摘要:

DALTON PERSPECTIVE J. Chem. Soc., Dalton Trans., 1999, 259–270 259 Structural families in nitride chemistry Duncan H. Gregory Department of Chemistry, The University of Nottingham, Nottingham, UK NG7 2RD. E-Mail: Duncan.Gregory@Nottingham.ac.uk Received 5th October 1998, Accepted 2nd November 1998 Recent progress in the area of nitride chemistry has been rapid and increasingly diverse. Advances in ternary transition metal nitride research have been significant in the last several years to the point where the chemist can begin to correlate these compounds in terms of certain favoured structure types.It is a facet of the often intermediate bonding behaviour in these solids that the structure types so far exhibited are an intriguing mix of covalent (carbide-like), ionic (oxide-like) and unique motifs. While new motifs continue to be discovered with astonishing regularity, an attempt is made to summarise and classify just some of the major emerging ternary nitride structural families. 1 Introduction The study of complex nitride chemistry is still very much in its infancy, with accelerated progress being made only within the last several years. This belies the fact that much of the pioneering work in nitride chemistry was first performed in the ’20s and ’30s and then over a period of several decades by Juza and coworkers, 1,2 among others. As a result, binary nitrides of numerous metals and non-metals are relatively well characterised.The properties of these compounds are often interesting and useful. Many have now found application, for example as high temperature refractory ceramics and coatings (e.g. BN, AlN, TaN, TiN), semiconductors (e.g. GaN, InN), etc. Progress beyond binary systems has, until recently, been hampered primarily by synthetic diYculties, but also by the previous limitations of analytical methods. Advances in the handling of air-sensitive materials together with the huge improvements in diVraction techniques (and associated structure-solution software) are major contributors to new solid state materials output, generally, and nitride output, specifically. Oxides are by far the most numerous, most extensively studied and best characterised group of compounds known to solid state chemists.Despite the relative abundances of oxygen and nitrogen in the Earth’s atmosphere, known oxides outnumber nitrides by orders of magnitude. This contrast and irony has been highlighted in many previous publications and Duncan H.Gregory graduated in Chemistry with Physics at the University of Southampton in 1989, where he remained to obtain a PhD in solid state chemistry in 1993. He took up a position as a postdoctoral research assistant at the University of Nottingham in 1994, where he now holds an EPSRC Advanced Fellowship (since 1997). His current research interests centre around new synthetic routes to predominantly non-oxide materials (principally nitrides) and their structure–property relations.Duncan H. Gregory will not be dwelt on here, although the thermodynamic implications of making and breaking N]] ] N bonds in nitride synthesis are all too apparent.3–6 Furthermore, the relatively immense energy of formation of the N32 anion from atomic N (over three times that of O22 from O) accounts for the reluctance of nitrogen to form predominantly ionic bonds with all but the most electropositive of elements. These thermodynamic factors, therefore, explain both the rarity of nitrides and also their tendency to form unusual and often unique structure types.Previous reviews in the literature have covered the many different aspects of nitride chemistry. We have now reached a point where it is perhaps no longer appropriate or practical to review nitride chemistry as one specific topic. Over the last several years, nitride sub-areas have flourished in their own right, embracing non-metal and nitridosilicate chemistry,7–9 oxynitrides and nitride halides,4,5,10–12 sub-nitrides and nitride clusters 13 and the development of binary nitrides as industrial materials.14,15 This is in addition to the research conducted into the ternary and higher order transition metal nitrides.3–6,16–18 In this respect, this article will concern itself only with the nitride chemistry of the transition metals and the ternary and complex compounds formed with electropositive elements.Earlier discussions of this group of compounds have focused on a summary of crystal chemistry by element or by co-ordination environment.As the number of new nitrides grows, one can begin to delineate broad structural classes. Such classification is, of course, already well established in chalcogenide and halide chemistry 19 and many of the structure types observed in these areas are also seen in nitrides. The tendency to covalency in nitride bonding is also reflected in several carbide structural analogues, although, for example, the formation of dinitrogen bonds in ternary and higher compounds is extremely rare.20 Perhaps most interesting is the existence of a significant number of unique structures (i.e.without chalcogenide, halide or carbide analogues), many of which indicate the existence of several bonding types (ionic, covalent, metallic) within one compound. The eclectic mix of bonding types frequently gives rise to rare and unexpected formal metal oxidation states, borne out by often unusual co-ordination environments to nitrogen.This article considers the crystal chemistry of ternary and higher order transition metal nitrides in terms of the classifi- cation of key structural types. While there are still (and continue to be) many examples of nitride compounds that are unique with respect to their crystal structures (and stoichiometries), as new compositions emerge, so structural trends can begin to be established.Composition–structure–property relationships are not yet well elucidated in such a young area, but with the increased information available from synthetic chemists, theorists, physicists and materials scientists one can envisage this situation changing in the not-too-distant future. The following two sections briefly consider first the synthetic challenges facing the solid state chemist and secondly how one might approach concepts of valence and bonding in nitrides. The main section is devoted to the description of some of the major structural260 J.Chem. Soc., Dalton Trans., 1999, 259–270 families, known properties and the potential for exploiting structure–property relationships to develop new materials. 2 Synthetic approaches Synthetic methodology is being designed to overcome a number of significant challenges. On the one hand, the preparation of nitrides demands high temperatures and long reactions times, as is typical with solid state ceramic materials.However, in many cases the thermodynamic restrictions brought about by the facile formation of N2 from nominally N32 dictate that reactions may need to be designed to prevent decompositions of nitride products, sometimes at relatively low temperatures. Furthermore, the often instantaneous reaction of predominantly ionic higher nitrides with water (or air) to yield hydroxides and ammonia necessitates that reactions and handling of products are performed in inert environments.Preparative techniques are further complicated by matters of stoichiometry and purity. The unusual stoichiometries of many alkaline earth and transition metal binary nitrides prevent synthesis of single phase ternary (or higher) compounds by direct solid state reaction unless additional nitrogen is present. There are several alternatives to using relatively unreactive (at ambient pressure) N2 gas as a source of nitrogen. Besides increasing the nitrogen pressure, solid sources of N such as sodium azide, NaN3, have been successfully used to produce new nitrides often in combination with a sodium flux to encourage crystallinity of nitride products.21 One can move further away from traditional high temperature ceramic approaches by treating molten metals with nitrogen either directly e.g.treating molten alkaline earth metals with transition metals under nitrogen,22 or indirectly by forming a molten alloy of Na and alkaline earth metal.23 This latter approach has been used extensively for producing highly pure, crystalline alkaline earth binary phases and exploits the nonreactivity of sodium with nitrogen to good eVect.24 In these instances the unchanged sodium can be removed by distillation under vacuum or by washing with liquid ammonia. Other non-ceramic-type approaches have focused on employing “softer” conditions, often at lower temperatures, and/or on exploiting thermodynamics by the use of suitable precursors.The ammonolysis of oxide, halide or sulfide precursors has proved a fruitful route to binary and ternary compounds.2,4,25,26 These reactions are often highly temperature-dependent; below a narrow temperature band either no reaction occurs or products are poorly crystalline, above the band ternary phase decomposition is observed. Often an added complication in these reactions is the formation of “partial nitrides” (oxynitrides, nitride halides, etc.) in preference to nitrides.While partial nitriding to oxynitrides etc. is interesting in itself, one has adequately to diVerentiate between N and other anionic species in ammonolysis products, in many cases. As Brese and O’KeeVe 5 point out in their earlier review, there are numerous examples of probable misinterpreted imides, amides and oxynitrides in the literature. Solid-solid phase metathesis reactions have proved useful for synthesizing binary transition metal and lanthanide nitrides.These reactions are often rapid, highly exothermic and selfpropagating. 27,28 Initiation of these reactions can be carried out by various means, including by conventional heating or by microwaves. There are also reports of nitride formation by ignition in air.29 Reaction temperatures in excess of product decomposition temperatures exclude these metathesis routes as means to binary transition metal nitrides of elements to the right hand side of the d block. It has also yet to be used eVectively for ternary nitride synthesis, although the exchange reaction at 400 8C between the existing ternary nitride NaTaN2 and the binary halide CuI yielding the new ternary phase CuTaN2 is an exception.30 However, a modification to the technique, exploiting the vapour pressures of physically separated reactants, has yielded a number of nitride halide phases from the reactions of Li3N with alkaline earth metal halides.31 The fleeting or non-existence of the heavier alkali metal binary nitrides (i.e.those besides lithium nitride) renders formation of the relevant ternary alkali metal–transition metal phases impossible by direct solid state reaction of binary nitrides. Along with ammonolysis of non-nitride precursors, perhaps the most profitable route to these compounds has been via reactions of transition metals with an excess of alkali metal amide melt often at high pressures in the presence of supercritical ammonia.32 These relatively harsh reaction conditions (with pressures often �1 kbar) have produced a variety of compounds, predominantly nitrido-tungstates, -molybdates, -niobates and -tantalates, although other compounds of this type have been prepared under milder conditions starting with the alkali metal and the transition metal nitride.33,34 Other techniques have been employed, sometimes specific to only one or two nitride products.These include Chemical Vapour Transport (CVT) routes to Group IV nitride halides, 12,35,36 ammonolysis of molecular precursors to binary and ternary transition metal nitrides 25,37,38 and the use of transition metal melts [both in an inert (solvent) and reactive capacity] to selected ternary transition metal compounds.39 It seems likely that new routes will continue to be developed, and indeed need to be developed, to access new and diverse materials. 3 Bonding considerations The coexistence of diVerent bonding types within ternary nitrides accounts for the often curious and unexpected coordination behaviour and crystal chemistry observed in these compounds.The bonding within ternary nitrides is by no means well understood. Even in ternary compositions containing high proportions of electropositive elements, such as lithium and alkaline earth metals, metal–nitrogen bonds show significant covalent character. The tendency to covalency is exhibited in binary nitrides of the late transition metals where low valence (I) states are the norm.Even early transition metals preferentially form nitrides below their highest oxidation states. The compound Ti2N is known in several forms and TiN is extremely stable whereas Ti3N4 is poorly characterised. Similarly, Group V elements form many compositions with nitrogen, yet neither V3N5 nor Nb3N5 is known. Many of the transition metal nitrides (and especially those of the early transition metals) are interstitial compounds with structures based on the eutaxy of metal atoms with varying quantities of nitrogen partially filling available sites.This accounts for the many non-stoichiometric transition metal nitride phases in existence and for the essentially metallic properties that are often observed (e.g. hardness, lustre, conductivity, etc.). The metallic nature of bonding in binary nitrides is not restricted to transition metals. Only the most electropositive metals form nitrides which one might regard as classically ionic.The heavier, less electropositive alkaline earth metals preferentially form sub-nitrides. The existence of the expected stoichiometric compounds such as Ba3N2 and Sr3N2 is doubtful whereas “reduced” compositions such as Sr2N40 and Ba3N41 are well characterised. Calcium forms stoichiometric Ca3N2 42 and the sub-nitride Ca2N.43 Only Mg and Be show no evidence for low oxidation state compounds. One can interpret the bonding in the Group II binary nitrides in several ways: first as in sub-oxide and -nitride metal clusters where the normal formal charges are assigned to cations and anions and the excess of charge is attributed to metal valence electrons contributing to metallic bonding, e.g.Ba3N º Ba21 3N32?(3e2).41 Alternatively, one can assign low formal oxidation states to alkaline earth metals and N. Bond valence and Madelung potential calculations tend to favour this latter representation as do density functional theory (DFT) calculations performed on relatedJ.Chem. Soc., Dalton Trans., 1999, 259–270 261 Fig. 1 The broad structure types adopted by binary metal nitrides by element. nitride hydrides (e.g. Sr2NH, Ba2NH).44 These calculations imply that while the charge on the hydride anions in these materials is very close to 21, significant covalent interactions between metal and nitrogen lead to charges of ª11.3 and ª21.7 for cation and anion respectively. The classification of binary metal nitrides is depicted in Fig. 1. Theoretical studies of bonding in ternary and higher nitrides have focused so far on several of the more well characterised groups of compounds.45,46 What is already apparent from calculations performed on these systems and from experimentally observed structures and bond lengths is that covalent bonding contributions, and p interactions especially, are significant in the majority of transition metal compounds. This is often manifested in low co-ordination numbers for transition metals and unexpectedly short bonds to nitrogen.It has been noted previously that cation co-ordination numbers are almost routinely low in metal-rich compositions but become higher as the metal :N ratio decreases.5 One alternative representation of ternary nitride bonding is to consider bonds as s–p hybrid “banana” bonds incorporating dp æÆ pp bonding from filled transition metal d orbitals to empty nitrogen p orbitals.47 This scheme provides for what can essentially be regarded as double and triple metal–nitrogen bonds.Various methods have been employed to model nitride bonding and to test the validity of nitride structures. These include, among others, Madelung potential and bond valence calculations 5 in addition to extended Hückel-type calculations 45 and various aspects of density funconal theory [e.g. local density approximation (LDA)48 and generalized gradient approximation (GGA)44]. The strong covalent character of metal– nitrogen bonds in ternary and higher nitrides often apparently devalues the use of bond valence calculations in validating structural models.While quantitatively bond valence sums are often high for transition metals and low for Group I and II262 J. Chem. Soc., Dalton Trans., 1999, 259–270 metals, this indirectly points to deviations from ionic bonding and becomes a useful qualitative tool. Just as with oxides and fluorides some years ago, as more nitrides continue to be discovered one can envisage the progressive refinement of the relevant bond valence parameters. 4 Structural families Common structural aspects As has been highlighted in previous reviews of this area, the crystal chemistry of the ternary and higher nitrides is often dominated by the low co-ordination environments of nitrogen around transition metals. Ternary transition metal nitrides commonly form low dimensional structures containing isolated complex anions or one-dimensional chains of anionic groups.While the number of examples of layered extended structures in nitrides is steadily increasing, three-dimensional frameworks are largely restricted to non-metal compounds such as nitridosilicates and nitridophosphates. These latter materials will not be covered here, but are evolving a rich and unique crystal chemistry on a par with their oxygen-containing counterparts. 7,8 Transition metal compounds are typically, but not exclusively, metal-rich and with nitrogen frequently coordinated to 5 or 6 metal atoms.Metal co-ordination numbers rarely exceed 6, even in compounds containing large alkaline earth or lanthanide ions. Unusually low metal co-ordination numbers, coupled with the stoichiometric restrictions imposed by the nominal 32 charge of the nitride ion, limit to a certain extent structural analogues based on ternary chalcogenides and halides. Nonetheless, many of these are already known. Furthermore, since nitride anion co-ordination numbers are often higher than those of the cations, anti structures are commonplace in binary and ternary nitride systems alike (e.g.Ca2N43 and Sr2N40 with the anti-CdCl2 structure, CaMg2N2 49 with the anti-La2O3 structure and LiMgN and LiZnN with the anti-fluorite structure 50). Compounds of this nature emphasise the usefulness of the anion-centred polyhedral approach in the structural description of nitrides described by Brese and O’KeeVe.5 Given the above features of metal–nitrogen co-ordination in nitrides, it is not surprising that a significant number of structures are unique to nitrides.There are some examples of carbide and pnictide (e.g. phosphide) isotypes but these are, if anything, less common than chalcogenide and halide analogues. This suggests that, despite the significance of covalency in nitride bonding, the relative electronegativity of N32 (and the lack of available d orbitals for bonding) plays a significant part in determining crystal structure.“Well established” nitride structural groups While, the definition of “well established” is perhaps rather loose in connection to ternary nitride crystal chemistry, there exist several well characterised structural groups which have been known for some years and discussed previously. These will be mentioned briefly here but in most cases no major developments in these areas have occurred recently and the reader will be directed to the appropriate existing reviews of the relevant literature.Much of the fundamental research into the ternary transition metal nitrides of lithium was conducted by Juza and co-workers over several decades. Two significant structural families emerged from this research and broadly these divide the left and right of the top row transition metals. Metals from Ti to Mn (and additionally Nb, Ta, Al, Ga, Mg and Zn) form ordered variants or superstructures of the anti-fluorite structure whereas Co, Ni and Cu form substituted Li3N-type structures (Fig. 2) as solid solutions (Li3 2 xMxN; x £ 0.6).1 Interestingly, the former group of compounds all contain transition metals in high oxidation states (in most cases, the highest possible) while the latter contains metals in the univalent state.These two sets of compounds typify the oxidation state behaviour of the first row transition metals in ternary nitrides. Intriguingly, iron does not fit conveniently into either one of these two structural groups, forming an anti-fluorite superstructure (Li3FeN2; FeIII) 51 and a defect variant of the Li3N-type structure (Li4FeN2; FeII) (Fig. 3).52 More recent studies have shown other structures to exist for ternary nitride compounds of lithium with first row transition metals and some of these are discussed in more detail in later sections. A third class of lithium ternary compounds was first established approximately 25 years ago, crystallising with the anti-La2O3 structure.These Li2MN2 compositions exist for transition metals, M = Zr or Hf (in addition to Ce, Th or U).53 Lithium and M are tetrahedrally and octahedrally co-ordinated to nitrogen respectively. The only ternary alkaline earth metalcontaining example of this structure contains no transition metal, Mg2CaN2.49 Alkaline earth–transition metal ternary phases of this stoichiometry form very diVerent structures, typically with later transition metals in low co-ordination environments. These are covered in more detail in later sections.The perovskite (ABX3) structure and its distorted variants are a common feature in ternary oxide chemistry. Examples of nitride perovskites are rare (although the structure and various distortions are observed in transition metal oxynitrides of the lanthanides or alkaline earth metals). A larger structural group is that of compounds crystallising in the anti-perovskite structure, Ca3XN. To date, however, this class of compounds has been mostly restricted to main group elements,54 with Ca3Au- N22a being the sole transition metal-containing example.Unlike the main group anti-perovskites, however, whose bonding can be represented ionically as Ca21 3X32N32 (where X here is P, As, Sb, Bi, Ge, Sn or Pb), the gold compound can be considered as a sub-nitride (Ca21 3Au2N32?2e2). In bonding terms, the auride nitride has more in common with the sub-nitrides of Group I/ Group II elements (e.g.NaBa3N º Na1Ba21 3N32?4e2) 23a than with the main group anti-perovskite nitrides. There are also earlier reported examples of anti-perovskites Mn3MN (M = Ni, Cu, Zn, Ga, Rh, Ag, Sn, Sb or Pt) (and associated Mn3M1 2 xM9xN phases) based on Mn4N.55 Many of these compounds are either ferrimagnetic or antiferromagnetic with significant hybridisation of the manganese d orbitals and nitrogen p orbitals proposed. Emerging structural families in nitrides The stoichiometry AMN. This stoichiometry exists for combinations of lithium or alkaline earth metals (A) with transition metals (M) but embraces several diVerent structure types, some unique to nitrides.The compound LiZnN forms an ordered anti-fluorite structure, as mentioned above, with Li and Zn distributed over the available tetrahedral sites (LiMgN forms a Fig. 2 Structure of Li3 2 xMxN (M = Co, Ni or Cu): Li, small red spheres; Ni, small green spheres; N, large blue spheres.J. Chem. Soc., Dalton Trans., 1999, 259–270 263 similar cubic cell but with an alternative, disordered, cation distribution).50 Among more recently reported compounds is (Li,Mn)2N which crystallises not in the anti-fluorite structure but in the disordered anti-rutile structure.56 While this is not an uncommon structure type for binary transition metal nitrides, M2N (e.g.e-Ti2N),57 this is the first example of a ternary compound adopting this structure. There is evidence of a range of Li/Mn compositions in this system and as with the earlier reported Li3 2 xMxN compounds this displays the ease with which Li can (partially) replace transition metal ions and form solid solutions. The metals are co-ordinated to nitrogen in distorted triangular planes, a motif seen surprisingly widely in ternary nitrides (as described later).The compound appears to be semiconducting with a possible antiferromagnetic transition at 115 K. The compound LiNiN is stoichiometric and forms a structure related to Li3N.58 In contrast to earlier reported Li–Ni–N compositions, however, the structure is not a simple substituted derivative of lithium nitride, Li3 2 xNixN, but a new lithium vacancy-ordered structure closer in nature to Li4FeN2 52 than Li3N.Unlike the iron nitride, however, metal vacancies lie Fig. 3 The defect variant of the Li3N-type structure formed by Li4FeN2; (a) ball and stick representation with Li as small red spheres, Fe as small purple spheres and N as large blue spheres, (b) perspective plot showing the linking of seven-co-ordinate N-centred polyhedra and [Li2N] and [Fe/h] layers (h = vacancy). within Li2N planes rather than between them, reducing the coordination of nitrogen from hexagonal bipyramidal to trigonal bipyramidal.This would appear to have implications with respect to the Li1 ionic conductivity of the nitride. There is also evidence of other vacancy-ordered compounds in the Li–Ni–N system (Li3 2 x 2 yNixhyN; h = vacancy) which show an evolution of nitrogen co-ordination with vacancy concentration.Alkaline earth metal compositions exist for transition metals in Groups IX, X and XI in a univalent state. The structures appear to show a significant dependence on the size of the alkaline earth metal cation. A common structural motif throughout is the extended –N–M–N– chain. These 1 •[MN2/2 22] infinite chains tend to be straight when the alkaline earth metal cation is small (Ca) and bent as the ionic radius increases (Sr, Ba). The compound CaNiN59 [and also LiSrN 60 and the Lisubstituted compounds, Ca(Ni,Li)N 61] crystallises with the YCoC structure.In this relatively simple, tetragonal structure, layers of straight –N–Ni–N– chains are sandwiched between layers of Ca (Fig. 4). Alternate layers of these chains are aligned perpendicular to one another in the ab plane. (A similar arrangement exists in LiSr2[CoN2], but alternate Co atoms are “replaced” by Li in –N–Li–N–Co–N–Li– chains and hence the linear [CoN2]52 anions become discrete.62) Completely replacing Ca by Sr or Ba causes the –N–Ni–N– chains to bend into zigzag formations and also leads to distortion of the tetrahedral alkaline earth metal–nitrogen environment (Fig. 5).63 The –N–Ni–N– chains in these orthorhombic structures are similar to –O–Cu–O– chains in CsCuO,64 but in the nitrides the “kinks” (changes in direction) in the chains occur less frequently. The same chain conformation is observed for Cu–N in isostructural SrCuN65 [and for Li/Cu in Sr(Li1 2 xCux)N; 0.33 £ x £ 0.52].66 Partial replacement of Ca by Sr is also possible, in Ca1 2 xSrxNiN, with retention of the YCoC structure for x £ 0.5.63a The structure of BaCoN is closely related to BaNiN but here the zigzag chains are “stepped” so that the chain length on each side of a kink is not equal.67 The chains distort further still in the related compound Ba8Ni6N7, twisting Fig. 4 Crystal structure of CaNiN.(The unit cell is delineated by the solid line.)264 J. Chem. Soc., Dalton Trans., 1999, 259–270 in infinite helices.68 Interestingly, the metal–metal distances are significantly shorter in the zigzag chain structures than in the YCoC-type compounds (e.g. Ni–Ni in CaNiN ª 3.5 Å, Ni–Ni in BaNiN ª 2.4 Å). The electrical and magnetic properties of CaNiN, SrNiN and also Ca1 2 xSrxNiN compositions have been studied. All show evidence of metallic conductivity and Pauliparamagnetic behaviour.The stoichiometry AMN2. This stoichiometry is widespread among chalcogenides and is developing rapidly now in nitride chemistry. In many cases, the observed nitride structure types are also those found in chalcogenide chemistry. This broad structural class consists mainly of 2-D layered-type compounds, although the nature of the layers and the pattern in which the layers stack vary with cation size and transition metal type. The stoichiometry already covers a wide range of nitride compositions with combinations of A = alkali metal, alkaline earth metal or transition metal with M = transition metal known.Most commonly, these layered nitrides contain six-coordinate metals in either octahedral or trigonal prismatic geometry. These structures often diVer, principally, in terms of their layer stacking sequence, which leads to predominantly hexagonal unit cells with varying c parameters. The a-NaFeO2 type structure, favoured by many oxides of the AMX2 stoichiometry, is also found in nitrides.This structure is adopted by predominantly 2nd and 3rd row transition metals with alkali and alkaline earth metals including NaTaN2, NaNbN2,32a,33,69 SrZrN2, SrHfN2 70 and CaTaN2.71 The hexagonal a-NaFeO2 structure is an NaCl-type superlattice and both A and M (and N) are octahedrally co-ordinated (Fig. 6). The structure can alternatively be described as an O3 type, where O represents the octahedral co-ordination of A and 3 represents the number of [MX2] layers in the unit cell.Chalcogenides with this structure type often undergo ready (de)intercalation by adding/removing A cations from between [MX2] layers. There is evidence that similar reactions can be performed in these layered nitrides (despite, in some cases, apparent stoichiometric/oxidation state restrictions). The compound Ca1 2 xTaN2 (xª0.26) forms after prolonged heating of CaTaN2 at ª1200 8C.71 Sodium can be removed from NaTaN2 using NO2PF6 to produce nominal Na1 2 xTaN2 compositions with x as high as ª0.9.33 The stoichiometric CaTaN2 compound (and, perhaps, also Ca1 2 xTaN2; x > 0 compositions) appears to superconduct at ca. 9 K. The deintercalated Na1 2 xTaN2 compounds, however, are coloured insulators. The bonding in these latter nitrides has been rationalised in terms of oxidation of N or the formation of N]] ] N bonds, but neither hypothesis has yet been verified. An alternative structure type observed in AMN2 nitrides is the P3 (Na0.6CoO2) structure. This diVers from the a-NaFeO2- Fig. 5 Structure of Ba(Sr)NiN, illustrating the zigzag chains of –N– Ni–N– units. type materials in the trigonal prismatic (P) (as opposed to octahedral) co-ordination of the A cations (Fig. 7). This structure is adopted by LiMoN2 25 and LiWN2 72 and also by the Fig. 6 The a-NaFeO2 structure as adopted by SrZrN2. Alternate layers of octahedrally co-ordinated Sr(A) and Zr(M) are stacked along the c axis.Fig. 7 Structure of LiMoN2 (P3 structure) with alternating layers of metal-centred octahedra and trigonal prisms. Note the edge-sharing of polyhedra within layers but also face-sharing between layers.J. Chem. Soc., Dalton Trans., 1999, 259–270 265 mixed transition metal species CrWN2.38 The compounds CoWN2 and NiWN2 are probably isostructural with these three compounds, although their crystal structures have not been refined.73 The compounds MnMoN2, Li0.84W1.16N2 (both anti-TiP type; P2b type), FeWN2, (Fe0.8Mo0.2)MoN2 and (Fe0.8W0.2)WN2 (P2a) form structures closely related to the P3 structure.The structures of this latter group diVer from the P3 type in (a) the number of layers within the unit cell (hence, smaller c parameters) and (b) the way in which octahedra are linked to trigonal prisms (face-sharing, P2a or edge-sharing, P2b).26a,c,74,75 It is unclear whether MnWN2 falls into the former (P3) or latter (P2) group.73 Up to 64% lithium can be deintercalated from LiMoN2.The fully intercalated compounds LiMoN2 and LiWN2 are metallic and paramagnetic. Electronic structure calculations show LiMoN2 to be a three dimensional metal with strongly covalent [MoN2] sheets. There is also evidence of a direct inter-layer interaction between N atoms in [MoN2] sheets.46a Those of the mixed transition metal species that have been investigated are weakly metallic or semiconducting and paramagnetic.The mixed transition metal compound CuTaN2 is to date the only reported example of a ternary nitride with the delafossite structure.30 This structure type is widely known for oxides AIMIIIO2 (A = Cu, Ag, Pd or Pt; M = Al, Ga, In, Fe, Co, etc.).76 Copper(I) is linearly co-ordinated to N and the hexagonal structure is thus made up of alternating octahedral Ta–N and linear Cu–N layers stacked in the c direction. The heavier alkali metals form non-hexagonal nitride structures with transition metals.The compounds KTaN2 and RbTaN2 prepared at 5 kbar of NH3 pressure form orthorhombic structures apparently isotypic to KGaO2; CsTaN2 prepared under similar conditions crystallises with the cubic b-crystobalite structure 32a (a structure also adopted by LiPN2, for example).77 At ambient pressure, KTaN2 and KNbN2 are reported to form cubic structures which may be analogous to CsTaN2.33 The KCoO2 structure is not a common structure type in oxide chemistry, yet it is reported for several alkaline earth metal–Group IV metal nitrides, BaZrN2, BaHfN2, BaZr1 2 x- HfxN2 and SrTiN2.78 These are tetragonal structures containing layers of edge-sharing [MN2]22 square-based pyramids with alternate pyramids aligned “up” and “down” parallel to the c axis (Fig. 8). The barium compounds are reported to be paramagnetic, although this may be due to impurity phases rather than the ternary nitrides themselves (containing nominally d0 ZrIV or HfIV). The vast array of existing AMO2 structures have been classi- fied in terms of, for example, ion size, ionicity of metal–oxygen Fig. 8 Perspective polyhedral representation of the KCoO2 structure adopted by Ba(Zr,Hf)N2 and SrTiN2. Layers of edge-sharing [MN2]22 square pyramidal anions align perpendicular to the c axis. Green spheres are Ba (Sr). bonds and interlayer distances.79 Similarly detailed studies of nitrides have yet to be performed, although the number of AMN2 compounds is now relatively considerable.It is already apparent that, in many cases, MN2 layers exhibit significant covalent character and that this is likely to have a profound eVect in determining crystal structure. The stoichiometry A2MN3. Two classes of compounds are known with this stoichiometry, one with A = Ce which will be detailed later and the other with A = alkaline earth metals. With respect to transition metals, this latter grouping consists exclusively of Group V elements in their highest oxidation state (M = V, Nb or Ta).These compounds form one of two structures which are closely related and share the common structural motif of infinite one-dimensional chains of corner-sharing [MN3]42 tetrahedral anions (1 •[MN2N2/2]). The compound Ba2VN3 80 (Fig. 9) adopts the orthorhombic Rb2TiO3 structure (also formed by Ca2PN3);81 Sr2MN3 (M = V, Nb or Ta) and Ba2MN3 (M = Nb or Ta) adopt the monoclinically distorted, Ba2ZnO3 structure.78a,80,82 There are no structural data reported for A = Mg, Ca nitrides although there are reports of a green, cubic Ca2TaN3 phase formed at high temperatures (ca. 1200 8C).71 Interestingly, Mg2PN3 is isostructural with Li2SiO3, Fig. 9 Structure of Ba2VN3 showing chains of vertex-sharing [VN3]42 tetrahedra (red and blue spheres are Ba). Fig. 10 Structure of Sr(Ba)3Cr(M)N3. Layers of discrete [CrN3]62 triangles, alternating by 1808 in direction, align perpendicular to the c axis.266 J. Chem. Soc., Dalton Trans., 1999, 259–270 crystallising in a smaller orthorhombic unit cell, with a diVerent tetrahedral anion chain conformation.83 Magnetic measurements of Ba2VN3 and Sr2VN3 indicate diamagnetic behaviour.Measurements performed on Sr2NbN3 and Ba2NbN3 show weak temperature independent paramagnetism, probably from impurities. The stoichiometry A3MN3. There are two distinct classes of nitride with this stoichiometry (313): those with A = alkali metal and those with A = alkaline earth metal.The former class embraces compositions with the heavier Group VI metals (M = Mo or W), the latter with the first row transition metals from M = V to Fe. The nitrides A3MN3 (M = Mo or W) are only known for A = Na and have been successfully synthesized only by reaction of the transition metals (or their nitrides) with sodium amide, NaNH2, and ammonia (either at ambient or high pressures).32c,34,84 The nitridomolybdate and nitridotungstate are isostructural and contain tetrahedral 1 •[MN2N2/2]32 anion chains similar in nature to the 1 •[MVN2N2/2]42 chains in the A2MN3 compounds described above.As in the A2MN3 compounds, there is evidence of significant M–N p bonding in the chains, notably to terminal (non-linking) nitrogens. Partial substitution of sodium in the 313 nitridotungstate by potassium (Na2K[WN3]) or rubidium (Na11Rb[(WN3)4]) leads to new structures also containing 1 •[WN2N2/2]32 anion chains, although the chain conformations are subtly diVerent.32d The above com- Fig. 11 Structure of Ca6Mn(M)N5. Layers of discrete [MN3]62 triangles are sandwiched between [NCa3]31 slabs stacked along the c direction. pounds form only a small part of a growing number of alkali metal nitridotungstates and nitridomolybdates which commonly feature anionic tetrahedral M–N units either as discrete units, oligomeric species, chains, layers or networks. The A = alkaline earth metal 313 compounds crystallise with structures which are unique to nitrides.Two structure types are Fig. 13 Crystal structure of K2NiF4. Fig. 12 Structure of Sr2Li[Fe2N3] containing chains of vertex-linked [Fe2N3]52 “dimeric” anions (N, large blue spheres; Sr, medium red spheres; Li, small yellow spheres; Fe, small green spheres).J. Chem. Soc., Dalton Trans., 1999, 259–270 267 Fig. 14 Crystal structure of Sr(Ba)2ZnN2: (a) ball and stick representation (N, large blue spheres; Sr, medium orange spheres; Zn, small green spheres); (b) polyhedral representation showing edge and vertex linking of N-centred octahedra.known and these are commonly linked by the trigonal planar [MN3]62 anion. The anion is isoelectronic with carbonate [CO3]22, but analogous metal species are very rare in other areas of solid state chemistry (RbNa7[CoO3]2 is an example).85 The calcium ternary compounds Ca3MN3 (M = V, Cr or Mn) have orthorhombic structures containing isolated [MN3]62 anions with C2v symmetry.86 The strontium and barium compounds Sr(Ba)3MN3 (M = Cr, Mn or Fe) have hexagonal structures (Fig. 10) with [MN3]62 anions with D3h symmetry.22b,87 The change in anion symmetry and crystal structure was originally postulated to be a consequence of Jahn–Teller distortion in the low spin [MN3]62 anion. However, recent experimental evidence suggests the changes in symmetry are more a consequence of the size of the counter cation, A21. While the nitrodovanadate Ca3VN3 appears to be essentially intrinsically diamagnetic and insulating, measurements conducted on Ca3CrN3 suggest a paramagnetic, insulating material with an antiferromagnetic transition at 240 K.The exchange energy in Ca3CrN3 is surprisingly large (185 K) given the distance between Cr atoms (ca. 5 Å) and this was attributed to large covalent bonding contributions. Although theoretical investigations tend to confirm the unusual, low spin state of these anions,45a measurements of the magnetic properties of these materials are not extensive. New detailed structural and magnetic studies of some of these materials validate earlier spin state arguments, but the evidence for co-operative magnetic interactions is less certain.88 It is valuable at this point to mention a number of compounds related to these AII 3MN3 compounds, that are more268 J. Chem.Soc., Dalton Trans., 1999, 259–270 usefully described in this context than as stoichiometric groups in their own right. These include Ca6MN5 [M = Mn or Fe (and Ga)] (615),89 A2FeN2 (A = Ca or Sr) 90 and A2LiFe2N3 (A = Sr or Ba).91 The Ca6MN5 compounds have structures in which layers of isolated trigonal planar [MN3]62 anions are stacked alternately along the c axis between [NCa3]31 layers (Fig. 11). These N–Ca layers resemble the edge-sharing octahedral N-Ca layers (“[NCa2]1”) in Ca2N. The three 615 compounds are isostructural, with [MN3]62 anions of D3h symmetry. No detailed magnetic measurements have yet been performed on these materials.The compounds Ca2FeN2 and Sr2FeN2 both contain “dimeric” [Fe2N3]82 units constructed from two distorted trigonal planar units sharing one triangular edge. These units are isosteric with the [In2P4]62 anion. The compound Sr2FeN2 also contains discrete linear [FeIIN2]42 anions which are isoelectronic with CO2. (Similar, but not isoelectronic, discrete [CuIN2]52 units are observed in Sr6Cu3N5 which additionally contains [Cu2N3]72 V-shaped anions.)65 In the lithium compounds A2LiFe2N3 (A = Sr or Ba) Fe is also co-ordinated to N in a trigonal planar geometry, here interlinking to form infinite 1 •[(FeN3/2)2]52 chains.These staggered chains are essentially made up of dimeric [M2N4]-like links joined by the “free” (nonbridging) vertices (Fig. 12). The Fe ? ? ?Fe distances in the A2MN2 and A2LiFe2N3 compounds are both short across “dimers” (ca. 2.4 Å). The physical properties of both groups of materials have yet to be investigated.The recently reported nitridomanganate Li6Ca2[Mn2N6] 48 is also related to the above compounds in that Mn is, again, trigonally co-ordinated to N [as in A3MnN3 (A = Ca, Sr or Ba) and Ca6MnN5]. Importantly, however, the co-ordination sphere of Mn is completed with an unbridged Mn–Mn bond [2.358(1) Å] linking two triangular MnN3 units to form [MnIV 2N6]102 anions. Stoichiometries A2MN2 and A2MN3; relations of the K2NiF4 structure. Whereas compounds of both these stoichiometries have been covered separately above, a convergent structural class “A2MN4 2 y (y = 1 or 2)” exists containing both these stoichiometries and based on the K2NiF4 structure.Compounds with y = 2 exist for A = Ca, Sr or Ba and M = Zn.92 Compounds for y = 1 exist for A = Th, U or Ce and M = Cr or Mn.93 The K2NiF4 structure (Fig. 13) is well known in oxides and halides, with a perovskite-like “core” (interposed with rock-salt type layers) and structural flexibility in terms of possible combinations of substituent cations (or anions).The zinc nitrides A2ZnN2 (A = Ca, Sr or Ba) contain divalent zinc co-ordinated to 2 nitrogen atoms to form isolated, dumbbell [ZnN2]22 anions. In terms of the K2NiF4 structure, Zn is bound to the axial N (F) equivalents of the NiF6 octahedron with all equatorial N (F) positions vacant (Fig. 14). These nitrides are isostructural with Na2HgO2 (a structure also adopted by the non-oxides Na2PdH2 and U2IrC2). The compound Ca2ZnN2 is a diamagnetic insulator.The properties of the strontium and barium compounds have not been studied. Interestingly, BaO rock-salt layers can be inserted into Ba2ZnN2 to create Ba3ZnN2O by treating Ba metal with ZnO, Na and NaN3 at 750 8C.94 This is the first sign that the “building block” methodology used in oxide perovskite chemistry (e.g. in the superconducting cuprates) can be similarly pursued in (oxy)- nitride chemistry. The A2MN3 nitrides contain Mn or Cr co-ordinated to N in an unusual square planar arrangement.Referring to the K2NiF4 “parent” structure, now the axial and two of the four equatorial anion positions of the NiF6 octahedron are occupied and the structure distorts from tetragonal (I4/mmm) to orthorhombic (Immm) (Fig. 15). The squares are vertex linked to form infinite, straight –N–MN2–N–MN2– (1 •[MN2N2/2]) chains parallel to the a axis. Recent studies of the A = Ce compounds suggest that the nitrides are best expressed as MI species, (Ce41)2M1(N32)3, and therefore contain [MN3]82 anions.93b,c Extended Hückel calculations performed on Ce2MnN3 illustrate that the bridging Mn–N bonds in the 1 •[MnN2N2/2] chains are strongly covalent p interactions whereas terminal Mn–N bonds are essentially single bonds.93c The apparent diYculty in synthesizing equivalent Ln31-containing nitrides would also suggest these materials are highly unusual manganese(I) and chromium(I) compounds. The compound Ce2MnN3 is metallic but the properties of the other A2MN3 compounds have not been investigated. 5 Concluding remarks Despite its youth, it is clear that the swelling contemporary momentum in nitride chemistry is driving the subject rapidly into new areas. The unique characteristics of “N32” and the nature of its bonding to metals continue to deliver surprising new compounds with unusual crystal structures and unexpected valence and co-ordination behaviour. As our feeling for nitride crystal chemistry develops, so we can establish the structure–property relationships that are the mainstay of solid state chemistry.Exciting times lie ahead as ternary nitrides begin to show evidence of intriguing and useful physical properties. Control of anion type and stoichiometry may well be crucial in this respect. Perovskite-related oxynitrides already show some of the flexibility that enables structural fine tuning to potentially useful properties. Interstitial nitrides exhibit ferromagnetic behaviour that can be modified via nitrogen stoichiometry.Superconducting lithium-doped, Group IV nitride Fig. 15 Perspective representation of Ce2Mn(M)N3 containing chains of unusual vertex-linked square planar [MN3]82 anions (Ce, medium orange spheres; N, large blue spheres; Mn, small green spheres).J. Chem. Soc., Dalton Trans., 1999, 259–270 269 fluorides demonstrate maximum critical temperatures (Tc) approaching high Tc superconductors. This article has highlighted some of the developing relationships in nitride crystal chemistry. The solid state chemist can usefully apply knowledge of existing structural relationships, for example in chalcogenide and halide chemistry, in his or her quest to establish an understanding of nitrides.What is also obvious, however, is that only by continuing to explore the unknown reaches of nitride chemistry itself will we ultimately gain a suYcient grasp of structure and bonding in these materials.It is both exciting and encouraging that completely new nitride structures and motifs continue to be unearthed with such regularity. Many of these structures have no analogues in solid state chemistry and cannot yet be placed into any kind of structure-type classification. Despite the growing number of such examples, one suspects we have merely scratched the surface of what may yet exist. 6 Acknowledgements The author would like warmly to acknowledge his colleagues of the last several years, Professor Peter P.Edwards and Dr Marten G. Barker for many useful discussions which have contributed to the views expressed herein. He would also like to extend thanks to the numerous other postgraduate and postdoctoral workers involved in various aspects of our work. Finally, the author would like to thank the EPSRC for the award of an Advanced Fellowship and for funding associated research projects. 7 References 1 R. Juza, K. Langer and K. Von Benda, Angew.Chem., Int. Ed. Engl., 1968, 7, 360. 2 See for example: R Juza and H. Hahn, Z. Anorg. Allg. Chem., 1938, 239, 282; R. Juza and W. Sachsze, Z. Anorg. Allg. Chem., 1945, 253, 95; R. Juza, A. Gabel, H. Rabenau and W. Klose, Z. Anorg. Allg. Chem., 1964, 329, 136. 3 F. J. DiSalvo, Science, 1990, 247, 649. 4 R. Marchand, Y. Laurent, J. Guyader, P. L’Haridon and P. Verdier, J. Eur. Ceram. Soc., 1991, 8, 197. 5 N. E. Brese and M. O’KeeVe, Struct. Bonding (Berlin), 1992, 79, 307. 6 F. J. DiSalvo and S. J. Clarke, Curr. Opin. Solid State Mater. Sci., 1996, 1, 241. 7 W. Schnick, Angew. Chem., Int. Ed. Engl., 1993, 32, 806. 8 W. Schnick and H. Huppertz, Chem. Eur. J., 1997, 3, 679. 9 J. V. Badding, Adv. Mater., 1997, 9, 877. 10 See for example, Ch. Wachsmann and H. Jacobs, Z. Anorg. Allg. Chem., 1996, 622, 885. 11 See for example, R. Marx, J. Solid State Chem., 1997, 128, 241. 12 See for example, S. Yamanaka, K. Hotehama and H. Kawaji, Nature (London), 1998, 392, 580. 13 See for example, A.Simon, J. Alloys Comp., 1995, 229, 158. 14 L. E. Toth, Transition Metal Carbides and Nitrides, Academic Press, New York, 1971. 15 See for example, C. C. Wang, S. A. Akbar, W. Chen and V. D. Patton, J. Mater. Sci., 1995, 30, 1627. 16 R. Metselaar, Pure Appl. Chem., 1994, 66, 1815. 17 R. Kniep, Pure Appl. Chem., 1997, 69, 185. 18 H. Jacobs, R. Niewa, T. Sichla, A. Tenten and U. Zachwieja, J. Alloys Comp., 1997, 246, 91. 19 A. F. Wells, Structural Inorganic Chemistry, Oxford University Press, Oxford, 5th edn., 1984. 20 L. LeGendre, R. Marchand and B. Piriou, Eur. J. Inorg. Solid State Chem., 1997, 34, 973. 21 See for example: R. Karam and R. Ward, Inorg. Chem., 1970, 9, 1849; U. Steinbrenner and A. Simon, Angew. Chem., Int. Ed. Engl., 1996, 35, 552; S. J. Clarke, G. R. Kowach and F. J. DiSalvo, Inorg. Chem., 1996, 35, 7009; S. J. Clarke and F. J. DiSalvo, Inorg. Chem., 1997, 36, 1143. 22 See for example: (a) J. Jäger, D. Stahl, P.C. Schmidt and R. Kniep, Angew. Chem., Int. Ed. Engl., 1993, 32, 709; (b) A. Tennstedt, C. Röhr and R. Kniep, Z. Naturforsch., Teil B, 1993, 48, 794. 23 See for example: (a) P. E. Rauch and A. Simon, Angew. Chem., Int. Ed. Engl., 1992, 31, 1519; (b) G. J. Snyder and A. Simon, J. Am. Chem. Soc., 1995, 117, 1996; (c) H. Yamane and F. J. DiSalvo, J. Alloys Comp., 1996, 240, 33; (d ) S. J. Clarke and F. J. DiSalvo, J. Alloys Comp., 1997, 259, 158. 24 See for example, C. C.Addison, R. J. Pulham and E. A. Trevellion, J. Chem. Soc., Dalton Trans., 1975, 2080. 25 S. H. Elder, L. H. Doerrer, F. J. DiSalvo, J. B. Parise, D. Guyomard and J. M. Tarascon, Chem. Mater., 1992, 4, 928. 26 See for example: (a) D. S. Bem and H.-C. zur Loye, J. Solid State Chem., 1993, 104, 467; (b) J. Grins, P.-O. Käll and G. Svensson, J. Mater. Chem., 1995, 5(4), 571; (c) D. S. Bem, C. M. LampeÖnnerud, H. P. Olsen and H.-C. zur Loye, Inorg. Chem., 1996, 35, 581; (d ) P. S.Herle, M. S. Hegde, N. Y. Vasathacharya, S. Philip, M. V. R. Rao and T. Sripathi, J. Solid State Chem., 1997, 134, 120. 27 E. G. Gillan and R. B. Kaner, Inorg. Chem., 1994, 33, 5693; R. E. Treece, E. G. Gillan and R. B. Kaner, Comments Inorg. Chem., 1995, 16, 313; E. G. Gillan and R. B. Kaner, Chem. Mater., 1996, 8, 333. 28 See for example: J. C. Fitzmaurice, A. L. Hector and I. P. Parkin, J. Chem. Soc., Dalton Trans., 1993, 2435; J. C. Fitzmaurice, A. Hector, A. T. Rowley and I.P. Parkin, Polyhedron, 1994, 13, 235; A. L. Hector and I. P. Parkin, Chem. Mater., 1995, 7, 1728; S. Ali, M. D. Aguas, A. L. Hector, G. Henshaw and I. P. Parkin, Polyhedron, 1997, 16, 3635. 29 P. E. Bacon, J. G. Hou, A. W. Sleight and R. L. Nielsen, J. Solid State Chem., 1995, 119, 207. 30 U. Zachwieja and H. Jacobs, Eur. J. Solid State Inorg. Chem., 1991, 28, 1055. 31 P. M. O’Meara, M. G. Barker, D. A. Rees and D. H. Gregory, in preparation. 32 See for example: (a) H. Jacobs and E.von Pinkowski, J. Less- Common Met., 1989, 146, 147; (b) Th. Brokamp and H. Jacobs, J. Alloys Comp., 1991, 176, 47; (c) D. Ostermann, U. Zachwieja and H. Jacobs, J. Alloys Comp., 1992, 190, 137; (d ) R. Niewa and H. Jacobs, J. Alloys Comp., 1996, 233, 61; (e) R. Niewa and H. Jacobs, Z. Anorg. Allg. Chem., 1996, 622, 881. 33 P. E. Rauch and F. J. DiSalvo, J. Solid State Chem., 1992, 100, 160. 34 P. E. Rauch, F. J. DiSalvo, N. E. Brese, D. E. Partin and M. O’KeeVe, J.Solid State Chem., 1994, 110, 162. 35 M. Ohashi, S. Yamanaka, M. Sumihara and M. Hattori, J. Solid State Chem., 1988, 75, 99; M. Ohashi, S. Yamanaka and M. Hattori, J. Solid State Chem., 1988, 77, 342. 36 S. Yamanaka, H. Kawaji, K. Hotehama and M. Ohashi, Adv. Mater., 1996, 8, 771; H. Kawaji, K. Hotehama and S. Yamanaka, Chem. Mater., 1997, 9, 2127. 37 See for example: D. V. Baxter, M. H. Chisholm, G. J. Gama, V. F. DiStasi, A. L. Hector and I. P. Parkin, Chem. Mater., 1996, 8, 1222; N.A. K. Hansen and W. A. Herrmann, Chem. Mater., 1998, 10, 1677. 38 K. S. Weil and P. N. Kumta, J. Solid State Chem., 1997, 128, 185. 39 D. J. Siddons, P. M. O’Meara, M. G. Barker and D. H. Gregory, in preparation. 40 N. E. Brese and M. O’KeeVe, J. Solid State Chem., 1990, 87, 134. 41 U. Steinbrenner and A. Simon, Z. Anorg. Allg. Chem., 1998, 624, 228. 42 Y. Laurent, Y. Lang and M. T. LeBihan, Acta Crystallogr., Sect. B, 1968, 24, 494. 43 E. T. Keve and A. C. Skapski, Inorg.Chem., 1968, 7, 1757. 44 H. Smolinski and W. Weber, J. Phys. Chem. Solids, 1998, 59, 915. 45 (a) K. A. Yee and T. Hughbanks, Inorg. Chem., 1992, 31, 1921; (b) M. T. Green and T. Hughbanks, Inorg. Chem., 1993, 32, 5611. 46 See for example: (a) D. J Singh, Phys. Rev. B, 1992, 46, 9332; (b) L. F. Mattheis, Phys. Rev. B, 1993, 47, 8224. 47 R. B. King, Can. J. Chem., 1995, 73, 963. 48 O. Hochrein, Y. Grin and R. Kniep, Angew. Chem., Int. Ed. Engl., 1998, 37, 1582. 49 V.Schultz-Coulon and W. Schnick, Z. Naturforsch., Teil B, 1995, 50, 619. 50 R. Juza and F. Hund, Z. Anorg. Allg. Chem., 1948, 257, 1. 51 A. Gudat, R. Kniep, A. Rabenau, W. Bronger and U. Reschewitz, J. Less-Common Met., 1990, 161, 31. 52 A. Gudat, R. Kniep and A. Rabenau, Angew. Chem., Int. Ed. Engl., 1991, 30, 199. 53 A.-P. Palisaar and R. Juza, Z. Anorg. Allg. Chem., 1971, 384, 1; M. G. Barker and I. C. Alexander, J. Chem. Soc., Dalton Trans., 1974, 2166; R. Niewa, H. Jacobs and H.M. Mayer, Z. Kristallogr., 1995, 210, 513. 54 M. Y. Chern, D. A. Vennos and F. J. DiSalvo, J. Solid State Chem., 1992, 96, 415. 55 D. Fruchart, E. F. Bertaut, R. Madar, G. Lorthioir and R. Fruchart, Solid State Commun., 1971, 9, 1793; D. Fruchart and E. F. Bertaut, Physica, 1977, 86–88B, 423; J. Phys. Soc. Jpn., 1978, 44, 781. 56 R. Niewa, F. J. DiSalvo, D.-K. Yang, D. B. Zax, H. Luo and W. B. Yelon, J. Alloys Comp., 1998, 266, 32. 57 B. Holmberg, Acta Chem. Scand., 1962, 16, 255.270 J.Chem. Soc., Dalton Trans., 1999, 259–270 58 D. J. Siddons, S. E. Smith, M. G. Barker, A. J. Blake, P. P. Edwards, D. H. Gregory and T. A. Hamor, in preparation. 59 M. Y. Chern and F. J. DiSalvo, J. Solid State Chem., 1990, 88, 459. 60 G. Cordier, A. Gudat, R. Kniep and A. Rabenau, Angew. Chem., Int. Ed. Engl., 1989, 28, 201. 61 A. Gudat, R. Kniep and J. Maier, J. Alloys Comp., 1992, 186, 339. 62 P. Höhn and R. Kniep, Z. Naturforsch., Teil B, 1992, 47, 434. 63 (a) T. Yamamoto, S. Kikkawa and F. Kanamaru, J. Solid State Chem., 1995, 115, 353; (b) A. Gudat, S. Haag, R. Kniep and A. Rabenau, J. Less-Common Met., 1990, 159, L29. 64 H. Klassen and R. Hoppe, Z. Anorg. Allg. Chem., 1983, 497, 70. 65 F. J. DiSalvo, S. S. Trail, H. Yamane and N. E. Brese, J. Alloys Comp., 1997, 255, 122. 66 J. Jäger and R. Kniep, Z. Naturforsch., Teil B, 1992, 47, 1290. 67 A. Tennstedt and R. Kniep, Z. Anorg. Allg. Chem., 1994, 620, 1781. 68 A. Gudat, W. Milius, S.Haag, R. Kniep and A. Rabenau, J. Less- Common Met., 1991, 168, 305. 69 H. Jacobs and B. Hellmann, J. Alloys Comp., 1993, 191, 51. 70 D. H. Gregory, M. G. Barker, P. P. Edwards and D. J. Siddons, Inorg. Chem., 1996, 35, 7608. 71 V. Balbarin, R. B. Van Dover and F. J. DiSalvo, J. Phys. Chem. Solids, 1996, 57, 1919. 72 P. S. Herle, M. S. Hegde, N. Y. Vasanthacharya, J. Gopalakrishnan and G. N. Subbanna, J. Solid State Chem., 1994, 112, 208. 73 P. S. Herle, N. Y. Vasanthacharya, M. S. Hegde and J. Gopalakrishnan, J. Alloys Comp., 1995, 217, 22. 74 D. S. Bem, H. P. Olsen and H.-C. zur Loye, Chem. Mater., 1995, 7, 1824; J. D. Houmes, S. Deo and H.-C. zur Loye, J. Solid State Chem., 1997, 131, 374. 75 S. Kaskel, D. Hohlwein and J. Strähle, J. Solid State Chem., 1998, 138, 154. 76 J.-P. Doumerc, A. Ammar, A. Wichainchai, M. Pouchard and P. Hagenmuller, J. Phys. Chem. Solids, 1987, 48, 37. 77 W. Schnick and J. Lücke, Z. Anorg. Allg. Chem., 1990, 588, 19. 78 (a) O. Seeger, M. Hofmann, J. Strähle, J. P. Laval and B. Frit, Z. Anorg. Allg. Chem., 1994, 620, 2008; (b) D. H. Gregory, M. G. Barker, P. P. Edwards, M. Slaski and D. J. Siddons, J. Solid State Chem., 1998, 137, 62; (c) D. H. Gregory, M. G. Barker, P. P. Edwards and D. J. Siddons, Inorg. Chem., 1998, 37, 3775. 79 See for example: C. Delmas, C. Fouassier and P. Hagenmuller, Mater. Res. Bull., 1976, 11, 1483; C. Delmas, J.-J. Braconnier, C. Fouassier and P. Hagenmuller, Z. Naturforsch., Teil B, 1981, 36, 1368. 80 D. H. Gregory, M. G. Barker, P. P. Edwards and D. J. Siddons, Inorg. Chem., 1995, 34, 3912. 81 W. Schnick and V. Schultz-Coulon, Angew. Chem., Int. Ed. Engl., 1993, 32, 280. 82 F. K.-J. Helmlinger, P. Höhn and R. Kniep, Z. Naturforsch., Teil B, 1993, 48, 1015; X. Z. Chen, H. A. Eick and W. Lasocha, J. Solid State Chem., 1998, 138, 297. 83 V. Schultz-Coulon and W. Schnick, Z. Anorg. Allg. Chem., 1997, 623, 69. 84 H. Jacobs and R. Niewa, Eur. J. Solid State Inorg. Chem., 1994, 31, 105. 85 J. Birx and R. Hoppe, Z. Anorg. Allg. Chem., 1990, 588, 7. 86 D. A. Vennos, M. E. Badding and F. J. DiSalvo, Inorg. Chem., 1990, 29, 4059; D. A. Vennos and F. J. DiSalvo, J. Solid State Chem., 1992, 98, 318; A. Tennstadt, C. Röhr and R. Kniep, Z. Naturforsch., Teil B, 1993, 48, 1831. 87 P. Höhn, R. Kniep and A. Rabenau, Z. Kristallogr., 1991, 196, 153; M. G. Barker, M. J. Begley, P. P. Edwards, D. H. Gregory and S. E. Smith, J. Chem. Soc., Dalton Trans., 1996, 1. 88 D. H. Gregory, R. W. Strange and M. Slaski, in preparation. 89 G. Cordier, P. Höhn, R. Kniep and A. Rabenau, Z. Anorg. Allg. Chem., 1990, 591, 58; D. H. Gregory, M. G. Barker, P. P. Edwards and D. J. Siddons, Inorg. Chem., 1995, 34, 5195. 90 P. Höhn and R. Kniep, Z. Naturforsch., Teil B, 1992, 47, 477. 91 P. Höhn, S. Haag, W. Milius and R. Kniep, Angew. Chem., Int. Ed. Engl., 1991, 30, 831. 92 M. Y. Chern and F. J. DiSalvo, J. Solid State Chem., 1990, 88, 528; H. Yamane and F. J. DiSalvo, J. Solid State Chem., 1995, 119, 375. 93 (a) R. Benz and W. H. Zachariasen, J. Nucl. Mater., 1970, 37, 109; (b) S. Broll and W. Jeitschko, Z. Naturforsch., Teil B, 1995, 50, 905; (c) R. Niewa, F. J. DiSalvo, H. Luo and W. B. Yelon, Z. Naturforsch., Teil B, 1998, 53, 63. 94 H. Yamane and F. J. DiSalvo, J. Alloys Comp., 1996, 234, 203. Paper 8/07732K

ISSN:1477-9226    

DOI:10.1039/a807732k

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1999, 271–272 271 An unprecedented Í2N,H bonding mode for a hydridotris(pyrazolyl)- borato ligand François Malbosc,a Philippe Kalck,*a Jean-Claude Daran b and Michel Etienne *b a Laboratoire de Catalyse, Chimie Fine et Polymères, ENSCT, 118 Route de Narbonne, 31077 Toulouse Cedex 4, France. E-mail: pkalck@ensct.fr b Laboratoire de Chimie de Coordination du CNRS, UPR 8241, 205 Route de Narbonne, 31077 Toulouse Cedex 4, France. E-mail: etienne@lcc-toulouse.fr Received 30th November 1998, Accepted 11th December 1998 An unprecedented Í2N,H bonding mode for a hydridotris(pyrazolyl)borato (TpMe2,4-Cl) ligand is observed in [Rh(CO)(PMePh2)2(TpMe2,4-Cl)]; the complex, which features two dangling pyrazole rings and a B–(Ï-H)–Rh agostic bond, is highly fluxional in solution.Hydridotris(pyrazolyl)borato ligands (Tp9) have a strong preference for adopting a k3N,N9,N0 bonding mode.1 On some occasions, the electronic nature of the metal dictates a different coordination behaviour.A textbook example is the d8 configuration in square-planar Rh complexes where k2N,N9 coordination may also be observed, the two bonding modes interconverting in several cases.2 The third pendant pyrazolyl arm may then fulfill subsequent electronic deficiency: oxidation of [Rh(CO)(PPh3)(k2N,N9-TpMe2)] yields a pentacoordinated k3N,N9,N0 rhodium(II) complex [Rh(CO)(PPh3)(k3N,N9,N0- TpMe2)]1.3 Also, stable species containing k2N,N9 forms have been shown to be intermediates leading to k3N,N9,N0 coordination as exemplified by the thermal loss of phosphine in octahedral [RuH(PPh3)2(k2N,N9-Tp)] which aVords [RuH- (PPh3)(k3N,N9,N0-Tp)].4 Isomerization from a k2N,N9 to a k3N,N9,H situation has been recently observed in [RuH- (COD)Tpi-Pr2] (COD = cycloocta-1,5-diene).5 In this communication, we report yet another bonding mode for a hydridotris- (pyrazolyl)borato ligand featuring two pendant pyrazolyl rings in the complex [Rh(CO)(PMePh2)2(TpMe2,4-Cl)].Only one pyrazolyl ring is N-bound to rhodium, and the B-bound hydrogen agostically interacts with rhodium leading to an unprecedented k2N,H bonding mode. A very recent example of a bridging (m3-1k1N:2k1N9:3k1N0) Tp9 has been reported in a trinuclear silver complex.6 Treatment of [Rh(CO)2(TpMe2,4-Cl)] 1a with one equivalent of PMePh2 leads to [Rh(CO)(PMePh2)(k2N,N9-TpMe2,4-Cl)] 2a [n(CO) = 1996 cm21] in high yield.The k2 bonding mode is ascertained in the solid state by the result of an X-ray diVraction analysis † (Fig. 1). The square-planar arrangement around rhodium is similar to that observed in related k2N,N9-TpMe2 [Rh(CO)(L)(TpMe2)] (L = PMe3,2c PPh3,3 PMePh2 2b 7). A clean reaction converts 1a and excess PMePh2 into [Rh(CO)(PMePh2)2(TpMe2,4-Cl)] 3a‡ [n(CO) = 1979 cm21] at 273 K. At 233 K, a single 31P-{1H} NMR doublet [d 23.7 (d, JPRh = 125 Hz)] and a single 13C-{1H} NMR doublet of triplets (d 190.6, JCRh = 69, JCP = 16 Hz) indicate two equivalent trans phosphines and a cis carbonyl group bound to rhodium.This formulation is confirmed by the result of an X-ray structure determination § (Fig. 2). A single N-bound pyrazolyl ring, trans to the carbonyl, completes the coordination sphere of the rhodium in a slightly distorted square-planar geometry. Distances and angles within the square-plane are in the classical range. A potential axial site is occupied by the X-ray located B-bound hydrogen leading to a somewhat loose agostic B– (m-H)–Rh system.The Rh(1) ? ? ? H(1) distance of 2.35(3) Å seems long as compared to a typical Rh(I)–H bond length of ca. 1.55 Å.8 Also, the low frequency shift of n(BH) to 2350 cm21 (vbr, w) is modest as compared to a n(BH) of 2477 cm21 (sharp, m) in 2a. Thus the agostic interaction could be driven by steric interactions between the unbound pyrazolyl rings and the phenyl groups on phosphorus.9 Indeed, the most peculiar feature of the crystal structure is the presence of two unbound pyrazolyl rings.To our knowledge there is no example of such a bonding mode for a hydridotris(pyrazolyl)borato ligand. The k3N,N9,H bonding mode (i.e. one pendant pyrazolyl ring) has been observed in [RuH(COD)Tpi-Pr2] 5 and [RuMe(TpMe2)- (COD)].10 Also the dihydridobis(3,5-trifluoromethylpyrazolyl)- borato (Bp(CF3)2) is k3N,N9,H in [RuH(COD)(Bp(CF3)2)] (Ru–H 1.43(3) Å, Ru ? ? ? H 1.97 Å) and [RuH(PPh3)2(Bp(CF3)2)], a somewhat classical bonding situation in Bp9 chemistry.11 The bonding mode is k2N,H in [RuH(H2)(Bp(CF3)2)(PCy3)2] [n(BH) = 2514, 2149, 2007 cm21], the change being attributed to the diVerent cone angle of the phosphines.11 Inter- and intra-molecular dynamic processes are observed in solution.First, 3a partially dissociates into 2a and free phosphine (3a/2a ª 4 at 293 K). 31P NMR shows characteristic broadening of the signals of the three species above room temperature, and 2a is converted into 3a upon addition of excess PMe2Ph.Complex 3a is also highly fluxional. In [2H]8-toluene, the 31P NMR doublet broadens below 233 K to ultimately give a doublet of doublets (193 K, d 22.5, 33.2, 1JRhP = 125, 2JPP = 320 Hz). In the 1H NMR spectrum at 183 K, five Fig. 1 Plot of the molecular structure of [Rh(CO)(PMePh2)- (TpMe2,4-Cl)] 2a (30% probability ellipsoids). Relevant bond distances (Å) and angles (8): Rh(1)–P(2) 2.2529(5), Rh(1)–N(1) 2.092(2), Rh(1)–N(3) 2.092(2), Rh(1) ? ? ? N(5) 3.800(2), Rh(1)–C(1) 1.809(2); N(1)–Rh(1)– C(1) 174.80(9), N(3)–Rh(1)–P(2) 172.30(6), N(1)–Rh(1)–N(3) 82.58(7).272 J.Chem. Soc., Dalton Trans., 1999, 271–272 pyrazole methyl signals out of six are well resolved. Between 213 and 273 K, four peaks in a 1:1:2:2 ratio for the pyrazolyl methyls together with an unresolved multiplet for the phosphine methyls are observed which now accounts for a symmetry plane on the NMR time scale.When the temperature is further raised, pyrazolyl methyl signals broaden and merge, first within each set of 1 : 2 peaks to give two very broad peaks, then altogether. The signal of the hydrogen bound to boron remains large in the d 5 region, except below 233 K where it vanishes into the base line, questioning the presence of a strong agostic interaction in solution.5,10,11 These observations are consistent with a low temperature asymmetric structure akin to that observed in the solid state.In the intermediate temperature range, exchange of the unbound pyrazolyl rings occurs via rotation about the B–N bonds. This includes rotation about B(1)–N(2) and opening of the B–(m-H)–Rh interaction. At higher temperatures, we cannot as yet diVerentiate unambiguously mechanisms in which the three pyrazolyl rings interconvert intramolecularly in 3a or intermolecularly via the equilibrium with 2a and free phosphine. Finally the ease with which 3a is formed from 1a and PMePh2 is striking.We have observed that, under comparable conditions, neither 1b nor 2b reacts with PMePh2 to give putative [Rh(CO)(PMePh2)2(TpMe2)]. Pyrazolylborates TpMe2,4-Cl and TpMe2 have similar steric requirements but markedly diVer in their electron withdrawing properties.12 Thus although steric eVects are undoubtedly responsible for the observed structure of 3a, the reduced electron density at the rhodium allows the reaction to occur in the case of 1a only.Fig. 2 Plot of the molecular structure of [Rh(CO)(PMePh2)2- (TpMe2,4-Cl)] 3a (30% probability ellipsoids). Relevant bond distances (Å) and angles (8): Rh(1)–P(1) 2.3273(8), Rh(1)–P(2) 2.3317(8), Rh(1)–N(1) 2.105(3), Rh(1)–C(1) 1.802(2), Rh(1) ? ? ? H(1) 2.35(3); N(1)–Rh(1)– C(1) 178.0(2), P(1)–Rh(1)–P(2) 167.54(3), P(1)–Rh(1)–N(1) 89.97(7), P(2)–Rh(1)–N(1) 90.21(7), Rh(1)–H(1)–B(1) 126(2). Notes and references † Crystal data for 2a: C29H32BCl3N6OPRh, M = 731.7, triclinic, P1� , a = 9.026(1), b = 10.434(2), c = 17.290(2) Å, a = 88.91(2), b = 85.96(2), g = 77.91(2)8, U = 1556.8(3) Å3, Z = 2, m = 8.657 cm21, T = 180(2) K, reflections collected/ique/used: 15694/5810 (Rint = 0.0338)/4794 [I > 2s(I )], 384 parameters, R/Rw 0.0237/0.0273.‡ Preparation of 3a. Addition of PMePh2 (0.335 ml, 1.82 mmol) to a cooled (273 K) pentane solution (30 ml) of [Rh(CO)2(k3-TpMe2,4-Cl)] (0.505 g, 0.90 mmol) yielded an orange precipitate. Recrystallisation from pure pentane at 273 K aVorded orange crystals of the product (0.55 g, 68 mmol, 76%) (Found: C, 53.9; H, 5.0; N, 8.9.C42H45- BCl3N6OP2Rh requires C, 54.1; H, 4.9; N, 9.0%). IR (KBr): n(CO) 1979, n(BH) 2442–2405 (vbr) cm21. NMR (233 K, 400 MHz for 1H, except phenyl resonances, all s unless specified). 1H ([2H]8-toluene): d 2.63 (6 H, C3N2ClMe2), 2.55 (3 H, C3N2ClMe2), 2.24 (3 H, C3N2ClMe2), 2.13 (6 H, C3N2ClMe2), 1.84 (6 H, PPh2Me). 31P-{1H} ([2H]8-toluene): d 23.7 (d, JRhP 125 Hz). 13C-{1H} (CD2Cl2): d 190.6 (dt, RhCO, JCRh 69, JCP 16 Hz), 144.4, 143.9, 143.0, 139.6 (CN2ClC2Me2), 107.9, 106.4 (C2N2Me2CCl), 11.3 (t, MeP, JPC 14 Hz), 12.6, 11.0, 9.8, 8.5 (C3N2ClMe2). 103Rh-{1H} (CD2Cl2): d 344.3 (d, JRhP 125 Hz). § Crystal data for 3a: C42H45BCl3N6OP2Rh, M = 931.88, triclinic, P1� , a = 11.164(2), b = 12.006(2), c = 17.097(3) Å, a = 101.07(2), b = 102.28(2), g = 92.46(2)8, U = 2189.1(6) Å3, Z = 2, m = 6.785 cm21, T = 180(2) K, reflections collected/unique/used: 17587/6532 (Rint = 0.049)/5120 [I > 2s(I)], 510 parameters, R/Rw 0.0407/0.0425.CCDC reference number 186/1279. See http://www.rsc.org/suppdata/dt/1999/ 271/ for crystallographic files in .cif format. 1 (a) S. Trofimenko, Chem. Rev., 1993, 93, 943; (b) in this paper we use the nomenclature proposed by Trofimenko,1a Tp9 referring to the generic ligand. 2 Selected examples: (a) U. E. Bucher, A. Currao, R. Nesper, H. Ruegger, L. M. Venanzi and E. Younger, Inorg.Chem., 1995, 34, 66; (b) R. G. Ball, C. K. Ghosh, J. K. Hoyano, A. D. McMaster and W. A. G. Graham, J. Chem. Soc., Chem. Commun., 1989, 341; (c) V. Chauby, C. Serra Le Berre, P. Kalck, J.-C. Daran and G. Commenges, Inorg. Chem., 1996, 35, 6354. 3 N. G. Connelly, D. J. H. Emslie, B. Metz, A. G. Orpen and M. J. Quayle, Chem. Commun., 1996, 2289. 4 I. D. Burns, A. F. Hill, A. J. P. White, D. J. Williams and J. D. E. T. Wilton-Ely, Organometallics, 1998, 17, 1552. 5 Y. Takahashi, M.Akita, S. Hikichi and Y. Moro-oka, Organometallics, 1998, 17, 4884. 6 E. R. Humphrey, N. C. Harden, L. H. Rees, J. C. JeVrey, J. A. McCleverty and M. D. Ward, J. Chem. Soc., Dalton Trans., 1998, 3353. 7 F. Malbosc, unpublished work. 8 (a) Rh–H 1.51(4) Å in [RhH(PPh3){trans-(PPh2CH2)2ZrCp2}], R. Choukroun, A. Iraqui, D. Gervais, J.-C. Daran and Y. Jeannin, Organometallics, 1987, 6, 1197; (b) Rh–H 1.58(2) Å in [RhH- (PiPr3)3], T. Yoshida, D. L. Thorn, T. Okano, S. Otsuka and J. A. Ibers, J. Am. Chem. Soc., 1980, 102, 6451. 9 (a) G. Ujaque, A. C. Cooper, F. Maseras, O. Eisenstein and K. G. Caulton, J. Am. Chem. Soc., 1998, 120, 361; (b) J. JaVart, R. Mathieu, M. Etienne, J. E. McGrady, O. Eisenstein and F. Maseras, Chem. Commun., 1998, 2011. 10 A. E. Corrochano, F. A. Jalon, A. Otero, M. M. Kubicki and P. Richard, Organometallics, 1997, 16, 145. 11 V. Rodrigez, J. Full, B. Donnadieu, S. Sabo-Etienne and B. Chaudret, New J. Chem., 1997, 21, 847. 12 (a) J. JaVart, C. Nayral, R. Choukroun, R. Mathieu and M. Etienne, Eur. J. Inorg. Chem., 1998, 425; (b) F. J. Lalor, T. J. Desmond, G. M. Cotter, C. A. Shanahan, G. Ferguson, M. Parvez and B. Ruhl, J. Chem. Soc., Dalton Trans., 1995, 1709. Communication 8/0932

ISSN:1477-9226    

DOI:10.1039/a809323g

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1999, 273–274 273 A short platinum to silver dative bond and its application in the construction of extended structures: syntheses and structures of Ag2[Pt(ox)2]?2H2O and [Ag(H2O)]2[Ag2(CF3SO3)4][Pt(acac)2]2 Tadashi Yamaguchi,* Fumie Yamazaki and Tasuku Ito* Department of Chemistry, Graduate School of Science, Tohoku University, Sendai 980-8578, Japan. E-mail: ito@agnus.chem.tohoku.ac.jp Received 6th November 1998, Accepted 4th January 1999 Two new coordination polymers, Ag2[Pt(ox)2]?2H2O (ox 5 oxalate) and [Ag(H2O)]2[Ag2(CF3SO3)4][Pt(acac)2]2, which are formed by short platinum to silver dative bonds, were synthesized and structurally characterized. In recent years, the synthesis of coordination compounds with extended structures has attracted much attention from the view point of nano-scale science, supramolecular chemistry, crystal engineering, and solid state properties (chemistry).1,2 Weak interactions such as hydrogen-bonding,2 charge transfer interactions, 3 and weak metal–metal homonuclear bonding 4 as well as strong chemical bonds have been used to make such systems.The present study has been undertaken to explore the construction of an extended structure using metal–metal dative bonds. It has been shown that the occupied dz2 orbital in a d8 transition metal ion with square-planar coordination geometry acts as a potential donor to another metal ion (M) to form a dative M(d8)–M bond.A few examples of a PtII–M dative bond have been reported.5–8 In the published literature on PtII–M dative bonds, no one has focused on making an extended structure. We report, here, three-dimensional (3-D) structures of Ag2[Pt(ox)2]?2H2O 1 (ox = oxalate) and two-dimensional structures (2-D) of [Ag(H2O)]2[Ag2(CF3SO3)4][Pt(acac)2]2 2, which were constructed through the formation of a relatively short Pt–Ag bond. The colorless compound 1 was prepared by the slow diVusion of aqueous solutions containing K2[Pt(ox)2] and [Ag(py)2]- CF3SO3, and characterized by X-ray crystallography.† Compound 1 has a stacked 2-D layer structure (Fig.1). The 2-D layer is composed of [Pt(ox)2]22, water, and a silver ion, which are connected by three types of Ag–O bond [Fig. 1(a)].‡ The stacking of the layer is depicted in Fig. 1(b) and (c). There are two interlayer interactions. One is a Pt to Ag dative bond and the other is a water O(3)-to-Ag coordination bond. The platinum in [Pt(ox)2]22 coordinates to two silver ions at its two axial sites with a relatively short Pt–Ag distance of 2.943(1) Å.In view of the separation, it is evident that weak Pt to Ag dative bonds are formed. Although the silver–water coordination bond [2.502(9) Å] exists between layers, the platinum to silver dative bond undoubtedly plays a significant role in the interlayer interaction. Complex 2 was prepared by slow evaporation of a nitromethane solution containing [Pt(acac)2] and AgCF3SO3 in a 1 : 2 molar ratio.§ The structure of 2 is described as a 2-D sheet comprised of three units, [Ag(H2O)], [Ag2(CF3SO3)4], and [Pt(acac)2].The repeating unit is shown in Fig. 2(a) and (b). There is an inversion center at the midpoint of two Ag(1) ions. Two of the silver ions, Ag(1) and Ag(19), are quadruply bridged by triflates in a “m-triflate-O,O9” fashion. The [Ag2(CF3SO3)4] moiety has the so called “lantern” type structure. The two terminal [Pt(acac)2] units are connected to this [Ag2(CF3SO3)4] moiety through the [Ag(H2O)] groups.The repeating units are further connected to each other by the Ag(1)–Pt bond to form the 2-D sheet [Fig. 2(c)]. The Ag(1)–Pt bond distance is 2.814(1) Å, and is shorter than that in 1. It is obvious that the Pt to Ag dative bonds, shown by filled bonds in Fig. 2(c), play a primary role in making the sheet structure and it should be emphasized that the Ag(1)–Pt bond in 2 essentially does not have any other supporting interactions. There is no obvious interlayer interaction.The shortest interlayer contacts are the F ? ? ? F contacts of 2.95(1) and 2.89(2) Å. This is consistent with the fact that the crystal of 2 has a cleavage plane. Fig. 1 ORTEP10 drawings of Ag2[Pt(ox)2]?2H2O 1. (a) Structure of 2-D sheet (50% probability). The repeating unit is shown in solid bonds; (b) and (c) piled layer structure. Relevant bond lengths (Å): Pt–Ag9 2.943(1), Pt–O(1) 2.003(4), Ag–O(2) 2.559(5), Ag–O(20) 2.632(6), Ag– O(3) 2.424(8), Ag9–O(3#) 2.502(9) Å.(Key to symmetry operation: 9 = 2x, 2y, 1 2 z; 0 = 21/2 2 x, 1/2 2 y, 1 2 z; # = 1 1 x, y, z; * = 2x, y, 2z).274 J. Chem. Soc., Dalton Trans., 1999, 273–274 The Pt–Ag distances in 1 [2.943(1) Å] and in 2 [2.814(1) Å] are relatively short, and the latter, especially, is below the sum of the metal radii (2.83 Å).¶ These short distances indicate the formation of Pt–Ag bonds. These short metal–metal dative bonds give rise to the 3-D and 2-D structures in 1 and 2.Comparison of the Pt–Ag distances in 1 and 2 shows that the platinum in 2 makes a stronger metal–metal dative bond than that in 1. The reason may be attributed to the following two factors. One is that platinum in 2 coordinates to only one silver whereas in 1 it coordinates to two silver ions; in the latter the electron pair in the dz2 orbital is shared between the two axial sites. The other reason is the ligand field strength of the ligand coordinated to the platinum.Possibly, the ligand field strength of acetylacetonate ion coordinated to Pt21 is stronger than that of oxalate ion, and the stronger donation of the acac2 ligand may cause an increase in the electron density of the dz2 orbital to make a stronger metal–metal dative bond. Fig. 2 ORTEP10 drawings of [Ag(H2O)]2[Ag2(CF3SO3)4][Pt(acac)2]2 2. (a) Top view of repeating unit (20% probability, atoms of CF3 moieties are drawn as small circle for clarity); (b) side view of repeating unit; (c) 2-D layer structure.Relevant bond lengths (Å): Pt–Ag 2.814(1), Pt–O(1) 1.968(9), Pt–O(2) 1.985(9), Pt–O(3) 1.982(9), Pt–O(4) 1.976(8), Ag(1)–O(5) 2.41(1), Ag(1)–O(8) 2.40(1), Ag(19)–O(7) 2.52(1), Ag(19)– O(9) 2.38(1), Ag(2)–O(1) 2.557(9), Ag(2)–O(4) 2.575(9), Ag(2)–O(6) 2.52(1), Ag(2)–O(10) 2.47(1), Ag(2)–O(11) 2.28(1). (Key to symmetry operation: 9 = 2x, 2y, 2z.) Acknowledgements This work was supported by Grants-in-Aid for Scientific Research (No. 10740299 and Priority Area No. 10149102) from the Ministry of Education, Science and Culture, Japan. Notes and references † Preparation of complex 1. An aqueous solution of K2[Pt(ox)2] (5 mg in 3 cm3) and [Ag(py)2]CF3SO3 (4 mg in 3 cm3) were allowed to slowly diVuse in an H-tube at ambient temperature. Colorless crystals of Ag2[Pt(ox)2]?2H2O were obtained after 1 week. Crystal data: 1 = PtAg2- C4H4O10, M = 623, monoclinic, space group C2/m (no. 12), a = 9.745(2), b = 7.913(1), c = 6.954(1) Å, b = 117.05(1)8, V = 477.6(1) Å3, Z = 2, m = 18.64 mm21, T = 293 K. With the use of 585 unique reflections [I > 3s(I)] out of 625 reflections, the final R and Rw values were 0.024 and 0.030. ‡ The asymmetric unit consists of six non-hydrogen atoms with unprimed labeling except for hydrogens. Crystallographic 2/m symmetry exists at Pt atoms: a crystallographic C2 axis passes through Pt and bisects the O(1)–Pt–O(1*) angle; there exists a mirror plane perpendicular to the [Pt(ox)2] plane which passes through Pt, Ag and O(3).§ Preparation of complex 2. [Pt(acac)2] (4 mg) and AgCF3SO3 (5 mg) were dissolved in nitromethane (3 cm3) and the solution was slowly evaporated in a refrigerator for 1 week. Colorless crystals of [Ag- (H2O)]2[Ag2(CF3SO3)4][Pt(acac)2]2 were collected. Crystal data: 2 = Pt2- Ag4C24H32O22F12S4, M = 1850, monoclinic, space group P21/n (no. 14), a = 13.039(7), b = 12.287(3), c = 15.000(3) Å, b = 102.90(2)8, V = 2342(1) Å3, Z = 2, m = 7.86 mm21, T = 293 K.With the use of 2985 unique reflections [I > 3s(I)] out of 5872 reflections, the final R and Rw values were 0.046 and 0.037. CCDC reference number 186/1301. See http://www.rsc.org/suppdata/dt/1999/273/ for crystallographic files in .cif format. ¶ The Cambridge Structural Database 9 shows a total of 27 complexes having a PtII-to-Ag dative bond shorter than 2.82 Å. All the complexes, however, have a pentahalogenophenyl ligand attached to Pt, and one ortho halogen atom of each C6X5 group makes a close contact (2.75– 2.90 Å) with the silver ion. The interaction can possibly facilitate an attractive force that contributes to the shortening of the Pt–Ag bonds.Some of these complexes also have a bridging hydride, chloride, or thioether between Pt and Ag other than the dative bond. 1 O. M. Yaghi and H. Li, J. Am. Chem. Soc., 1996, 118, 295; M. Fujita, Y. J. Know, S. Washizu and K.Ogura, J. Am. Chem. Soc., 1994, 116, 1151; M. Kondo, T. Yoshomi, K. Seki, H. Matsuzaka and S. Kitagawa, Angew. Chem., Int. Ed. Engl., 1997, 36, 1725. 2 C. B. Aakeröy and A. M. Beatty, Chem. Commun., 1998, 1067; C. B. Aakeröy, A. M. Beatty and D. S. Leinen, J. Am. Chem. Soc., 1998, 120, 7383. 3 R. D. Bailey, L. L. Hook and W. T. Pennington, Chem. Commun., 1998, 1181. 4 F. Robinson and M. J. Zaworotko, J. Chem. Soc., Chem. Commun., 1995, 2413; M.-L. Tong, X.-M. Chen, B.-H.Ye and S. W. Ng, Inorg. Chem., 1998, 37, 5278. 5 G. Aullón and S. Alvarez, Inorg. Chem., 1996, 35, 3137; L. R. Falvello, J. Forniés, A. Martín, R. Navarro, V. Sicilia and P. Villarroya, Inorg. Chem., 1997, 36, 6166; C. Mealli, F. Pichierri, L. Randaccio, E. Zangrando, M. Krumm, D. Holtenrich and B. Lippert. Inorg. Chem., 1995, 34, 3418 and refs. therein. 6 M. P. Brown, S. J. Cooper, A. A. Frew, L. Manojlovic-Muir, K. W. Muir, R. J. Puddephatt, K. R. Seddon and M. A. Thomson, Inorg. Chem., 1981, 20, 1500; M. Krumm, E. Zangrando, L. Randaccio, S. Menzer and B. Lippert, Inorg. Chem., 1994, 32, 700. 7 F. A. Cotton, L. R. Falvello, R. Uson, J. Fornies, M. Tomas, J. M. Casas and I. Ara, Inorg. Chem., 1987, 26, 1366; R. Uson, J. Fornies, M. Tomas, I. Ara, J. M. Casas and A. Martin, J. Chem. Soc., Dalton Trans., 1991, 2253; J. Fornies, R. Navarro, M. Tomas and E. P. Urriolabeitia, Organometallics, 1993, 12, 940. 8 F. D. Rocnon and R. Melanson, Acta Crystallogr., Sect. C, 1988, 44, 474. 9 F. H. Allen and O. Kennard, Chem. Des. Autom. News, 1993, 8, 31. 10 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. Communication 8/08678H

ISSN:1477-9226    

DOI:10.1039/a808678h

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1999, 275–277 275 Synthesis and structures of photodecarbonylated ruthenium(II) complexes—potential intermediates for mixed ligand complexes Glen B. Deacon,a Christopher M. Kepert,a Norma Sahely,a Brian W. Skelton,b Leone Spiccia,*a Nicholas C. Thomasa and Allan H. White b a Department of Chemistry, Monash University, Clayton 3168, Australia. E-mail: leone.spiccia@sci.monash.edu.au b Department of Chemistry, University of Western Australia, Nedlands 6907, Australia Received 15th October 1998, Accepted 8th December 1998 Irradiation of solutions containing complexes of the type [RuL(CO)2Cl2] where L is a 2,29-bipyridine analogue leads to monodecarbonylation and the formation of dimeric ruthenium(II) complexes, [RuL(CO)2Cl]2, for which two different structures have been established, viz.with a trans disposition of bridging and terminal chlorides [L = di(2- pyridyl) ketone] or with CO trans to bridging chlorides (L = 1,10-phenanthroline).Ruthenium(II) complexes of 2,29-bipyridine and 2,29-bipyridine analogues have attracted much attention, originally as photoredox catalysts for water splitting 1 and as sensitisers for photovoltaic cells.2 The photoluminescent and redox properties of polynuclear ruthenium complexes containing analogous ligands 3 have also been the subject of intense study. Additionally, complexes with mixed bipyridine and carbonyl ligands have been investigated as catalysts for the reduction of carbon dioxide.3,4 Although redox and hence catalytic activity depend on the ligands,5–8 few strategies are available which enable the controlled sequential addition of three diVerent bidentate ligands to ruthenium and only one general route to [RuL(L9)L0]21 complexes (where L, L9 and L0 are inequivalent bidentate diimines) exists.7 Irradiation of ruthenium(II) carbonyl complexes with visible or UV light has long been known to result in decarbonylation.For example, photodecarbonylation of ruthenium(II) carbonyl porphyrin complexes in various substitution reactions has been investigated extensively.9,10 In contrast, the photodecarbonylation of [RuL(CO)2Cl2] complexes, where L = bipyridine or related diimine, as a synthetic route to mixed ligand complexes has been largely overlooked except for a recent application to Ru(II) terpyridine complexes.11 We report the application of photodecarbonylation in the synthesis of complexes of the form [RuL(CO)Cl2]2.These complexes are ideal precursors to monocarbonylruthenium(II) complexes containing dissimilar bidentate ligands, [RuL(L9)(CO)Cl]1.12 Irradiation of [RuL(CO)2Cl2]n complexes, formed by reaction of polymeric [Ru(CO)2Cl2]n with appropriate bidentate ligands, 7 in poorly coordinating solvents results in photo-monodecarbonylation and the subsequent formation of low solubility monocarbonyl complexes [eqn. (1), L = di(2-pyridyl) ketone 2 [Ru(L)(CO)2Cl2] hn [Ru(L)(CO)Cl2]2 1 2 CO (1) (dpk), 2,29-bipyridine (bpy), 4,49-dimethyl-2,29-bipyridine (Me2bpy) or 1,10-phenanthroline (phen)].Single n(CO) and terminal n(Ru–Cl) stretching frequencies indicate symmetric structures. The correspondence in n(CO) for most complexes (ca. 1945 cm21 for bpy, Me2bpy and phen) is suggestive of similar structures. For the dpk complex, a diVerent structure is indicated by a n(CO) frequency of 1985 cm21. The lowering of n(CO) from 1990–2010 and 2050–2060 cm21 for the [RuL(CO)2Cl2] precursors reflects an increased bond order between the carbonyl and metal centre and thus a diminished likelihood of further decarbonylation.Hence, long term storage presents no problems enhancing their values as synthetic precursors. Deposition of [RuL(CO)Cl2]2 from reaction mixtures in CH2Cl2 generally gave powders. However, the phen complex precipitated as small orange moderate quality crystals, whilst well formed red single crystals were obtained for dpk.The X-ray structures† revealed two diVerent, dichloro-bridged, centrosymmetric dimers [Fig. 1(a), (b)]. In each case, the Ru atoms and bridging Cl are coplanar, and each Ru has distorted octahedral stereochemistry. In [Ru(phen)(CO)Cl2]2 the carbonyls are trans to the bridging Cl and the terminal Cl trans to pyridyl nitrogens, whilst in [Ru(dpk)(CO)Cl2]2 the terminal and bridging Cl are mutually trans, and the carbonyls are trans to the pyridyls.The Ru ? ? ?Ru separation in [Ru(phen)(CO)Cl2]2 [3.67(1) Å] is slightly shorter than the value in [Ru(dpk)(CO)Cl2]2 [3.5644(5) Å]. The latter is shorter than the 3.741 Å separation reported for a supported di-m-chloride complex [Ru3(S2- CNEt2)4(CO)3Cl2].13 The angle subtended by the dpk ligand, N–Ru–N angle 86.7(1)8 compares well with that subtended by dpk in other related Ru(II) dpk complexes; 86.6(2)8 in [Ru- (bpy)(dpk)(CO)Cl]1 and 87.4(3)8 in [Ru(dpk)(Me2phen)(CO)- Cl]1.12 By forming a six-membered chelate ring dpk allows the Ru(II) center to achieve a geometry that is closer to octahedral than is possible for diimines which form five-membered chelates with more acute N–Ru–N angles {e.g. 79.4(4)8 in [Ru(phen)- (CO)Cl2]2}. The Ru–C distance of 1.856(4) Å matches those found in other monocarbonyl Ru(II) complexes {e.g. 1.86(3) Å for [Ru(bpy)2(CO)Cl]1}.14 The pyridyl entities of the dpk ligands have undergone significant distortion from coplanarity in order to conform to a configuration more conducive to the requirements of an octahedral Ru geometry.The interplanar dihedral angle between the pair of C5N planes is 44.8(2)8, the ruthenium deviations from the two planes being 0.225(6), 0.289(7) Å. Consequently, the behaviour of dpk is not typical of the bipyridine analogue ligands investigated. While substituted bipyridines are unable to complex in a similar manner to dpk owing to geometric restriction caused by the smaller chelate ring size, potential conformations in phen and its derivatives are also limited by a greater structural rigidity.The smaller N–Ru–N angle in the case of [Ru(phen)(CO)Cl2]2 provides evidence for this. The contrast in coordination geometries of dpk, cf. the other diimine ligands, may account for structural diVerences in [RuL(CO)Cl2]2, and hence the variance in n(CO) frequencies. Given the distinctions between the two structurally characterised isomers, the greater p-bonding capacity of pyridyl rings (L = dpk) over bridging chlorides (L = phen) trans to the carbonyls accounts for the higher n(CO) frequency for L = dpk.Since the n(CO) frequencies of the other diimine complexes examined are similar to that of [Ru(phen)(CO)Cl2]2 it is likely that they are also dimeric with CO trans to the chloride bridge. The progress of the decarbonylation of trans-[Ru(dpk)- (CO)2Cl2] was examined by 1H NMR and IR spectroscopy.276 J. Chem. Soc., Dalton Trans., 1999, 275–277 (a) (b) IR spectra of [Ru(dpk)(CO)2Cl2] recorded in near saturated dichloromethane solution after various irradiation times show the attrition of n(CO) frequencies at 2072 and 2013 cm21 (cf.Nujol mull 2059, 1992 cm21) coupled with the appearance and intensification of a single peak at 1980 cm21. The 1H NMR spectrum of a near saturated solution of [Ru(dpk)(CO)2Cl2] in CDCl3 showed only four resonances at d 9.25, 8.25, 8.17 and 7.73 relative to SiMe4, consistent with the expected C2 symmetrical cis-(CO)2 trans-Cl2 arrangement.After 15 min irradiation with a xenon arc lamp, eight new resonances having a similar Fig. 1 (a) The centrosymmetric dimer [Ru(phen)(CO)Cl2]2; 25% thermal envelopes are shown, atoms refined with anisotropic thermal parameters being shaded and hydrogen atoms having arbitrary radii of 0.1 Å. Selected distances (Å) and angles (8): Ru–Cl(1) 2.422(4), Ru– Cl(2) 2.402(1), Ru–C(1) 1.94(1), Ru–N(1) 2.05(1), Ru–N(2) 2.068(10), Ru–Cl(19) 2.489(3); Cl(1)–Ru–Cl(2) 90.8(1), Cl(1)–Ru–C(1) 83.4(1), Cl(1)–Ru–N(1) 171.8(3), Cl(1)–Ru–N(2) 94.7(3), Cl(1)–Ru–Cl(19) 83.4(1), Cl(2)–Ru–C(1) 89.0(3), Cl(2)–Ru–N(1) 94.9(3), Cl(2)–Ru–N(2) 174.3(3), Cl(2)–Ru–Cl(19) 91.4(1), C(1)–Ru–N(1) 91.9(5), C(1)–Ru– N(2) 92.1(4), C(1)–Ru–Cl(1) 177.4(4), N(1)–Ru–N(2) 79.4(4), N(1)– Ru–Cl(19) 90.6(3), N(2)–Ru–Cl(19) 87.7(3), Ru–Cl(1)–Ru9 96.6(1). (b) [Ru(dpk)(CO)Cl2]2, projected normal to the Ru(m–Cl)2Ru plane; 20% thermal ellipsoids are shown for non–hydrogen atoms.Selected distances (Å) and angles (8): Ru–Cl(1) 2.429(1), Ru–Cl(2) 2.376(1), Ru–C(1) 1.856(4), Ru–N(11) 2.129(3), Ru–N(21) 2.038(3), Ru–Cl(19) 2.410(1); Cl(1)–Ru–Cl(2) 92.88(4), Cl(1)–Ru–C(1) 90.6(1), Cl(1)–Ru– N(11) 89.50(9), Cl(1)–Ru–N(21) 174.6(1), Cl(1)–Ru–Cl(19) 85.11(4), Cl(2)–Ru–C(1) 92.2(2), Cl(2)–Ru–N(11) 88.6(1), Cl(2)–Ru–N(21) 90.8(1), Cl(2)–Ru–Cl(19) 175.77(6), C(1)–Ru–N(11) 179.1(2), C(1)–Ru– N(21) 93.2(2), C(1)–Ru–Cl(1) 91.5(2), N(11)–Ru–N(21) 86.7(1), N(11)– Ru–Cl(19) 87.6(1), N(21)–Ru–Cl(19) 91.0(1), Ru–Cl(1)–Ru9 94.89(4).integration to the reactant signals were observed and corresponded to those observed for a solution obtained by partial dissolution of [Ru(dpk)(CO)Cl2]2 in CDCl3. Although consistent with the formation of the dimer, assignment to a chloroform solvate or a five-coordinate intermediate is more likely given the insoluble nature of the final product. In such a chloroform complex, the trans orientation of the chloro ligands must be preserved, as the stereochemistry of a cis-chloride complex should give rise to a greater number of signals.Photoisomerism is also unlikely due to the necessity of maintaining the transchloride configuration. This is not the case in the photodecarbonylation of trans-chloride [Ru(phen)(CO)2Cl2], for which the stereochemistry of the product necessitates isomerisation. Subsequent work has shown that regardless of their structure, [RuL(CO)Cl2]2 complexes display similar reactivity towards further substitution by diimine ligands and yield [RuL(L9)(CO)Cl]1 (L, L9 = chelating diimines).For example, monodecarbonylated bpy and dpk complexes have been successfully used to prepare new mixed ligand Ru(II) complexes, {e.g., [Ru(bpy)(Me2phen)(CO)Cl]1, [Ru(bpy)(dpk)(CO)Cl]1 and [Ru(dpk)(Me2phen)(CO)Cl]1}, which have in turn been used as precursors to tris-heteroleptic complexes.12 Thus, the photodecarbonylation of [Ru(L)(CO)2Cl2] is a key step in a new synthetic route to tris(heteroleptic) ruthenium(II) complexes.Acknowledgements We acknowledge the receipt of an Australian Postgraduate Award (C. M. K.), and a grant from the Monash Research Fund. Notes and references † Preparation of [RuL(CO)Cl2]n: in a typical preparation [Ru(bpy)- (CO)2Cl2] 7 (1.00 g, 2.6 mmol) was dissolved in dichloromethane (70 cm3) and filtered through diatomaceous earth. The filtrate was decanted into a stoppered Pyrex conical flask.The flask was then irradiated with a 50 W halogen lamp at a distance of 15 cm. Irradiation of the stirred solution was sustained for a duration of 48 h after which time an orange precipitate had formed. The precipitate was collected by filtration, washed thoroughly with ethanol and dried at 70 8C; yield (0.85 g, 87%) (Found: C, 36.2; H, 2.2; N, 7.6. [Ru(bpy)(CO)Cl2]?0.25CH2Cl2 requires C, 35.8; H, 2.3; N, 7.4%), nmax/cm21 (CO) 1944s (Nujol). Analogous preparations for L = dpk (Found : C, 32.9; H, 2.0; N, 5.9.[Ru(dpk)- (CO)Cl2]2?2CH2Cl2 requires C, 33.3; H, 2.1; N, 6.0%), nmax/cm21 (CO) 1985s, deposited from reaction mixture as red single crystals. L = 4,49- Me2bpy (Found: C, 37.8; H, 3.0; N, 6.6. [Ru(Me2bpy)(CO)Cl2]?CH2Cl2 requires C, 38.0; H, 3.1; N, 6.6%), nmax/cm21 (CO) 1945s. L = phen (Found: C, 40.9; H, 2.3; N, 7.2. [Ru(phen)(CO)Cl2] requires C, 41.1; H, 2.1; N, 7.4%), nmax/cm21 (CO) 1944s. ‡[Ru(dpk)(CO)Cl2]2: Data were collected at room temperature on a four-circle/single counter diVractometer.Solution was eVected by Patterson methods. Anisotropic thermal parameters were refined for the non-hydrogen atoms in the full matrix least-squares refinement, (x, y, z, Uiso)H being constrained at estimated values. DiVerence map residues were modelled in terms of a pair of dichloromethane solvent molecules, site occupancies set at unity after trial refinement. Neutral atom complex scattering factors were employed, computation using the Xtal 3.4 program system.15 Crystal data: C24H16Cl4N4O4Ru2?2CH2Cl2 M = 938.2, monoclinic, space group P21/c (C5 2h, no. 14), a = 9.800(1), b = 12.192(2), c = 14.503(2) Å, b = 100.94(1)8, U = 1701.4(5) Å3, T = 295 K, Dc (Z = 2) = 1.83 g cm23, 2qmax = 608, 2q–q scan mode, F(000) = 920, mMo = 15.5 cm21 (no correction), Mo-Ka graphite monochromated radiation (l = 0.71073 Å), 4556 independent reflections, 3880 observed [I �3s(I)], R = 0.041, Rw = 0.064 [on |F|o statistical weights, derivative of s2(I) = s2(Idiff) 1 0.0004s2(Idiff)].[Ru(phen)(CO)Cl2]2: a data set collected on a Enraf-Nonius Kappa CCD diVractometer was used in the full matrix least squares refinement after solution of the structure by direct methods. Anisotropic thermal parameters were refined for ruthenium and chlorine, other atoms being refined using isotropic thermal parameters except hydrogens which were constrained at calculated positions.DiVerence map residues were modelled in terms of dichloromethane solvent molecules, site occupancies set at 0.5. Calculations were performed using Texsan 16 employing neutral atom complex scattering factors. Crystal data: C22H16Cl4N4O2Ru2?5/2CH2Cl2 M = 924.7, monoclinic, space group C2/c (no. 15), a = 18.153(2), b = 10.544(2), c = 17.092(2) Å,J. Chem. Soc., Dalton Trans., 1999, 275–277 277 b = 115.66(1)8, U = 2948.7(6) Å3, T = 123 K, Dc (Z = 4) = 2.01 g cm23, 2qmax = 558, 1808 f-scan, F(000) = 1767, mMo = 18.7 cm21 (no correction), 3395 unique reflections, 1247 observed [I � 3s(I)], R = 0.083, Rw = 0.069 on |F|.CCDC reference number 186/1273. See http:// www.rsc.org/suppdata/dt/1999/275/ for crystallographic files in .cif format. 1 K. Kalyanasundaram, Coord. Chem. Rev., 1982, 46, 159. 2 A. Hagfeldt and M. Gratzël, Coord. Chem. Rev., 1995, 95, 49. 3 V. Balzani, A. Juris and M. Venturi, Chem. Rev., 1996, 96, 759. 4 H. Tanaka, B.-C. Tzeng, H.Nagao, S.-M. Peng and K. Tanaka, Inorg. Chem., 1993, 32, 1508. 5 M. Haukka, J. Kiviaho, M. Ahlgrén and T. A. Pakkanen, Organometallics, 1995, 14, 825. 6 H. Ishida, K. Fujiki, T. Ohba, K. Ohkubo, K. Tanaka, T. Terada and T. Tanaka, J. Chem. Soc., Dalton Trans., 1990, 2155. 7 P. A. Anderson, G. B. Deacon, K. H. Haarmann, F. R. Keene, T. J. Meyer, D. A. Reitsma, B. W. Skelton, G. F. Strouse, N. C. Thomas, J. A. Treadway and A. H. White, Inorg. Chem., 1995, 34, 6145. 8 G. Denti, L. Sabatino, G. De Rosa, A. Bartolotta, G. Di Marco, V. Ricevuto and S. Campagna, Inorg. Chem., 1989, 28, 3309. 9 M. Hoshino and Y. Kashiwagi, J. Phys. Chem., 1990, 94, 673. 10 S. C. Jeoung, D. Kim, D. W. Cho, M. Yoon and K. Ahn, J. Phys. Chem., 1996, 100, 8867. 11 N. C. Fletcher and F. R. Keene, J. Chem. Soc., Dalton Trans., 1998, 2293. 12 G. B. Deacon, C. M. Kepert and L. Spiccia, unpublished work. 13 C. L. Raston and A. H. White, J. Chem. Soc., Dalton Trans., 1975, 2422. 14 J. M. Clear, J. M. Kelly, C. M. O’Connell, J. G. Vos, C. J. Cardin, S. R. Costa and A. J. Edwards, J. Chem. Soc., Chem. Commun., 1980, 750. 15 The Xtal 3.4 User’s Manual, ed. S. R. Hall, G. S. D. King and J. M. Stewart, University of Western Australia, Lamb, Perth, 1995. 16 Texsan Crystal Structure Analysis Package, Molecular Structure Corporation, The Woodlands, TX, 1985 and 1992. Communication 8/08

ISSN:1477-9226    

DOI:10.1039/a808006b

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 279–284 279 Confinement of C60 in an extended saddle shaped nickel(II) macrocycle Paul D. Croucher, Peter J. Nichols and Colin L. Raston* Department of Chemistry, Monash University, Clayton, Melbourne, Victoria 3168 Australia Received 28th July 1998, Accepted 16th December 1998 Saddle shaped (5,14-dihydro-2,3,6,8,11,12,15,17-octamethyldibenzo[b,i][1,4,8,11]tetraazacyclotetradecine)nickel(II), Ni(OMTAA), acts as a divergent receptor molecule with C60 forming a 1 : 1 complex in solution and in the solid state, isolated as [Ni(OMTAA)C60]?2CS2.The extended supramolecular array is based on linear chains of close contact C60 molecules [fullerene ? ? ? fullerene centroids 10.03(1) and 10.04(1) Å], and linear chains of p-stacked alternating molecules of C60 and Ni(OMTAA) with adjacent chains running in opposite directions. Introduction Host–guest chemistry of globular molecules including fullerenes,1–19 carboranes,20 and P4E3, E = S, Se,1,21 has recently gained prominence in forming inclusion nano-structures, crystal engineering, and in their purification.A major challenge in forming supermolecules involving such molecules is gaining control over the inherently weak host–guest interactions. Complementarity of curvature and maximising the number of points per area of van der Waals contact are important factors in the formation of stable host–guest complexes, at least in the absence of hydrogen bonding, and electrostatic and coordination interactions.Competition between the host–host, guest–host and guest–guest interactions is important in determining the structure of the resulting supramolecular array. This is particularly evident in fullerene chemistry where interfullerene interactions play a major role in the structures of inclusion complexes.5–16 The interactions of solvent with all species and crystal packing forces provide additional variables.Recently, we reported that Ni(TMTAA) (see below, R = H), acts as a divergent heterotopic receptor with C60, 1,2-dicarbadodecaborane and the chalcogenides P4S3 and P4Se3;1,21 the saddle shape for the host arises from otherwise unfavourable interactions between the {NC(Me)}2CH methyl groups and the hydrogen atoms on the aromatic rings. Herein we report the preparation and structural elucidation of a divergent receptor nickel(II) macrocycle which has an extended concave surface, Ni(OMTAA), (see above, R = Me) together with the structural elucidation and UV/vis studies of its supramolecular complex with C60.The larger area of contact of the host with the surface of the fullerene should in principle lead to a complex with diminished fullerene–fullerene interactions per fullerene,6,7,12 and indeed this was found to be the case. In contrast to [Ni(TMTAA)C60] which has a corrugated two dimensional sheet of close contact fullerenes, Ni(OMTAA) forms linear, single column chains of close contact C60 molecules in its 1: 1 complex with the fullerene, [Ni- (OMTAA)C60]?2CS2.This columnar motif of inter-fullerene contacts is unusual and has few precedents; Sugawara and Crane have prepared inclusion complexes of C60 with redox active hosts with a view to preparing molecular conduction devices, these complexes containing columnar stacks of C60 molecules.17–19 Related to these are the zig-zag chains of C60 molecules surrounded by a sheath of bowl shaped cyclotriveratrylene (CTV) molecules in [(CTV)C60].11 Results and discussion Synthesis of (5,14-dihydro-2,3,6,8,11,12,15,17-octamethyldibenzo[ b,i][1,4,8,11]tetracyclotetradecine)nickel(II), Ni(OMTAA) and [N,N9-bis(2-amino-4,5-dimethylphenyl)pentanediiminato] nickel(II) acetate, I, and [Ni(OMTAA)C60]?2CS2 The synthesis of Ni(OMTAA) is a modification of an existing preparation.22 Two equivalents of 1,2-diamino-4,5-dimethylbenzene were condensed with two equivalents of acetylacetone and one equivalent of nickel(II) acetate tetrahydrate as a highly concentrated mixture in anhydrous, anoxic methanol.The Ni(OMTAA) precipitates from solution and is isolated in 76% yield which is considerably higher than that obtained in the original preparation, 45%, Scheme 1. The reaction requires a high concentration of reagents for optimum yield and careful exclusion of air due to the oxygen sensitive nature of the “threequarter complex” intermediate I which has been isolated and structurally characterised, Scheme 1.This was achieved when Scheme 1280 J. Chem. Soc., Dalton Trans., 1999, 279–284 the reaction mixture was more dilute and filtered hot to remove a small amount of Ni(OMTAA) and on cooling gave I in 12% yield. This intermediate complex is indefinitely air stable in the solid state but once in solution, decomposes rapidly in the presence of oxygen to an as yet unknown product. A related intermediate has been isolated, but not structurally authenticated, from the preparation of Ni(TMTAA).23 [Ni- (OMTAA)C60]?2CS2 was prepared by the slow evaporation of CS2 solutions of C60 and excess Ni(OMTAA).Structure of I?CH3OH The asymmetric unit consists of one molecule of the cation/ anion complex which has the nickel atom in a square planar environment, as expected. The tetradentate ligand has a saddle shape with respect to the two phenyl groups, and has a shallow pitch, 159.4(3)8, defined as the dihedral angle between the planes of the two aromatic rings, Fig. 1. This is much shallower than in the closed macrocycle itself, Ni(OMTAA), pitch angle 137.9(3)8, and the related macrocycle Ni(TMTAA), pitch angle 128.3(3)–135.4(3)8 (see below).24–27 The shallower pitch is due to the presence of only one pair of CH3 ? ? ?H–Caromatic repulsions, compared to two pairs in Ni(TMTAA) and Ni(OMTAA). To a lesser extent it may be associated with longer Ni–N distances for the two terminal amine centres, and thus a more open side of the cation.The Ni–N bond lengths in the {NC(Me)}2CH section of the complex are shorter [1.88(1) and 1.88(1) Å] than in the amine portion [1.91(1) and 1.92(1) Å], in accordance with the diVerence in valency of the N centres, sp2 versus sp3. The angle between the least-squares planes through the four nitrogen atoms and the least-squares planes through the phenyl rings is 169.3(3)8 while the angle between the least-squares planes through the four nitrogen atoms and the least-squares planes through the {NC(Me)}2CH section is 161.3(3)8.Hydrogen atoms were placed in calculated positions and refined, except for those on the methanol molecule which were not included. Hydrogen bonding was deduced by inspection of the inter-atomic distances and the geometries of interacting species. The acetate moiety is hydrogen-bonded to the amine groups, the N ? ? ? O distances being 2.78(1) and 2.95(1) Å, with associated N–H ? ? ? O distances, 1.83(1) and 2.10(1) Å.The acetate oxygen associated with the longer N–H ? ? ? O hydrogen bond is also involved in hydrogen bonding to a molecule of methanol, O ? ? ? O 2.67(2) Å although the hydrogen atom on the methanolic oxygen atom was not located in the analysis of the X-ray diVraction data. The methanol molecule is also within hydrogen bonding distance [O? ? ? N 2.91(2) Å, N–H ? ? ?O 2.00(1) Å] of the amine nitrogen participating in the shorter hydrogen bond to acetate.Structure of Ni(OMTAA) Despite the synthesis of and the use of Ni(OMTAA) as a precursor for metal ion complexes of OMTAA22,22,28–30 and the crystal structure of the protonated metal-ion free ligand Fig. 1 Structure of I showing the cation/anion hydrogen bonding interplay. having been described,31,32 the structure of Ni(OMTAA) has not previously been reported. The asymmetric unit consists of one molecule of Ni(OMTAA) which adopts a saddle-like conformation due to steric interactions between the 2,3,6,8- methyl groups ({NC(Me)}2CH) and the hydrogen atoms on the aromatic rings, as observed for Ni(TMTAA)24,27 and H2(OMTAA).31,32 Molecules of Ni(OMTAA) are assembled into dimers but the nature of the dimeric motif diVers considerably from that observed for Ni(TMTAA).Ni(TMTAA) has been structurally characterised in three crystalline forms, as monomeric,26 dimeric24 and tetrameric units.25 The tetrameric form of Ni(TMTAA), isostructural with Cu(TMTAA), self assembles around a central dimer in which the Ni(TMTAA) macrocycles interlock at 908, driven by complementarity of curvature of the components as well as by Ni ? ? ? H interactions, with the nickel atom of one macrocycle residing below a phenyl ring of the other.Additional interlocking of the dimers occurs through the ({NC(Me)}2CH)2Ni concave surfaces to two other Ni(TMTAA) molecules through the same surfaces.25 Similarly, in dimeric Ni(TMTAA), the self assembly is driven by favourable complementarity of curvature of the interlocking components and favourable interactions between the {NC(Me)}2CH moieties.24 Saddle shaped Ni(OMTAA) possesses two divergent concave surfaces, one face of which is comprised of the NiN4 plane and the phenyl groups while the NiN4 plane and the {NC(Me)}2CH moieties and metal centres make up the opposing surface, Fig. 2.1,21,24–28 The angles between the least-squares planes through the four nitrogen atoms and the least-squares planes through the phenyl rings are 158.6(3) and 159.4(3)8, while the angles between the least-squares planes through the four nitrogen atoms and the least-squares planes through the {NC- (Me)}2CH moieties are 156.0(3) and 155.3(3)8.The overall pitch angle between the least-squares planes of the phenyl rings is 137.9(3)8 which is comparable to that observed for H2(OMTAA) [139.2(3)8] and the Ni(TMTAA) dimer [135.4(3) and 134.4(3)8] but somewhat less than that observed for the Ni(TMTAA) momomer [128.3(3)8].There is a slight decrease in the pitch angle for the ({NC(Me)}2CH)2 face in Ni(OMTAA), the least-squares planes intersecting at 131.3(3)8 while in dimeric Ni(TMTAA) the same planes intersect at 130.5(3)8.24,26,31–34 In the present structure the phenyl group of one Ni(OMTAA) molecule sits over the phenyl of another in a skewed p-stacked arrangement with the two rings separated by 3.52(1) to 3.71(1) Å (Fig. 2). This distortion from a conventional oVset p-stacked arrangement is due to otherwise unfavourable Mephenyl–Mephenyl and Mephenyl–{NC(Me)}2CH interactions and results in an o-aromatic proton residing unusually close to the nickel atom, Ni ? ? ? H 2.81(1) Å (Fig. 2). Fig. 2 Projection of the self associated dimers of Ni(OMTAA); the dotted line represents the close Ni ? ? ? H–C contact.J. Chem. Soc., Dalton Trans., 1999, 279–284 281 Structure of [Ni(OMTAA)C60]?2CS2 The compound crystallises in the space group C2/c with the asymmetric unit comprised of half a molecule of the two supramolecular synthons and a molecule of CS2.The fullerene is in the larger saddle of one Ni(OMTAA) molecule, i.e. the Fig. 3 Host–guest–host–guest ? ? ? interplay in the structure of [Ni- (OMTAA)C60]?2CS2. saddle built up by the phenyl groups, and in the opposite smaller saddle of another Ni(OMTAA) molecule. The overall host–guest contacts form a continuous linear chain parallel to the b axis within which the host molecules are aligned unidirectionally, Fig. 3. There are no significant interactions between hosts, the closest host ? ? ? host distances being greater than 4.0 Å. The closest fullerene contact to the nickel centre in the phenyl lined face of the host is to the midpoint of an edge shared by two C6 rings which lies “parallel” to Ni(OMTAA), Ni ? ? ?C60 3.23(1) Å (C2—mid point). Although the Ni ? ? ?C60 distance is similar on the other face there is not such a clear geometrical alignment.Despite this apparent preference of C60 for the phenyl face of the host, the subtle interplay between the guest and host manifests itself in the tendency for the fullerene to bind to both faces of Ni(OMTAA). In the presence of excess host, this array still forms despite the apparent advantage of C60 binding to the larger surface area of the phenyl containing face, and thus greater potential van der Waals interactions. Contiguous chains are slightly oVset to achieve eYcient packing within the array, Fig. 3, and run in opposite directions, thus eVectively cancelling dipoles. The pitch of the phenyl groups in Ni(OMTAA) decreases upon complexation with the fullerene, the unique angle between Fig. 4 Inter-fullerene contacts in [Ni(OMTAA)C60]?2CS2 showing the linear chains of fullerenes along the c axis in the (a) ac and (b) ab planes.282 J. Chem. Soc., Dalton Trans., 1999, 279–284 the least-squares planes through the four nitrogen atoms and the least-squares plane through the phenyl ring is 148.5(3)8, a decrease of 10.1(3)–10.9(3)8; the corresponding angle between the phenyl groups (pitch angle) is 116.9(3)8.In contrast, the angle between the least-squares plane through the four nitrogen atoms and the least-squares plane through the {NC(Me)}2CH, 156.9(3)8, moieties remains virtually unchanged relative to uncomplexed Ni(OMTAA). The corresponding pitch angle between the two {NC(Me)}2CH moieties is 135.8(3)8.A similar change is observed on complexation of C60 with Ni(TMTAA). The least-squares planes through the phenyl rings of the Ni(TMTAA) dimer intersect at 135.4(3), 134.4(3)8 and decrease to 127.9(3)8 upon complexation to C60.1 Importantly, fullerene–fullerene interactions are now restricted to one dimensional chains, with fullerene at the van der Waals limit, centroid ? ? ? fullerene centroid distances 10.03(1) and 10.04(1) Å; the chains of fullerenes run parallel to the c axis and are associated with a sheath of Ni(OMTAA) and CS2 molecules, Fig. 4. Each fullerene makes intimate contact, C ? ? ? C 3.45(1) to 3.84(1) Å, with another by typical oVset p–p interactions involving six membered rings of adjacent fullerenes. This is in contrast to [Ni(TMTAA)C60] which crystallises as two dimensional corrugated sheets of interacting fullerenes with the sheets separated by Ni(TMTAA) molecules.1 Here the dihedral angle defined by the two nickel centres and the C60 centroid is not linear but is markedly stepped at 83.48 whereas in the present structure it is 1808.For [Ni(TMTAA)C60] the host is small enough to skew from linearity without disrupting the host–guest interactions or engaging in destabilising host–host interactions, while allowing the fullerene a suitable surface for inter-fullerene interactions. A diVerence in this angle is expected because of the larger contact surface area of the Ni(OMTAA) with C60 for the phenyl lined face relative to the same face for Ni(TMTAA), especially given that the two host molecules associated with each fullerene in the Ni(TMTAA) are already at the van der Waals limit with respect to each other.However, the magnitude of the change is larger than expected just on steric considerations alone, and most certainly arises from the interplay of the CS2 molecules and the overall cohesion of the supramolecular array. The Ni(OMTAA) molecules occupy opposite longitudinal poles of the fullerene, intra-chain C60 molecules occupy opposite latitudinal poles, while CS2 molecules and phenyl– methyl groups from adjacent chains complete the sphere of van der Waals contacts to each C60.The CS2 molecules occupy voids created by the face of the phenyl rings not involved in host–guest binding and make contact with both C60 and Ni(OMTAA). The closest CS2 ? ? ?C60 contact involves the central carbon of the carbon disulfide [C? ? ? C 3.41(1) Å] while the sulfur atoms interact with the underside of the Ni- (OMTAA) phenyl face [C? ? ? S 3.48(1) Å] and the {NC(Me)}2- CH methyl groups [C? ? ? S 3.69(1) Å], Figs. 3 and 4.Ni(OMTAA)–C60 Solution studies The UV/vis spectra indicate the formation of a 1 : 1 association of C60 and Ni(OMTAA) in solution. Carbon disulfide solutions were prepared with Ni(OMTAA) and C60 in ratios varying from 10: 2 to 10:30 whilst keeping the concentration of Ni(OMTAA) constant (1 × 1024 mol dm23).The C60 concentration in the fullerene deficient solutions was increased incrementally and showed the presence of three isosbestic points at 467, 562 and 619 nm, indicating the presence of one host–guest species, Fig. 5. Further, prior to stoichiometric equality, the addition of C60 led to a linear decrease in the intensity of the nickel d–d transition as well as a slight shift to longer wavelength (lmax 595 to 600 nm). A similar perturbation was observed with the ligand to metal charge transfer band (lmax 447 to 456 nm).Once the C60 was in stoichiometric excess absorbance at these wavelengths increased markedly. This linear relationship between the concentration of C60 and the change in absorbance prior to stoichiometric equality, indicates the presence of a tightly bound 1: 1 host–guest species with an association constant in excess of 105 mol21 dm3. It appears most likely that in solution the fullerene is interacting almost exclusively with the phenyl containing face due to the greater van der Waals forces involved/greater surface area contact.For an excess of C60 a dramatic change in absorbance at 600 nm is observed. This may be due to the formation of micelle like species,4,10 involving clusters of fullerenes shrouded by a sheath of Ni(TMTAA) molecules. Conclusion In attempting to further understand the inclusion chemistry of the [Ni(TMTAA)C60] system we have prepared a C60 complex of Ni(OMTAA) as a CS2 adduct.The eVect of extending the phenyl arms of the macrocycle and the interplay of the CS2 molecules gives a structure with infinite linear chains of host– guest–host– species. Further, and more importantly, this array manifests itself into linear chains of close contact fullerenes. In the context of generating arrays containing fullerenes, reducing the inter-fullerene interactions into a directional array is of fundamental importance, and metal-ion doping of such an array is an attractive endeavour,17–19 as is changing the metal in the macrocycle with a view of forming reduced fullerenes via electron transfer.Experimental Measurements Elemental analyses (C,H,N) were performed by the Campbell Microanalytical laboratory, University of Otago. Infrared spectra were recorded on a Perkin-Elmer FT 1600 spectrometer, UV/vis measurements were recorded on a Varian Cary 5 spectrophotometer, and NMR spectra on a Varian Mercury 300 MHz spectrometer. Materials All reagents used were purchased from Aldrich and used without further purification.Methanol was dried over magnesium methoxide and distilled immediately prior to use. Carbon disulfide was used without further purification. Syntheses (5,14-Dihydro-2,3,6,8,11,12,15,17-octamethyldibenzo[b,i] [1,4,8,11]tetraazacyclotetradecine)nickel(II), Ni(OMTAA) and I. Nickel acetate tetrahydrate (8.4 g, 0.034 mol), 1,2-diamino- 4,5-dimethylbenzene (9.2 g, 0.068 mol), and acetylacetone (6.8 g, 0.068 mol) were combined and the mixture flushed with Fig. 5 UV/visible titration of Ni(OMTAA) with C60.J. Chem. Soc., Dalton Trans., 1999, 279–284 283 nitrogen. Dry methanol (50 mL) was added and the resulting mixture was then stirred and brought to the boil. After 48 hours at reflux, the green mixture was cooled and filtered to remove the product as a green precipitate (11.8 g, 76%). Single crystals suitable for X-ray diVraction studies were obtained by slow evaporation of carbon disulfide–hexane solutions.(Found: C, 68.12; H, 6.78; N, 11.96. NiC26H30N4 requires C, 68.30; H, 6.61; N, 12.25%). 1H NMR (solvent CDCl3, standard SiMe4): d 2.05 (s, 12H), 2.06 (s, 12H), 4.78 (s, 2H), 6.48 (s, 4H). 13C NMR (solvent CDCl3, standard CDCl3): d 19.70, 22.13, 110.37, 121.92, 129.41, 145.19, 154.68. lmax (e) (CS2): 595 (7180), 447 (16 500), 407 (20 300) nm (dm3 mol21 cm21). m/z 456.2, 471.3, 489.2, 519.2, 912.4, 927.2, 975.2. When nickel(II) actetate tetrahydrate (12.4 g, 0.05 mol), acetylacetone (10.1 g, 0.10 mol) and 1,2-diamino-4,5-dimethylbenzene (13.6 g, 0.10 mol) were reacted using 250 mL of methanol, Ni(OMTAA) was isolated in reduced yield (2.9 g, 13%) and purple crystals of I as the methanol solvate, formed on cooling (3.0 g, 12% based on 1,2-diamino-4,5-dimethylbenzene) (Found: C, 58.80; H, 7.15; N, 11.61.NiC26H30N4 requires C, 59.41; H, 7.06; N, 11.55%). 1H NMR (solvent CDCl3, standard SiMe4): d 2.12 (s, 6H), 2.15 (s, 6H), 2.26 (s, 6H), 4.99 (s, 1H), 6.78 (s, 2H), 6.96 (s, 2H). 13C NMR (solvent CDCl3, standard CDCl3): d 19.18, 20.03, 23.43, 50.75, 62.39, 109.36, 122.07, 125.22, 131.35, 133.10, 133.82, 146.76, 156.48. m/z 393.2, 785.3. [Ni(OMTAA)C60]?2CS2. To a purple solution of C60 (0.021 g, 3.0 × 1025 mol) in carbon disulfide (3 mL) was added Ni(OMTAA) (0.030 g, 6.6 × 1025 mol). The resulting green solution was evaporated to dryness depositing large black crystals. These were then washed with dichloromethane until the filtrate ran colourless (0.031 g, 78% yield) (Found: C, 78.82; H, 2.36; N, 4.48.NiC88H30S4N4 requires C, 79.46; H, 2.27; N, 4.21%). lmax (e) (CS2): 600 (6 150), 456 sh (14 100), 409 (28 290) nm (dm3 mol21 cm21). X-Ray crystallography Data were collected on an Enraf-Nonius CCD diVractometer at 123 K using graphite-monochromated Mo-Ka radiation (l = 0.71013 Å). Data were corrected for Lorentzian and polarisation but not absorption. The structure of Ni(OMTAA) was solved by direct methods using maXus,35 whilst the structures for I and [Ni(OMTAA)C60]?2CS2 were solved using TEXSAN.36 All non-hydrogen atoms were refined anisotropically using a full matrix least squares refinement against F.Hydrogen atoms were included at calculated positions with a riding model. I. C24H34N4O3Ni. Mr = 485.26 monoclinic, a = 7.9740(6), b = 23.452(2), c = 12.267(1) Å, b = 90.911(5)8, U = 2293.8 Å3, space group P21/c, Z = 4, m(Mo-Ka) = 8.70 mm21, 4917 reflections, 289 parameters, R1 = 0.0673, wR = 0.0810.Ni(OMTAA). C26H30N4Ni. Mr = 457.24. monoclinic, a = 20.2852(5), b = 11.5462(4), c = 19.0594(5) Å, b = 94.8529 (10)8, U = 4447.8 Å3, space group C2/c , Z = 8, m(Mo-Ka) = 8.935 mm21, 3164 reflections, 280 parameters, R1 = 0.042, wR = 0.046. [Ni(OMTAA)C60]?2CS2. C88H30N4S4Ni. Mr = 1330.11, monoclinic, a = 21.8510(5), b = 13.2492(4), c = 19.6201(5) Å. b = 105.7109(10), U = 5468 Å3, space group C2/c , Z = 3, m(Mo- Ka) = 6.20 mm21, 7032 reflections, 438 parameters, R1 = 0.088, wR = 0.089.CCDC reference number 186/1287. See http://www.rsc.org/suppdata/dt/1999/279/ for crystallographic files in .cif format. Acknowledgements This work is supported by the Australian Research Council. References 1 P. C. Andrews, J. L. Atwood, L. J. Barbour, P. J. Nichols and C. L. Raston, Chem. Eur. J., 1998, 4, 8, 1382. 2 J. L. Atwood, G. A. Koustantonis and C. L. Raston, Nature (London), 1994, 368, 229. 3 T.Haino, M. Yanase and Y. Fukazawa, Tetrahedron Lett., 1997, 38, 3739; A. Ikeda and S. Shinkai, Chem. Rev., 1997, 97, 1713; A. Ikeda, M. Yoshimura and S. Shinkai, Tetrahedron Lett., 1997, 38, 2107; K. Akao, K. Ikeda, T. Suzuki and S. Shinkai, ibid., 1996, 37, 73; T. Suzuki, K. Nakashima and S. Shinkai, Tetrahedron Lett., 1995, 36, 249; T. Suzuki, K. Nakashima and S. Shinkai, Chem. Lett., 1994, 699; R. M. Williams, J. M. Zwier and J. W. Verhoeven, J. Am. Chem. Soc., 1994, 116, 6965; R.M. Williams and J. W. Verhoeven, Recl. Trav. Chim. Pays-Bas, 1992, 111, 531. 4 A. Drljaca, C. Keppert, L. Spiccia, C. L. Raston, C. A. Sandoval and T. D. Smith, Chem. Commun., 1997, 195. 5 J. L. Atwood, L. J. Barbour, C. L. Raston and I. B. N. Sudria, Angew. Chem., Int. Ed. Engl., 1998, 37, 981. 6 Z. Yoshida, H. Takekuma, S. Takekuma and Y. Matsubara, Angew. Chem., Int. Ed. Engl., 1994, 33, 1597. 7 T. Haino, M. Yanase and Y. Fukozawa, Angew. Chem., Int. Ed. Engl., 1997, 36, 3, 259. 8 K. Tsubaki, K. Tanaka, T. Kinoshita and K. Fuji, Chem. Commun., 1998, 895. 9 T. Haino, M. Yanase and Y. Fukozawa, Angew. Chem., Int. Ed. Engl., 1998, 37, 7, 997. 10 C. L. Raston, J. L. Atwood, P. J. Nichols and I. B. N. Sudria, Chem. Commun., 1996, 2615. 11 J. W. Steed., P. C. Junk, J. L Atwood, M. J. Barnes, C. L. Raston and R. S. Burkhalter, J. Am. Chem. Soc., 1994, 116, 10346; J. L. Atwood, M. J. Barnes, M. G. Gardiner and C. L. Raston, Chem. Commun., 1996, 1449. 12 N.S. Isaacs, P. J. Nichols, C. L Raston, C. A. Sandoval and D. Young, Chem. Commun., 1997, 1839. 13 L. Y. Chiang, J. W. Swirczewski, K. Liang and J. Millar, Chem. Lett., 1994, 981. 14 Q. Zhu, D. E. Cox, J. E. Fischer, K. Kniaz, A. R. McGhie and O. Zhou, Nature (London), 1992, 355, 712. 15 J. D. Crane, P. B. Hitchcock, H. W. Kroto, R. Taylor and D. R. M. Walton, J. Chem. Soc., Chem. Commun., 1992, 1764. 16 P. W. Stephens, D. Cox. J. W. Lauher, L. Milhaly, J. B. Wiley, P. Allemand, A.Hirsch, K. Holczer, Q. Li, J. D. Thompson and F. Wudl, Nature (London), 1992, 355, 331. 17 A. Izuoka, T. Tachikawa, T. Sugawara, Y. Saito and H. Shinohara, Chem. Lett., 1992, 1049. 18 A. Izuoka, T. Tachikawa, T. Sugawara, Y. Suzuki, M. Konno, Y. Saito and H. Shinohara, J. Chem. Soc., Chem. Commun., 1992, 1472. 19 J. D. Crane and P. B. Hitchcock, J. Chem. Soc., Dalton Trans., 1993, 2537. 20 A. Harada and S. Takahashi, J. Chem. Soc., Chem. Commun., 1988, 1352; P. D.Godfrey, W. J. Grigsby, P. J. Nichols and C. L. Raston, J. Am. Chem. Soc., 1997, 119, 9283; R. J. Blanch, M. Williams, G. D. Fallon, M. G. Gardiner, R. Kaddour and C. L. Raston, Angew. Chem., Int. Ed. Engl., 1997, 36, 504. 21 P. C. Andrews, P. D. Croucher, P. J. Nichols and C. L. Raston, unpublished work. 22 F. A. L’Eplattenier and A. Pugin, Helv. Chim. Act., 1975, 58, 917. 23 V. L. Goedkin, M. C. Weiss, D. Place and J. Dabrowiak, Inorg. Synth., 1980, 20, 115. 24 P. D. Croucher, P. J. Nichols and C. L. Raston, unpublished work. 25 R. L. Paul, S. F. Ghellor, G. A. Heath, D. C. R. Hockless, L. M. Redina and M. Sterns, J. Chem. Soc., Dalton Trans., 1997, 4143. 26 Y. Wang, S.-M. Peng, Y.-L. Lee, M.-C. Chuang, C.-P. Tang and C.-J. Weng, J. Chin. Chem. Soc., 1982, 29, 217. 27 F. A. Cotton and J. Czuchajowska-Wiessinger, Acta Crystallogr., Sect. C, 1992, 48, 1434; G. Ricciardi, A. Bavoso, A. Rosa, F. Lelj and Y. Gizov, J. Chem. Soc., Dalton Trans., 1995, 2385. 28 M. C. Kutcha and G. Parkin, J. Am. Chem. Soc., 1994, 116, 8372. 29 M. C. Kutcha and G. Parkin, J. Chem. Soc., Chem. Commun., 1994, 1351. 30 M. C. Kutcha and G. Parkin, Chem. Commun., 1996, 1669. 31 D. A. Busatu, S. P. Nolan and E. D. Stevens, Acta Crystallogr., Sect. C, 1995, 51, 1855.284 J. Chem. Soc., Dalton Trans., 1999, 279–284 32 N. W. Alcock, J. C. Cannadine, W. Errington, P. Moore and M. G. H. Wallbridge, Acta Crystallogr., Sect. C, 1994, 50, 2037. 33 F. A. Cotton and J. Czuchajowska, Polyhedron, 1990, 9, 2553. 34 P. J. Lukes, J. A. Crayston, D. J. Ando, M. E. Harman and M. B. Hursthouse, J. Chem. Soc., Perkin 2, 1991, 1845. 35 S. Mackay, C. J. Gilmore, C. Edwards, M. Tremayne, N. Stewart and K. Shankland, maXus: a computer program for the solution and refinement of crystal structures from diVraction data, University of Glasgow, Scotland, UK, Nonius BV, Delft, The Netherlands and MacScience Co. Ltd., Yokohama, Japan, 1998. 36 TEXAN, Crystal Structure Analysis Package, Molecular Structure Corporation, Houston, TX, 1985 and 1992. Paper 8/05893H

ISSN:1477-9226    

DOI:10.1039/a805893h

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 285–291 285 The first co-ordinatively unsaturated Group 8 allenylidene complexes: insights into Grubbs’ vs. Dixneuf–Fürstner olefin metathesis catalysts Karsten J. Harlow, Anthony F. Hill * and James D. E. T. Wilton-Ely Centre for Chemical Synthesis, Department of Chemistry, Imperial College of Science, Technology and Medicine, South Kensington, London, UK SW7 2AY. E-mail: a.hill@ic.ac.uk Received 6th October 1998, Accepted 30th November 1998 The reactions of [MCl2(PPh3)3] with HC]] ] CCPh2OH provided [MCl2(]] C]] C]] CPh2)(PPh3)2] (M = Ru 1a or Os 1b) the first examples of co-ordinatively unsaturated allenylidene complexes of Group 8 metals. The phosphine ligands of 1a are labile and readily replaced by PCy3 to give [RuCl2(]] C]] C]] CPh2)(PCy3)2] 1c.Heating 1a with NaPF6 in chloroform gave the known bimetallic salt [Ru2(m-Cl)3(]] C]] C]] CPh2)2(PPh3)4]PF6 2?PF6. The reaction of 1a with carbon monoxide provided [RuCl2(]] C]] C]] Ph2)(CO)(PPh3)2] 3 which may also be prepared from [RuCl2(CO)(dmf)(PPh3)2] and HC]] ] CCPh2OH. The first macrocycle coligated allenylidene complex [RuCl(]] C]] C]] CPh2)(PPh3)([9]aneS3)]Cl 4?Cl ([9]aneS3 = 1,4,7-trithiacyclononane) was obtained from the reaction of 1a with [9]aneS3.Alternatively, 4?PF6 is also obtained by treating [RuCl2(PPh3)([9]aneS3)] sequentially with NaPF6 in acetonitrile followed by HC]] ] CCPh2OH. The reaction of 1a with dppe and NaPF6 yielded the known salt trans-[RuCl(]] C]] C]] CPh2)(dppe)2]PF6 5?PF6.The complex [RuCl(]] C]] C]] CPh2)(PCy3){HB(pz)3}] 6 (pz = pyrazol-1-yl) was obtained from the reaction of 1c with K[HB(pz)3], whilst the related benzylidene complex [RuCl(]] CHPh)(PCy3){HB(pz)3}] 7 was obtained similarly from [RuCl2(]] CHPh)(PCy3)2] and K[HB(pz)3]. Heating [Ru2(m-Cl)2Cl2(h-cym)2] (cym = iPrC6H4Me-4) with PCy3 and HC]] ] CCPh2OH in refluxing benzene provided a mixture of 1c and the bimetallic complex [Ru2(m-Cl2)Cl2(]] C]] C]] CPh2)- (h-cym)] 8 and other unidentified products. The complex 8 may however be obtained quantitatively from the reaction of 1c with [Ru2(m-Cl)2Cl2(h-cym)2].These results suggest that the active species in ring-closure olefin metathesis processes mediated by the allenylidene pre-catalyst [RuCl(]] C]] C]] CPh2)(PCy3)(h-cym)]1 in non-polar arene solvents may be allenylidene analogues of the Grubbs’ alkene metathesis catalyst, viz. 1c and 8. Introduction Reviews on the rapidly emerging chemistry of allenylidene complexes of Groups 8 1 and 9 2 reveal noteworthy points of distinction. Most significantly, the rich chemistry of Group 9 allenylidenes developed by Werner centres primarily on neutral co-ordinatively unsaturated examples, typically of the electronically advantageous d8-ML4 square planar geometry. In marked contrast, since the isolation by Selegue 3 of the archetypal complex [Ru(]] C]] C]] CPh2)(PMe3)2(h-C5H5)]1 all Group 8 examples have been co-ordinatively saturated, in the main cationic, 1b and of the ubiquitous pseudo-octahedral d6-ML6 geometry.This reflects the more strict adherence to the 18-electron rule by complexes of low-valent metals on moving to the left across the transition series. However, the advent and unprecedented utility of Grubbs’ catalysts [RuCl2(]] CHR)(PR93)2] (R = CH]] CPh2 or Ph; R9 = Ph or Cy)4 and the highly successful application of bulky phosphines (PiPr3, PCy3, PMeBut 2) within Group 85 have made d6-ML5 geometries an increasingly prevalent feature of the organometallic chemistry of ruthenium and osmium.Herein we report the synthesis of the first examples of co-ordinatively unsaturated allenylidene complexes of ruthenium and osmium, viz. [MCl2(]] C]] C]] CPh2)- (PPh3)2] (M = Ru or Os), [RuCl2(]] C]] C]] CPh2)(PCy3)2] and [Ru2(m-Cl)2Cl2(]] C]] C]] CPh2)(PCy3)(h-MeC6H4 iPr-4)]. These 16- electron complexes are conveniently accessible from commercially available starting materials and serve as versatile precursors to a range of other allenylidene complexes via facile ligand addition and/or exchange reactions. Results and discussion The vinylidene complex [RuCl2(]] C]] CHCMe3)(PPh3)2] has been shown to result from the reaction of HC]] ] CtBu with [RuCl2- (PPh3)3]. 6 A related ‘parent’ vinylidene complex [RuCl2(]] C]] CH2)(PCy3)2] was subsequently shown by Grubbs to be accessible via the reaction of [RuCl2(]] CHPh)(PCy3)2] with allene 4 whilst Katayama and Ozawa7 showed that this class of complex was capable of catalysing ring-opening metathesis polymerisation of strained bicyclo-olefins.We were prompted by the reports of this and Grubbs’ alkylidene analogues 4 to attempt the synthesis of related allenylidene complexes [RuCl2- (]] C]] C]] CPh2)(PR3)2] (R = Ph or Cy). Grubbs’ alkylidene complexes ultimately appear to lose their catalytic activity in metathesis processes via bimolecular decomposition processes involving alkylidene coupling.In the case of our proposed allenylidene target complexes, the analogous formation and elimination of a hexapentaene appeared unlikely, given that stable binuclear bis(allenylidene) complexes have been reported.8 The majority of late transition metal allenylidene complexes 1,2 arise from variants of Selegue’s ground-breaking propynol dehydration approach.3 Fortunately, this strategy has also proven successful here.The reaction of [RuCl2(PPh3)3] with HC]] ] CCPh2OH and NaPF6 in dichloromethane has been previously reported to provide the bimetallic complex [Ru2(m-Cl)3- (]] C]] C]] CPh2)2(PPh3)4]PF6. Although no intermediates were identified, it was suggested that the complex [RuCl2- (]] C]] C]] CPh2)(PPh3)2] was a plausible intermediate.8 We find that when [RuCl2(PPh3)3] is treated with an excess of HC]] ] CCPh2OH in refluxing thf (2 h) the deep red-brown complex [RuCl2(]] C]] C]] CPh2)(PPh3)2] 1a may be obtained in high yield.In a similar manner the osmium analogue [OsCl2(]] C]] C]] CPh2)(PPh3)2] 1b may be obtained in 98% yield from [OsCl2- (PPh3)3], although refluxing toluene is the solvent of choice (Scheme 1). Both complexes 1a and 1b may also be recrys-286 J. Chem. Soc., Dalton Trans., 1999, 285–291 tallised from mixtures of dichloromethane and methanol, indicating that they are comparatively stable with respect to nucleophilic attack by alcohols.This may be attributed to the neutrality of the complexes which naturally deactivates the p-acidic allenylidene towards nucleophilic attack, in contrast to many cationic ruthenium allenylidene complexes.1a Spectroscopic data for the two complexes are broadly comparable and confirm the mononuclear formulation and stereochemistry. Of particular note are the following observations. (i) Whilst 1b displays an abundant molecular ion (confirmed by isotopic simulation) in addition to a minor fragment ion due to halide ionisation, the molecular ion for 1a is very weak relative to the [M 2 Cl]1 peak.(ii) Low field carbon-13 nuclear magnetic resonances for the metal bound allenylidene carbon nuclei [1a, d 301.2; 1b, d 266.0] are observed as triplets due to cis coupling to two chemically equivalent and mutually trans phosphorus nuclei. Whilst this is also consistent with a squarebased pyramid with apical allenylidene and cis-basal phosphines, such a geometry is highly unlikely on steric grounds, and may also be discounted by the virtual triplet multiplicity of the phenyl resonances for the phosphine substituents.(iii) Characteristic infrared absorptions due to the allenylidene groups are observed for both complexes at 1939 cm21 in solution (CH2Cl2). The solid state infrared spectra of samples of 1a crystallised under various conditions reveal three bands of varying intensities in this region (Nujol: 1968, 1929, 1902 cm21) however all such samples on dissolution in dichloromethane give rise to a single absorption (1939 cm21) suggesting that solid state eVects are responsible for the splitting observed.This is perhaps not surprising, given that the allenylidene ligand protrudes considerably from the trans-Ru(PPh3)2 double cone, exposing it to packing eVects. As noted above, the majority of known ruthenium allenylidene complexes are cationic and are characterised by very intense n(C]] C]] C) infrared absorptions.With both neutral and related cationic examples (see below) in hand it becomes clear that there is a very substantial loss of relative intensity for this absorption in neutral complexes, compromising somewhat its diagnostic utility. We have also observed similar though less dramatic behaviour for the complexes [Ru(]] C]] C]] CPh2)(PPh3)2{HB(pz)3}]1 (pz = pyrazol-1-yl) and [RuCl(]] C]] C]] CPh2)(PPh3){HB(pz)3}].1b The gross mechanism by which complex 1a forms, i.e.metalmediated propynol dehydration, has considerable precedent.1,2 The more detailed pathway by which the mononuclear complex 1a is obtained under thermal conditions, whilst the binuclear salt [Ru2(m-Cl)3(]] C]] C]] CPh2)2(PPh3)4]PF6 2?PF6 is reported to form under ambient conditions, calls for comment. We find that Scheme 1 Ru Cl Cl PPh3 Ph3P CPh2 C C Ru Cl PPh3 CPh2 C C Ru Cl Ph3P Ph2C C C Cl Cl Cl Ru Cl Cl PPh3 Ph3P CPh2 C C Cl Ru Cl Ph3P PPh3 Ph2C C C Os Cl Cl PPh3 Ph3P CPh2 C C thf, heat, 2h HCºCCPh2OH thf 25 °C 2 h HCºCCPh2OH 1a RuCl2(PPh3)3 thf heat X X = Cl 2•Cl OsCl2(PPh3)3 1b HCºCCPh2OH NaPF6 X = PF6 2•PF6 NaPF6 "A" + toluene, heat, 4h if the reaction of [RuCl2(PPh3)3] and HC]] ] CCPh2OH is carried out at room temperature in thf (2 h) three species may be observed [Fig. 1(a), Scheme 1]: small amounts (25%) of 1a are observed in addition to two compounds which give rise to ABquartet patterns in the 31P-{1H} NMR spectrum.The first of these corresponds to the chloride salt [Ru2(m-Cl)3(]] C]] C]] CPh2)2(PPh3)4]Cl 2?Cl [50%: d 40.0, 37.8; J (PAPB) = 26.8 Hz], with the small variation in chemical shift relative to 2?PF6 [d 42.2, 40.8; J(PAPB) = 30 Hz] 8 being attributed to ion-pairing eVects. This could be confirmed by anion metathesis with NaPF6 or KPF6. The second compound A also giving rise to an AB system [d 49.0, 47.7; J(PAPB) = 36.9 Hz] and formed in 25% yield has yet to be identified unambiguously but possibly also corresponds to a binuclear complex which is however neutral, possessing only two halide bridges {cf.[Ru2(m-Cl)2Cl2(CS)2- (PPh3)4], see below}. Thus, the presumed intermediacy of 1a in the formation of 2?PF6 is not quite as straightforward as might have been initially supposed. Heating isolated 1a with KPF6 in chloroform provides 2?PF6, however if [NH4][PF6] is employed as the anion source a complex mixture of unidentified products results, presumably via allenylidene aminolysis.Furthermore, a solution of 1a in dichloromethane or chloroform stirred for 2 weeks at room temperature undergoes slow and only partial reversion (ca. 50%) to a mixture of 2?Cl and the unknown binuclear complex A. The formation of KCl presumably contributes to the driving force for the formation of 21 in the presence of KPF6. The reaction of [RuCl2(PPh3)3] and HC]] ] CCMe2OH was also briefly investigated. Under identical conditions to those for the synthesis of 1a (refluxing thf, 3 h) six inseparable products were observed by 31P NMR spectroscopy: the two major products corresponded to analogues of the binuclear salt 2?Cl [d 39.7, 41.9; J(PAPB) = 25.1] and the neutral binuclear complex A [d 47.5, 49.8; J(PAPB) = 37.1 Hz].The remaining four singlet resonances all occurred suYciently close to that for 1a such that unambiguous assignment could not be made. Thus, although the reaction appears to proceed in a similar manner, optimum conditions for the isolation of single products were not established.It should be noted that parallels exist in the behaviour of formally isoelectronic thiocarbonyl complexes: whilst heating [OsCl2(PPh3)3] with carbon disulfide and an excess of phosphine provides [OsCl2(CS)(PPh3)3] quantitatively,9 the analogous chemistry based on ruthenium is considerably more complicated by the facile formation of binuclear species with two or three halide bridges.10 Indeed it is only when thermally forcing Fig. 1 31P NMR spectra of the products of the reaction of [RuCl2(PPh3)3] with HC]] ] CCPh2OH in thf (a) at room temperature; (b) under reflux.J. Chem. Soc., Dalton Trans., 1999, 285–291 287 conditions are employed (refluxing xylene) that the mononuclear complex [RuCl2(CS)(OH2)(PPh3)2] is obtained in useful amounts.11 Under a variety of milder conditions however, the complexes [Ru2(m-Cl)3Cl(CS)(PPh3)3], [Ru2(m-Cl)3(CS)2- (PPh3)4]1 (analogous to 21) and [Ru2(m-Cl)2Cl2(CS)2(PPh3)4] could each be isolated.The first of these was presumed to arise from the trapping of transiently formed [RuCl2(CS)(PPh3)2] (isoelectronic with 1a) by an excess of [RuCl2(PPh3)3]. The last of these, [Ru2(m-Cl)2Cl2(CS)2(PPh3)4], is the most likely analogue of complex A, which was, however, insuYciently soluble for comparative 31P NMR data to be available.10 The complex 1a proves to be a highly versatile synthetic entry point into a range of allenylidene complexes of ruthenium(II) (Scheme 2) by virtue of (i) its co-ordinative unsaturation, (ii) the lability of one halide and (iii) the lability of one or both phosphine ligands.Carbonylation under very mild conditions (CH2Cl2, 1 atm) provides the new complex all-trans-[RuCl2- (]] C]] C]] CPh2)(CO)(PPh3)2] 3. Alternatively, we also find that complex 3 may be prepared directly via the reaction of [RuCl2- (CO)(dmf)(PPh3)2] with HC]] ] CCPh2OH in dichloromethane at room temperature. The allenylidene-associated absorption in the infrared spectrum of 3 appears at 1953 cm21 (CH2Cl2) and is considerably more intense than that of the precursor, although it is highly likely that there is a degree of coupling between the n(C]] C]] C) and n(CO) modes [n(CO) 2007 cm21].The increase in intensity may also arise from co-ordination trans to a strong p acid (CO). The allenylidene ligand gives rise to three resonances of note in the 13C-{1H} NMR spectrum at d 310.7, 198.0 and 163.4 corresponding to the a, b and g carbons of the allenylidene spine, the former showing coupling to the two cis phosphorus nuclei (13.5 Hz).The gross composition was confirmed by the appearance of a molecular ion in the FAB mass spectrum in addition to identifiable fragmentations. The first example of a macrocycle co-ligated allenylidene complex [RuCl(]] C]] C]] CPh2)(PPh3)([9]aneS3)]Cl 4?Cl results from the reaction of 1a with 1,4,7-trithiacyclononane ([9]ane- S3).We also find that this chiral complex may be obtained in two steps from the reaction of [RuCl2(PPh3)([9]aneS3)] 12,13 with KPF6 (MeOH) followed by HC]] ] CCPh2OH to provide 4?PF6. As expected, the chirality at ruthenium results in a complex 1H NMR spectrum due to the 12 diVerent chemical environments of the macrocycle protons.14 The 31P-{1H} NMR spectrum Scheme 2 Ru Ph2P PPh2 PPh2 Ph2P CPh2 C C Cl Ru Cl Cl L L CPh2 C C OC S S S Ru Cl L CPh2 C C Ru Cl Cl PCy3 Cy3P CPh2 C C Ru Cl Cy3P CPh2 C C HB(pz)3 Ru Cl Cy3P C HB(pz)3 H Ph Ru Cl Cl PCy3 Cy3P C H Ph K[HB(pz)3] 3 PCy3 5 1c CO [9]aneS3 dppe, KPF6 4+ 1a K[HB(pz)3] 6 7 L = PPh3 however consists of a singlet resonance at d 34.9 in addition to the characteristic high field PF6 heptet.The most abundant (base) peak in the FAB mass spectrum corresponds to the molecular ion, and is accompanied by assignable fragmentations involving allenylidene dissociation and ethylene eliminations from the macrocycle.The latter is a common feature of the FAB-mass spectra of [9]aneS3 complexes. The reaction of 1a with dppe and KPF6 in a refluxing 1 : 1 mixture of tetrahydrofuran and methanol provides the salt [RuCl(]] C]] C]] CPh2)- (dppe)2]PF6 5?PF6. This salt was previously obtained by heating 2?PF6 with dppe in toluene for 12 h.8 The formation of 5?PF6 from the more reactive 1a is however complete within 4 h at lower temperatures. Amongst the variants of Grubbs’ catalyst, those with bulky PCy3 coligands are found to be the most eVective.4 Accordingly, the reaction of complex 1a with PCy3 was investigated and found cleanly to provide [RuCl2(]] C]] C]] CPh2)(PCy3)2] 1c in high yield at room temperature.Although the preparation of 1c must be carried out under anaerobic conditions, once formed the complex appears stable towards aerial oxidation both in solution and in the solid state. The FAB mass spectrum features abundant peaks due to the molecular ion (10%), chloride ionisation (10%) and dissociation of one phosphine (6%).A low-field carbon-13 resonance is observed at d 293.6 showing coupling to the two phosphorus nuclei, although this is marginally smaller (7.5 Hz) than that observed for the corresponding resonances of the PPh3 derivatives 1a, 1b and 3 (10.8–13.5 Hz). We have recently described the synthesis of the allenylidene complexes [Ru(]] C]] C]] CPh2)(PPh3)2{HB(pz)3}]PF6 and [RuCl- (]] C]] C]] CPh2)(PPh3){HB(pz)3}] via the reactions of [RuCl- (PPh3)2{HB(pz)3}] 15 with HC]] ] CCPh2OH in the presence or absence, respectively, of AgPF6.1b Surprisingly, treating 1a with K[HB(pz)3] does not appear (31P NMR) to provide either of these allenylidene complexes.In contrast, the reaction of 1c with K[HB(pz)3] gives the complex [RuCl(]] C]] C]] CPh2)(PCy3)- {HB(pz)3}] 6 in 83% yield. Although a complex reaction does ensue between [RuCl(]] C]] C]] CPh2)(PPh3){HB(pz)3}] and PCy3, 6 which might be anticipated via simple phosphine exchange could not be identified (31P NMR) amongst the plethora of products.The related benzylidene complex [RuCl(]] CHPh)- (PCy3){HB(pz)3}] 7 could also be prepared via the reaction of [RuCl2(]] CHPh)(PCy3)2] with K[HB(pz)3].† Most notable amongst the characteristic spectroscopic data for 7 are the low field NMR resonances due to the alkylidene proton [d 20.06; J(PH) = 9.3 Hz] and carbon nuclei [d 333.8; J(PC) = 14.3 Hz]. The complexes 6 and 7 therefore complement the growing range of ‘C1’ p-acidic ligands multiply bonded to the ‘RuCl(PR3){HB- (pz)3}’ fragment in the complexes [RuCl(CA)(PR3){HB(pz)3}] (R = Ph or Cy; CA = CO,16,17 CS,17 CNCMe3,18 C]] CHPh19 or C]] C]] CPh2 1b).Dixneuf, Fürstner and co-workers 20 recently showed that an allenylidene salt of ruthenium [RuCl(]] C]] C]] CPh2)(PCy3)- (h-cym)]PF6 (h-cym = MeC6H4 iPr-4) could serve as a precatalyst for the ring-closure olefin metathesis of a,w-dienes.Simultaneously Grubbs 21 showed that his catalysts could be further activated by treatment with [Ru2(m-Cl)2Cl2(h-cym)2] to provide [Ru2(m-Cl)2Cl2(]] CHR)(PCy3)(h-cym)] and [RuCl2- (PCy3)(h-cym)]. This prompted us to question whether a complex of the form 1c or (in situ) [RuCl2(]] C]] C]] CPh2)(PCy3)] might be the active species arising from the Dixneuf–Fürstner pre-catalyst. Although arene dissociation could in principle generate the 12-electron species “RuCl(]] C]] C]] CPh2)(PCy3)1” (as suggested by recent photochemical studies 22) this would be unlikely to be particularly long lived unless stabilised by recoordination of cymene or the arene solvent. Disproportionation via halide transfer could however in principle lead to species akin to Grubbs’ binuclear alkylidene complexes, or altern- † Note added in proof: During the processing of this manuscript, Grubbs reported the synthesis of 7 via an identical procedure.26288 J.Chem. Soc., Dalton Trans., 1999, 285–291 atively the 14-electron species “RuCl2(]] C]] C]] CPh2)(PCy3)” directly observed in the FAB-MS studies above.These neutral species might be more likely to persist and dissolve in the non-polar arene solvent. The reaction of [Ru2(m-Cl)2Cl2(h-cym)2] with HC]] ] CCPh2OH and PCy3 (stoichiometry 1:4:6) in refluxing benzene (4 h) was therefore investigated. Amongst the species present in the crude reaction mixture it was apparent (31P NMR; Fig. 2) that complex 1c does indeed form, however in only approximately 2–3% yield.Notably, no [RuCl2(PCy3)(h-cym)] (d 25.8) was detected. However, two other major products were observed with resonances at d 41.1 and 20.3. The first of these was identified as the new binuclear complex [Ru2(m-Cl)2Cl2(]] C]] C]] CPh2)(PCy3)- (h-cym)] 8 following the unequivocal and high yield synthesis via the reaction of 1c with [Ru2(m-Cl)2Cl2(h-cym)2] (Scheme 3). Although this reaction was spectroscopically quantitative (31P NMR), the complex 8 could only be obtained free from the side-product [RuCl2(PCy3)(h-cym)] (31P NMR: d 25.9) in 73% isolated yield due to sacrificial losses incurred during separ- Fig. 2 31P NMR spectra of the products of the reaction of [Ru2(m- Cl)2Cl2(h-cym)2] with HC]] ] CCPh2OH and PCy3 in refluxing benzene. Scheme 3 Ru Cl Cl PCy3 Cy3P CPh2 C C Ru Cl Ru Cl Cl Cl Ru Cl Cl Cl Ru Cl PCy3 C C CPh2 Ru Cl Cl PCy3 Cy3P CPh2 C C Ru Cl Ru Cl Cl Cl Ru Cl Cl Cl Ru Cl PCy3 C C CPh2 Ru Cl Cl PCy3 1c HCºCCPh2OH PCy3, C6H6, heat, 4 h + 8 (+ .. . .,) 1c 8 + CH2Cl2 25 °C + ation. It also transpires that the phosphine of [RuCl2(PCy3)- (h-cym)] is suYciently labile for it also to react with 1c to provide 8, however this reaction is considerably slower than that between 1c and [Ru2(m-Cl)2Cl2(h-cym)2]. Furthermore, we find that the binuclear complex 8, once formed, is stable under ambient conditions in the presence of an excess of PCy3. The formulation of 8 follows from spectroscopic data and by analogy with the complex [Ru2(m-Cl)2Cl2(]] CHPh)(PCy3)(h-cym)] described by Dias and Grubbs.21 Most notably the allenylidene ligand gives rise to a weak infrared absorption (KBr: 1945 cm21) and a low field doublet resonance [d 310.4, J(PC) = 15.1 Hz] in the 13C-{1H} NMR spectrum.The two resonances for the (diastereotopic) methyl constituents of the cymene iPr group indicate that the molecule does not possess any element of symmetry suggesting the stereochemistry depicted in Scheme 3.The 31-P{1H} NMR peak at d 41.2 is moved approximately 8.3 ppm to low field from the precursor, and a similar low field shift was observed for the conversion of [RuCl2(]] CHPh)- (PCy3)2] into [Ru2(m-Cl)2Cl2(]] CHPh)(PCy3)(h-cym)].21 The gross formulation was also confirmed by FAB-mass spectrometry which revealed an intense molecular ion (14%) in addition to fragmentations due to loss of chloride (8%) and also cleavage of the ‘RuCl2(h-cym)’ group (33%), i.e.the presumed catalytically active species ‘RuCl2(]] C]] C]] CPh2)(PCy3)’. It is noteworthy in this context that very recently the gas phase alkene metathesis chemistry of the complex [RuCl2(]] CHPh)- (PCy2C2H4NMe3)]1 has been investigated through electrospray ionisation tandem MS experiments.23 Conclusion Co-ordinatively unsaturated allenylidene complexes of osmium and ruthenium, 1, are both stable and easily accessible from commercially available starting materials.Their synthetic utility as precursors to a range of other allenylidene complexes of ruthenium has been demonstrated with the synthesis of 16 or 18 electron and mono- or bi-nuclear derivatives. One of these, the binuclear complex [Ru2(m-Cl)2Cl2(]] C]] C]] CPh2)- (PCy3)(h-cym)] 8, has been shown also to result from the direct combination of [Ru2(m-Cl)2Cl2(h-cym)2], PCy3 and HC]] ] CCPh2OH under the conditions used for ring closure metathesis of a,w-diolefins mediated by the Dixneuf–Fürstner pre-catalyst [RuCl(]] C]] C]] CPh2)(PCy3)(h-cym)]1.These observations lead us to contend that under these conditions (80 8C, non-polar arene solvent), the Dixneuf–Fürstner pre-catalyst may well provide allenylidene analogues of both class of Grubbs’ catalyst, viz. 1c and 8. Although allenylidene analogues of both classes of complex are obtained under these conditions, the higher activity of Grubbs’ binuclear catalysts relative to his monometallic precursors 21 suggests that 8 may be the more active catalyst in the Dixneuf–Fürstner system.In this context, preliminary and on-going studies on the ring-closure metathesis of a,w-diolefins by our isolated and well defined complexes 1c and 8 indicate similar activities to those of the Grubbs’ catalysts.24 We will report on these promising results subsequently. Experimental General comments All experiments were routinely carried out under anaerobic conditions using conventional Schlenk-tube and vacuum line techniques unless otherwise stated.Solvents were distilled from appropriate drying agents and degassed prior to use. The complexes [RuCl2(PPh3)3],25 [RuCl2(]] CHPh)(PCy3)2],4 [RuCl2- (PPh3)([9]aneS3)] 13 and [RuCl2(CO)(dmf)(PPh3)2] 10a were prepared according to published procedures; [Ru2(m-Cl)2Cl2- (h-cym)2] was obtained commercially (Aldrich). All other reagents were used as received from commercial sources. Infrared, NMR and FAB-MS data were obtained using a MattsonJ.Chem. Soc., Dalton Trans., 1999, 285–291 289 Research Series IR spectrometer, JEOL JNM-EX270, and Autospec Q instruments, respectively. Phosphine-associated infrared data are not reported. For phosphine-derived 13C NMR resonances “tv” denotes a virtual triplet with ‘apparent’ coupling constants given, indicative of a trans bis(phosphine) arrangement. The FAB-mass spectra were obtained from 3-nitrobenzyl alcohol matrices and assignments are denoted by the most intense peak of isotopic envelopes confirmed by simulation; for salts M1 refers to the cationic complex.Microanalytical data were obtained from the Imperial College and University of North London Microanalytical (S.A.C.S.) services. Crystal solvates were confirmed by 1H NMR integration for dichloromethane, however this was not always possible for chloroform solvates due to overlap with phosphine resonances, or adventitious CHCl3 present in the deuteriated NMR solvent.Light petroleum refers to that fraction of boiling range 40– 60 8C. The majority of complexes reported could be recrystallised from mixtures of dichloromethane or chloroform and hexane or methanol. Preparations [RuCl2(]] C]] C]] CPh2)(PPh3)2] 1a. The complex [RuCl2(PPh3)3] (1.00 g, 1.04 mmol) and HC]] ] CCPh2OH (0.33 g, 1.59 mmol) were degassed under vacuum and then dissolved under nitrogen in degassed tetrahydrofuran (80 cm3). The mixture was heated with stirring under reflux for 2 h.All solvent was then removed under reduced pressure and the resulting oil dissolved in dichloromethane (10 cm3) to which hexane (60 cm3) was then slowly added. The resulting red-brown precipitate was filtered oV, washed with hexane (40 cm3) and dried in vacuo. Yield: 0.85 g (92%). Similar yields were obtained when the reaction was carried out on twice the scale: 2.0 g of [RuCl2(PPh3)3] provided 1.65 g (89%) of 1a. IR: (Nujol) 1968, 1929, 1902 [n(C]] C]] C)]; (CH2Cl2) 1939 cm21 [n(C]] C]] C)].NMR (CDCl3, 25 8C): 1H, d 6.61 [t, 4 H, H3,5(C6H5), J(HH) = 6.7], 7.05 [d, 2 H, H2,6(C6H5), J(HH) = 7.2 Hz] and 6.9–7.6 (m, 34 H, C6H5); 13C-{1H}, d 301.2 [t, Ca, J(PC) = 12.9], 225.4 (s, Cb), 145.0 (s, Cg), 135.0 [tv, C2,6(PC6H5), J(PC) = 5.4], 130.9 [tv, C1(PC6H5), J(PC) = 21.6], 130.2 [s, C4(PC6H5)], 128.1 [tv, C3,5(PC6H5), J(PC) = 5.4 Hz], 141.5–118.1 (C6H5); 31P-{1H}, d 29.0. FABMS: m/z (%) = 851 (12) [M 2 Cl]1; 696 (2), [M 2 C3Ph2]1; 660 (2), [M 2 Cl 2 C3Ph2]1, 625 (2), [M 2 2Cl 2 C3Ph2]1; 589 (2), [M 2 Cl 2 PPh3]1; 553 (12), [M 2 2Cl 2 PPh3]1; 363 (6), [M 2 2Cl 2 C3Ph2 2 PPh3]1; and 327 (4), [M 2 Cl 2 2PPh3]1 (Found: C, 63.5; H, 4.5.Calc. for C51H40Cl2P2Ru?1.25CH2Cl2: C, 63.2; H, 4.3%). [OsCl2(]] C]] C]] CPh2)(PPh3)2] 1b. The complex [OsCl2(PPh3)3] (0.20 g, 0.19 mmol) and HC]] ] CCPh2OH (0.09 g, 0.43 mmol) were degassed under vacuum, then dissolved under nitrogen in degassed toluene (20 cm3) and the mixture heated under reflux for 3 h.All solvent was then removed and the resulting oil dissolved in dichloromethane (3 cm3) and hexane (40 cm3) slowly added. The brown-red precipitate was filtered oV, washed with hexane (20 cm3) and dried in vacuo. Yield: 0.19 g (98%). IR: (Nujol) 1986(sh), 1933 [n(C]] C]] C)]; (CH2Cl2) 1939 cm21 [n(C]] C]] C)]. NMR (CDCl3, 25 8C): 1H, d 6.95 [t, 4 H, H3,5(C6H5), J(HH) = 7.9 Hz], 7.22, 7.49, 7.77 (m × 3, 36 H, C6H5); 13C-{1H}, d 266.0 [t, Ca, J(PC) = 10.8], 210.5 (s, Cb), 162.1 (s, Cg), 135.2 [tv, C2,6(PC6H5), J(PC) = 4.9], 134.4–128.5 [CC6H5 1 C1(PC6H5)], 130.1 [s, C4(PC6H5)] and 127.5 [tv, C3,5(PC6H5), J(PC) = 5.4 Hz]; 31P-{1H}, d 214.7.FAB-MS: m/z (%) = 977 (18), [M]1; 941 (5), [M 2 Cl]1; 750 (2), [M 2 Cl 2 C3Ph2]1; 715 (6), [M 2 PPh3]1; 677 (3), [M 2 Cl 2 PPh3]1; and 641 (12), [M 2 2Cl 2 PPh3]1 (Found: C, 66.1; H, 4.4. Calc. for C51H40Cl2OsP2?1.5C7H8: C, 66.4; H, 4.6%). [RuCl2(]] C]] C]] CPh2)(PCy3)2] 1c.The complex [RuCl2- (]] C]] C]] CPh2)(PPh3)2] 1a (0.80 g, 0.90 mmol) and PCy3 (0.68 g, 2.43 mmol) were degassed under vacuum, then dissolved under nitrogen in degassed dichloromethane (60 cm3) and the mixture stirred for 30 min at room temperature. All solvent was then removed under vacuum and the resulting oil triturated ultrasonically in methanol (20 cm3) to give a brick-red solid which was filtered oV, washed with methanol (20 cm3) and hexane (40 cm3) and dried in vacuo.Yield: 0.70 g (84%). A further crop could be obtained from the filtrate. Repeating the reaction employing 1.20 g of 1a provided similar yields, 1.10 g (88%). The phosphine exchange reaction may also be carried out conveniently in diethyl ether suspension. IR: (Nujol) 1969(sh) and 1929 cm21 [n(C]] C]] C)]. NMR (CDCl3, 25 8C): 1H, d 1.22, 1.51, 1.65, 1.91, 2.59 (m × 5, 66 H, Cy), 7.25, 7.33, 7.39, 7.49, 7.72, 8.65 (m × 6, 10 H, C6H5); 13C-{1H}, d 293.6 [t, Ca, J(PC) = 7.5], 210.0 (s, Cb), 174.1 (s, Cg), 144.5–117.2 (C6H5), 32.7 [tv, C1(C6H11), J (PC) = 8.6], 29.8 [d, C3,5(C6H11), J (PC) = 3.2], 27.8 [m, C2,6(C6H11)] and 26.5 [C4(C6H11)]; 31P-{1H}, d 32.7.FABMS: m/z (%) = 922 (10), [M]1; 887 (10), [M 2 Cl]1; 851 (0.5), [M 2 2Cl]1; 644 (6), [M 2 PCy3]1; 605 (2), [M 2 PCy3 2 Cl]1; 569 (4), [M 2 PCy3 2 2Cl]1; and 470 (100), [M 2 PCy3 2 C3Ph2] (Found: C, 66.3; H, 8.3. Calc. for C51H76Cl2P2Ru: C, 66.4; H, 8.3%). [Ru2(Ï-Cl)3(]] C]] C]] CPh2)(PPh3)4]PF6 2?PF6.The mixture of 1a and dimeric complexes A and 2?Cl obtained from the room temperature reaction of [RuCl2(PPh3)3] and HC]] ] CCPh2OH in thf (2 h) was treated with an excess (3 equivalents) of KPF6 in a mixture of dichloromethane and methanol (1 : 1) and stirred for 10 h. The solvent was removed and the residue extracted with dichloromethane. The combined extracts were filtered through diatomaceous earth and then freed of volatiles. Inspection of the phosphorus-31 NMR spectrum of the residue indicated that it was the previously reported complex 2?PF6. 31P-{1H} NMR (CDCl3, 25 8C): d 41.1, 42.4; J(PAPB) = 30.0 Hz [cf. d 40.8, 42.2; J(PAPB) = 30 Hz8]. all-trans-[RuCl2(]] C]] C]] CPh2)(CO)(PPh3)2] 3. The complex [RuCl2(CO)(dmf)(PPh3)2] (0.25 g, 0.31 mmol) was stirred with HC]] ] CCPh2OH (0.10 g, 0.45 mmol) in dichloromethane (50 cm3) at room temperature for 8 h. The solvent volume was reduced under vacuum to ca. 10 cm3. The deep pink product was precipitated by addition of light petroleum and was isolated by filtration.The crude product was recrystallised from a mixture of dichloromethane and diethyl ether, from which it was obtained as a dichloromethane hemisolvate. Yield: 0.25 g (87%). IR: (Nujol) 2001 [n(CO)], 1945 [n(C]] C]] C)]; (CH2Cl2) 2007 [n(CO)], 1953 cm21 [n(C]] C]] C)]. NMR (CDCl3, 25 8C): 1H, d 7.10, 7.31, 7.46, 7.60, 7.74, 7.88 (m × 6, 40 H, C6H5); 13C-{1H}; d 310.7 [t, Ca, J(PC) = 13.5], 198.0 (Cb), 194.3 [t, CO, J(PC) = 12.4], 163.4 (Cg), 142.3 [C1(CC6H5)], 134.4 [tv, C3,5(PC6H5), J(PC) = 5.4], 133.2 [tv, C1(PC6H5), J(PC) = 23.7], 127.7 [tv, C2,6(PC6H5), J(PC) = 4.3 Hz], 132.4 [C4(CC6H5)], 131.8 [C2,6/3,5(CC6H5)], 129.9 [C4(PC6H5)] and 128.9 [C3,5/2,6- (CC6H5)]; 31P-{1H}, d 19.0.FAB-MS: m/z (%) = 915 (1), [M]1; 879 (9), [M 2 Cl]1; 851 (6), [M 2 Cl 2 CO]1; 689 (29), [M 2 C3Ph2 2 Cl]1; 653 (4), [M 2 PPh3]1; 625 (13), [M 2 C3Ph2 2 2Cl 2 CO]1; 553 (13), [M 2 PPh3 2 2Cl 2 CO]1; and 363 (16), [M 2 PPh3 2 C3Ph2 2 2Cl 2 CO]1 (Found: C, 65.3; H, 4.3.Calc. for C52H40Cl2OP2Ru?0.5CH2Cl2: C, 65.9; H, 4.3%). [RuCl(]] C]] C]] CPh2)(PPh3)([9]aneS3)]PF6 4?PF6. A mixture of [RuCl2(]] C]] C]] CPh2)(PPh3)2] 1a (0.20 g, 0.25 mmol), 1,4,7- trithiacyclononane (0.050 g, 0.27 mmol), NaPF6 (0.08 g, 0.48 mmol) in dichloromethane (20 cm3) and ethanol (20 cm3) was stirred for 5 h and then freed of solvent under reduced pressure. The residue was extracted with dichloromethane (2 × 10 cm3) and the combined extracts filtered through diatomaceous earth to remove NaCl.The filtrate was diluted with ethanol (20 cm3) and then concentrated slowly under reduced pressure to provide290 J. Chem. Soc., Dalton Trans., 1999, 285–291 crystals of the salt. These were filtered oV, washed with cold ethanol (5 cm3) and hexane (10 cm3) and dried in vacuo. An analytical sample was obtained by recrystallisation of the crude product from a mixture of chloroform and ethanol as a chloroform monosolvate. Yield 0.13 g (62%).IR: (CH2Cl2) 1947 cm21 [n(C]] C]] C)]; (Nujol) 1941 [n(C]] C]] C)], 1311, 1282, 1263, 1128, 935, 840 cm21 (PF6). NMR (CDCl3, 25 8C): 1H, d 0.89, 1.27, 2.30, 2.59, 2.87, 3.21, 3.49 (m × 7, 12 H, SCH2), 7.30, 7.60, 7.76 (m × 3, 25 H, C6H5); 31P-{1H}, d 34.9. FAB-MS: m/z (%) = 769 (100), [M]1; 705 (8), [M 2 Cl 2 C2H4]1; 581 (17), [M 2 C3Ph2]1; 551 (55), [M 2 C2H4 2 C3Ph2]1; 515 (6), [M 2 Cl 2 C2H4 2 C3Ph2]1; and 479 (15), [M 2 C2H4 2 PPh3]1 (Found: C, 47.0; H, 3.7.Calc. for C39H37ClF6P2RuS3?CHCl3 requires C, 46.5; H, 3.7%). trans-[RuCl(]] C]] C]] CPh2)(dppe)2]PF6 5?PF6. Complex 1a (0.20 g, 0.23 mmol), dppe (0.23 g, 0.58 mmol) and KPF6 (0.08 g, 0.43 mmol) in tetrahydrofuran (10 cm3) and methanol (10 cm3) were heated under reflux for 4 h. The solvent was removed under reduced pressure and the residue extracted with dichloromethane (3 × 5 cm3). The combined extracts were filtered through diatomaceous earth.The filtrate was diluted with methanol (20 cm3) and then slowly concentrated under reduced pressure to provide red crystals which were filtered oV, washed with diethyl ether (10 cm3) and dried in vacuo. Yield 0.20 g (70%). The salt was identified by comparison of spectroscopic data [IR (Nujol) 1922 cm21; 31P-{1H} NMR d 38.3] with those previously published [IR (Nujol) 1923 cm21; 31P-{1H} NMR d 37.8].8 [RuCl(]] C]] C]] CPh2)(PCy3){HB(pz)3}] 6. A mixture of [RuCl2(]] C]] C]] CPh2)(PCy3)2] 1c (0.30 g, 0.32 mmol) and K[HB(pz)3] (0.097 g, 0.38 mmol) in dichloromethane (20 cm3) was stirred for 12 h and then freed of volatiles.The residue was extracted with hexane (3 × 15 cm3), the combined extracts were filtered through diatomaceous earth and the filtrate concentrated to ca. 5 cm3 and cooled to 230 8C overnight to provide red crystals. The isolated yield and the obtention of satisfactory elemental microanalytical data were compromised by the high solubility of the product.Yield 0.22 g (83%). A second crop of less pure material (containing small amounts of PCy3 and OPCy3) could be isolated on further cooling of the filtrate. IR (CH2Cl2): 2479 [n(BH)] and 1962 cm21 [n(C]] C]] C)]. NMR (CDCl3, 25 8C): 1H, d 0.87, 1.26, 1.42, 1.59, 1.80 (m × 5, 33 H, Cy), 3.1 [s(br), 1 H, BH], 5.40, 5.43, 5.78, 6.06, 6.35, 6.43 (1 H × 6, pz), 7.12–7.91 (m, 12 H, Ph 1 2pz) and 8.66 (s, 1 H, pz); 31P-{1H}, d 26.8(br). FAB-MS: m/z (%) = 820 (72), [M]1; [785 (81), [M 2 Cl]1; 632 (4), [M 2 C3Ph2]1; 593 (17), [M 2 Cl 2 C3Ph2]1; 540 (84), [M 2 PCy3]1; and 505 (51), [M 2 Cl 2 PCy3]1.[RuCl(]] CHPh)(PCy3){HB(pz)3}] 7. The complex [RuCl2- (]] CHPh)(PCy3)2] (0.30 g, 0.36 mmol) was treated as described above for the synthesis of 6 to provide green crystals. Yield 0.19 g (74%). IR (Nujol): 2494 [n(BH)], 1310, 1253, 1215, 1118, 1049, 888, 851 cm21. NMR (CD2Cl2, 25 8C): 1H, d 0.85, 1.21, 1.58, 1.80, 1.96 (m × 5, 33 H, Cy), 5.81, 6.04, 6.24, 6.39, 6.40, 7.57, 7.83, 7.85, 8.54 (1 H × 9, pz), 7.05, 7.48 (m × 2, 5 H, C6H5) and 20.06 [d, 1 H, Ru]] CH, J(PH) = 9.3 Hz]; 13C-{1H}, d 333.8 [d, Ru]] C, J(PC) = 14.3], 150.8 [C1(C6H5)], 145.8, 144.6, 143.4 [C3/5(pz)], 136.7, 135.5, 133.9 [C5/3(pz)], 131.5, 128.5 [C2,3,5,6(C6H5)], 130.8 [C4(C6H5)], 106.1 (2C), 105.3 [C4(pz)], 34.3 [d, C1(C6H11), J(PC) = 16.1], 28.8 [d(br), C3,5(C6H11), J(PC) = 12.5], 28.0, 27.7 [d × 2, C2,6(C6H11), J(PC) = 8.9, 10.7 Hz] and 26.2 [C4(C6H11)]; 31P-{1H}, d 33.1 (Found: C, 56.5; H, 6.6; N, 11.8.Calc. for C34H49BClN6PRu: C, 56.7; H, 6.9; N, 11.7%).† [Ru2(Ï-Cl)2Cl2(]] C]] C]] CPh2)(PCy3)(Á-cym)] 8. Complex 1c (0.16 g, 0.17 mmol) and [Ru2(m-Cl)2Cl2(h-cym)2] (0.11 g, 0.18 mmol) were dissolved in dichloromethane (10 cm3) and the mixture stirred for 1 h. All solvent was then removed under reduced pressure and propanone (5 cm3) added. The mixture was triturated in an ultrasound bath for 5 min and then the resulting suspension was filtered and the dark brown product washed with cold acetone (2 × 2 cm3) and hexane (10 cm3) and dried in vacuo.NB The filtrate and washings contain traces of 8. Yield: 0.12 g (73%). The reaction was repeated using 0.70 g of 1c to provide comparable yields: 0.55 g (76%). IR: (KBr) 1945 [n(C]] C]] C)], 1619, 1587, 1444, 1363, 1270, 1172, 1114, 1072, 1056, 1027, 1004, 916, 887 and 850; (CH2Cl2) 1951 cm21 [n(C]] C]] C)]. NMR (CDCl3, 25 8C): 1H, d 1.01, 1.55, 1.75, 1.88, 1.92 (m × 5, 33 H, Cy), 1.14, 1.29 [d × 2, 3 H × 2, CHCH3, J(HH) = 6.7], 2.15 (s, 3 H, C6H4CH3), 2.75 (h, 1 H, CHCH3), 4.92, 5.12, 5.16, 5.49 [d × 4, 4 H, C6H4, J(HH) = 5.7 Hz], 6.68, 6.97, 7.32, 7.56, 8.90 (m × 5, 10 H, C6H5); 31P-{1H}, d 41.2; 13C- {1H}, d 310.4 [d, Ca, J(PC) = 15.1], 140.9 [d, Cb, J(PC) = 4.3], 144.8–96.7 (cym 1 Ph 1 Cg), 36.9 [tv, C1(C6H11), J(PC) = 20.5], 29.4 [C3,5(C6H11)], 28.3, 27.5 [d × 2, C2,6(C6H11), J(PC) = 8.6, 10.8 Hz], 25.5 [C4(C6H11)], 22.9, 21.4, 18.3 (CH3).FAB-MS: m/z (%) = 950 (14), [M]1; 913 (9), [M 2 Cl]1; 758 (0.5), [M 2 C3Ph2]1; and 642 (31), [M 2 RuCl2(cym)]1 (Found: C, 54.3; H, 5.9. Calc. for C43H57Cl4PRu2: C, 54.4; H, 6.1%). The yellow filtrate was diluted with hexane (20 cm3) and then concentrated to ca. 5 cm3 and cooled to provide yellow crystals of [RuCl2- (PCy3)(h-cym)]. 31P-{1H} NMR (CDCl3, 25 8C): d 25.8. Reaction of [Ru2(Ï-Cl)2Cl2(Á-cym)2] with PCy3 and HC]] ] CCPh2OH A suspension of [Ru2(m-Cl)2Cl2(h-cym)2] (0.10 g, 0.16 mmol), PCy3 (0.14 g, 0.49 mmol) and HC]] ] CCPh2OH (0.070 g, 0.32 mmol) in benzene (20 cm3) was heated under reflux for 4 h and then freed of solvent. The residue was washed with hexane (2 × 10 cm3) and then dissolved in CDCl3 and the 31P-{1H} NMR spectrum measured.The results are shown in Fig. 2, which confirm the presence of complexes 1c and 8. Acknowledgements A. F. H. gratefully acknowledges the Leverhulme Trust and the Royal Society for the award of a Senior Research Fellowship.This work was supported by the EPSRC. Ruthenium salts were generously provided by Johnson Matthey Chemicals Ltd. References 1 (a) A. F. Hill, in Comprehensive Organometallic Chemistry, II, eds. E. W. Abel, F. G. A. Stone and G. Wilkinson, Pergamon, Oxford, 1995, vol. 7, pp. 348–356; (b) neutral examples have, however, recently been described, H. Werner, C. Grunwald, P. Steinert, O. Gevert and J. Wolf, J. Organomet.Chem., 1998, 565, 231; K. J. Harlow, A. F. Hill, T. Welton, A. J. P. White, D. J. Williams and J. D. E. T. Wilton-Ely, J. Organomet. Chem., 1998, in the press. 2 H. Werner, Chem. Commun., 1997, 903 and refs. therein. 3 J. P. Selegue, Organometallics, 1982, 1, 217. 4 S. T. Nguyen, R. H. Grubbs and J. W. Ziller, J. Am. Chem. Soc., 1993, 115, 9858; P. Schwab, M. B. France, J. W. Ziller and R. H. Grubbs, Angew. Chem., Int. Ed. Engl., 1995, 34, 2039; E. L. Dias, S. T. Nguyen and R.H. Grubbs, J. Am. Chem. Soc., 1997, 119, 3887. 5 M. A. Esteruelas and L. A. Oro, Adv. Organomet. Chem., 1999, 46, in preparation. 6 Y. Wakatsuki, H. Yamazaki, N. Kumegawa, T. Satoh and J. Y. Satoh, J. Am. Chem. Soc., 1991, 113, 9604. 7 H. Katayama and F. Ozawa, Chem. Lett., 1998, 67. 8 D. Touchard, S. Guesmi, M. Bouchaib, P. Haquette, A. Daridor and P. H. Dixneuf, Organometallics, 1996, 15, 2579. 9 T. J. Collins, K. R. Grundy and W. R. Roper, J. Organomet. Chem., 1982, 231, 161. 10 (a) P. W. Armit, W. J. Sime, T. A. Stephenson and L. Scott, J. Organomet. Chem., 1978, 161, 391; (b) W. J. Sime and T. A. Stephenson, J. Chem. Soc., Dalton Trans., 1979, 1045; (c) P. W. Armit, W. J. Sime and T. A. Stephenson, J. Chem. Soc., DaltonJ. Chem. Soc., Dalton Trans., 1999, 285–291 291 Trans., 1976, 2121; (d ) T. A. Stephenson, E. S. Switkes and P. W. Armit, J. Chem. Soc., Dalton Trans., 1974, 1134. 11 P. J. Brothers, C. E. L. Headford and W. R. Roper, J. Organomet. Chem., 1980, 195, C29. 12 A. F. Hill, N. W. Alcock, J. C. Cannadine and G. R. Clark, J. Organomet. Chem., 1992, 426, C40. 13 N. W. Alcock, J. C. Cannadine, G. R. Clark and A. F. Hill, J. Chem. Soc., Dalton Trans., 1993, 1131. 14 J. C. Cannadine, A. F. Hill, A. J. P. White, D. J. Williams and J. D. E. T. Wilton-Ely, Organometallics, 1996, 15, 5409. 15 N. W. Alcock, I. D. Burns, K. S. Claire and A. F. Hill, Inorg. Chem., 1992, 31, 2906; A. F. Hill and J. D. E. T. Wilton-Ely, Inorg. Synth., 1999, 33, in the press. 16 N.-Y. Sun and S. J. Simpson, J. Organomet. Chem., 1992, 434, 341. 17 I. D. Burns, A. F. Hill, A. J. P. White, D. J. Williams and J. D. E. T. Wilton-Ely, Organometallics, 1998, 17, 1552. 18 B. Buriez, I. D. Burns, A. F. Hill, A. J. P. White, D. J. Williams and J. D. E. T. Wilton-Ely, Organometallics, in the press. 19 C. Slugovc, K. Mereiter, E. Zobetz, R. Schmid and K. Kirchner, Organometallics, 1996, 15, 5275. 20 A. Fürstner, M. Picquet, C. Bruneau and P. H. Dixneuf, Chem. Commun., 1998, 1315. 21 E. L. Dias and R. H. Grubbs, Organometallics, 1998, 17, 2758. 22 M. Picquet, C. Bruneau and P. H. Dixneuf, Chem. Commun., 1998, 2249. 23 C. Hinderling, C. Aldhart and P. Chen, Angew. Chem., Int. Ed. Engl., 1998, 37, 2685. 24 A. Fürstner, M. Liebl, A. F. Hill and J. D. E. T. Wilton-Ely, unpublished work. 25 P. S. Hallman, T. A. Stephenson and G. Wilkinson, Inorg. Synth., 1970, 12, 237. 26 M. S. Sanford, L. M. Henling and R. H. Grubbs, Organometallics, 1998, 17, 5384. Paper 8/08817I

ISSN:1477-9226    

DOI:10.1039/a808817i

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J. Chem. Soc., Dalton Trans., 1997, Pages 287–292 287 DALTON Insertion of COS into Group 2 metal–ethoxide bonds; crystal structures of [Mg(OCSOEt)2(EtOH)4] and [Sr3(OCSOEt)6(EtOH)8] * Izoldi K. Bezougli,a Alan Bashall,b Mary McPartlin b and D. Michael P. Mingos a a Department of Chemistry, Imperial College of Science Technology and Medicine, South Kensington, London SW7 2AY, UK b Department of Applied Chemistry, University of North London, Holloway Road, London, UK Some compounds of the alkaline-earth metals, which result from the insertion of COS into metal–ethoxide bonds, have been synthesized and characterised.The crystal structure of the first linear trinuclear metal strontium complex [Sr3(OCSOEt)6(EtOH)8] 2, which resulted from the reaction of [{Sr(OEt)2(EtOH)4}n] with COS gas, has been determined. The octahedral monomer [Mg(OCSOEt)2(EtOH)4] 1 resulted from the insertion of a molecule of COS into the ethoxide bonds of [{Mg(OEt)2(EtOH)4}n]. The structures demonstrate three alternative co-ordination modes of the OCS(OR)2 ligand.In the last few years there has been a resurgence of interest in the chemistry of alkaline-earth-metal complexes, which can be attributed to their potential application as molecular precursors for metal oxide thin films via metal organic chemical vapour deposition (MOCVD) and sol–gel techniques. The insertion reactions of alkaline-earth-metal alkoxides and alkyls with small molecules such as SO2, CO2, COS and CS2 have been relatively neglected compared to those associated with transition metals.1,2 Darensbourg et al.3 investigated the insertion reactions of tungsten aryloxide complexes and showed that they undergo facile COS and CO2 insertion reactions.The insertion of COS into the M]O bond was irreversible whereas that of CO2 was reversible. We have previously provided the first example of insertion of COS into the calcium–methoxide bond which resulted in the formation of the thiocarbonatobridged dimer [{Ca(OCSOMe)2(MeOH)3}2].4 In this paper we describe the extension of this work to other Group 2 metal alkoxides.Specifically, we describe the COS insertion into the M]O bonds of magnesium and strontium ethoxides. The crystal structures of the products [Mg(OCSOEt)2(EtOH)4] 1 and [Sr3(OCSOEt)6(EtOH)8] 2 are described. Results and Discussion The crystalline ethanol-solvated metal ethoxides [{Mg(OEt)2- (EtOH)4}n] and [{Sr(OEt)2(EtOH)4}n] were suspended in ethanol and COS bubbled through the suspension at room temperature.An exothermic reaction occurred, reaching completion within 10 min and the products were isolated by reducing the volume. In both cases crystallisation from ethanol gave colourless crystals. On the basis of single-crystal X-ray studies, analytical data and spectroscopic measurements, the products have been formulated as the mono- and tri-meric thiocarbonato complexes [Mg(OCSOEt)2(EtOH)4] 1 and [Sr3(OCSOEt)6- (EtOH)8] 2, respectively.These characterisational results have confirmed that insertion of COS into the metal–alkoxide bonds has occurred [equations (1) and (2)]. Complexes 1 and 2 are moisture sensitive, but may be stored indefinitely under an inert atmosphere at room temperature without losing COS, although some reversible desolvation occurs. They are soluble in alcohols and co-ordinating and polar organic solvents, but have poor solubilities in hydrocarbons.* Non-SI unit employed: cal = 4.184 J.288 J. Chem. Soc., Dalton Trans., 1997, Pages 287–292 [{Mg(OEt)2(EtOH)4}n] + 2nCOS æÆ n[Mg(OCSOEt)2(EtOH)4] (1) 3[{Sr(OEt)2(EtOH)4}n] + 6nCOS æÆ n[Sr3(OCSOEt)6(EtOH)8] + 4nEtOH (2) The IR spectra of complexes 1 and 2 (as Nujol mulls between CsI plates) were studied. The band at 1554 cm21 of 1 has been assigned to the C]] O stretching mode of the COS moiety and that at 1174 cm21 to the C]] S stretching mode. Complex 2 exhibits CO stretching vibrations at 1620 and 1568 cm21, implying two distinct co-ordination modes.The C]] S and C]S stretching frequencies featured at 1176 and 930 cm21 respectively. These bands have been assigned on the basis of previously published IR data for similar compounds.1,5 Proton NMR spectroscopic studies in (CD3)2SO for complexes 1 and 2 clearly differentiate the ethanol and ethyl thiocarbonato groups. A doublet of quartets for the CH2 of the coordinated ethanol molecules was observed at d 3.41 for 1 and at 3.42 for 2, while a quartet at d 3.74 for 1 and 3.80 for 2 was assigned to the CH2 of the ethyl thiocarbonato moieties.The two different CH3 environments of the ethanol and the ethyl thiocarbonato moieties appeared in the same region as two superimposed triplets at approximately d 1.04 for both complexes 1 and 2. Similar assignments were possible in the 13C- {1H} NMR spectra of 1 and 2. A characteristic signal for the thiocarbonato carbon, OCS(OEt)2, was observed at low field, d 184.5, in the 13C-{1H} NMR spectra for both complexes.Mass spectroscopic studies using fast atom bombardment (FAB) (positive ion) techniques for complexes 1 and 2 yielded complicated fragmentation patterns which were not very helpful in their characterisation. Molecular ions were not observed, primarily because of the poor volatilities and mass-transport properties of these complexes. The DSC trace for the magnesium complex 1 [Fig. 1(a)] shows two exotherms between 23 and 140 8C.This is mirrored in the TGA plot with a 72.5% weight loss, corresponding to loss of the ethanol and the COS molecules. The subsequent weight loss of ca. 18% represents the decomposition of Mg(OEt)2 to MgO between 140 and 700 8C leaving a residue of 10.27%. The DSC trace of compound 2 [Fig. 1(b)] shows three endotherms. The first two reveal two overlapping reaction processes between 40 and 197 8C. These two endotherms are mirrored in the TGA Fig. 1 The TGA/DSC curves of complexes (a) 1 and (b) 2 spectrum with a weight loss of 52.1% representing the loss of ethanol and COS gas. The third endotherm on the DSC trace is observed at ca. 240 8C; this feature is coupled with a simultaneous loss of 11.3% on the TGA curve corresponding to the loss of the ethyl groups, to yield a strontium oxide residue by 900 8C (residue 30.0%). The observation of exotherms for 1 and endotherms for 2 with closely related complexes suggests that the thermodynamics of the decomposition process is sensitive to the nuclearity of the complex and the mode of co-ordination of the ethyl thiocarbonate ligand.Crystal structures [Mg(OCSOEt)2(EtOH)4] 1. The insertion product 1 has the centrosymmetric monomeric structure shown in Fig. 2. Selected bond lengths and angles are listed in Table 1. The insertion of COS into the two metal–ethoxide bonds has resulted in transethyl thiocarbonate ligands with four equatorial oxygen atoms from ethanol ligands completing an octahedral co-ordination geometry round the magnesium atom with O]Mg]O angles in the range 86.59–93.41(9)8.The bond distance from the metal to the oxygen donor of the ethyl thiocarbonate ligand [Mg]O(1a) 2.036(2) Å] is markedly shorter than the mean of those to the Fig. 2 Crystal structure of [Mg(OCSOEt)2(EtOH)4] 1 Table 1 Selected bond lengths (Å) and angles (8) for complex 1 Mg]O(1a) Mg]O(1b) C(2a)]O(3a) 2.036(2) 2.095(2) 1.337(5) Mg]O(1c) O(1a)]C(2a) C(2a)]S(21a) 2.071(2) 1.235(4) 1.718(4) O(1a)]Mg]O(1a9) O(1a)]Mg]O(1c9) O(1a)]Mg]O(1b9) O(1a)]Mg]O(1b) O(1b9)]Mg]O(1b) O(1a)]C(2a)]O(3a) O(3a)]C(2a)]S(21a) 180.0 86.59(9) 91.66(9) 88.34(9) 180.0 119.7(3) 113.4(3) O(1a)]Mg]O(1c) O(1c)]Mg]O(1c9) O(1c)]Mg]O(1b9) O(1c)]Mg]O(1b) C(2a)]O(1a)]Mg O(1a)]C(2a)]S(21a) C(2b)]O(1b)]Mg 93.41(9) 180.0 91.27(10) 88.73(10) 142.8(2) 126.8(3) 124.9(2) Symmetry transformation used to generate equivalent atoms: 2x, 2y, 2z.J.Chem.Soc., Dalton Trans., 1997, Pages 287–292 289 Fig. 3 Crystal structure of [Sr3(OCSOEt)6(EtOH)8] 2 two independent ethanol groups, 2.083(2) Å, the latter being similar to the value of 2.069(3) Å reported for the Mg]O bond length in hexa(ethanol)magnesium(II) chloride.6 The monomeric structure observed for [Mg(OCSOEt)2- (EtOH)4] 1, with two monodentate thiocarbonate ligands, provides a marked contrast to the trinuclear structure of 2 (see below) and to the dimeric eight-co-ordinate calcium complex [{Ca(OCSOMe)2(MeOH)3}2] 3, obtained by insertion of COS into calcium–methoxy bonds (the only previous example of this type of insertion product to be structurally characterised).4 In the calcium complex 3 the two thiocarbonate ligands adopt a bidentate bonding mode, with one also bridging to the second metal atom to give a centrosymmetric dimer; this is very similar to that in the linear polymer resulting from insertion of SO2 into the Ca]OMe bond.7 The preference of the ethyl thiocarbonate ligand to adopt a monodentate bonding mode in 1, resulting in a discrete six-co-ordinate magnesium complex, may be attributed to the small size of the magnesium(II) ion.[Sr3(OCSOEt)6(EtOH)8] 2. The product of COS insertion into the Sr]OEt bond is a novel linear trinuclear molecule 2, which is centred on a site of Ci symmetry in the crystal with a Sr(1) ? ? ? Sr(2) distance of 3.959 Å. The molecular structure is shown in Fig. 3 while selected bond lengths and angles are given in Table 2.The structure is quite unlike those observed for either the magnesium or calcium analogues. Remarkably, the central strontium atom, located on the inversion centre, is bonded only to ethyl thiocarbonate ligands, all six being involved in an eight-co-ordinate geometry at strontium [Sr(1)]O range 2.521–2.764 Å]. The two outer strontium atoms are also eight-co-ordinated, each having four ethanol ligands [Sr(2)]O 2.533–2.566 Å]; the remaining four sites are occupied by donation from bridging thiocarbonate ligands via three oxygen atoms [Sr]O 2.577–2.785(5) Å] and one sulfur atom [Sr(2)]S(1b) 3.025(3) Å].In the overall structure four of the thiocarbonate ligands are chelating, and one oxygen atom of each of the six thiocarbonate ligands is involved in forming bridges between the central and outer strontium atoms. The two symmetry-related outer metal atoms are each chelated via an oxygen and a sulfur atom of one ligand [Sr(2)]O(1b) 2.785(5), Sr(2)]S(1b) 3.025(3) Å]; the oxygen atom of both of these symmetry-related chelate rings also bridges to the central strontium atom [Sr(1)]O(1b) 2.521(5) Å].These two bridging atoms adopt trans sites in what may be envisaged as a very distorted hexagonal-bipyramidal coordination geometry at the central atom Sr(1); the ‘equatorial’ Table 2 Selected bond lengths (Å) and angles (8) for complex 2 Sr(1)]O(1b) Sr(1)]O(1a) Sr(1) ? ? ? O(3c) Sr(2)]O(1g) Sr(2)]O(1e) Sr(2)]O(1c9) Sr(2)]S(1b) O(1a)]C(2a) O(3a)]C(4a) O(1b)]C(2b) O(1c)]C(2c) C(2c)]S(1c) 2.521(5) 2.604(5) 3.157(5) 2.549(7) 2.566(6) 2.614(6) 3.025(3) 1.239(10) 1.447(10) 1.250(9) 1.254(9) 1.710(9) Sr(1)]O(1c) Sr(1)]O(3a) Sr(2)]O(1f) Sr(2)]O(1d) Sr(2)]O(1a) Sr(2)]O(1b) S(1a)]C(2a) C(2a)]O(3a) S(1b)]C(2b) C(2b)]O(3b) C(2c)]O(3c) O(3c)]C(4c) 2.570(6) 2.764(6) 2.533(6) 2.557(5) 2.577(6) 2.785(5) 1.710(10) 1.361(9) 1.708(8) 1.335(10) 1.352(9) 1.451(9) O(1a)]Sr(1)]O(1a9) O(1c)]Sr(1)]O(1c9) O(1b)]Sr(1)]O(1c) O(1b)]Sr(1)]O(1a9) O(1b)]Sr(1)]O(1a) O(1b)]Sr(1)]O(3a) O(1a)]Sr(1)]O(3a) O(1c)]Sr(1)]O(3a9) O(1b)]Sr(1)]O(3c) O(1f)]Sr(2)]O(1d) O(1f)]Sr(2)]O(1e) O(1d)]Sr(2)]O(1e) O(1g)]Sr(2)]O(1a) O(1e)]Sr(2)]O(1a) O(1g)]Sr(2)]O(1c9) O(1e)]Sr(2)]O(1c9) O(1f)]Sr(2)]O(1b) O(1d)]Sr(2)]O(1b) O(1a)]Sr(2)]O(1b) O(1f)]Sr(2)]S(1b) O(1d)]Sr(2)]S(1b) O(1a)]Sr(2)]S(1b) O(1b)]Sr(2)]S(1b) Sr(1)]O(1b)]Sr(2) 180.0 180.0 106.3(2) 108.2(2) 71.8(2) 107.5(2) 47.3(2) 104.4(2) 72.9(2) 76.3(2) 90.7(2) 86.5(2) 136.1(2) 88.6(2) 71.2(2) 73.1(2) 108.1(2) 130.5(2) 68.1(2) 98.1(2) 77.33(14) 112.27(14) 53.16(12) 96.4(2) O(1b)]Sr(1)]Sr(1b9) O(3a)]Sr(1)]O(3a9) O(1b)]Sr(1)]O(1c9) O(1c)]Sr(1)]O(1a9) O(1c)]Sr(1)]O(1a) O(1c)]Sr(1)]O(3a) O(1b)]Sr(1)]O(3a9) O(1a)]Sr(1)]O(3a9) O(1f)]Sr(2)]O(1g) O(1g)]Sr(2)]O(1g) O(1g)]Sr(2)]O(1e) O(1f)]Sr(2)]O(1a) O(1d)]Sr(2)]O(1a) O(1f)]Sr(2)]O(1c9) O(1d)]Sr(2)]O(1c9) O(1a)]Sr(2)]O(1c9) O(1g)]Sr(2)]O(1b) O(1e)]Sr(2)]O(1b) O(1c9)]Sr(2)]O(1b) O(1g)]Sr(2)]S(1b) O(1e)]Sr(2)]S(1b) O(1c9)]Sr(2)]S(1b) Sr(2)]O(1a)]Sr(1) Sr(1)]O(1c)]Sr(29) 180.0 180.0 73.7(2) 65.3(2) 114.7(2) 75.6(2) 72.5(2) 132.7(2) 149.2(2) 75.4(2) 75.6(2) 69.3(2) 145.2(2) 131.5(2) 144.2(2) 65.0(2) 99.4(2) 141.0(2) 68.8(2) 87.4(2) 159.1(2) 113.25(14) 99.7(2) 99.6(2) Symmetry transformation used to generate equivalent atoms: 2x + 1, 2y, 2z + 1.290 J.Chem. Soc., Dalton Trans., 1997, Pages 287–292 co-ordination sites are occupied by oxygen atoms from two pairs of symmetry-related ethyl thiocarbonate ligands, one pair of which is chelating [Sr(1)]O(1a) 2.604(5), Sr(1)]O(3a) 2.764(6) Å], and one is monodentate [Sr(1)]O(1c) 2.570(6), Sr(1) ? ? ? O(3c) 3.157(5) Å].One oxygen atom from each of these four ethyl thiocarbonate groups also bridges to one of the two symmetry-related outer strontium atoms [Sr(2)]O(1a) 2.577(6), Sr(2)]O(1c9) 2.614(6) Å]. A number of oligomeric complexes of strontium with oxygen ligands have been reported,8 including triangular trimers but to the best of our knowledge 2 is the first example of a linear trinuclear complex.The OCS(OEt)2 ligand in complexes 1 and 2 exhibits in total five alternative co-ordination modes which can be divided into three different classes based on the arrangement around specific metal centres as shown in Scheme 1. In the magnesium complex 1 the OCS(OEt)2 ligand is monodentate with only the terminal oxygen directly bonded to Mg [Scheme 1(a)]; in the strontium compound 2 it exhibits three different co-ordination modes. First, it acts as a bidentate ligand with both the terminal oxygen and sulfur atoms co-ordinated to the central strontium [Scheme 1(d)].Secondly, there is an alternative bidentate mode, chelating via two oxygen donors [(e)]. Finally, it adopts a monodentate co-ordination as in the magnesium complex 1, but in this case the terminal oxygen is also involved in bridging two strontium atoms [(b) and (c)].The two bidentate modes adopted by the OCS(OEt)2 ligand in the strontium complex 2 are analogous to those adopted by the OCS(OMe)2 ligand in the calcium complex 3. The terminal oxygen of the OCS(OR)2 group is the most nucleophilic. The different co-ordination modes observed in these complexes (Scheme 1) suggest that the nucleophilic character decreases in the order shown in Scheme 2. A metal ion such as Mg2+ which is a hard or class ‘a’ metal ion has a preference to co-ordinate exclusively through the terminal oxygen, i.e. in a monodentate fashion [Scheme 1(a)]; it usually adopts a sixco- ordinate geometry with small oxygen ligands.Moving down Scheme 1 Various co-ordination modes adopted by the OCS(OR)2 ligand in complexes 1–3. The metal atoms indicated in bold type emphasise alternative co-ordination modes for that particular metal ion Scheme 2 the group to the heavier alkaline-earth metals Ca2+ and Sr2+, the increase in ionic radii and basicity results in the preference for higher co-ordination numbers and the introduction of the S and OR groups into the co-ordination sphere in order to make up the co-ordination number.The O? ? ? HO hydrogen bonding between the two halves of the molecule in 3, which appears to assist oligomerisation, is absent in 2 where there is an increased number of bridging thiocarbonato groups. Comparison of the alkyl thiocarbonato-complexes 1, 3 and 2 of Mg, Ca and Sr demonstrates clearly how the increasing size of the metal ion leads to more oligomerisation and increased utilisation of the S and OR donor groups.Magnesium is able to achieve coordinative saturation by utilising only the terminal oxygen of the ligand and the ethanol donor groups. Each of the calcium ions in the dimeric methyl thiocarbonato-complex is eight-coordinate with two ligands bridging the two metals. The trinuclear strontium complex also exhibits an eight-co-ordinate geometry at the metals with all ethyl thiocarbonate ligands adopting bridging modes.This research has demonstrated that the insertion of COS into Group 2 metal–alkoxide bonds leads to a range of crystalline derivatives which are reasonably air stable and soluble in organic solvents. The crystallographic structural determinations of complexes 1–3 have revealed interesting alternatives. The OCS(OR)2 ligand appears to be a flexible ligand which can use its alternative donor sites and bridging modes to coordinate to a wide range of metals and thereby form a series of crystalline derivatives with low molecular weights.The thermogravimetric results indicate that the insertion of the COS molecule may be reversed at higher temperatures. The reaction of COS with the Group 2 alkoxides has shown a sequence going down the Group. The weakly bound sulfur atoms of the terminal alkyl thiocarbonate ligands suggest that complexes 2 and 3 may react further with class ‘b’ metal ions to form mixedmetal complexes.Experimental General procedures All manipulations were carried out under an atmosphere of dry nitrogen using standard glove-box (Miller-Howe FF160) and Schlenk techniques. All solvents were rigorously dried and deoxygenated by standard procedures. The samples for NMR and infrared studies were handled in a glove-box, but those for microanalysis and thermogravimetric analysis were not. This has lead to some differences concerning the extent to which ethanol molecules of crystallisation were detected by the spectroscopic and analytical techniques.Instrumentation Infrared spectra were recorded on a Perkin-Elmer FTIR 1720 spectrometer using Nujol mulls between 25 × 4 mm CsI plates. The Nujol was dried with 4 Å molecular sieves prior to use (and stored in a glove-box); the samples were protected from the atmosphere by an O-ring-sealed Presslok holder (Aldrich Chemicals). The NMR spectra were recorded on a JEOL GS 270 MHz spectrometer, 1H referenced internally to the residual 1H impurity present in the deuteriated solvent.Chemical shifts are recorded in parts per million (d) relative to SiMe4 (d = 0) using (CD3)2SO (d = 2.52). The 13C NMR spectra are referenced to (CD3)2SO (d = 40.6). Controlled thermal analyses of the complexes were investigated using a Polymer Laboratories 1500H simultaneous thermal analyser, controlled by an Omni- Pro 486DX-33 personal computer. The mass of the samples investigated was between 10 and 25 mg.The measurements were carried out in alumina crucibles under an atmosphere of flowing (25 cm3 min21) nitrogen gas, using heating rates of 5 8C min21.J. Chem. Soc., Dalton Trans., 1997, Pages 287–292 291 Table 3 Crystal data and structure refinement details for compounds 1 and 2 1 2 Formula Mr Crystal system Space group a/Å b/Å c/Å a/8 b/8 g/8 U/Å3 Z Dc/Mg m23 F(000) Crystal size/mm l/Å m(Mo–Ka) or m(Cu–Ka)/mm21 q Range/8 hkl Ranges Reflections collected Unique reflections Minimum and maximum transmission Data, restraints, parameters Goodness of fit on F2, S Final R1, wR2 I > 2s(I) All data Weighting, w21 Largest peak and hole/e Å23 C14H34MgO8S2 418.84 Monoclinic P21/c (no. 14) 7.408(1) 16.253(1) 9.864(1) 104.96(1) 1147.4(1) 2 1.212 452 0.40 × 0.16 × 0.14 1.541 78 2.660 5.4–55.0 27 to 7, 217 to 17, 210 to 10 3132 1404 0.430, 0.676 1404, 25, 128 1.031 0.0645, 0.1719 0.0714, 0.1801 [s2(Fo)2 + (0.193P)2 + 0.73P] 0.388, 20.537 C34H78O20S6Sr3 1262.18 Triclinic P1� (no. 2) 11.290(3) 11.761(2) 13.138(4) 73.78(2) 87.78(30) 62.30(2) 1474.6(6) 1 1.421 652 0.32 × 0.40 × 0.50 0.710 73 2.976 1.6–25.0 21 to 12, 212 to 13, 215 to 15 5839 5004 0.512, 0.584 5003, 139, 322 1.003 0.0677, 0.1124 0.1637, 0.1491 [s2(Fo)2 + (0.0414P)2 + 2.16P] 0.604, 20.626 S = [ow(Fo 2 2 Fc 2)2/(n 2 p)]� �� , R1 = o |Fo| 2 |Fc| /o |Fo|, wR2 = ow(Fo 2 2 Fc 2)2/o [w(Fo 2)2]� �� , P = [max(Fo 2, 0) + 2(Fc 2)]/3 where n = number of reflections and p = total number of parameters.Starting materials Strontium granules and dibutylmagnesium were obtained from Aldrich Chemicals Co. and were used as received. Preparations Tetra(ethanol)bis(ethyl thiocarbonato)magnesium, [Mg(OCSOEt) 2(EtOH)4] 1. Dibutylmagnesium in heptane (20 cm3, 20 mmol) was added to ethanol (30 cm3) at 240 8C resulting in an exothermic reaction. The reaction mixture was slowly warmed to room temperature, the solvent reduced in volume until all the heptane was removed and precipitation of the white solid [{Mg(OEt)2(EtOH)4}n] was observed.Addition of ethanol (30 cm3) resulted in the formation of a suspension of the magnesium ethoxide. Carbonyl sulfide gas was bubbled through the suspension at room temperature. This resulted in a vigorous exothermic reaction and dissolution of the ethoxide to yield a yellow solution, which was stirred for 1 h. A crystalline solid was isolated after cooling the solution to 220 8C (yield 6.21 g, 74.3%) (Found: C, 31.2; H, 5.8.Calc. for C6H10MgO4S2: C, 30.8; H, 4.3%). The analysis is based on the unsolvated molecular formula. IR (cm21) (Nujol): 3076m, 1554s, 1462s, 1376m, 1262s, 1174s, 1091s, 1050s, 884s, 803s, 724m, 690m, 525w, 462m, 396s and 336m. NMR [(CD3)2SO, 20 8C]: 1H (270 MHz), d 4.41 (t, CH3CH2OH, 4 H), 3.74 (q, OCSOCH2CH3, 4 H), 3.41 (m, CH3CH2OH, 8 H) and 1.04 (t, CH3, OCSOR/ROH, 18 H); 13C (67.94 MHz), d 183.42 (s, OCSOR), 59.53 (s, OCSOCH2CH3), 56.66 (s, CH3CH2OH), 19.19 (s, CH3CH2OH) and 15.62 (s, OCSOCH2CH3).Mass spectrum (positive-ion FAB): m/z 235, [Mg(OCSOEt)2]+; 282, [Mg(OCSOEt)2(EtOH)]+. Octakis(ethanol)hexakis(ethyl thiocarbonato)tristrontium, [Sr3(OCSOEt)6(EtOH)8] 2. Strontium metal (1.2 g, 13.7 mmol) was suspended in ethanol (50 cm3) and the mixture refluxed for 2 h resulting in dissolution of the metal and evolution of hydrogen gas, yielding a clear solution. Carbonyl sulfide gas was bubbled through the solution at room temperature resulting in an exothermic reaction to give a yellow solution, which was stirred for 1 h.A crystalline solid was isolated by cooling the solution to 220 8C (yield 3.87 g, 67%) (Found: C, 21.6; H, 3.2. Calc. for C18H38O16S6Sr3: C, 22.4; H, 3.9%) [analysis based on Sr3(OCSOEt)6?4H2O]. IR (cm21) (Nujol): 3195w, 1620m, 1568s, 1466s, 1377s, 1302w, 1261w, 1176s, 1152w, 1085s, 1051s, 930w, 890m, 801m, 721w, 693m, 681m, 557w and 529m.NMR [(CD3)2SO, 20 8C]: 1H (270 MHz), d 4.39 (t, CH3CH2OH, 8 H), 3.80 (q, OCSOCH2CH3, 12 H), 3.42 (m, CH3CH2OH, 16 H) and 1.04 (t, CH3, OCSOR/ROH, 42 H); 13C (67.94 MHz), d 184.76 (s, OCSOR), 60.13 (s, OCSOCH2CH3), 56.67 (s, CH3- CH2OH), 19.19 (s, CH3CH2OH) and 15.15 (s, OCSOCH2CH3). Mass spectrum (positive-ion FAB): m/z 193, [Sr(OCSOEt)]+; 299, [Sr(OCSOEt)2]+; 346, [Sr(OCSOEt)2(EtOH)]+ and 596, [Sr2(OCSOEt)4]+. X-Ray crystallography Data were collected using a Siemens P4 diffractometer equipped with a Siemens LT2 low-temperature device with graphite-monochromated radiation using w–2q scans at 173 K.No significant decay in the intensity of three standard reflections measured after every 100 was observed. The data were corrected for Lorentz-polarisation factors and for absorption (y scans). The crystal data, data collection and refinement details are summarised in Table 3. Both structures were solved by direct methods and in each case all non-hydrogen atoms were located from subsequent Fourier-difference syntheses.All non-hydrogen atoms were assigned anisotropic thermal parameters and refined using fullmatrix least squares on Fo 2.9 The hydrogen atoms for each of the compounds were included at calculated positions with C]H bond distances of 0.99 and 0.98 Å for the methylene and methyl groups respectively. The hydroxylic hydrogens in 1 were located in a Fourier-difference synthesis but those in 2 were not located.During refinement all the hydrogens were allowed to292 J. Chem. Soc., Dalton Trans., 1997, Pages 287–292 ride on their parent atom and assigned isotropic thermal parameters equal to 1.2Ueq of the parent atom for the methylene groups and 1.5Ueq for the methyl and hydroxyl groups. In complex 1 the methyl group of one ethanol C(3c)/C(4c) was disordered (60/40) over two sites. In 2 two of the ethanol groups were disordered; in one case the methylene carbon was disordered (50/50) over two sites C(2e)/C(2e9) and for the second ethanol two distinct conformations (50/50) were resolved C(2g), C(3g) and C(4g), C(5g).Atomic cooral parameters, and bond lengths and angles have been deposited at the Cambridge Crystallographic Data Centre (CCDC). See Instructions for Authors, J. Chem. Soc., Dalton Trans., 1997, Issue 1. Any request to the CCDC for this material should quote the full literature citation and the reference number 186/318. Acknowledgements The EPSRC is thanked for financial support and BP plc for endowing D.M. P. M.’s Chair. References 1 K. K. Pandey, Coord. Chem. Rev., 1995, 140, 37 and refs. therein. 2 M. H. Chisholm, Abstr. Papers Am. Chem. Soc., 1984, 187, 205; A. Wojcicki, Adv. Organomet. Chem., 1974, 12, 31; W. Kitching and C. W. Fong, Organomet. Chem. Rev. A, 1970, 5, 281; G. J. Kubas, Acc. Chem. Res., 1994, 27, 183; W. A. Schenk, Angew. Chem., Int. Ed. Engl., 1987, 26, 98; R. R. Ryan, G.J. Kubas, D. C. Moody and P. G. Eller, Struct. Bonding (Berlin), 1981, 46, 47. 3 D. J. Darensbourg, K. M. Sanchez and A. L. Rheingold, J. Am. Chem. Soc., 1987, 109, 290; D. J. Darensbourg, B. L. Mueller, C. J. Bischoff, S. S. Chojnacki and J. H. Reibenspies, Inorg. Chem., 1991, 30, 2418; D. J. Darensbourg, B. L. Mueller and J. H. Reibenspies, J. Organomet. Chem., 1993, 451, 83; D. J. Darensbourg, K. M. Sanchez, J. H. Reibenspies and A. L. Rheingold, J. Am. Chem. Soc., 1989, 111, 7094. 4 V. C. Arunasalam, D. M. P. Mingos, J. C. Plakatouras, I. Baxter, M. B. Hursthouse and K. M. A. Malik, Polyhedron, 1995, 14, 1105. 5 A. J. Goodsel and G. Blyholder, J. Am. Chem. Soc., 1972, 94, 6725. 6 G. Valle, G. Baruzzi, G. Paganetto, G. Depaoli, R. Zannetti and A. Marigo, Inorg. Chim. Acta, 1989, 156, 157. 7 V. C. Arunasalam, I. Baxter, M. B. Hursthouse, K. M. A. Malik, D. M. P. Mingos and J. C. Plakatouras, J. Chem. Soc., Chem. Commun., 1994, 2695. 8 S. R. Drake, S. A. S. Miller, M. B. Hursthouse and K. M. A. Malik, Polyhedron, 1993, 12, 1621; S. R. Drake, M. B. Hursthouse, K. M. A. Malik and D. J. Otway, J. Chem. Soc., Dalton Trans., 1993, 2883; S. R. Drake, W. E. Streib, M. H. Chisholm and K. G. Caulton, Inorg. Chem., 1990, 29, 2707; K. G. Caulton, M. H. Chisholm, S. R. Drake, K. Folting, J. C. Huffman and W. E. Streib, Inorg. Chem., 1993, 32, 1970; K. G. Caulton, M. H. Chisholm, S. R. Drake, K. Folting and J. C. Huffman, Inorg. Chem., 1993, 32, 816; R. D. Rogers, M. L. Jezl and C. B. Bauer, Inorg. Chem., 1994, 33, 5682; S. R. Drake, M. B. Hursthouse, K. M. A. Malik and S. A. S. Miller, J. Chem. Soc., Chem. Commun., 1993, 478 and refs. therein. 9 SHELXTL, PC version 5.03, Siemens Analytical Instruments Inc., Madison, WI, 1994. Received 13th August 1996; Paper 6/05678D

ISSN:1477-9226    

DOI:10.1039/a605678d

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 293–299 293 Arylsulfonylnitrene and arenesulfonyl azide complexes of palladium Isabella Foch, László Párkányi, Gábor Besenyei,* László I. Simándi and Alajos Kálmán Institute of Chemistry, Chemical Research Center, Hungarian Academy of Sciences, H-1525 Budapest, PO Box 17, Hungary. E-mail: besenyei@cric.chemres.hu Received 14th July 1998, Accepted 18th November 1998 Reaction of [Pd2Cl2(dppm)2] 1 with arenesulfonyl azides RSO2N3 in CH2Cl2 at ambient temperature resulted in the formation of dimeric palladium complexes with bridging arylsulfonylnitrene ligands [Pd2Cl2(dppm)2(m-NSO2R)] 2 (R = phenyl a, 4-methylphenyl b, 4-nitrophenyl c, 2-nitrophenyl d, 2-naphthyl e or ferrocenyl f).According to their NMR spectra, complexes 2a–2f are typical A-frame adducts, which has been verified by single crystal X-ray diVraction studies on 2d–2f. The short N(1)–S(1) distances [1.541(3)–1.565(3) Å] reflect the double bond character of these bonds and indicate a dp–pp interaction of the sulfur atom and the imido nitrogen atom.With 2-nitrobenzenesulfonyl azide a by-product was observed in a yield of 10–15%. Based on spectroscopic properties (1H and 31P NMR, IR) and elemental analysis, this compound contains a bridging arenesulfonyl azide ligand, [Pd2Cl2(dppm)2(m-2-O2NC6- H4SO2N3)] 3d. Irradiation with light of l > 455 nm caused selective dissociation of 3d into 1 and 2-O2NC6H4SO2N3. The azide complex 3d is not an intermediate in the formation of the corresponding nitrene complex 2d.Introduction Chemically reactive species, such as carbenes and nitrenes, can often be stabilised as ligands in metal complexes.1 The stability of nitrene complexes is especially apparent with high-valent transition metals, where multiple bonding is facilitated also by the strong p-donor properties of the imido (nitrene) ligand. The isolation of stable nitrene complexes of certain transition metals under a given set of conditions may indicate that reactive nitrene intermediates are possibly involved in catalytic reactions occurring under similar conditions. This may be a useful tool in elucidating the mechanisms of some complex reactions. Mechanistic considerations and the search for new synthetic strategies are the major factors that have contributed to the exceptional growth of research activity associated with transition metal imido complexes.In the course of our studies on the chemical properties of palladium complexes depicted by the general formula of [Pd2Cl2(P–P)2] [where P–P denotes bis(diphenylphosphino)- methane, dppm, or 1,1-bis(diphenylphosphino)ethane, dppmMe] we have observed that the palladium dimers [Pd2Cl2- (dppm)2], [Pd2Cl2(dppm)(dppmMe)] and syn-[Pd2Cl2(dppm- Me)2] can preserve the short-lived sulfur monoxide and selenium monoxide in their co-ordination spheres.2 Earlier, we prepared a series of these complexes and characterised them by spectroscopic methods.The structure of the [Pd2Cl2(dppm)2- (m-SO)] adduct has been determined by X-ray crystallography. Later, we became interested in palladium-stabilised nitrene complexes. This interest was initiated by our observations showing that the catalytic carbonylation of certain arenesulfonamide derivatives oVers new phosgene-free routes for the industrially important arenesulfonyl isocyanates.3 We supposed palladium nitrene complexes to be intermediates in these catalytic transformations. This turned our attention toward the synthesis, structural and chemical properties of palladium complexes containing co-ordinated arylsulfonylnitrene ligands. The search for a suitable synthetic route to these complexes revealed that the interaction of the dimeric complex [Pd2Cl2(dppm)2] with arenesulfonyl azides results in the formation of dinuclear, sulfonylnitrene bridged adducts [Pd2Cl2(dppm)2(m-NSO2R)] (R = aryl).4 Owing to their theoretical interest and potential synthetic utility, the chemistry of nitrene complexes is well explored. The literature data accumulated over the past decades have been reviewed from various aspects.1,5–7 While metal centres with empty d orbitals can eVectively stabilise imido ligands, nitrene complexes of metals belonging to Group 10 (VIII) are rare,1d,8 and we are not aware of any palladium sulfonylnitrene complexes other than those described in our preliminary report.4 Complexes with arylsulfonylnitrene moieties ligated to mononuclear centres of Ru, W and Mo have been described, however.9–11 In this paper we report the synthesis and structural characterisation of a series of novel [Pd2Cl2(dppm)2(m-NSO2R)] A-frame derivatives, including the results of single crystal X-ray diVraction studies of complexes where R = 2-nitrophenyl, 2-naphthyl or ferrocenyl.Spectroscopic evidence will also be presented supporting the formation of an arenesulfonyl azide complex, [Pd2Cl2(dppm)2(m-N3SO2C6H4NO2-2)], as a by-product in the reaction of [Pd2Cl2(dppm)2] with 2-nitrobenzenesulfonyl azide.Results and discussion Reaction of the palladium complex [Pd2Cl2(dppm)2] 1 with arenesulfonyl azides, RSO2N3 results in the formation of novel A-frame adducts [Pd2Cl2(dppm)2(m-NSO2R)] 2. The reaction takes place smoothly in CH2Cl2 at ambient temperature and is accompanied with evolution of stoichiometric amounts of nitrogen gas, as determined by volumetric experiments [eqn.(1)] (R = C6H5 a, 4-CH3C6H4 b, 4-O2NC6H4 c, 2-O2NC6H4 d, 2-naphthyl e or ferrocenyl f). [Pd2Cl2(dppm)2] 1 RSO2N3 æÆ 1 [Pd2Cl2(dppm)2(m-NSO2R)] 1 N2 (1) 2 The 1H NMR spectra are consistent with the apical position of the arylsulfonylimido moiety by showing the expected doublet/quintet splitting pattern of the methylene hydrogen atoms (see Table 1). The chemical shift values and the magnitude of the J(HH) and J(HP) coupling constants are similar to those observed in related spin systems.12 The novelty of dimetallic arylsulfonylnitrene complexes prompted us to carry out detailed structural investigations.The294 J. Chem. Soc., Dalton Trans., 1999, 293–299 Table 1 The NMR spectroscopic data of compounds 2a–2f and 3d a Complex 2a 2b 2c 2d 2e 2f 3d R Phenyl 4-Methylphenyl b 4-Nitrophenyl 2-Nitrophenyl 2-Naphthyl Ferrocenyl 2-Nitrophenyl dH 4.01 (2 H, dq) 2.58 (2 H, dq) 4.10 (2 H, dq) 2.57 (2 H, dq) 2.22 (3 H, s) 3.89 (2 H, dq) 2.66 (2 H, dq) 4.17 (2 H, dq) 2.65 (2 H, dq) 4.13 (2 H, dq) 2.53 (2 H, dq) 4.27 (2 H, dq) 4.08 (2 H, m) 4.04 (5 H, s) 3.71 (2 H, m) 2.53 (2 H, dq) 2.93 (2 H, dq) 2.56 (2 H, dq) JHH/Hz 12.8 12.8 12.8 12.8 12.8 12.8 12.9 12.9 12.7 12.7 12.5 12.5 13.6 13.6 JHP/Hz 5.8 3.4 5.9 3.4 5.8 3.5 5.8 3.3 5.9 3.4 6.2 3.3 5.1 3.1 dP 7.1 (s) 7.2 (s) 6.5 (s) 6.0 (s) 7.1 (s) 7.6 (s) 13.5 c 15.0 a Spectra were recorded in CDCl3 and referenced to internal TMS or external 85% H3PO4, respectively; d = doublet, q = quintet, m = multiplet, s = singlet.b 13C NMR: d 21.4 (CH3Ph) and 24.5 (CH2). c AA9BB9 system. geometrical arrangement deduced from the spectroscopic features of the complexes has been confirmed by single crystal X-ray diVraction studies in the case of 2d–2f. Selected bond lengths and angles are presented in Table 2. Figs. 1–3 show typical A-frame structures in which the diphosphine ligands are arranged in trans-trans positions.The N-arylsulfonyl moieties occupy the apical sites and the methylene carbon atoms are bent toward the bridging nitrogen atom, forming thereby the well known “boat” conformation. An overview of the structural data reveals the close structural similarity of 2d–2f. The Pd(1) ? ? ? Pd(2) distances vary between 3.231 and 3.350 Å, excluding any bonding metal–metal interaction {cf. the Pd–Pd distance is only 2.699(5) Å in the related [Pd2Br2(dppm)2] complex13}. The Pd–N bond lengths are in the range of 2.007(4) and 2.029(3) Å.To our knowledge, there is only one crystallographically characterised palladium complex with a bridging nitrene unit,8 in which the respective distances lie between 2.009 and 2.071 Å. The Pd–N bond lengths in 2d–2f are in line with the single bond character of the palladium– nitrogen interaction as a metal–nitrogen single bond is expected to be in the region of 1.9–2.15.14 In contrast, with other metals the imido nitrogen atom tends to form an additional dp–pp bond with the metal centres, and bond orders of higher than one are common.1d Fig. 1 Molecular structure of [Pd2Cl2(dppm)2(m-NSO2C6H4NO2-2)] 2d. The bond lengths and angles around the palladium centres do not diVer significantly from the corresponding data of other [Pd2Cl2(dppm)2(m-X)] derivatives 15 and reflect a slightly distorted square planar geometry around the metal ions. The short S(1)–N(1) bond distance [1.541(3)–1.565(3) Å] is a further common feature of the structures 2d–2f.The sum of covalent radii of sulfur and nitrogen atoms is 1.70 Å therefore the S–N bonds [and, as indicated by the similar values of n(SO2) stretching bands, in 2a–2c as well, see Table 3] should be regarded as having strong double bond character. This structural feature suggests an extensive dp–pp interaction between the bridging imido nitrogen atom and the sulfur atom. The sum of angles around the nitrogen atoms in 2d–2f, which approaches 3608, gives further evidence for this additional bond formation. The shift of electron density toward the sulfonyl group is also reflected by the position of the nasym(SO2) and nsym(SO2) bands, showing a shift of 60–80 cm21 to lower wavenumbers, as compared with the same bands recorded for sulfonamides. A planar trigonal symmetry about the bridging nitrogen atom has been observed in the structurally characterised [Rh2(CO)2(dppm)2(m-NC6H4NO2-4)] 16 and [Ir2(CO)2(dppm)2- (m-NC6H4CH3-4)] 17 A-frame adducts.An analysis of the crystallographic data published on [Pd3(PEt3)3(m-NPh)2(m-NHPh)] reveals, however, that none of the nitrogen atoms seem to follow this behaviour.8 Fig. 2 Molecular structure of complex 2e, molecule 2.J. Chem. Soc., Dalton Trans., 1999, 293–299 295 To our knowledge, structurally characterised derivatives of ferrocenesulfonamide are rare. A comparison of the C(Cp)–S and S–N bond lengths in 2f with the respective data of ferrocenesulfonamide and the sulfonylurea RSO2NHC(O)NHBu (R = ferrocenyl) 18 reveals that electron donating properties of the substituents on the nitrogen atom bring about changes of opposite directions in these distances.This observation is in conformity with the conclusions drawn upon a statistical analysis of derivatives containing the CSO2N fragment.19 Interestingly, the disposition of the ferrocenesulfonyl unit is Fig. 3 Molecular structure of [Pd2Cl2(dppm)2(m-NSO2R)] 2f (R = ferrocenyl).Table 2 Selected bond lengths (Å) and angles (8) of complexes 2d–2f 2e Pd(1) ? ? ? Pd(2) Pd(1)–N(1) Pd(2)–N(1) Pd(1)–Cl(1) Pd(2)–Cl(2) N(1)–S(1) S(1)–O(1) S(1)–O(2) S(1)–C(3) Pd(1)–P(1) Pd(1)–P(2) Pd(2)–P(3) Pd(2)–P(4) Pd(1)–N(1)–Pd(2) Pd(1)–N(1)–S(1) Pd(2)–N(1)–S(1) C(3)–S(1)–N(1) O(1)–S(1)–O(2) Cl(1)–Pd(1)–N(1) Cl(2)–Pd(2)–N(1) P(1)–Pd(1)–P(2) P(3)–Pd(2)–P(4) Pd(2)–N(1)–S(1)–O(2) Pd(2)–N(1)–S(1)–O(1) P(1)–Pd(1)–Pd(2)–P(4) P(3)–Pd(1)–Pd(2)–P(2) O(4)–N(2)–C(4)–C(3) 2d 3.281(1) 2.019(3) 2.029(3) 2.316(1) 2.306(1) 1.541(3) 1.444(3) 1.446(3) 1.801(4) 2.327(1) 2.333(1) 2.330(1) 2.316(1) 108.3(2) 125.7(2) 122.3(2) 108.2(2) 115.0(2) 177.1(1) 175.1(1) 174.4(1) 175.1(1) 126.6(2) 26.4(3) 175.7(1) 168.3(1) 57.8(7) molecule 1 3.288(1) 2.016(4) 2.013(4) 2.330(1) 2.328(1) 1.556(4) 1.434(4) 1.445(4) 1.759(6) 2.323(1) 2.357(1) 2.318(1) 2.312(2) 109.4(2) 125.5(2) 123.8(2) 105.7(2) 114.5(3) 177.5(1) 177.2(1) 174.4(1) 175.5(1) 101.3(3) 229.6(4) 169.8(1) 169.5(1) molecule 2 3.231(1) 2.007(4) 2.019(4) 2.342(2) 2.318(2) 1.562(4) 1.422(4) 1.450(4) 1.790(7) 2.336(2) 2.331(2) 2.313(2) 2.324(2) 106.8(2) 125.0(2) 124.2(2) 108.3(2) 116.2(3) 175.5(1) 179.1(1) 174.1(1) 177.3(1) 121.5(3) 210.4(4) 172.7(1) 172.7(1) 2f 3.350(1) 2.010(2) 2.010(2) 2.328(1) 2.334(1) 1.565(3) 1.436(2) 1.442(2) 1.762(3) 2.315(1) 2.315(1) 2.333(1) 2.326(1) 112.8(1) 118.5(1) 124.7(1) 107.5(2) 116.2(2) 174.1(1) 177.1(1) 171.6(1) 173.6(1) 145.9(2) 15.0(2) 2169.4(1) 2178.8(1) the most symmetrical among the three arylsulfonyl moieties.This is reflected by the almost equal O(1) ? ? ? P(3) and O(2) ? ? ? P(1) distances (3.267 and 3.409 Å) and also by the Pd(2)–N(1)–S(1)–O torsion angles (see Table 2). The central arrangement of the ferrocenyl moiety is apparently due to the lack of any asymmetric substitution on the Cp ring. The other two aryl groups are unsymmetrically substituted with respect to the S(1)–C(3) bonds and this results in twisted orientations of the arylsulfonyl moieties to achieve minimum steric interactions.When the aryl group is naphthyl there are two independent molecules in the asymmetric unit. Molecule 1 is diVerent from 2 in that the naphthyl group is rotated by 1808 about the S(1)– C(3) axis. The diVerent orientation of the naphthyl moiety is accompanied by other small structural changes like twisting about the N(1)–S(1) bond and modifications in the Pd(1) ? ? ? Pd(2) distance and Pd(1)–N(1)–Pd(2) angle.Sharp and co-workers 16 have observed that an imido/amido equilibrium is established readily with [Rh2(CO)2(dppm)2- (m-NR)] complexes. The ratio of these two components depends on the nature of R, and only the imido form could be detected when R was the electron-withdrawing substituent 4-O2NC6H4. In agreement with these findings, our complexes do not exhibit any sign of imido/amido tautomerism and the 1H NMR spectra show a doublet/quintet splitting pattern for the methylene hydrogen atoms, whose integration is always correct for two protons in both resonances. Additional proof for the lack of interaction of the bridging nitrogen atom with the methylene group of dppm ligands comes from the C(1)–P and C(2)–P bond lengths [1.811(6)–1.845(4) Å], which are ca. 0.1 Å longer than those observed in P–CH–P bridges. A further analogy with the rhodium compounds is that complexes 2b and 2d, when subjected to carbonylation, did not show any tendency for CO insertion and no isocyanate formation was observed.Our original aim was to study the carbonylation of palladium nitrene complexes to shed light on the mechanistic details of two-step oxidative carbonylation reactions.3 Although reaction (1) shows selectivities better than 95% for most of the sulfonyl azides (1H NMR), we observed a byproduct with 2-nitrobenzenesulfonyl azide, which formed in 10–15%. EVorts to isolate and identify this compound have caused an unexpected turn in our research.If the raw product obtained from complex 1 and 2-O2NC6- H4SO2N3 (via evaporation of the reaction mixture to dryness) was analysed by 1H NMR spectroscopy resonances of weak intensity were observed at d 2.56 and 2.93. The 31P NMR spectrum was also indicative of the presence of a by-product having non-equivalent phosphorus nuclei. While purification of the raw reaction mixture to obtain the pure nitrene adduct could be accomplished easily, the by-product proved to be elusive and our first attempts to isolate and characterise it were unsuccessful.By taking advantage of the better solubility of the nitrene complex, we enriched the samples to have the unknown component as a major constituent. A series of the IR spectra illustrates the spectral changes occurring during this operation (Fig. 4). We have been unable to obtain crystals of the unknown component suitable for X-ray diVraction studies, therefore its characterisation is based on NMR and IR spectroscopy and elemental analysis.Spectra in Fig. 4 clearly show that not only the nitrene complex 2d but also the by-product contains a sulfonyl group. In the former case it was characterised by the nasym(SO2) and nsym(SO2) stretching vibrations at 1252 and 1110 cm21, while bands at 1311 and 1152 cm21 were assigned to the same vibrations of the unknown component. The shift of n(SO2) bands to lower wavelengths indicates that conjugation of the sulfonyl group to its substituents is by far not as extensive as that in the nitrene complex.Other peaks at 1536, 1484, 1435, 1356 and 1098296 J. Chem. Soc., Dalton Trans., 1999, 293–299 Table 3 The UV–VIS and IR data of complexes 2a–2f and 3d UV–VIS IR (KBr, cm21) Compound 2a 2b 2c 2d 2e 2f 3d l/nm 262 (sh) 329 262 (sh) 329 262 (sh) 331 271 (sh) 331 265 (sh) 328 262 (sh) 328 329 b 417 b e/M21 cm21 41600 19300 42200 19100 43400 21300 35400 18700 44500 19900 44700 19300 26400 5200 nasym(NO2) 1524 1536 1540 nsym(NO2) 1346 1365 1369 nasym(SO2) 1242 1241 1249 1252 1245 1233 1241 1311 nsym(SO2) 1106a 1105 a 1107 a 1110 (sh) 1106 (sh) 1098 a 1152 a Owing to overlap with other vibrations, assignments of nsym(SO2) bands were done by a deconvolution of the 1260–1050 cm21 region.b Obtained after subtracting contribution of complex 2d, as determined from the 1H NMR spectcrum of a mixture of 3d and 2d. cm21 did not vary on going from the pure nitrene complex (top spectrum) to the mixtures containing the unknown component.These observations allow us to conclude that the by-product is also a derivative of the palladium dimer and the sulfonyl azide, and both the nitrene complex 2d and the unknown component exhibit the same basic structural features. An important fact is that the increasing percentage of the unknown component in the sample was accompanied with an increasing nitrogen content (see Experimental section). A good agreement between the experimental and calculated elemental analysis data was found when the unknown component in the mixtures was assumed to be the azide complex [Pd2Cl2(dppm)2- (2-O2NC6H4SO2N3)] 3d.The 1H NMR spectrum of the by-product clearly shows the presence of two doublets of quintets, which can readily be attributed to non-equivalent methylene hydrogen atoms. The diVerence of the chemical shifts of the methylene protons is remarkably less than that in the spectra of the nitrene complexes but the values of the J(HH) and J(HP) coupling constants, which are more informative with respect to the structural relationship, emphasise the structural similarities of the azide Fig. 4 The IR spectra of complex 2d (A), and mixtures of 2d and 3d: B, 13% (raw reaction product); C, 38% and D, 85% of 3d. and nitrene adducts. Unexpectedly, the 31P NMR spectrum reflects the non-equivalence of the phosphorus nuclei, which form an AA9BB9 spin system. In M2(P–P)2Ln complexes, the non-equivalence of the phosphorus nuclei is typically attained in two ways: (i) the phosphorus atoms of a diphosphine ligand are equivalent but diVerent from those of the other diphosphine, or (ii) the trans phosphorus atoms are equivalent but the P nuclei of the bridging ligands are magnetically diVerent.The former case results in J(HP) couplings of 10–15 Hz and the methylene hydrogen atoms appear as doublets of triplets.2 In the latter version the methylene hydrogen atoms are coupled not only to the adjacent but to the remote P atoms as well thereby keeping the quintet splitting pattern.20,21 As for the structure of 3d, it can be concluded that the magnetic inequivalence of the P nuclei is induced by the azide ligand that is supposed to have a fixed orientation similar to that of the diazonium ligand in the related platinum dimer.20 The NMR spectral data of 3d unambiguously demonstrate that this complex has a typical A-frame structure and the position of the apical azide ligand is asymmetric with respect to the phosphorus nuclei attached to the same metal centre.Although azido complexes of transition metals are common, complexes containing a co-ordinated organic azide are rare 22–24 and, to our knowledge, no complexes ligating an arenesulfonyl azide ligand have previously been reported. Therefore, our aim was to acquire further evidence to support our conclusion. Although mass spectroscopy with fast atom bombardment ionisation proved to be a convenient way of characterising sulfonylnitrene complexes, we have not been able to detect the molecular ion of the proposed azide complex.Instead, it was the nitrene adduct that was observed in MS measurements (M1, m/z 1252). The appearance of this peak was not surprising as we know that the sample analysed did indeed contain some (ª15%) nitrene complex (see Experimental section). The lack of the molecular ion corresponding to the azide complex can be explained by assuming that this compound decomposes to the nitrene adduct during the procedure of sample preparation or, alternatively, nitrogen is lost (and perhaps other fragmentation processes occur) upon ionisation of the sample.The MS spectrum of 3d, however, does provide some information with respect to the azide adduct by displaying an ion cluster around m/z 1094, completely absent in the MS spectrum of 2d. We ascribed this group of ions to the fragment [Pd2Cl2(dppm)2- N3]1.The experimental and simulated distribution of peak intensities showed good agreement. The appearance of the fragment at m/z 1094 coincides with the behaviour of 2- nitrobenzenesulfonyl azide under conditions of the MS experiments. Spectra recorded by using both EI and FAB techniquesJ. Chem. Soc., Dalton Trans., 1999, 293–299 297 showed that bond breaking between the sulfonyl and the azide groups is the major route of fragmentation, because the peak at m/z 186 was more intense than that at m/z 228 in both spectra.In order to obtain further evidence that the by-product is an azide complex, we made attempts at transforming 3d to the nitrene adduct. When searching for the right conditions for this reaction, we found absorption bands of the free palladium dimer [Pd2Cl2(dppm)2] in the UV–VIS spectrum of a solution of the enriched mixture. Evaporation of an aliquot of this solution to dryness and analysis of the solid residue by IR spectroscopy revealed the presence of free 2-nitrobenzenesulfonyl azide [nasym(N3) 2145 cm21].This observation unequivocally shows that the by-product is an azide complex and its formulation as [Pd2Cl2(dppm)2(2-O2NC6H4SO2N3)] is correct. Further examination of this remarkable reaction of complex 3d has revealed that the dissociation of the azide complex is a photolytic process. A solution of 3d in CH2Cl2 did not show any detectable change in the UV–VIS spectrum for 22 h at ambient temperature, if it was kept in the dark.Irradiation of the solution by sunlight or by the filtered light (l > 455 nm) of a xenon lamp resulted in the rapid decomposition of the azide complex to [Pd2Cl2(dppm)2] and 2-O2NC6H4SO2N3. The products of photolysis were identified by 1H NMR, UV–VIS and IR spectroscopy and TLC. If the photolysis was accomplished in the presence of 4-CH3C6H4SO2N3 or 4-O2NC6H4SO2N3, the respective nitrene complexes 2b and 2c could be readily recognised in the 1H NMR spectra.The details of the experiments are described in the Experimental section. The 1H NMR spectrum of a photolysed sample of complex 3d is consistent with the presence of 1, 2d and 3d in a ratio of 45 : 33 : 22. Detection of 3d does not necessarily mean, however, that the decomposition of the photolysed sample was not complete. A comparison of the results obtained under varying conditions allows to conclude that 3d in this sample is not a residue of the original complex but a product of a secondary reaction of 1 and 2-O2NC6H4SO2N3 obtained photolytically.Complex 3d may be formed during the evaporation of the solvent dichloromethane, and also during sample preparation and NMR analysis. This is shown by the experiments carried out in the presence of a 10 molar excess of 4-CH3C6H4SO2N3 or a 5 molar excess of 4-O2NC6H4SO2N3. In these cases no 3d could be observed in the photolysed samples because the free added azides trapped 1, thereby preventing the reaction between the primary photolytic decomposition products 1 and 2-O2NC6- H4SO2N3.Following an identical reasoning, we may conclude that the increased proportion of the nitrene complex 2d should also be attributed to a secondary reaction of 1 and 2-O2NC6- H4SO2N3 and not a direct transformation of 3d to 2d. This conclusion may be deduced from the observation that photolysis of 3d in the presence of 4-O2NC6H4SO2N3 results in the predominant formation of 2c and leads to only a negligible, if any, increase of the percentage of 2d, as shown by a comparison of the composition of the starting mixture (15% 2d) with that of the photolysed product (18% 2d).The nitrene complex 2d is, however, the major constituent of the reaction mixture if a photolysed solution of 3d is allowed to stand for 2 d. Although the azide complex 3d can be converted into the nitrene adduct 2d this is not a straightforward process but is preceded by the photolytic decomposition of 3d to its constituents 1 and 2-O2NC6H4SO2N3.The chemical behaviour of 3d is consistent with the conclusion that the azide complex is not an intermediate but a product of a side reaction. Conversion of 3d into 2d represents a pathway which is mechanistically diVerent from those described for related transformations.22–24 The ability of organic azides to react with double and triple bonds with the formation of N-heterocycles is well known from organic chemistry.25 The analogous interaction of a triple bonded molybdenum dimer with organic azides has been studied as an organometallic model reaction of (213) cycloaddition. 26 Although conversion of the organic azide into both nitrene and azide ligands was equally detected, formation of the expected five-membered dimetallatriazene ring could not be observed.Reaction (1) can also be considered as an example of cycloaddition across a palladium–palladium bond (Scheme 1). Organic azides are known to have amphiphilic character, but in this reaction they will certainly behave as electrophiles with respect to the Pd–Pd bond.As a 1,3-dipole, the sulfonyl azide carries a positive charge on the remote nitrogen atom in one of its major resonance forms. It seems likely that the sulfonyl azide approaches the reactive part of the dimer with the rod-like azide group to minimise steric interactions. The electrophilic azide group can cause polarisation of the Pd–Pd bond, resulting in the formation of an ionic intermediate A, which may rearrange to B or, in sterically controlled cases, to C.Such steric hindrance may emerge when an ortho substituent is present on the aryl group. We believe that the formation of complex 3d is due to the presence of the ortho substituent on the phenyl ring in 2-O2NC6H4SO2N3. Species B is the intermediate from which nitrogen can be extruded with ease, resulting in the formation of a nitrene adduct, as has been demonstrated in organic reactions 25 and also postulated with organometallic reagents.24,26 The structure type of species C is stable in the case of dimetallic complexes and is thought to be the right description of 3d as well.Spectroscopic features of 3d are in line with this proposal. Photodissociation of complex 3d is likely to occur through an intermediate diVerent from B. Therefore, formation of 2d does not necessarily take place during decomposition to 1 and 2-O2NC6H4SO2N3.We know of only one light-driven dissociation among [Pd2Cl2(dppm)2(m-X)] type adducts, viz. for X representing co-ordinated acetylenes.27 In our case, however, breaking of the Pd–N bonds can be accomplished by irradiation with photons of much lower energy. Work on the synthesis and characterisation of complexes with sterically more demanding substituents on the aryl group is in progress in our laboratory. Experimental Palladium dimer 1 and the sulfonyl azides were prepared by established procedures.28,29 Solvents were of analytical grade used without further purification. The NMR spectra were recorded at 30 8C on a Varian Unity Inova spectrometer at 400 (1H), 101 (13C) and 162 MHz (31P).Chemical shifts are reported in ppm units and are referenced to TMS and 85% H3PO4. The IR spectra were obtained on a Scheme 1 Pd—Pd represents 1; R is the same as in a–f; in C, R denotes 2-nitrophenyl. RSO2 N N N N Pd Pd N N(d–) SO2R N SO2R N N Pd Pd N Pd N N Pd SO2R SO2R N Pd Pd – – – –(–) (+) Pd Pd + (d+) A B C hn –N2298 J.Chem. Soc., Dalton Trans., 1999, 293–299 Table 4 Crystal data, data collection and refinement parameters for complexes 2d–2f Empirical formula Formula weight Crystal system Space group a/Å b/Å c/Å b/8 V/Å3 Z Dc/Mg m23 m/mm21 F(000) Crystal colour Crystal description Crystal size/mm Maximum and minimum transmission Reflections collected Independent reflections Reflections >2s(I) Data/restraints/parameters Goodness of fit on F2 Final R1, wR2 [I > 2s(I)] (all data) Largest diVerence peak and hole/e Å23 2d C56H48Cl2N2O4P4Pd2S?CHCl3 1371.97 Monoclinic P21/n 12.827(1) 19.622(1) 23.149(6) 98.74(2) 5758.7(16) 4 1.582 1.051 2760 Yellow Block 0.45 × 0.30 × 0.08 1.0000 and 0.8451 15932 15932 [R(int) = 0.0182] 7997 15932/88/760 0.909 0.0471, 0.1046 0.1618, 0.12010 1.199 and 21.012 2e C60H51Cl2NO2P4Pd2S?0.5CHCl3 1344.02 Monoclinic P21/n 21.714(4) 14.489(3) 38.342(7) 90.39(1) 12063(4) 8 1.480 0.937 5427 Yellow Block 0.30 × 0.20 × 0.18 1.0000 and 0.9759 25644 25644 10598 25644/1703/1407 0.813 0.0469, 0.0915 0.1604, 0.1092 0.816 and 20.937 2f C60H53Cl2FeNO2P4Pd2S?C3H6O 1373.60 Orthorhombic Pbca 12.414(2) 22.267(2) 44.002(5) 12163(3) 8 1.500 1.093 5568 Yellow Block 0.50 × 0.40 × 0.18 1.000 and 0.903 18026 17628 [R(int) = 0.0095] 8915 17628/154/696 0.843 0.0401, 0.0859 0.1215, 0.0984 0.795 and 21.039 Nicolet 205 FTIR spectrometer in KBr pellets, UV–VIS spectra on a HP 8453 diode array spectrophotometer using CH2Cl2 as solvent.The FABMS analyses were carried out on a VG ZAB2-SEQ tandem mass spectrometer on samples prepared in a 3-nitrobenzyl alcohol or thioglycerol matrix. Elemental analyses were performed on vacuum dried samples (ethanol reflux, ca. 0.1 Torr). Crystal structure determinations of complexes 2d–2f The determination of the unit cell parameters and intensity data collections were performed on an Enraf-Nonius CAD4 diVractometer at 293(2) K with Mo-Ka radiation (l 0.71073 Å).Crystal data and refinement details are listed in Table 4. Unit cell parameters were determined by least-squares of the setting angles of 25 reflections in the q ranges 15.49–15.92 (2d), 12.43–12.95 (2e) and 11.25–11.978 (2f). The w–2q scan method was used for the data collection for 2d and w scans were applied for 2e and 2f. The structures were solved by direct methods using the SHELXS 9730 (2d, 2f) and the SIR 92 31 (2e) program and subsequent diVerence syntheses.The SHELXL 9732 program was used for anisotropic full-matrix leastsquares refinement on F2. Hydrogen atoms were included in structure factor calculations but not refined. Their isotropic displacement parameters were approximated from the U(eq) value of the atom to which they were bonded [U(H) = 1.3U(eq)]. High displacement parameters observed for the C14p–C18p and the C43p–C48p phenyl rings in complex 2f may indicate disorder, but no attempt was made to resolve the splitting of the atomic positions.Neutral atomic scattering factors and anomalous scattering factors were taken from ref. 33. CCDC reference number 186/1258. See http://www.rsc.org/suppdata/dt/1999/293/ for crystallographic files in .cif format. Preparation of nitrene complexes [Pd2Cl2(dppm)2(Ï-NSO2R)] 2a–2f To a solution of 0.21 g (0.2 mmol) of complex 1 in 4 cm3 of CH2Cl2 0.4 mmol of RSO2N3 was added in 1 cm3 of CH2Cl2 in one portion.The solution was stirred at ambient temperature (ca. 20 8C) until all the palladium dimer was consumed (1–2 h). The reaction can be monitored by TLC (dichloromethane–ethyl acetate 20 :1). When the reaction was complete, the solvent was reduced to half volume and the product precipitated with methanol. The microcrystalline solid was collected on a glass filter and washed several times with methanol and diethyl ether to remove the excess of sulfonyl azide. Dissolution of the product in CH2Cl2 and reprecipitation with methanol gave azide-free samples with isolated yields better than 75% (Found: C, 55.21; H, 3.97; N, 1.20; S, 2.35.C56H49Cl2- NO2P4Pd2S 2a requires C, 55.69; H, 4.09; N, 1.16; S, 2.66. Found: C, 55.78; H, 4.08; N, 1.20; S, 2.59%; [M 1 H]1, m/z 1222. C57H51Cl2NO2P4Pd2S 2b requires C, 56.04; H, 4.21; N, 1.15; S, 2.62%; M, m/z 1221. Found: C, 53.31; H, 3.81; N, 2.31; S, 2.55. C56H48Cl2N2O4P4Pd2S 2c requires C, 53.69; H, 3.86; N, 2.24; S, 2.56.Found: C, 53.55; H, 3.86; N, 2.27; S, 2.20%. M1, m/z 1252. C56H48Cl2N2O4P4Pd2S 2d requires C, 53.69; H, 3.86; N, 2.24; S, 2.56. Found: C, 57.02; H, 4.02; N, 1.16; S, 2.35. C60H51Cl2NO2P4Pd2S 2e requires C, 57.30; H, 4.09; N, 1.11; S, 2.55. Found: C, 54.30; H, 3.98; N, 1.13; S, 2.28. C60H53Cl2FeNO2P4Pd2S 2f requires C, 54.78; H, 4.06; N, 1.06; S, 2.44%). Isolation of the azide complex [Pd2Cl2(dppm)2(2-O2NC6H4SO2- N3)] 3d The complex [Pd2Cl2(dppm)2] (1.05 g, 1.0 mmol) in 12 cm3 of CH2Cl2 was allowed to react with 0.5 g (2.2 mmol) of 2-nitrobenzenesulfonyl azide, added in 2 cm3 of CH2Cl2 in one portion.A Schlenk tube with the reaction mixture was attached to a bubbler to release the nitrogen evolved and was covered with an aluminium foil to protect the solution from light. When the palladium dimer could not be detected by TLC (in ca. 2–3 h) two thirds of the solvent was removed under vacuum and the product precipitated with methanol.The yellow crystals were collected on a glass filter and the nitrene complex removed by fractional dissolution. This procedure was carried out by adding acetone in 5 cm3 portions, which were removed from the solid by suction after a few minutes of stirring the slurry on the filter with a spatula. The change of the ratio of the nitreneJ. Chem. Soc., Dalton Trans., 1999, 293–299 299 and azide complex could be monitored by TLC (chloroform– acetone 30 : 1) or by IR spectroscopy.An intermediate stage of this process is shown in Fig. 4, spectrum C. Integration of the 1H NMR spectrum gave an azide content of 38%. This ratio of 2d and 3d requires a nitrogen content of 3.06%, in good agreement with the experimentally found value of 3.02%. When the concentration of the azide complex reached ca. 60% the remaining solid was dissolved in CHCl3, filtered through a Celite layer, concentrated and reprecipitated with methanol. Two or three fractions were collected, from which the first one (80–100 mg) generally contained more than 80% of azide complex.It is advisable to protect solutions of 3d from sunlight during manipulations. Spectrum D in Fig. 4 was recorded on a sample with 85% azide complex (Found: C, 52.07; H, 3.79; N, 4.11; S, 2.11. A 85 : 15 mixture of 3d and 2d requires C, 52.72; H, 3.79; N, 4.06; S, 2.50%). Photolysis of complex 3d A 400 W xenon lamp was used as a light source, which was filtered with a 420 nm filter at the exit of the beam.A 455 nm filter was applied to cover the height of the Schlenk tube filled with the solution of the azide complex. The distance between the lamp and the reaction mixture was 50 cm. The solution was magnetically stirred and the decomposition was monitored by UV–VIS spectroscopy. Physical measurements and reactions were carried out with samples containing about 85% of the azide complex 3d (Fig. 4, spectrum D). The azide complex (11.0 mg) was dissolved in 18.0 cm3 of CH2Cl2 and irradiated as described above. UV–VIS spectra showed a strong decrease at 330 nm, while bands at 360 and 418 nm increased in intensity.After 40 min of irradiation the solution was cooled to ca. 210 8C and evaporated to dryness using an oil pump and a liquid nitrogen trap. The solid residue was dissolved quantitatively in CDCl3 and the 1H NMR spectrum recorded. Three palladium complexes could be identified, the main component being the palladium dimer 1 (45%), accompanied with 2d (33%) and 3d (22%) (based on integration of the 1H NMR spectra).Photolysis of complex 3d in the presence of 4-nitrobenzenesulfonyl azide The photolysis was carried out in the same way as described above, but in the presence of 5 molar equivalents of the azide. The UV–VIS spectra did not show the formation of 1, which is attributable to the fast reaction of [Pd2Cl2(dppm)2] with the azide. After the photolysis was stopped the reaction mixture was stirred for 25 min.The 1H NMR spectrum was consistent with the presence of two A-frame adducts, 2c (82%) and 2d (18%). No azide complex could be observed in the 1H NMR spectrum. If the photolysis was carried out in the presence of 10 molar equivalents of 4-CH3C6H4SO2N3, nitrene complexes 2b and 2d were observed in a ratio of 3 : 2. Acknowledgements This work was supported by the Hungarian Research Fund (OTKA Grant 16213). We thank Professor A. Neszmélyi (Institute of Chemistry, Chemical Research Center, Budapest) and Dr G.Szalontai (University of Veszprém, Hungary) for discussions related to the structure of the azide complex. References 1 (a) F. A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry. A comprehensive text, Wiley, New York, 1980, ch. 4; (b) W. A. Nugent and J. M. Mayer, Metal–Ligand Multiple Bonds: The Chemistry of Transition Metal Complexes Containing Oxo, Nitrido, Imido, Alkylidene, or Alkylidyne Ligands, Wiley, New York, 1988; (c) W.P. GriYth, Coord. Chem. Rev., 1972, 8, 369; (d) D. E. Wigley, Prog. Inorg. Chem., 1994, 42, 239. 2 G. Besenyei, L. Párkányi, L. I. Simándi and B. R. James, Inorg. Chem., 1995, 24, 6118; G. Besenyei, C. L. Lee, Y. Xie and B. R. James, Inorg. Chem., 1991, 20, 2446; G. Besenyei, C. L. Lee, J. Gulinski, S. J. Rettig, B. R. James, D. A. Nelson and M. A. Lilga, Inorg. Chem., 1987, 26, 3622. 3 G. Besenyei, S. Németh and L. I. Simándi, Tetrahedron Lett., 1994, 35, 9609; G.Besenyei and L. I. Simándi, Tetrahedron Lett., 1993, 34, 2839; G. Besenyei, S. Németh and L. I. Simándi, Angew. Chem., Int. Ed. Engl., 1990, 29, 1147. 4 G. Besenyei, L. Párkányi, I. Foch, L. I. Simándi and A. Kálmán, Chem. Commun., 1997, 1143. 5 W. A. Nugent and B. L. Haymore, Coord. Chem. Rev., 1980, 31, 123. 6 K. Dehnicke and J. Strähle, Chem. Rev., 1993, 93, 981. 7 S. Cenini and G. La Monica, Inorg. Chim. Acta, 1976, 18, 279. 8 S. W. Lee and W. C. Trogler, Inorg. Chem., 1990, 29, 1099. 9 W. H. Leung, M. C. Wu, J. L. C. Chim and W. T. Wong, Inorg. Chem., 1996, 35, 4801. 10 P. J. Pérez, P. S. White, M. Brookhart and J. L. Templeton, Inorg. Chem., 1994, 33, 6050. 11 E. W. Harlan and R. H. Holm, J. Am. Chem. Soc., 1990, 112, 186. 12 A. L. Balch, Homogeneous Catalysis with Metal Phosphine Complexes, ed. L. Pignolet, Plenum, New York, 1983, p. 167. 13 R. G. Holloway, B. R. Penfold, R. Colton and M. J. McCormick, J. Chem. Soc., Chem. Commun., 1976, 485. 14 B. R. Davis, N. C. Payne and J. A. Ibers, Inorg. Chem., 1969, 8, 2719. 15 A. L. Balch, L. S. Benner and M. M. Olmstead, Inorg. Chem., 1979, 18, 2996. 16 Y. W. Ge, F. Peng and P. R. Sharp, J. Am. Chem. Soc., 1990, 112, 2632. 17 C. Ye and P. R. Sharp, Inorg. Chem., 1995, 34, 55. 18 G. Besenyei, L. Párkányi, S. Németh and L. I. Simándi, J. Organomet. Chem., 1998, 563, 81. 19 P. Bombitz, M. Czugler, A. Kálmán and I. Kapovits, Acta Crystallogr., Sect. B, 1996, 52, 720. 20 F. Neve, M. Ghedini, G. De Munno and A. Crispini, Inorg. Chem., 1992, 31, 2979. 21 B. F. Hoskins, R. J. Steen and T. W. Turney, J. Chem. Soc., Dalton Trans., 1984, 1831. 22 G. Proulx and R. G. Bergman, J. Am. Chem. Soc., 1995, 117, 6382. 23 M. G. Fickes, W. M. Davis and C. C. Cummins, J. Am. Chem. Soc., 1995, 117, 6384. 24 T. A. Hanna, A. M. Baranger and R. G. Bergman, Angew. Chem., Int. Ed. Engl., 1996, 35, 653. 25 J. March, Advanced Organic Chemistry, Wiley, New York, 4th edn., 1992, pp. 833, 836. 26 M. D. Curtis, J. J. D’Errico and W. M. Butler, Organometallics, 1987, 6, 2151. 27 C. L. Lee, C. T. Hunt and A. L. Balch, Inorg. Chem., 1981, 20, 2498. 28 A. L. Balch and L. S. Benner, Inorg. Synth., 1982, 21, 47. 29 M. Regitz, J. Hocker and A. Liedhegener, Org. Synth., 1968, 48, 36. 30 G. M. Sheldrick, SHELXS 97, Program for Crystal Structure Solution, University of Göttingen, 1997. 31 A. Altomare, G. Cascarano, C. Giacovazzo and A. Guagliardi, J. Appl. Crystallogr., 1993, 26, 343. 32 G. M. Sheldrick, SHELXL-97 Program for Crystal Structure Refinement, University of Göttingen, 1997, Germany. 33 International Tables for X-Ray Crystallography, ed. A. J. C. Wilson, Kluwer, Dordrecht, 1992, vol. C, Tables 6.1.1.4 (pp. 500–502), 4.2.6.8 (pp. 219–222) and 4.2.4.2 (pp. 193–199). Paper 8/05467C

ISSN:1477-9226    

DOI:10.1039/a805467c

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J. Chem. Soc., Dalton Trans., 1997, Pages 305–311 305 DALTON Synthesis, structure and magnetism of iron-(III) and -(II) complexes of 1-thia-4,7-diazacyclononane and N,N9-dimethyl-1-thia-4,7- diazacyclononane † Vincent A. Grillo,a Graeme R. Hanson,b Trevor W. Hambley,c Lawrence R. Gahan,*,a Keith S. Murray *,d and Boujemaa Moubaraki d aDepartment of Chemistry, The University of Queensland, Brisbane, QLD 4072, Australia bCentre for Magnetic Resonance, The University of Queensland, Brisbane, QLD 4072, Australia cSchool of Chemistry, The University of Sydney, Sydney, NSW 2006, Australia dChemistry Department, Monash University, Clayton, Victoria 3168, Australia The mononuclear complexes [Fe([9]aneN2S)Cl3] and [Fe(Me2[9]aneN2S)Cl3] ([9]aneN2S = 1-thia-4,7- diazacyclononane, Me2[9]aneN2S = N,N9-dimethyl-1-thia-4,7-diazacyclononane) have been prepared by addition of the cyclononane to an ethanolic solution of FeCl3.m-Oxo-bis(m-acetato)diiron(III) complexes [Fe2O(O2CMe)2([9]aneN2S)2][PF6]2 and [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2 have been synthesised by addition of sodium acetate to suspensions of the mononuclear complexes and isolated as the hexafluorophosphate salts.The iron(II) dimer [Fe2(OH)(O2CMe)2(Me2[9]aneN2S)2]ClO4 was prepared under anaerobic conditions. The four iron(III) complexes were characterised by crystal structural studies. On the bases of the isomer shift and quadrupole splitting observed in the Mössbauer spectra of the dimers (4.2 K) the iron-(III) and -(II) ions were determined to be in the high-spin configuration.The magnetic susceptibility (300–4.2 K) of [Fe2O(O2CMe)2([9]aneN2S)2][PF6]2 (* = 22JS1?S2) indicated that the iron(III) sites were antiferromagnetically coupled (J = 2125 cm21). In the case of the iron(II) dimer J = 27.4 cm21. The differences observed in the redox behaviour of [Fe2O(O2CMe)2([9]aneN2S)2][PF6]2 and [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2 are attributed to the presence of the sterically demanding ligand methyl substituents. 1,4,7-Triazacyclononane ([9]aneN3) and N,N9,N0-trimethyl- 1,4,7-triazacyclononane (Me3[9]aneN3) and their complexes with iron-(III) and -(II) have been shown to be particularly effective in the development of potential models for proteins containing the bridged iron–oxo structural motif.1 Monomeric complexes such as [Fe([9]aneN3)Cl3] and [Fe(Me3[9]aneN3)Cl3] have been shown to be useful precursors in the development of this chemistry and also to have wider synthetic application.2–15 Derivatives of the complex [Fe(Me3[9]aneN3)Cl3] have been shown to be efficient DNA cleavage agents 16 and Wieghardt and co-workers 17 have shown the utility of the same complex in the generation of new complexes such as m-nitrido-diiron systems.We have now prepared and characterised monomeric and bimetallic iron-(II) and -(III) complexes of 1-thia-4,7- diazacyclononane ([9]aneN2S) and N,N9-dimethyl-1-thia-4,7- diazacyclononane (Me2[9]aneN2S), analogues of the [9]aneN3 and Me3[9]aneN3 systems.We were interested to investigate the chemical influences of the thioether in the bridged diiron systems in an extension of earlier investigations related to the redox, stereochemical and spectroscopic influences on metal complexes by the replacement of nitrogen donors by thioethers.18–24 Experimental Methanol and ethanol were dried over magnesium methoxide and stored under dinitrogen. Anaerobic manipulations were carried out under dry dinitrogen using standard Schlenk techniques with a double-manifold vacuum line, or in a VAC Vacuum/Atmospheres (HE-43-2) controlled-atmospheres laboratory.Electronic absorption and infrared spectra were recorded with Beckman DU7500 and Perkin-Elmer FT1600 † Non-SI unit employed: mB ª 9.27 × 10224 J T21. spectrophotometers, respectively. The infrared spectra were recorded as KBr pellets. 1-Thia-4,7-diazacyclononane was prepared as previously described,25 N,N9-dimethyl-1-thia-4,7- diazacyclononane following a previously published procedure employed for N-methylation of secondary amines.26 Syntheses [Fe([9]aneN2S)Cl3].A methanol solution (40 cm3) of FeCl3? 6H2O (2.8 g, 10.4 mmol) and 1-thia-4,7-diazacyclononane (1.46 g, 10.4 mmol) was refluxed gently for 1 h and then permitted to cool to room temperature. The orange crystalline material which precipitated was filtered off, washed with ethanol and diethyl ether and dried in air (2.9 g, 90%).The complex was recrystallised from warm acetonitrile solution (Found: C, 23.05; H, 4.6; N, 9.0. C6H14Cl3FeN2S requires C, 23.35; H, 4.6; N, 9.1%). [Fe(Me2[9]aneN2S)Cl3]. The compound was prepared using a similar procedure to that employed for [Fe([9]aneN2S)Cl3] using N,N9-dimethyl-1-thia-4,7-diazacyclononane. A bright orange microcrystalline product resulted (90%) (Found: C, 28.15; H, 5.5; N, 8.05. C8H18Cl3FeN2S requires C, 28.55; H, 5.4; N, 8.35%).[Fe2O(O2CMe)2([9]aneN2S)2][PF6]2. An ethanol solution (26 cm3) of [Fe([9]aneN2S)Cl3] (0.5 g, 1.6 mmol) and sodium acetate (0.34 g, 4.25 mmol) was stirred for 2 h at room temperature. The resulting solution was filtered through Celite to remove precipitated sodium chloride. To the intense brown filtrate was added dropwise an ethanol solution (5 cm3) of ammonium hexafluorophosphate (0.52 g, 3.2 mmol) producing a brown precipitate which was filtered off. The filtrate was cooled to 4 8C and upon standing for 48 h brilliant dark brown crystals deposited. The crystals were filtered off, washed with ethanol306 J.Chem. Soc., Dalton Trans., 1997, Pages 305–311 and dried in air (0.40 g, 60%) (Found: C, 23.25; H, 3.85; N, 6.0; S, 7.8. C16H34F12Fe2N4O5P2S2 requires C, 23.2; H, 4.15; N, 6.75; S, 7.75%). Absorption spectrum (MeCN): lmax/nm (e/dm3 mol21 cm21) 251 (6848), 347 (4128), 380 (sh), 430 (sh), 480 (807), 518 (sh), 564 (sh) and 740 (83).IR (KBr, cm21): nasym(CO) 1561, nsym(CO) 1420. [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2. This complex was prepared using a similar procedure to that for [Fe2O(O2- CMe)2([9]aneN2S)2][PF6]2 using [Fe(Me2[9]aneN2S)Cl3] as the starting material. Upon stirring with sodium acetate the solution turned red-brown, precipitating a pale red product upon addition of ammonium hexafluorophosphate. The compound was recrystallised by slow evaporation of an acetonitrile solution, yielding brilliant red needle-like crystals (22%) (Found: C, 27.0; H, 4.65; N, 6.2; S, 7.15.C20H42F12Fe2N4O5P2S2 requires C, 27.15; H, 4.8; N, 6.35; S, 7.25%). Absorption spectrum (MeCN): lmax/nm (e/dm3 mol21 cm21) 252 (6855), 358 (3732), 496 (873), 533 (sh) and 769 (88). IR (KBr, cm21): nasym(CO) 1542, nsym(CO) 1437. [Fe2(OH)(O2CMe)2(Me2[9]aneN2S)2]ClO4. All operations were carried out under anaerobic conditions. To a stirred methanol solution (25 cm3) of N,N9-dimethyl-1-thia-4,7-diazacyclononane (1.04 g, 6 mmol) was added dropwise a methanol solution (10 cm3) of iron(II) perchlorate hexahydrate (0.72 g, 2.0 mmol).The mixture was stirred for 1.5 h whereupon a pale green precipitate formed. Sodium acetate (0.4 g, 4.8 mmol) was then added, resulting in dissolution of the green precipitate. Upon stirring for 2 h a golden solution resulted. A methanol solution (10 cm3) of sodium perchlorate monohydrate (0.8 g, 2.8 mmol) was added. The volume was reduced to approximately 20 cm3 and the solution placed in a freezer at 225 8C, whereupon pale yellow crystals formed within 24 h.The product was collected and subsequently recrystallised from methanol to give pale yellow to clear crystals (Found: C, 34.65; H, 6.5; N, 7.75. C20H43ClFe2N4O9S2 requires C, 34.55; H, 6.25; N, 8.05%). Electrochemistry Cyclic voltammetric studies were undertaken with a BAS100B Electrochemical Analyser using a three-compartment cell. A glassy carbon working electrode, platinum-wire auxiliary and a Ag–Ag+ (0.01 mol dm23 AgNO3 in acetonitrile) reference electrode were employed.All solutions were degassed by purging with nitrogen for at least 15 min prior to the experiment. Acetonitrile were dried and distilled prior to use. The ferrocenium–ferrocene couple was employed as an internal reference. Magnetic studies Magnetic susceptibility studies were made using a Quantum Design MPMS SQUID magnetometer with an applied field of 1 T. The powdered sample was contained in a calibrated gelatine capsule held in the centre of a soda straw fixed to the end of the sample rod.The magnetisation values of the instrument were calibrated against a standard palladium sample, supplied by Quantum Design, and also chemical calibrants such as CuSO4?5H2O and [Ni(en)3][S2O3] (en = ethane-1,2-diamine). Mössbauer spectroscopy Mössbauer spectra were measured in the Physics Department at Monash University with a standard electromechanical transducer operating in a symmetrical constant-acceleration mode.A conventional helium-bath cryostat was employed for temperature control with the sample maintained in exchange gas. Data were collected with an LSI-based 1000-channel multichannel analyser. Velocity calibration was made with respect to iron foil. Spectra were fitted with a Lorentzian lineshape. Crystallography Data collection, structure solution and refinement. Crystal data and refinement details for the complexes [Fe([9]aneN2S)- Cl3], [Fe(Me2[9]aneN2S)Cl3] and [Fe2O(O2CMe)2([9]aneN2S)2]- [PF6]2, [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2 are reported in Table 1.For diffractometry crystals were mounted on glass fibres with cyanoacrylate resin. Lattice parameters at 294 K were determined by least-squares fits to the setting parameters of 25 independent reflections, measured and refined on an Enraf-Nonius CAD4-F four-circle diffractometer with a graphite monochromator (Mo-Ka, l 0.710 69 Å).Intensity data were collected in the range 1 < q < 258. Data were reduced and Lorentz-polarisation and numerical absorption corrections applied using the SDP package.27 The structures were solved using direct methods in SHELXS 86 28 and refined (on F) by full-matrix least-squares analysis with SHELX 76.29 Neutral complex scattering factors were used.30 Hydrogen atoms were included at calculated sites with fixed isotropic thermal parameters. All other atoms were refined anisotropically.Plots were drawn using ORTEP.31 Selected bond lengths and angles are given in Tables 2–5. Atomic coordinates, thermal parameters, and bond lengths and angles have been deposited at the Cambridge Crystallographic Data Centre (CCDC). See Instructions for Authors, J. Chem. Soc., Dalton Trans., 1997, Issue 1. Any request to the CCDC for this material should quote the full literature citation and the reference number 186/319. Results and Discussion Syntheses The mononuclear complexes [Fe([9]aneN2S)Cl3] and [Fe- (Me2[9]aneN2S)Cl3], employed as starting materials for the syntheses of the dimers, were prepared by either addition of the desired cyclononane to an ethanolic solution of hydrated iron(III) chloride, or gentle reflux of a methanolic solution of the chloride and the macrocycle.15 Recrystallisation of [Fe([9]- aneN2S)Cl3] and [Fe(Me2[9]aneN2S)Cl3] by slow evaporation of either acetonitrile or methanol solutions afforded crystals of X-ray crystallographic quality.Addition of sodium acetate to suspensions of [Fe([9]aneN2S)Cl3] or [Fe(Me2[9]aneN2S)Cl3] in ethanol, removal of any precipitates, and addition of ammonium hexafluorophosphate resulted in crystallisation of [Fe2O- (O2CMe)2([9]aneN2S)2][PF6]2 and [Fe2O(O2CMe)2(Me2[9]ane- N2S)2][PF6]2, respectively. The synthetic methodology is similar to that reported for the m-oxo-bis(m-acetato)diiron(III) compounds [Fe2O(O2CMe)2([9]aneN3)2]I2 and [Fe2O(O2CMe)2{HB- (pz)3}2] (pz = pyrazolyl).12,32 The lability of the chlorides, and the stability of the m-oxo-bis(m-acetato) core, ensured the production of both [Fe2O(O2CMe)2([9]aneN2S)2][PF6]2 and [Fe2O- (O2CMe)2(Me2[9]aneN2S)2][PF6]2 in moderate to good yields.The synthesis of the iron(II) complex [Fe2(OH)(O2CMe)2- (Me2[9]aneN2S)2]ClO4 was achieved under anaerobic conditions although crystals suitable for X-ray analysis were not obtained. Crystal structures Crystal data for the iron(III) complexes are given in Table 1.The structures and atom numbering schemes for [Fe([9]aneN2S)Cl3] and [Fe(Me2[9]aneN2S)Cl3] are shown in Figs. 1 and 2, respectively, while selected bond lengths and angles are in Tables 2 and 3. The structures are composed of the cyclononane ligand, an iron(III) atom and three chloride atoms. In both, the cyclononane ligands occupy three co-ordination sites about iron(III), with the three chloride ions completing the distorted octahedron.The addition of the N-methyl substituents makes very little difference to the N]Fe]N and N]Fe]S bond angles which are similar in both structures, as are the Cl]Fe]Cl bond angles. A slight elongation (0.08 Å) is observed in the Fe]N bondJ. Chem. Soc., Dalton Trans., 1997, Pages 305–311 307 distance on going from [Fe([9]aneN2S)Cl3] to the N-methylated analogue, [Fe(Me2[9]aneN2S)Cl3], but both are shorter than those observed for [Fe(Me3[9]aneN3)Cl3] (2.232–2.264 Å).16 The FeIII]S bond lengths exhibited by these complexes appear typical of the few examples reported previously.33–35 The Fe]Cl bond distances are within the range of previously reported Fe]Cl bond distances.36,37 The related complex [Fe{HB- (pz)3}Cl3]2 exhibits Fe]Cl bond distances of 2.319(1), 2.316(1) and 2.305(1) Å, the shorter being associated with the longer trans Fe]N bond.36 The Fe]Cl bond lengths in [Fe([9]ane- N2S)Cl3] and [Fe(Me2[9]aneN2S)Cl3] are essentially equivalent (average 2.303 and 2.291 Å, respectively), although a small elongation of that adjacent to the thioether is observed, possibly attributed to steric effects between the thioether and the chloride.The Fe]Cl bond lengths for [Fe(Me3[9]aneN3)Cl3] range from 2.300 to 2.309 Å.16 The structures and atom numbering schemes for [Fe2O- (O2CMe)2([9]aneN2S)2][PF6]2 and [Fe2O(O2CMe)2(Me2[9]ane- Fig. 1 Structure of [Fe([9]aneN2S)Cl3] with relevant atoms labelled. 30% Probability ellipsoids are shown Fig. 2 Structure of [Fe(Me2[9]aneN2S)Cl3]. Details as in Fig. 1 N2S)2][PF6]2 are shown in Figs. 3 and 4, respectively, while selected bond lengths and angles are given in Tables 4 and 5. It is generally observed that the Fe ? ? ?Fe distances for the m-oxo-m-acetato binuclear complexes with the cyclononane ligands are shorter than for other complexes containing this bridging core.14 The Fe ? ? ?Fe distances for [Fe2O(O2CMe)2- ([9]aneN2S)2][PF6]2 and [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2 [3.057(2) and 3.076(3) Å, respectively] reflect a lengthening upon addition of the N-methyl groups.The effect is not as pronounced as for the [9]aneN3 analogues where the increase in Fe ? ? ?Fe distance between [Fe2O(O2CMe)2([9]aneN3)2][PF6]2 and [Fe2O(O2CMe)2(Me3[9]aneN3)2][PF6]2 was 3.063(5) 2 and 3.12(1) Å.14 The Me2[9]aneN2S ligand appears to offer less steric repulsion than the Me3[9]aneN3 analogue. It was anticipated that a distribution of products might be observed with the mixed sulfur–nitrogen ligands, with the thioether donors being cis and/or trans with respect to the bridging oxo moiety, and in a gauche, anti or syn arrangement with respect to the Fe]O]Fe projection. However, for the [Fe2O(O2CMe)2([9]aneN2S)2]2+ complex the crystal structure indicated that the product isolated displayed the thioethers trans with respect to the bridging oxo unit, with S]Fe]O bond angles of 175.5(2) and 178.2(2)8 and in a syn configuration with respect to the Fe]O]Fe plane.However, for the analogous [Fe2O(O2CMe)2(Me2[9]aneN2S)2]2+ complex the structural analysis indicated that the thioethers were located cis to the bridging oxo group, and in a gauche configuration with respect to the Fe]O]Fe plane. The solid-state structure, of course, does not necessarily reflect that which exists in solution. Fig. 3 Structure of [Fe2O(O2CMe)2([9]aneN2S)2][PF6]2. Details as in Fig. 1 Fig. 4 Structure of [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2?MeCN.Details as in Fig. 1308 J. Chem. Soc., Dalton Trans., 1997, Pages 305–311 Table 1 Crystal data for the complexes [Fe([9]aneN2S)Cl3] [Fe(Me2[9]aneN2S)Cl3] [Fe2O(O2CMe)2([9]aneN2S)2][PF6]2 [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2? MeCN Empirical formula C6H14Cl3FeN2S C8H18Cl3FeN2S C16H34F12Fe2N4O5P2S2 C22H45F12Fe2N5O5P2S2 M 308.46 336.52 828.23 926.06 Crystal system Monoclinic Monoclinic Monoclinic Orthorhombic Space group P21/n P21/n P21/c Pcab a/ Å 7.994(1) 7.326(1) 11.039(2) 14.512(4) b/ Å 13.443(2) 12.611(3) 14.218(4) 17.569(4) c/ Å 11.122(1) 14.628(3) 40.648(10) 29.612(6) b/o 92.50(1) 91.06(1) 90.41(2) U/Å3 1194.1(3) 1351.3(4) 6379(3) 7550(3) Z 4 4 8 8 Dc/g cm23 1.716 1.654 1.725 1.630 m/cm21 19.99 17.70 15.02 10.49 F(000) 628 692 3360 3696 Crystal colour Orange Orange Brown Red-brown Habit Prismatic Prism Needles Needles Dimensions/mm 0.22 × 0.15 × 0.13 0.13 × 0.14 × 0.23 0.11 × 0.30 × 0.13 0.25 × 0.07 × 0.05 Scan mode w–2q w–q w–q w–q q Range/o 1.0–27.5 1–27.5 1–22.5 1.0–22 Reflections measured 3003 3360 8956 5023 hkl Ranges 210 to 10, 0–17, 0–14 29 to 9, 0–16, 0–19 211 to 11, 0–15, 0–43 0–15, 0–18, 0–31 Merging R 0.021 0.016 0.032 Reflections used [I > 2.5s(I)] 2400 2269 4116 1377 Number of variables 175 209 797 483 R(Fo) 0.021 0.031 0.064 0.051 R9 0.025 0.034 0.069 0.051 g, k in w = g/[s2(Fo ) + kFo 2] 1.58, 8.0 × 1025 1.53, 1.55 × 1024 3.04, 5.4 × 1024 1.49, 4.0 × 1024 Shift/e.s.d. 0.007 0.006 0.2 0.06 Residual extrema/ e Å23 0.37, 20.23 0.42, 20.32 0.52, 20.43 0.41, 0.28 For [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2 the average Fe]S bond distance is 2.516(6) Å increasing to an average of 2.573(4) Å for the Fe]S bond trans to the m-oxo in [Fe2O(O2CMe) 2([9]aneN2S)2][PF6]2, a difference ascribed to the trans influence of the oxo group.32 A smaller change is observed for [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2, which contains Fe]N bonds both cis [Fe]N 2.230(14) Å] and trans [2.256(16) Å] to the m-oxo unit.The Fe]S and Fe]N bond lengths in the dimeric com- Table 2 Selected bond lengths (Å) and angles (8) for [Fe([9]ane- N2S)Cl3] Cl(1)]Fe 2.331(0) Cl(2)]Fe 2.279(0) Cl(3)]Fe 2.298(0) S]Fe 2.549(0) N(1)]Fe 2.195(1) N(2)]Fe 2.199(2) Cl(2)]Fe]Cl(1) 96.3(0) Cl(3)]Fe]Cl(1) 99.1(0) Cl(3)]Fe]Cl(2) 100.3(0) S]Fe]Cl(1) 88.7(0) S]Fe]Cl(2) 171.1(0) S]Fe]Cl(3) 86.1(0) N(1)]Fe]Cl(1) 91.2(0) N(1)]Fe]Cl(2) 92.9(0) N(1)]Fe]Cl(3) 162.2(0) N(1)]Fe]S 79.6(0) N(2)]Fe]Cl(1) 165.2(0) N(2)]Fe]Cl(2) 92.6(0) N(2)]Fe]Cl(3) 90.9(0) N(2)]Fe]S 81.1(0) N(2)]Fe]N(1) 76.5(1) Table 3 Selected bond lengths (Å) and angles (8) for [Fe(Me2[9]ane- N2S)Cl3] Cl(1)]Fe 2.274(1) Cl(2)]Fe 2.312(1) Cl(3)]Fe 2.288(1) S]Fe 2.516(1) N(1)]Fe 2.273(2) N(2)]Fe 2.279(2) Cl(2)]Fe]Cl(1) 99.0(1) Cl(3)]Fe]Cl(1) 98.2(1) Cl(3)]Fe]Cl(2) 97.6(1) S]Fe]Cl(1) 173.0(1) S]Fe]Cl(2) 87.2(1) S]Fe]Cl(3) 84.2(1) N(1)]Fe]Cl(1) 92.4(1) N(1)]Fe]Cl(2) 162.5(1) N(1)]Fe]Cl(3) 93.9(1) N(1)]Fe]S 80.8(1) N(2)]Fe]Cl(1) 96.0(1) N(2)]Fe]Cl(2) 88.7(1) N(2)]Fe]Cl(3) 163.3(1) N(2)]Fe]S 80.7(1) N(2)]Fe]N(1) 76.8(1) plexes are similar to those observed in the mononuclear complexes, [Fe([9]aneN2S)Cl3] and [Fe(Me2[9]aneN2S)Cl3], although lengthening of the trans m-oxo Fe]N in [Fe2O(O2CMe)2(Me2[9]- aneN2S)2][PF6]2 and the Fe]S trans m-oxo in the [Fe2O(O2CMe) 2([9]aneN2S)2][PF6]2 is observed.These differences are again ascribed to the trans influence of the m-oxo compared to the mcarboxylato group, and similar effects have been reported previously for the related [Fe2O(O2CMe)2{HB(pz)3}2] complex.32 The Fe]O]Fe angles for both compounds are similar, 118.5(4)8 for [Fe2O(O2CMe)2([9]aneN2S)2]2+ and 118.4(5)8 for [Fe2O(O2CMe)2(Me2[9]aneN2S)2]2+, and fall within the range reported for related complexes.2,3 Table 4 Selected bond lengths (Å) and angles (8) for [Fe2O(O2CMe)2- ([9]aneN2S)2][PF6]2 Fe(2) ? ? ?Fe(1) 3.057(2) S(1)]Fe(1) 2.590(4) N(1)]Fe(1) 2.150(9) N(2)]Fe(1) 2.155(9) O(1)]Fe(1) 1.774(7) O(2)]Fe(1) 2.002(8) O(4)]Fe(1) 2.004(7) S(2)]Fe(2) 2.563(4) N(3)]Fe(2) 2.164(11) N(4)]Fe(2) 2.141(11) O(1)]Fe(2) 1.783(7) O(3)]Fe(2) 2.038(9) O(5)]Fe(2) 2.005(8) C(16)]C(15) 1.507(17) C(13)]O(2) 1.257(14) C(14)]C(13) 1.552(19) C(15)]O(5) 1.267(14) C(13)]O(3) 1.235(16) C(15)]O(4) 1.265(14) N(1)]Fe(1)]S(1) 80.8(3) N(2)]Fe(1)]Fe(2) 115.9(3) O(1)]Fe(1)]N(1) 98.7(3) O(1)]Fe(1)]S(1) 175.5(2) O(2)]Fe(1)]N(1) 87.0(3) O(1)]Fe(1)]N(2) 93.4(3) O(2)]Fe(1)]O(1) 96.6(3) O(2)]Fe(1)]S(1) 87.8(2) O(4)]Fe(1)]S(1) 81.4(2) O(2)]Fe(1)]N(2) 163.9(3) O(4)]Fe(1)]N(2) 88.8(3) O(4)]Fe(1)]O(1) 98.2(3) O(4)]Fe(1)]O(2) 102.1(3) O(4)]Fe(1)]N(1) 159.7(3) N(4)]Fe(2)]N(3) 78.9(4) N(3)]Fe(2)]S(2) 80.5(3) O(1)]Fe(2)]S(2) 178.2(2) N(4)]Fe(2)]S(2) 81.7(3) O(1)]Fe(2)]N(4) 97.5(4) O(1)]Fe(2)]N(3) 97.8(4) O(3)]Fe(2)]S(2) 84.5(2) O(3)]Fe(2)]N(3) 87.9(4) O(3)]Fe(2)]N(4) 162.2(4) O(3)]Fe(2)]O(1) 96.0(3) O(5)]Fe(2)]S(2) 83.2(2) O(5)]Fe(2)]O(3) 100.2(3)J.Chem.Soc., Dalton Trans., 1997, Pages 305–311 309 The Fe]O bond distances for the complex [Fe2O(O2CMe)2- ([9]aneN2S)2][PF6]2, (average 1.78 Å) and [Fe2O(O2CMe)2- (Me2[9]aneN2S)2][PF6]2 (average 1.89 Å) are similar to those reported for [Fe2O(O2CMe)2([9]aneN3)2]I2?0.5MeCN and the methylated analogue [Fe2O(O2CMe)2(Me3[9]aneN3)2][ClO4]2? H2O [1.781(4) and 1.800(3) Å, respectively].13,14 Complexes with both symmetrically and unsymmetrically substituted m-oxo-m-carboxylato cores have been prepared and in each case the dimensions of the core remain essentially the same, indicating that the nature of the terminal ligands does not affect the dimensions of this core.38,39 Mössbauer spectroscopy The Mössbauer spectra of [Fe2O(O2CMe)2([9]aneN2S)2][PF6]2 and [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2 at 4.2 K and zero field consist of a symmetric quadrupole doublet with an isomer shift of 0.48 and 0.49 mm s21, respectively.The observed isomer shifts are in the range 0.35–0.60 mm s21, characteristic of fiveor six-co-ordinate high-spin iron(III) m-oxo compounds.3,14,40 The isomer shift indicates a similar electron density around the iron atoms, and appears insensitive to the terminal ligands.The quadrupole doublet is not distinctive with values of 1.23 and 1.52 mm s21 for [Fe2O(O2CMe)2([9]aneN2S)2][PF6]2 and [Fe2O- (O2CMe)2(Me2[9]aneN2S)2][PF6]2, respectively. The quadrupole splitting for the two compounds agree well with those of other Fe]O]Fe compounds in the high-spin state.14,40 The isomer shift and quadrupole splitting for [Fe2(OH)- (O2CMe)2(Me2[9]aneN2S)2]ClO4 were 1.19 and 2.67 mm s21, similar to those reported for [Fe2(OH)(O2CMe)2(Me3[9]ane- N3)2]ClO4 (1.16 and 2.83 mm s21, respectively).14 The values are consistent with iron(II) in the high-spin configuration and are in agreement with values reported for other binuclear high-spin iron(II) complexes.3,40,41 Magnetic susceptibility The magnetic susceptibility of the binuclear iron compound [Fe2O(O2CMe)2([9]aneN2S)2][PF6]2 was analysed using the spinexchange Hamiltonian, H = 22JS1?S2 (S1 = S2 = 5�2 ).Additional terms for the temperature-independent paramagnetism and the mole percentage of paramagnetic impurity, p, were included in the susceptibility expression.42 Plots of molar susceptibility and effective moment versus temperature are given in Fig. 5. The best least-squares fit obtained gave J = 2125 cm21, p = 0.37%, g = 2.078 and q = 230 K, where q is incorporated in the T 2 q term.Table 5 Selected bond lengths (Å) and angles (8) for [Fe2O(O2CMe)2- (Me2[9]aneN2S)2][PF6]2 Fe(2) ? ? ?Fe(1) 3.076(3) N(1)]Fe(1) 2.230(14) O(1)]Fe(1) 1.799(10) S(2)]Fe(2) 2.520(5) O(3)]Fe(1) 1.946(12) O(4)]Fe(2) 2.034(11) N(2)]Fe(1) 2.256(16) N(3)]Fe(2) 2.265(14) O(1)]Fe(2) 1.782(10) C(1)]O(2) 1.194(25) O(5)]Fe(2) 2.006(11) C(1)]O(4) 1.294(23) N(4)]Fe(2) 2.240(15) C(3)]O(3) 1.239(25) C(3)]O(5) 1.259(23) C(2)]C(1) 1.470(29) C(4)]C(3) 1.548(29) S(1)]Fe(1) 2.511(6) O(2)]Fe(1) 1.984(13) O(2)]Fe(1)]S(1) 85.6(4) O(2)]Fe(1)]O(1) 98.3(5) O(1)]Fe(1)]S(1) 91.7(4) O(3)]Fe(1)]S(1) 168.6(4) O(3)]Fe(1)]O(1) 99.3(5) O(3)]Fe(1)]O(2) 95.6(5) N(1)]Fe(1)]O(1) 93.9(5) N(1)]Fe(1)]S(1) 83.0(4) N(1)]Fe(1)]O(3) 93.3(5) N(1)]Fe(1)]O(2) 163.5(5) N(2)]Fe(1)]S(1) 82.6(4) N(2)]Fe(1)]O(1) 171.8(5) N(2)]Fe(1)]O(2) 87.2(6) N(2)]Fe(1)]O(3) 86.1(6) N(2)]Fe(1)]N(1) 79.6(6) O(1)]Fe(2)]S(2) 93.2(4) O(4)]Fe(2)]O(1) 97.5(5) O(4)]Fe(2)]S(2) 169.3(4) O(5)]Fe(2)]O(1) 98.5(5) Fe(2)]O(1)]Fe(1) 118.4(5) O(5)]Fe(2)]O(4) 94.2(5) Msurements on two samples of [Fe2O(O2CMe)2(Me2[9]- aneN2S)2][PF6]2 unfortunately gave magnetisation values very similar to those of the sample holder and thus the resultant molar susceptibilities were not as well defined as in the [9]aneN2S example. Nevertheless, a J value of ca. 2130 cm21 could be deduced from the high-temperature data but with a g value much lower than expected, viz.ª1.5. The possibility of cocrystallisation of some diamagnetic diluent seems unlikely in view of the very good microanalytical data obtained. Invariably, the magnetic exchange coupling in the binuclear moxo- diiron(III) complexes is antiferromagnetic giving rise to an S = 0 ground state. For binuclear m-oxo-bis(m-acetato)diiron(III) complexes the magnitude of the coupling falls in the range 284 cm21 for [Fe2O(O2CMe)2([9]aneN3)2]I2?0.5NaI?3H2O to 2132 cm21 for [Fe2(bipy)2O(O2CMe)2Cl2] (bipy = 2,29- bipyridyl),15,39,43 with values of 2120 to 2130 cm21 being most common39 for these model compounds and for the diiron(III) forms of iron–oxo proteins.1 The apparent invariance of the magnitude and the size of the electron exchange between the two high-spin iron(III) ions through the m-oxo-bis(m-acetato) core, even upon introduction of a thioether donor to the terminal ligand, illustrates both the stability of the core and its dominance in the mediation of the electron exchange.The large negative q value was required to obtain the fit and reproduce the small decrease in moment below 20 K. A similar constraint was observed in the magnetic studies with [Fe2O(O2CMe)2- (Me3[9]aneN3)2][PF6]2 (J = 2119 cm21, q = 237 K).14 The mechanism of exchange, and the correlation of structural features to the magnitude and the sign of the exchange in binuclear iron(III) complexes, has in recent years received considerable attention.44–52 In general, evidence points to a correlation, especially in multiply bridged systems, between the Fe]O bond distance (or distance with the shortest superexchange pathway), rather than the Fe]O]Fe bond angle, and the exchange coupling constant.A recent study by Wieghardt and co-workers 2 proposed a theory for the observed invariance of the Fe]O]Fe angle with the exchange coupling parameter. It was suggested that the major pathways are the orbital interactions of byz–yz and bxz–z2 = bz2–xz producing antiferromagnetic interactions Jz2–xz AF, Jxz–z2 AF and Jyz–yz AF.Using Hückel calculations for the N5FeIII]FeIIIN5 model it was shown that these interactions were the major terms and the bz2–z2 was approximately four times weaker.2 The FeIII]O]FeIII system remains strongly antiferromagnetic because the Syz–yz (Sij being the orbital overlap integral) is not angular dependent. As the M]O]M angle increases the value of Jxz–z2 AF decreases, but the value of Jxz–xz AF increases and maintains the strong antiferro- Fig. 5 Plot of cm and meff vs. T for [Fe2O(O2CMe)2([9]aneN2S)2][PF6]2. The solid line represents the best least-squares fit to the experimental data using the parameters given in the text310 J. Chem. Soc., Dalton Trans., 1997, Pages 305–311 magnetic coupling. Similarly, the Sz2–z2 overlap is expected to increase, further compensating for the decrease in Sxz–z2.2 Gorun and Lippard47 also investigated the antiferromagnetic coupling of di- and tri-bridged high-spin iron(III) complexes.It was concluded that for binuclear iron(III) complexes with m-oxo (m-hydroxo, alkoxo) and at least one other bridging ligand (carboxylate, sulfate, etc.), a correlation existed between the antiferromagnetic exchange coupling and P, defined as half the shortest superexchange pathway between the two iron(III) ions. The relationship proposed was 2J = A exp(BP), where A = 8.763 × 1011 and B = 212.663 and are numerical constants. 47 No correlation of the coupling constant with the Fe]O]Fe angle was found. Application of this correlation to [Fe2O(O2CMe)2([9]aneN2S)2][PF6]2, using the determined Fe]O bond distance of 1.779 Å, produced a value of J = 2144 cm21, 10% greater than the experimentally determined value (2125 cm21). The exchange coupling for [Fe2(OH)(O2CMe)2(Me2[9]ane- N2S)2]ClO4 was analysed using S1 = S2 = 2 (* = 22JS1?S2) model.42 The best least-squares fit obtained gave J = 27.4 cm21, g = 2.23 and q = 21.5 K.Plots for molar susceptibility and effective moment versus temperature are given in Fig. 6. The maximum in cm at ca. 50 K is indicative of weak antiferromagnetic coupling. Below this maximum the observed data reproducibly show a change in slope at ca. 20 K, and this is not well reproduced by the model. Others have used a similar spin– spin model to that employed here, on related species.53 Strictly, octahedral iron(II) centres have 5T2g ground states and the Heisenberg model is not appropriate. However, under ligand- field distortions of the type anticipated in the present structure, it is reasonable to assume a 5B2 ground state well separated from the 5E, possibly with zero-field splitting (D) of the 5B2 state.The simple model used here assumes J>>D. This seemed a reasonable approach since the complex was EPR silent when measured in transverse detection mode at both X-band at 4 K and Q-band at 100 K in the solid and frozen solution [acetonitrile– toluene (1 : 1)] states.However, use of high frequency (n > 95 GHz) and/or parallel detection mode for even-spin-state signals, of the type pioneered by Hendrich and co-workers,54 would be required to confirm this. The inflection in the cm plot at ca. 20 K probably relates to zero-field splitting or to the presence of some m-oxo-diiron(III) impurity. There was no evidence for the m-oxo species in the Mössbauer spectrum. Proof of zero-field splitting would require variable-field measurements in the temperature range 50–2 K, in the manner used recently for S = 2 dimers 55 and monomers,56 and analysis by a spin Hamiltonian containing J and D terms.54–56 This was not done in the present Fig. 6 Plots of cm and meff vs.T for [Fe2(OH)(O2CMe)2(Me2[9]ane- N2S)2]ClO4. Details as in Fig. 5 study since the primary aim was to deduce the J value which is clearly defined by the position of cmax and by the use of the simple 22JS1?S2 model.The modest antiferromagnetic coupling observed for [Fe2- (OH)(O2CMe)2(Me2[9]aneN2S)2]ClO4 is consistent with the exchange coupling observed in OR-bridged binuclear iron(II) complexes (OR = hydroxide, alkoxide or phenoxide),51,57,58 for example [Fe2(OH)(O2CMe)2(Me3[9]aneN3)2]ClO4 (J = 213 cm21) 13 and with deoxy proteins, e.g. deoxyhaemerythrin (213 cm21).1 Contrastingly, complexes that contain a m-aqua bridge, such as [Fe2(tmen)2(H2O)(O2Ph)4] and [Fe2(tmen)2- (H2O)(O2Ph)4] (tmen = Me2NCH2CH2NMe2), exhibit small ferromagnetic or weak exchange.59 The weak exchange in the protonated binuclear iron(II) complex is in line with the concept that the m-hydroxo moiety is a weak mediator of the p-superexchange pathway.This is clearly illustrated in the attenuation of the exchange coupling between binuclear m-oxo-diiron(III) complexes upon protonation of the bridging oxo group.60–62 The mechanism of the exchange interaction in binuclear iron(II) complexes has not been investigated as extensively as for the binuclear m-oxo-diiron(III) complexes.Recently, Hendrich and co-workers 54 prepared a series of binuclear complexes bridged by a m-phenoxo and bis(m-carboxylato) or bis(mphosphato) core. The m-phosphato-bridged complex [Fe2L- {O2P(OPh)2}2]BF4 (L = 2,6-bis{[bis(2-pyridylmethylamino]- methyl}-4-methylphenol) had a larger Fe]O]Fe angle [122.7(2)8] and displayed antiferromagnetic coupling, compared to the m-propionato analogue [108.93(6)8] which exhibited ferromagnetic coupling.54 The change from ferro- to antiferromagnetic behaviour was attributed to the larger M]O]M angle, the modulation of the coupling in the diiron(II) system being apparently similar to that for dicopper(II) complexes.54,63 Whilst it is clear that the Fe]O]Fe angle is important for the diiron(II) complexes, the limited number of examples of these leaves the exact influence of the FeII]m-O bond distance on the exchange interaction to be established.54 Zero-field splitting effects, of the type alluded to above, further complicate the picture.Electrochemistry The electrochemical properties of [Fe2O(O2CMe)2([9]ane- N2S)2][PF6]2 and [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2 were investigated using cyclic voltammetry. The former complex exhibited a reduction wave at 20.760 V (vs. ferrocene– ferrocenium, 0.050 V s21) with a shoulder to positive potential at a scan rate of 0.100 V s21. As the scan rate was reduced the magnitude of the peak at 20.760 V decreased and the shoulder became the major peak at a scan rate of 0.010 V s21, with a peak potential at 20.610 V.Both processes were irreversible, although a small anodic inflection at 20.520 V was observed at high scan rates (<0.500 V s21). As the scan range is increased to 21.5 V another peak is observed at 21.050 V. The redox chemistry of [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2 is less complicated and displays a quasi-reversible electron-transfer process in acetonitrile at E2� 1 of 20.556 V with DE of 0.140 V and Ic/Ia 1.26 (0.020 V s21).The electrochemical behaviour of [Fe2O(O2CMe)2([9]aneN2- S)2][PF6]2 suggested that there were multiple species present in the electron transfer, with the second species (Ec = 20.610 V) being observed when the scan rate was decreased. The behaviour is indicative of a mechanism where the heterogenous electron transfer forming the reduced species is followed by a chemical reaction, with the reduced species reacting to form an electroinactive species, leaving only a small amount of the reduced species available for oxidation on the reverse sweep.64 The peak observed at 21.050 V is assigned to the second reduction of the binuclear iron(III) dimer, i.e.the FeIII/II FeII/II redox couple. Additionally, a small peak observed at approximately 0.4 V corresponds to the redox couple of [Fe([9]aneN2S)2]2+/3+ arising from reductive decom-J.Chem. Soc., Dalton Trans., 1997, Pages 305–311 311 position of [Fe2O(O2CMe)2([9]aneN2S)2][PF6]2. Continuous scans failed to produce an increase of the peak current indicating this was not the major product after reduction, in contrast to the situation for [Fe2O(O2CMe)2{HB(pz)3}2] which displayed an irreversible process in acetonitrile at 21.182 V and a couple assigned to the mononuclear complex [Fe{HB(pz)3}2]+/2+ (E2� 1 0.192 V ) as a result of decomposition of the dimer.65 The quasi-reversible electron-transfer process displayed by [Fe2O(O2CMe)2(Me2[9]aneN2S)2][PF6]2 at 20.556 V is assigned to the FeIIIFeIII–FeIIFeIII couple, the reversibility being attributed to the fact that the methylated derivative is less likely to form mononuclear complexes upon reduction due to the steric constraints of the methyl substituents.Similar electrochemical behaviour was reported for [Fe2O(O2CMe)2(Me3[9]aneN3)2]2+.14 Very recently,66 the mixed-valence complex [FeIIIFeII(m-OH)- (m-O2CCMe3)2(Me3[9]aneN3)2]2+ has been structurally characterised and found to have properties similar to semimethaemerythrin.It forms the m-oxo-diiron(III) analogue in air, and thus protonation equilibria probably occur simultaneously with electron transfer. Incorporation of the thioether in these tridentate cyclononane macrocycles shifts the redox couple to positive potential, stabilising the lower oxidation state, FeII. Acknowledgements We thank Mr.Yasser Korbatieh and Professor John D. Cashion for the measurement and analysis of the Mössbauer spectrum. References 1 K. K. Andersson and A. Gräslund, Adv. Inorg. Chem., 1995, 43, 359. 2 R. Hotzelmann, K. Wieghardt, U. Flörke, H.-J. Haupt, D. C. Weatherburn, J. Bonvoisin, G. Blondin and J.-J. Girerd, J. Am. Chem. Soc., 1992, 114, 1681. 3 R. Hotzelmann, K. Wieghardt, J. Ensling, H. Romstedt, P. Gütlich, E. Bill, U. Flörke and H.-J. Haupt, J. Am. Chem. Soc., 1992, 114, 9470. 4 B. Mauerer, J. Crane, J. Schuler, K. Wieghardt and B. Nuber, Angew. Chem., Int. Ed. Engl., 1993, 32, 289. 5 P. Chaudhuri, M. Winter, P. Fleischhauer, W. Haase, U. Flörke and H.-J. Haupt, Inorg. Chim. Acta, 1993, 212, 241. 6 S. Drüeke, P. Chaudhuri, K. Pohl, K. Wieghardt, X.-Q. Ding, E. Bill, A. Sawaryn, A. X. Trautwein, H. Winkler and S. J. Gurman, J. Chem. Soc., Chem. Commun., 1989, 59. 7 P. Chaudhuri, M. Winter, K. Wieghardt, S. Gehring, W. Haase, B. Nuber and J. Weiss, Inorg.Chem., 1988, 27, 1564. 8 A. Spool, I. D. Williams and S. J. Lippard, Inorg. Chem., 1985, 24, 2156. 9 K. Wieghardt, I. Tolksdorf and W. Herrmann, Inorg. Chem., 1985, 24, 1230. 10 J. L. Sessler, J. W. Sibert, V. Lynch, J. T. Markert and C. L. Wooten, Inorg. Chem., 1993, 32, 621. 11 J. L. Sessler, J. W. Sibert and V. Lynch, Inorg. Chem., 1990, 29, 4143. 12 P. Poganiuch, S. Liu, G. C. Papaefthymiou and S. J. Lippard, J. Am. Chem. Soc., 1991, 113, 4645. 13 K. Wieghardt, K. Pohl and W.Gebert, Agnew. Chem., Int. Ed. Engl., 1983, 22, 727. 14 J. R. Hartman, R. L. Rardin, P. Chaudhuri, K. Pohl, K. Wieghardt, B. Nuber, J. Weiss, G. C. Papaefthymiou, R. B. Frankel and S. L. Lippard, J. Am. Chem. Soc., 1987, 109, 7387. 15 K. Wieghardt, K. Pohl and D. Ventur, Angew. Chem., Int. Ed. Engl., 1985, 24, 392. 16 G. C. Silver and W. C. Trogler, J. Am. Chem. Soc., 1995, 117, 3983. 17 T. Jüstel, T. Weyhermüller, K. Wieghardt, E. Bill, M. Lengen, A. X. Trautwein and P.Hildebrandt, Angew. Chem., Int. Ed. Engl., 1995, 34, 669. 18 J. I. Bruce, T. M. Donlevy, L. R. Gahan, C. H. L. Kennard and K. A. Byriel, Polyhedron, 1995, 15, 49. 19 T. M. Donlevy, L. R. Gahan, T. W. Hambley, G. R. Hanson, K. L. McMahon and R. Stranger, Inorg. Chem., 1994, 33, 5131. 20 T. M. Donlevy, L. R. Gahan and T. W. Hambley, Inorg. Chem., 1994, 33, 2668. 21 J. I. Bruce, L. R. Gahan, T. W. Hambley and R. Stranger, Inorg. Chem., 1993, 32, 5997. 22 T. M. Donlevy, L. R.Gahan, R. Stranger, S. E. Kennedy, K. A. Byriel and C. H. L. Kennard, Inorg. Chem., 1993, 32, 6023. 23 T. M. Donlevy, L. R. Gahan, T. W. Hambley, K. L. McMahon and R. Stranger, Aust. J. Chem., 1993, 46, 1799. 24 J. I. Bruce, L. R. Gahan, T. W. Hambley and R. Stranger, J. Chem. Soc., Chem. Commun., 1993, 702. 25 L. R. Gahan, G. A. Lawrance and A. M. Sargeson, Aust. J. Chem., 1982, 35, 1119. 26 E. K. Barefield and F. W. Wagner, Inorg. Chem., 1973, 12, 2435. 27 SDP Structure Determination Package, Enraf-Nonius, Delft, 1985. 28 G. M. Sheldrick, SHELXS 86, in Crystallographic Computing 3, eds. G. M. Sheldrick, C. Krüger and R. Goddard, Oxford University Press, Oxford, 1985, pp. 175–189. 29 G. M. Sheldrick, SHELX 76, Program for Crystal Structure Determination, University of Cambridge, 1976. 30 D. T. Cromer and J. T. Waber, International Tables for X-Ray Crystallography, Kynoch Press, Birmingham, 1974, vol. 4. 31 C. K. Johnson, ORTEP, A Thermal Ellipsoid Plotting Program, Oak Ridge National Laboratory, Oak Ridge, TN, 1965. 32 W. H. Armstrong and S. J. Lippard, J. Am. Chem. Soc., 1983, 105, 4837. 33 P. Chakraborty, S. K. Chandra and A. Chakravorty, Inorg. Chem., 1994, 33, 6429. 34 A. J. Blake, A. J. Holder, T. I. Hyde and M. Schröder, J. Chem. Soc., Chem. Commun., 1989, 1433. 35 T. Mashiko, C. A. Reed, K. J. Haller, M. E. Kastner and W. R. Scheidt, J. Am. Chem. Soc., 1981, 103, 5758. 36 S.-H. Cho, D. Whang, K.-N. Han and K. Kim, Inorg. Chem., 1992, 31, 519. 37 J. K. Beattie and C. J. Moore, Inorg. Chem., 1982, 21, 1292. 38 L. Que, jun. and A. E. True, Prog. Inorg. Chem., 1990, 38, 97. 39 D. M. Kurtz, jun., Chem. Rev., 1990, 90, 585. 40 A. L. Feig and S. J. Lippard, Chem. Rev., 1994, 94, 759. 41 C. Cheng and W. M. Reiff, Inorg. Chem., 1977, 16, 2097. 42 C. J. O’Connor, Prog. Inorg. Chem., 1982, 29, 203. 43 J. B. Vincent, J. C. Huffman, G. Christou, Q. Li, M. A. Nanny, D. N. Hendrickson, R. H. Fong and R. H. Fish, J. Am. Chem. Soc., 1988, 110, 6898. 44 B. J. Kennedy, K. S. Murray, P. R. Zwack, H. Homborg and K. Kalz, Inorg. Chem., 1985, 24, 3302. 45 C. Ercolani, M. Gardini, K. S. Murray, G. Pennesi and G. Rossi, Inorg. Chem., 1986, 25, 3972. 46 R. N. Mukherjee, T. P. D. Stack and R. H. Holm, J. Am. Chem. Soc., 1988, 110, 1850. 47 S. M. Gorun and S. J. Lippard, Inorg. Chem., 1991, 30, 1625. 48 S. uml;eke, K. Wieghardt, B. Nuber and J. Weiss, Inorg. Chem., 1989, 28, 1414. 49 K. Tatsumi and R. Hoffman, J. Am Chem. Soc., 1981, 103, 3328. 50 P. J. Hay, J. C. Thibeault and R. Hoffman, J. Am. Chem. Soc., 1975, 97, 4884. 51 B. S. Synder, G. S. Patterson, A. J. Abrahamson and R. H. Holm, J. Am. Chem. Soc., 1989, 111, 5214. 52 P. Gomez-Romero, E. H. Witten, W. M. Reiff and G. B. Jameson, Inorg. Chem., 1990, 29, 5211. 53 Y. Hayashi, T. Kayatani, H. Sugimoto, M. Suzuki, K. Inomata, A. Uehara, Y. Mizutani, T. Kitagawa and Y. Maeda, J. Am. Chem. Soc., 1995, 117, 11 220. 54 H. G. Jang, M. P. Hendrich and L. Que, jun., Inorg. Chem., 1993, 32, 911. 55 Y. Dong, S. Ménage, B. A. Brennan, T. E. Elgren, H. G. Jang, L. L. Pearce and L. Que, jun., J. Am. Chem. Soc., 1993, 115, 1851. 56 B. J. Kennedy and K. S. Murray, Inorg. Chem., 1985, 24, 1552. 57 A. Stassinopoulos, G. Schulte, G. C. Papaefthymiou and J. P. Caradonna, J. Am. Chem. Soc., 1991, 113, 8686. 58 S. Menage, B. A. Brennan, C. Juarez-Garcia, E. Münck and L. Que, jun., J. Am. Chem. Soc., 1990, 112, 6423. 59 K. S. Hagen and R. Lachicotte, J. Am Chem. Soc., 1992, 114, 8741. 60 W. H. Armstrong and S. J. Lippard, J. Am. Chem. Soc., 1984, 106, 4632. 61 M. Rapta, P. Kamaras, G. A. Brewer and G. B. Jameson, J. Am. Chem. Soc., 1995, 117, 12 865. 62 R. Hazell, K. B. Jensen, C. J. McKenzie and H. Toftlund, J. Chem. Soc., Dalton Trans., 1995, 707. 63 W. E. Hatfield, in Magneto-structural Correlations in Exchange Coupled Systems, eds. R. D. Willet, D. Gatteschi and O. Kahn, Reidel, New York, 1983. 64 A. J. Bard and L. R. Faulkner, Electrochemical Methods, Wiley, New York, 1980. 65 W. H. Armstrong, A. Spool, G. C. Papaefthymiou, R. B. Frankel and S. J. Lippard, J. Am. Chem. Soc., 1984, 106, 3653. 66 U. Bossek, H. Hummel, T. Weyhermüller, E. Bill and K. Wieghardt, Angew. Chem., Int. Ed. Engl., 1995, 34, 2642. Received 22nd July 1996; Paper 6/05058A

ISSN:1477-9226    

DOI:10.1039/a605058a

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 311–321 311 Electrospray mass spectral study of isopolyoxomolybdates † Daud K. Walanda, Robert C. Burns, GeoVrey A. Lawrance and Ellak I. von Nagy-Felsobuki * Department of Chemistry, The University of Newcastle, Callaghan, NSW 2308, Australia. E-mail: chvo@cc.newcastle.edu.au. Fax: 61-2-49215472 Received 19th August 1998, Accepted 26th November 1998 Electrospray mass spectrometry (ESMS) has been performed on aqueous solutions of dilute (1023 M) isopolyoxomolybdate systems.There is direct evidence that the evaporation process in the ESMS technique involves significant chemical eVects, resulting in the detection of many new anions and cations. For ammonium polyoxomolybdate systems, negative-ion ESMS yields ions of the form [HMomO3m 1 1]2, [MomO3m 1 1]22, [MomO3m 1 2]42 as well as [Mo7O24]62, whereas for alkali metal polyoxomolybdate systems ions of the form [MomO3m 1 1A]2 and [MomO4mA2m 2 2]22 (where A = Li1, Na1 or K1) were observed.In positive-ion mode two series of polyoxomolybdate cations, namely [MomO4mA2m 1 1]1 and [MomO4mA2m 1 2]21 could be assigned. Aggregates of both the [HMomO3m 1 1]2 and [MomO3m 1 1]22 series in the ammonium polyoxomolybdate system can be classified in terms of open-chained structures of tetrahedra that are corner shared, whereas the highly charged anions [MomO3m 1 2]42 and [MomO3m 1 3]62 are consistent with closed-packed structures. For the alkali metal polyoxomolybdate anion and cation systems the spectra are consistent with open-chained structures of octahedral units that are edge shared, with a terminating tetrahedral unit. Linear correlations suggest that the assembly of these aggregates occurs via an addition polymerization mechanism.This model, consistent with the ESMS data, may identify the additive moieties (MoO3, MoO2 21 and Mo2O8A4) required for aggregation of polyoxomolybdate species in aqueous solution.The chemistry of molybdenum(VI) oxo-complexes has been investigated for several decades because of its importance in biochemical and chemical applications.1–4 To exemplify this diversity, [SiMo12O40]42 has been used as an electron acceptor in studies of photosynthesis and photophosphorylation, whereas the mixed metal polyanions [PVxMo12 2 xO40]n2 coupled with palladium(II) salts have been used to catalyse the aerial oxidation of olefins.2,3 It is well known that in alkaline solution molybdate is present as the monomeric anion [MoO4]22.However, on acidification polymerization occurs which leads to the formation of a range of isopolyoxomolybdate species. The polymerization reaction is thought to proceed mainly via condensation following protonation.2–5 This yields isopolyoxomolybdate ions in aqueous solution of the form [Hp 2 2rMoqO4q 2 r](2q 2 p)2, where p and q are the number of moles of H1 and [MoO4]22 in the reaction scheme respectively and r is the number of moles of water product.Hence each polyoxomolybdate species can be expressed according to the ratio H1 : [MoO4]22. However, there is uncertainty concerning the speciation and formulation of some isopolyoxomolybdate species in solution, since they are diYcult to characterize under such conditions.6–16 Several studies of alkali metal molybdates have confirmed that the monomolybdate [MoO4]22 has a tetrahedral structure in both the solid state and in solution.17,18 On addition of acid, [MoO4]22 is protonated to yield [HMoO4]2 and subsequently H2MoO4.17,19 The [HMoO4]2 and [MoO4]22 ions are the predominant species at high pH (> ª 6) and low concentration.19 However, it has been suggested that [HMoO4]2 is probably octahedral due to an expansion of the co-ordination sphere 20–22 and some investigators 22–24 have proposed that [HMoO4]2 and † Supplementary data available: Negative- and positive-ion ESMS data of alkali metal polyoxomolybates at pH 6.0.For direct electronic access see http://www.rsc.org/suppdata/dt/1999/311/, otherwise available from BLDSC (No. SUP 57467, 3 pp.) or the RSC Library. See Instructions for Authors, 1999, Issue 1 (http://www.rsc.org/dalton). H2MoO4 are more appropriately designated as [MoO(OH)5]2 and Mo(OH)6 respectively. Investigations of the structures of isopolyoxomolybdates by Knöpnadel et al.25 and Armour et al.26 have shown that the crystal structure of (NH4)2Mo2O7 contains an infinite chain anion in which there are present both MoO4 tetrahedra and MoO6 octahedra.In contrast, the discrete [Mo2O7]22 anion consists of two corner-shared MoO4 tetrahedra,27 while Cooper and Salmon28 have shown that the [Mo4O13]22 anion is a flexible chain of four linked MoO4 tetrahedra. The [Mo6O19]22 anion was first identified and analysed structurally by Allcock et al.,29 but in this case the building block of the structure is formed from six distorted MoO6 octahedra that are condensed so that all share a common vertex.30 Lindqvist 16,31 originally reported the structures of both the [Mo7O24]62 and [Mo8O26]42 anions, which consist of seven or eight MoO6 edge-shared octahedra, respectively.The large [Mo36O112(H2O)18]82 anion has also been structurally characterized,32 and consists of two octadecamolybdate (18-molybdate) sub-units related to each other by a center of inversion. The structure involves both six- and seven-co-ordinated molybdenum atoms, while the co-ordinated water molecules are bound directly to the molybdenum atoms in both terminal and bridging locations.The structures of alkali metal isopolyoxomolybdates with general formula [MomO3m 1 1A2] (where A is an alkali metal ion and m is a positive integer) have also been investigated over the past two decades.33 Structural studies have shown that the shapes of these compounds depend on the number of molybdenum atoms, m, and the size of the counter cation.34 Electrospray mass spectrometry (ESMS) has displayed its utility in characterizing polyoxometalates dissolved in organic, aqueous–organic and purely aqueous solutions.35–38 The first ESMS study of dilute aqueous oligomeric anions was that of Howarth and co-workers,38 who investigated aqueous solutions of isopolytungstates, peroxotungstates and heteropolymolybdates.Their study detected new species which had not been previously reported in water, including [W6O19]22 and [W2O7]22.312 J.Chem. Soc., Dalton Trans., 1999, 311–321 For mixed-metal species, heteropolyoxoanion species of the form [HxPWnMo12 2 nO40](3 2 x)2 were also reported. Moreover, Howarth and co-workers 38 were able to show that desolvation by the drying agent may lead to interference in the equilibrium process, since it rapidly changes the pH and concentrations of the solutes in the eventual formation of the analytes. Hence species extremely sensitive to either condition (i.e.presence of other species or pH) may diVer in concentration from bulk measurements (where changes in such conditions are less rapid). A number of investigations have been performed on polyoxomolybdates using organic solvents. For example, Colton and Traeger 35 have shown that negative ion ESMS was amenable to analysing intact heteropolymolybdate species such as the [S2Mo18O62]22 ion in acetonitrile. Similarly, Le Quan Tuoi and Muller36 and Siu and co-workers 37 were successful in identifying the intact ions of heteropolyacids and polyoxoanions.They found that the negative-ion mass spectral data observed were in agreement with their respective calculated isotopic mass distributions. The study of Le Quan Tuoi and Muller 36 on heteropolyacids in diVerent solvents (i.e. methanol–water and acetonitrile–water) found anion peaks associated with the mono-, di- and tri-anions of [H3PMo12O40], [H4PMo11VO40], [H3PW12O40] and [H4SiW12O40]. Siu and co-workers 37 performed experiments using ES tandem mass spectrometry on similar compounds as well as on a number of tetrabutylammonium salts of the isopolyoxometalates (both molybdates and tungstates) in either acetone or methanol–water mixtures.They observed the parent ions [Mo2O7]22, [Mo6O19]22 and [Mo8O26]42 and fragment ions such as [Mo2O7]22, [Mo3O10]22, [Mo4O13]22 and [Mo5O16]22. As an extension of our previous studies on the kinetics and mechanism of formation of heteropolyoxomolybdates 39–41 we report here the ESMS study of ammonium and alkali metal isopolyoxomolybdates in aqueous solution.The aggregates, which have been characterized by ESMS, are important to delineate the chemistry and suggest possible structures of the isopolyoxomolybdates. Experimental Sample preparation Only lithium and ammonium molybdate compounds were synthesized. The other compounds in the series were available commercially [e.g. sodium molybdate, Na2MoO4?2H2O (Univar, 99%), ammonium paramolybdate, (NH4)6Mo7O24?4H2O (Univar, 99%) and potassium molybdate, K2MoO4 (Aldrich, 98%)].The ammonium molybdate was prepared by the action of aqueous NH3 on (NH4)6Mo7O24. Three grams of (NH4)6Mo7- O24?4H2O were dissolved in 20 mL of deionized water. To this solution was added, dropwise, a solution of concentrated NH3. The mixture was stirred vigorously with a motor stirrer to release NH3 gas. On standing for several days at room temperature colorless crystals precipitated.The product was recrystallized from water. IR spectrum: n(N–H) 3100m, d(NH2) 1409s, n(Mo–O) 839s cm21. The lithium molybdate was prepared by the action of LiOH? H2O on MoO3. Equimolar proportions of reagent grade LiOH?H2O (5.04 g) and MoO3 (8.64 g) were mixed in 30 mL of deionized water. The resultant syrupy solution was allowed to stand for some days. The colorless crystalline product, Li2MoO4, was isolated and recrystallized from water. IR spectrum: n(Mo–O) 833s cm21.A Hanna HI 8521 pH meter (Hanna Instruments, Italy) was used for measuring pH values. It was calibrated at room temperature using buVers of pH 4.0 and 7.0 (BDH, Australia). Sample solutions were prepared by dissolving the colorless crystals in deionized water. The ESMS mobile phase was of similar pH to that of the sample solutions. The concentration of all sample solutions was 1023 M. A solution of 0.1% acetic acid was added to adjust the pH of each solution and the mobile phase to 4.5.A solution of pH 6.0 was prepared by boiling the deionized water and bubbling a stream of nitrogen gas through it while it cooled. The solutions were infused directly into the electrospray source by a 10 mL syringe assisted by the flow of the mobile phase (which was fed by a binary pump into the HPLC system). Mass spectrometry All the ESMS experiments were performed on a VG Platform II single quadrupole mass spectrometer (Micromass Ltd., Altrincham, UK) coupled to a HPLC binary pump system.For all spectral acquisitions, the tip of the capillary was at a potential of 13.5 and 23.5 kV relative to ground for positive-ion and negative-ion mode respectively. The source temperature was maintained at 353 K. The mobile phase consisted of water with appropriate concentrations of acetic acid. An optimum flow rate of 10 mL min21 was employed. The cone voltage, CV, was varied between ±20 and ±60 V relative to the skimmer (which was set to ground).Raising the sample CV increases the collisional energy, which in turn enhances the formation of daughter ions by collisions with solvent molecules within the rotary-pumped region between the sample cone and the skimmers (focus cone and skimmer plate), or between the skimmers and RF lens (the hexapole).35 A Rheodyne injector fitted with a 10 mL loop was used to inject the sample solution into the flow of the mobile phase. Mass spectra were acquired by scanning the quadrupole mass filter from m/z 1500 to 100 and ions were detected by means of a scintillator detector.Approximately 25 scans were summed to give a mass spectrum. For the calibration spectra, a 3.0 mg mL21 NaI solution was used with a 50 : 50 mixture of water and acetonitrile as the mobile phase. In positive-ion mode the spectrum exhibited sodium iodide clusters of the form [NaxIy]1 (where y = x 2 1), whereas in negative-ion mode sodium iodide clusters of the form [NaxIy]2 (where y = x 1 1) were used.All data were acquired and processed using the Micromass MassLynx system. Results In the negative-ion investigation of the (NH4)6Mo7O24 and (NH4)2MoO4 solutions, members of three ion series were observed, namely [HMomO3m 1 1]2 (where m = 1 to 6), [MomO3m 1 1]22 (where m = 2 to 20) and [MomO3m 1 2]42 (where m = 9 to 23 and is only odd), as well as [Mo7O24]62. Figs. 1 and 2 are typical of the spectra of ammonium polyoxomolybdate solutions at pH 4.5 and 6.0 respectively for CV 220 and 260 V.The inserts of Figs. 1(a) and 2(a) show the quadruply charged series of isopolyoxomolybdates (i.e. the [MomO3m 1 2]42 species) at CV 220 V for pH 4.5 and 6.0 respectively. The inserts of Figs. 1(b) and 2(b) show the higher members of the [MomO3m 1 1]22 series and the [Mo7O24]62 ion respectively. The [MoO3]2 anion, formally containing MoV, is assigned as a fragment ion since it only appears when the CV is set at 260 V. Notably, the [HMoO4]2 anion is the most abundant peak in nearly all sample solutions.A detailed assignment for all isopolyoxomolybdate species is given in Table 1. The negative-ion spectra of the alkali metal polyoxomolybdate systems are given in Fig. 3. Apart from the most abundant ion (i.e. [HMoO4]2), two series were observed which have the form [MomO3m 1 1A]2 and [MomO4mA2m 2 2]22 (where A = Li1, Na1 or K1). The detailed assignments of the ESMS data for both CV 220 and 260 V at pH 4.5 are given in Table 2. Data measured at CV 220 and 260 V at pH 6.0 are available as SUP 57467.In the positive-ion mode investigations of the ammonium and lithium polyoxomolybdate systems only a background signal was observed at pH 4.5 and 6.0 when the CV was set toJ. Chem. Soc., Dalton Trans., 1999, 311–321 313 Fig. 1 Negative-ion ESMS of ammonium molybdate at pH 4.5: (a) CV set at 220 V; (b) CV set at 260 V. Fig. 2 Negative-ion ESMS of ammonium molybdate at pH 6.0: (a) CV set at 220 V; (b) CV set at 260 V. 120 V. However, the lithium, sodium and potassium polyoxomolybdate systems did show formation of aggregates at various pH and CV values. Figs. 4(a)–4(c) show the lithium, sodium and potassium molybdate mass spectra at pH 4.5 with CV 160 V. The isopolyoxomolybdate cations exhibit two series, [MomO4mA2m 1 1]1 and [MomO4mA2m 1 2]21 (where A = Li1, Na1 or K1 and m varies from 1 to 13). The doubly charged ions were observed only for m odd. Detailed assignments of the ESMS data for pH 4.5 at CV 120 and 160 V are given in Table 3, with the results for pH 6.0 available as SUP 57467.The above assignments are based purely on the m/z values of the peaks detected. Given the resolution of the instrument, it is possible that some minor peaks may overlap with peaks of the species listed above. Indeed, the product of a reductive protonation will have a similar mass to that of the parent species. It should be noted that no evidence of the reduced species314 J.Chem. Soc., Dalton Trans., 1999, 311–321 Table 1 Negative-ion ESMS data of ammonium isopolyoxomolybdates pH 4.5 pH 6.0 Ion observed [MoO3]2 [MomO3m 1 1] [HMoO4]2 [Mo2O7]22 [HMo2O7]2 [Mo3O10]22 [HMo3O10]2 [Mo4O13]22 [HMo4O13]2 [Mo5O16]22 [HMo5O16]2 [Mo6O19]22 [HMo6O19]2 [Mo7O22]22 [Mo8O25]22 [Mo9O28]22 [Mo10O31]22 [Mo11O34]22 [Mo12O37]22 [Mo13O40]22 [Mo14O43]22 [Mo15O46]22 [Mo16O49]22 [Mo17O52]22 [Mo18O55]22 [Mo19O58]22 [Mo20O61]22 [MomO3m 1 2] [Mo9O29]42 [Mo11O35]42 [Mo13O41]42 [Mo15O47]42 [Mo17O53]42 [Mo19O59]42 [Mo21O65]42 [Mo23O71]42 [MomO3m 1 3] [Mo7O24]62 (p,q) a — 1,1 2,2 3,2 4,3 5,3 6,4 7,4 8,5 9,5 10,6 11,6 12,7 14,8 16,9 18,10 20,11 22,12 24,13 26,14 28,15 30,16 32,17 34,18 36,19 38,20 14,9 18,11 22,13 26,15 30,17 34,19 38,21 42,23 8,7 m/z Theoretical 143.9 160.9 151.9 304.8 223.9 448.8 295.9 592.8 367.8 736.7 439.8 880.6 511.8 583.8 655.7 727.7 799.7 871.6 943.6 1015.6 1087.5 1159.5 1231.4 1303.4 1375.4 1447.4 331.8 403.8 475.8 547.7 619.7 691.7 763.7 835.6 175.9 CV = 220 V m/z exptl., %BPI — 162.2, 100 151.8, 10.80 304.8, 6.81 223.4, 17.51 448.7, 2.45 295.8, 31.36 591.0, 9.73 368.7, 2.85 730.6, 1.85 440.7, 4.80 876.9, 1.11 511.9, 3.83 — 657.7, 1.58 — 799.3, 1.46 — 944.3, 0.68 1015.4, 0.54 1085.7, 0.43 1159.3, 0.20 ———— — 404.8, 0.62 476.7, 0.84 548.1, 0.90 620.5, 0.63 692.0, 0.47 760.2, 0.46 835.8, 0.38 — CV = 260 V m/z exptl., %BPI 145.2, 64.50 162.1, 100 152.2, 8.66 306.0, 7.22 233.7, 25.77 449.1, 2.45 296.9, 14.45 592.6, 16.55 367.7, 7.27 736.1, 4.59 438.9, 4.32 877.8, 3.84 510.4, 2.73 — 655.2, 2.57 — 799.9, 3.18 — 943.8, 2.06 1017.7, 1.25 1088.3, 0.86 1158.6, 0.87 1233.7, 0.72 1302.4, 0.71 1376.3, 0.58 1448.2, 0.52 ———————— — CV = 220 V m/z exptl., %BPI — 160.1, 100 152.1, 16.05 306.2, 9.35 224.3, 29.00 450.2, 2.37 296.2, 40.03 594.0, 10.17 366.5, 5.81 735.7, 0.96 440.9, 4.38 877.0, 0.81 511.2, 3.36 — 653.3, 1.01 — 797.4, 1.06 — 944.6, 0.53 1014.1, 0.35 1090.5, 0.34 ————— — 407.4, 0.92 476.9, 1.29 548.4, 0.61 620.6, 0.70 692.5, 0.26 761.1, 0.29 — 179.2, 1.44 CV = 260 V m/z exptl., %BPI 143.1, 46.99 162.2, 91.02 152.7, 32.77 305.2, 20.84 224.0, 100 450.8, 5.20 295.0, 28.18 591.8, 23.78 367.2, 11.65 736.8, 3.03 439.1, 5.34 880.1, 3.36 513.1, 2.86 — 662.2, 6.50 — 803.2, 2.83 — 946.8, 1.26 1018.3, 0.65 1088.3, 0.65 1157.8, 0.49 1232.7, 0.43 1305.0, 0.41 —— 333.7, 0.67 404.0, 0.91 —————— 176.0, 6.37 a See eqn. (1).[MoO3]2 was observed at CV 220 V, so that the data at this CV would not include such contributions.Thus much of the analysis described below will concentrate on the species and the observed percentage base peak intensity (%BPI) at this CV. Nevertheless, a more definitive analysis of the %BPI and/or the inferred structures of species (see below) with similar m/z ratios but diVerent z values awaits a capillary electrophoresis MSMS investigation. Finally, the formulae given in Tables 1–3 are symbolic and are not based on experimentally determined structures.Discussion The formation scheme for the possible polyoxomolybdate species can be represented by the general condensation or protonation reaction (1).2,4,28 The observed species may or may not pH1 1 q[MoO4]22 [Hp 2 2rMoqO4q 2 r](2q 2 p)2 1 rH2O (1) have the associated protons, but this will depend on the pH of the solution and the range of stability of the particular isopolyoxomolybdate. Equilibrium studies have shown that apart from the protonated forms of the molybdate ion, that is [HMoO4]2 and H2MoO4, no species are observed until [Mo7O24]62, while to lower pH values [HMo7O24]52, b-[Mo8O26]42 and a more condensed species, perhaps related to the [Mo36O112(H2O)18]82 anion, are present.19 Recent 95Mo and 17O studies have identi- fied [Mo7O24]62, [HMo7O24]52, b-[Mo8O26]42 and the intermediate species [H3Mo8O28]52 from pH 6 to 1.42 Under the conditions employed in the present studies, however, the major species in the acidified molybdate solutions would have been [Mo7O24]62, as well as the protonated forms of the monomer.It is evident from the ESMS data presented above that considerable rearrangement has occurred, with the major species detected in all solutions being [HMoO4]2, while polymeric species larger than [Mo7O24]62 were present. This is a result of the extremely labile nature of these systems, which can undergo fast addition and/or loss of aggregation units.In general, less labile systems remain intact and only undergo limited dehydration and related reactions during passage through the gas phase, over a range of cone voltages.43 As noted above, in bulk investigations, the [Mo7O24]62 ion is the dominant polymeric species in the solutions employed. Hence it would be anticipated that, as ESMS is a “soft” ionization technique, this species should be dominant at the pHJ. Chem. Soc., Dalton Trans., 1999, 311–321 315 Fig. 3 Negative-ion ESMS of alkali metal molybdate at pH 4.5 and CV set at 220 V: (a) lithium; (b) sodium; (c) potassium system. values examined. It is present at pH 6.0 and at both CV 220 and 260 V, but with a low %BPI. At pH 4.5 it is not detected at either CV value. Thus the large abundance of [HMoO4]2 in all of the ESMS spectra of the ammonium and alkali metal molybdate systems suggests that the [Mo7O24]62 ion dissociates according to eqn. (2). This cannot have proceeded to [Mo7O24]62 1 4H2O 7[HMoO4]2 1 H1 (2) completion as the presence of this species is required as a precursor to the formation of the [MomO3m 1 2]42 series of anions, as shown below.It is known from bulk kinetic studies that concentration terms for [HMoO4]2 appear in the rate laws for both the assembly and dissociation of (hetero)polyoxomolybdate anion frameworks.40,41 Therefore, copious quantities of [HMoO4]2 in the charged droplets provide a source of {MoO3} aggregation units via eqn.(3) for the assembly of polymeric species in [HMoO4]2 1 {H1} {MoO3} 1 H2O (3) aqueous solution, which gives rise to the large range of aggregates observed in all series in the present studies, many with m > 7. Here {H1} represents a protonation site on an isopolyoxomolybdate anion suitable for addition of {MoO3} by the [HMoO4]2 ion with elimination of a water molecule. Thus, the ESMS technique should provide valuable insights into the polymerization reactions of isopoly- and heteropoly-oxometalates and their intermediate fragments. 1 The nature of protonated [MoO4]22 in solution The formation of polyoxomolybdates in the bulk (e.g. [Mo7O24]62) is a rapid reaction and requires an expansion of the co-ordination number of molybdenum(VI) from 4 to 6. It is assumed that monoprotonation of [MoO4]22 also enables this reaction to occur. However, the (%BPI, m/z ratio) of (tetrahedral) [HMoO4]2 and (octahedral) [MoO(OH)5]2 {or an alternative form such as [MoO3(OH)(H2O)2]2} are (100%, 162) and (0%, 197) respectively.On the basis of the mass peaks measured in this investigation, monoprotonated monomolybdate is of the form [HMoO4]2 {or better [MoO3(OH)]2} and not as [MoO(OH)5]2. Nevertheless, it should be noted that the ESMS results could be also consistent with [MoO(OH)5]2 rapidly dissociating into [HMoO4]2 through the process (4) where A is the inert collision medium such as nitrogen. Such a process may readily occur because of the lability of these systems.316 J.Chem. Soc., Dalton Trans., 1999, 311–321 Table 2 Negative-ion ESMS data of alkali metal polyoxomolybdates (pH 4.5) A = Lithium Sodium Potassium Ion observed [MoO3]2 [AMomO3m 1 1]2 [HMoO4]2 [AMoO4]2 [AMo2O7]2 [AMo3O10]2 [AMo4O13]2 [AMo5O16]2 [MomO4mA2m 2 2]22 [Mo3O12A4]22 [Mo4O16A6]22 [Mo5O20A8]22 [Mo6O24A10]22 [Mo7O28A12]22 [Mo8O32A14]22 [Mo9O36A16]22 [Mo10O40A18]22 [Mo11O44A20]22 [Mo12O48A22]22 [Mo13O52A24]22 [Mo14O56A26]22 m/z Theoretical 143.9 160.9 166.9 310.9 454.8 598.7 742.7 253.9 340.8 427.8 514.8 601.8 688.8 775.7 862.7 949.7 1036.6 1123.6 1210.6 CV = 220 V m/z exptl., %BPI — 162.5, 100 168.2, 10.25 — 456.4, 0.83 600.2, 0.91 — 253.6, 1.85 340.0, 2.05 426.6, 0.57 514.6, 1.49 600.2, 0.62 687.3, 0.62 775.9, 0.40 862.0, 0.56 949.5, 0.40 1036.1, 0.31 1122.3, 0.27 1208.9, 0.20 CV = 260 V m/z exptl., %BPI 145.2, 53.06 162.4, 100 166.1, 21.04 309.1, 0.91 454.8, 0.59 601.0, 1.52 746.9, 0.56 — 340.9, 7.58 427.7, 0.26 512.2, 1.95 601.0, 1.52 688.0, 2.49 774.8, 0.84 863.8, 1.55 949.3, 0.91 1035.6, 1.12 1120.9, 0.65 1210.4, 0.67 m/z Theoretical 143.9 160.9 182.9 326.9 470.8 614.7 758.7 285.9 388.9 491.9 594.8 697.8 800.8 903.8 1006.7 1109.7 1212.6 1315.6 1418.6 CV = 220 V m/z exptl., %BPI — 162.5, 100 184.2, 14.42 ———— 285.2, 4.21 387.9, 4.13 491.5, 1.19 595.0, 1.91 697.9, 0.87 798.7, 0.90 902.3, 0.34 1005.2, 0.57 1109.8, 0.31 1211.0, 0.31 1313.9, 0.21 1418.0, 0.20 CV = 260 V m/z exptl., %BPI 145.3, 41.22 161.1, 100 184.2, 13.41 ———— 285.0, 0.18 388.8, 7.36 492.4,0.27 594.9, 2.20 697.2, 0.76 801.2, 2.07 903.4, 0.73 1008.1, 1.36 1109.3, 0.72 1211.1, 0.85 1314.2, 0.45 1416.4, 0.51 m/z Theoretical 143.9 160.9 198.9 342.9 486.8 630.8 774.7 317.9 436.9 555.8 674.8 793.8 912.8 1031.7 1150.7 1269.7 1388.6 1508.8 1627.8 CV = 220 V m/z exptl., %BPI — 162.5, 100 198.1, 19.19 ———— 318.3, 2.65 436.7, 3.83 556.3, 1.06 674.0, 1.82 793.6, 0.84 912.3, 0.89 1030.5, 0.45 1151.2, 0.49 — 1385.5, 0.25 —— CV = 260 V m/z exptl., %BPI 145.5, 45.82 160.2, 100 200.8, 25.41 ———— — 436.5, 6.68 — 672.9, 1.45 794.5, 0.26 911.8, 1.58 1031.9, 0.68 1149.3, 1.31 1269.7, 0.60 1388.7, 0.94 ——J.Chem. Soc., Dalton Trans., 1999, 311–321 317 Fig. 4 Positive-ion ESMS of alkali metal molybdate at pH 4.5 and CV set at 160 V: (a) lithium; (b) sodium; (c) potassium system. 2 The ammonium molybdate systems The [HMomO3m 1 1]2 and [MomO3m 1 1]22 series.The two series, [HMomO3m 1 1]2 and [MomO3m 1 1]22, were found in studies of the ammonium molybdate system. The former was observed from m = 1 to 6 and the second from m = 2 to 20. The recent investigation of Siu and co-workers 37 of the three tetrabutylammonium salts of [Mo2O7]22, [Mo6O19]22 and [Mo8O26]42 in aqueous–organic solvents identified the parent ion as the base peak for each solution. Other anions, such as [Mo3O10]22, [Mo4O13]22 and [Mo5O16]22, all members of the [MomO3m 1 1]22 series, were regarded as fragment ions due to the multiple removal of formal {MoO3} units from precursor ions.The fragment ions observed had relative %BPI values ranging from 0.1 to 19%. Additionally, the quadruply charged species, [Mo6O20]42 (which has a similar m/z to [Mo3O10]22), was likewise regarded as a fragment ion of the precursor species [Mo8O26]42. Both types of species (doubly and quadruply charged ions) that were detected have also been observed in this investigation.However, the presence of protonated O Mo OH O –O OH2 OH2 A O Mo –O OH O + 2H2O (4) forms of several of these polyoxomolybdate ions such as the [HMomO3m 1 1]2 series, as well as [HMoO4]2 (m = 1), were not reported. The latter is the most abundant peak in nearly all of the spectra measured in this investigation, while no species with an aggregation greater than an octamolybdate were reported in the study by Lau et al.37 It should be noted that the evaporation process in ESMS can have significant chemical eVects.The removal of ammonia (as is evident from our ESMS data of the ammonium molybdate systems) results in a rapidly changing pH of the charged droplets. Thus protonated forms ([HMomO3m 1 1]2) of the early members of the [MomO3m 1 1]22 series are observed, but terminate at m = 6. As the Lewis basicity of the anions rapidly diminishes with increasing polymerization the latter members of this series (m > 6) remain unprotonated.For the [MomO3m 1 1]22 series, the peak at m/z 296 in Table 1 can be assigned as either the [Mo4O13]22 or [Mo8O26]42 anion. It is possible to distinguish between these species by comparing their calculated line shapes with the experimental data. Fig. 5(a) shows the experimental spectrum of the peak at m/z 296, whereas Fig. 5(b) and 5(c) show the theoretical spectra line shapes of the [Mo4O13]22 and [Mo8O26]42 anions respectively. The ratios of the full width at half-maximum height (FWHM) of the theoretical spectral peaks versus experiment for the [Mo4O13]22 and [Mo8O26]42 ions are 0.98 and318 J.Chem. Soc., Dalton Trans., 1999, 311–321 Table 3 Positive-ion ESMS data of alkali metal polyoxomolybdates (pH 4.5) A = Lithium Sodium Potassium Ion observed [MomO4mA2m 1 1]1 [MoO4A3]1 [Mo2O8A5]1 [Mo3O12A7]1 [Mo4O16A9]1 [Mo5O20A11]1 [Mo6O24A13]1 [Mo7O28A15]1 [MomO4mA2m 1 2]21 [Mo3O12A8]21 [Mo5O20A12]21 [Mo7O28A16]21 [Mo9O36A20]21 [Mo11O44A24]21 [Mo13O52A28]21 m/z Theoretical 180.9 354.9 528.8 702.8 876.7 1050.6 1224.6 267.7 441.5 647.8 789.7 963.7 1136.8 CV = 120 V m/z exptl., %BPI ——————— —————— CV = 160 V m/z exptl., % BPI 179.9, 68.48 354.8, 61.41 528.3, 48.70 703.1, 100 878.4, 80.09 1046.5, 70.28 1222.3, 53.16 —— 648.2, 21.18 789.1, 51.00 964.3, 47.60 — m/z Theoretical 228.9 434.9 640.8 846.8 1052.7 1258.6 1464.6 331.9 537.8 743.8 949.7 1155.7 1361.6 CV = 120 V m/z exptl., %BPI 229.8, 0.99 434.5, 0.63 641.0, 0.65 845.9, 1.04 1052.3, 0.86 1258.6, 0.69 1461.2, 0.56 —— 742.9, 0.76 949.7, 0.68 —— CV = 160 V m/z exptl., %BPI 229.9, 59.68 434.9, 64.98 641.1, 64.51 846.5, 100 1051.4, 90.81 1257.0, 70.54 1463.9, 54.41 —— 744.1, 30.84 948.1, 45.61 1154.5, 36.55 1356.8, 30.95 m/z Theoretical 276.9 514.9 752.8 990.8 1228.7 1466.6 1704.6 395.9 633.8 871.8 1109.7 1347.7 1585.6 CV = 120 V m/z exptl., %BPI 278.0, 37.09 514.9, 46.44 754.1, 18.72 989.4, 16.75 1227.9, 7.39 1456.6, 6.88 — 396.0, 16.91 634.6, 19.52 871.2, 10.98 1109.7, 10.29 —— CV = 160V m/z exptl., %BPI 278.0, 35.84 514.8, 21.48 752.2, 7.49 990.9, 10.80 1227.8, 12.59 1464.4, 13.13 1702.2, 11.26 —— 872.1, 1.99 1110.0, 6.07 1348.2, 4.17 1582.0, 5.74J.Chem. Soc., Dalton Trans., 1999, 311–321 319 0.70 : 1 respectively. Hence, the ESMS datum is more consistent with the assignment of this peak to the [Mo4O13]22 species. This is also the case for all assigned doubly charged ions. Table 1 shows that the %BPI of [Mo4O13]22 at pH 4.5 and 6.0 with CV 220 V are 31 and 40% respectively. At CV 260 V the %BPI are reduced to 14 and 28% respectively.This indicates that fragmentation of this species has occurred at the higher CV. This is consistent with the concomitant relative increases in the %BPI values of [Mo3O10]22 and [Mo5O16]22, which are produced via the dissociation reaction (5). Thus the following 2[MomO3m 1 1]22 æÆ [Mom 2 1O3m2 2]22 1 [Mom 1 1O3m 1 4]22 (5) analysis will center on the spectra measured at CV 220 V, which are not subject to such fragmentation reactions.Addition polymerization. The open-chain polymerization or aggregation of the isopolyoxomolybdate systems may occur either by edge or corner sharing of the oxygen atoms. There are other possible combinations, but these are less likely. It should be noted that the polymerization or aggregation may not be associated with the ions in bulk solution due to the ESMS process, which dries charged droplets in the production of the charged molecular aggregates.This is especially important for these labile systems, where water molecules that may have initially occupied axial positions in molybdate octahedra are removed, leaving a tetrahedral residue [eqn. (4)]. To assist in the interpretation of the mass spectra of the [MomO3m 1 1]22 series, investigation of the %BPI values of the members is informative. In an addition polymerization mechanism, once the “seed” has been constructed, polymerization is driven by the addition of a monomer or moiety for the “growth” of the chain.The degree of polymerization, q, is given Fig. 5 Comparison of negative-ion ESMS line shapes: (a) experimental line shape of [Mo4O13]22; (b) calculated line shape of [Mo4O13]22 (with FWHM = 0.8 Dalton and isotope separation = 1.0 Dalton); (c) calculated line shape of [Mo8O26]42 (with FWHM = 0.8 Dalton and isotope separation = 1.0 Dalton). by N0/Nt (where N0 is the number of molybdate monomers available for addition at time zero and Nt is the number available at time t), and must be proportional to the quantity of polymers present, that is to the %BPI.Hence ln q ª p µ 2m, where m is the number of molybdate units in the polymer chain and p is the extent of the reaction, (N0 2 Nt)/N0. Therefore a plot of the logarithm of the %BPI versus the number of molybdate units in the aggregates should reveal a straight line with a negative slope.Such a correlation is observed with the correlation coeYcients for m (odd, even, total) being (0.8933, 0.9081, 0.8945) respectively. This suggests that the assembly of these aggregates in this series is via an addition polymerization mechanism. Inferred structures. To characterize the ESMS aggregates detected in this investigation the following classification will be invoked. In the case of the ammonium molybdate system members of the [MomO3m 1 1]22 series (and related protonated forms) were observed.Two diVerent formulae can be invoked to depict the inferred structures and molybdenum valencies of the species. Aggregates with even or odd molybdate units may have a structural formula denoted as [Tc]i, where the square brackets enclose the “i” repeating tetrahedral unit “T”, and the superscript “c” indicates that corner oxygens are shared. It should be noted that “i” must equal m. For example, the largest aggregate observed at CV 220 V and pH 4.5 is [Mo16O49]22, which is denoted as [Tc]16.In the corner-shared tetrahedral structure the {MoO3} moiety is the additive unit for the polymerization process, generated from the presence of [HMoO4]2 [see eqn. (3)]. The polymerization process may involve corner-shared octahedral structures which are desolvated via eqn. (4) during the ESMS process, leaving tetrahedral residues. Nevertheless, the analytes detected do not exhibit the presence of coordinated water ligands. Fig. 6(a) gives a schematic of possible Fig. 6 Possible structures for: (a) [MomO3m 1 1]22, [Tc]i for i = 2, 4 or 16; (b) [MomO3m 1 1]22, [(TeOe)i]Te for i = 1, 3 or 7; (c) [MomO4m- A2m 2 2]22, [(Oe)i]Te for i = 2–4.320 J. Chem. Soc., Dalton Trans., 1999, 311–321 structures for three members of the series. Also, it should be noted that closed-packed structures are generally associated with larger-sized highly negatively charged aggregates 16,31,44 and are not applicable nor consistent with the O :Mo ratios observed in the ESMS spectra for this series.There is experimental evidence, according to crystallographic studies of isopolyoxomolybdates,16,31,44 of a repeating unit of the form [(OcTc)i], where MoO6 octahedra (denoted as “O”) share corners with MoO4 tetrahedra and with the adjacent MoO6 octahedra in the chain as exemplified by ([Mo2O7]22)• in [NH4]2[Mo2O7]. The gas-phase ESMS data do not support these alternating structures. For example, for a 5-molybdate fragment derived from the infinite chain ([Mo2O7]22)• in [NH4]2- [Mo2O7] there are several fragments that may occur in the gas phase.The least-charged fragments have the mass formulae [Mo5O19]82 and [Mo5O20]102, but neither is observed in the negative-ion ESMS spectra. The ESMS data for m odd are also consistent with the structural formula [(TeOe)i]Te, with i now being equal to (m 2 1)/2, and the superscript “e” implying that the oxygens are now edge shared, with the four oxygens in the non-terminating tetrahedral units exhibiting partial double bonding to molybdenum.The “T” outside of the square bracket is a terminating tetrahedral residue. The largest polymer observed at CV 220 V and pH 4.5 is for m = 15, which yields a structural formula of [(TeOe)7]Te corresponding to [Mo15O46]22. Fig. 6(b) gives a schematic of these structures with m odd for three members of the series. To assist in delineating which structures may be appropriate for the [MomO3m 1 1]22 series, an analysis of the correlation coeYcients from the addition polymerization analysis is informative.As each component correlation coeYcient is similar to the overall value there appears to be no reason to assume that a diVerent assembly mechanism should occur for m even or odd. In the absence of more definitive experimental evidence, the corner-shared tetrahedral structure representation, [Tc]i, with the {MoO3} moiety being the additive unit, appears to be a more consistent interpretation of the ESMS data. The [MomO3m 1 2]42 series.The species of the series [MomO3m 1 2]42 are highly charged and so would be expected to have closed-packed structures. In fact, the structure of [Mo8O26]42 has been established crystallographically, and consists of eight MoO6 octahedra sharing edges (i.e. [Oe]8).16,31,44 The [Mo7O24]62 ion appears to be the likely seed for the [MomO3m 1 2]42 series, since the series begins with m > 7, and the addition of a formal {MoO2 21} unit together with {MoO3} would yield the first member of the series.The {MoO2 21} unit can be generated from [HMoO4]2 via eqn. (6). The [Mo9O29]42 [HMoO4]2 1 3{H1} {MoO2 21} 1 2H2O (6) structure may be simply generated by the addition of {MoO2 21} and {MoO3} aggregation units in adjacent sites on an [Mo7O24]62 ion, such that the {MoO2 21} unit shares four oxygen sites and the {MoO3} unit three sites on the seed. Further additions of {MoO3} units occur on edge-sharing positions of available MoO6 octahedra, thereby generating this series.Alternatively, the source of the additive {MoO2 21} and {MoO3} units may be the [HMo2O7]2 ion, which is present with a reasonably large %BPI (see Table 1) for CV 220 V at pH 4.5 and 6 via eqn. (7). [HMo2O7]2 1 3{H1} {MoO2 21} 1 {MoO3} 1 2H2O (7) The structure of the [Mo7O24]62 ion is well known, and consists of seven MoO6 octahedra sharing edges (i.e. [Oe]7).16,31,44 At pH 6.0 there is an approximate fourfold increase in the %BPI of this species between CV 220 and 260 V, which suggests that the increase is due to fragmentation via the dissociation reaction (8) where m = 9.However, the low %BPI 2[MomO3m 1 2]42 æÆ [Mom 2 2O3m 2 3]62 1 [Mom 1 2O3m 1 7]22 (8) for the [Mo7O24]62 ion has been attributed to the generation of [HMoO4]2 [see eqn. (2)]. Nevertheless, the absence of other members of a [MomO3m 1 3]62 series suggests that even if small quantities of the [Mo7O24]62 ion are present it may act as a seed to generate larger sized polymers. The rapid dissociation reaction (9) where m is even ensures a low %BPI of these even-membered species when compared with the m odd ions.Nevertheless, within the resolution of the instrument, the m/z overlap of the [MomO3m 1 2]42 and [MomO3m 1 1]22 (for m even) precludes a definitive analysis as to the extent of dissociation according to eqn. (9). 2[MomO3m 1 2]42 æÆ [Mom 2 1O3m 2 1]42 1 [Mom 1 1O3m 1 5]42 (9) 3 The alkali metal molybdate systems The [MomO3m 1 1A]2 and [MomO4mA2m 2 2]22 series.The alkali metal molybdate solutions yield two sequences in negative-ion mode due to the incorporation of the alkali metal cations. The [MomO3m 1 1A]2 series is similar in structure to that observed in the ammonium molybdate system except for the addition of an alkali metal cation. The series is severely truncated in the sodium and potassium systems and, because of the lack of data, no attempt was made to apply the addition polymerization model.There is no evidence for a series related to the more highly charged series found in the ammonium molybdate system. The second series observed has the general formula [MomO4mA2m 2 2]22. Plots of the logarithm of the %BPI for the [MomO4mA2m 2 2]22 series versus the number of molybdate units (m) in the aggregates yield correlation coeYcients for the lithium, sodium and potassium systems for m (odd, even, total) of (0.8049, 0.9744, 0.8284), (0.9221, 0.9890, 0.9083) and (0.9512, 0.9985, 0.8783) respectively.There are significant diVerences between the component correlation coeYcients and the overall correlation, which suggests that the addition polymerization begins from diVerent seeds. For example, starting with an odd-seed the addition polymerization sequence is (10) whereas [Mo2q 1 1O8q 1 4A4q]22 1 [Mo2O8A4] [Mo2q 1 3O8q 1 12A4q 1 4]22 for q = 1,2, ? ? ? 5 (10) starting with an even-seed the addition polymerization sequence is (11).An additive [Mo2O8A4] moiety for both the even and [Mo2qO8qA4q 2 2]22 1 [Mo2O8A4] æÆ [Mo2q 1 2O8q 1 8A4q 1 2]22 for q = 2, ? ? ? 6 (11) odd seeds is therefore consistent with the component correlation coeYcients for odd and even sequences being better correlated than for the total correlation. The open-chained structures for both odd and even m can be designated as [(Oe)i]Te, where i = m 2 1. For example, in all cases, aggregation first occurs for i = 2 which corresponds to the [Mo3O12A4]22 ion for each alkali metal system.The tetrahedral unit may have had initially two axial water molecules which were desolvated during the drying process via eqn. (4). Fig. 6(c) depicts three members of the sequence. On electrostatic grounds the negatively charged oxygen ligands would distort the octahedral shape (which is not shown in this schematic). Moreover, the structures have been drawn exhibiting cis-dioxo MoO2 units which are consistent with this structural feature asJ. Chem.Soc., Dalton Trans., 1999, 311–321 321 found in the known structures of isopolymolybdates.2 The ESMS data are also supportive of structures in which the tetrahedral unit occupies positions other than the terminal position, although this would not lead to a facile polymer condensation mechanism. The additive unit described above, [Mo2O8A4], must be dimeric and not two monomeric A2[MoO4] units since the addition of the latter is inconsistent with the correlations observed in this work.The additive unit structure is therefore designated as (OeT) which supports the polymerization process suggested by Kepert.45 The addition polymerization process given by eqns. (10) and (11) suggests that the tetrahedral residue is in the terminal position since the terminal oxygens are perpendicular to the plane of the adjoining octahedra and the insertion of the octahedral residue of the additive [Mo2O8A4] moiety would sweep these oxygen atoms into axial positions as required by the [(Oe)i]Te structural respresentation. The [MomO4mA2m 1 1]1 and [MomO4mA2m 1 2]21 series.The positive-ion ESMS data for the singly charged cations are consistent with the structural formula of the anions, i.e. [(Oe)i]Te, where i = m 2 1, with the open-chain interleaved by the alkali metal cations, stabilizing the negatively charged oxygens. For the sodium and potassium systems the doubly charged cations have the same structural formula as the anions except that there is an additional alkali metal cation in the structure.A schematic of a singly charged [Mo3O12K7]1 and doubly charged [Mo3O12K8]21 may be represented as shown below. For simplifi- cation the alkali metal cations have been omitted. They most likely occupy edge-bridging or face-capping positions on the individual MoO6 octahedra. Acknowledgements The acquisition of the electrospray mass spectrometer was made possible because of the support of the Australian Research Council, Research Infrastructure Equipment Facility grant.D. K. W. acknowledges support from AusAID for a University of Newcastle postgraduate scholarship and from Tadulako University for granting him leave to study abroad. We wish to thank Mr Brian Mason of the Advanced Mass Spectrometry Unit for his helpful advice in obtaining the ESMS spectra. References 1 C. L. Hill, Chem. Rev., 1998, 98, 1. 2 M. T. Pope, Inorganic Chemistry Concepts 8, Heteropoly and Isopoly Oxometalates, Springer, Berlin, 1983. 3 K. F. Jahr and J. Fuchs, Angew. Chem., Int. Ed. Engl., 1966, 5, 689. 4 K. H. Tytko and O. Glemser, Adv. Inorg. Chem. Radiochem., 1976, 19, 239. O Mo O O O O O Mo O O O O Mo O O 6– 5 H. T. Evans, Jr., Perspect. Struct. Chem., 1971, 4, 2. 6 E. Darmois and J. Périn, C. R. Acad. Sci., 1923, 177, 762. 7 Y. Doucet and S. Bugnon, J. Chim. Phys., 1957, 54, 155. 8 E. Richardson, J.Inorg. Nucl. Chem., 1959, 9, 267. 9 C. Heitner-Wirguin and R. Cohen, J. Inorg. Nucl. Chem., 1964, 26, 161. 10 D. S. Honig and K. Kustin, Inorg. Chem., 1972, 11, 65. 11 N. Kiba and T. Takeuchi, J. Inorg. Nucl. Chem., 1974, 36, 847. 12 K. Y. S. Ng and E. Gulary, Polyhedron, 1984, 3, 1001. 13 K. H. Tytko, G. Baethe and J. J. Cruywagen, Inorg. Chem., 1985, 24, 3132. 14 T. Ozeki, H. Kihara and S. Ikeda, Anal. Chem., 1988, 60, 2055. 15 S. Himeno, H. Niiya and T. Ueda, Bull. Chem. Soc. Jpn., 1997, 70, 631. 16 I. Lindqvist, Ark. Kemi, 1950, 2, 349. 17 G. Schwarzenbach and J. Meier, J. Inorg. Nucl. Chem., 1958, 8, 302. 18 R. H. Busey and O. L. Keller, J. Chem. Phys., 1964, 41, 215. 19 Y. Sasaki and L. G. Sillén, Acta Chem. Scand., 1964, 18, 1014. 20 E. F. C. H. Rohwer and J. J. Cruywagen, J. S. Afr. Chem. Inst., 1963, 16, 26. 21 J. Aveston, E. W. Anacker and J. S. Johnson, Inorg. Chem., 1964, 3, 735. 22 E. F. C. H. Rohwer and J. J. Cruywagen, J. S. Afr. Chem. Inst., 1964, 17, 145. 23 D. V. S. Jain, Indian J. Chem., 1970, 8, 945. 24 G. Wiese and D. Böse, Z. Naturforsch., Teil B, 1972, 27, 897. 25 I. Knöpnadel, H. Hartl, W.-D. Hunnius and J. Fuchs, Angew. Chem., Int. Ed. Engl., 1974, 13, 823. 26 A. W. Armour, M. G. B. Drew and P. C. H. Mitchell, J. Chem. Soc. Dalton Trans., 1975, 1493. 27 V. W. Day, M. F. Fredrich, W. G. Klemperer and W. Shum, J. Am. Chem. Soc., 1977, 99, 6146. 28 M. K. Cooper and J. E. Salmon, J. Chem. Soc., 1962, 2009. 29 H. R. Allcock, F. C. Bissell and E. T. Shawl, J. Am. Chem. Soc., 1972, 94, 8603. 30 W. Clegg, G. M. Sheldrick, C. D. Garner and I. B. Walton, Acta Crystallogr., Sect. B, 1982, 38, 2906. 31 I. Lindqvist, Ark. Kemi, 1950, 2, 325. 32 B. Krebs, S. Stiller, K. H. Tytko and J. Mehmke, Eur. J. Solid State Inorg. Chem., 1991, 28, 883. 33 J. Marrot and J.-M. Savariault, Acta Crystallogr., Sect. C, 1995, 51, 2201. 34 B. M. Gatehouse and P. Leverett, J. Chem. Soc. A, 1971, 2107. 35 R. Colton and J. C. Traeger, Inorg. Chim. Acta, 1992, 201, 153. 36 J. Le Quan Tuoi and E. Muller, Rapid Commun. Mass Spectrom., 1994, 8, 692. 37 T.-C. Lau, J. Wang, R. Guevremont and K. W. M. Siu, J. Chem. Soc., Chem. Commun., 1995, 877. 38 M. J. Deery, O. W. Howarth and K. R. Jennings, J. Chem. Soc., Dalton Trans., 1997, 4783. 39 S. J. Dunne, R. C. Burns and G. A. Lawrance, Aust. J. Chem., 1992, 45, 1943. 40 S. J. Dunne, J. A. Irwin, R. C. Burns, G. A. Lawrance and D. C. Craig, J. Chem. Soc., Dalton Trans., 1993, 2717. 41 A. L. Nolan, R. C. Burns, G. A. Lawrance and D. C. Craig, J. Chem. Soc., Dalton Trans., 1996, 2629. 42 O. W. Howarth, P. Kelly and L. Pettersson, J. Chem. Soc., Dalton Trans., 1990, 81. 43 M. J. Deery, T. Fernandez, O. W. Howarth and K. R. Jennings, J. Chem. Soc., Dalton Trans., 1998, 2177. 44 G. A. Tsigdinos and C. J. Hallada, Isopoly Compounds of Molybdenum, Tungsten and Vanadium, Bulletin Cdb-14, Climax Molybdenum Company, New York, 1969. 45 D. L. Kepert, Prog. Inorg. Chem., 1962, 4, 199. Paper 8/06541A

ISSN:1477-9226    

DOI:10.1039/a806541a

出版商:RSC    

年代:1999

数据来源: RSC