. Results and discussion Synthesis of [RuVI(por*)O2] Preparation of dioxoruthenium(VI) porphyrin through oxidation of its ruthenium(II) carbonyl precursor by PhIO or mchloroperoxybenzoic acid in CH2Cl2 or CH2Cl2–alcohol has previously been reported.3b The literature method works for a variety of porphyrin ligands. For the non-bulky octaethylporphyrin alcohol is needed to suppress the m-oxo dimer formation. 3c The H2por* ligand used in this work is bulky and hence dimerization via Ru–O–Ru formation is not favored. Therefore only dichloromethane was used as the solvent for the synthesis. The [RuVI(por*)O2] complex was obtained in a high yield by treating m-chloroperoxybenzoic acid with [RuII(por*)(CO)- (EtOH)]; the structures for both [RuVI(por*)O2] 12 and [RuII- (por*)(CO)(EtOH)] 6 have been determined by X-ray crystallography.The [RuVI(por*)O2] complex is air stable, diamagnetic and shows no manifest spectroscopic changes when dissolved in purified CH2Cl2 for hours at room temperature. It can be stored as a solid at 220 8C for months. It was characterized by 1H, 13C NMR, IR and UV/VIS spectroscopy. The nasym (O]] Ru]] O) occurs at 822 cm21 which falls into the range reported for other achiral trans-dioxoruthenium(VI) porphyrins and the oxidation marker of [RuVI(por*)O2] at 1019 cm21 is in accordance with the ruthenium(VI) formation.13 Comparing the 1H NMR spectra of [RuII(por*)(CO)(EtOH)] with [RuVI- (por*)O2], the double doublet of the pyrrolic protons of the former converge to a singlet signal in the latter, as expected for the change of symmetry of the molecule from C4 to D4.Stoichiometric oxidation Oxygen atom transfer from [RuVI(por*)O2] to alkenes occurs readily at room temperature with the organic epoxides being the major products. Results of stoichiometric alkene epoxidations are summarized in Table 1.The reactions were carried out in two solvents, CH2Cl2 and benzene, with higher product yields found in the latter solvent. Lowering the temperature was found to lower the epoxide yield. The enantiomeric excess (e.e.) of the epoxides ranged from 40 to 77% with the highest value for 1,2- dihydronaphthalene oxide obtained in the reaction of [RuVI- (por*)O2] with 1,2-dihydronaphthalene in dichloromethane at 215 8C. To our knowledge, this is the highest e.e.attainable in stoichiometric alkene epoxidation using well characterizedJ. Chem. Soc., Dalton Trans., 1998, 3559–3564 3561 Table 1 Stoichiometric epoxidation of alkenes by [RuVI(por*)O2]. Entry 1 2 3 4 5 Substrate Product Solvent CH2Cl2 a C6H6 a CH2Cl2 C6H6 a CH2Cl2 C6H6 CH2Cl2 CH2Cl2 C6H6 C6H6 a CH2Cl2 CH2Cl2 C6H6 Temperature r.t. r.t. r.t. r.t. r.t. r.t. r.t. 215 8C r.t. r.t. r.t. 215 8C r.t. % Yield 43 61 32 71 35 41 72 (cis/trans = 9.9) 68 (cis/trans = 11) 70 (cis/trans = 8.3) 64 (cis/trans = 11) 32 30 41 % e.e. 59 (R) 65 (R) 50 (R) 45 (R 41 40 67 (1R,2S) 70 (1R,2S) 67 (1R,2S) 72 (1R,2S) 71 77 72 a In the presence of 20 mg pyrazole. chiral oxometal complexes.9 The enantioselectivity shows little solvent dependence while lowering the reaction temperature tends to give higher e.e. values. In general, addition of pyrazole to the reaction mixture has little eVect on the enantioselectivity except for the epoxidation of styrene to styrene oxide.In the epoxidation of para-substituted styrenes the side products were benzaldehydes and arylacetaldehydes. Changing the substrate from styrene to p-chlorostyrene lowered the e.e. from 65 to 45%. The absolute configurations of styrene and chlorostyrene epoxides were determined to be (R). In the reaction with cis-bmethylstyrene, cis-b-methylstyrene oxide was obtained in a high yield and with good % e.e. Besides, the epoxidation was highly stereospecific with a stereoretention of more than 90%.The absolute configuration of cis-b-methylstyrene epoxide was determined to be (1R,2S), thus, the configuration of the carbons bearing the phenyl group are the same for both styrene and cis-b-methylstyrene. Kinetics of the reduction of [RuVI(por*)O2] by alkenes When [RuVI(por*)O2] reacted with styrene in dichloromethane in the absence of pyrazole the spectral trace did not exhibit isosbestic points. This can be attributed to the accumulation and/or disproportionation of the [RuIV(por*)O] intermediate formed during the oxidation (Scheme 1).Indeed, Groves and co-workers 13 had reported that [RuIV(tmp)O] (H2tpm = 5,10, 15,20-tetramesitylporphyrin) is unstable and undergoes disproportionation in solution at room temperature. However, in the presence of pyrazole, isosbestic spectral changes for the reaction of [RuVI(por*)O2] with styrene in dichloromethane were found (Fig. 1). The final ruthenium product was [RuIV(por*)(pz)2] when the reaction was carried out in the presence of 3 equivalents of pyrazole.As expected for paramagnetic ruthenium(IV) complexes, the meff of [RuIV(por*)(pz)2] was found to be 2.9 mB. Similar reaction of [RuVI(dpp)O2] (H2dpp = 2,3,5,7,8,10,12,13,15,17,18,20-dodecaphenylpor- Scheme 1 phyrin) with alkenes in the presence of pyrazole was also found to give [RuIV(dpp)(pz)2], the structure of which was determined by X-ray crystallography.14 The observation of isosbestic spectral traces indicates no accumulation of intermediate in the conversion from [RuVI(por*)O2] to [RuIV(por*)- (pz)2].We propose that [RuIV(por*)O] was the immediate product of the oxygen atom transfer from [RuVI(por*)O2] to alkene, which rapidly reacted with pyrazole to give [RuIV(por*)- (pz)2] (Scheme 2). The kobs values, however, are independent of the pyrazole concentrations. Under the condition [alkene] @ [RuVI] and [alkene] < 2 mol dm23, pseudo-first order decay of Fig. 1 Spectral trace of the reaction of styrene (0.20 mol dm23) with [RuVI(por*)O2] in the presence of pyrazole (3.0%) in 1,2-dichloroethane at 298 K; scan interval 90 s. Scheme 23562 J. Chem. Soc., Dalton Trans., 1998, 3559–3564 the absorbance at 424 nm (Soret band) was observed. The observed first order rate constants, kobs, display a linear dependence on [alkene] and the rate law (1) was established. 2d[RuVI]/dt = k2[RuVI] [alkene] (1) The second order rate constants (k2) were determined by measuring the slope of the plots of kobs vs.[alkene]. The kinetic data are summarized in Table 2. The k2 for the oxidation of styrene is 2.19 × 1023 dm3 mol21 s21. It is comparable to that by [RuVI(oep)O2] (k2 = 1.55 × 1023 dm3 mol21 s21) and [RuVI- (tpp)O2] (k2 = 4.30 × 1023 dm3 mol21 s21). With the exception of p-bromo- and p-methoxy-styrenes, the k2 values for the oxidation of para-substituted styrenes by [RuVI(por*)O2] fall into a narrow range.Similar observations have previously been reported for the two achiral [RuVI(oep)O2] and [RuVI(tpp)O2] systems.7 Previous studies showed that oxidation of alkenes by Table 2 Second order rate constants for the oxidation of alkenes by [RuVI(por*)O2] in 1,2-dichloroethane at 298 K. Entry 1 2 3 4 5 6 7 8 9 10 11 12 Alkene k2/dm3 mol21 s21 0.00219 0.00262 0.00459 0.00361 0.00965 0.00720 0.000340 0.00173 0.00289 0.00560 0.0012 0.0016 log krel 0 0.0779 0.3214 0.2171 0.6441 0.5169 s1 0 20.0073 0.114 20.311 20.778 TE 0 0.16 0.23 0.42 1.08 krel = k2(para-substituted styrene)/k2(styrene) and k2 is the rate of oxidation of various alkenes by the dioxo complex.the achiral dioxoruthenium(VI) porphyrin [RuVI(tpp)O2] exhibited a linear free-energy plot of log k2 vs. E2� 1 (one electron potential of alkenes) with a slope of 21.1 V21. Furthermore, the oxidation of substituted styrenes by [RuVI(tpp)O2] and [RuVI(oep)O2] exhibited U-shaped Hammett plots. These observations were attributed to little degree of charge transfer in the transition state and a mechanism involving a continuum of transition states was suggested.In this report, we adopt the (TE = total substituent eVect parameter) values developed by Wu et al.15 to construct the Hammett plots. Hammett plots for three trans-dioxoruthenium( VI) porphyrin complexes are shown in Fig. 2(a)–(c). The data for the [RuVI(tpp)O2] and [RuVI(oep)O2] complexes were taken from the literature paper.For each case, reasonable obedience to a linear correlation is obtained. The best linear fit is found for the [RuVI(oep)O2] system probably due to minimum steric complication. Thus, we propose that the rate-determining step of the oxygen atom transfer involves a loosely bound radical intermediate as depicted in Scheme 3. The involvement Fig. 2 Hammett plots for the oxidation of para-substituted styrenes by (a) [RuVI(por*)O2], (b) [RuVI(tpp)O2] and (c) [RuVI(oep)O2].J.Chem. Soc., Dalton Trans., 1998, 3559–3564 3563 Table 3 Catalytic epoxidation of alkenes by PhIO with [RuII(por*)(CO)(EtOH)] (A) and [RuVI(por*)O2] (B) as catalyst. Yield (%) Entry 1 2 3 4 5 6 7 8 Substrate Catalyst AAB AAB AAB AA AAB AA AAB AB Solvent CH2Cl2 C6H6 CH2Cl2 CH2Cl2 C6H6 CH2Cl2 CH2Cl2 C6H6 CH2Cl2 CH2Cl2 C6H6 CH2Cl2 C6H6 CH2Cl2 CH2Cl2 C6H6 CH2Cl2 C6H6 CH2Cl2 CH2Cl2 CH2Cl2 Arylacetaldehyde 5.5 34 7 20 11 Trace Trace Trace 0 ——— —— — —— Benzaldehyde 21 42 23 30 40 42 32 38 35 38 12 19 21 —— 47 52 61 —— Epoxide 71 57 52 51 35 41 51 66 53 40 59 (cis/trans = 9.9) 52 (cis/trans = 6.3) 53 (cis/trans = 11) 62 46 45 41 31 55 61 % e.e. 55 (R) 63 (R) 51 (R) 40 40 38 41 (R) 51 (R) 54 52 58 cis (1R,2S) 52 cis (1R,2S) 55 cis (1R,2S) 30 62 16 trans 17 trans 13 trans 87 of this radical intermediate can be used to explain the production of trans-b-methylstyrene oxide in the epoxidation of cis-bmethylstyrene as bond rotation of this carbon center radical can lead to loss of stereochemistry (Scheme 4).Its existence can also explain the rearrangement product, 2-arylacetaldehyde, in the epoxidation of styrene-type substrates through a hydrogenatom 1,2-shift process (Scheme 3). A radical mechanism has recently been suggested to be involved in the Mn(salen) catalysed epoxidation reaction.16 The oxidation of (1) and (2)-limonenes by [RuVI(por*)O2] Scheme 3 Scheme 4 were studied to see if this chiral oxidant is capable of diVerentiating a pair of chiral substrates (Table 2, entries 11 and 12).In this work the rate constant for the reduction of [RuVI(por*)O2] by (2)-limonene was found to be higher than that by (1)- limonene by a factor of ª1.3. Catalytic epoxidation using iodosylbenzene as oxidant Both [RuII(por*)(CO)(EtOH)] and [RuVI(por*)O2] were examined as catalysts for enantioselective alkene epoxidation using PhIO as a terminal oxidant. Both catalysts showed enantioselectivity for a wide range of alkene substrates.The best e.e. values were observed with styrene (63% e.e.) (Table 3). In the epoxidation of para-substituted styrenes, benzaldehydes and arylacetaldehydes were found as side products. The epoxide yields are in the range of 31–71%. Unlike the stoichiometric reactions the yields are higher in CH2Cl2 than in benzene. The absolute configuration of styrene and chlorostyrene oxides were determined to be (R). The epoxidation of cis-b-methylstyrene was highly stereospecific with up to 91% stereoretention in dichloromethane. The absolute configuration of cis-b-methylstyrene oxide was determined to be (1R,2S).In the cases of trans-b-methylstyrene and 1-phenylcyclohexene, low e.e. values, 17 and 8% respectively, were obtained. Unlike the results obtained with the ruthenium(II) D2-porphyrin catalyst,17 there is only a little diVerence in e.e. of the organic epoxides when changing the solvent from CH2Cl2 to benzene. Comparison between stoichiometric and catalytic epoxidation by [RuVI(por*)O2] The comparison of the e.e.values of organic epoxides obtained in stoichiometric and in catalytic alkene epoxidations is an objective of this work. In the catalytic styrene epoxidation, arylacetaldehyde and benzaldehyde were obtained as side products. This is similar to the stoichiometric reaction. Of the five alkenes (styrene, 4-chlorostyrene, 4-methylstyrene, cis-bmethylstyrene and 1,2-dihydronaphthalene) studied in this work similar e.e.values for both the catalytic and stoichiometric3564 J. Chem. Soc., Dalton Trans., 1998, 3559–3564 reactions were obtained. This strongly argues that the [RuVI- (por*)O2] complex is the major oxidizing active intermediate in the catalytic epoxidation. Conclusion In this work the first detailed study of a well defined chiral trans-dioxoruthenium(VI) porphyrin complex as a stoichiometric oxidant for epoxidation of alkenes is presented. Its use as a catalyst for alkene epoxidation by PhIO has also been examined.The linearity of Hammett plots (log krel vs. TE) observed for the oxidation of para-substituted styrenes by the [RuVI- (por*)O2] complex supports a carbon center radical intermediate formed during the course of oxygen atom transfer reactions. The parallel enantioselectivity between the catalytic and stoichiometric epoxidations suggests that [RuVI(por*)O2] is the major oxidizing intermediate in the catalytic alkene epoxidation by PhIO.Acknowledgements We acknowledge support from the Hong Kong Research Grants Council, the University of Hong Kong and City University of Hong Kong. References 1 (a) D. Dolphin, T. G. Traylor and L. Y. Xie, Acc. Chem. Res., 1997, 30, 251; (b) J. P. Collman, X. Zhang, V. J. Lee, E. S. UVelman and J. I. Brauman, Science, 1993, 261, 1404; (c) K. Drauz and H. Waldmann, Enzyme Catalysis in Organic Synthesis, A comprehensive handbook, VCH, Weinheim, 1995, vol. 2, p. 680. 2 J. R. Valbert, J. G. Zajacek and D. I. Orenbuch, in Encyclopedia of Chemical Processing and Designing, ed. J. McKetta, Marcel Dekker, New York, 1993, vol. 45, pp. 88–137. 3 (a) Z. Gross, S. Ini, M. Kapon and S. Cohen, Tetrahedron Lett., 1996, 40, 7325; (b) J. T. Groves and R. Quinn, J. Am. Chem. Soc., 1985, 107, 5790; (c) W.-H. Leung, C.-M. Che, C.-H. Yeung and C.-K. Poon, Polyhedron, 1993, 12, 2331. 4 C.-M. Che and M.-C. Cheng, unpublished work. 5 (a) W. C. Lo, C. M. Che, K. F. Cheng and T. C. W. Mak, Chem. Commun., 1997, 1205; (b) E. Galardon, P. Le Maux and G. Simonneaux, Chem. Commun., 1997, 927. 6 T. Shingaki, K. Miura, T. Higuchi, M. Hirobe and T. Nagano, Chem. Commun., 1997, 861; J. T. Groves, M. Bonchio, T. Carofiglio and K. Shalyaev, J. Am. Chem. Soc., 1996, 118, 8961. 7 C. Ho, W.-H. Leung and C.-M. Che, J. Chem. Soc., Dalton Trans., 1991, 2933. 8 W.-H. Leung, T. S. M. Hun, H.-W. Hou and K.-Y. Wong, J. Chem. Soc., Dalton Trans., 1997, 237; S. M. Au, W. H. Fung, M. C. Cheung, C.-M. Che and S.-M. Peng, Chem. Commun., 1997, 1655. 9 W.-H. Fung, W.-C. Cheng, W.-Y. Yu, C.-M. Che and T. C. Mak, J. Chem. Soc., Chem. Commun., 1995, 2007. 10 A. Berkessel and M. Frauenkron, J. Chem. Soc., Perkin Trans. 1, 1997, 2265. 11 R. L. Halterman d S.-T. Jan, J. Org. Chem., 1991, 56, 5253. 12 T.-S. Lai, R. Zhang, K.-K. Cheung, H.-L. Kwong and C.-M. Che, Chem. Commun., 1998, 1583. 13 W.-H. Leung and C.-M. Che, J. Am. Chem. Soc., 1989, 111, 8812; J. T. Groves and K.-H. Ahn, Inorg. Chem., 1987, 26, 3831. 14 C.-J. Liu, W.-Y. Yu, S.-M. Peng, T. C. W. Mak and C.-M. Che, J. Chem. Soc., Dalton Trans., 1998, 1805. 15 Y.-D. Wu, C.-L. Wong, K. W. K. Chan, G.-Z. Ji and X.-K. Jiang, J. Org. Chem., 1996, 61, 747 and refs. therein. 16 C. Linde, M. Arnold, P. O. Norrby and B. Åkermark, Angew. Chem., Int. Ed. Engl., 1997, 36, 1723. 17 Z. Gross and S. Ini, J. Org. Chem., 1997, 62, 5514. Paper 8/02587H

ISSN:1477-9226    

DOI:10.1039/a802587h

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3565–3576 3565 Kinetics and mechanisms of the complexation of aqueous lanthanide ions by 4-(2-pyridylazo)resorcinol † Yanlong Shi,a Edward M. Eyring *a and Rudi van Eldik *b a Department of Chemistry, University of Utah, Salt Lake City, Utah 84112, USA b Institute for Inorganic Chemistry, University of Erlangen-Nürnberg, 91058 Erlangen, Germany Received 24th July 1998, Accepted 14th August 1998 Three kinetic steps were observed for the complexations of Eu31 and UO2 21 by PAR [4-(2-pyridylazo)resorcinol] or PAN [1-(2-pyridylazo)-2-naphthol] in diVerent buVered solutions. The first step can be assigned to the co-ordination of the nitrogen donor from the pyridine moiety of the ligand based on the dependencies of [Eu31], pH, pressure, nature and concentration of the buVer.The rate-determining step is the release of water molecules from the co-ordination sphere of the lanthanide ion. Variations in the rate of the first step with diVerent lanthanide ions indicated that a co-ordination number changeover is involved in this lanthanide series.For the second step the formation of a “hydroazone–Ln31 chelate” intermediate accounts for all of the observed kinetic behaviors. The kinetic investigations of the third step show that there is a deprotonation preequilibrium preceding the transition state of the final product with two chelated 5-membered rings involved. Surprisingly, the rate constants of the three steps for the complexation of UO2 21 by PAR are very close to those for 18-crown-6 and diaza-18-crown-6 reacting with uranyl ion.The diVerences in the kinetics between PAR and PAN can be related to the diVerence in their structures. The fifteen trivalent lanthanide, or f-block, ions ranging from La3 to Lu31 represent the most extended series of chemically similar metal ions. The progressive filling of the 4f orbitals from La31 to Lu31 is accompanied by a smooth decrease in the cation radius rM with increasing atomic number because of the increasingly strong nuclear attraction for the electrons in the diVuse f orbitals (the lanthanide contraction).In an ideal situation, smooth variation of rate parameters with radii might be expected. However, the solution chemistry of the lanthanides displays more interesting variation than a simple linear correlation of rate and/or thermodynamic parameters with shrinking cation radius.1,2 A changeover in the co-ordination number of the lanthanide complexes from nine to eight near the middle of the series 3–6 gives rise to kinetics for the complexation or solvent exchange of lanthanide ions in solution that has been studied extensively using high-pressure NMR relaxation techniques. 7–13 Interest in these kinetic phenomena has increased with the development of some lanthanide complexes as contrast agents in magnetic imaging (MRI).14–16 Another signifi- cant feature of lanthanide element behavior in aqueous solution is the very high stability of the trivalent state although cerium(IV) and, in strongly reducing solutions, divalent samarium, europium and ytterbium can be formed.17,18 A third important characteristic is the strongly ionic character of lanthanide bonding. Thus, the lanthanides are typically “hard acids”.The kinetics of complexation is normally quite fast for Ln31 cations reacting with simple ligands compared to the rates of complexation for analogous complexes of the transition metal ions in the same oxidation state.The kinetics of complexation of Ln31(aq) by many monodentate or multidentate ligands has been studied using fast kinetic techniques. The ligands include NO3 2 (ultrasonic relaxation),19,20 SO4 22 (ultrasonic relaxation),21,22 acetate (ultrasonic relaxation),23 picolinic acid (pyridine-2-carboxylic acid) (pulse-radiolytic pH-jump),24 † Supplementary data available: rate constants as a function of buVer concentration for the Eu31–PAR reaction.For direct electronic access see http://www.rsc.org/suppdata/dt/1998/3565/, otherwise available from BLDSC (No. SUP 57431, 2 pp.) or the RSC Library. See Instructions for Authors, 1998, Issue 1 (http://www.rsc.org/dalton). murexide (E-jump),25 methyl red (pulse-radiolytic pH-jump),26 anthranilate (temperature-jump),26 malonate (ultrasonic relaxation), 27 oxalate (pressure-jump),28 arsenazo III [3,6-bis- (o-arsonophenylazo)-4,5-dihydroxynaphthalene-2,7-disulfonic acid] (stopped-flow) 13 and acyclic aminopolycarboxylates, such as EDTA (ethylenedinitrilotetraacetate) and DTPA [carboxymethyliminobis( ethylenenitrilo)tetraacetate].18,29 The rate of complexation is aVected either by the size of the central ion or by the nature of the ligand.However, the complexations of Ln31 by cyclic aminopolycarboxylates, such as DOTA (1,4,7,10-tetraazacyclododecane-N,N9,N0,N--tetraacetic acid), are slower and susceptible to study by traditional UV/VIS spectrophotometric techniques.18,30–39 The rate-determining step of the complexation of Ln31 by cyclic aminocarboxylates is proton loss from the ligand and the rearrangement of the intermediate.The rate of complexation is also aVected by the ring size of the ligands. The well known analytical reagents PAR [4-(2-pyridylazo) resorcinol] and PAN [1-(2-pyridylazo)-2-naphthol], like arsenazo III, have been studied extensively for the colorimetric determination of lanthanides and uranium(VI) 40 since they form stable, intensely colored complexes with a molar absorptivity of (3–8) × 104 M21 cm21. The compound PAR has been used widely in analytical chemistry because both it and the lanthanides and uranium(VI) complexes are water soluble, thus simplifying the analysis since no expensive, or toxic, organic solvents are required. Although the IR spectra,41,42 Raman spectra,43 acid–base equilibria 44,45 and HMO (Hückel molecular orbital) quantum calculations 45 of PAR and PAN, and some structural chemistry 46–49 and stability constants 50–52 of their lanthanide complexes have been investigated, neither bonding information N N N HO N N N HO OH PAR PAN3566 J.Chem. Soc., Dalton Trans., 1998, 3565–3576 between the lanthanide ion and the ligand nor kinetic studies have been reported for the complexation of lanthanides by PAR and PAN. However, some kinetics of the complexation of transition metal ions by PAR and PAN has been reported.53–66 DiVerences between transition metal ions and lanthanide or actinide ions make it interesting to study the complexation kinetics of lanthanide ions and uranium(VI) with PAR and PAN.We have sought a clear understanding of the nature of the diVerent contributions to the complexation kinetics and mechanism (e.g., ligand geometry, size of the central metal ion, pH, buVer, pressure, etc.). In the present paper, we mainly focus on the complexation of Eu31 by PAR and PAN, under diVerent buVer environments in the pH range of 1.8–8.1 using either conventional or high-pressure stopped-flow spectrophotometric techniques.In addition, we also studied the kinetics and mechanism of the complexation of other lanthanide(III) ions and UO2 21 by PAR for comparison. Experimental Materials The compounds PAR, PAN and Sudan Orange G [4-(phenylazo) resorcinol] were obtained from Aldrich and recrystallized from methanol, LaCl3?6H2O, CeCl3?7H2O, PrCl3?6H2O, NdCl3? 6H2O, SmCl3?6H2O and DyCl3?6H2O from Aldrich, EuCl3? 6H2O, GdCl3?6H2O, ErCl3?6H2O, YbCl3?6H2O and LuCl3? 6H2O from Strem and TbCl3?6H2O from Alfa.All these lanthanides were used as received (purity >99.9%); HoCl3?6H2O and YbCl3?6H2O were prepared from Ho2O3 (Strem) and Yb2O3 (Sigma) and UO2(ClO4)2?6H2O (Alfa) was used as received. N9- 2-Hydroxyethylpiperazine-N-3-propanesulfonic acid (HEPPS), MES [2-(morpholino)ethanesulfonic acid] and Tris [tris- (hydroxymethyl)aminomethane] were obtained from ICN Biochemicals.Imidazole (Eastman Kodak) was recrystallized from benzene. Succinic acid, sodium salt (A. R.) was from Aldrich. Acetate buVer solution was prepared by treating acetic acid (J. T. Baker) with sodium hydroxide. Succinate buVer solution was made by mixing succinic acid dipotassium salt (Eastman Kodak) with perchloric acid (Fisher Scientific). The pH of HEPPS, MES and Tris buVer solutions was adjusted with NaOH.Distilled water was purified using a Barnstead “EPure” purification system. 1,4-Dioxane (spectrophotometric grade) was from Aldrich. Stock solutions of the lanthanides and ligands were prepared by weight; NaClO4 (Aldrich) was used to maintain the ionic strength. All glassware was first treated with an EDTA solution and then cleaned with successive detergent, ammonia, and distilled water rinses. The pH was adjusted by adding HClO4 (Fisher Scientific, ACS reagent) or NaOH solutions (Aldrich).Instrumentation Spectrophotometric measurements were made with a Hewlett- Packard 8452A diode array spectrophotometer equipped with a thermostatted cell holder. pH-Metric measurements were made with an Orion Research 701 A Digital Ionanalyzer equipped with glass and calomel combined electrodes. Kinetic studies Kinetic measurements were made either at atmospheric pressure on a Durrum stopped-flow spectrophotometer or on a home-made, high pressure stopped-flow system 67 for pressures up to 1000 bar. n-Heptane was used as the pressurizing medium.An Edmund Scientific f/3.9 monochromator (1 nm per division) and a Hamamatsu photomultiplier tube (R376) were employed in all kinetic measurements. Transmitted light intensity versus time signals were recorded on a Tektronix (model 7D20) storage oscilloscope and transferred to a PC, on which data were fitted with the On Line Instrument System (OLIS) KINFIT (Bogart, GA) programs.Several experimental traces were averaged in the determination of each rate constant. The complexation of Ln31 or UO2 21 by PAR or PAN was studied at 25 8C. Constant temperature was maintained with a Forma Scientific model 2006 constant temperature bath and circulator system for the ambient Durrum D-110 Stopped- Flow Spectrophotometer, and a Brinkmann Instrument Lauda K-2/RD constant temperature apparatus for the high-pressure stopped-flow spectrophotometer at 25.0 8C.Temperature control precision was ±0.1 8C. All kinetic data were measured after not less than 1 h of temperature equilibriation. Experimental rate constants reported in the Results section are the average of at least 5 replicate determinations. The optimum observation wavelength of 502 nm was determined from preliminary observations on a HP 8452A spectrophotometer. Calculations All experimental runs for the three consecutive kinetic steps were best described by a single exponential.Observed pseudo- first-order rate constants were obtained from a least-squares fit of at least 3 half-lives of the reactions. Volumes of activation were obtained by a fit of the natural logarithm of the observed pseudo-first-order rate constants using eqn. (1). Here k0 denotes ln k = ln k0 2 (DV‡P/RT) (1) the rate constant at ambient pressure. Errors reported in the Tables correspond to one standard deviation. Results and discussion Structure, acid–base equilibria and tautomeric equilibria of PAR The visible spectra of 4-(2-pyridylazo)resorcinol were studied by Geary et al.44 as a function of pH from 1.0 to 13.0 and the chromophoric species were identified as follows: protonated form (A), lmax = 420 nm, e = 14 750 dm3 mol21 cm21, pH 1.06 and 1.52; free base form (B), lmax = 392 nm, e = 15 240 dm3 mol21 cm21, pH 3.19, 4.35 and 5.56; monoionic form (C), lmax = 414 nm, e = 23 100 dm3 mol21 cm21, pH 7.56–13.56; diionic form (D), lmax = 502 nm, e = 17 800 dm3 mol21 cm21, pH 12.96 and 13.56.The relationship among all the species is summarized in Scheme 1. Zhao et al.45 suggested that there are five forms for PAR in aqueous alcohol solution. The equilibria are: H4L21 2H1 Ka1 H3L1 2H1 Ka2 H2L 2H1 Ka3 HL2 2H1 Ka4 L22. The pKa values are pKa1 = 22.30, pKa2 = 3.00, pKa3 = 5.58 and pKa4 = 12.03, respectively. They found that the H3L1 is present in a hydrazone form (see Scheme 1); H2L exists in both the hydrazone and azo forms.The results 45 are in agreement with those obtained by Drozdzewski 43 from resonance Raman spectra. The anions HL2 and L22 exist in the azo forms, and are consistent with the result of Geary et al. shown in Scheme 1. Bonding and stability constants of the LnIII–PAR complexes When a 1 × 1024 mol dm23 PAR solution is mixed with a 1 × 1023 mol dm23 Eu31 solution at pH 4.35 in 0.1 mol dm23 acetate buVer using a tandem cuvette the lmax immediately shifts from 390 to 502 nm.This indicates that complexation of Eu31 (aq) by PAR does take place and raises the question what the structure of the formed complex could be. N N O O Sudan Orange G H HJ. Chem. Soc., Dalton Trans., 1998, 3565–3576 3567 Scheme 1 lmax and pKa values, measured in water solution, cited from ref. 44; lmax* and pKa* values, measured in water–alcohol solution, cited from ref. 45. + N N N OH2 + O H H [H2PAR]2+ N N N OH O H+ H H+ PAR N N N O HO PAR H lmax = 470, 401 nm* lmax = 494 nm* + N N N OH O H H N N N OH O H N N N O– –O N N N O– O H lmax = 420 nm lmax = 392 nm, lmax* = 393 nm lmax = 502 nm lmax* = 493, 430 nm lmax = 414 nm, lmax* = 413 nm H+PAR PAR PAR2– PAR– –H+, pKal* = –2.30 –H+ pKal = 2.35 pKa2* = –3.00 pKa2 = 7.01 pKa3* = 5.58 –H+ pKa3 = 13.0 pKa4* = 12.0 –H+ In order to study the bonding between PAR and Eu31, we used Sudan Orange G to react with Eu31 as a function of pH.We found that the metal complex absorbed at almost the same wavelength as the ligand itself at diVerent pH values below 7.For example, when [Sudan Orange G] = 1 × 1024 mol dm23 at pH 5.4 in 0.05 mol dm23 acetate buVer lmax is 374 nm for free Sudan Orange G, and lmax is still at 374 nm after mixing with aqueous 1 × 1023 mol dm23 Eu31 solution (1 : 1 v/v). The role of the pyridine nitrogen atom of PAR in the colorimetric reaction with Eu31 ion is evident from these results. The fact that there is no shift in lmax on chelation by the benzene analogue of PAR, Sudan Orange G, must mean that the nitrogen atom from the pyridine moiety of PAR is involved in the bonding to Eu31.To test which one of the nitrogen atoms of the azo group participates in the bonding, Geary et al.44 found that when the copper(II) ion reacts with 2-(o-hydroxyphenyliminomethyl)- pyridine, the complex absorbs at a wavelength of 395 nm at pH values above 8.50. This represents a shift of 15 nm away from the maximum wavelength of the ligand at this pH.This shift is considerably less than the shift for the copper or europium complexes of PAR, yet the co-ordinating system is the same as in the PAR system except that the azo nitrogen nearest the heterocycle is replaced by a CH]] group. The removal of this nitrogen has a profound eVect on the color reaction with metal ions, and it seems clear that in PAR the azo nitrogen farthest from the heterocycle must play a greater role in the chromophoric reaction than its neighbors.This conclusion is further supported by the visible spectra of the metal complexes of 2-(salicyclideneamino)pyridine.44 This ligand gave strongly absorbing red complexes of transition N C H N HO 2-(o-hydroxyphenyliminomethyl)pyridine metal ions similar to those with PAR. For example, the main peak of the ligand at 350 nm at pH 4.88 is shifted to 453 nm at this pH in the presence of copper(II). These results demonstrate that in the Eu31–PAR complex the chromophoric reaction is due to co-ordination by the pyridine nitrogen, the azo nitrogen farthest from the heterocycle, and the o-hydroxyl group, even though there is an “intramolecular hydrogen bonding” as shown in free PAR. The partial coordination of Eu31(aq) by PAR can decrease the pKa values for the deprotonation of both the o- and the p-hydroxyl groups, and therefore the deprotonation takes place at much lower pH.The sensitivity of the color reaction of this ligand with metal ions is therefore explained by the combination of a pseudophenanthroline system and o,o9-disubstituted azo dyestuV.The stability constants of Ln31–PAR have been measured by Ohyoshi 51,52 using UV/VIS spectrophotometry. Both Ln31– PAR2 and Ln31–PAR22 complexes form in the pH range 5–6. The stability constants, log K (Ln31–PAR2) and log K (Ln31– PAR22), range from 3.78 ± 0.02 (Ce) to 4.39 ± 0.02 (Lu), and from 9.61 ± 0.06 (Ce) to 10.70 ± 0.05 (Lu), respectively. The acidity of the Ln31–PAR2 complexes parallels the order of stability of the Ln31–PAR22 complexes.In an attempt to elucidate the co-ordination structure of the N N C H HO 2-salicylideneaminopyridine N N N O OH PAR with "intramolecular hydrogen bonding" H3568 J. Chem. Soc., Dalton Trans., 1998, 3565–3576 PAR complexes, extraction studies of the 1 : 2 Ni–PAR complexes were carried out by Hoshino et al.68 They suggested that in the chelation of Ni(PAR2)2?2H2O and [Ni(PAR22)2]22 the PAR2 and PAR22 are acting as bidentate and terdentate ligands, respectively.They suggested that the basic change in the chelate structure from Ni(PAR2)2?2H2O to [Ni(PAR22)2]22 and deprotonation of the PAR2 ligand both cause a substantial increase in the absorptivity (from 3.73 × 104 to 8.08 × 104 dm3 mol21 cm21). Although the 1 : 1 lanthanide–PAR complexes differ in type from the Ni–PAR complexes, a considerable increase in the absorptivity [from (0.95–1.15) × 104 to 3.0 × 104 dm3 mol21 cm21] was similarly observed with increasing pH.The Ln31–PAR22 complexes may have a more stable chelate structure which gives rise to a larger diVerence in stability than for the Ln31–PAR2 complexes. Based on all the information given above it appears that the structures of the 1 : 1 Eu31–PAR2 and Eu31–PAR22 complexes are those shown below. Overview of the observed kinetics All kinetic runs were made with at least a 10-fold excess of Eu31. The reaction of aqueous Eu31 with PAR occurs in two steps in succinate buVered solution when pH < 2.65. Three steps were observed at 502 nm on diVerent timescales when pH > 2.65. The first is much faster than the second and third with a half-life of 2 ms to 100 s, depending upon pH, Eu31 concentration, amine buVer concentration, the nature of the amine buVer, pressure, and the nature of the lanthanide ions and UO2 21.A typical kinetic trace is shown in Fig. 1(a). The first step is assigned to the complexation of Eu31(aq) by the nitrogen atom from pyridine since the kinetics traces could be obtained when the pH was as low as 1.80 at a rate which depended on the pH.Fig. 1(b) is a typical kinetic trace for the second step showing that the absorbance at 502 nm decreases with increasing time with a half-life of ca. 500 ms. The second step is independent of pH, Eu31 concentration, amine buVer concentration, the nature of the amine buVer and the choice of the lanthanide ions, but the rate increases with increasing pressure. We assign the second step to the formation of the “hydrazone–Eu31 chelate” intermediate of a “phenanthroline style”.In the third slowest step the absorbance at 502 nm increases with increasing time. A typical kinetic trace for this step at pH 7.55 in HEPPS buVer solution is shown in Fig. 1(c). The third step depends on pH, amine buVer concentration, the nature of the amine buVer and pressure, but is independent of the concentration of Eu31.This step is attributed to the formation of the final 1 : 1 complex, Eu31–PAR22, shown above. Kinetic studies of the first step Eu31 Concentration dependence. Table 1 presents the rate constants k1 for the first step of the reaction of Eu31 with PAR as a function of Eu31 concentration either in a pH 2.08 succinate buVer or a pH 4.35 acetate buVer. The plots (Fig. 2) of k1 versus [Eu31] in both succinate and acetate buVers are linear with no significant intercept. The rate constants calculated from the slopes are 6.15 × 101 and 4.12 × 104 dm3 mol21 s21 in succinate and acetate buVers, respectively.The large diVerence is attributed not only to the diVerence in pH but also to the nature of the buVers used (see EVect of pH). This kinetic behavior suggests that the first step follows the Eigen–Wilkins mechanism depicted in Scheme 2 69 where A represents H2O, acetate or other buVer molecules which are involved in the co-ordination N N N HO O– 1:1 Eu3+-PAR–complex Eu3+ N N N –O O– Eu3+ 1:1 Eu3+-PAR2–complex to Eu31.From Scheme 2, d[product]/dt = k1[b] but d[b]/dt = 2k1[b] 2 k210[b] 1 k10[a] = 0, [b] = k10[a]/(k1 1 k210) and d- [product]/dt = k1k10[a]/(k1 1 k210) = {K19k1k10[Eu(H2O)8(A)x1]- [PAR]}/(k1 1 k210), but [PAR] = Ka1[H1PAR]/[H1]. Thus, d[product]/dt = {Ka1K19k1k10[Eu(H2O)8(A)x1][H1PAR]}/{(k1 1 k210)[H1]} where K19 = k19/k219. This rate law is consistent with the dependence on Eu31 concentration shown in Table 1 and Fig. 2, and also agrees with the observed eVect of pH discussed below. EVect of pH. Measurements made in succinate and acetate buVer solutions covered the pH range 1.80–5.40. All runs were made at a constant ionic strength of [NaClO4] = 0.1 mol dm23 and a constant concentration of 0.05 mol dm23 of the basic component of the buVer. Table 2 presents the observed rate Fig. 1 Typical kinetic traces recorded for the first (a), second (b) and third step (c) of the reaction of Eu31(aq) with PAR at ambient pressure and 25.0 8C in 0.01 mol dm23 HEPPS buVer solution (pH 7.55, [Eu31] = 1 × 1023 mol dm23, [PAR] = 1 × 1024 mol dm23 and [NaClO4] = 0.1 mol dm23).J.Chem. Soc., Dalton Trans., 1998, 3565–3576 3569 constants as a function of pH in either succinate buVer or acetate buVer at ambient pressure and 25.0 8C. Fig. 3 is a linear plot of k1 versus 1/[H1] (k1 increases with increasing 1/[H1]) that proceeds through the origin in the succinate buVer in the pH range 1.80–3.31.The rate constants of the first step in acetate buVered solution in the pH range 3.61–5.40 (see Table 2) are much faster, and are independent of the pH or 1/[H1]. Although the acetate anion forms only weak complexes with lanthanide cations, for Fig. 2 Plots of k1 versus Eu31 concentration for the first step of the reaction of Eu31 with PAR at ambient pressure and 25.0 8C ([Eu31] = 1 × 1023 mol dm23, [PAR] = 1 × 1024 mol dm23 and [NaClO4] = 0.1 mol dm23): n, pH 4.35, [acetate] = 0.05 mol dm23; h, pH 2.08, [succinate] = 0.05 mol dm23.Scheme 2 [Eu(H2O)8(A)]x+ + PAR(aq) [Eu(H2O)8(A)]x+ • [PAR(aq)] [Eu(H2O)8(A)x+----PAR(aq)] [Eu(H2O)7(A)x+----PAR(aq)] + H2O product H+PAR(aq) k1¢ k–1¢ k1¢¢ k–1¢¢ k1 Ka1 –H+ a b Table 1 Rate constants as a function of concentration of Eu31 for the three steps of the reaction of Eu31 (aq) with PAR at 25.0 8C and ambient pressure ([PAR] = 1 × 1024 mol dm23 and [NaClO4] = 0.1 mol dm23) [Eu31]/mol dm23 10k1/s21 10k2/s21 10k3/s21 pH 2.08 (succinate 0.05 mol dm23) 0.001 0.002 0.003 0.004 0.005 0.50 ± 0.02 1.42 ± 0.01 2.12 ± 0.13 2.53 ± 0.19 3.02 ± 0.14 3.90 ± 0.15 3.76 ± 0.09 3.68 ± 0.15 3.84 ± 0.10 3.59 ± 0.08 3.79 ± 0.09 3.58 ± 0.11 3.81 ± 0.12 3.72 ± 0.05 3.65 ± 0.11 pH 4.35 (acetate 0.05 mol dm23) 0.001 0.002 0.003 0.005 0.010 50 ± 3 89 ± 7 127 ± 6 204 ± 8 420 ± 12 10.1 ± 0.4 9.82 ± 0.2 9.53 ± 0.4 9.95 ± 0.3 9.30 ± 0.2 18.5 ± 0.4 19.0 ± 0.6 17.8 ± 0.6 18.9 ± 0.2 17.6 ± 0.5 example, log b1 ª 1.9 for EuAc21,16 the concentration of acetate is very large compared with the other species present.Therefore, the Ac2 involvement in co-ordination must be considered. Calculations from the thermodynamic equilibrium data indicate that the concentration of EuAc21 is about 40% of that of Eu31(aq) in solutions of pH 4.5 when [HAc] 1 [Ac2] @ [Eu31].70 Some kinetic studies 71–73 indicate that the catalytic eVect results from a trans labilization eVect by co-ordinated acetate and involves the attack by a solvent water molecule on the metal–carbon bond for the heterolysis reaction of (a-hydroxyalkyl)chromium(III). The co-ordinated acetate accelerates dissociation of water molecules from the inner coordination sphere of the metal ions, resulting in kinetic diVerences with reactions in unbuVered or other buVered solutions. 13,74 No protonated pyridine nitrogen exists in the acetate buVered pH range (3.61–5.40) since the pKa1 is as low as 2.35,44 thus [PAR]tot = [PAR], and k1 is independent of pH in acetate buVer.Pressure dependence. The pressure dependence for the first Fig. 3 Plot of k1 versus [H1]21 for the first step of the reaction of Eu31 with PAR at ambient pressure and 25.0 8C in 0.05 mol dm23 succinate buVered solution ([Eu31] = 1 × 1023 mol dm23, [PAR] = 1 × 1024 mol dm23 and [NaClO4] = 0.1 mol dm23). Table 2 Rate constants as a function of pH for the first step of the reaction of Eu31(aq) with PAR* pH 10k1/s21 [succinate] = 0.05 mol dm23 1.80 2.08 2.30 2.65 2.89 3.31 0.07 ± 0.06 0.83 ± 0.04 1.90 ± 0.16 5.00 ± 0.04 8.96 ± 0.13 21.2 ± 0.5 [acetate] = 0.05 mol dm23 3.62 4.14 4.35 4.58 5.02 5.40 48.9 ± 1.3 50.2 ± 1.3 50.2 ± 1.8 48.5 ± 1.7 46.9 ± 1.1 49.6 ± 2.3 * Experimental conditions: [Eu31] = 1 × 1023 mol dm23; [PAR] = 1 × 1024 mol dm23; 25.0 8C; [NaClO4] = 0.1 mol dm23.3570 J.Chem. Soc., Dalton Trans., 1998, 3565–3576 Table 3 Rate constants as a function of pressure for the three steps of the reaction of Eu31(aq) with PAR in diVerent buVer solutions at 25.0 8C and [Eu31] = 1 × 1023 mol dm23 and [NaClO4] = 0.1 mol dm23 10k1/s21 DV‡ 1/cm3 mol21 P/bar 1 50 250 500 750 1000 pH 2.65 [succinate] = 0.05 [PAR] = 1 × 1024 0.50 ± 0.04 0.49 ± 0.03 0.44 ± 0.04 0.40 ± 0.04 0.36 ± 0.02 0.33 ± 0.02 pH 4.35 [acetate] = 0.05 [PAR] = 5 × 1025 — 25.7 ± 1.4 23.0 ± 2.3 20.7 ± 0.4 12.6 ± 0.4 10.1 ± 0.3 pH 4.35 [acetate] = 0.05 [PAR] = 1 × 1024 50.2 ± 1.8 47.3 ± 2.6 40.0 ± 1.9 32.7 ± 2.2 25.0 ± 0.6 19.8 ± 1.4 pH 2.65 [succinate] = 0.05 [PAR] = 1 × 1024 110.1 ± 0.7 pH 4.35 [acetate] = 0.05 [PAR] = 5 × 1025 125.4 ± 3.8 pH 4.35 [acetate] = 0.05 [PAR] = 1 × 1024 122.8 ± 0.56 k2/s21 102k3/s21 DV‡ 2/cm3 mol21 DV‡ 3/cm3 mol21 P/bar 1 50 250 500 750 1000 pH 4.35 [acetate] = 0.05 [PAR] = 1 × 1024 1.01 ± 0.08 1.02 ± 0.11 1.23 ± 0.07 1.39 ± 0.12 1.67 ± 0.05 1.91 ± 0.08 pH 4.35 [acetate] = 0.05 [PAR] = 1 × 1024 1.85 ± 0.12 1.92 ± 0.19 2.25 ± 0.12 2.40 ± 0.14 2.70 ± 0.20 2.96 ± 0.24 pH 7.10 [imidazole] = 0.05 [PAR] = 1 × 1024 — 29.4 ± 2.5 35.4 ± 2.0 40.3 ± 3.4 48.1 ± 3.5 55.3 ± 4.5 pH 4.35 [acetate] = 0.05 [PAR] = 1 × 1024 215.9 ± 0.6 pH 4.35 [acetate] = 0.05 [PAR] = 1 × 1024 211.2 ± 0.9 pH 7.10 [imidazole] = 0.05 [PAR] = 1 × 1024 215.9 ± 0.8 step of the reaction of Eu31 with PAR was studied at diVerent pH values, PAR concentrations and in diVerent buVers at 25.0 8C.The k1 values as a function of pressure under diVerent reaction conditions are summarized in Table 3.Fig. 4 clearly shows a linear relationship between ln k1 and pressure, from which it follows (see Table 3) that all DV‡ 1 values are positive. The DV‡ 1 values in acetate buVered solution (>120 cm3 mol21) are much larger than those found in succinate buVered solution (110.1 cm3 mol21). In general the pKa values of buVers depend on pressure, typically for HAc H1 1 Ac2, DV ª 212 cm3 mol21. It means that the buVer becomes more acidic (increase in Ka) with increasing pressure. In our system this will not aVect the data in the acetate buVer since we found no pH dependence in this range.However, at lower pH in succinate buVer, a part of the observed DV‡ 1 could be due to the Fig. 4 Plots of ln k1 versus pressure for the first step of the reaction of Eu31 with PAR at 25.0 8C under diVerent reaction conditions ([Eu31] = 1 × 1023 mol dm23 and [NaClO4] = 0.1 mol dm23): h, pH 4.35, [PAR] = 1 × 1024 mol dm23, [acetate] = 0.05 mol dm23; s, pH 4.35, [PAR] = 5 × 1025 mol dm23, [acetate] = 0.05 mol dm23; ,, pH 2.66, [PAR] = 1 × 1024 mol dm23, [succinate] = 0.05 mol dm23.pressure dependence of the buVer, i.e. a lowering in pH due to an increase in Ka, which will slow down the reaction and show up as a positive DV‡ value. These DV‡ 1 values demonstrate that when other reaction conditions are held constant the ratedetermining step is the release of water molecules from the first co-ordination sphere. In succinate buVered solution an Id mechanism apparently prevails, whereas in acetate buVered solution a D mechanism controls the first step because of the “acetate eVect” mentioned above.Influence of buVer. Since weak organic acids are frequently used as buVers, their ability to complex lanthanide cations should not be ignored. The stability constant, log bn, of the lanthanide cation with the buVering anion increases with the pKa of the buVer acid. Acetate buVer is a typical example.In neutral solutions buVers are often amine compounds. Complexation by these buVers may be of less concern. Thus our measurements of buVer concentration dependence were made in MES, HEPPS and Tris buVers over the 6.15–8.1 pH range. In this pH range the pyridinium ion of the ligand is completely dissociated (even in acetate buVer as discussed above). The rate of the first step is pH-independent and is subject to specificand general-base catalysis so that k1 = k0 1 kOH[OH2] 1 kb[B].Fig. 5 shows typical plots of k1 as a function of buVer base concentration. The linear plot for the MES buVers at pH 6.15 is evidence of general-base catalysis by this buVer when [MES] < 0.1 mol dm23. The non-linear plot of k1 vs. [B2] at pH 7.55 with HEPPS buVer suggests that specific interaction with this buVer obscures any base catalysis. The same phenomenon was observed by Reeves 55 for the complex formation of NiIIsulfonated 1-(2-pyridylazo)-2-naphthol (b-PAN) in HEPES buVer [N9-(2-hydroxyethyl)-N-piperazineethanesulfonic acid].Reeves noted that absorption of b-PAN in a pH 6.89 HEPES buVer ([HEPES] = 0.01 mol dm23) in the absence of NiII has a significantly higher absorptivity than the curve for a MES buVer of the same pH and [B2]. The interaction is apparently between the buVer and the dye and not with the metal ion. Similar evidence for a specific dye–buVer interaction between piperazine buVer and a cyano keto azo dye ligand has been observed.60 In the linear plot (Fig. 5) for the Tris buVer k1 decreases with increasing Tris buVer concentration. This could be due to the multiple complexation equilibria between the buVer and Eu31.J. Chem. Soc., Dalton Trans., 1998, 3565–3576 3571 This kind of complexation between NiII and Tris has been reported.75 The log b1 for Eu(Tris)31 is ca. 2.5. For [Tris] = 0.005 mol dm23, [EU(Tris)31] would be 60% larger than [Eu31]. In addition, the formation of the hydrolysed europium species Eu(OH)n 32n, at such high pH causes a decrease in reactive europium concentration that cannot be neglected.DiVerent Ln31. The second order rate constants for the first step of the reaction of Ln31 with PAR, k1/dm3 mol21 s21, in logarithmic form, as a function of diVerent lanthanide ions, and versus reciprocal metal ionic radius in 0.05 mol dm23 MES buVer at pH 6.15, ambient pressure and 25.0 8C ([Ln31] = 1 × 1023 mol dm23, [PAR] = 1 × 1024 mol dm23 and [NaClO4] = 0.1 mol dm23), are given in Fig. 6. For comparison, the complex-formation rate constants of aqueous Ln31 ions with some other ligands are also given in Fig. 6. The rates of complexation not only depend on the metal ion but also on the nature of the incoming ligand. The rate constants k1 are seen to reach a maximum around samarium for our Ln31–PAR system. The same trends were observed for the complexations of lanthanide ions by sulfate,22 acetate 23 and CyDTA.30 Other workers 12,13,76,77 have concluded that there is probably a change in co-ordination number (CN) occurring from Sm31 to Gd31.This may explain the observed change along the lanthanide series in the rate of almost all the complexations included in Fig. 6, as well as by sulfate and acetate. These results suggest a very easy substitution pathway for these ions, due to the almost identical energies of the octa- and nona-co-ordinated species. Moreover, the observation of an associative water-exchange 12 or complex-formation 13 mechanism on the octahydrated heavy Ln31 ions (the smallest of the series according to their ionic radii) leads to the presumption of a larger co-ordination number for the lighter elements, thus reinforcing the idea of a co-ordination number change in the middle of the series.The co-ordination numbers of the Ln31 ions in water have been the subject of substantial debate,78–80 but it is now established from neutron scattering,76,81 X-ray scattering,82 extended X-ray absorption fine structure (EXAFS),5 density 83 and spectrophotometric 6 studies that the lighter La31–Nd31 ions are predominantly nine-co-ordinate, Pm31–Gd31 exist in equilibria between eight- and nine-co-ordinate states, and the heavier Tb31–Lu31 are predominantly eight-co-ordinate.5,6,76,81–83 The systematic decrease in k1 for L = PAR, oxalate and murexide shown in Fig. 6, and the decreasing water exchange Fig. 5 Plot of the rate constant as a function of the concentration of buVer base in MES (s, pH 6.15), HEPPS (h, pH 7.55), and Tris (,, pH 8.10) for the first step of the reaction of aqueous Eu31 with PAR. rate,12,84 as the ionic radius decreases from Gd31 to Yb31, is consistent with increasing steric crowding, hindering the entry of the incoming ligand and dominating the variation of k1.The corresponding increase in surface charge density might be expected to provide an increased electrostatic attraction between the entering ligand and Ln31, and thereby accelerate the co-ordination rate, but this is evidently not important here.The k1 values shown in Fig. 6, and the other rate constants of aqueous Ln31 complexations by H2O,5,12 NO3 2,19,20 SO4 22,21,22 acetate,23 picolinic acid,24 methyl red,26 malonate,27 arsenazo III,13 and acyclic aminopolycarboxylates, such as EDTA and DTPA,18,29 vary by several orders of magnitude for the same metal ion. The ligands do not have the same steric properties and electronic charge and will have very diVerent outer-sphere parameters.However, the diVerence in the k1 values is probably due to various other factors as well, such as diVerences in experimental reaction conditions and measurement methods. It should also be recalled that the ligands in Fig. 6 are all weak bases, and competition may occur between protonation and metal bond formation on the ligand basic sites. This is exempli- fied by monoprotonated CyDTA complexation on the lanthanide ions where the complexation rate constants (ª3 dm3 mol21 s21 at 25.0 8C) are governed by the final deprotonation step of the ligand.In many studies 19–29 only one step was observed for the complexations of Ln31 by various multidentate ligands on a short timescale because of limitations of the kinetic techniques, such as ultrasonic relaxation and NMR. Subsequent slower reaction steps were not observed. An advantage of the stopped-flow technique is the relatively slower accessible timescales permitting a more complete kinetic picture of the complexation of aqueous Ln31 by multidentate ligands.Based on the information given above, Scheme 3 provides a reasonable description of the first step of the complexation of aqueous Eu31 by PAR in diVerent buVered solutions. Comparison between PAR and PAN. Owing to the low solubility of PAN in aqueous solution, a water–1,4-dioxane (3 : 1 v/v) mixed solvent was used to compare the kinetics of complexation of Eu31(aq) by PAN and PAR.The rate constant, k1 (see Fig. 6 Comparison of the complex-formation rate constants k1 (second order rate constant) for some reactions of Ln31 ions with different ligands in water, versus reciprocal metal ionic radius: e, L = anthranilate, see ref. 27; n, L = murexide, see refs. 25 and 26; h, L = oxalate, see ref. 29; ,, L = CyDTA (cyclohexane-1,2-diyldinitrilotetraacetate), see ref. 31; s, L = PAR, this work.3572 J.Chem. Soc., Dalton Trans., 1998, 3565–3576 Table 4), for the first step of the complexation of aqueous Eu31 by PAR is 12 times greater than for PAN under the same reaction conditions. Scheme 1 shows that PAR exists as the monoanion PAR2 at pH 6.15 in MES buVer; PAN is a neutral molecule at the same pH. The negative charge on PAR gives rise to a higher electron density on the nitrogen atom of the pyridine moiety through conjugation. Consequently, the first complexation step takes place at a faster rate for PAR than for PAN.In addition, the k1 value (71.5 ± 0.9 s21) for PAR obtained in the water–1,4-dioxane mixed solvent is smaller than in pure water (126 ± 2 s21) under the same reaction conditions. Cusumano61 also observed that rate constants for the complexation of nickel(II) by PAR and PAN in diVerent non-aqueous solvents depend strongly on the nature of the solvent. Comparison between Eu31 and UO2 21. Although the number of actinide elements is the same as the number of lanthanide elements, the availability of the former and their chemical characteristics have so far largely restricted the study of their ligand substitution mechanisms to dioxouranium(VI), the ionic form of uranium most amenable to such studies in solution.Commonly observed solvated species have the stoichiometry [UO2- (solvent)5]21, for example, [UO2(H2O)5]21, characterized by two oxo ligands bound in axial sites with average axial U–O distances in the range 1.71–1.75 Å.Five water molecules occupying the equatorial plane have an average equatorial distance of 2.45 Å (see structure below). In solution the two oxo atoms undergo slow exchange, whereas the equatorial solvent molecules experience fast exchange. Thus the [UO2(solvent)5]21 system oVers the opportunity to study solvent exchange in a single plane of a solvated metal ion.2,85–87 Kinetic data of the complexations of UO2(aq)21 and Eu31(aq) by PAR in 0.1 mol dm23 acetate buVer at pH 4.35 are summarized in Table 5.The k1 value for the complexation of Scheme 3 + N N N OH O H H N N N OH O H H+ PAR PAR N N N OH O H [Eu(H2O)8(A)]x+ [Eu(H2O)7(A)]x+ + H2O k1 k–1 + –H+ pKa1 [H2O)7(A)Eu]x+ k A = H2O, x = 3 A = acetate, x = 2 O U O H2O H2O OH2 OH2 OH2 2+ UO2(H2O)5 2+ Table 4 Rate constants for the three steps of the complexation of Eu31(aq) by PAR or PAN in MES buVered, water–1,4-dioxane (3 : 1 v/v) at pH = 6.15, ambient pressure and 25.0 8C ([Eu3] = 1 × 1023 mol dm23, [L] = 1 × 1024 mol dm23 (L = PAR or PAN), [MES] = 0.05 mol dm23 and [NaClO4] = 0.1 mol dm23) L PAR PAN lmax/nm 502 532 k1/s21 71.6 ± 0.90 5.89 ± 0.30 10k2/s21 1.85 ± 0.12 4.58 ± 0.12 102k3/s21 7.01 ± 0.73 1.01 ± 0.10 UO2 21 by PAR is about the same as that of the Eu31–PAR system.This suggests that the first step of the UO2 21 complexation by PAR follows the same mechanism proposed for the lanthanides (see Schemes 2 and 3), even though there are two oxo ligands bound in the axial sites on UO2 21.Fux et al.87 found that the rate constants of the observed first step of the complexation of UO2 21 by 18-crown-6 and diazo-18-crown-6 in propylene carbonate are 930 ± 50 and 23 ± 1 s21, respectively. The mechanism is very similar to our mechanism proposed in Scheme 2. Comparing our rate constant k1 with those found by Fux et al.87 for reactions with UO2 21, we found that: k1(18- crown-6, in propylene carbonate, 930 s21) > k1(PAR, in water, 46.7 s21) > k1(diazo-18-crown-6, in propylene carbonate, 23 s21). Our k1 value is much closer to that for the complexation of UO2 21 by the nitrogen atom in diazo-18-crown-6, which means that the complexation of UO2 21 by an oxygen donor is faster and stronger than that by a nitrogen donor.This is attributed to the “hard acid” character of UO2 21. Kinetic studies of the second step In a typical kinetic trace shown in Fig. 1(b) for the observed second step the absorbance at 502 nm decreases with increasing time with a half-life of ca. 500 ms. Mochizuki et al.88 observed a similar phenomenon in the complexation of Co21 and Ni21 by PAN in aqueous 1,4-dioxane. They proposed that the absorbance decrease is caused by the formation of the insoluble neutral intermediate [CoII(PAN2)2]0, and the slowest step, namely, the absorbance increase with increasing time is due to the formation of the soluble final product [CoIII- (PAN2)2]1. If the same process holds true for our Ln31–PAR (or PAN) system, the intermediates should be [LnII(PAR22)]0 or [LnII(PAR2)2]0, and [LnII(PAN2)2]0, respectively.However, if it is kept in mind that for our kinetic studies aqueous Ln31 ions were always in excess concentration and only samarium, europium and ytterbium have 12 oxidation states,70 it is obvious that the suggestion by Mochizuki et al.88 cannot apply to our Ln31–PAR (or PAN) system. We propose that the formation of the “hydrazone–Ln31 chelate” intermediate (see the structure below) is more reasonable because it destroys the whole conjugated structure which is the basis of many azo dyes.It therefore becomes interesting to explore whether the formation of a “hydrazone–Mn1 chelate” intermediate is a common phenomenon during multi-step complexations of metal cations by many azo dyes. Eu31 Concentration dependence. The rate constants for the second step of the reaction of Eu31 with PAR, k2, as a function of Eu31 concentration, either in a pH 2.08 succinate buVer or a pH 4.35 acetate buVer, are summarized in Table 1.One sees that k2 is independent of Eu31 concentration either in succinate N N N O OH(O–) H [(H2O)5(A)Eu]x+ Hydrogen bonding H2O "Hydrazone-Eu3+ chelate" Table 5 Rate constants for the three steps of the complexations of UO2 21(aq) and Eu31(aq) by PAR in acetate buVer at pH 4.35, ambient pressure and 25.0 8C ([M] = 1 × 1023 mol dm23, M = UO2 21 or Eu31; [PAR] = 1 × 1024 mol dm23, [acetate] = 0.1 mol dm23 and [NaClO4] = 0.1 mol dm23) M UO2 21 Eu31 1021k1/s21 4.67 ± 0.10 5.02 ± 0.18 k2/s21 0.94 ± 0.40 1.00 ± 0.03 102k3/s21 0.74 ± 0.50 1.85 ± 0.11J.Chem. Soc., Dalton Trans., 1998, 3565–3576 3573 or acetate buVers which suggests that “intramolecular ring closure” must be rate-determining. pH Dependence. Measurements made in acetate buVer solutions covered the pH range 3.61–5.40. All runs were made at a constant ionic strength of [NaClO4] = 0.1 mol dm23 and a constant (0.05 mol dm23) concentration of the basic component of the buVer.The observed rate constants as a function of pH in acetate buVered solution at ambient pressure and 25.0 8C are shown in Table 6. The k2 values are independent of pH. This result is consistent with the formation of a “hydrazone–Eu31 chelate” intermediate proposed above which does not involve any deprotonation or acid–base equilibrium. Pressure dependence. The pressure dependence for the second step of the reaction of Eu31 with PAR was studied in 0.05 mol dm23 acetate buVer at 25.0 8C (see Table 3).The plots of ln k2 versus pressure yield a DV‡ 2 value 215.9 ± 0.6 cm3 mol21. The negative sign suggests that the rate-determining process of the second step has an associative character. This result is also consistent with the intermediate formation of the “hydrazone–Eu31 chelate” proposed above, which will involve a ring compact transition state.Influence of buVer. The k2 values as a function of buVer concentration in MES (pH 6.15), HEPPS (pH 7.55) and Tris (pH 8.10) buVers at ambient pressure and 25.0 8C over the concentration range of 0.01–0.10 mol dm23 are given in SUP 57431. In all three buVers the rate constants are independent of buVer concentration. Absence of a buVer eVect in this step suggests that the mechanism does not involve complex formation between the buVer and Eu31 species. The k2 values obtained from MES and HEPPS are very similar.On the other hand, the values obtained with Tris buVer (higher pH) are smaller. The diVerence is probably caused by a decrease in concentration of the reactive Eu31 species at such a high pH (8.10) in the Tris buVer due to hydrolysis to form Eu(OH)n 3 2 n as mentioned before. Comparison between PAR and PAN. Measurements were made under the same reaction conditions as for the first step except for the timescale. Table 4 indicates that the k2 value (0.458 s21) for the complexation of Eu31 by PAN is larger than that for PAR (0.185 s21).The extra benzene ring of the Eu31– PAN intermediate may make it more stable than the Eu31–PAR intermediate because of further conjugation. N N N O O– H Eu H2O N N N O H Eu H2O Eu3+-PAR hydrazone chelate Eu3+-PAN hydrazone chelate Table 6 Rate constants as a function of pH for the second step of the reaction of Eu(aq)31 with PAR in acetate buVer ([acetate] = 0.05 mol dm23) solution at ambient pressure and 25.0 8C ([Eu31] = 1 × 1023 mol dm23, [PAR] = 1 × 1024 mol dm23 and [NaClO4] = 0.1 mol dm23) pH 3.62 4.14 4.35 4.58 5.02 5.40 k2/s21 1.01 ± 0.04 1.03 ± 0.03 1.00 ± 0.03 0.99 ± 0.05 0.96 ± 0.04 1.02 ± 0.05 Comparison between Eu31 and UO2 21.Table 5 shows that the k2 value for the complexation of UO2 21 (0.94 s21) by PAR is about the same as that of the Eu31–PAR system (1.00 s21). This suggests that the second step of the UO2 21 complexation by PAR follows the same pathway as for the lanthanides.Desolvation and formation of the “UO2 21–PAN hydrazone intermediate” must be completed within the equatorial plane. The two axial U]] O bonds do not aVect the complexing process. Again, it is interesting to compare our k2 value for the complexation of UO2 21 by PAR with observed rate constants for the complexation of uranyl ion by 18-crown-6 or diazo-18-crown-6.86 Our k2 value is close to that of the UO2 21–diazo-18-crown-6 system (1.3 s21) and smaller than that of the UO2 21–18-crown-6 system (18 s21).Fux et al.87 proposed that the rate-determining rearrangement reaction in the UO2L “external” complex consists of metal and ligand cavity desolvations with a simultaneous rotation of the uranyl group to give the UO2L “exclusive” complex. DiVerent Ln31. Measurements were carried out under the same reaction conditions as for the first step except for the timescale. All the k2 values for the complexation of the lanthanides from Sm31 to Lu31 by PAR are close to 2.0 s21, and parallel the stability constants for Ln31–PAR2.The second step was not observed for the complexation of La31, Ce31, Pr31 and Nd31 by PAR. Our observations here are consistent with the work by Merbach and co-workers.7–12,76,80,84,85 Based on the above, the second step of the complexation of aqueous Eu31 by PAR is adequately represented by Scheme 4. Kinetic studies of the third step pH and Eu31 Concentration dependence.pH Dependence studies were carried out in succinate, acetate, imidazole, MES, HEPPS and Tris buVered solutions covering the pH range 1.80–8.80. All runs were made at a constant ionic strength of [NaClO4] = 0.1 mol dm23 and a constant 0.05 mol dm23 concentration of the basic component of the buVer. The observed rate constants as a function of pH in all buVered solutions at ambient pressure and 25.0 8C are given in Table 7. Fig. 7 shows that a plot of k3 versus pH has a typical “titration curve” from pH 4 to 7.This demonstrates that there must be a deprotonation Fig. 7 Plot of k3 versus pH for the third step of the reaction of Eu31 with PAR at ambient pressure and 25.0 8C in diVerent buVer solutions ([Eu31] = 1 × 1023 mol dm23, [PAR] = 1 × 1024 mol dm23, [NaClO4] = 0.1 mol dm23 and [buVer] = 0.05 mol dm23). BuVers: s, succinate; h, acetate; n, HEPPS; , imidazole; e, MES; ,, Tris.3574 J. Chem. Soc., Dalton Trans., 1998, 3565–3576 Scheme 4 N N N O OH(O–) H [(H2O)6(A)Eu]x+ Hydrogen bonding N N N O OH(O–) H [(H2O)5(A)Eu]x+ H2O N N N O OH(O–) H [(H2O)7(A)Eu]x+ Hydrogen bonding N N N O OH(O–) H [Eu(H2O)5(A)x+ k2 slow fast –H2O –H2O fast 'Hydrazone-Eu3+" intermediate pre-equilibrium prior to the rate -determining step for the third step of the complexation of aqueous Eu31 by PAR.The preequilibrium should be the deprotonation of o-hydroxyl with the hydrogen bonding shown. An approximate value of the pKa for the pre-equilibrium, i.e. 5.4, can be deduced from the data of Table 7/Fig. 7. Rate constants k3 for the third step of the reaction of Eu31 with PAR as a function of Eu31 concentration either in a pH 2.08 succinate buVer or a pH 4.35 acetate buVer are given in Table 1. They are seen to be independent of the Eu31 concentration either in succinate or acetate buVers, which means that an “intramolecular rearrangement” or “intramolecular ring closure” is the rate-determining step. Pressure dependence.The pressure dependence for the third step of the reaction of Eu31 with PAR was studied either in 0.05 mol dm23 acetate buVer at pH 4.35 or in 0.05 mol dm23 imidazole buVer (pH 7.10) at 25.0 8C. The k3 values as a function of Table 7 Rate constants as a function of pH for the third step of the reaction of aqueous Eu31 with PAR in diVerent buVer solutions at ambient pressure and 25.0 8C ([Eu31] = 1 × 1023 mol dm23, [PAR] = 1 × 1024 mol dm23 and [NaClO4] = 0.1 mol dm23) BuVer Succinate Acetate MES Imidazole HEPPS Tris pH 2.65 3.31 3.61 4.14 4.35 4.58 5.02 5.40 5.40 5.70 5.95 6.15 6.21 6.50 6.81 7.10 7.55 7.20 7.60 8.00 8.10 8.40 8.80 102k3/s21 0.37 ± 0.02 0.64 ± 0.03 0.68 ± 0.03 1.54 ± 0.09 1.85 ± 0.11 2.64 ± 0.18 5.74 ± 0.27 14.0 ± 0.9 15.0 ± 1.2 24.0 ± 1.2 26.0 ± 0.2 27.8 ± 2.2 28.2 ± 3.0 28.4 ± 1.3 28.6 ± 2.4 29.4 ± 2.3 30.6 ± 2.7 30.0 ± 1.5 29.6 ± 2.4 30.4 ± 1.5 29.8 ± 1.7 30.6 ± 2.6 30.6 ± 3.0 pressure are summarized in Table 3.The plots of ln k3 versus pressure indicate that the DV‡ 3 value in 0.05 mol dm23 acetate buVer at pH 4.35 is 211.2 ± 0.9 cm3 mol21 and the DV‡ 3 value in 0.05 mol dm23 imidazole buVer at pH 7.10 is 215.9 ± 0.8 cm3 mol21.These DV‡ values suggest that the rate-determining process in the third step of the complexation, like the second step described above, is dominated by an associative character and a compact transition state. Influence of buVer. The k3 values as a function of buVer concentrations in MES (pH 6.15), HEPPS (pH 7.55) and Tris (pH 8.10) buVered solutions at ambient pressure and 25.0 8C over the concentration range of 0.01–0.10 mol dm23 are given in SUP 57431. They show that in HEPPS and Tris buVers the observed rate constants for the third step of the complexation are independent of the concentrations of the buVers used.The absence of a buVer eVect in the third step suggests that the mechanism does not involve complex formation between the buVers and “hydrazone–Eu31 chelate” or the following “azo– Eu31, pseudo-phenanthroline style” chelate species.The k3 values obtained in both HEPPS and Tris are very close. However, the k3 values obtained from MES buVer (lower pH) decrease with increasing concentration of the MES buVer. We do not know the cause of this diVerence. Correlating with the pH-dependent studies shown in Fig. 7, we found that in MES buVered solution at pH 6.15 the k3 value does not reach the saturation value. Probably, the higher the concentration of the MES buVer, the more diYcult is the deprotonation of the o-hydroxyl because of intramolecular hydrogen bonding, and therefore the smaller are the k3 values.DiVerent Ln31. Kinetic measurements were performed under the same reaction conditions as for the first and the second steps except for the timescale. All the k3 values for the complexation of the lanthanides by PAR are close to 0.25 s21. Thus the third step of the complexation of aqueous Ln31 by PAR is not aVected by the nature of the lanthanide. Comparison between PAR and PAN.Table 4 indicates that the k3 value (0.01 s21) for the complexation of Eu31 by PAN is smaller than that by PAR (0.07 s21). The two negative charges due to deprotonation of both the o- and p-hydroxyl groups of N N N O O– H [Eu(H2O)5(A)]x+ N N N –O O– [Eu(H2O)6(A)]x+ + H+J. Chem. Soc., Dalton Trans., 1998, 3565–3576 3575 PAR at pH 6.15 in MES buVer make the co-ordination by the o-oxyl anion of PAR22 much easier than by PAN2 which has only one negative charge as a result of deprotonation.Therefore, the third step of the complexation of Eu31 by PAR is faster than that of PAN. Comparison between Eu31 and UO2 21. Table 5 shows that the k3 value for the complexation of UO2 21 (0.0074 s21) by PAR is smaller than that of the Eu31–PAR system (0.0185 s21). This means that the third step for UO2 21 complexation by PAR is slowed down by the two axial U]] O bonds.In order to complete the slowest step, namely, the co-ordination of the UO2 21–PAR intermediate from the second step by the o-oxyl anion of PAR, the two axial U]] O bonds must rotate to some degree. The formation of the transition state cannot be completed within the equatorial plane. The k3 value of the UO2 21–PAR system is also smaller than that in the third observed step of the complexation of UO2 21 by either 18-crown-6 (0.022 s21) or by diazo-18-crown-6 (0.283 s21).87 On the basis of the above kinetic results for the third step of the complexation of aqueous Eu31 by PAR we propose the mechanism depicted in Scheme 5.Acknowledgements Financial support by the Department of Energy, OYce of Basic Energy Sciences (Y. S. and E. M. E.) and by the Volkswagen Foundation (R. v. E.) is gratefully acknowledged. References 1 S. F. Lincoln, Adv. Inorg. Bioinorg. Mech., 1983, 4, 217. 2 S. F. Lincoln and A. E. Merbach, in Advances in Inorganic Chemistry, ed. A.G. Sykes, Academic Press, San Diego, 1995, vol. 42. 3 L. Helm and A. E. Merbach, Eur. J. Solid State Inorg. Chem., 1991, 28, 245. 4 A. Habenschuss and F. H. Spedding, J. Chem. Phys., 1980, 73, 442. 5 T. Yamaguchi, M. Nomura, H. Wakita and H. Ohtaki, J. Chem. Phys., 1988, 89, 5153. 6 K. Miyakawa, Y. Kaizu and H. Kobayashi, J. Chem. Soc., Faraday Trans., 1988, 84, 1517. 7 N. Graeppi, D. H. Powell, G. Laurenczy, L. Zekány and A. E. Merbach, Inorg. Chim.Acta, 1995, 235, 311. 8 C. Cossy, L. Helm and A. E. Merbach, Inorg. Chem., 1988, 27, 1973. 9 D. Pubanz, G. González, D. H. Powell and A. E. Merbach, Inorg. Chem., 1995, 34, 4447. 10 Th. Kowall, F. Foglia, L. Helm and A. E. Merbach, J. Am. Chem. Soc., 1995, 117, 3790. 11 K. Micskei, L. Helm, E. Brücher and A. E. Merbach, Inorg. Chem., 1993, 32, 3844. 12 C. Cossy, L. Helm and A. E. Merbach, Inorg. Chem., 1989, 28, 2699. 13 Y. Shi, E. M. Eyring and R. van Eldik, J. Chem. Soc., Dalton Trans., 1998, 967. 14 R. B. LauVer, Chem. Rev., 1987, 87, 901. 15 S. Jurisson, D. Berning, W. Jia and D. Ma, Chem. Rev., 1993, 93, 1137. Scheme 5 N N N O– [Eu(H2O)5(A)]x+ N N N –O O– [Eu(H2O)5(A)]x+ N N N –O O– [Eu(H2O)4(A)]x+ N N N O– O– [Eu(H2O)4(A)]x+ H2O + H+ k3 slow fast –H2O O H 16 M. F. Tweedle, in Lanthanide Probes in Life, Chemical and Earth Sciences: Theory and Practices, eds. J.-C. G. Bünzli and G. R. Choppin, Elsevier, Amsterdam, 1989. 17 V. Alexander, Chem. Rev., 1995, 95, 273. 18 G. R. Choppin and P. J. Wong, ACS Symp. Ser., 1994, 565. 19 H. B. Silber, N. Scheinin, G. Atkinson and J. Grecsek, J. Chem. Soc., Faraday Trans., 1972, 68, 1200. 20 R. Garnsey and D. W. Ebdon, J. Am. Chem. Soc., 1969, 91, 50. 21 N. Purdie and C. A. Vincent, Trans. Faraday, Soc., 1967, 63, 2745. 22 D. P. Fay, D. Litchinsky and N. Ourdie, J. Phys. Chem., 1969, 73, 544. 23 T. E. Eriksen, I. Grenthe and I. Puigdomenech, Inorg. Chim. Acta, 1987, 126, 131. 24 M. M. Farrow, N.Purdie and E. M. Eyring, Inorg. Chem., 1974, 13, 2024. 25 T. E. Eriksen, I. Grenthe and I. Puigdomenech, Inorg. Chim. Acta, 1986, 121, 63. 26 H. B. Silber, R. D. Farina and J. H. Swinehart, Inorg. Chem., 1969, 8, 819. 27 M. M. Farrow and N. Purdie, Inorg. Chem., 1974, 13, 2111. 28 A. J. GraVeo and J. L. Bear, J. Inorg. Nucl. Chem., 1968, 30, 1577. 29 G. R. Choppin, J. Alloys Compd., 1995, 225, 242. 30 G. A. Nyssen and D. W. Margerum, Inorg. Chem., 1970, 9, 1814. 31 K.Kumar, T. Jin, X. Wang, J. F. Desreus and M. F. Tweedle, Inorg. Chem., 1994, 33, 3823. 32 É. Tóth, E. Brücher, I. Lázár and I. Tóth, Inorg. Chem., 1994, 33, 4070. 33 K. Kumar and M. F. Tweedle, Inorg. Chem., 1993, 32, 4193. 34 E. Brücher, S. Cortes, F. Chavez and A. D. Sherry, Inorg. Chem., 1991, 30, 2092. 35 M. Kodama, T. Koike, A. B. Mahatma and E. Kimura, Inorg. Chem., 1991, 30, 1270. 36 K.-Y. Choi, K. S. Kim and J. C. Kim, Polyhedron, 1994, 13, 567. 37 E. Brücher and A. D.Sherry, Inorg. Chem., 1990, 29, 1555. 38 K. Kumar, C. A. Chang and M. F. Tweedle, Inorg. Chem., 1993, 32, 587. 39 C. A. Chang, H.-L. Peter, V. K. Manchanda and S. P. Kasprzyk, Inorg. Chem., 1988, 27, 3786. 40 J. Geoges, Spectrochim. Acta Rev., 1991, 14, 337. 41 D. Betteridge and D. Hohn, Analyst (London), 1973, 98, 377. 42 I. M. Rao, D. Satyanarayana and U. Agarwala, Bull. Chem. Soc. Jpn., 1979, 52, 588. 43 P. M. Drozdzewski, Spectrochim. Acta, Part A, 1985, 41, 1035. 44 W. J. Geary, G.Nickless and F. H. Pollard, Anal. Chim. Acta, 1962, 26, 575. 45 F. Zhao, Y. Liang, Y. Zhang and D. Tong, Chin. Chem. Lett., 1992, 3, 107. 46 A. Corsini, I. M. Yih, Q. Fernando and H. Freiser, Anal. Chem., 1962, 34, 1090. 47 A. Corsini, Q. Fernando and H. Freiser, Inorg. Chem., 1963, 2, 224. 48 H. J. Méndez, B. M. Cordero, J. L. Pavón and J. C. Miralles, Inorg. Chim. Acta, 1987, 140, 245. 49 A. Cladera, E. Gómez, J. M. Estela and V. Cerdà, Anal. Chem., 1993, 65, 707. 50 K. Ueda, H. Matsuda and O. Yoshimura, Anal. Sci., 1989, 5, 725. 51 E. Ohyoshi, Talanta, 1984, 31, 1129. 52 E. Ohyoshi, Anal. Chem., 1983, 55, 2404. 53 M. Tanaka, S. Funahashi and K. Shirai, Anal. Chim. Acta, 1967, 39, 437. 54 B. Perlmutter-Hayman and R. Shinar, Inorg. Chem., 1976, 15, 2932. 55 R. L. Reeves, Inorg. Chem., 1986, 25, 1473. 56 R. H. Hoyler, C. D. Hubbard, S. F. A. Kettle and R. G. Wilkins, Inorg. Chem., 1966, 5, 622. 57 R. L. Reeves, G. S. Calabrese and S. A. Harkaway, Inorg. Chem., 1983, 25, 3076. 58 E. Mentasti and C. Baiocchi, J. Chem. Soc., Dalton Trans., 1985, 2615. 59 H. L. Fritz and J. H. Swinehart, Inorg. Chem., 1975, 14, 1935. 60 G. Meyers, F. M. Michaels, R. L. Reeves and P. P. Trotter, Inorg. Chem., 1985, 24, 731. 61 M. Cusumano, Inorg. Chim. Acta, 1977, 25, 207. 62 S. Funahashi and M. Tanaka, Inorg. Chem., 1969, 8, 2159. 63 C. D. Hubbard and A. D. Pacheco, J. Inorg. Nucl. Chem., 1977, 39, 1373. 64 C. F. Shaw III, J. E. Laib, M. M. Savas and D. H. Petering, Inorg. Chem., 1990, 29, 403. 65 S. Funahashi and M. Tanaka, Bull. Chem. Soc. Jpn., 1976, 49, 2481. 66 H. Wada and G. Nakagawa, Bull. Chem. Soc. Jpn., 1979, 52, 3559. 67 R. Van Eldik, W. Gaede, S. Wieland, J. Kraft, M. Spitzer and D. A. Palmer, Rev. Sci. Instrum., 1993, 34, 1355.3576 J. Chem. Soc., Dalton Trans., 1998, 3565–3576 68 H. Hoshino, T. Yotsuyanagi and K. Aomura, Anal. Chim. Acta, 1976, 83, 317. 69 M. Eigen and R. G. Wilkins, Adv. Chem. Ser., 1965, 49, 55. 70 G. R. Choppin, in Lanthanide Probes in Life, Chemical and Earth Sciences: Theory and Practices, eds. J.-C. G. Bünzli and G. R. Choppin, Elsevier, Amsterdam, 1989. 71 H. Cohen, W. Gaede, A. Gerhard, D. Meyerstein and R. van Eldik, Inorg. Chem., 1992, 31, 3805. 72 W. Gaede, R. van Eldik, H. Cohen and D. Meyerstein, Inorg. Chem., 1993, 32, 1997. 73 W. Gaede, A. Gerhard, R. van Eldik, H. Cohen and D. Meyerstein, J. Chem. Soc., Dalton Trans., 1992, 2065. 74 T. Shi and L. I. Elding, Inorg. Chem., 1996, 35, 735. 75 K. S. Bai and A. E. Martell, J. Inorg. Nucl. Chem., 1969, 31, 1697. 76 C. Cossy, A. C. Barnes, A. E. Merbach and J. Enderby, J. Chem. Phys., 1989, 90, 3254. 77 K. Miyakawa, Y. Kaizu and H. Kobayashi, J. Chem. Soc., Faraday Trans., 1988, 90, 3254. 78 T. W. Swaddle, Adv. Inorg. Bioinorg. Mech., 1983, 2, 95. 79 S. F. Lincoln, Adv. Inorg. Bioinorganic. Mech., 1986, 4, 217. 80 C. Cossy and A. E. Merbach, Pure. Appl. Chem., 1988, 60, 1785. 81 L. Helm and A. E. Merbach, Eur. J. Solid State Inorg. Chem., 1991, 28, 245. 82 A. J. Habenschuss and F. H. Spedding, J. Chem. Phys., 1980, 73, 442. 83 F. Spedding, L. E. Shiers, M. A. Brown, J. L. Derer, D. L. Swanson and A. J. Habenschuss, J. Chem. Eng. Data, 1975, 20, 81. 84 K. Micskei, D. H. Powell, L. Helm, E. Brücher and A. E. Merbach, Magn. Reson. Chem., 1993, 31, 1011. 85 A. Abou-Hamdan, N. Burki, S. F. Lincoln, A. E. Merbach and S. J. F. Vincent, Inorg. Chim. Acta, 1993, 207, 27. 86 G. L. Honan, S. F. Lincoln and E. H. Williams, J. Chem. Soc., Dalton Trans., 1979, 320. 87 P. Fux, J. Lagrange and P. Lagrange, J. Am. Chem. Soc., 1985, 107, 5927. 88 K. Mochizuki, T. Imamura, T. Ito and M. Fujimoto, Chem. Lett., 1977, 1239. Paper 8/05808C

ISSN:1477-9226    

DOI:10.1039/a805808c

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3577–3585 3577 Kinetic control of reactions of a sterically hindered platinum picoline anticancer complex with guanosine 59-monophosphate and glutathione Yu Chen, Zijian Guo, John A. Parkinson and Peter J. Sadler * Department of Chemistry, University of Edinburgh, King’s Buildings, West Mains Road, Edinburgh, UK EH9 3JJ Received 19th August 1998, Accepted 11th September 1998 Kinetic studies (296 K, 0.1 M NaClO4, pH 6–7) of reactions of the anticancer complex cis-[PtCl2(NH3)(2-pic)] 1 (AMD473) (2-pic = 2-picoline) with guanosine 59-monophosphate (59-GMP) and the tripeptide glutathione (GSH) using 2D [1H, 15N] HSQC NMR spectroscopy have been made, and compared to reactions of the isomeric complex cis-[PtCl2(NH3)(3-pic)] 2.Reactions with 59-GMP followed two pathways with either hydrolysis trans to NH3 or picoline as the first step, with subsequent formation of Cl/GMP and H2O/GMP intermediates, and cis-[Pt(15NH3)(pic)(59-GMP-N7)2]21 as the final product. Eight rate constants were determined for each starting platinum complex 1 and 2.The rates of ligand substitution (Cl2 by H2O and H2O by 59-GMP) cis to 2-picoline were 2–12 times slower than the same ligand substitution cis to 3-picoline. This was also the case for ligand substitution trans to 2-picoline (2–3 times slower), except that when 59-GMP was present as the cis ligand (second stage of substitution) the rate of substitution was enhanced for the 2-picoline complex.Slow rotation about the Pt–N picoline bond (0.62 s21) and fast rotation about Pt-N7 GMP bonds on the NMR timescale were observed at 296 K for the bis(GMP) adduct of complex 1, while these were both fast for the analogous adduct of complex 2. Reactions of GSH with 1 were ca. 3 times slower than those with 2, and appeared to proceed via aquated intermediates with initial substitution trans to 2-picoline for 1 and trans to NH3 for 2, but no kinetic analyses were attempted due to the complexity of the reactions.Both mono- and bis-GMP adducts were observed during competitive reactions of GSH and 59-GMP with complex 1 (molar ratio: 2:2:1) at pH 7, 296 K. These features of the chemistry of 1 may play an important role in its altered spectrum of biological activity compared to cisplatin. Introduction Cisplatin, cis-[PtCl2(NH3)2], is a widely used anticancer drug. Because of its serious toxic side-eVects and the spontaneous development of drug resistance in tumours, there is a need for new drugs which circumvent these drawbacks.Investigations have been made of the activity of many other platinum complexes, most of which belong to the structural class cis- [PtX2(ammine)2] (X = anionic leaving group; ammine = ammonia, primary or secondary amine).1 An exception is the 2-picoline (2-methylpyridine) complex cis-[PtCl2(NH3)(2-pic)] 1 (AMD473), which is currently in phase I clinical trials.2,3 It possesses activity against cisplatin-resistant cell lines and against an acquired cisplatin-resistant subline of a human ovarian carcinoma xenograph both by injection and by oral administration.2 It appears to circumvent thiol-mediated resistance mechanisms whilst still maintaining the ability to form cytotoxic lesions with DNA.3,4 In order to gain an understanding of the chemical reactivity of this complex, we have recently investigated its crystal structure and hydrolysis behaviour in comparison with the isomeric 3-picoline complex cis-[PtCl2(NH3)(3-pic)] 2.5 The most notable feature of the structures is the orientation of the picoline ring with respect to the platinum square plane.The 2-picoline ligand is almost perpendicular (1038) such that the 2-methyl group lies directly over the square plane, while the 3-picoline ligand is tilted by 498. This steric eVect plays an important role in determining the hydrolysis rates of the Cl2 ligands.For complex 2 the rate of hydrolysis of Cl2 trans to NH3 is similar to that of cisplatin (t1/2 = 1.75 h, 310 K),6 while for complex 1 the rate of hydrolysis for the Cl2 ligand trans to NH3 (cis to 2-picoline) is about 5 times slower (t1/2 = 8.7 h). The slow hydrolysis of 1 can be attributed to the axial steric hindrance provided by the 2-methyl group. Complex 1 appears to form interstrand DNA cross-links and to bind to plasma proteins much more slowly than cisplatin.7 Besides steric eVects, electronic factors may also play a role in determining the reactivity of pyridine complexes. For example the presence of planar pyridine ligands in cis- or trans- [PtCl2(py)2] complexes can reduce the rates of DNA binding compared to ammine complexes.8 DNA platination is a key event in the mechanism of action of platinum anticancer drugs, and the major adduct formed by attack of cisplatin on DNA is the intrastrand cross-link between N7 atoms of two adjacent guanine (G) residues.1 However, platinum can also interact with many other biomolecules, especially those containing methionine and cysteine residues.Glutathione (GSH), a cysteine-containing tripeptide (g-L-Glu-L-Cys-Gly), is the predominant intracellular thiol with concentrations typically ranging from 0.5 to 10 mM. At physiological pH, platinum(II) complexes usually show a kinetic preference for the thiols3578 J. Chem. Soc., Dalton Trans., 1998, 3577–3585 cysteine and glutathione over 59-GMP, even in the presence of excess of nucleotide.9 Reactions of thiols with platinum complexes are often considered to have a negative eVect on antitumor activity and to be responsible for drug inactivation and the development of drug resistance.10 GSH is overexpressed in cisplatin-resistant cells and Pt–GS adducts can be pumped out of cells.11 In this work we have investigated reactions of complex 1 and the less-hindered isomeric complex 2 with 59-GMP and glutathione by both 1D 1H and 2D [1H, 15N] HSQC (heteronuclear single quantum coherence) NMR spectroscopy, in order to investigate the influence of both steric and electronic eVects on reactions of potential biological importance. Experimental Materials 2- and 3-Picoline and GSH were purchased from Aldrich. The sodium salt of 59-GMP was obtained from Sigma.The complex cis-[PtCl2(15NH3)2] was prepared according to the reported procedure.12 The 15N labelled complexes 1 and 2 were prepared from cis-[PtCl2(15NH3)2] by a similar procedure to that described in the literature for natural abundance mixed-ligand ammine–amine platinum(II) complexes.13 The diaqua complex cis-[Pt(NH3)(2-pic)(H2O)2]21 was prepared in situ by addition of slightly less than 2 mol equivalents of AgNO3 to an aqueous solution of complex 1 followed by removal of the AgCl precipitate.NMR spectroscopy The NMR spectra were recorded on a Bruker DMX500 spectrometer operating at 500.13 MHz using a TBI probehead.All data processing, including the integration routines described below, was carried out using XWIN-NMR, version 1.3 (Bruker Spectrospin Ltd.). The chemical shift references were as follows: 1,4-dioxane (internal, d 3.767) for 1H, and 1 M 15NH4Cl in 1.5 M HCl for 15N (external). All spectra were recorded at 296 K unless otherwise stated. Typical acquisition conditions for 1H spectra were: 45–608 pulses, 2.5 s relaxation delay, 64–256 transients, final digital resolution 0.2 Hz per point.The water resonance was suppressed by presaturation, or via the WATERGATE pulsed-field-gradient sequence.14 The 2D [1H, 15N] HSQC NMR spectra (optimised for 1JNH = 72 Hz) were recorded by using the sequence of Stonehouse et al.15 The 15N spins were decoupled by irradiating with the GARP-1 decoupling sequence during acquisition. The 2D exchange experiment was performed using phase-sensitive nuclear Overhauser eVect spectroscopy (NOESY) with a mixing time 1 s, at 296 K.Rates were calculated using the program D2DNMR kindly supplied by Dr K. G. Orrell.16 Inputs for each calculation consisted of volume integrals from diagonaland cross-peaks together with population estimates based on 1D 1H NMR spectra. All samples were in 90% H2O–10% D2O (0.6 ml). The reactions of complex 1 or 2 (3 mM) with 59-GMP were conducted at a 1 : 2 molar ratio. Samples contained 0.1 M NaClO4 to maintain a constant ionic strength. BuVer was not used in the reactions of complexes 1 and 2 with 59-GMP in order to avoid possible interference, and the pH values were adjusted to 6.85 and 6.55 (for complexes 1 and 2, respectively) at the beginning of the reaction.The pH value for the reaction of [Pt(NH3)- (2-pic)(H2O)2]21 with 59-GMP (1: 2) was adjusted to 6.47. In the reactions of complex 1 or 2 with GSH (3 mM, 1 : 1 molar ratio), 100 mM phosphate buVer (pH 7) was used. Competitive reactions between 59-GMP and GSH with complexes 1 or 2 (2 mM) were carried out at 2:2:1 molar ratios, with 100 mM phosphate buVer present to maintain neutral pH.After mixing, argon was bubbled through the solution to minimise GSH oxidation and the NMR samples were carefully sealed. pH Measurements These were made using a Corning 145 pH meter equipped with an Aldrich micro combination electrode calibrated with Aldrich buVer solutions of pH 4, 7 and 10. Values of pH were adjusted with 1 M HClO4 or NaOH as appropriate.Kinetic measurements For kinetic analysis of NMR spectra, peak volumes were measured and the relative concentrations of each species were calculated at each time point. The appropriate diVerential equations were integrated numerically, and the rate constants were determined by a non-linear optimisation procedure using the program SCIENTIST (version 2.01, MicroMath Inc.). The errors represent one standard deviation. Results Labelling the ammine ligand with 15N, allowed investigation of the specificity and rates of substitution of the chloride ligands in complexes 1 and 2 by 59-GMP and GSH to be made for the first time using 2D [1H, 15N] NMR spectroscopy.The 15N chemical shift is diagnostic of the co-ordinating atom of the ligand in the position trans to the ammine.17 For an ammine ligand trans to an oxygen donor in a platinum(II) complex, the 15N shift lies between d 275 and 290, for an ammine trans to nitrogen/chloride, between d 255 and 270, and for ammine trans to sulfur, between d 240 and 250.Reactions of complexes 1 and 2 with 59-GMP Both 1D 1H and 2D [1H, 15N] HSQC NMR spectroscopy were used to monitor the reactions between complex 1 or 2 (3 mM) and 59-GMP in a 1 : 2 molar ratio at 296 K and pH 6.85 or 6.55, respectively. The 2D [1H, 15N] HSQC NMR spectra recorded 40 h after mixing for complex 1 and 20 h for complex 2 are shown in Fig. 1A and 1B, respectively. For the 3-picoline complex 2, the peak assignable to 2 at d 4.26/262.68 (peak 2a) was accompanied by two new peaks (2b and 2c) with chemical shifts of d 4.12/279.62 and 4.33/261.31 after 1 h.The 15N shift of the former peak is typical of NH3 trans to O and is consistent with assignment to [PtCl(15NH3)(3-pic)(H2O)]1 2b,5 and the latter peak to the aqua complex 2c with 15NH3 trans to Cl, Table 1 (for the structures of complexes, see Scheme 1). The intensity of the cross-peak from the monoaqua complex 2b, in which the H2O is trans to NH3, is much stronger than that of complex 2c with H2O trans to 3-picoline. After 2 h, two other new peaks (2d and 2e) appeared at d 4.27/264.21 and 4.53/ 260.81, which are both in the region of 15NH3 trans to N or Cl.By comparing the intensities of these two peaks with those for the two monoaqua species (2b, 2c), the stronger cross-peak 2d can be assigned to [PtCl(15NH3)(3-pic)(59-GMP-N7)]1 (GMP trans to NH3), and cross-peak 2e to [PtCl(15NH3)- (3-pic)(59-GMP-N7)]1 (GMP trans to 3-picoline).These two peaks reached a maximum intensity after 5 to 6 h and then gradually decreased in intensity. Cross-peaks 2f at d 4.44/ 260.93 (NH3 trans to N) and 2g at d 4.47/275.92 (NH3 trans to O) diVer greatly in intensity, comparable to the diVerence observed for cross-peaks 2d and 2e (Fig. 1B). Therefore, crosspeak 2f is assignable to [Pt(15NH3)(3-pic)(59-GMP-N7)(H2O)]21 (GMP trans to NH3) and cross-peak 2g to [Pt(15NH3)(3-pic)- (59-GMP-N7)(H2O)]21 (GMP trans to 3-picoline).Cross-peak 2h at d 4.63/262.60, which was observable from soon after the beginning of the reaction, increased in intensity with time, and became the major cross-peak after 24 h. This is assignable to the final bis 59-GMP product cis-[Pt(15NH3)(3-pic)(59-GMPN7) 2]21 2h. An unassigned cross-peak 2i at d 4.33/259.55 appeared after 8 h of reaction, but accounted for only a very small amount of the total Pt–NH3 species present (<4%), andJ.Chem. Soc., Dalton Trans., 1998, 3577–3585 3579 disappeared after 1 week. The assignments and chemical shifts of all peaks are listed in Table 1. The reaction pathway is summarised in Scheme 1. The reaction of the 2-picoline complex 1 with 59-GMP was much slower in comparison with complex 2. Only peaks for Fig. 1 The 2D [1H, 15N] HSQC NMR spectra from reactions of 2 mol equivalents of 59-GMP (0.1 M NaClO4, 296 K) with (A) (15NH3)-1, pH 6.85, after 40 h and (B) (15NH3)-2 pH 6.55, after 20 h.Peak assignments are given in Table 1. Table 1 Proton and 15N NMR Pt–NH3 chemical shifts at 296 K for cis-[Pt(15NH3)(2-pic)]21 and cis-[Pt(15NH3)(3-pic)]21 complexes Compound [PtCl2(15NH3)(2-pic)] [PtCl(15NH3)(2-pic)(H2O)]1 [PtCl(15NH3)(2-pic)(H2O)]1 [PtCl(15NH3)(2-pic)(59-GMP-N7)]1 [PtCl(15NH3)(2-pic)(59-GMP-N7)]1 [Pt(15NH3)(2-pic)(59-GMP-N7)(H2O)]21 [Pt(15NH3)(2-pic)(59-GMP-N7)(H2O)]21 [Pt(15NH3)(2-pic)(59-GMP-N7)2]21 [PtCl2(15NH3)(3-pic)] [PtCl(15NH3)(3-pic)(H2O)]1 [PtCl(15NH3)(3-pic)(H2O)]1 [PtCl(15NH3)(3-pic)(59-GMP-N7)]1 [PtCl(15NH3)(3-pic)(59-GMP-N7)]1 [Pt(15NH3)(3-pic)(59-GMP-N7)(H2O)]21 [Pt(15NH3)(3-pic)(59-GMP-N7)(H2O)]21 [Pt(15NH3)(3-pic)(59-GMP-N7)2]21 Unassigned Peak * 1a 1b 1c 1d 1e 1f 1g 1h 2a 2b 2c 2d 2e 2f 2g 2h 2i d(1H) 4.19 3.99 4.23 4.22 4.47 4.35 4.36 4.59 4.58 4.26 4.12 4.33 4.27 4.53 4.44 4.47 4.63 4.33 d(15N) (15NH3 trans to) 266.75 (Cl) 282.61 (O) 265.34 (Cl) 267.43 (N) 264.51 (Cl) 264.18 (N) 278.60 (O) 265.87 (N) 266.15 (N) 262.68 (Cl) 279.62 (O) 261.31 (Cl) 264.21 (N) 260.81 (Cl) 260.93 (N) 275.92 (O) 262.60 (N) 259.55 (N) * See Fig. 1. complex 1 (peak 1a) (d 4.19/266.75) and two monoaqua hydrolysis adducts (peaks 1b and 1c) with similar intensity (d 3.99/282.61, 4.23/265.34, respectively) were observed after 2 h at 296 K. The chemical shift distribution of the cross-peaks for the 2-picoline complexes in Fig. 1A is very similar to that for the 3-picoline complexes in Fig. 1B, except that all the 15N chemical shifts of the signals in Fig. 1B are shifted to low frequency by ca. 3 ppm. However, the relative intensities of the cross-peaks for the intermediates in the reaction of complex 1 with 59-GMP are diVerent from those for complex 2. Crosspeaks 1e at d 4.47/264.51 and 1g at d 4.36/278.60 are stronger than the peaks 1d at d 4.22/267.43 and 1f at d 4.35/264.18, respectively, whereas, for complex 2, cross-peaks 2d and 2f are much stronger than 2e and 2g. The two cross-peaks 1b and 1c for the monoaqua adducts of complex 1 had similar intensities during the reaction.The final bis(GMP) adduct is formed much more slowly for complex 1 compared to complex 2. The broad cross-peak (1h) for the bis(GMP) adduct of complex 1 appeared to be composed of two overlapping cross-peaks at d 4.59/265.96 and 4.58/266.15. The peak intensity versus time profiles (Fig. 2) allowed similar assignments to be made for the cross-peaks obtained from reactions between complexes 1 Scheme 13580 J.Chem. Soc., Dalton Trans., 1998, 3577–3585 Fig. 2 Plots of the relative concentrations of species a to h (based on Pt–NH3 peak integrals) versus time for the reactions shown in Fig. 1. (A) 2- Picoline complex 1, (B) 3-picoline complex 2. Peak labels: (a) r, (b) h, (c) m, (d) ×, (e) 1, (f) j, (g) d, (h) n. The curves are the computer best fits for the rate constants given in Table 2. and 2 with 59-GMP (Table 1). Cross-peaks 1g, 1f and 1h were also observed during reaction of the diaqua complex cis- [Pt(15NH3)(2-pic)(H2O)2]21 (3 mM) with 59-GMP (1: 2) at 296 K and pH 6.47.Kinetic fits to the reaction profile in Scheme 1 for complexes 1 and 2 are shown in Fig. 2A and 2B, and the rate constants for each step are listed in Table 2. Most of the rate constants for the reaction steps of complex 2 with 59-GMP are more than twice as large as those for complex 1, except k5 and k7. The largest diVerences are for k6 and k8 (substitution of Cl2 cis to picoline by H2O, and then by 59-GMP in the second stage), which are five and twelve times larger for complex 2.Rotation of 59-GMP and 2-picoline in cis-[Pt(NH3)(2-pic)- (59-GMP)2]21 The 1H NMR spectra for the reactions between complex 1 or 2 and 59-GMP (1: 2 reaction ratio) at 296 K recorded after Table 2 Rate constants for reactions of cis-[PtCl(NH3)(pic)] with 59- GMP at 296 K (0.1 M NaClO4). The errors represent one standard deviation Rate Complex Ratio k3-pico/ constants k1/s21 k2/s21 k3/M21 s21 k4/M21 s21 k5/s21 k6/s21 k7/M21 s21 k8/M21 s21 2-Picoline 6.87 (± 0.18) × 1026 5.87 (± 0.18) × 1026 7.97 (± 0.33) × 1023 6.67 (± 0.37) × 1023 8.53 (± 1.11) × 1025 2.77 (± 0.31) × 1025 4.2 (± 0.6) × 1022 5.92 (± 0.74) × 1023 3-Picoline 2.60 (± 0.03) × 1025 1.17 (± 0.03) × 1025 1.60 (± 0.03) × 1022 1.87 (± 0.08) × 1022 1.78 (± 0.04) × 1025 1.48 (± 0.14) × 1024 3.2 (± 0.1) × 1022 6.85 (± 0.75) × 1022 k2-pico 4223 0.2 5 0.8 12 2 h and 1 week are shown in Fig. 3. For complex 1, after 1 week of reaction, the H8 resonance of free 59-GMP at d 8.19 and the methyl signal of complex 1 at d 3.15 have completely disappeared (Fig. 3A). Four major H8 signals were observed in two sets of two singlets between d 8.7 and 8.3, and two methyl signals appeared at d 3.30 and 3.27. The H19 signals of bound 59-GMP were also observed in two sets of two doublets, with vicinal coupling constants 3J(H19–H29) = 4.8, 4.4 and 5.1, 4.9 ± 0.1 Hz.However, for complex 2 after 1 week of reaction, only two H8 signals at d 8.66 and 8.52 were observed together with one methyl signal at d 2.36 and two H19 doublets [3J(H19– H29) = 4.6 and 4.9 ± 0.1 Hz] for bound 59-GMP (Fig. 3B). The 1H NMR data for cis-[Pt(NH3)(2-pic)(59-GMP)2]21 1h and cis-[Pt(NH3)(3-pic)(59-GMP)2]21 2h are listed in Table 3. The temperature dependence of the H8 signals of [Pt(NH3)- (2-pic)(59-GMP)2]21 is shown in Fig. 4. The four H8 singlets coalesced into two singlets when the temperature was raised to 338 K.Based on the coalescence temperature (Tc) and the chemical shift diVerences (Dn in Hz) between the two signals in the slow exchange limit, a rate constant kc of 25.5 s21 (kc = 2.22 Dn) and activation free energy DG‡ of 73.2 kJ mol21 for the exchange process at 338 K were calculated.18 At 338 K, the two 1H, 15N cross-peaks which constituted peak 1h (Fig. 1A) for bis(GMP) adducts of complex 1 merged into one cross-peak.The two methyl signals at d 3.30 and 3.27 also became closer when the temperature was raised to over 333 K. Because of the temperature limitation of the NMR probe, the spectrum was recorded at a maximum temperature of 348 K, where no coalescence of the two methyl signals was observed. When the spectra were recorded at 250 MHz the two methyl signals were observed to coalescence at 343 K, from which similar kc (25.6 s21) and DG‡ (75.15 kJ mol21) values were obtained. A 2D EXSY (exchange spectroscopy) experiment with a mixing time of 1 s (Fig. 5) showed clear exchange cross-peaks between theJ. Chem. Soc., Dalton Trans., 1998, 3577–3585 3581 Table 3 1H Chemical shifts and coupling constants (Hz) for bis(59-GMP) adducts of complexes 1 (pH 6.85) and 2 (pH 6.55) at 296 K d Complex 1 cis-[PtCl2(NH3)(2-pic)] [Pt(NH3)(2-pic)(59-GMP)2]21 2 cis-[PtCl2(NH3)(3-pic)] [Pt(NH3)(3-pic)(59-GMP)2]21 H8 8.66, 8.63 8.42, 8.40 8.66, 8.52 CH3 3.15 3.30, 3.27 2.37 2.36 H19 5.93, 5.92 5.80, 5.78 5.93, 5.82 3J(H19–H29)/Hz 4.8, 4.4 5.1, 4.9 4.6, 4.9 two methyl signals at 296 K.An average rate constant for the exchange process of 0.62 s21 at this temperature was determined. Interestingly, one of the four H8 signals at d 8.42 (Fig. 3A) for the bis(GMP) adduct of complex 1 became broader when the temperature was below 296 K, as did one of the CH3 signals at d 3.30. Reactions of complexes 1 and 2 with glutathione (1 : 1, pH 7) Reactions of complexes 1 and 2 with GSH were followed by 2D [1H, 15N] HSQC NMR spectroscopy.Spectra recorded after 3.5 h are shown in Fig. 6. In both cases, peaks assignable to hydrolysis products (1b, 1c and 2b, 2c) were observed before the appearance of peaks for GSH adducts. For complex 2, crosspeaks from GSH adducts began to appear after 45 min in the d(15N) region of 235 to 240 which corresponds to NH3 trans to S, while for complex 1 new peaks were observed only after 2 h and in the d(15N) region of 262 to 267 which corresponds to NH3 trans to N or Cl.Therefore, the first binding site for GS2 is trans to 2-picoline for complex 1, but cis to 3-picoline for complex 2. Due to the complicated nature of the reactions (Table 4), no specific assignments for the products of the reactions between complex 1 and 2 and GSH have been made. Time Fig. 3 The 1H NMR spectra recorded after 2 h and 1 week of reactions at 296 K of 59-GMP (2 mol equivalents) with (A) complex 1 and (B) complex 2.The peaks labelled * are from the aromatic protons of picoline ligands. Fig. 4 The temperature dependence of the H8 1H NMR signals of cis- [Pt(NH3)(2-pic)(59-GMP)2]21. Fig. 5 The 2D EXSY spectrum (mixing time = 1 s) for cis-[Pt(NH3)- (2-pic)(59-GMP)2]21 showing exchange cross-peaks between the two methyl signals, indicative of the slow rotation of co-ordinated 2- picoline.3582 J. Chem. Soc., Dalton Trans., 1998, 3577–3585 dependences of the concentrations of complexes 1 and 2 and their monoaquachloro adducts during the reactions with GSH Fig. 6 Two-dimensional [1H, 15N] HSQC NMR spectra for reactions of GSH (1 mol equivalent) with (A) complex 1 and (B) complex 2, recorded at 296 K, pH 7 (100 mM phosphate buVer) after 3.5 h. Peaks (1a to 1c, and 2a to 2c) are assigned according to Table 1; the others are unassigned. Satellites (195Pt) of peaks 1a and 2a are labelled with *. Peaks with 15N chemical shifts near d 240 are characteristic of NH3–Pt trans to sulfur.Table 4 Proton and 15N NMR Pt–NH3 chemical shifts at 296 K for the major peaks observed during the reactions of complexes 1 and 2 with GSH (1: 1 molar ratio, pH 7) Peak a 1a 1b 1c 1j 1p 1q ***** 2a 2b 2c 2j 2k 2m 2n 2p d(1H) 4.21 3.89 4.23 4.27 3.60 3.97, 3.95 4.29 4.25 3.96 4.05 3.69 4.27 3.93 4.30 4.32 4.10 4.05 3.97 3.68 d(15N) (15NH3 trans to) 266.84 (Cl) 281.27 (O) 265.49 (Cl) 262.64 (N or Cl) 240.85 (S) 265.84, 265.95 (N or Cl) 262.52 (N or Cl) 262.89 (N or Cl) 243.24 (S) 242.76 (S) 242.12 (S) 262.82 (Cl) 276.77 (O) 261.74 (Cl) 258.14 (N or Cl) 240.09 (S) 239.01 (S) 239.37 (S) 236.70 (S) a For peak labels, see Fig. 6. Peaks labelled * were not observed in the same spectrum as the other peaks but appeared at later times. are shown in Fig. 7. The rate of reaction of complex 2 is about 3 times as fast as that of complex 1. After 2 weeks’ standing at ambient temperature, yellow precipitates formed in both reactions, and all the signals had disappeared from the 2D [1H, 15N] HSQC NMR spectra.Competitive reactions of 1 or 2 with GSH and 59-GMP The reaction of complex 1 with GSH in the presence of 59- GMP (1:2:2) was followed by 2D [1H, 15N] NMR spectroscopy at 296 K, pH 7 (100 mM phosphate buVer). A spectrum recorded after 14.5 h of reaction is shown in Fig. 8. Only the peaks for the starting complex 1 (peak 1a) and hydrolysis products (peaks 1b, 1c) were present after 0.5 h. Peak 1q, which was also observed during the reaction with GSH (Fig. 6A), began to appear after 1 h of reaction. The mono(GMP) adduct (peak 1e) began to form after 2.5 h, and the bis(GMP) adduct (peak 1h) was present after 8 h of reaction. Peaks due to both 59-GMP adducts (1d, 1e, 1f and 1h) and GSH adducts (peaks Fig. 7 Time dependence of the concentrations of the starting complex 1 (j), its monoaquachloro adducts (1b 1 1c) (m), and complex 2 (s), its monoaquachloro adducts (2b 1 2c) (×), during reactions with GSH (1 : 1 molar ratio, pH 7, 296 K).Fig. 8 The 2D [1H, 15N] HSQC NMR spectra for the competitive reaction of GSH and 59-GMP with complex 1 (molar ratio: 2:2:1) at 296 K, pH 7.0 (100 mM phosphate buVer) recorded after 14.5 h. Peaks of 59-GMP adducts (1d, 1e, 1h) and GSH adducts (1j, 1q) are labelled according to those in Tables 1 and 4. Satellites of peak 1a are labelled with *. The peaks in the dashed box are due to the formation of GSH adducts (15N chemical shifts near d 240 are characteristic of NH3–Pt trans to sulfur), but no specific assignments were made.Peak 1r is tentatively assigned to an adduct with one GMP and one GSH ligand in the cis position (NH3–Pt trans to N7).J. Chem. Soc., Dalton Trans., 1998, 3577–3585 3583 1q, 1j, 1r and those in dashed box) are present in Fig. 8. Due to the complicated nature of GSH reactions, no attempt was made to assign the peaks 1q, 1j or those in the region for NH3–Pt trans to S (dashed box, Fig. 8). Most of these peaks for GSH adducts were also observed in the reaction of complex 1 with GSH in the absence of 59-GMP. Peak 1r has a 15N chemical shift in the region of NH3–Pt trans to N or Cl, and was not observed in the reactions of 1 with 59-GMP or with GSH. It can be tentatively assigned to a mixed-ligand adduct containing GMP trans to NH3 and a GS2 ligand. Spectra recorded after 3 d of reaction at 296 K showed that only one major peak (1h) in the region of NH3–Pt trans to N or Cl [d(15N) 260 to 270] together with some other peaks [d(15N) 240 to 245] in the region of NH3–Pt trans to S.The presence of a strong H8 signal at d 8.13 due to free 59-GMP in 1D spectra recorded after 3 d showed that most of the 59-GMP was still unreacted at this stage, while signals for free GSH had nearly disappeared. The 2-methyl signal at d 3.15 from co-ordinated 2-picoline had shifted to d 2.57, which implied that the 2-picoline ligand had been mostly displaced.Signals for the bis(GMP) adduct were only just visible in 1D spectra because of their low intensity and overlap with broad signals. Compared with 1, the competitive reaction of 2 with GSH and 59-GMP (1:2:2) was much faster and peaks assignable to GSH adducts were detectable after 0.5 h. Mono- and bis-GMP products started to appear after 1.5 and 5 h of reaction, respectively. Much less of the bis(GMP) adduct of complex 2 was formed in the competitive reaction with GSH compared to complex 1.Discussion The complex cis-[PtCl2(NH3)(2-pic)] 1 is a new anticancer drug which has a diVerent spectrum of biological activity compared to cisplatin. In particular it is active against cisplatin-resistant cell lines and xenographs.2–4 In order to shed light on the chemical reactivity of 1, and to elucidate the potential role of steric hindrance,5 we have made kinetic studies of reactions of complex 1 with 59-GMP and GSH in comparison to the isomeric 3-picoline complex 2.By labelling complexes with 15NH3, it was possible to determine for the first time the specificity of substitution of the chloride ligands by 59-GMP and GSH using 2D [1H, 15N] HSQC NMR spectroscopy. Because of the sensitivity of the 15N chemical shifts to the ligand trans to H3N–Pt, and of 1H NH3 shifts to cis eVects, it has been possible to resolve peaks for a variety of intermediates in the reactions, detect them at micromolar concentrations, and make reasonable assignments.Reactions with 59-GMP Concentration versus time profiles for reactions of complexes 1 and 2 with 59-GMP were fitted well by the kinetic scheme shown in Scheme 1, showing that they are two-step and two stage: initial hydrolysis followed by 59-GMP substitution to give the 1 : 1 Pt–59GMP adduct, and then further hydrolysis and 59-GMP substitution to give the final product, the bis- (59-GMP) complex. For each complex there are two routes to the final product, depending on whether initial substitution is trans to picoline or trans to NH3.From the rate constants listed in Table 2, the following conclusions can be drawn. (i) The rates of ligand substitution (Cl2 by H2O, or H2O by 59-GMP) cis to picoline (k1, k3, k6, k8) are 2–12 times slower for the 2-picoline complexes compared to 3-picoline. (ii) The rates of substitution trans to picoline are also slower for the 2-picoline complex (2–3 times slower) except when 59-GMP is present as the cis ligand (compare k5 and k7 values for complexes 1 and 2, Table 2).However, it should be noted that the errors for some of the rate constants determined for the second stage of the reaction are rather large on account of the low concentrations of the intermediates. The relative rates of the first step of the GMP reactions are similar to those we determined previously for hydrolysis of the complexes, and for complex 2 these follow the order expected from the higher trans influence of NH3 (pKa 9.29) compared to 3-picoline (pKa 6.0),19 which is the weaker s donor.Since 2- picoline has a similar pKa value (6.1) to that of 3-picoline, the similarity in the hydrolysis rates for Cl2 trans to NH3 and trans to 2-picoline in complex 1 can be attributed to steric hindrance provided by the 2-methyl group. In the crystal structure of complex 1, the 2-picoline ligand is almost perpendicular to the platinum square plane (102.78) and the methyl group is close to Pt (H3C? ? ? Pt 3.224 Å).5 Such steric hindrance is well known to slow down the rates of ligand substitution reactions in square-planar metal complexes. 20 For example, the rate constants (s21) for the hydrolysis of cis-[PtCl(L)(PEt3)2] complexes (the replacement of Cl2 by H2O) decrease in the order of (L): pyridine (800 × 1024) > 2-methylpyridine (2.0 × 1024) > 2,6-dimethylpyridine (0.01 × 1024).21 In the latter case the methyl groups block entry of incoming nucleophiles both above and below the square plane.The steric eVect is more prominent in the position cis to 2-picoline which is consistent with the CH3 group being further from the entering and leaving groups in the trigonal bipyramidal transition state if the pyridyl ligand is in the trigonal plane.21,22 The plane of the 3-picoline ring in complex 2 is tilted by 48.98 with respect to the platinum square plane, and the rates of substitution for 2 are determined mainly by the trans eVect.The rate of binding of the second 59-GMP ligand in the position cis to picoline is much slower for complex 1 compared with 2, which gives rise to a long-lived monofunctional adduct for the 2-picoline complex. This may explain the brief report that complex 1 forms DNA cross-links extremely slowly.7 There appear to be no comparable kinetic data available in the literature for reactions of 59-GMP with cisplatin. Compared to the rate constants for the second stage of binding of 59-GMP to [Pt(NH3)2(59-GMP)(H2O)]21 (298 K, 0.24 M21 s21),23 those for complexes 1 and 2 are an order of magnitude smaller. The rates of reactions of aquachloro cisplatin with DNA oligonucleotides 24,25 are about two orders of magnitude faster in comparison to those for reaction of the chloroaqua species of the picoline complexes 1 and 2 with 59-GMP.Rotation of ligands in bis(59-GMP) adducts Restricted rotation about the Pt–N7 bonds can potentially lead to three diVerent bis(GMP) stereoisomers for square-planar complexes having two cis ligands with C2 local symmetry: two head-to-tail (HT) and one head-to-head (HH) species, with the H8s of 59-GMP on the opposite side or on the same side of the platinum co-ordination plane, respectively. Because of the chiral ribose of 59-GMP, the H/T enantiomers become diastereomers and should be distinguishable by NMR.If the cis-PtA2 moiety lacks local C2 symmetry [e.g. two diVerent A (A, A9) or an unsymmetrical chelate], four diVerent stereoisomers (two HT and two HH) are possible when there is restricted rotation about the Pt–N7 bonds in bis(GMP) complexes. 26 Normally only HT conformations are expected to be thermodynamically favoured and detectable in solution, as is found in most solid-state crystallographic studies.27–29 Each H8 is non-equivalent for both HT stereoisomers,28,30,31 giving a total of four H8 signals for the HT stereoisomers.32 For the 2-picoline complex 1, the 2-methyl group can be on the upper side or the lower side of the platinum plane for each HT stereoisomer, so theoretically eight H8 signals should be observed for the two HT isomers together with four methyl signals (Scheme 2) when there is slow rotation (on the NMR timescale) about both the Pt–N7 (GMP) and Pt–N (picoline) bonds.Usually fast rotation of bound GMP about the Pt–N7 bond in cis-[PtA2(GMP)2] (where A2 are two unidentate or3584 J. Chem. Soc., Dalton Trans., 1998, 3577–3585 one bidentate amine ligand) is expected if the substituents on the two cis amine ligands are not very bulky.30,32 In our NMR experiments only four H8 signals in two sets of two singlets were observed for complex 1, together with two methyl signals.This suggests that rotation of 59-GMP about the Pt–N7 bond is fast on the NMR timescale but slow about the Pt–N bond for 2-picoline. This slowness of the rotation was evident from the exchange rate (0.62 s21, 296 K) determined from the 2D EXSY spectrum.The activation free energy (DG‡) measured from the coalescence temperature of the H8 signals represents the barrier for rotation about the Pt–N (2-picoline) bond. The coalescence of the two close cross-peaks comprising peak 1h in the 2D [1H, 15N] HSQC NMR spectrum of the bis(GMP) adduct of complex 1 at 65 8C can also be attributed to faster rotation about the Pt–N (2-picoline) bond at higher temperature. The broadening of one H8 signal and one CH3 signal at low temperature (<296 K) suggests that one of the bound GMP ligands (probably that cis to 2-picoline) in a HT stereoisomer is more aVected by the rotation of Pt–N (2-picoline) than the other.Compared with complex 1, the rotation of 3- picoline in complex 2 is facile, there being little steric hindrance from the methyl group in this case. Slow rotation about Pt–N (2-picoline) bonds has been reported previously for cis- [Pt(2-pic)2(Guo)2]21 at room temperature, whilst the rotation about Pt–N7 (Guo) was still fast on the NMR timescale.33 Hydrogen-bond interactions between ammine hydrogens and O(6) or the 59-phosphate of GMP can give rise to highfrequency 1H chemical shifts of Pt–NH3 groups in bis(GMP) adducts (1h and 2h). Platinum co-ordination to nucleotides is known to induce a change in the sugar-ring conformation from S-type (C29-endo) to N-type (C39-endo), which changes the H19–H29 coupling constant,34 consistent with the small decrease observed in the present work.The facile rotation of guanine about the Pt–N7 bond is thought to be important in the binding of cisplatin to DNA. From the results obtained here, several isomers would be expected when complex 1 reacts with G bases of DNA because of the steric eVect of 2-picoline and the non-C2-symmetrical structure, which is similar to cis-[PtCl2(NH3)(C6H11NH2)].35 These isomers may be stabilised by diVerent hydrogen-bonding patterns and have contrasting reactivities. Such studies may allow the further development of strategies for the systematic design of platinum antitumour complexes.Reactions with GSH The reactions of cisplatin and related platinum complexes with S-containing nucleophiles usually occur via direct substitution without prior aquation,36 as has been observed for substitution of Cl2 by methionine, GSH and metallothionein.37–39 However, in the reactions described here, aqua adducts were Scheme 2 Diastereomers of cis-[Pt(15NH3)(2-pic)(59-GMP-N7)2]21.The arrows (æÆ) represent 59-GMP molecules, with the head of the arrow denoting H8. the first species observed during reactions of both complexes 1 and 2 with GSH. There could be two reasons for this: (1) reactions of complexes 1 and 2 with GSH do proceed via hydrolysis, or (2) because of the steric eVect of the picoline ligands, binding of GSH is slowed and hydrolysis is fast enough to compete with direct substitution. Unfortunately the reactions were too complicated to allow a kinetic analysis to be carried out.The thiolate sulfur of GSH has a high trans eVect and this can lead to labilisation and release of the ammine ligand in the trans position.40 However, despite an overall loss of general signal intensity, no cross-peak for free 15NH4 1 was observed in the 2D spectra, probably because of fast proton exchange with H2O at pH 7. The yellow solids formed during reactions of thiols with platinum complexes are usually thought to be sulfur-bridged polymers.39–41 From the 2D [1H, 15N] HSQC NMR spectrum (Fig. 6), the initial binding site for GSH on complex 1 is in the position trans to 2-picoline, and attributable to the steric eVect in the cis position, which is similar to the situation for GMP binding. No rate constants were determined for the reactions of complexes 1 and 2 with GSH due to their complicated nature, but the time-dependent plots (Fig. 7) showed that complex 1 reacted much more slowly than complex 2.The half life (t1/2) of complex 1 was about three times longer than that of complex 2. Competitive binding of GSH and 59-GMP At neutral pH, GMP cannot usually compete with thiols for binding to PtII when both ligands are present at equal concentrations. 9 The kinetic preference of [Pt(dien)(H2O)]21 at neutral pH is exclusively toward thiols when in competition with 59-GMP.9 Complex 1 increases steric crowding in the square plane and circumvents resistance caused by reactions with glutathione and other cellular thiols, but still maintains the ability to form cytotoxic lesions with DNA.4 From the experimental results obtained above, the rate of binding of the thiol ligand GSH to complex 1 is greatly slowed but is still faster than that of 59-GMP.Substitution of Cl2 by the thiol ligand appeared to proceed via prior aquation because of the steric hindrance by the 2-methylpyridine ligand. The presence of hydrolysis products in the system provides a pathway for the formation of mono- and bis-GMP adducts which are formed via hydrolysis steps.The bis(GMP) adduct was stable in the presence of free GSH even after 3 d, but was present in only small amounts. Conclusion Labelling of the new anticancer drug cis-[PtCl2(15NH3)(2-pic)] has allowed detailed insight to be gained into the kinetics and mechanisms of its reactions with 59-GMP and glutathione. The steric eVect of 2-picoline reduces the reactivity of complex 1 towards 59-GMP compared to its 3-picoline analogue, especially towards substitution in the position cis to 2-picoline.The low reactivity of complex 1 towards 59-GMP may explain why it forms interstrand DNA cross-links much more slowly than cisplatin.2 Two HT isomers were observed for cis- [Pt(15NH3)(2-pic)(59-GMP-N7)2]21 due to the slow rotation of 2-picoline and non-C2-symmetrical structure. Reactions of GSH with complex 1 were ca. three times slower than those with 2, and appeared to proceed via aquated intermediates, with initial binding of GS2 trans to 2-picoline for 1 and trans to NH3 for 2.The bis(GMP) adduct of complex 1 was able to form in the presence of GSH at neutral pH in a competitive reaction. The steric eVect of 2-picoline and asymmetric structure of 1 may give rise to several isomers when it binds to DNA. This, together with its low reactivity towards GSH, may play an important role in its high activity against cisplatin-resistant cell lines.J. Chem.Soc., Dalton Trans., 1998, 3577–3585 3585 Acknowledgements This research was supported by the Biotechnology and Biological Sciences Research Council, Engineering and Physical Sciences Research Council (Biomolecular Sciences Programme), and EC COST programme. We are grateful to the Committee of Vice-Chancellors and Principals for an ORS Award, University of Edinburgh for a Research Studentship for Y. Chen and Johnson Matthey plc for the loan of some Pt.References 1 J. Reedijk, Chem. Commun., 1996, 801. 2 F. I. Raynaud, F. E. Boxall, P. M. Goddard, M. Valenti, M. Jones, B. A. Murrer, M. Abrams and L. R. Kelland, Clin. Cancer Res., 1997, 3, 2063. 3 J. Holford, S. Y. Sharp, B. A. Murrer, M. Abrams and L. R. Kelland, Br. J. Cancer, 1998, 77, 366. 4 J. Holford, F. Raynaud, B. A. Murrer, K. Grimaldi, J. A. Hartley, M. Abrams and L. R. Kelland, Anti-Cancer Drug Design, 1998, 13, 1. 5 Y. Chen, Z. J. Guo, S. Parsons and P. J.Sadler, Chem. Eur. J., 1998, 4, 672. 6 S. E. Miller and D. A. House, Inorg. Chim. Acta, 1989, 161, 131. 7 F. Raynaud, F. Boxall, P. Goddard, M. Valenti, M. Jones, B. Murrer, C. Giandomenico and L. Kelland, Proc. 88th Annual Meeting American Assoc. Cancer Res., San Diego, CA, 12–16 April 1997, vol. 38, no. 2085, p. 311. 8 Y. Zou, B. Van Houten and N. Farrell, Biochemistry, 1993, 32, 9632. 9 R. N. Bose, S. Moghaddas, E. L. Weaver and E. H. Cox, Inorg. Chem., 1995, 34, 5878. 10 J. Reedijk, in Handbook of Metal–Ligand Interactions in Biological Fluids, ed. G. Berthon, Marcel Dekker, New York, 1995, vol. 2, p. 967. 11 T. Ishikawa, C. D. Wright and H. Ishizuka, J. Biol. Chem., 1994, 29085. 12 S. J. S. Kerrison and P. J. Sadler, J. Chem. Soc., Chem. Commun., 1977, 861. 13 S. J. Barton, K. J. Barnham, A. Habtemariam, P. J. Sadler and R. E. Sue, Inorg. Chim. Acta, 1998, 273, 8. 14 M. Piotto, V. Saudek and V. Sklenar, J. Biomol. NMR, 1992, 2, 661. 15 J.Stonehouse, G. L. Shaw, J. Keeler and E. D. Laue, J. Magn. Reson., Ser. A, 1994, 107, 178. 16 E. W. Abel, T. P. J. Coston, K. G. Orrell, V. Sik and D. Stephenson, J. Magn. Reson., 1986, 70, 34. 17 S. J. Berners-Price and P. J. Sadler, Coord. Chem. Rev., 1996, 151, 1. 18 H. Friebolin, Basic One- and Two-Dimensional NMR Spectroscopy, VCH, Weinheim, 1991. 19 G. Pettit and L. D. Pettit, IUPAC Stability Constants Database, IUPAC and Academic Software, Otley, 1993. 20 F. Basolo, J. Chatt, H. B. Gray, R. G. Pearson and B. L. Shaw, J. Chem. Soc., 1961, 2207. 21 D. F. Shriver, P. W. Atkins and C. H. Langford, Inorganic Chemistry, 2nd edn., Oxford University Press, 1994, p. 627. 22 R. Romeo, D. Minniti and M. Trozzi, Inorg. Chem., 1976, 15, 1134. 23 S. S. Eapen, M. Green and I. M. Ismail, J. Inorg. Biochem., 1985, 24, 233. 24 K. J. Barnham, S. J. Berners-Price, T. A. Frenkiel, U. Frey and P. J. Sadler, Angew. Chem., 1995, 34, 1874. 25 S. J. Berners-Price, K. J. Barnham, U. Frey and P. J. Sadler, Chem. Eur. J., 1996, 2, 1283. 26 D. Kiser, F. P. Intini, Y. Xu, G. Natile and L. G. Marzilli, Inorg. Chem., 1994, 33, 4149. 27 B. Lippert, Prog. Inorg. Chem., 1989, 37, 1. 28 M. D. Reily and L. G. Marzilli, J. Am. Chem. Soc., 1986, 108, 6785. 29 Y. Xu, G. Natile, F. P. Intini and L. G. Marzilli, J. Am. Chem. Soc., 1990, 112, 8177. 30 R. E. Cramer and P. L. Dahlstrom, Inorg. Chem., 1985, 24, 3420. 31 S. J. Berners-Price, U. Frey, J. D. Ranford and P. J. Sadler, J. Am. Chem. Soc., 1993, 115, 8649. 32 K. Inagaki, F. J. Dijt, E. L. M. Lempers and J. Reedijk, Inorg. Chem., 1988, 27, 382. 33 A. T. M. Marcelis, J. L. Van Der Veer, J. C. M. Zwetsoot and J. Reedijk, Inorg. Chim. Acta, 1983, 78, 195. 34 K. Okamoto, V. Behnam, M. T. Phan Viet, M. Polissiou, J. Y. Gauthier, S. Hanessian and T. Theophanides, Inorg. Chim. Acta, 1986, 123, L3. 35 J. F. Hartwig and S. J. Lippard, J. Am. Chem. Soc., 1992, 114, 5646. 36 A. J. Repta and D. F. Long, in Cisplatin Current Status and New Developments, eds. A. W. Prestayko, S. T. Crooke and S. K. Carter, Academic Press, New York, 1980, p. 285. 37 M. I. Djuran, E. L. M. Lempers and J. Reedijk, Inorg. Chem., 1991, 30, 2648. 38 B. J. Corden, Inorg. Chim. Acta, 1987, 137, 125. 39 T. G. Appleton, J. W. Connor, J. R. Hall and P. D. Prenzler, Inorg. Chem., 1989, 28, 2030. 40 K. J. Barnham, M. I. Djuran, P. del S. Murdoch, J. D. Ranford and P. J. Sadler, Inorg. Chem., 1996, 35, 1065. 41 B. Odenheimer and W. Wolf, Inorg. Chim. Acta, 1982, 66, L41. Paper 8/06544F

ISSN:1477-9226    

DOI:10.1039/a806544f

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3587–3600 3587 Oxovanadium(IV) complexes of the dipeptides glycyl-L-aspartic acid, L-aspartylglycine and related ligands; a spectroscopic and potentiometric study João Costa Pessoa,*a Tamás Gajda,b Robert D. Gillard,*c Tamás Kiss,*b Susana M. Luz,a José J. G. Moura,d Isabel Tomaz,a João P. Telo a and Ibolya Török e a Instituto Superior Técnico, Departamento de Engenharia Química, Av. Rovisco Pais, 1096 Lisboa, Portugal. E-mail: pcjpessoa@alfa.ist.utl.pt b Attila József University, Department of Inorganic and Analytical Chemistry, H-6701 Szeged, POB 440, Hungary c Department of Chemistry, University of Wales, POB 912, Cardiff, UK CF1 3TB d Centro de Química Fina, Universidade Nova de Lisboa, Quinta da Torre, 2825 Monte da Caparica, Portugal e Research Group for Biocoordination Chemistry of the Hungarian Academy of Sciences, Attila József University, Department of Inorganic and Analytical Chemistry, H-6701 Szeged, POB 440, Hungary Received 9th March 1998, Accepted 4th September 1998 The equilibria in the systems VO21 1 L (L = Gly-L-Asp, L-Asp-Gly, N-acetyl-L-aspartic acid or succinic acid) have been studied at 25 8C and 0.2 mol dm3 K(Cl) medium by a combination of potentiometric and spectroscopic methods (ESR, circular dichroism and visible absorption). Formation constants were calculated from pH-metric data with total oxovanadium(IV) concentrations of (0.6–4) × 1023 mol dm23 and ligand-to-metal (L : M) ratios of 2–8 (AspGly) or 4–15 : 1 (other systems).The position of the Asp residue in the peptide chain aVects the co-ordination mode of the ligands: while in the GlyAsp system bis complexes start to form at pH less than 2, for AspGly only 1 : 1 complexes form, with relatively high CD signal. The co-ordination behaviour of N-acetyl-L-aspartic and succinic acids is similar. The results of potentiometric and spectroscopic methods are self consistent.Isomeric structures are discussed for each stoichiometry proposed and the results compared with those for L-aspartic acid and dipeptides with non-coordinating side chains. Introduction To model potential binding sites for the oxovanadium(IV) cation, complexation by several a-amino acids and simple peptides has been investigated.1–21 For dipeptides containing Gly and/or Ala, at VO21 concentrations of ª1022 mol dm23, the hydroxide precipitates at pH ª 4–5, even when using high amino acid : metal (L :M) ratios, e.g. 150–180: 1, and at pH > 7.5–8 the hydroxide slowly dissolves to give brown solutions, indicating that oxovanadium(IV) is extensively hydrolysed. The present study of the systems VO21 1 L (L = Gly-L-Asp, L-Asp-Gly, N-acetyl-L-aspartic acid and succinic acid) combines the results of potentiometric and spectroscopic techniques (ESR, circular dichroism and visible absorption). These ligands, as compared with simple oligopeptides, are expected to be more eYcient VO21 binders, due to their extra carboxylate group. Part of this work has been presented in a preliminary form.22 Precipitation of VO(OH)2 is clearly not so important in these systems, as was also the case for L-aspartic acid,5 and much lower L :M ratios may be used.If lower oxovanadium(IV) concentrations are used e.g. ª (1–4) × 1023 mol dm23, results of pH-potentiometric titrations with L :M ratios of 1 to 15 : 1 up to pH ª 5.0–6.5 (depending on the ligand) may be used to calculate formation constants.Potentiometry must be ruled out for pH > ª8; pKa(NH3 1) ª 8.4 and 8.0 for Gly-L-Asp and L-Asp- Gly, respectively, limiting the use of samples with high L :M ratios in the pH range 8–9, and oxovanadium(IV) is extensively hydrolysed, particularly at pH > 10. Further, the very high absorbance values (especially for 450 < l < 650 nm) of the [(VO)n(OH)m] species present also preclude the use of visible spectra. Minor oxidation of VIV may also aVect spectral measurements. The ESR spectra may give important information about the groups co-ordinated to oxovanadium(IV).23–26 The CD and ESR spectra for solutions containing L-Ala-Gly, Gly-L-Ala and L-Ala-L-Ala with high L :M ratios gave clear evidence of peptidic Namide co-ordination,16 suggesting the formation of 1 (Y = H2O or OH2) as the important species contributing to these spectra in the pH range 7.5–8.5.The crystal-structure characterisation of [NEt4][VVO(O2)(Gly-Gly)]?1.58 H2O,27 of [VIVO(Gly-Tyr)(phen)] 28 [ Gly-Tyr = glycyl-L-tyrosinate(22), phen = 1,10-phenanthroline] and of [NHEt3][VIVO(mpg)- (phen)] 29 [H3mpg = N-(2-sulfanylpropionyl)glycine], where NH2, Namide and CO2 2 are equatorial, and the isolation of [VIVO(Gly-Gly)(phen)]?2CH3OH and [VIVO(Gly-Ala)(phen)]? CH3OH,29 and their characterisation by continuous wave ESR and 14N electron spin echo envelope modulation also indicate that the dipeptides Gly-Gly and Gly-Ala are bonded as in 1.Circular dichroism (CD) spectra are more informative than the corresponding isotropic absorption spectra 30–33 for systems containing optically active amino acids and peptides. Only the vanadium–peptide complexes contribute to DA values (DA = diVerential absorption) in CD spectra, and the sign patterns N V Y NH2 O Y O C O COO V W W W O R 1 23588 J. Chem. Soc., Dalton Trans., 1998, 3587–3600 of Cotton eVects can be compared for bands I, II (and III) of the diVerent oxovanadium(IV) peptide and/or amino acid systems 23 as pH is varied.Since three pKa values (terminal CO2H, b-CO2H of the aspartic side chain and terminal NH3 1) can be determined for these dipeptides (H3L1), and two for the carboxylic groups of N-acetyl-L-aspartic acid and succinic acid (H2L9), the formation constants correspond to the general reaction (1). pM21 1 qL22 1 rH1 MpLqHr (1) We abbreviate Gly-L-Asp as GlyAsp, L-Asp-Gly as AspGly and N-acetyl-L-aspartic acid as NacAsp.For the formulations (VO)p(ligand)qHr 2p 2 2q 1 r we normally use the abbreviation MpLqHr in the case of GlyAsp, AspGly and L-Asp, and MpL9qHr in the case of NacAsp, and other H2L9 ligands (e.g. alanine, succinic acid). For each MpLqHr or MpL9qHr proposed in our speciation models possible structures are discussed. Basicity corrected formation constants are used when needed to help in this discussion as in previous publications.4,5,7,9 Experimental All solutions were prepared and manipulated in an inert atmosphere (high purity dinitrogen or purified argon).Amino acids (from Sigma) were dried for several days (in a desiccator with silica gel in vacuo), and succinic acid was a Fluka product of puriss p.a. quality for pH-metric studies and Merck p.a. (GR) for spectroscopic measurements. The KOH and HCl solutions used were Reanal products of the highest purity. KOH was standardized with potassium hydrogen phthalate and HCl with KOH potentiometrically.The purity of peptides was checked as described below and the exact concentrations of solutions were determined by the Gran method.34 The experimental diYculties in the present systems were discussed in refs. 1 and 35. Stock solutions of VO21 were prepared and standardised as described in ref. 1. The studies have been performed in 0.2 mol dm23 K(Cl) ionic medium. For potentiometric measurements the temperature was 25.0 ± 0.1 8C, and for CD and VIS spectra 25.0 ± 0.3 8C.TLC Thin-layer chromatography was performed on Merck plates (Art. 5626, 10 × 20 cm). The compounds and solutions used for spectroscopic measurements were monitored throughout the whole pH range to check their purity and test for reactions (e.g. hydrolysis): neither contamination nor decomposition was detected. Usually 2 ml samples were applied to the plates and the eluent was butanol–ethanol–propionic acid–water (10:10:2:5). The chromatogram was developed with a ninhydrin–collidine (2,4,6-trimethylpyridine)–copper solution prepared according to MoVat and Lytle.36 In some cases, after such development, the plate was placed in an enclosed chamber for development with iodine vapour.Typically, samples were taken for TLC after dissolution of the ligand, and addition of VO21 at several pH values. pH Measurements Spectroscopic measurements. For preparation of the solutions and pH calibrations we used a special glass vessel with a double wall, with entries for the glass electrode (Orion Ross 81-01) and reference electrode (Orion Ross 80-05), thermometer, nitrogen and reagents (e.g.base). A computerised system developed locally (for an IBM-PCXT 286 computer) was used to control the titration conditions for pH calibrations. The emf measurements were made with a Crison 517 pH meter. Potentiometric titrations. Stability constants were determined by pH-metric titration of 10.0 cm3 samples. The ligand concentration was 0.004 and 0.008 mol dm23 and ligand-to metal ion molar ratios: 2 : 1, 4 : 1, 6 : 1 and 8 : 1 (AspGly), and 2 : 1, 4 : 1, 6 : 1, 10 : 1 and 15 : 1 (other ligands).Titrations were performed from pH 2.0 until precipitation, very extensive hydrolysis or slow equilibration: these problems occurred in the pH range 5.0–6.7, depending on ligand and L :M ratio (see Table 1). Titrations were with KOH solution of known concentration (ca. 0.2 mol dm23) under a purified argon atmosphere.In some cases pH equilibrium could not be reached within 10 min due either to precipitation or very slow complex formation. Those titration points were omitted. The reproducibility of the included points was within 0.005 pH unit over the whole pH range. The pH was measured with an Orion 710A precision digital pH meter equipped with an Orion Ross 8103BN type combined glass electrode, calibrated for hydrogen ion concentration as described earlier.37 The ionic product of water was pKw = 13.76.The concentration stability constants bpqr = [MpLqHr]/[M]p[L]q[H]r were calculated with the aid of the PSEQUAD computer program.38 The formation of the hydroxo complexes of VO21 was taken into account. The following species were assumed: [VO(OH)]1 (log b1–1 = –5.94), [{VO- (OH)}2]21 (log b2–2 = 26.95), with stability constants calculated from the data of Henry et al. 39 and corrected for the diVerent ionic strength using the Davis equation. Spectroscopic measurements The CD spectra were recorded with a JASCO 720 spectropolarimeter with a red-sensitive photomultiplier (EXWL-308), visible spectra with a Perkin-Elmer lambda 9 spectrophotometer. Unless otherwise stated, by visible (VIS) and circular dichroism (CD) spectra we mean a representation of em or Dem values vs.l [em = absorption/bCVO and Dem = diVerential absorption/bCVO where b = optical path and CVO = total oxovanadium( IV) concentration]. The spectral range covered was usually 400–900 (VIS) and 400–1000 nm (CD).The ESR spectra were usually recorded at 77 K with a Bruker ESR-ER 200D X-band spectrometer. The CD, VIS and ESR spectra for GlyAsp, NacAsp and succinic acid systems were recorded by varying the pH with approximately fixed total vanadium and ligand concentration, at L :M ratios of 15 and 30; for AspGly this was done for solutions with L :M = 9.8 : 1. Several CD and ESR spectra were also recorded at fixed pH at varying L :M ratios by addition of VO21 stock solution.the GlyAsp solutions at pH 6.1 (L :M = 35.0, 24.7, 14.5, 9.4 : 1) as well as at 7.5 (35.0, 15.0, 9.4 : 1), 7.3 (20.1 :1), and AspGly solutions at pH 4.9 (L:M = 10, 7.1, 6.3, 5.6, 4.6 : 1) as well as at 6.8 (L :M = 7.1:1) and 2.6 (L :M = 4.6 : 1) were used for these measurements. Results and discussion Protonation and formation constants calculated for the systems studied 38 are in Table 1. The protonation constants agree well with earlier results: 40–43 the presence of the N-acetyl group in NacAsp increases the acidity of the CO2H groups relative to succinic acid.These ligands all contain two carboxylate binding sites and VO21 has a strong aYnity for oxygen containing ligands,23 so their complexes may diVer from those of simple dipeptides such as GlyGly, AlaGly, GlyAla or AlaAla.16 Further, relatively low L:M ratios (e.g. 10 : 1) may be adequate to avoid precipitation of VO(OH)2 in the case of GlyAsp and AspGly.Complexation starts through monodentate carboxylate co-ordination (as in 2): MLH2 for GlyAsp and AspGly (one proton belongs to the terminal NH3 1, the other to the non-co-ordinated CO2H group) and ML9H for NacAsp and succinic acid (the proton belongs to the non-co-ordinated CO2H group). With high excesses of ligand, the bis complexes M(LH2)2 or M(L9H)2 may also be formed. However, their formation can hardly be detected by pH-metry due to the overlap between the pro-J.Chem. Soc., Dalton Trans., 1998, 3587–3600 3589 Table 1 Formation constants (log values a) of species formed in VO21–ligand systems at T = 298 K and I = 0.2 mol dm23 KCl MpLqHr HL2 H2L H3L1 pKa1 pKa2 pKa3 MLH2 21 MLH1b ML MLH21 2 ML2H3 1b ML2H2 ML2H2 pH range studied L:VO ratio GlyAsp 8.36(1) 12.58(2) 15.24(3) 2.66 4.22 8.36 15.1(5) 11.52(8) 1.7(1) 26.6(2) 22.5(3) 17.1(2) 2.0–5.5 4–15 AspGly 7.93(1) 11.50(2) 14.32(3) 2.82 3.57 7.93 13.4(2) 10.46(3) 6.42(2) 0.83(4) 2.0–5.5 2–8 MpL9qHr HL92 H2L9 pKa1 pKa2 ML9H1 ML9 b ML9H22 22 ML92H2b ML92 22 NacAsp 4.52(2) 7.61(5) 3.09 4.52 6.2(2) 2.7(2) 27.32(5) 9.56(6) 5.0(3) 2.0–6.2 4–15 Succinic acid 5.19(1) 9.17(3) 3.98 5.19 7.2(2) 3.20(4) 27.25(3) 10.76(4) 5.6(1) 2.0–6.7 4–15 a Formation constants correspond to bpqr = [MpLqHr]/[M]p[L]q[H]r where L is the ligand in its deprotonated form (L22).Oxovanadium(IV) hydrolysis products are VO(OH)1 and [(VO)2(OH)2]21 with log b10 2 1 = 25.94 and log b20 2 2 = 26.95.The numbers in parentheses apply to the last digit included; it defines the range of log b values in refinements with PSEQUAD for the several plausible models obtained. b Formation constants for stoichiometries MLH (or ML9) and ML2H3 (or ML92H) could not be refined simultaneously (see text). cesses of co-ordination and deprotonation of the carboxylate group and to the high buVering eVect of large excesses of ligand. Accordingly, most of the resultant pH change belongs to ligand deprotonations and only a relatively small part to VO21 complexation.It is also worth mentioning that b111 and b123 could not be refined simultaneously; if both were included in the same calculation; one was always rejected, even if only data obtained at high excesses of ligand were included in the calculation. Probably ML2H3 does not exist in significant concentration at low L:M and the same applies to MLH at high L :M. For the corresponding stoichiometries ML9 and ML29H for the NacAsp and succinic acid systems, it was similarly not possible to refine b110 and b121 simultaneously.This is because, in the pH range 3–4.5 in which bis complexes are mostly formed, the ligand has already lost a proton from its more acidic carboxylic function, hence the complex formation involves no H1, eqns. (2A) and (2B). Under the conditions used for pHVOLH1 1 H2L VOL2H3 1 (2A) VOL9 1 HL92 VOL92H2 (2B) potentiometry (except for AspGly) the formation of 2 : 1 species is assumed above pH ª 2–3; when using high L:M, no polymeric complexes are expected till pH ª5.Fig. 1 gives calculated 44 distributions of concentration. The X-band ESR spectra of frozen ’solutions’ may be simulated as axial spectra. The field region corresponding to A|| and MI = 5/2 and 7/2 gives more information about the type and number of species. Fig. 2 shows ESR spectra in this range. When the pH is increased from ª1 to ª3.2 (AspGly), to ª3.5 (GlyAsp), to ª5.0 (NacAsp) or to ª 6.5 (succinic acid, not seen), the peaks shift slightly to lower field, so more than one species contribute to each spectrum.For higher pH, distinct species are detected, but for pH > 6–7 (for NacAsp and succinic acid) or > 8 (for GlyAsp or AspGly) the signal weakens significantly, increasing again at pH > 12, due to the formation of VO(OH)3 2. For VO21 and succinic acid with L :M = 30 : 1 and CVO ª 0.008 mol dm23, a new component appears from pH > ª6.7; it may correspond to ML9H23 (or indeed another stoichiometry, e.g.ML92H22/23). At pH 7.3 the components ascribed to ML9H22 and ML9H23 have relative intensities approximately 3 : 2. At pH 7.9 most VO21 is precipitated and the ESR signal is weak. Table 2 summarizes the spin Hamiltonian parameters obtained by simulating the whole spectrum using program EPRPOW,45 ascribing the ESR-active components to specific stoichiometries. Superscripts ’exptl’ and ’est’ refer to ’experimental’ and estimated parameters [eqn.(3)] where A||,i A|| est = S 4 i = 1 A||,i/4 (3) are the contributions to A|| est of each of the four equatorial groups (most presented by Chasteen 24 with estimated accuracy ±3 × 1024 cm21). Fig. 3 includes visible spectra for (A) GlyAsp 1 VO21 with L:M ratio 30 : 1 and (B) AspGly 1 VO21 with L:M ratio 9.8 :1. For pH < 2.5, they resemble (except for em values) oxovanadium(IV) solutions. Visible spectra for NacAsp or succinic acid 1 VO21 are similar and show the same trend as VIVO–GlyAsp.For pH > 2.5, as pH increases, band II gradually separates from band I revealing a progressive increase of the ligand field around VO21. However, while spectra for solutions of GlyAsp, NacAsp or succinic acid 1 VO21 are similar and their changes are like those found for dipeptides with non-coordinating side chains, the spectra for AspGly 1 VO21 diVer: band II shifts ª30–40 nm to the UV and its em values are about the same as those for band I.This means that the type of complexes formed and their co-ordination geometries diVer in the system with AspGly. In the pH range 1–4.5 the VIS spectra for solutions of AspGly resemble those for L-Asp.5 The CD spectra for the GlyAsp, AspGly and NacAsp systems modify as the pH is increased. In the pH range 1.5–5, all diVer from those for L-Asp.5 Fig. 4 includes spectra for (A) GlyAsp 1 VO21 with L :M 15 : 1 and (B) AspGly 1 VO21 with L:M 9.8. In the pH range 1–4, CD spectra of NacAsp 1 VO21 with L :M 30 : 1 are similar and show the same trend as those of GlyAsp: maximum values of Dem for band I are found in the pH range 4.1–4.6.The decrease in intensity of this band for pH > 4.8 is apparently due to the formation of ML92 and is more significant when ML9H22 forms [Fig. (1C)]. For pH > 4.5 the CD spectra for GlyAsp, AspGly and NacAsp diVer (Fig. 5) and also modify as pH is increased. For pH > 10 optical activity is low and becomes almost zero for pH > 11. The pattern of the CD bands and approximate lmax values ascribed to each stoichiometry and comparison with species distribution is in Table 2.The CD (e.g. Fig. 5) and ESR (Fig. 2) spectra of VO21 1 GlyAsp in the pH range 8–9.5 have the same profile and very similar spin-Hamiltonian parameters as those for simple dipeptides such as GlyAla, AlaAla, etc.16 As with them, this GlyAsp complex probably involves Namide equatorial co-ordination.While for GlyAsp with L:M 30 : 1,3590 J. Chem. Soc., Dalton Trans., 1998, 3587–3600 Fig. 1 Concentration distribution of the complexes formed in the (A) GlyAsp–, (B) AspGly–, (C) NacAsp– and (D) succinic acid–VO21 systems in solutions with CVO = 0.008 mol dm23 and L :M = 10 (AspGly) or 30 : 1 (other systems), calculated44 using bpqr of the equilibrium models of Table 1. The dashed lines indicate that the curves only represent an approximate estimation of the relative concentrations. The possible relative proportion of [(VO)m(OH)n] hydrolysis products is estimated assuming the formation of M2H25 with b2–5 = 10221.8.1,23 For MLH22 (GlyAsp) and ML9H23 (succinic acid) no b values are available, and concentrations were roughly estimated from the ESR spectra.maximum |Dem| values are for pH ª9.5, for AlaAla and GlyAla these are at pH ª7.5. The CD and ESR spectra have also been recorded at fixed pH for varying but high L :M ratios. The profile for the grey solution containing GlyAsp 1 VO21 at pH 6.1 with L :M = 35:1 and CVO = 6.0 × 1023 mol dm23 is very similar to spectrum 2 in Fig. 5(A), but the |Dem| are ª1.5 times larger. The corresponding ESR spectrum is like that in Fig. 2(A) (pH 5.97). On adding VO21 stock solution till L :M = 9.4 : 1 the solution becomes greenish (L :M = 24.7 : 1) and dark green (L :M = 14.5 or 9.4 : 1); the CD and ESR spectral profiles remain the same but the ESR signal increases and values of |Dem| decrease: at L :M = 9.4 :1 to ª20% of those for L:M = 35 : 1. These significant changes in Dem as CVO is varied can be explained only by equilibria between species with diVerent degrees of polymerisation.At pH 6.1 these could be as in Scheme 1. As long as the L :M ratios are high (e.g. > 12 : 1) the relative concentrations of monomeric complexes vary little with CVO but the % of vanadium in the form of monomeric vs. oligomeric species varies significantly. Since the pattern of the CD spectra changes little with CVO (and ill defined isodichroic points are observed at l ª 560 and 790 nm with Dem ª 0), the oligomeric species present at pH 6.1 must be optically inactive.Other Scheme 1 Equilibria and main species involved at pH ª 6.1 in the GlyAsp system. amino acid systems behave similarly.1–6 In similar experiments at pH 7.5, when extra VO21 was added to a sample of L :M = 35:1 (light yellowish brown) to change L :M to ª14.5 : 1 (pH 7.50, dark green) the CD profile changed, in particular, the Cotton eVect associated with the band at 700 nm became negative and the band at ca. 800 nm shifted to the red. Overall |Dem| decrease to ª30% of those for L :M = 35 : 1. On addition of more VO21 till L :M = 9.4 : 1 (pH 7.50, very dark green), the CD spectrum becomes very noisy but now the profile apparently changes little. The ESR profile for these experiments at pH 7.5 is always the same: the dominant components are those designated by MLH21 and/or MLH22. Therefore at this pH the changes in CD profile and Dem with CVO can be explained only by assuming equilibria between monomeric complexes (MLH21 and MLH22), optically and ESR active, and oligomeric species (e.g.M2L2H23, M2L2H24) which are optically active but ESR-inactive. Optically inactive oligomeric species, e.g. [(VO)n- (OH)m], probably also have significant concentrations at pH 7.5. Scheme 2 summarizes the processes expected to occur as pH is increased. For GlyAsp 1 VO21 the distribution of Fig. 1(A) describes the ESR spectra of Fig. 2(A) reasonably well till at least pH ª6. Scheme 2 Overall description of processes as pH is increased in the range 6.5–13 for the GlyAsp system. The optically active oligomeric species is expected to form in the pH range 6.5–10.J. Chem. Soc., Dalton Trans., 1998, 3587–3600 3591 Fig. 2 High field range (3800–4400 G) of the first derivative ESR spectra at 77 K of frozen “solutions” containing (A) GlyAsp and VO21 with L:M = 30.0 : 1 and CVO ª 0.008–0.011 mol dm23, (B) AspGly and VO21 with L :M = 9.8 : 1 and CVO ª 0.006–0.010 mol dm23 and (C) NacAsp and VO21 with L :M = 30.0 : 1 and CVO ª 0.008–0.014 mol dm23. pH Values, colours of the corresponding solutions and stoichiometries corresponding to the ESR components are indicated.For pH > 6–6.5 at least two diVerent ESR-inactive oligomers form, as well as an ESR and CD active complex: possibly MLH22. Some results for it are in Table 2.No value of b is available and its concentration in Fig. 1(A) was roughly estimated from the ESR results. For AspGly 1 VO21, till at least pH 6 the distribution of Fig. 1(B) also describes reasonably well the relative intensities of the ESR-active species in Fig. 2(B). The CD signal is weak and noisy under conditions corresponding to these ESR spectra for pH < 2.7. At pH 1.2 VO(OH2)5 21 predominates and for pH < 1.7 the dominant optically active species is MLH2, with negative De, as for amino acid complexes containing a mono-co-ordinated carboxylate function.1–8,16 Since the CO2 2 group is now from an achiral glycine residue the CD signal is very weak.The structure of this complex probably corresponds to 3 (Table 3) so the vicinal eVect is transmitted mainly through the axial (NH)C]] Oamide, not as eYcient as an equatorial CO2 2 group.16 For pH > ª2.0 Dem becomes positive due to the formation of MLH. The CD profile in the pH range 2.7–4.1 is approximately the same so the relevant spatial factors that determine the CD signal are similar for MLH and ML.Although the VIS spectra of solutions containing AspGly and VO21 for pH > 3 suggest increased ligand field strength, no such increase occurs for GlyAsp or NacAsp. This suggests that in ML the NH2 group is also co-ordinated (e.g. 12 and 13 in Table 3). For AspGly and VO21 with L:M = 9.8 :1, at pH 6 [Fig. 5(B)], the overall CD signals are intense with the band pattern: 1,1,2 for bands II, IB, IA and: Dem band II ª 2Dem band IB @ Dem band IA.As pH is increased in the range 6–8 the pattern changes to 1,1,1, bands II and IA now being about equal. These modifications indicate that besides MLH21 a new optically active species forms. In the same pH range the ESR signal decreases: in the range 8–10.5 oligomeric ESR inactive species {e.g. [(VO)n(OH)m]} form, causing further decrease. The spin-Hamiltonian parameters for the ESR-active species correspond to MLH21 (Table 2).For AspGly 1 CuII, bis complexes formed, presumably involving amino acid-like co-ordination.41 Our pH-metric results (till L :M = 8 : 1) suggest no formation of bis complexes till at least pH 5.2. The CD decreases for pH > 5 which indicates no significant formation of bis complexes at higher pH. Therefore, the ESR and CD active complex that forms for pH > 5–6 probably corresponds to a stoichiometry MLH22. For NacAsp 1 VO21 it is not straightforward to compare the distribution of Fig. 1(C) with the ESR [Fig. 2(C)] because ML9H, ML92H and ML92 cannot be detected separately. However, the agreement is reasonable: e.g. at pH 5.78 the distribution predicts ª55% of ML92 and ª37% of ML9H22, and this is in good accord with the intensities of the ESR components. The CD and VIS profiles for the solutions containing NacAsp corresponding to the ESR spectra of Fig. 2(C) at pH < 3.5 are very similar to those for GlyAsp under similar conditions.The main contributors to CD are ML9H and ML92H. These correspond to stoichiometries MLH2 and ML2H3 in the GlyAsp system. So, dominant factors that contribute to the optical activity must be similar for the corresponding complexes. For pH > 4 the CD spectra for GlyAsp and NacAsp diVer. For NacAsp 1 VO21 [Fig. 5(C)] the pattern of the CD is the same in the pH3592 J. Chem. Soc., Dalton Trans., 1998, 3587–3600 Table 2 Hyperfine coupling constants, g values, lmax of CD spectra (and corresponding signal) or of VIS spectra (and corresponding e value for succinic acid) for several stoichiometries of each of the systems studied.Most A and g values were obtained by simulation with the program EPRPOW45 of EPR spectra of frozen “solutions” at 77 K lmax (CD signals or e) a Stoichiometry A|| exptl × 104 g|| exptl A^ exptl × 104 g^ exptl band II band I b GlyAsp MLH2 21c ML2H3 1 ML2H2 c ML2H2c MLH21 2 MLH22 22c (?) ª177 174.0 ª172 ª171 160.7 162.5 ª1.936 1.938 ª1.940 ª1.942 1.953 1.951 ª65.0 63.0 ª61 ª60 53.5 53.7 ª1.978 1.977 ª1.975 ª1.977 1.981 1.985 590 (2) 600 (2) 550 (2) ?? (2) ª510 (2) 510 (2) 510 (2) 760 (2) 820 (2) 820 (2) ??? (1), ??? (2) d 730 (1) ? 700 (1),e 850 (2) e ? 730 (1) AspGly M21f MLH2 21c MLH1c ML MLH21 2 MLH22 22c (?) 181.2 ª177.0 173.5 169.8 165.0 163.5 1.934 ª1.937 1.939 1.944 1.950 1.949 68.2 ª65–66 62.8 60.8 57.5 54.0 1.978 ª1.977 1.977 1.977 1.979 1.980 ??? (2) 575 (1) 565 (1) 550 (1) 525 (1) ??? (2) d 750 (1) 730 (1) 700 (1), 860 (2) 735 (+). ª880 (1) NacAsp M21g ML9H1c ML92H2 ML92 22c ML9H22 22c (?) MH23 2h 181.5 ª177 173.5 ª172 ª168 162 1.934 ª1.937 1.938 ª1.940 ª1.945 1.955 68.8 ª65.2 62.4 ª61.5 ª60.0 49.5 1.978 ª1.977 1.977 ª1.978 ª1.977 1.977 610 (2) 600 (2) 575 (1) 570 (1) 770 (2) 810 (2) 813 (2) 810 (2) Succinic acid i M 1 ML9H1 (pH 2.2) M 1 ML9H1 1 ML29H2 pH 3.0 ML29H2 ML92 22 ML9H22 22 ML9H23 32 (?) 179 178 174 172 172 171 ª165 1.934 1.938 1.940 1.942 1.942 1.942 ª1.944 69 67 63 61 61 61 ª62 1.978 1.980 1.981 pH 3.80 1.980 pH 5.75 ª1.980 pH 6.32 j ª1.980 pH 6.95 j ª1.982 ?? d ?? d ª625 (ª12) 605 (18.5) 565 (21) 565 (113 k) 768 (19.0) 773 (22.6) 780 (26.1) 800 (32.5) 830 (23.0) 830 (64 k) a Based on qualitative analysis of experimental CD spectra and distribution diagrams.For succinic acid the lmax/nm and em/dm3 mol21 cm21 presented are those for VIS spectra at pH values where each stoichiometry is expected to be largely predominating.b When splitting of band I is clearly observed, two lmax values are given: bands IB and IA. c The ESR spectrum was diYcult to simulate due to noise or presence of a significant amount of more than one species. d The lmax cannot be estimated satisfactorily. e This species could have this pattern or the same as for MLH22 (see text). f From ESR spectra of solutions containing AspGly and VO21 (L:M = 9.8 and CVO ª 0.010 mol dm23) at pH 1.0 and 1.2.g From ESR spectra of solutions containing NacAsp 1 VO21 (L:M = 30 and CVO ª 0.010 mol dm23) at pH 0.6 and 1.2. h From the ESR spectra of a solution containing NacAsp and VO21 at pH 12.9 (L :M = 30 and CVO ª 0.008 mol dm23). i Each spectrum seems to correspond to only one component, but all correspond to at least two species. Only parameters for ML92H, ML92, ML9H22 can be determined with reasonable accuracy. j The lmax and e for ML9H22 are only rough approximations as at pH 6.32 the contribution of oligomeric species e.g.[(VO)n(OH)m] is significant. k The em of solutions containing succinic acid and VO21 (L:M = 30 and CVO ª 0.008 mol dm23) at pH ª 7 decreases continuously from 350 (e ª 606) to 900 nm (em ª 61 dm3 mol21 cm21). Here em values are given at 565 and 830 nm, the lmax observed at pH 6.32. range 4.2–6.3, with lmax 550–580 (band II) and 805–815 nm (band I), see Table 2. The value of |Dem| increases till pH ª4.7 (band I) or 5.8 (band II).At pH > 3.5, for NacAsp Dem > 0 in the region of band II, but negative till pH ª6.7 for GlyAsp. Formation constants could be calculated from titrations up to pH 6.2 (Table 1) but the ESR of Fig. 2(C) and the CD of Fig. 5(C) reflect no new ESR or CD active complexes till pH ª 9.3. However, intensities decrease markedly for pH > 7 indicating the formation of [(VO)n(OH)m] inactive in ESR and CD. So Fig. 1(C) includes [{(VO)2(OH)5 2}m] with m = 1, in accord with previous practice,1–9 but in reality m is probably more than 1.The ESR and VIS spectra for VO21–succinic acid at high L:M as pH is varied resemble those for NacAsp but at pH > ª3.5 bands I and II are more distinctly separate. This system was previously studied by spectroscopic 18,46 and pHpotentiometric 43,46 methods at L :M ratios of 1 and 2 : 1, and by ESR47 using L :M 100 : 1. The formation of a single complex ML9 (with stability constant 103.66) was assumed at 30 8C and I = 0.1 mol dm23 KNO3.43 At low excess of ligand, in the pH range 4.5–6, spectroscopy (VIS and ESR) gave evidence only for the monodentate carboxylato complex 2, besides the aqua ion and VO21 hydroxo complexes.It was characterised by absorption maxima at 590–600 (band II) and 780–810 nm (band I).18,46 Our VIS spectra at high excess of ligand show features apparently not then observed 18,46 at low L:M ratios. Two bands are distinct in the range 450–900 nm for pH > 3.5, with lmax(band II) ª565 nm and lmax(band I) ª830 nm: this suggests complexes with bidentate succinate (or the monodentate co-ordination of three or more succinate ligands).The conclusions of Ferrari et al. 18 based on ESR and VIS spectra are similar, except that they assumed 2 to form at lower pH. In their ESR study with CVO = 1023 mol dm3 and L:M 100 : 1, McPhail and Goodman47 detected four components in the pH range 3–6, finding A to decrease and g to increase as pH is increased.Our ESR results agree. With succinic acid at high L :M the distribution of Fig. 1(D) describes the spectroscopic results up to pH 6.7. The new ESR-active species at pH ª7 could correspond to a ML9H23 stoichiometry: its estimated spin-Hamiltonian parameters are in Table 2. As at least two distinct components are detected in the ESR spectra; the parameters were obtained by simulation of spectra but using equations 48 based on Chasteen9s24 iterative method.Lacking a b value, the relative concentration was estimated from ESR.J. Chem. Soc., Dalton Trans., 1998, 3587–3600 3593 Since the vanadyl ion is lop-sided many of its chelates may give rise to more isomers than do those of apparently analogous monoatomic metal centres such as CuII refs. 1 and 23 briefly discuss this. For monodentate ligands, such as an amino acid bonded via its carboxylate, this extra isomerism would be absent. However, if there are two such ligands geometric isomers may be present.We now discuss, on the basis of Figs. 2–5 and Tables 1 and 2, the dominant structure in solution for each stoichiometry. For a given stoichiometry and similar co-ordination geometry, CD spectra for GlyAsp, AspGly and NacAsp may diVer. Table 3 gives plausible structures and, for each, the main mechanisms for inducing optical activity. These systems are labile. Optical activity consequently arises from the chiral arrangement of the chelate rings, i.e.the conformational effect,30 and, since the ligands contain asymmetric carbon atoms, from the transmission through space and by way of the chemical bonds linking the asymmetric centre to the chromophore, i.e. the vicinal effect.30 The order of magnitude of the conformational eVect depends on the chelate-ring puckering and that of the vicinal eVect on the number of atoms between the asymmetric centre and the metal, and on the eYciency of the donor groups in transmitting dissymmetry, which for peptides has been suggested to be: 16,31 Namide > (CO)Namide > CO2 2 > C]] Oamide > (NH)C]] Oamide > NH2.We also assume here and elsewhere 1–8,16 that optical activity transmits better from the chiral centres of the ligand into optical transitions through Fig. 3 Visible absorption spectra of solutions containing (A) GlyAsp and VO21 with L :M = 30.0 : 1 and CVO ª 0.008–0.011 mol dm23 and, (B) AspGly and VO21 with L :M = 9.8 : 1 and CVO ª 0.006–0.010 mol dm23.The pH corresponding to each spectrum is indicated. Spectrum 1 in B approximately coincides with that of VO(OH2)5 21 (]] ] M21). The ESR spectra corresponding to some of these solutions are in Fig. 2. equatorial rather than axial donor ligands. Asymmetric deviations of ligating atoms from regular polyhedra (asymmetric distortions) may also make a significant contribution to the CD spectrum,32,33 particularly in VO21 complexes. In these systems several optically active species may contribute at each pH, so extracting structural information from CD is not straightforward.In particular, the fact that in the pH range 3–8 |Dem|AspGly are much greater than |Dem|GlyAsp cannot be easily understood. One possibility is that the dominant complexes in the AspGly system (ML and MLH21) are more rigid, because of extra ‘chelation’ 49 (e.g. via hydrogenbonding of a solvent water between two donor groups or the dipeptide behaving as tetradentate with one donor group axial): the configurational eVect may then contribute to optical activity.Another explanation is a significant contribution to optical activity arising from inherent dissymmetry and asymmetric distortions for complexes ML and MLH21 in the AspGly system. These eVects may 32 be much stronger than the conformational or the vicinal eVects. Vanadium in VO21 may become an optically active centre, so most complexes here contain two dissymmetric centres. The structures presented correspond to one of two possible diastereomers: absolute configurations for oxovanadium, either C or A,50 and for the a-carbon, L (e.g.discussion later for structure 17a). Their stability may diVer as found51 by 51V NMR for e.g. [VVO(sal-L-aa)(dl2)] (sal-L-aa = N-salicylidene-L-amino acidate of Gly, Ala, Val or Phe; dl2 = monoanion of glycerol, ethane-1,2-diol or propane-1,3-diol), the trends observed being consistent with steric control. If this is the case for complexes ML and MLH21 in the AspGly system, the Fig. 4 Circular dichroism spectra of solutions containing (A) GlyAsp and VO21 with L :M = 15.0 : 1 and CVO ª 0.018–0.020 mol dm23 and, (B) AspGly and VO21 with L :M = 9.8 : 1 and CVO ª 0.008–0.010 mol dm23. The pH corresponding to each spectrum is indicated. The ESR and VIS spectra corresponding to some of the AspGly solutions are in Figs. 2 and 3.3594 J. Chem. Soc., Dalton Trans., 1998, 3587–3600 contributions of inherent dissymmetry, asymmetric distortions and finally that of the configurational eVect may become important.MLH2 and ML9H The stoichiometry MLH2 (GlyAsp, AspGly) or ML9H (NacAsp, succinate) corresponds to structures with equatorially co-ordinated carboxylate (the terminal one in the case of GlyAsp and AspGly). The pattern of the VIS and CD spectra of this stoichiometry resembles spectra 2 of Fig. 3(A) and 4(A), and those for the corresponding VO21 complexes of various amino acids 1–6 and simple dipeptides.9,16 For GlyAsp and NacAsp, co-ordination involves the CO2 2 group near the Fig. 5 Circular dichroism spectra of solutions containing: (A) GlyAsp and VO21 with L :M = 15.0 : 1 and CVO ª 0.016–0.018 mol dm23, (B) AspGly and VO21 with L :M = 9.8 : 1 and CVO ª 0.006–0.008 mol dm23 and, (C) NacAsp and VO21 with L :M = 30.0 : 1 and CVO ª 0.008–0.012 mol dm23. The pH corresponding to each spectrum is indicated. The ESR spectra corresponding to some of these solutions are in Fig. 2. asymmetric centre, the CD showing its characteristic pattern (Fig. 3 and Table 2): for AspGly it involves the Gly residue, the CD being therefore extremely weak at pH 1–2. Chasteen’s method 24 (eqn. 3) gives g|| est ª 1.953 and A|| est ª 180 × 1024 cm21 for this geometry. Consequently, the values in Table 2 agree with those obtained for solutions containing mostly MLH2 (and ML2H4) in the case of GlyAsp and AspGly, or ML9H (and ML29H2) in the case of NacAsp and succinic acid.VO21 1 H2L KCO2 2 VO(H2L)21 (4A) VO21 1 HL92 K9CO2 2 VO(HL9)1 (4B) MLH or ML9 and ML2H3 or ML92H In both MLH and ML9 the basic binding mode is probably chelation via both carboxylate groups. Derived equilibrium constants b1 1 and b1* for MLH (GlyAsp, AspGly and Asp) or ML9, and also for corresponding complexes of some reference ligands, are in the second row of Table 4. The data show that 7- membered chelate formation is less favoured, log b1* = 23.8, than the 6-membered dicarboxylate chelate in the malonic acid system (log b1* = 21.99).The appropriate derived constants log b1* for NacAsp and succinic acid are ª25 and 26, respectively, revealing important stabilisation for GlyAsp and AspGly complexes. This can very likely be explained by the co-ordination of Oamide of the peptides. For several plausible structures no strain is apparent and the amide group remains planar (e.g. 4–9). This is the case of 4 for GlyAsp and 5–7 for AspGly.For AspGly, structures 6 and 7 and others containing two “free” equatorial positions are discarded; in fact, if they corresponded to MLH there would be no obvious reason for the absence of ML2H3 (or dimeric complexes where the equatorial positions occupied by W and/or Y could be shared with donor groups from the ligand of a second complex). Therefore MLH probably corresponds to 5 although the tridentate equatorial co-ordination is not apparent in the values of b1 1 and b1*.For ML9 (NacAsp or succinic) the structures can correspond to 8 and/or 9. Structures for ML2H3 (GlyAsp) or ML29H (NacAsp or succinic) can be regarded as comprising those for MLH or ML9 plus a monodentate carboxylate group from a second ligand. Therefore, in the schematic structures here (e.g. 4, 8 or 9), X represents either H2O or a RCO2 2 ligand. The spin- Hamiltonian parameters for ML2H3 or ML29H [ g|| exptl ª 1.938 and A|| exptl ª (173–174) × 1024 cm21] indicate equatorial coordination of three carboxylate groups, or two carboxylate groups and one Oamide; therefore structure 4, 8 or 9 (and other isomers) is consistent with the ESR data.The pattern of CD expected for MLH (GlyAsp) and ML9 (NacAsp), and ML2H3 (GlyAsp) and ML29H, is the same; therefore, the dominant dissymmetric factors persist in corresponding stoichiometries in these systems, i.e. if, say, structure 4 is correct for GlyAsp, then for NacAsp the structure would correspond to 9. However it is not yet possible to decide which geometry (4, 8, 9 or some other) actually predominates.ML (AspGly) The pK values for the formation of ML stoichiometries are included in the fourth row of Table 4. While for NacAsp, succinic and malonic acids the pK values correspond to the deprotonation/co-ordination of a carboxylate group, for AspGly and L-Asp this corresponds to a similar process for the NH3 1 group. The equatorial co-ordination of the NH2 group of AspGly is indicated by the ESR spectra of Fig. 2(B), and by the spin-Hamiltonian parameters obtained for ML (Table 2).TheJ. Chem. Soc., Dalton Trans., 1998, 3587–3600 3595 Table 3 Structures, corresponding A|| est, g|| est [eqn. (3)],a stoichiometries and comment about expected origin for the optical activity for complexes that may form in solutions containing oxovanadium(IV) and GlyAsp, AspGly, NacAsp or succinic acid (see text) Schematic representation b A|| est × 104/cm21 ( g|| est) Stoichiometry (and comments) c a X = H2O 180 (1.935) b X = Ocarboxylate 177 (1.937) MLH2 (GlyAsp or AspGly) (optical activity induced by the vicinal eVect through the CO2 2 group) ML2H4 (GlyAsp or AspGly) a X = H2O 177 (1.937) b X = Ocarboxylate 174 (1.939) MLH (GlyAsp) (optical activity induced by the vicinal eVect through the CO2 2 groups, and by the conformational eVect) ML2H3 (GlyAsp) a Y = H2O 175 (1.939) d b Y = OH2 168 (1.946) d MLH (AspGly) (optical activity induced by the vicinal eVect throught the b-CO2 2 and C]] Oamide, and by the conformational eVect) MLH21 (AspGly) a Y = H2O 174 (1.939) d b Y = OH2 167 (1.946) d MLH (AspGly) [Optical activity induced by the vicinal eVect through the b-CO2 2 (low) and C]] Oamide (very low) and by the conformational eVect] MLH21 (AspGly) a Y = H2O 174 (1.939) b Y = OH2 167 (1.946) MLH (AspGly) [optical activity induced by the vicinal eVect through the b-CO2 2 (low) and C]] Oamide (very low), and by the conformational eVect] MLH21 (AspGly) a X = H2O 180 (1.935) b X = Ocarboxylate 174 (1.939) MLH (AspGly) or ML9 (NacAsp or succinic e) [optical activity induced by the vicinal eVect through the a-CO2 2 and b-CO2 2 (low) and by the conformational eVect] ML2H3 (AspGly) or ML29H (NacAsp or succinic e) a X = H2O 177 (1.937) b X = Ocarboxylate 170 (1.944) MLH (AspGly) or ML9 (NacAsp or succinic e) [optical activity induced by the vicinal eVect through the a-CO2 2 and b-CO2 2 (low) and by the conformational eVect] ML2H3 (GlyAsp) or ML29H (NacAsp or succinic e) a Y = H2O 172 (1.942) b Y = OH2 165 (1.950) ML2H (GlyAsp) [optical activity induced by the vicinal eVect through the NH2 (low), C]] Oamide (low) and axial CO2 2 (low), and by the conformational eVect] ML2 (GlyAsp) a Y = H2O 172 (1.942) b Y = OH2 165 (1.950) ML2H (GlyAsp) [optical activity induced by the vicinal eVect through the NH2 (low), C]] Oamide (low) and a-CO2 2 and by the conformational eVect] ML2 (GlyAsp)3596 J.Chem. Soc., Dalton Trans., 1998, 3587–3600 Table 3 (Contd.) Schematic representation b A|| est × 104/cm21 ( g|| est) Stoichiometry (and comments) c a Y = H2O 171 (1.942) b Y = OH2 164 (1.950) ML (AspGly) [optical activity induced by the vicinal eVect through the NH2 and b-CO2 2 (low) and by the conformational eVect] MLH21 (AspGly) a Y = H2O 171 (1.942) b Y = OH2 164 (1.950) ML (AspGly) [optical activity induced by the vicinal eVect through b-CO2 2, NH2 and C]] Oamide (low), and by the conformational, asymmetric distortion f and configurational eVects] MLH21 (AspGly) a Y = H2O 175 (1.940) b Y = OH2 168 (1.947) MLH21 (AspGly) (optical activity induced by the vicinal eVect through C]] Oamide and NH2, and by the conformational eVect) a Y = H2O ª163 g (ª1.950) g b Y = OH2 ª158 g (ª1.960) g MLH21 (AspGly) [optical activity induced by the vicinal eVect through (CO)Namide and NH2, by the conformational (?), asymmetric distortion f and configurational eVects] MLH22 (AspGly) a Y = H2O ª165 g (ª1.950) g b Y = OH2 ª158 g (ª1.957) g MLH21 (AspGly) [optical activity induced by the vicinal eVect through (CO)Namide and b-CO2 2, conformational, configurational and asymmetric distortion f eVects] MLH22 (AspGly) a Y = H2O ª162 g (ª1.953) g b Y = OH2 ª158 g (ª1.960) g MLH21 (GlyAsp) [optical activity induced by the vicinal eVect through Namide and a-CO2 2 and by the conformational (?), configurational and asymmetric distortion f eVects] MLH22 (GlyAsp) a Y = H2O ª162 g (ª1.953) g b Y = OH2 ª158 g (ª1.960) g MLH21 (GlyAsp) [optical activity induced by the vicinal eVect through Namide and b-CO2 2 (low), by the conformational eVect and asymmetric distortions f] MLH22 (AspGly) a The A|| est presented were calculated using eqn. (3) and the g|| est using an equivalent equation.In these estimates we assume that ligands co-ordinated in axial position have no influence on the spin-Hamiltonian parameters.The A|| and g|| donor group contributions presented by Chasteen 24 assume axial co-ordination of water. bThe glycine residue is normally indicated with a G; the CO2 2 group of the side chain of the aspartic residue with a b. c The optical activity induced by the vicinal eVect through NH2 or CO2 2 of glycine residues, or from groups co-ordinated in axial position, is expected to be low. d Assuming the contribution of Oamide in eqn. (3) is 174.7 × 1024 cm21 as presented by Pecoraro and co-workers.26 e No optical activity for the succinato complexes.f The eVect of asymmetric distortions (see text) may be important whenever there are distortions in the otherwise symmetric regular structures of complexes. As a whole these distortions must be dissymmetric. 1–6,32,33 g Assuming that the contributions of Namide to A|| and g|| are 136 × 1024 cm21 and 1.983, respectively.16,28,29 CD and VIS spectra for ML approximately correspond to spectra 7 of Figs. 4(B) and 3(B), respectively. Structure 12a or 13a could correspond to ML. For 12a the axial co-ordination of a water molecule is apparently precluded by the a-proton of the Asp residue. There is no significant strain in tetradentate co-ordination of AspGly and we propose 13a as the structure of the AspGly complex corresponding to ML stoichiometry. ML2H2 and ML29 The pK values for the formation of ML29 from ML29H or ML2H2 from ML2H3 are included in the fifth row of Table 4, being in the range 3.2–5.2. Comparing these values with the corresponding pKa2 of the ‘free’ ligands, no significant decrease in the pK of the carboxylic groups is observed in ML92H or ML2H3 complexes.The changes observed in the ESR, VIS andJ. Chem. Soc., Dalton Trans., 1998, 3587–3600 3597 Table 4 Derived equilibrium constants for VO21 complex formation, partial processes of the ligands studied and some related compounds Monodentate CO2 2 co-ordination a Bidentate 2O2C? ? ?CO2 2 co-ordination b Bidentate (2O2C? ? ?CO2 2)2 co-ordination c MLH ML 1 H1 or ML9H ML9 1H1 ML2H3 ML2H2 1 H1 or ML92H ML92 1 H1 ML2H2 ML2H 1 H1 or ML92 ML92H21 1 H1 ML MLH21 1 H1 or ML9 ML9H21 1 H1 log KCO2 2 log b1 1 log b1* log b2 1 log b2* pK pK pK pK GlyAsp 2.5 3.16 (7-m.r.) 23.72 5.8 (7-m.r.) 27.98 — 4.1 5.4 — AspGly 1.9 2.53 (7-m.r.) 23.86 — 4.0 — — 5.6 NacAsp 1.7 2.7 (7-m.r.) 24.9 5.0 (7-m.r.) 210.2 3.5 4.6 — — Succinic acid 2.0 3.20 (7-m.r.) 25.97 5.6 (7-m.r.) 212.7 4.0 5.2 — — GlyGly9 1.8 — — — — — pKamide ª7 Ala1 1.19 — 25.62 212.6 4.3 5.2 8.0 ª5.3 Gly10 1.17 — 25.60 212.4 4.3 4.8 7.7 ª5.2 Malonic acid35 1.18 5.6 (6-m.r.) 21.99 9.5 (6-m.r.) — ª0.6 — — 5.1 Aspartic acid5 ª1.9 2.8 (7-m.r.) 23.22 ª5.0 (7-m.r.) 27.0 ª3.5 ª3.2 ª3.5 6–7 AspGly (Cu)41 — 2.1 — — 3.5 — — pKamide 4.9 GlyAsp (Cu)40 — 2.1 25.04 — 3.8 — — pKamide 4.8 a Ligand is protonated except at the carboxylate which co-ordinates; KCO2 2 is defined for reactions corresponding to eqns.(4A) (GlyAsp, AspGly, Asp) and (4B) for the rest of the ligands. b Ligand is protonated at amino group: b1 is defined for reactions VO 1 LH VOLH (GlyAsp, AspGly and Asp) or VO 1 L9 VOL9 (rest of the ligands). The basicity corrected b1* is defined for VO 1 H3L VOHL 1 2H1 (GlyAsp, AspGly and Asp) or VO 1 H2L9 VOL9 1 2H9 (rest of the ligands); 7-m.r., 6-m.r. = 7- and 6-membered ring respectively. c The same as for footnote b but for bis complexes: b2 1 is defined for reactions VO 1 2HL VO(LH)2 (GlyAsp, AspGly and Asp) or VO 1 2L9 VOL92 (rest of the ligands).The basicity corrected b2* is defined for VO 1 2H3L VO(HL)2 1 4H1 (GlyAsp, AspGly and Asp) or VO 1 2H2L9 VOL92 1 4H1 (rest of the ligands).3598 J. Chem. Soc., Dalton Trans., 1998, 3587–3600 CD spectra as pH is increased for GlyAsp, NacAsp and succinic acid do however suggest that the co-ordination geometries of the complexes alter. This may be explained assuming that the O atoms of the second CO2 2 group in ML92 or ML2H2 also coordinate to VO21.The co-ordination geometries either involve all carboxylate O atoms equatorial or three equatorial and one axial (in the case of GlyAsp with possibly Oamide equatorial). The ESR parameters for ML92 or ML2H2 cannot be determined accurately but the approximate values presented in Table 2 are consistent with these geometries if the contributions of Ocarboxylate and Oamide are similar. Formation constants b2 1 (corrected for the protonation of the NH3 1 group: b2 1 = b122/b011 2) for complexes involving two 7-membered chelate rings co-ordinated by carboxylate groups (assuming these are the donor groups) are ª105 (third row of Table 4), while for malonate, which involves two 6-membered chelate rings, b120 ª 109.5.Basicity corrected formation constants b*, defined assuming ligands totally protonated (for H3L ligands: b2* = b122/b013 2 and for H2L9 ligands b2* = b120/b012 2) are significantly higher for GlyAsp and Asp than for NacAsp and succinic acid.As for MLH and ML2H3 there is extra stabilisation for GlyAsp complexes possibly due to the co-ordination of Oamide of one of the ligands. The VIS spectra for ML2H2 (GlyAsp) and ML29 (NacAsp and succinic acid) approximately correspond to spectrum 7 of Fig. 3(A). The CD spectrum is dominated by a relatively intense and negative band I (GlyAsp and NacAsp: lmax ª 820–830 nm), a negative band II at lmax ª 550 nm (GlyAsp) or a relatively weak and positive band II at lmax ª 575 nm (NacAsp); all these diVer from the corresponding bands in the L-Asp system: this is also in accordance with its diVerent binding mode.ML2H (GlyAsp) The pK for the process ML2H2 æÆ ML2H 1 H1 (sixth row of Table 4) is 5.4 (ª3.5 for L-Asp). This is not the deprotonation of equatorial water as ML2H2 has no such group. While this pK gives little change in ESR, the pattern of CD shows drastic changes from pH 4.5 to 6.5 due to the formation of ML2H and MLH21.This may involve changes in the co-ordination of the equatorial and axial donor functions. The spin-Hamiltonian parameters for ML2H (Table 2) practically coincide with those expected for four equatorially co-ordinated Ocarboxylate and/or Oamide atoms,25,26 so the geometry may be either (CO2 2)4, (CO2 2)3(Oamide) or (CO2 2)2(Oamide)2. Analysing the CD spectra for L :M 15 and 30 : 1 in the pH range 4.5–6.5 (Fig. 5) we expect for ML2H a pattern 2,1,2 with lmax ª530 ± 20 nm (band II), ª630 ± 20 (band IB) and ª840 ± 30 nm (band IA).Therefore the symmetry for this complex is low and the structure such that bands IA and IB are separate. Axial co-ordination of OH2 is not expected to promote change in the CD. Alternative coordination geometries for ML2H2 are 10a and 11a (Table 3: Y = H2O). These correspond to A|| est ª 172 × 1024 cm21 and could help explain the low Dem values for the GlyAsp system relative to AspGly, as the equatorial NH2 and C]] Oamide are far from the asymmetric carbon and a low vicinal eVect is expected.It might be surprising that formation of bis complexes is disfavoured with AspGly, although N-terminal Asp oVers a b-Ala chelation site. However, this is not favoured for VO21 which binds N-donors rather weakly. Unlike GlyAsp, chelating O-donor sites, including carboxylate groups (strong binding functions for VO21), are not similarly available in AspGly. MLH21, ML9H22 and MLH22 The pK values for the formation of MLH21 (seventh row in Table 4) for the dipeptides can either correspond to the deprotonation/co-ordination of Npeptide, or to the deprotonation of an equatorially coordinated water molecule.These two processes correspond to the formation of complexes such as 5b, 6b, 7b, 12b, 13b, or 14–18 (Table 3). The stoichiometry MLH22 (GlyAsp) probably corresponds to a co-ordination geometry such as 17b (possibly OH2 axial instead of CO2 2).For L :M = 30 :1, maximum |Dem| are found at pH ª 9.5, conditions where MLH22 would dominate CD spectra for GlyAsp; the ESR and CD spectra similarly resemble those for the corresponding complexes with AlaAla and GlyAla [ESR: A|| exptl = (161 ± 1) × 1024 cm21, g|| = 1.954 ± 0.004. CD spectra: band II, lmax ª 500 nm, De < 0; band I, lmax ª 720–750 nm, De > 0).16 This is why structure 17b is assigned to MLH22 and not 18. This is also consistent with 5-membered rings being more stable than 6-membered ones.Over the pH range 6.5–8 where MLH21 (GlyAsp) is predominant, as pH is increased the ESR spectra change little apart from decreasing intensity, but the CD shows changes due to the processes of Scheme 2. Some of these equilibria take time to establish making the exact pattern of CD bands for MLH21 elusive. The band pattern 2,1,2 at L:M = 15 : 1 may be due to the optically active oligomeric species: increasing L :M, the CD spectra in the pH range 7–8.5 change towards a 2,1 pattern.So MLH21 and MLH22 may well have identical CD band patterns, i.e. both like the corresponding complexes in the AlaAla and GlyAla systems; 16 the co-ordination geometry for MLH21 therefore probably corresponds to 17a, although it is not possible to rule out geometries such as (CO2 2, NH2, CO2 2, OH2)equatorial. The few results available for the oligomeric optically active species that form make it impossible to predict their binding modes.Assuming 17a corresponds to the geometry of MLH21 in the GlyAsp 1 VO21 system, and taking into account that now the vanadium atom is an asymmetric centre, two diastereomers may be considered. The geometry for both complexes corresponds to a distorted polyhedron and the contributions of inherent dissymmetry and asymmetric distortions to the optical activity will roughly cancel if both diastereomers are present in equal concentrations; however, if in 17a the b-CO2 2 is co-ordinated axial, this makes 17a slightly more stable than 17a*, these contributions to optical activity and its magnitude then depending on the relative concentration of the diastereomers.Similar comments apply to most structures included in Table 3. For the MLH21 (and MLH22) stoichiometries in the AspGly system, besides 5b, 6b, 7b, the following basic binding modes may be envisaged all compatible with the ESR: (CO2 2, N2 amide, NH2, Y)equatorial, e.g. 15, (CO2 2, NH2, CO2 2, Y)equatorial, e.g. 12, 13 or (CO2 2, N2 amide, CO2 2, Y)equatorial, e.g. 16. The present results and earlier data do not suYce to define the geometry for MLH21 and MLH22 stoichiometries, 15 and 16 corresponding to what is normally assumed for Cu21 complexes.30,31 Structure 16 is more in agreement with the high |Dem| values obtained for the AspGly–VO21 system [Fig. 5(B)]. In the GlyAsp system MLH21 starts to form at pH higher than for AspGly; this may be due to its higher pKa2, pKa3, and to some steric factor making tetradentate co-ordination (e.g. 17a) unfavorable. In solutions containing NacAsp and VO21 [L:M = 15 or 30:1 e.g. Fig. 5(C)] the CD spectra for pH > 5 diVer from those of GlyAsp in similar conditions, and the |Dem| values decrease for pH > 5–5.5. If N2 amide co-ordinates, e.g.: (CO2 2, N2 amide, CO2 2, Y)equatorial, we would expect an increase in |Dem| with pH and lower values of A|| exptl for ML9H22 (Table 2). These observations indicate that, for the NacAsp system, N2 amide does not C O O V* NH2 N W O HC C CH2 O CH2 – OOCb * W *V N H2N O O C O O H2C C CH Cb CH2 O * ? O 17a (V chirality50: C.a-C:L) 17a* (V chirality50: A. a-C:L)J. Chem. Soc., Dalton Trans., 1998, 3587–3600 3599 co-ordinate, and that NH2, although not a good anchor for VO21, is involved in co-ordination for the corresponding stoichiometries in the GlyAsp system (in agreement with our conclusions there). Therefore, the stoichiometry ML9H22 (NacAsp and succinic) possibly involves complexes where co-ordination is (OH2, OH2, CO2 2, H2O)equatorial and (CO2 2)axial, which corresponds to A|| est ª 168 × 1024 cm21.This and the extensive hydrolysis of oxovanadium(IV) are also compatible with the very low CD and ESR signal at pH > 7 for NacAsp 1 VO21. The only evidence for the stoichiometry ML9H23 (succinic acid) is that a new ESR active species appears at pH ª 7 (see above) with A|| exptl ª (165 ±2) × 1024 cm21; this possibly corresponds to a geometry involving (2CO2 2, 2OH2)equatorial (OH2)axial.This would have A|| est ª163 × 1024 cm21 while geometries of the type (CO2 2, 3OH2)equatorial (CO2 2)axial would give A|| est ª159 × 1024 cm21. Summarizing the behaviour of the studied dipeptides, in the weakly acidic range only the carboxylate, the peptide CO and amino groups (this latter being somewhat less favoured) participate in the metal ion binding. The position of the tridentate Asp residue in the peptide chain aVects the co-ordination mode of the ligands: when Asp is C-terminal (GlyAsp) it rather behaves as a succinic acid favouring equatorial carboxylate chelation of two neighbouring groups with some involvement of the peptide CO in metal binding. In the N-terminal Asp dipeptide, AspGly, the involvement of either the peptide carbonyl or the terminal amino groups seems more essential, since the two carboxylates are much further apart.For pH > 6–7 the dominant stoichiometries for the ESR and CD active complexes are MLH21 and MLH22.Further, for GlyAsp the results indicate the basic binding mode: (CO2 2, N2 amide, NH2, Y)equatorial with Y = H2O or OH2. The axial coordination of the b-CO2 2 group is possible but unproven. The binding mode for AspGly diVers, particularly the fact that, at least till L :M = 8 : 1, only 1 : 1 complexes form. The distinction in the mode of 1 : 1 attachment to the VO21 ion of these isomeric dipeptides is a remarkable new property.Whereas with the ‘spherical’ copper(II) ion, the chelating stabilities of N-terminal GlyAsp40 and C-terminal AspGly41 are the same, with the lop-sided vanadyl ion they diVer markedly. Acknowledgements We thank Fundo Europeu para o Desenvolvimento Regional (FEDER), project PRAXIS 2/2.1/QUI/151/94), the National Research Fund (project OTKA T2273/97) and the Hungarian Ministry of Culture and Education (project FKFP OOB/97) for financial support, and the Hungarian–Portuguese Intergovernmental S & T Co-operation Programme for 1998–1999 for travelling funds.We thank B. Herold, L. Alcácer and R. T. Henriques for the use of their ESR facilities and Fundação Calouste Gulbenkian for travel grants. References 1 J. Costa Pessoa, L. F. Vilas Boas, R. D. Gillard and R. J. Lancashire, Polyhedron, 1988, 7, 1245. 2 J. Costa Pessoa, L. F. Vilas Boas and R. D. Gillard, Polyhedron, 1989, 8, 1173. 3 J. Costa Pessoa, L. F. Vilas Boas and R.D. Gillard, Polyhedron, 1989, 8, 1745. 4 J. Costa Pessoa, L. F. Vilas Boas and R. D. Gillard, Polyhedron, 1990, 9, 2101. 5 J. Costa Pessoa, R. L. Marques, L. F. Vilas Boas and R. D. Gillard, Polyhedron, 1990, 9, 81. 6 J. Costa Pessoa, J. L. Antunes, L. F. Vilas Boas and R. D. Gillard, Polyhedron, 1992, 11, 1449. 7 J. Costa Pessoa, S. M. Luz, I. Cavaco and R. D. Gillard, Polyhedron, 1994, 13, 3177. 8 J. Costa Pessoa, S. M. Luz and R. D. Gillard, Polyhedron, 1995, 14, 1495. 9 J. Costa Pessoa, S. M. Luz and R. D. Gillard, Polyhedron, 1993, 12, 2857. 10 I. Nagypál and I. Fábián, Inorg. Chim. Acta, 1982, 62, 193. 11 H. Sakurai, Y. Hamada, S. Shimomura, S. Yamashita and K. Ishizn, Inorg. Chim. Acta, 1980, 46, L119; H. Sakurai, S. Shimomura and K. Ishizn, Inorg. Chim. Acta, 1981, 55, L67; H. Sakurai, Z. Taira and N. Sakai, Inorg. Chim. Acta, 1988, 151, 85. 12 M. D. Sifri and T. L. Riechel, Inorg. Chim. Acta, 1988, 142, 229. 13 A. Dess, G. Micera and D Sanna, J.Inorg. Biochem., 1993, 52, 275. 14 T. Kiss, P. Buglyó, G. Micera, A. Dessi and D. Sanna, J. Chem. Soc., Dalton Trans., 1993, 1849. 15 K. Knüttel, A. Müller, D. Rehder, H. Vilter and V. Wittneben, FEBS, 1992, 302, 11. 16 J. Costa Pessoa, S. M. Luz and R. D. Gillard, J. Chem. Soc., Dalton Trans., 1997, 569. 17 M. Delfini, E. Gaggelli, A. Lepri and G. Valensin, Inorg. Chim. Acta, 1985, 107, 87. 18 R. P. Ferrari, E. Laurenti, S. Poli and L. Casella, J. Inorg. Biochem., 1992, 45, 99. 19 E. G. Ferrer, P. A. Williams and E. J. Baran, Biol. Trace Elem. Res., 1991, 30, 175; J. Inorg. Biochem., 1993, 50, 253; P. A. Williams and E. J. Baran, J. Inorg. Biochem., 1994, 54, 75. 20 E. G. Ferrer, P. A. Williams and E. J. Baran, Biol. Trace Elem. Res., 1996, 55, 153; Transition Met. Chem., 1997, 22, 589. 21 L. Xiaoping and X. Yuanzhi, Kexue Tongbao, 1985, 30, 340; L. Xiaoping and Z. Kangjing, J. Crystallogr. Spectrosc. Res., 1986, 16, 681. 22 J. Costa Pessoa, T.Gajda, S. M. Luz, T. Kiss, J. J. G. Moura and I. Torok, J. Inorg. Biochem., 1997, 67, 388. 23 L. F. Vilas Boas and J. Costa Pessoa, in Comprehensive Coordination Chemistry, eds. G. Wilkinson, R. D. Gillard and J. A. McCleverty, Pergamon, Oxford, 1987, vol. 3, pp. 453–583 and refs. therein. 24 N. D. Chasteen, in Biological Magnetic Resonance, eds. J. Lawrence, L. J. Berliner and J. Reuben, Plenum, New York, 1981, vol. 3, p. 53. 25 C. R. Cornman, E. P. Zovinka, Y. D. Boyajian, K. M. Geiser-Bush, P. D. Boyle and P. Singh, Inorg. Chem., 1995, 34, 4213. 26 B. J. Hamstra, A. L. P. Houseman, G. J. Colpas, J. W. Kampf, R. LoBrutto, W. D. Frasch and V. L. Pecoraro, Inorg. Chem., 1997, 36, 4866. 27 F. W. B. Einstein, R. J. Batchelor, S. J. Angus-Dunne and A. S. Tracey, Inorg. Chem., 1996, 3, 1680. 28 A. J. Tasiopoulos, Y. G. Deligiannakis, J. D. Woolins, A. M. Z. Slawin and T. A. Kabanos, Chem. Commun., 1998, 569. 29 A. J. Tasiopoulos, A. T. Vlahos, A. D. Keramidas, T. A. Kabanos, Y. G. Deligiannakis, C. P. Raptopoulou and A. Terzis, Angew. Chem., Int. Ed. Engl., 1996, 35, 2531. 30 R. Kuroda and Y. Saito, in Circular Dichroism. Principles and Applications, eds. K. Nakanishi, N. Berova and R. W. Woody, VCH, New York, 1994, ch. 9, pp. 223–225. 31 H. Sigel and R. B. Martin, Chem. Rev., 1982, 82, 385 and refs. therein. 32 N. C. Payne, Inorg. Chem., 1973, 12, 1151; K. Z. Suzuki, Y. Sasaki, S. Ooi and K. Saito, Bull. Chem. Soc. Jpn., 1980, 53, 1288; S. Okazaki and K. Saito, Bull. Chem. Soc. Jpn., 1982, 55, 785. 33 I. Cavaco, J. Costa Pessoa, M. T. Duarte, R. T. Henriques, P. M. Matias and R. D. Gillard, J. Chem. Soc., Dalton Trans., 1996, 1989. 34 G. Gran, Acta Chem. Scand., 1950, 4, 559. 35 I. Nagypál and I. Fábián, Inorg. Chim. Acta, 1982, 61, 109. 36 E. D. MoVat and R. I. Lytle, Anal. Chem., 1959, 31, 926. 37 T. Gajda, B. Henry and J. J. Delpeuch, J. Chem. Soc., Dalton Trans., 1992, 2301. 38 L. Zekany and I. Nagypal, in Computational Methods for the Determination of Stability Constants, ed. D. Leggett, Plenum, New York, 1985. 39 R. P. Henry, P. C. H. Mitchell and J. E. Prue, J. Chem. Soc., Dalton Trans., 1973, 1156. 40 A. Gergeley and E. Farkas, J. Chem. Soc., Dalton Trans., 1982, 381. 41 I. Sóvágó, E. Farkas, T. Jankowska and H. Kozlowski, J. Inorg. Biochem., 1993, 51, 715. 42 A. E. Martell and R. M. Smith, Critical Stability Constants, Plenum, New York, 1974, vol. 1; 1977, vol. 3; 1982, vol. 5; 1989, vol. 6. 43 S. P. Singh and J. P. Tandon, Acta Chim. Acad. Sci. Hung., 1974, 80, 425. 44 A. E. Martell and R. J. Motekaitis, Determination and Use of Stability Constants, VCH, Weinheim, 1988, p. 197. 45 EPRPOW, L. K. White and R. L. Belford, University of Illinois (modified by L. K. White, N. F. Albanese and N. D. Chasteen, University of New Hampshire to include both Lorentzian and Gaussian line shape functions, an I = 7/2 nucleus, a 4th hyperfine interaction and multiple sites having diVerent linewidths), 1978. 46 G. Micera and A. Dessi, J. Inorg. Biochem., 1988, 33, 99.3600 J. Chem. Soc., Dalton Trans., 1998, 3587–3600 47 D. B. McPhail and B. A. Goodman, J. Chem. Soc., Faraday Trans., 1987, 3513. 48 L. Casella, M. Gullotti and A. Pintar, Inorg. Chim. Acta, 1988, 144, 89. 49 R. D. Gillard, J. Inorg. Nucl. Chem., 1964, 26, 657; J. P. Mathieu, Contributions to the Study of Molecular Structure, (Victor Henri Commem. Vol), Désoer, Liège, 1947, p. 111. 50 G. J. Leigh (Editor), Nomenclature of Inorganic Chemistry, Blackwell Scientific Publications, Oxford, 1990, p. 186. 51 S. Mondal, S. P. Rath, K. K. Rajak and A. Chakravorty, Inorg. Chem., 1998, 37, 1713. Paper 8/01888J

ISSN:1477-9226    

DOI:10.1039/a801888j

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3601–3607 3601 Neutral and cationic ruthenium hydrotris(pyrazolyl)borate derivatives containing bulky monodentate phosphines. Crystal structures of [RuTp(H2O)(PPri 2Me)2][CF3SO3]?EtOH and [RuTp(N2)(PEt3)2][BPh4] Miguel A. Jiménez Tenorio, Manuel Jiménez Tenorio, M. Carmen Puerta * and Pedro Valerga Departamento de Ciencia de Materiales e Ingeniería Metalúrgica y Química Inorgánica, Facultad de Ciencias, Universidad de Cádiz, Apartado 40, 11510 Puerto Real, Cádiz, Spain.E-mail: carmen.puerta@uca.es Received 26th June 1998, Accepted 28th August 1998 The complexes [RuTp(Cl)(PPri 2Me)2] 1 and [RuTp(Cl)(PEt3)2] 2 were prepared by thermal displacement of PPh3 from [RuTp(Cl)(PPh3)2] by the corresponding phosphine. A series of cationic complexes of the type [RuTp(L)- (PR3)2]1 (L = H2O, N2 or CNBut; PR3 = PPri 2Me or PEt3) was obtained by chloride abstraction from 1 or 2 in the presence of L. As consequence of the steric crowding, one of the PPri 2Me ligands in 1 is labile, and it was readily replaced by neutral molecules such as MeCN or CNBut to yield the neutral complexes [RuTp(Cl)L(PPri 2Me)] (L = MeCN or CNBut).The monohydrides [RuTp(H)(PPri 2Me)2] and [RuTp(H)(PEt3)2] were obtained by treatment of 1 or 2 with NaBH4 in MeOH. Protonation of these monohydrides led to the cationic dihydrogen complexes [RuTp(H2)(PPri 2Me)2]1 and [RuTp(H2)(PEt3)2]1 which were isolated as [BPh4]2 salts, and characterized by determination of their (T1)min and 1JHD coupling constants in the corresponding isotopomers. The neutral hydride(dihydrogen) complex [RuTp(H)(H2)(PPri 2Me)], which resulted from the reaction of [RuTp(Cl)(MeCN)- (PPri 2Me)] with NaBH4 in MeOH, was characterized in analogous fashion.The crystal structures of the complexes [RuTp(H2O)(PPri 2Me)2][CF3SO3]?EtOH and [RuTp(N2)(PEt3)2][BPh4] were determined. Introduction The chemistry of ruthenium hydrotris(pyrazolyl)borate (Tp) complexes, which had been rather underdeveloped compared to that of the related and formally homologous cyclopentadienyl and pentamethylcyclopentadienyl ruthenium derivatives, has attracted increasing attention in recent years.New TpRu complexes containing a variety of co-ligands, particularly nitrogen donors such as Me2NCH2CH2NMe2, MeCN, pyridine, etc. as well as cycloocta-1,5-diene (COD) and hemilabile phosphino amine ligands of the type Ph2PCH2CH2NR2, have been reported very recently,1–6 and many of these have shown to be active catalysts for processes such as the dimerization of terminal acetylenes or the coupling of phenylacetylene with benzoic acid or allyl alcohols.1,7 Tertiary phosphines have also been used as co-ligands,8–15 and complexes such as [RuTp(X)- (PPh3)2] (X = Cl or H) 8,12,15 and [RuTp(MeCN)(PPh3)2][BF4] 12 shown to be quite eYcient catalyst precursors.7,12,16 Interestingly, complexes containing bidentate phosphine ligands, e.g.[RuTp(Cl)(dppe)],1 are catalytically inactive.7 It seems that the catalyst precursor must contain at least one weakly bound ligand which upon dissociation generates the co-ordinatively unsaturated, catalytically active species. We have previously prepared TpRu complexes containing the bulky diphosphine 1,2-bis(diisopropylphosphino)ethane (dippe), and studied its reactivity towards small molecules,15 and also towards terminal alkynes and alkynols.17 Although these complexes exhibit a chemical reactivity similar to that of their half-sandwich homologues, 18 they are catalytically inactive. For this reason, and continuing our studies on the chemistry of ruthenium complexes with bulky phosphine ligands, we have now focused our attention on PPri 2Me, a bulky phosphine which can be considered as the monodentate equivalent of dippe.Dissociation of one monodentate PPri 2Me ligand is expected to occur more easily than the chelate ring opening in complexes containing bidentate phosphines such as dippe or dppe, and hence such derivatives are more likely to act as suitable precursors for catalytically active species.In this work we describe the synthesis, properties and chemical reactivity of a range of neutral and cationic TpRu complexes containing PPri 2Me, including dinitrogen adducts as well as “classical” and “non-classical” hydrides, and also related derivatives with PEt3, a less sterically demanding phosphine ligand having electron-donating capabilities similar to those of PPri 2Me.Experimental All synthetic operations were performed under a dry dinitrogen or argon atmosphere following conventional Schlenk or drybox techniques. Tetrahydrofuran, diethyl ether and light petroleum (boiling point range 40–60 8C) were distilled from the appropriate drying agents. All solvents were deoxygenated immediately before use. Triethylphosphine was purchased from Aldrich, whereas PPri 2Me was obtained by reaction of PClPri 2 (Aldrich) with MgMeI in diethyl ether; KTp19 and [RuTp(Cl)- (PPh3)2] 8 were obtained according to published procedures.Infrared spectra were recorded in Nujol mulls on Perkin- Elmer FTIR Spectrum 1000 spectrophotometers, NMR spectra on Varian Unity 400 MHz or Gemini 200 MHz equipment. Chemical shifts are given in ppm from SiMe4 (1H and 13C-{1H}) or 85% H3PO4 (31P-{1H}). The coupling constants 3JHH for the Tp ligand were all in the range 2–2.5 Hz.Microanalyses were by the Serveis Científico-Tècnics, Universitat de Barcelona. Preparations [RuTp(Cl)(PPri 2Me)2] 1. To a solution of [RuTp(Cl)(PPh3)2] (1.75 g, 2 mmol) in toluene (15 ml), PPri 2Me (0.75 ml, ca. 5 mmol) was added via a syringe. The mixture was refluxed for 1 h, then allowed to cool to room temperature and light petrol-3602 J. Chem. Soc., Dalton Trans., 1998, 3601–3607 eum (15 ml) added. A pale yellow, crystalline precipitate was obtained. It was filtered oV, washed with diethyl ether and light petroleum, and dried in vacuo.Yield: 0.74 g, 60% (Found: C, 45.2 ; H, 7.28; N, 13.4. Calc. for C23H44BClN6P2Ru: C, 45.0; H, 7.17; N, 13.7%). IR: n(BH) 2456 cm21. NMR (C6D6): 1H, d (298 K) 20.23, 0.74, 1.34, 1.54 {m, P[CH(CH3)2]2}; 1.14 (d, JHP = 3.2 Hz, PCH3); 2.31, 2.95 {m, P[CH(CH3)2]2}; 5.85 (t, 1 H), 5.93 (t, 2 H), 7.29 (d, 1 H), 7.51 (d, 2 H), 7.54 (d, 1 H) and 8.39 (d, 2 H); d (198 K, CD2Cl2) 20.63, 0.78, 0.92, 1.14, 1.25, 1.32, 1.39, 2.10, 3.35 (br, PPri 2Me); 6.05 (s, 1 H), 6.13 (br, 2 H), 7.14 (s, 1 H), 7.65 (br, 2 H), 7.70 (s, 1 H) and 7.91 (br, 2 H); 31P-{1H}, d (298 K) 27.5 (s); d (183 K, CD2Cl2) 23.9 (d), 25.9 (d), 2JPP = 30.9 Hz; 13C-{1H}, d (298 K) 6.8 (m, PCH3); 16.2, 17.7, 19.1, 19.2 {s, P[CH(CH3)2]2}; 24.1, 28.1 {m, P[CH(CH3)2]2}; 104.7, 105.0, 134.5, 136.2, 145.1 and 147.1 [s, HB(C3H3N2)3].[RuTp(Cl)(PEt3)2] 2. Complex 2 was prepared in a fashion analogous to that for 1, starting from [RuTp(Cl)(PPh3)2] (1.75 g, 2 mmol) and PEt3 (0.75 ml, ca. 5 mmol). Yield: 0.68 g, 65% (Found: C, 43.0; H, 6.96; N, 14.1. Calc. for C21H40BClN6P2Ru: C, 43.1; H, 6.83; N, 14.4%). IR: n(BH) 2469 cm21. NMR (CDCl3): 1H, d 0.75 [m, P(CH2CH3)3], 1.87 [m, P(CH2CH3)3]; 6.07 (t, 1 H), 6.13 (t, 2 H), 7.41 (d, 1 H), 7.61 (d, 2 H), 7.71 (d, 1 H) and 7.97 (d, 2 H); 31P-{1H}, d 26.8 (s); 13C-{1H}, d 7.8 [s, P(CH2CH3)3], 18.4 [m, P(CH2CH3)3]; 104.0, 105.2, 134.9, 136.2, 143.9 and 147.6 [s, HB(C3H3N2)3].[RuTp(H2O)(PPri 2Me)2][CF3SO3]?EtOH 3. To a tetrahydrofuran solution (15 ml) of complex 1 (0.3 g, 0.5 mmol) under argon, AgO3SCF3 (0.12 g, ca. 0.5 mmol) was added. The mixture was stirred at room temperature for 1 h, then, filtered through Celite in order to remove the precipitate of AgCl. The solvent was removed in vacuo, and the residue dissolved in 96% EtOH. Concentration and cooling to 220 8C for several days aVorded well formed crystals, which were filtered oV, washed with light petroleum and dried in vacuo.Yield: 0.29 g, 77% (Found: C, 41.0; H, 6.83; N, 10.9. Calc. for C26H52F3N6O5P2RuS: C, 41.1; H, 6.85; N, 11.1%). IR: n(BH) 2463 cm21. NMR (CDCl3): 1H, d 20.23, 1.01, 1.23, 1.34 {m, P[CH(CH3)2]2}; 1.45 (d, JHP = 6 Hz, PCH3); 2.26, 2.44 {m, P[CH(CH3)2]2}; 3.10 (br, H2O); 6.08 (t, 1 H), 6.25 (t, 2 H), 7.29 (d, 1 H), 7.63 (d, 1 H), 7.73 (br, 2 H) and 7.84 (br, 2 H); 31P-{1H}, d 28.5 (s); 13C-{1H}, d 7.7 (m, PCH3); 16.9, 18.7, 19.8, 20.0 {s, P[CH(CH3)2]2}; 25.1, 28.9 {m, P[CH(CH3)2]2}; 105.6, 106.0, 135.6, 137.3, 145.5 and 148.6 [s, HB(C3H3N2)3].[RuTp(H2O)(PEt3)2][BPh4] 4. Complex 4 was obtained following a procedure identical to that for 3, starting from 2 (0.29 g, 0.5 mmol). It was converted into its tetraphenylborate salt by addition of NaBPh4 (0.3 g, excess) to an ethanol solution. Cooling to 220 8C aVorded white crystals, which were filtered oV, washed with light petroleum and dried in vacuo.Yield: 0.35 g, 80% (Found: C, 61.0; H, 7.14; N, 9.2. Calc. for C45H62B2N6OP2Ru:C, 60.9; H, 6.99; N, 9.5%). IR: n(BH) 2479 cm21. NMR (CDCl3): 1H, d 0.70 [m, P(CH2CH3)3], 1.68 [m, P(CH2CH3)3]; 5.30 (s, H2O); 6.19 (t, 1 H), 6.24 (t, 2 H), 7.38 (d), 7.44 (s br), 7.69 (d, 2 H) and 7.77 (d, 1 H); 31P-{1H}, d 25.0 (s); 13C-{1H} [(CD3)2CO], d 6.9 [s, P(CH2CH3)3], 17.9 [m, P(CH2CH3)3]; 106.0, 106.2, 136.0, 137.2, 143.2 and 147.9 [s, HB(C3H3N2]3. [RuTp(N2)(PPri 2Me)2][BPh4] 5. To a tetrahydrofuran solution (15 ml) of complex 1 (0.3 g, 0.5 mmol) under dinitrogen, AgO3SCF3 (0.12 g, ca. 0.5 mmol) was added. The mixture was stirred at room temperature for 2 h then filtered through Celite or centrifuged. The solvent was removed in vacuo, and the residue dissolved in MeOH. Addition of solid NaBPh4 (0.3 g, excess), concentration and cooling to 220 8C for several days aVorded red crystals, which were filtered oV, washed with ethanol and light petroleum and dried in vacuo.They were recrystallized from a mixture of dichloromethane and light petroleum. Yield: 0.31 g, 68% (Found: C, 61.2; H, 7.03; N, 11.8. Calc. for C47H64B2N8P2Ru: C, 61.0; H, 6.92; N, 12.1%). IR: n(BH) 2491, n(N]] ] N) 2159 cm21. NMR [(CD3)2CO]: 1H, d 20.32, 1.04, 1.32 {m, P[CH(CH3)2]2}; 1.68 (d, JHP = 7.2 Hz, PCH3); 2.36, 2.72 {m, P[CH(CH3)2]2}; 6.23 (t, 1 H), 6.33 (t, 2 H), 7.75 (d, 1 H), 7.84 (d, 2 H), 7.92 (d, 1 H) and 7.99 (d, 2 H); 31P-{1H}, d 25.4 (s); 13C-{1H}, d 6.9 (m, PCH3); 16.5, 18.2, 18.8, 19.4 {s, P[CH(CH3)2]2}; 24.8, 27.0 {m, P[CH(CH3)2]2}; 106.9, 107.5, 137.3, 139.0, 146.5 and 150.7 [s, HB(C3H3N2)3].[RuTp(N2)(PEt3)2][BPh4] 6. A procedure identical to that for the preparation and recrystallization of complex 5 was followed, starting from 2 (0.29 g, 0.5 mmol). Yield: 0.33 g, 73% (Found: C, 59.9; H, 7.01; N, 12.1. Calc. for C45H60B2N8P2Ru: C, 60.2; H, 6.69; N, 12.5%). IR: n(BH) 2505 n(N]] ] N) 2163 cm21. NMR [(CD3)2CO]: 1H, d 0.78 (m, PCH2CH3), 1.93 (m, PCH2CH3); 6.20 (t, 1 H), 6.34 (t, 2 H), 7.70 (d, 1 H), 7.82 (d, 2 H), 7.91 (d, 1 H) and 7.98 (d, 2 H); 31P-{1H}, d 25.8 (s); 13C- {1H}, d (CD2Cl2) 7.9 [s, P(CH2CH3)3], 18.5 [m, P(CH2CH3)3]; 107.5, 106.5, 137.5, 138.8, 143.2 and 143.3 [s, HB(C3H3N2)3].[RuTp(CNBut)(PPri 2Me)2][BPh4] 7. Method A. To a suspension of complex 1 (0.15 g, ca. 0.25 mmol) in EtOH (10 ml), NaBPh4 (0.2 g, excess) and a few drops of CNBut were added. The mixture was heated smoothly for ca. 2 h using a water-bath. Then, the resulting yellow solution was concentrated and cooled to 220 8C. The white solids were collected by filtration, washed with EtOH and light petroleum and dried in vacuo. Yield: 0.2 g, 82%. Method B. A dichloromethane solution (10 ml) of complex 5 (0.1 g, ca. 0.11 mmol) was treated with a few drops of CNBut. The solution became pale yellow. It was stirred at room temperature for 10 min. Concentration and addition of light petroleum yielded a white precipitate, which was filtered oV, washed with light petroleum and dried in vacuo.Yield: 0.1 g, quantitative (Found: C, 63.8; H, 7.55; N, 9.8. Calc. for C52H73- B2N7P2Ru: C, 63.7; H, 7.45; N, 10.0%). IR: n(BH) 2487, n(C]] ] N) 2124 cm21. NMR (CDCl3): 1H, d 20.33, 0.91 1.28, 1.36 {m, P[CH(CH3)2]2}; 1.29 (d, JHP = 6.8 Hz, PCH3); 1.44 [s, RuCNC- (CH3)3]; 1.94, 2.41 {m, P[CH(CH3)2]2}; 6.20 (t, 2 H), 6.33 (t, 1 H), 7.63 (d, 2 H), 7.65 (d, 2 H), 7.69 (d, 1 H) and 7.83 (d, 1 H); 31P-{1H}, d 26.0 (s); 13C-{1H}, d 7.2 (m, PCH3); 19.5, 19.1, 18.1, 16.2 {s, P[CH(CH3)2]2}; 30.8, 24.9 {m, P[CH(CH3)2]2}; 31.1 [s, RuCNC(CH3)3]; 49.3 [s, CNC(CH3)3]; 106.7, 136.8, 137.5, 144.6, 145.7 [s, HB(C3H3N2)3]; RuCNC(CH3)3 not observed.[RuTp(CNBut)(PEt3)2][BPh4] 8. This complex was obtained by either of the two procedures outlined for the preparation of 7, starting either from 2 (method A) or from 6 (method B), with similar yields (Found: C, 63.1; H, 7.33; N, 10.1.Calc. for C50H69B2N7P2Ru: C, 63.1; H, 7.25; N, 10.3%). IR: n(BH) 2471, n(C]] ] N) 2123 cm21. NMR (CDCl3): 1H, d 0.72 [m, P(CH2CH3)3]; 1.43 [s, CNC(CH3)3]; 1.69 [m, P(CH2CH3)3]; 6.02 (t, 2 H), 6.31 (t, 1 H), 7.48 (d, 2 H), 7.62 (d, 1 H), 7.66 (d, 2 H) and 7.84 (d, 1 H); 31P-{1H}, d 24.0 (s); 13C-{1H}, d 7.5 [s, P(CH2CH3)3], 19.1 [m, P(CH2CH3)3]; 30.8 [s, RuCNC(CH3)3]; 49.5 [s, RuCNC( CH3)3]; 106.4, 106.5, 136.5, 137.1, 143.4 and 145.5 [s, HB(C3H3N2)3]; RuCNC(CH3)3 not observed.[RuTp(Cl)(MeCN)(PPri 2Me)] 9. A solution of complex 1 (0.3 g, 0.5 mmol) in MeCN (15 ml) was heated under reflux for 2 h. Then, the solvent was removed in vacuo. The resulting yellow microcrystalline material was washed with several portions of light petroleum and dried thoroughly. Yield: 0.24 g, 95% (Found: C, 41.2; H, 5.89; N, 18.5. Calc. for C18H30BClN7PRu: C, 41.4; H, 5.74; N, 18.8%). IR: n(BH) 2479; n(C]] ] N) 2269, 2245 cm21. NMR [(CD3)2CO]: 1H, d 0.15, 1.10, 1.35 {m, P[CH- (CH3)2]2}; 1.52 (d, JHP = 8 Hz, PCH3); 2.24 {m, P[CH(CH3)2]2}; 2.56 (s, RuNCCH3); 6.11 (t, 1 H), 6.16 (t, 1 H), 6.17 (t, 1 H),J.Chem. Soc., Dalton Trans., 1998, 3601–3607 3603 7.56 (d, 1 H), 7.68 (d, 1 H), 7.69 (d, 1 H), 7.74 (d, 1 H), 7.80 (d, 1 H) and 7.88 (d, 1 H); 31P-{1H}, d 42.1 (s); 13C-{1H}, d 4.1 (s, RuNCCH3); 4.2 (d, JCP = 21.7, PCH3); 16.5 (d, JHP = 1.3), 16.7 (d, JHP = 1.3), 17.8, 17.9 {s, P[CH(CH3)2]2}; 24.7 {d, JCP = 22.7 Hz, P[CH(CH3)2]2}; 105.6, 105.7, 106.6, 134.4, 136.2, 136.5, 142.1, 146.0 and 146.2 [s, HB(C3H3N2)3]. [RuTp(Cl)(CNBut)(PPri 2Me)] 10.To a solution of complex 1 (0.15 g, ca. 0.25 mmol) in tetrahydrofuran a few drops of CNBut were added. The mixture was heated at 60 8C for 2 h. Then the solvent was removed in vacuo, leaving a yellow powder which was washed with several portions of light petroleum and dried in vacuo. Yield: 0.14 g, quantitative (Found: C, 44.3; H, 6.40; N, 17.1. Calc. for C21H36BClN7PRu: C, 44.7; H, 6.38; N, 17.4%).IR: n(BH) 2460, n(C]] ] N) 2071, 2099 cm21. NMR [(CD3)2CO]: 1H, d 0.31, 1.00, 1.04, 1.14 {m, P[CH(CH3)2]2}; 0.85 (d, JHP = 3.2 Hz, PCH3); 1.51 [s, CNC(CH3)3]; 1.99, 2.19 {m, P[CH(CH3)2]2}; 6.15 (t, 1 H), 6.16 (t, 1 H), 6.24 (t, 1 H), 7.50 (d, 1 H), 7.65 (d, 1 H), 7.79 (d, 1 H), 7.82 (d, 1 H), 7.84 (br, 1 H) and 7.94 (d, 1 H); 31P-{1H}, d 42.5 (s); 13C-{1H} (C6D6), d 4.2 (d, JCP = 40, PCH3); 15.9, 17.1, 17.7, 19.2 {s, P[CH(CH3)2]2}; 24.1 (d, JCP = 41), 28.5 {d, JCP = 49 Hz, P[CH(CH3)2]2}; 31.0 [s, RuCNC(CH3)]; 48.3 [s, RuCNC(CH3)3]; 105.0, 105.1, 105.5, 134.8, 135.5, 142.6, 143.9, 145.3 [s, HB(C3H3N2)3]; 233 [br, RuCNC(CH3)3].[RuTp(Cl)(PPri 2Me)(PEt3)] 11. To a tetrahydrofuran solution of complex 1 (0.15 g, 0.25 mmol) neat PEt3 (0.05 ml, slight excess) was added. The mixture was heated at 60 8C for 1 h. Removal of the solvent and washing with light petroleum aVorded a yellow solid, which was dried in vacuo. Yield: 0.13 g, quantitative (Found: C, 43.9; H, 7.15; N, 14.0.Calc. for C22H42BClN6P2Ru: C, 44.1; H, 7.01; N, 14.0%). IR: n(BH) 2473 cm21. NMR (CDCl3): 1H, d 0.78 [m, P(CH2CH3)3], 1.88 [m, P(CH2CH3)3]; 20.39, 0.97, 1.39, 1.47 {m, P[CH(CH3)2]2}; 1.37 (d, JHP = 7.2 Hz, PCH3); 2.71, 2.26 {m, P[CH(CH3)2]2}; 6.08 (t, 1 H), 6.13 (br, 2 H), 7.44 (d, 1 H), 7.61 (d, 1 H), 7.71(br, 2 H), 8.00 (d, 2 H) and 8.03 (d, 2 H); 31P-{1H}, d 31.1 (d); 23.3 (d), 2JPP = 32.3 Hz; 13C-{1H}, d 6.8 (d, JCP = 34.8, PCH3); 6.9 [s, P(CH2CH3)3]; 18.5 [d, JCP = 45.4, P(CH2CH3)3]; 19.3, 19.2, 17.8, 15.7 {s, P[CH(CH3)2]2}; 24.0, 28.9 {d, JCP = 42.4 Hz, P[CH(CH3)2]2}; 104.0, 105.1, 105.3, 134.8, 134.9, 136.4, 144.0, 144.4 and 147.4 [s, HB(C3H3N2)3].[RuTp(H)(PPri 2Me)2] 12. To a slurry of complex 1 (0.3 g, 0.5 mmol) in MeOH (15 ml) an excess of NaBH4 (0.15 g) was added. The mixture was heated using a water-bath, until efervescence ceased. After 45 min a yellow-orange solution was obtained. Removal of the solvent yielded an oily residue, which was dissolved in light petroleum and filtered through Celite.Concentration and cooling to 220 8C aVorded a yellow solid which was filtered oV and dried in vacuo. This compound slowly turns green on standing under dinitrogen or argon, even when stored in a freezer. Yield: 0.16 g, 58% (Found: C, 47.3; H, 7.92; N, 14.1. Calc. for C23H45BN6P2Ru: C, 47.7; H, 7.77; N, 14.5%). IR: n(BH) 2456, n(RuH) 1917 cm21. NMR (C6D6): 1H, d 215.36 (t, 2JHP = 28.8, RuH); 0.02, 0.64, 1.15, 1.24 {m, P[CH(CH3)2]2}, 1.05 (d, JHP = 5.8 Hz, PCH3); 2.16 {m, P[CH(CH3)2]2}; 5.70 (s br, 2 H), 5.99 (s br, 1 H), 7.00 (s br, 2 H), 7.37 (s br, 1 H), 7.59 (s br, 1 H) and 7.76 (s br, 2 H); 31P-{1H}, d 47.8 (s); 13C-{1H}, d 9.2 (m, PCH3); 17.8, 18.1, 18.2, 19.3 {s, P[CH(CH3)2]2}; 25.3, 31.9 {m, P[CH(CH3)2]2}; 104.5, 104.6, 135.0, 145.4 and 146.9 [s, HB(C3H3N2)3].[RuTp(H)(PEt3)2] 13. Complex 13 was obtained in a fashion analogous to that for 12, starting from 2 (0.29 g, 0.5 mmol).Yield: 0.16 g, 60% (Found: C, 45.4; H, 7.27; N, 14.9. Calc. for C21H41BN6P2Ru: C, 45.8; H, 7.44; N, 15.3%). IR: n(BH) 2456, n(RuH) 1915 cm21. NMR (C6D6): 1H, d 215.11 (t, 2JHP = 29.2 Hz, RuH); 0.74 [m, P(CH2CH3)3]; 1.50 [m, P(CH2CH3)3]; 5.86 (t, 2 H), 6.14 (t, 1 H), 7.52 (d, 2 H), 7.75 ( s br, 3H) and 8.25 (d, 1 H); 31P-{1H}, d 45.9 (s); 13C-{1H}, d 8.1 [s, P(CH2CH3)3], 21.2 [m, P(CH2CH3)3]; 105.0, 105.8, 135.2, 145.8 and 146.9 [s, HB(C3H3N2)3].[RuTp(H2)(PPri 2Me)2][BPh4] 14. To a solution of complex 12 (0.15 g, 0.26 mmol) in diethyl ether (10 ml) at 280 8C under argon, HBF4?OEt2 (2–3 drops, excess) was added. The mixture was allowed to warm to room temperature. Then the solvent was removed in vacuo, and the residue treated with a MeOH solution containing NaBPh4 (0.2 g, excess). A white, microcrystalline precipitate was obtained. It was filtered oV, washed with ethanol and light petroleum and dried in vacuo.This compound was recrystallized from a dichloromethane–ethanol mixture under argon. The isotopomer [RuTp(HD)(PPri 2Me)2]1 was generated in situ, by reaction of 12 with DBF4?OEt2 (obtained from D2O–HBF4?OEt2 3 : 1). Yield: 0.19 g, 81% (Found: C, 62.5; H, 7.46; N, 9.1. Calc. for C47H66B2N6P2Ru: C, 62.8; H, 7.34; N, 9.35%). IR: n(BH) 2489 cm21. NMR (CD2Cl2): 1H, d 29.55 [s br, Ru(H2); (T1)min 16 ms at 203 K, 400 MHz, 1JHD = 31.1, JHP = 7.3 Hz]; 20.08, 0.90, 1.35 {m, P[CH(CH3)2]2}; 1.42 (d, PCH3); 2.00, 2.22 {m, P[CH(CH3)2]2}; 6.23 (t, 2 H), 6.45 (t, 1 H), 7.68 (d, 2 H), 7.72 (d, 2 H), 7.77 (d, 1 H) and 7.96 (d, 1 H); 31P-{1H}, d 30.6 (s); 13C-{1H}, d 7.0 (m, PCH3); 16.6, 17.6, 18.5, 19.4 {s, P[CH(CH3)2]2}; 25.1, 30.3 {m, P[CH(CH3)2]2}; 107.0, 107.5, 137.4, 138.2, 146.7 and 147.0 [s, HB(C3H3N2)3]. [RuTp(H2)(PEt3)2][BPh4] 15.A procedure identical to that for complex 14 was followed for 15, starting from 13 (0.15 g, 0.27 mmol). Yield: 0.17 g, 75% (Found: C, 61.9; H, 7.18; N, 9.4.Calc. for C45H62B2N6P2Ru: C, 62.0; H, 7.12; N, 9.65%). IR: n(BH) 2495 cm21. NMR (CD2Cl2): 1H, d 29.83 [s br, Ru(H2), (T1)min 18 ms at 203 K, 400 MHz, 1JHD = 30.9, JHP = 7.1 Hz]; 0.78 [m, P(CH2CH3)3]; 1.74 [m, P(CH2CH3)3]; 6.23 (t, 2 H), 6.41 (t, 1 H), 7.59 (d, 2 H), 7.71 (d, 2 H), 7.78 (d, 1 H) and 7.93 (d, 1 H); 31P-{1H}, d 28.1 (s); 13C-{1H}, d 7.3 [s, P(CH2CH3)3], 19.4 [m, P(CH2CH3)3]; 107.0, 107.2, 146.9, 146.2, 137.2 and 138.0 [s, HB(C3H3N2)3].[RuTpH3(PPri 2Me)] 16. A solution of complex 9 (0.15 g, 0.29 mmol) in MeOH (10 ml) was treated with an excess of NaBH4 (0.15 g). The mixture was heated at 60 8C for 1 h. Then the solvent was removed in vacuo, the residue extracted with light petroleum, and the solution filtered through Celite. A sticky residue was obtained upon solvent removal and dissolved in warm MeOH. Filtration, concentration and cooling to 220 8C aVorded pale yellow crystals, which were filtered oV and dried in vacuo.A mixture of the isotopomers [RuTpH2D(PPri 2Me)] and [RuTp(H)D2(PPri 2Me)] was obtained by gentle heating of a CD3OD solution of 16. Yield: 0.1 g, 77% (Found: C, 42.3; H, 6.91; N, 18.5. Calc. for C16H30BN6PRu: C, 42.8; H, 6.68; N, 18.7%). IR: n(BH) 2475, n(RuH) 1946 cm21. NMR (C6D6): 1H, d 210.45 [d, 2JHP = 19.6, RuH3, (T1)min 41 ms at 201 K, CD2Cl2, 400 MHz, 1JHD = 7.4]; 0.69, 0.83 {m, P[CH(CH3)2]2}; 1.03 (d, JHP = 7.2 Hz, PCH3); 5.99 (s br), 7.56 (s br) and 7.84 (s br); 31P-{1H}, d 62.9 (s); 13C-{1H}, d 13.3 (d, JCP = 44, PCH3); 15.0, 18.3, 18.4, 18.5 {s, P[CH(CH3)2]2}; 22.6, 26.5 {d, JCP = 44 Hz, P[CH(CH3)2]2}; 105.7, 135.6 and 146.2 [s br, HB(C3H3N2)3]. Crystallography Crystals suitable for X-ray diVraction analysis were mounted onto a glass fiber and transferred to an AFC6S-Rigaku automatic diVractometer (T = 290 K, Mo-Ka radiation, graphite monochromator, l = 0.71073 Å).Accurate unit cell parameters and an orientation matrix in each case were determined by least-squares fitting from the settings of 25 high-angle reflections.Crystal data and details on data collection and refinements are given in Table 1. Data were collected by the w–2q scan method in both cases. Lorentz-polarization corrections3604 J. Chem. Soc., Dalton Trans., 1998, 3601–3607 were applied. Decay was monitored by measuring three standard reflections every 100 measurements. Decay and semiempirical absorption correction (y method) were also applied.The structures were solved by Patterson methods and subsequent expansion of the models using DIRDIF.20 Reflections having I > 3s(I) were used for structure refinement. For complex 3 all non-hydrogen atoms were anisotropically refined; H(1), H(2) and H(52) were localized in Fourier-diVerence maps and the remaining hydrogen atoms included at idealized positions and not refined. In the case of compound 6 all the non-hydrogen atoms in the cation except the phosphine ethyl groups were anisotropically refined, and the remaining non-hydrogen atoms were isotropically refined.Hydrogen atoms were included at idealized positions and not refined. Since the space groups were non-centrosymmetrical in both cases, the two possible enantiomorphs were checked and no significant diVerences found. All calculations for data reduction, structure solution, and refinement were carried out on a VAX 3520 computer at the Servicio Central de Ciencia y Tecnología de la Universidad de Cádiz, using the TEXSAN21 software system and ORTEP22 for plotting.Maximum and minimum peaks in the final Fourierdi Verence maps were 11.42 and 20.97 e Å23 for 3, and 10.56 and 20.43 e Å23 for 6. CCDC reference number 186/1139. See http://www.rsc.org/suppdata/dt/1998/3601/ for crystallographic files in .cif format. Results and discussion The complexes [RuTp(Cl)(PPri 2Me)2] 1 and [RuTp(Cl)(PEt3)2] 2 were obtained by thermal displacement of PPh3 from [Ru- Tp(Cl)(PPh3)2] by the corresponding phosphine in refluxing toluene, a procedure which has been previously used for the synthesis of [RuTp(Cl)(dippe)].15 The physical properties of these pale yellow, crystalline compounds match those of the dippe derivative.The presence of six separate pyrazole ring proton resonances in the 1H NMR spectra, and of one singlet in the 31P-{1H} NMR spectra, of these complexes at room temperature suggest an octahedral structure analogous to that found for the parent complex [RuTp(Cl)(PPh3)2] by X-ray crystallography.It is interesting that the complexes [Ru(C5R5)- Cl(PEt3)2] (R = H or Me), formal homologues of 2, are known. However, attempts made to synthesize organometallic counterparts of complex 1, namely [Ru(C5R5)Cl(PPri 2Me)2] (R = H or Me), have been unsuccessful.23 It seems that the bulk of PPri 2Me makes diYcult, or even prevents, the formation of half-sandwich species containing two of these phosphine ligands in a cisoid disposition, an observation which is consistent with the fact that complexes of the type [Ru(C5R5)Cl(PR3)2] (R = H or Me; PR3 = PPri 3 or PCy3) are also unknown.24 Instead, 16-electron species [(C5Me5)RuCl(PR3)] (PR3 = PCy3, PPri 3, PPhPri 2) have shown to be stable, yet reactive.25 Given the increased bulk of the Tp ligand compared to C5H5 or C5Me5, the formation of a complex containing two cis-PPri 2Me ligands such as 1 is remarkable.Compound 1 shows in its 1H NMR spectrum one multiplet at d 20.23 attributable to CH3 protons of isopropyl groups of the PPri 2Me ligand, which is somewhat unusual, since such a high-field resonance has not been observed in the 1H NMR spectra of other TpRu phosphine complexes, e.g. 2 or [RuTp(Cl)(dippe)]. This anomalous chemical shift for isopropyl groups casts some doubts on whether 1 is really a plain six-co-ordinate species having a k3-Tp ligand (A), or a highly fluxional five-co-ordinate molecule containing a k2-Tp group and phosphines in a trans disposition, being stabilized by means of an “agostic” interaction with the isopropyl group of one of the phosphine ligands (B).Ruthenium derivatives containing one k2-Tp ligand and two trans phosphine ligands have recently been reported, i.e. [Ru- (k2-Tp)H(CO)(PPri 3)2] 26 and [Ru(k2-Tp)H(CO)(PPh3)2],27 so this is a real possibility to be considered in our case. When the temperature is lowered the 31P-{1H} NMR resonance observed in the spectrum of 1 broadens, and splits into two separate signals which resolve into doublets at 183 K [Fig. 1(A)], suggesting the non-equivalence of the phosphorus atoms at this temperature. At this temperature, three of the pyrazole ring resonances in the 1H NMR spectrum broaden [Fig. 1(B)], although they do not split as would be expected if the slow exchange limit were reached to give rise to nine separate signals, indicative of the non-equivalence of the three pyrazole rings of the Tp ligand.Apart from broadening and shifting to d 20.63, lowering the temperature causes no eVect on the high-field phosphine proton resonance. Apparently, this spectral behaviour might be consistent with a formulation as a k2- Tp complex with “agostic” interaction. However, no reduced 1JCH coupling constants were observed in the proton-coupled 13C NMR spectrum, as would be expected if “agostic” hydrogen atoms were present. If instead we assume that 1 is just a sterically crowded six-co-ordinate complex, the unusual chemical shift observed might be caused by anisotropy due to the fact that some of the isopropyl hydrogen atoms of the phosphine, as result of the steric pressure, are forced near the magnetic ring current influence of a pyrazole ring.The inequivalent phosphines at low temperature may arise from freezing out conformers around single M–P and P–C bonds. No crystals of 1 suitable for X-ray diVraction could be obtained.However, crystals of the aqua complex [RuTp(H2O)(PPri 2Me)2][CF3SO3]? EtOH 3 were serendipitously obtained during an attempt to prepare the co-ordinatively unsaturated complex [RuTp(PPri 2- Me)2][CF3SO3] by reaction of 1 with AgO3SCF3 in EtOH, and its structure was determined. A structural ORTEP view of complex 3 is shown in Fig. 2. Selected bond lengths and angles are listed in Table 2. Com- Fig. 1 The NMR spectra of complex 1 at 183 K (CD2Cl2): (A) 31P-{1H}, (B) 1H.The peak marked with * corresponds to the solvent. N N N N N N N N N N N N B A H C H P P Cl B Ru P P Cl B Ru HJ. Chem. Soc., Dalton Trans., 1998, 3601–3607 3605 pound 3, which also shows a high field phosphine proton resonance at d 20.23, has a six-coordinate distorted octahedral structure, with phosphines in a cisoid arrangement showing no evidence for “agostic” interaction with the metal. The water molecule appears bound to ruthenium, and linked to an oxygen atom of the [CF3SO3]2 anion via a hydrogen bond.The OH group of the ethanol solvate also forms a hydrogen bond with the co-ordinated water molecule. The hydrogen bond distances O? ? ? O are 2.59(1) and 2.82(1) Å. Strong hydrogen bonds between water and the [CF3SO3]2 anion have also been observed in the recently reported crystal structures of [Ru- Tp(H2O){Ph2P(CH2)2NMe2}][CF3SO3]?0.5CH2Cl2 3 and [Ru- Tp(H2O)(COD)][CF3SO3],2 which consist of neutral dimeric units linked by hydrogen bonds.In these compounds the hydrogen bond O ? ? ? O distances range from 2.714 Å to 2.966 Å, comparing well with the O(1) ? ? ? O(2) separation, although the O(1) ? ? ? O(5) bond distance of 2.59(1) Å is indicative of a much stronger hydrogen bond between the water ligand and the ethanol solvate molecule. The Ru(1)–O(1) bond distance 2.142(6) Å is similar to the Ru–O separations observed for [RuTp- (H2O){Ph2P(CH2)2NMe2}][CF3SO3]?0.5CH2Cl2 [2.190(2) Å] 3 and [RuTp(H2O)(COD)][CF3SO3] [ 2.151(4) Å],2 and also for the complex [Ru{HC(pz)3}(H2O)3][p-MeC6H4SO3]2?1.5 H2O [2.131(1) Å; HC(pz)3 = tris(pyrazolyl)methane)].28 The angle P(1)–Ru(1)–P(2) of 98.54(10)8 is significantly larger than the values of ca. 858 found for complexes containing the bidentate phosphine dippe, such as [RuTp{]] C(OMe)CH2CO2Me}- (dippe)][BPh4] 17 or [RuTp(H2)(dippe)][BPh4],15 since the phos- Table 1 Summary of data for the crystal structure analysis of complexes 3 and 6 Formula M Crystal size/mm Crystal system Space group a/Å b/Å c/Å b/8 U/Å3 Z Dc/g cm23 m(Mo-Ka)/cm21 F(000) Unique reflections Observed reflections (I > 3sI) Number of parameters RR 9 (w = sF 22) 3 C26H52N6BF3O5P2RuS 791.62 0.32 × 0.25 × 0.14 Monoclinic P21/c (no. 14) 14.337(5) 22.077(6) 13.204(6) 90.44(2) 3734(1) 4 1.408 6.03 1648 6012 3434 406 0.062 0.075 6 C45H60B2N8P2Ru 897.66 0.40 × 0.36 × 0.22 Orthorhombic P212121 (no. 19) 15.908(4) 11.550(3) 22.555(4) 4637(2) 4 1.286 4.37 1880 3754 1998 338 0.066 0.077 Table 2 Selected bond distances (Å) and angles (8) for [RuTp(H2O)- (PPri 2Me)2][CF3SO3]?EtOH Ru(1)–P(1) Ru(1)–P(2) Ru(1)–O(1) Ru(1)–N(12) Ru(1)–N(22) Ru(1)–N(32) O(1) ? ? ? O(2) P(1)–Ru(1)–P(2) P(1)–Ru(1)–O(1) P(1)–Ru(1)–N(12) P(1)–Ru(1)–N(22) P(1)–Ru(1)–N(32) P(2)–Ru(1)–O(1) P(2)–Ru(1)–N(12) P(2)–Ru(1)–N(22) P(2)–Ru(1)–N(32) 2.342(3) 2.362(3) 2.142(6) 2.049(7) 2.147(8) 2.140(7) 2.82(1) 98.54(10) 92.3(2) 92.6(2) 170.8(2) 91.5(2) 96.1(2) 92.1(2) 90.5(2) 170.0(2) O(1) ? ? ? O(5) O(1)–H(1) O(1)–H(2) O(5)–H(3) O(1)–H(3) O(2)–H(1) O(1)–Ru(1)–N(12) O(1)–Ru(1)–N(22) O(1)–Ru(1)–N(32) N(12)–Ru(1)–N(22) N(12)–Ru(1)–N(32) N(22)–Ru(1)–N(32) Ru(1)–O(1)–H(1) Ru(1)–O(1)–H(2) H(1)–O(1)–H(2) 2.59(1) 1.21 0.95 0.97 1.70 1.83 169.7(3) 85.1(3) 83.6(3) 88.6(3) 87.3(3) 79.4(3) 120.6 110.0 88.4 phine ligands are monodentate in 3 and do not have the “bite angle” imposed by the backbone ethane chain in dippe.In this fashion the steric repulsions between the two PPri 2Me are minimized, but this forces some of the methyl groups on the isopropyl substituents [C(11) and C(3)] to move towards the gap between two pyrazole rings.As consequence, the hydrogen atoms attached to these methyl groups fall into the the magnetic ring current influence of the pyrazole rings, giving rise to the anomalous high field chemical shift observed for these protons in the 1H NMR spectra. From the crystal structure of 3 it is clear that two bulky PPri 2Me can appear simultaneously bound to a TpRu centre, and therefore the absence of known complexes of the type [Ru(C5Me5)Cl(PR3)2] (PR3 = bulky phosphine ligand: PPri 3, PCy3, PPri 2Me, PPri 2Ph etc.) might be due to electronic rather than to steric reasons.24,25 The aqua complex [RuTp(H2O)(PEt3)2][BPh4] 4 was also obtained by chloride abstraction from 2 under argon in the presence of water, using AgO3SCF3.The water protons in 3 and 4 appear respectively at d 3.1 and 5.30 respectively.If the chloride abstraction from complexes 1 and 2 is performed under dinitrogen instead of argon, then the dinitrogen adducts [RuTp(N2)(PPri 2Me)2]1 and [RuTp(N2)(PEt3)2]1 are obtained, which were isolated as their respective tetraphenylborate salts 5 and 6. These compounds display one strong n(N2) band at 2159 and 2163 cm21 respectively. Since our initial report of the synthesis of the dinitrogen complex [RuTp(N2)(dippe)][BPh4] [n(N2) 2165 cm21],15 two other TpRu dinitrogen adducts have been described: [RuTp(N2){Ph2P(CH2)2NMe2}][CF3SO3] [n(N2) 2182 cm21] 3 and [RuTp(N2)(PPh3)2][BF4] [n(N2) 2177 cm21],12 which suggest that, at variance with their cyclopentadienyl or pentamethylcyclopentadienyl homologues,18a TpRu dinitrogen complexes seem to be a well established class of compounds.The crystal structure of 6 was determined. An ORTEP view of the complex cation is shown in Fig. 3. Selected bond lengths and angles are listed in Table 3. As for compound 3, the coordination around the Ru atom is distorted octahedral.The dinitrogen ligand is bound in the end-on manner, with a Ru(1)– N(1)–N(2) angle of 166(3)8. The Ru(1)–N(1) and N(1)–N(2) separations are 1.91(2) and 1.01(2) Å respectively, which are fully consistent with the dimensions obtained for other ruthenium dinitrogen complexes including [RuTp(N2){Ph2P- (CH2)2NMe2}][CF3SO3] (Ru–N 1.943(4), N–N 1.097(5) Å].3 As Fig. 2 An ORTEP drawing of the compound [RuTp(H2O)(PPri 2- Me)2][CF3SO3]?EtOH with 50% probability thermal ellipsoids.Hydrogen atoms, except those of the water ligand, are omitted.3606 J. Chem. Soc., Dalton Trans., 1998, 3601–3607 in most other cases, the observed N(1)–N(2) bond distance is identical within the experimental error to that of the free N2 molecule, a fact which in the case of [RuTp(N2){Ph2P(CH2)2- NMe2}]1 has been interpreted in terms of a delicate compensatory influence of s-bond strengthening and p-bond weakening in the N2 molecule, as inferred from extended Hückel molecular orbital (EHMO) calculations.3 This also accounts for the high frequency at which the n(N2) IR band appears in these complexes.The dinitrogen ligand in 5 and 6 is labile, and easily replaceable by good neutral donors such as CNBut, leading to the cationic complexes [RuTp(CNBut)(PPri 2Me)2][BPh4] 7 [n(CN) 2124 cm21] and [RuTp(CNBut)(PEt3)2][BPh4] 8 [n(CN) 2123 cm21], which have octahedral structures as inferred from spectral data and do not require further comment.As a consequence of its bulkiness, PPri 2Me in complex 1 is substitutionally labile, at variance with PEt3 in 2. Thus, 1 reacts smoothly with neutral donors such as MeCN or CNBut furnishing the neutral complexes [RuTp(Cl)L(PPri 2Me)] (L = MeCN 9 or CNBut 10). These compounds display strong bands in their respective IR spectra at 2269 and 2245 cm21, and at 2071 and 2099 cm21, attributable to n(CN) in the ligands MeCN and CNBut.The three pyrazole rings of the Tp ligand in these species are inequivalent, and hence nine separate proton and carbon resonances appear in their 1H and 13C-{1H} NMR spectra. The characteristic high-field phosphine proton resonance observed for compounds of the type [RuTp(X)(PPri 2Me)2] is absent in the 1H NMR spectra of 9 and 10. The steric crowding around the metal decreases upon replacement of one of the Fig. 3 An ORTEP drawing of the cation [RuTp(N2)(PEt3)2]1 with 50% probability thermal ellipsoids.Hydrogen atoms are omitted. Table 3 Selected bond distances (Å) and angles (8) for [RuTp(N2)- (PEt3)2][BPh4] Ru(1)–P(1) Ru(1)–P(2) Ru(1)–N(1) Ru(1)–N(12) P(1)–Ru(1)–P(2) P(1)–Ru(1)–N(1) P(1)–Ru(1)–N(12) P(1)–Ru(1)–N(22) P(1)–Ru(1)–N(32) P(2)–Ru(1)–N(1) P(2)–Ru(1)–N(12) P(2)–Ru(1)–N(22) 2.362(5) 2.365(6) 1.91(2) 2.17(1) 99.8(2) 91.8(7) 90.7(5) 172.7(5) 91.6(5) 91.2(9) 169.3(5) 87.5(5) Ru(1)–N(22) Ru(1)–N(32) N(1)–N(2) P(2)–Ru(1)–N(32) N(1)–Ru(1)–N(12) N(1)–Ru(1)–N(22) N(1)–Ru(1)–N(32) N(12)–Ru(1)–N(22) N(12)–Ru(1)–N(32) N(22)–Ru(1)–N(32) Ru(1)–N(1)–N(2) 2.15(2) 2.07(1) 1.01(2) 88.9(5) 90(1) 87.5(9) 176.5(9) 82.0(7) 88.7(7) 89.0(7) 166(3) PPri 2Me by a less bulky ligand, so the methyl protons of the isopropyl substituents are not forced any more to remain under the magnetic ring current influence of the pyrazole rings.Triethylphosphine also displaces one PPri 2Me from 1, aVording the mixed phosphine derivative [RuTp(Cl)(PPri 2Me)(PEt3)] 11, which is characterized by the presence of two doublets in its 31P-{1H} NMR spectrum corresponding to an AB spin system, as expected. In this particular compound the substitution of one PPri 2Me by one PEt3 does not relieve the steric pressure and the high field Pri proton resonance is still observed in the 1H NMR spectrum.Other compounds of the type [RuTp(Cl)L(PR3)] (L = MeCN, py, CO, P(OMe)3 or PMe3; PR3 = PPh3 or PCy3) have been described recently, these being prepared either by displacement of DMF from [RuTp(Cl)(DMF)(PPh3)] by L,1 or by reaction of L with the ruthenium(III) complex [RuTp- (Cl)(OR)(PCy3)] (R = Me or Et).9 Synthesis and characterization of hydride complexes Complex 1 and 2 reacted with NaBH4 in MeOH aVording the neutral monohydride complexes [RuTp(H)(PPri 2Me)2] 12 and [RuTp(H)(PEt3)2] 13.These air-sensitive compounds display one strong n(RuH) IR band near 1915 cm21, and one triplet in their 1H NMR spectra attributable to the hydride proton.As result of the smaller size of hydrogen compared to chloride or other atoms, there is less steric pressure in the hydride complexes containing two PPri 2Me ligands than in other [RuTp(X)(PPri 2Me)2] derivatives, and therefore all phosphine proton resonances in the 1H NMR spectrum of 12 have positive chemical shifts relative to tetramethylsilane. Both complexes 12 and 13 are protonated by HBF4?OEt2 furnishing the dihydrogen adducts [RuTp(H2)(PPri 2Me)2]1 and [RuTp(H2)(PEt3)2]1 which were isolated as their tetraphenylborate salts 14 and 15.The dihydrogen ligand in these complexes is characterized by a broad resonance in the corresponding 1H NMR spectra, having short minimum longitudinal relaxation times (T1)min of 16 and 18 ms respectively (400 MHz). The coupling constants 1JHD observed for the isotopomers [RuTp(HD)(PPri 2Me)2]1 and [RuTp(HD)(PEt3)2]1 are 31.1 and 30.9 Hz. These values of (T1)min and 1JHD compare well with those previously found for other cationic TpRu dihydrogen complexes such as [RuTp(H2)(dippe)][BPh4],15 [RuTp(H2)(PPh3)2][BF4],12 and also [RuTp(H2)(CO)(PPri 3)]- [BF4],26 being consistent with the presence of a co-ordinated dihydrogen molecule. From the values of (T1)min, a separation d(H–H) of 0.95 Å for both 14 and 15 has been estimated, assuming fast spinning of the dihydrogen ligand.Complexes 14 and 15 are white, crystalline solids, indefinitely stable at room temperature under argon.Under dinitrogen the dihydrogen ligand is displaced very slowly by N2 to yield the corresponding dinitrogen complex 5 or 6. As for other cationic TpRu derivatives, equilibrium or irreversible tautomerization to the ruthenium(IV) dihydride form has not been observed, this being attributed to the particular electron donating capabilities of the Tp group (e.g. in comparison with those of the formally related C5H5 or C5Me5 ligands), as well as to the fact that this ligand favours six- over seven-co-ordinate species.12,15,26 We attempted to prepare the neutral monohydride complex [RuTp(H)(MeCN)(PPri 2Me)] by treatment of 9 with NaBH4 in MeOH.However, in the course of the reaction the MeCN ligand was lost, and the ultimate product obtained turned out to be the hydride(dihydrogen) complex [RuTp(H)(H2)(PPri 2Me)] 16, which was isolated in the form of pale yellow crystals. This compound displays one strong IR band at 1946 cm21 attributable to n(RuH).The 1H NMR spectrum shows one doublet at d 210.45 (2JHP = 19.6 Hz, 3 H) corresponding to the hydridic protons. No decoalescence of this resonance is observed as the temperature is lowered, suggesting that there is rapid atom exchange between hydride and dihydrogen sites, leading to the equivalence of these protons.J. Chem. Soc., Dalton Trans., 1998, 3601–3607 3607 Only three rather broad proton and carbon pyrazole ring resonances are observed for complex 16 even at low temperature, indicative of the equivalence of the three pyrazole rings of the Tp ligand due to fluxional behaviour.Accordingly, the 31P-{1H} NMR spectrum shows one singlet. The value of (T1)min for the hydride resonance is 40.8 ms (208 K, 400 MHz). As consequence of rapid chemical exchange with the terminal hydride, this relaxation time is averaged, but in agreement with the presence of one h2-H2 ligand within the complex. Complex 16 undergoes H–D exchange with CD3OD, leading to the isotopomers [RuTp(H)(HD)(PPri 2Me)] and [RuTp(D)(HD)- (PPri 2Me)].A 1JHD coupling constant of 7.4 Hz is observed, which is also averaged. Assuming a rapid hydride–dihydrogen fluxionality in a MH(H2) system, 1JHD of the dihydrogen ligand is equal to 3 times the observed 1JHD coupling constant,14 and hence in our case turns out to be 22.2 Hz, a value which is typical for dihydrogen co-ordinated to a transition metal. The values of (T1)min and 1JHD for 16 are very similar to those found for the compounds [RuTp(H)(H2)(PR3)] (PR3 = PCy3 13 or PPh3 29).The derivatives [RuTp*(H)(H2)(PCy3)] [Tp* = tris(3,5- dimethylpyrazolyl)hydroborate or tris(4-bromo-3-isopropylpyrazolyl) hydroborate] have also been described,14 and their spectral properties match those of 16. It is interesting that the compounds [RuTp*(H)(H2)(PCy3)] 14 and [RuTp(H)(H2)- (PPh3)] 29 have been obtained by hydrogenation of suitable precursor complexes, using pressures of H2 ranging from ca. 3 (PR3 = PCy3) to 6–40 atm (PR3 = PPh3). In our case the hydride–dihydrogen complex is formed smoothly just by reaction of 9 with NaBH4 in MeOH, the use of a H2 atmosphere not being required. Under similar conditions, but with longer reaction times, the reaction of [RuTp(Cl)(MeCN)(PPh3)] with NaBH4 in MeOH leads to the hydrido carbonyl complex [Ru- Tp(H)(CO)(PPh3)]. In general, the reaction of [RuTp(H)- (MeCN)(PPh3)] with NaBH4 and alcohols RCH2OH (R = Me, Et, Ph or tolyl) yields alkyl or aryl TpRu carbonyl derivatives [RuTp(R)(CO)(PPh3)] resulting from the decarbonylation of the alcohol.29 However, we have not detected similar species so far in our system.The chemical reactivity of hydride complexes 12–16 is currently being investigated in detail. In a forthcoming paper we will describe stoichiometric and catalytic C–C coupling reactions in 1-alkynes mediated by the complexes described in this work.N N N N N N N N N N N N N N N N N N H MeiPr2P Ru B H H H H MeiPr2P Ru B H H H H MeiPr2P Ru B H H H Acknowledgements We wish to thank the Ministerio de Educación y Ciencia of Spain (Dirección General de Investigación Científica y Técnica, Project PB94 -1306) and Junta de Andalucía (PAI-FQM 0188) for financial support. References 1 C. Gemel, G. Trimmel, C. Slugovc, S. Kremel, K. Mereiter, R. Schmid and K. Kirchner, Organometallics, 1996, 15, 3998. 2 C. Gemel, P. Wiede, K. Mereiter, V.N. Sapunov, R. Schmid and K. Kirchner, J. Chem. Soc., Dalton Trans., 1996, 4071. 3 G. Trimmel, C. Slugovc, P. Wiede, K. Mereiter, V. N. Sapunov, R. Schmid and K. Kirchner, Inorg. Chem., 1997, 36, 1076. 4 C. Slugovc, P. Wiede, K. Mereiter, R. Schmid and K. Kirchner, Organometallics, 1997, 16, 2768. 5 K. Kirchner, personal communication. 6 F. A. Jalón, A. Otero and A. Rodríguez, J. Chem. Soc., Dalton Trans., 1995, 1629. 7 C. Slugovc, D. Doberer, C. Gemel, R. Schmid, K. Kirchner, B.Winkler and F. Stelzer, Monatsh. Chem., 1998, 129, 221. 8 N. W. Alcock, I. D. Burns, K. S. Claire and A. F. Hill, Inorg. Chem., 1992, 31, 2906. 9 C. Gemel, G. Kickelbick, R. Schmid and K. Kirchner, J. Chem. Soc., Dalton Trans., 1997, 2113. 10 C. Slugovc, V. Sapunov, P. Wiede, K. Mereiter, R. Schmid and K. Kirchner, J. Chem. Soc., Dalton Trans., 1997, 4209. 11 N.-Y. Sun and S. J. Simpson, J. Organomet. Chem., 1992, 434, 341. 12 W.-C. Chan, C.-P. Lau, Y.-Z. Chen, Y.-Q. Fang, S.-M. Ng and G. Jia, Organometallics, 1997, 16, 34. 13 M. A. Halcrow, B. Chaudret and S. Trofimenko, J. Chem. Soc., Chem. Commun., 1993, 465. 14 B. Moreno, S. Sabo-Etienne, B. Chaudret, A. Rodríguez-Fernández, F. A. Jalón and S. Trofimenko, J. Am. Chem. Soc., 1995, 117, 7441. 15 M. Jiménez-Tenorio, M. A. Jiménez Tenorio, M. C. Puerta and P. Valerga, Inorg. Chim. Acta, 1997, 259, 77. 16 C. Slugovc, K. Mereiter, E. Zobetz, R. Schmid and K. Kirchner, Organometallics, 1996, 15, 5275. 17 M. Jiménez-Tenorio, M. A. Jiménez Tenorio, M. C. Puerta and P. Valerga, Organometallics, 1997, 16, 5528. 18 (a) I. de los Ríos, M. Jiménez Tenorio, J. Padilla, M. C. Puerta and P. Valerga, Organometallics, 1996, 15, 4565; (b) I. de los Ríos, M. Jiménez Tenorio, M. C. Puerta and P. Valerga, J. Am. Chem. Soc., 1997, 119, 6529 and refs. therein. 19 S. Trofimenko, Inorg. Synth., 1970, 12, 99. 20 P. T. Beurskens, DIRDIF, Technical Report 1984/1, Crystallography Laboratory, Toernooiveld, 1984. 21 TEXSAN, Single-Crystal Structure Analysis Software, version 5.0, Molecular Structure Coorporation, Houston, TX, 1989. 22 C. K. Johnson, ORTEP, A Thermal Ellipsoid Plotting Program, Oak Ridge National Laboratory, Oak Ridge, TN, 1965. 23 M. Jiménez Tenorio, J. Padilla, M. C. Puerta and P. Valerga, unpublished work. 24 B. K. Campion, R. H. Heyn and T. D. Tilley, J. Chem. Soc., Chem. Commun., 1988, 278; T. Arliguie, C. Border, B. Chaudret, J. Devillers and R. Poilblanc, Organometallics, 1989, 8, 1308. 25 T. J. Johnson, K. Foltig, W. E. Streib, J. D. Martin, J. C. HuVman, S. A. Jackson, O. Eisenstein and K. G. Caulton, Inorg. Chem., 1995, 34, 488. 26 C. Bohanna, M. A. Esteruelas, A. V. Gómez, A. M. López and M. P. Martínez, Organometallics, 1997, 16, 4464. 27 I. D. Burns, A. F. Hill, A. J. P. White, D. J. Williams and J. D. E. T. Wilton-Ely, Organometallics, 1998, 17, 1552. 28 A. Llobet, D. J. Hodgson and T. J. Meyer, Inorg. Chem., 1990, 29, 3760. 29 Y.-Z. Cheng, W.-C. Chan, C.-P. Lau, H.-S. Chu, H.-L. Lee and G. Jia, Organometallics, 1997, 16, 1241. Paper 8/04866E

ISSN:1477-9226    

DOI:10.1039/a804866e

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3609–3614 3609 Synthesis of new heterotridentate ligands comprising mixed hard–soft donor sets, and their complexation with Group 10 metals Pravat Bhattacharyya, Jonathan Parr * and Alexandra M. Z. Slawin Department of Chemistry, Loughborough University, Loughborough, Leics., UK LE11 3TU Received 27th July 1998, Accepted 28th August 1998 The new phosphorus-containing heterotridentate ligands 2-(2-Ph2P)C6H4CH]] NCH(Me)CH(OH)Ph-1S,2R (HL1) and 2-(2-Ph2P)C6H4N]] CHC6H4OH (HL2) were prepared by the condensation of 2-(diphenylphosphino)- benzaldehyde with 1S,2R-norephedrine (HL1) and 2-(diphenylphosphino)aniline with salicylaldehyde (HL2).The co-ordination chemistry of HL1 and HL2 with Group 10 metals was explored, together with that of the previously reported ligand 2-(2-Ph2P)C6H4CH]] NC6H4OH (HL3) which is isomeric with (HL2). Compound HL1 aVords cationic complexes of general formula [M(HL1)Cl]Cl, where M = Ni (1), Pd (2) or Pt (3).Spectroscopic, microanalytical and crystallographic data for 1–3 confirm that deprotonation of the hydroxyl group does not occur on complexation. In contrast, both HL2 and HL3 deprotonate on complexation to form neutral species of general formula [MLCl] where L = L2, M = Ni (4), Pd (5) or Pt (6); L = L3, M = Ni (7), Pd (8) or Pt (9). The crystal structures of 4, 5, 7 and 9 confirm tridentate PNO co-ordination of the deprotonated ligands to the metal centres, forming 5- and 6-membered rings.Over the past fifteen years or so there has been a growing interest in bidentate ligands which combine both hard and soft donor atoms. Typically these are heterofunctionalised phosphines, where the phosphorus is the soft donor, and the hard donor is either an oxygen or nitrogen atom.1 These ligands can exhibit partial lability, where the co-ordination mode alternates between bidentate and monodentate, leading to co-ordinative unsaturation at the metal centre.This is a particularily desirable characteristic for complexes which might have applications in homogeneous catalysis, and since the majority of metals used in such systems are late or middle transition metals it is usually the soft donor which is continually bound to the metal. The incorporation of a chiral centre into such ligands allows enantioselectivity in catalytic transformations mediated by complexes of these ligands.1c,i Furthermore, the unco-ordinated donor atom may be used to bind to a second metal centre, aVording bimetallic species.1e,f There are, however, few tridentate PNE (E = N9, O or S) donor ligands which have been prepared or studied in this regard.For E = O, the known examples can be subdivided into monobasic ligands containing ionisable (OH or NH) groups2–6 and those comprising ether donor groups.7–10 Ruthenium(II) complexes containing ligands in this second category have been shown to be active catalysts for the transfer hydrogenation of ketones in the presence of PriOH.9,10 There are in the literature even fewer examples reported of such ligands where E = N9 or S.11–13 We report here new tridentate ligands providing PNO donor sets which are readily accessible from simple condensation reactions and outline their complexation behaviour with the Group 10 metals.Results and discussion Ligand synthesis Condensation reactions of 2-(diphenylphosphino)benzaldehyde and 1S,2R-norephedrine [PhCH(OH)CH(Me)NH2] or of 2-(diphenylphosphino)aniline with salicylaldehyde, both performed in refluxing thf, lead to new functionalised heterotridentate ligands HL1 and HL2 in good yield.These have been characterised by mass spectrometry, NMR and IR spectroscopies; HL2 is isomeric with the known compound HL3, available from the condensation of 2-(diphenylphosphino)benzaldehyde and 2-aminophenol, the relative positions of the phosphine and aryloxy substituents of the C]] N bonds being reversed.6 Compounds HL1 and HL2 are, in common with HL3, soluble in a range of organic solvents and stable to both aerial oxidation and hydrolysis under ambient conditions.While HL1 is amenable to further reaction without purification, it is necessary to recrystallise HL2 from chloroform–diethyl ether to free the compound from the traces of salicylaldehyde which inevitably persist. In the 1H NMR spectrum of HL1 the imine proton is coupled to the 31P nucleus, J(31P–1H) 4 Hz, which is comparable with the value of 5.5 Hz seen for HL3.In the case of HL2 there is no observed (31P–1H) coupling for the corresponding imine proton. The molecular structure of HL2 has been determined using a crystal grown from chloroform– diethyl ether. Complexation reactions of HL1–3 The stoichiometric reaction of HL1 with either NiCl2?6H2O in ethanol or [M(cod)Cl2], where M = Pd or Pt, in dichloro- PPh2 N HO PPh2 PPh2 N HO HO CH3 N HL1 HL2 HL33610 J.Chem. Soc., Dalton Trans., 1998, 3609–3614 Table 1 Selected spectroscopic (31P-{1H} and 1H NMR, IR and MS) and microanalytical data (calc. values in parentheses) for 1–9. Data for HL1–3 included for comparison Analysis (%) Compound (HL1 1 [Ni(HL1)Cl]Cl 2 [Pd(HL1)Cl]Cl 3 [Pt(HL1)Cl]Cl (HL2 4 [NiL2Cl] 5 [PdL2Cl] 6 [PtL2Cl] (HL3 7 [NiL3Cl] 8 [PdL3Cl] 9 [PtL3Cl] C 60.4 (60.6) 55.9 (55.8) 48.6 (48.8) 63.2 (63.3) 57.3 (57.4) 48.8 (49.1) 63.1 (63.3) 57.8 (57.4) 48.7 (49.1) H 5.1 (4.9) 4.8 (5.7) 4.0 (4.1) 4.0 (4.0) 3.5 (3.7) 2.9 (3.1) 4.4 (4.0) 3.8 (3.7) 3.3 (3.1) N 2.2 (2.5) c 2.1 (2.2) d 1.8 (2.0) d 3.0 (2.9) 2.8 (2.7) 2.3 (2.3) 3.3 (2.9) 2.6 (2.7) 2.2 (2.3) d(31P) [J(PtP)/Hz] a 210.5 — 45.6 9.7 [4042] 214.6 — 47.1 12.9 [3814] 29.0 — 32.6 3.2 [3559] n(CN)/cm21 1638 1631s 1640m 1631m 1614 1608s 1606s 1607s 1624 1582m 1579m 1580m m/z b 516 564 654 489 536 626 474 522 611 d(CH]] N)a 8.62 — 8.36 8.66 8.40) — 8.65 9.03 8.92) — 8.44 8.75 [a]589 e /deg cm3 g21 dm21 11.1) 120.0 167.0 152.7 a No NMR observed for nickel complexes 1, 4 and 7.b M1 2 HCl for complexes 1–3, M1 for 4–9. c As 0.25 EtOH solvate. d As 0.5 EtOH solvate. e In 0.01 M solution in MeOH at 28 8C. methane gives the cationic complexes [ML1Cl]Cl 1–3 respectively in high yield. These complexes are all soluble in ethanol but largely insoluble in other common organic solvents. The equivalent reactions of HL2 and HL3 give the neutral complexes [MLCl] 4–9 which are all highly soluble in chlorinated solvents, thf and acetone. This is supported by elemental analysis and FAB1 mass spectrometry (Table 1), where the parent molecular ions give rise to the most prominent peak in the observed mass spectra.IR spectroscopy Selected IR data for the complexes 1–9 and the ligands HL1–3 are presented in Table 1. For 1–3 the most salient features are the broad bands of medium intensity seen at 2400–2500 cm21, assigned to n(OH), which are absent from the spectra of 4–9.The bands due to n(C]] N) are also readily identified, and are found to vary little from the free HL values for 1–6. For 7–9, however, there is a consistent and significant bathochromic shift of ca. 40 cm21 compared to the value for free HL3. NMR spectroscopy Selected NMR data for complexes 1–9 are collected in Table 1. For the complexes of NiII, 1, 4 and 7, there were no NMR spectra observed, indicating that in solution these samples are paramagnetic.Since complexes of NiII which have a square planar geometry are expected to be diamagnetic, this observation indicates that these complexes are distorted somewhat from their expected geometry. The 31P-{1H} NMR spectra show singlets for the palladium( II) and platinum(II) complexes, with resolved coupling to 195Pt for 3, 6 and 9. The 1J(195Pt–31P) couplings are consistent with a phosphine ligand disposed trans to oxygen, with the 1J values decreasing in the order 3 > 6 > 9.Interestingly there is a ca. 7% increase in 1J for [PtL2Cl] 6 over its isomeric form [PtL3Cl] 9. In the 1H NMR spectra of the palladium(II) and platinum(II) complexes no coupling of the imine proton with the 31P or 195Pt (in the case of 3, 6 and 9) nuclei was observed. The hydroxyl protons of 2 and 3 are not observed in spectra recorded in C2D5OD, which is perhaps unsurprising since there is likely to be rapid exchange of this hydroxyl proton with the (OD) deuteron. Single crystal X-ray diVraction studies The crystal structure of HL2 (Fig. 1) reveals that in the solid state the molecule exhibits the same trans planar configuration of the central C(2)–C(1)–N(1)–C(8) backbone and orientation of P(1) towards O(1) [P(1) ? ? ? O(1) 4.06 Å], which is consistent with that observed in the structure of HL3.6 Additionally for HL2 there is an intramolecular hydrogen-bonding interaction between the imine nitrogen atom N(1) and the phenolic hydrogen H(1o) of O(1) [N(1) ? ? ? H(1o) 1.59 Å, O(1)–H(1o)– N(1) 1558] to form a pseudo six membered ring, a feature not observed in HL3 but common in other related imine–phenol compounds.14 There are small distortions in the pyramidal geometry at P(1), with C–P–C angles between 100.6(1) and 103.1(2)8.The C(1)–N(1) length of 1.282(3) Å in HL2 is marginally longer than the corresponding distance in HL3 [1.269(3) Å],6 in which the substituents on the C]] N bond are reversed. The crystal structures of the Group 10 metal complexes 1–5, 7 and 9 (Figs. 2–4; selected bond lengths and angles are presented in Table 2) reveal in general a square planar geometry at the Fig. 1 Crystal structure of HL2.J. Chem. Soc., Dalton Trans., 1998, 3609–3614 3611 metal atom with tridentate PNO co-ordination of the ligand, where the remaining bound chloride ligand is obligatorily disposed trans to N(1). The complexes 1–3 are cations and the proximity of the unbound chloride Cl(2) to the innocent alcohol oxygen O(21) [Cl(2) ? ? ? O(21) 2.88–2.93 Å, O(21)– H(21) ? ? ? Cl(2) 147–1648] in the crystal lattices suggests a strong hydrogen-bonding interaction with the undissociated hydroxyl proton H(21), corroborating the IR spectral assignments.The complexes 4–9 are neutral owing to deprotonation of the oxygen atom of the tridentate ligand upon complexation, presumably due to the greater acidity of the phenolate protons compared to that of the alcohol. The ligand in complexes 4–6 is isomeric with that in 7–9 in that the relative positions of the six- and five-membered P,N and N,O chelate rings formed on complexation are reversed.Fig. 2 Crystal structure of [Ni(HL1)Cl]Cl 1 (hydrogen atoms omitted for clarity); [Pd(HL1)Cl]Cl 2 and [Pt(HL1)Cl]Cl 3 have the same geometry. Fig. 3 Crystal structure of [NiL2Cl] 4 (hydrogen atoms omitted for clarity); [PdL2Cl] 5 has the same geometry. All of the crystal structures exhibit distortions from idealised square-planar geometry at the metal, due to the bulk of the phosphine group and to the bite angle of the N,O chelate; the trans O–M–P [168.3(2)–176.14(8)8] and trans N–M–Cl [171.15(11)–176.63(6)8] axes are less than 1808 in all of the complexes. The cis P–M–N and cis N–M–O angles vary between the complexes and are seemingly dependent upon the sizes of the P,N and N,O chelate rings.Thus cis N–M–O for the HL1 and (L3)2 complexes 1–3, 7 and 9, which all have fivemembered N,O chelate rings, are in the range 81.6(3)–86.6(3)8, considerably smaller than those seen for 4 and 5, 96.09(7) and 94.02(8)8, which both contain a six-membered N,O ring.Conversely, the cis N–M–P angles exceed 908 for 1–3, 7 and 9, with six membered P,N chelate rings, as compared to 87.92(6) and 86.03(6)8 for 4 and 5. The structures of the nickel complexes 1, 4 and 7 allow comparisons between the bonding parameters for the tridentate ligands HL1, (L2)2 and (L3)2. In 1, 4 and 7 there is little variation among the Ni–N and Ni–P distances (Table 2), suggesting similar basicities for the nitrogen and phosphorus atoms in all of these PNO ligands, although the Ni–Cl distance in 4 is ca. 0.04 Å longer than in either 1 or 7. In [Ni(HL1)Cl]Cl 1 there is an elongation of the Ni–O distance by ca. 0.1 Å compared with 4 and 7 which reflects the combined eVects of O-protonation and the non-phenolic nature of the hydroxyl group in HL1. Similar lengthening by ca. 0.12 Å of the M–O distances in the HL1 complexes compared with those of (L2)2 and (L3)2 is observed with palladium and platinum [Pd–O 2.155(6) in 2, 2.030(2) Å in 5; Pt–O 2.147(6) Å in 3, 2.038(3) Å in 9].The M–N and M–P bond lengths in the palladium and platinum complexes, as seen for nickel, vary little with the phosphine employed, the Pd–N distance in [Pd(HL1)Cl]Cl 2 exceptionally being ca. 0.03 Å shorter than in [PdL2Cl] 5. The complex [NiL2Cl] 4 has the smallest cis O–M–Cl, cis N–M–P and cis P–M–Cl angles of all the nickel complexes [88.54(5), 87.92(6) and 87.65(2)8 respectively], whereas the corresponding values for 1 and 7 are all greater than 908.Conclusion New tridentate PNO donor ligands are readily synthesized from condensation reactions of 2-(diphenylphosphino)-aniline or -benzaldehyde, leading to stable complexes with the Group 10 metals. DiVerences in the acidity of the hydroxyl group and the sizes of the P,N and N,O chelate rings formed upon co-ordination dramatically aVect the co-ordination behaviour and spectroscopic parameters for complexes of HL1–HL3.Fig. 4 Crystal structure of [NiL3Cl] 7 (hydrogen atoms omitted for clarity); [PtL3Cl] 9 has the same geometry.3612 J. Chem. Soc., Dalton Trans., 1998, 3609–3614 Table 2 Selected bond lengths (Å) and angles (8) for complexes 1–5, 7 and 9 (estimated standard deviations in parentheses) M–P M–N M–O M–Cl(1) C]] N C–O O? ? ? Cl(2) N–M–O N–M–P O–M–Cl(1) P–M–Cl(1) O–M–P N–M–Cl(1) C–N–M C–O–M O–H? ? ? Cl(2) 1 2.138(1) 1.877(3) 1.975(2) 2.144(1) 1.297(4) 1.441(4) 2.93 84.67(11) 91.48(9) 91.98(8) 91.80(4) 176.14(8) 171.70(10) 134.1(3) 111.0(2) 164 2 2.190(2) 1.985(7) 2.155(6) 2.267(3) 1.323(9) 1.441(10) 2.89 81.6(3) 90.8(2) 94.3(2) 93.11(9) 172.3(2) 172.5(2) 109.1(5) 108.7(5) 147 3 2.185(2) 2.005(2) 2.147((6) 2.278(3) 1.270(12) 1.469(10) 2.88 82.0(3) 91.3(2) 92.2(2) 94.50(9) 173.2(2) 172.9(2) 106.7(6) 109.6(5) 160 4 2.132(1) 1.892(2) 1.863(2) 2.187(1) 1.311(3) 1.301(3) — 96.09(7) 87.92(6) 88.54(5) 87.65(2) 174.16(5) 174.75(6) 121.1(2) 111.2(6) — 5 2.194(1) 2.016(2) 2.030(2) 2.306(1) 1.442(3) 1.293(3) — 94.02(8) 86.03(6) 88.70(5) 91.40(3) 176.00(6) 176.63(6) 119.1(2) 125.4(2) — 7 2.142(2) 1.900(7) 1.875(5) 2.148(3) 1.260(11) 1.324(10) — 86.3(3) 91.7(2) 90.4(2) 92.25(10) 168.3(2) 175.1(2) 131.0(5) 111.2(6) — 9 2.198(1) 2.003(3) 2.038(3) 2.282(1) 1.289(5) 1.322(5) — 83.02(13) 96.66(11) 89.07(9) 91.65(4) 173.37(8) 171.15(11) 128.1(3) 110.7(3) — Experimental Ligand syntheses and complexation reactions were performed under an atmosphere of oxygen-free nitrogen; thf and dichloromethane were distilled under nitrogen from sodium– benzophenone and calcium hydride respectively, all other solvents were analytical grade and used without further purification.The compounds [M(cod)Cl2] (M = Pt or Pd, cod = cycloocta-1,5-diene),15 2-(diphenylphosphino)aniline 16 and 2-(2-Ph2P)C6H4CH]] NC6H4OH (HL3) 6 were prepared by literature methods; 2-(diphenylphosphino)benzaldehyde (Aldrich) was used as received.The 1H (250.13) and 31P-{1H} (36.21 MHz) NMR spectra, recorded as C2D5OD (1–3) or CDCl3 (4–9) solutions on Bruker AM250 and JEOL FX-90Q spectrometers, and were referenced to external tetramethylsilane (d 0) and 85% phosphoric acid (d 0) respectively using the high-frequency positive convention. Infrared spectra (pressed KBr discs) were recorded on a Perkin-Elmer System 2000 NIR FT-Raman spectrometer. Elemental analyses (Perkin-Elmer 2400 CHN elemental analyser) were performed by the University of Loughborough Analytical Service.Electron impact and fast atom bombardment (positive ionisation mode, 3-nitrobenzyl alcohol matrix) mass spectra were performed by the EPSRC National Mass Spectrometry Service Centre, Swansea, UK. Ligand syntheses 2-(2-Ph2P)C6H4CH]] NCH(Me)CH(OH)Ph-R,S HL1. A thf (20 cm3) solution of 2-(diphenylphosphino)benzaldehyde (0.51 g, 1.75 mmol) and 1S,2R-norephedrine (0.27 g, 1.75 mmol) was heated at reflux for 3 h, giving an orange solution.The solvent was removed under reduced pressure and the resulting oil pumped in vacuo for 30 min to give 0.53 g of HL1 as a pale pink solid. Yield 68% [Found (Calc. for C28H26NOP): C, 79.4 (79.4); H, 6.2 (6.2); N, 3.4 (3.3)%]. d(31P) 210.5 (s). d(1H) 8.62 [1 H, d, J(PH) 4, CH]] N], 7.7–6.7 (m, aryl H), 4.53 [1 H, d, J(HH) 4, CHOH], 3.40 (1 H, m, CHMe), 0.7 [3 H, d, J(HH) 6 Hz, CH3].IR (cm21): n(CN) 1638s. EI: m/z 424 (M1). 2-(2-Ph2PC6H4N]] CH)C6H4OH HL2. A thf solution (30 cm3) of 2-(diphenylphosphino)aniline (0.7 g, 2.5 mmol) and salicylaldehyde (0.33 g, 2.7 mmol) was heated at reflux for 12 h, during which time it became orange. The solvent was removed in vacuo and the solid product obtained recrystallised from chloroform– diethyl ether as 0.69 g of yellow crystals. Yield 67% [Found (Calc. for C25H20NOP): C, 78.4 (78.8); H, 5.2 (5.3); N, 3.3 (3.7)%].d(31P) 214.6 (s). d(1H) 12.50 (1 H, br, OH), 8.40 (1 H, s, CH]] N) and 7.5–6.8 (m, aryl H). IR (cm21): n(CN) 1614s. EI: m/z 381 (M1). Metal complexes 1–9 Complexes 1, 4 and 7 were all prepared according to a general method. An ethanolic (15 cm3) solution of HL (0.1 mmol) and NiCl2?6H2O (0.1 mmol) was stirred for 2 h to give a dark red solution. The solvent was removed in vacuo, the crude product extracted into dichloromethane (ca. 2–3 cm3) and precipitated with hexanes (25 cm3).Complex 1 was recrystallised from ethanol–diethyl ether, 4 and 7 from dichloromethane–diethyl ether. Complexes 2, 3, 5, 6, 8 and 9 were all prepared according to a general method. A dichloromethane (15 cm3) solution of HL (0.1 mmol) and [M(cod)Cl2] (M = Pd or Pt; 0.1 mmol) was stirred for 2 h to give an intensely coloured (yellow to red) solution. The solvent was removed in vacuo, the crude product extracted into dichloromethane (ca. 2–3 cm3) and precipitated with hexanes (25 cm3).Complexes 2 and 3 were recrystallised from ethanol–diethyl ether, 5, 6, 8 and 9 from dichloromethane –hexanes. Isolated yields of microcrystalline complexes 1–9 were typically in the range 25–60%; characterisation data are given in Table 1. X-Ray crystallography The crystal structures of complexes 1–5, 7 and 9 were determined at 298 K using a Siemens SMART diVractometer with graphite-monochromated Mo-Ka radiation (l = 0.71073 Å). The structure of HL2 was recorded on a Rigaku AFC7S instrument with Cu-Ka radiation (l = 1.54178 Å) and w scans.The crystal data, a summary of the data collections and the structure refinements are given in Table 3. All structures were solved by direct methods and all of the non-hydrogen atoms refined with anisotropic displacement parameters; the hydrogen atoms bound to carbon were included in calculated positions (C–H 0.95 Å) with a fixed isotropic displacement parameter. The hydrogen atom H(21) associated with O(21) in 1–3 and H(1o) associated with O(1) in HL2 were located using the diVerence maps and allowed to refine isotropically with no distance restraint. Structural refinements were by full-matrix least-squares methods on F2, calculations being performed using the program SHELXTL PC;17 for HL2, calculations were performed using TEXSAN18 and empirical absorption corrections (DIFABS)19 applied.CCDC reference number 186/1141.J. Chem. Soc., Dalton Trans., 1998, 3609–3614 3613 Table 3 Details of the X-ray data collections and refinements for compounds HL2, 1–5, 7 and 9 Empirical formula M Crystal system Space group a/Å b/Å c/Å a/8 b/8 g/8 U/Å3 Z m/mm21 Total data Unique data (Rint) No.of parameters Goodness of fit on F2 R1, wR2 [I > 2s(I)]a HL2 C25H20NOP 381.41 Monoclinic P21/n 11.038(5) 10.589(4) 16.919(5) 92.97(3) 1975(1) 4 1.327 3390 3136 (0.326) 254 2.87 0.041, 0.026 b 1?0.25EtOH C28.5H27.5Cl2NNiO1.25P 564.60 Orthorhombic P212121 10.513(1) 15.023(1) 19.019(1) 3003.75(10) 4 0.898 18041 6916 (0.1031) 324 0.635 0.0427, 0.0996 2?0.5EtOH C29H29Cl2NO1.5PPd 623.80 Orthorhombic P212121 10.534(1) 15.194(1) 18.805(1) 3010.0(3) 4 0.870 17938 7080 (0.1756) 324 0.818 0.0664, 0.1085 3?0.5EtOH C29H29Cl2NO1.5PPt 712.49 Orthorhombic P212121 10.573(1) 15.220(1) 18.695(1) 3008.34(12) 4 4.918 18619 7083 (0.0710) 323 0.879 0.0503, 0.0895 4 C25H19ClNNiOP 474.54 Monoclinic P21/c 9.702(1) 13.006(1) 17.296(1) 96.14(1) 2169.9(2) 4 1.108 13043 5157 (0.0286) 272 0.655 0.0298, 0.0793 5 C25H19ClNOPPd 522.23 Monoclinic P21/c 9.784(1) 13.054(1) 17.284(1) 96.81(1) 2191.91(8) 4 1.058 12916 5131 (0.0331) 272 0.914 0.0302, 0.0581 7 C25H19ClNNiOP 474.54 Triclinic P1� 10.326(1) 10.518(1) 12.144((1) 83.06(1) 73.58(1) 71.52(1) 1199.31(3) 2 1.002 5161 3353 (0.0745) 272 0.837 0.0849, 0.2390 9 C25H19ClNOPPt 610.92 Monoclinic P21/n 9.085(1) 16.349(1) 14.091(1) 92.36(1) 2091.18(3) 4 6.932 9057 3009 (0.0843) 272 0.602 0.0238, 0.0423 a R1 = S( Fo| 2 |Fc )/S|Fo|, wR2 = [Sw(Fo 2 2 Fc 2)2/Sw(Fo 2)2]� �� ; I > 3s(I) for HL2.b wR2 for HL2 as defined by TEXSAN.3614 J. Chem. Soc., Dalton Trans., 1998, 3609–3614 Acknowledgements We are grateful to Johnson Matthey plc for a loan of precious metals, Dr Steve Aucott for a kind gift of 2-(diphenylphosphino) aniline and to the EPSRC (P. B.) for financial support. References 1 See, for example, (a) G. R. Newkome, Chem. Rev., 1993, 93, 2067; (b) A.Bader and E. Lindner, Coord. Chem. Rev., 1991, 108, 27; (c) K. L. Bray, C. P. Butts. G. C. Lloyd-Jones and M. Murray, J. Chem. Soc., Dalton Trans., 1998, 1421; (d ) R. W. Wegman, A. G. Abatjoglou and A. M. Harrison, J. Chem. Soc., Chem. Commun., 1987, 1891; (e) W. H. Chan, Z. Z. Zhang, T. C. W. Mak and C. M. Che, J. Chem. Soc., Dalton Trans., 1998, 803; ( f ) G. Francio, R. Scopelliti, C. G. Arena, G. Bruno, D. Drommi and F. Faraone, Organometallics, 1998, 17, 338; (g) C.A. Ghilardi, S. Midollini, S. Moneti, A. Orlandini and G. Scappaci, J. Chem. Soc., Dalton Trans., 1992, 3371; (h) A. Del Zotto, G. Nardin and P. Rigo, J. Chem. Soc., Dalton Trans., 1995, 3343; (i) H. Brunner and A. F. M. M. Rahman, Chem. Ber., 1984, 117, 710. 2 H. J. Banbery, W. Hussain, T. A. Hamer, C. J. Jones and J. McCleverty, J. Chem. Soc., Dalton Trans., 1990, 657. 3 K. K. Hii, S. D. Perera and B. L. Shaw, J. Chem. Soc., Dalton Trans., 1994, 3589. 4 K.Mashima, Y. Kaneda, A. Fukumoto, M. Tanaka, K. Tani, H. Nakamo and A. Nakamura, Inorg. Chim. Acta, 1998, 270, 459. 5 J. Andrieu, B. R. Steele, C. G. Sorettas and C. J. Cardin, Organometallics, 1998, 17, 839. 6 J. R. Dilworth, S. D. Howe, A. J. Hutson, J. R. Miller, J. Silver, R. M. Thompson, M. Harman and M. B. Hursthouse, J. Chem. Soc., Dalton Trans., 1994, 3553. 7 K. Tani, M. Yakuta, S. Nakamura and T. Yamagata, J. Chem. Soc., Dalton Trans., 1993, 2781. 8 Y. Kataoka, Y. Tsuji, O. Matsumoto, M. Ohashi, T. Yamagata and K. Tani, J. Chem. Soc., Chem. Commun., 1995, 2099. 9 H. Yang, M. Alvarez, N. Lugan and R. Mathieu, J. Chem. Soc., Chem. Commun., 1995, 1721. 10 H. Yang, M. Alvarez Gressier, N. Lugan and R. Mathieu, Organometallics, 1997, 16, 140. 11 H. A. Ankersmit, N. Veldman, A. L. Spek, K. Vrieze and G. Van Koten, Inorg. Chim. Acta, 1996, 252, 339. 12 P. Wehman, R. E. Rulke, V. E. Kaasjager, P. C. J. Kamer, H. Kooijman, A. L. Spek, C. J. Elsevier, K. Vrieze and P. W. N. M. Van Leeuwen, J. Chem. Soc., Chem. Commun., 1995, 331. 13 S. D. Perera, M. Shamsuddin and B. L. Shaw, Can. J. Chem., 1995, 73, 1010. 14 N. Gündüz, T. Gündüz, M. B. Hursthouse, H. G. Parkes, L. S. Shaw (née Gözen), R. A. Shaw and M. Tüzün, J. Chem. Soc., Perkin Trans. 2, 1985, 899. 15 J. Chatt, L. M. Vallerino and L. M. Venanzi, J. Chem. Soc., 1957, 2496. 16 M. K. Cooper, J. M. Downes, P. A. Duckworth and E. R. T. Tiekink, Aust. J. Chem., 1992, 45, 595. 17 G. M. Sheldrick, SHELXTon 5.03, program for crystal structure solution, University of Göttingen, 1994. 18 TEXSAN, Crystal Structure Analysis Package, Molecular Structures Corporation, The Woodlands, TX, 1985 and 1992. 19 DIFABS, N. G. Walker and D. Stuart, Acta Crystallogr., Sect. A, 1983, 39, 158. Paper 8/05847D

ISSN:1477-9226    

DOI:10.1039/a805847d

出版商:RSC    

年代:1998

数据来源: RSC