摘要:

DALTON PERSPECTIVE J. Chem. Soc., Dalton Trans., 1999, 1185–1200 1185 The rational design of high symmetry coordination clusters† Dana L. Caulder and Kenneth N. Raymond* Department of Chemistry, University of California, Berkeley, CA 94720-1460, USA. E-mail: raymond@socrates.berkeley.edu Received 28th October 1998, Accepted 10th December 1998 There are numerous examples of supramolecular assemblies based on metal–ligand interactions. Many have been synthesized unintentionally, while others have been the result of systematic variation of metal and ligand components.Until recently, no one has provided a rational synthetic methodology for the synthesis of metal–ligand coordination clusters. We have provided a model for the assembly of naturally-occurring, high-symmetry protein assemblies such as ferritin (O symmetry) and viral protein coats (I symmetry) based on incommensurate symmetry interactions. This model has led us to develop a rational synthetic approach to the synthesis of high-symmetry clusters based on the metal–ligand coordinate bond.Herein we describe our design strategies, provide examples of triple helicates of stoichiometry M2L3 and tetrahedral clusters of stoichiometry M4L6 and M4L4, and discuss the solution dynamics of these clusters. High symmetry clusters in nature Supramolecular chemistry, which has been called a molecular information science, describes the spontaneous assembly of non-covalently linked molecular clusters of unique shape and composition.1 This requires both a driving force and a dynamic † This review covers papers in our series Coordination Number Incommensurate Cluster Formation. For the most recent paper in that series see ref. 82. system so that all possible molecular structures can be explored to generate the formation of the thermodynamically favored, ideally pre-designed, assembly. An example of such a structure in nature is the iron storage protein apoferritin (Fig. 1).2,3 The protein is composed of 24 non-covalently linked protein subunits that form a nearly spherical shell of octahedral symmetry. Inside the shell up to 4,500 iron atoms can be stored in the form FeO(OH). In eVect, ferritin operates to keep iron in solution as a small particle of rust by stopping the growth of the iron oxide particle before it reaches a size that would result in precipitation. Remarkably, when the apoprotein is dissociated into the individual subunits and allowed to reassemble, only the highly symmetric 24 subunit cluster forms.4 Intermediate assemblies of less than the full 24 complement are only transiently stable.Similar structures are seen in many viruses, in which noncovalently linked assemblies of protein subunits are used to protect the viral nucleic acid. Again, it is generally true that dissociation and reassembly of the protein coat does not give a random polymeric assembly, but instead only the highly symmetric cluster, in many cases 60-mers with icosahedral symmetry. 5,6 Viral coats of this stoichiometry and the ferritin cluster correspond to the pure rotation groups I and O, with 60 and 24 symmetry elements, respectively. That is, each of the protein subunits in the examples given constitutes an asymmetric unit of the cluster.‡ ‡ In the case of certain viruses, a four protein bundle makes up one of the 60 subunits.7 Kenneth N. Raymond was born on January 7, 1942 in Astoria, Oregon. He obtained a B.A.from Reed College in 1964 and Ph.D. in 1968 from Northwestern University (with Basolo and Ibers). He was appointed Assistant Professor at the University of California, Berkeley on July 1, 1967, becoming Associate Professor in 1974 and Dana L. Caulder was born in Columbia, South Carolina in 1972. She earned her B.S. in chemistry from the University of South Carolina in 1994 and Ph.D. from the University of California, Berkeley in 1998. Her Ph.D. thesis focused on the rational design of high symmetry coordination clusters and Professor in 1978.He has served as Vice Chair for the Berkeley Chemistry Department (1982–1984) and Chair (1993–1996). He was Chair of the ACS Division of Inorganic Chemistry in 1996. He received the Lawrence Award of the Department of Energy in 1984, a Humboldt Research Award in 1992, the ACS Bader Award in Bioinorganic Chemistry in 1994 and was elected to the National Academy of Sciences in 1997. He has a long-standing interest in coordination chemistry, both synthetic and biological.the investigation of the unique kinetic and thermodynamic properties of these clusters. Currently she is dividing her time between UCB and the Lawrence Berkeley National Laboratory. At UCB she is the assistant crystallographer and at LBNL she is investigating actinides by XAFS techniques. Kenneth N. Raymond Dana L. Caulder1186 J. Chem. Soc., Dalton Trans., 1999, 1185–1200 Consider again the octahedral symmetry ferritin cluster.Interaction of the protein subunits at the four-fold axis (the view direction in Fig. 1) can be considered a lock-and-key interaction in which the lock and key are 908 apart (Fig. 1). The interaction around the four-fold axis is both a symmetry and stoichiometry requirement: it requires formation of tetramers from the monomeric subunit. Similarly, the interaction of the protein subunits at the three-fold axis (the view direction in Fig. 1, top right) can be regarded as a lock-and-key interaction in which the lock and key are positioned 608 apart (Fig. 1, bottom right). The result is a stoichiometry requirement to form trimers. Simultaneous satisfaction of these two incommensurate n-fold symmetry axes can only be satisfied by formation of a cluster with octahedral symmetry. In a similar fashion the icosahedral cluster is formed through the combination of incommensurate lock-and-key interactions with five-fold and three-fold symmetry.The term “incommensurate” is used in the same sense as applied to incommensurate lattices. When a planar crystal layer with one lattice spacing is deposited on a structurally similar layer with a slightly diVerent lattice spacing, the two layers must curve in order for the unit cells to remain in registry. This is what drives the formation of some naturally-occurring tubular structures.9 In the case of a discrete closed cluster such as ferritin, the surface lattice must curve to satisfy the two incommensurate symmetry interactions.The information for this curvature is programmed into the subunits as the angle between the three-fold and four-fold axes. Rational design of high symmetry coordination clusters The protein–protein interactions described above are formed from many weak hydrogen bonding and van der Waals contacts along large regions or surfaces. The complicated sum of these many individual interactions can still be described by a single vectorial relationship that represents the geometry of the highly directional lock-and-key interactions described above.Metal– ligand interactions, on the other hand, are strong and highly directional, and can be used in place of many weak interactions to direct the formation of multi-metal coordination clusters. There are examples of one-, two- and three-dimensional polymeric assemblies using metals and ligands.10–16 There are also numerous discrete structures, including helices, tetrahedra, squares and cubes among others.17–50 Most of these interesting complexes have been discovered fortuitously by systematic variation of ligand and metal components.Only Fig. 1 Based on the crystal structure of human H chain ferritin,8 the octahedral 24-subunit iron storage protein as viewed down the four-fold (left) and three-fold (right) axes. The four helix bundle protein subunits that directly interact at these symmetry axes are highlighted in yellow and blue, respectively.The interaction at the four-fold axis, in which the lock and key are 908 apart, requires the formation of tetramers. Similarly, the interaction at the three-fold axis, in which the lock and key are 608 apart, requires the formation of trimers. recently has there been progress towards developing a rational synthetic approach to the design and synthesis of such architectures.17–19,21,25,36,38–40,48,51–53 In principle, the formation of clusters of any symmetry should be possible.To do so, the symmetry elements of a particular point group need to be considered. In order to design a cluster with D3 symmetry, an M2L3 triple helicate 41 for example, both the C2 and C3 axes of the point group must be taken into account. A C2-symmetric bis(bidentate) ligand can provide the 2-fold axis, while a metal ion with pseudo-octahedral coordination by three bidentate chelators can provide the 3-fold axis. These symmetry axes must, however, be oriented 908 to one another (Fig. 2). A cluster with T symmetry, an M4L6 tetrahedron17,27,37,50,54 for example, is also possible with the same combination of symmetry elements. In this type of cluster, however, the C2 and C3 axes must be oriented 54.78 from one another (Fig. 2). Design strategies The metal coordination geometry and the orientation of the interaction sites in a given ligand provide the instructions, or blueprint, for the self-assembly of the proposed cluster. As a result, there are several important considerations in designing these supramolecular assemblies based on metal–ligand interactions.Firstly, we choose to use multi-branched chelating ligands because of their increased preorganization and stronger binding as a result of the chelate eVect;55 although, there are numerous examples of supramolecular assemblies based upon multibranched monodentate ligands.25,26,36,40 Secondly, the orientation of the multiple binding units within a ligand must be rigidly fixed so that other, unwanted, cluster stoichiometries or geometries are avoided.Thirdly, because the selfassembly of the thermodynamically-favored cluster from the ligand and metal components involves the making of many metal–ligand bonds, the metals should be labile so that “mistakes” resulting from the initial formation of kinetic products can be corrected. Catecholamide and hydroxamate ligands are excellent choices for binding units in supramolecular complexes because of the high stability and lability of these chelates with 13 metal ions with octahedral coordination environments.56–61 Recently, hydroxypyridinone 62 (HOPO) and pyrazolone 63 ligands have also proven useful in synthesizing supramolecular clusters (Fig. 3). Three catecholamide units coordinating a 13 metal ion will generate a 23 charge for the M(catam)3 unit. In contrast, the hydroxamates, hydroxypyridinones and pyrazolones will form neutral M(ligand)3 units. As part of our approach, the feasibility of the proposed metal–ligand system is explored prior to ligand synthesis using molecular mechanics calculations.64 Although these calculations do not guarantee that the proposed structure will form, they do help eliminate unsuitable structures.If the metal coordination and ligand geometry are correctly chosen, the intended supramolecular cluster should be the only structure that satisfies the binding requirements of the metal, while not creating unfavorable steric interactions in the ligands. Fig. 2 The orientation of the C3 and C2 symmetry axes determines whether a T symmetry tetrahedron or a D3 symmetry triple helix will form.J. Chem. Soc., Dalton Trans., 1999, 1185–1200 1187 Definitions In order to describe this approach to rational design, it will be useful to define terms that more accurately describe the relevant geometric relationships. The vector that represents the interaction between a ligand and metal is the Coordinate Vector (Fig. 4).§ In the case of a monodentate ligand, this vector is simply the one directed from the coordinating atom of the ligand towards the metal ion.In the case of a bidentate ligand, this vector bisects the bidentate chelating group and is directed towards the metal ion. When using chelating ligands, the plane orthogonal to the major symmetry axis of a metal complex is the Chelate Plane (Fig. 5); all of the coordinate vectors of the chelating ligands lie in the chelate plane.Any symmetric coordination complex cluster can be described in terms of the relationships between these chelate planes. In principle, by careful pre-arrangement of coordinate vectors in a multibranched ligand, programming of a cluster of any symmetry or stoichiometry becomes feasible. Although the twist angle 55 is a common measure of the Fig. 3 Catecholamides, hydroxamates, hydroxypyridinones and pyrazolones are useful chelating units for synthesizing supramolecular clusters because of their high stability and lability with 13 metal ions such as Al(III), Ga(III) and Fe(III).NH O OH OH N R O OH N O OH O NH O N N OH X X X pyrazolone catecholamide hydroxamate hydroxypyridinone X Fig. 4 In the case of a monodentate ligand, the Coordinate Vector is the vector from the coordinating atom of the ligand directed towards the metal center. In the case of a bidentate ligand, the coordinate vector is the vector that bisects the chelating group and is directed toward the metal ion.§ Although we previously used the term chelate vector to describe the interaction between a bidentate ligand and a metal ion, we now use the term coordinate vector so that the term is more general and applicable to describing the numerous supramolecular systems based on monodentate ligands. arrangement of three bidentate chelators around a metal ion,¶ the Approach Angle has the advantage that it provides a measure that can be readily compared to angles generated by a given high symmetry cluster (Fig. 6). The approach angle is the angle between the vector connecting the two coordinating atoms of a bidentate ligand projected down the (pseudo) 2-fold axis of the chelate group and the major symmetry axis of the metal center. A twist angle of 608 corresponds to an approach angle of 35.38, while a twist angle of 08 corresponds to an approach angle of 08. M2L3 Complexes Triple helicates The simplest multi-metal cluster contains two metal sites linked by one or more ligands.When these two metal ions are linked by three identical, C2-symmetric ligand strands the resulting bimetallic cluster is called a triple helicate if both metal ions have the same chirality. This chiral M2L3 complex has idealized D3 symmetry: the C3 axis is coincident with, and the three C2 axes are perpendicular to, the helical axis of the complex. There are numerous examples of clusters of this type (Fig. 7).19,45,65–69 An early example of a triple helicate is the iron(III) complex of the dihydroxamate siderophore rhodotorulic acid (H211), which is produced by the yeast Rhodotorula mucilaginosa (previously R.piliminae).70,71 Rhodotorulic acid enantioselectively forms a D-cis complex of Fe2113 stoichiometry at neutral pH, and this complex is currently the only known naturally-occurring example of a triple helicate (Fig. 8). Later, the Fe(III) complex of a hydroxypyridinone (H212) analogue of rhodotorulic acid Fig. 5 The plane orthogonal to the major symmetry axis of the metal complex is the Chelate Plane. In the case of bidentate chelators, all of the coordinate vectors lie in the chelate plane. Fig. 6 An alternative measure of the arrangement of three bidentate chelators around a metal ion is the Approach Angle, which is the angle between the vector connecting the two coordinating atoms of a bidentate ligand projected down the (pseudo) 2-fold axis of the chelate group and the major symmetry axis of the metal center.¶ A twist angle of 608 corresponds to that of a perfect octahedral metal complex, while a twist angle of 08 corresponds to that of a trigonal prismatic metal complex.1188 J. Chem. Soc., Dalton Trans., 1999, 1185–1200 Fig. 7 Ligands which form M2L3 triple helicates. Ligands 1,68 H4245 and H43 67 form helicates with Co(III), Ga(III) and Ti(IV), respectively. Ligands H44–H410 were designed to make M2L3 triple helicates with 13 metal ions like Ga(III), Al(III) and Fe(III).19,65,66 The arrows represent the coordinate vectors, which should be parallel (or be able to become parallel upon complexation), pointing in the same direction if a triple helicate is to form.OH OH OH OH N N N N N N OH OH O NH NH O OH OH NH NH O O OH OH OH OH NH O OH OH NH O OH OH NH O OH OH NH O OH OH NH NH O O OH OH OH OH O NR1 R2 O NR1R2 R1 =R2=(S)-CHMePh * * 1 R1=R2=Pri H43 R1=Pri, R2=Me H44 H49 H48 H47 H46 H45 H410 R1=R2=H H42 was structurally characterized (Fig. 8); the crystalline L,LFe2 123 complex encapsulates a molecule of water.72 As mentioned above, in order to rationally design an M2L3 triple helicate with idealized D3 symmetry, both C2 and C3 axes must be encoded. Using a metal ion with pseudo-octahedral coordination and a C2 symmetric bis(bidentate) ligand, these symmetry axes can be generated (vide supra). These symmetry axes must, however, be oriented 908 to one another. Because the two metal centers share the same C3 helical axis, the two chelate planes in a triple helix must be parallel (Fig. 9). Although a flexible linker may allow for the formation of an M2L3 triple helicate, a rigid linker can direct the formation of an M2L3 triple helicate. Based on this design strategy, a series of M2L3 triple helicates based on ideally planar bis(bidentate) catecholamide ligands has been synthesized (H44–H410, Fig. 7).19,65,66 The rigid aromatic linkers serve to maintain preorganization of the ligand, since other topologies are possible when flexible linkers are used (vide infra).|| The chelate vectors, indicated as arrows, are parallel || Note, however, that a rigid backbone is not necessary if the linker is short enough to prevent both ends of the ligand from coordinating a single metal.45,67 and point in the same direction within each ligand.Molecular mechanics calculations indicated that for each of these ligands the chiral helicate was lower in energy than the meso-M2L3 cluster.64 The M2L3 stoichiometry was confirmed by both fast atom bombardment (FAB) and electrospray mass spectrometry. The crystal structure of the Ga(III) complex of H45 is shown in Fig. 10 and confirms that the rigid ligand forms a racemic mixture of homochiral triple helicates with Ga(III).Triple mesocates A non-chiral M2L3 cluster has a D-configuration at one and a L-configuration at the other metal center, and, therefore, will be called a meso-complex or a mesocate.** This type of cluster has idealized C3h symmetry: rather than having three C2 axes perpendicular to the C3 axis, there is an orthogonal mirror plane that relates the D- to the L-configured metal center.What factors control the formation of a mesocate versus a helicate? Although it has been proposed that the length of an alkyl spacer between two chelating moieties may direct helicate versus mesocate formation 73,76 or that a chiral ligand may be able to ** Since a helix by definition is chiral, the term meso-helicate 73–75 is an oxymoron and will not be used.J. Chem.Soc., Dalton Trans., 1999, 1185–1200 1189 Fig. 8 The dihydroxamate siderophore rhodotorulic acid forms an enantiomerically pure D,D-triple helicate with Fe(III).70,71 The bis(hydroxypyridinone) ligand H212 was synthesized as a rhodotorulic acid analogue and forms a rac-(DD/LL) Fe2123 triple helicate encapsulating a molecule of H2O.72 HN HN O O N N OH O O OH NH HN O O N HO O N O OH O N O O N O Fe Fe O N O O N O Fe Fe Fe2113 3 2FeIII H211 3 2FeIII 3 3 H212 H2O Fe2123 H2O induce a helical twist in an M2L3 complex these explanations do not seem convincing.45 Recently we have presented the first example of a ligand (H213) that makes both a helicate and a mesocate.77 Remarkably, the X-ray analysis showed that in the solid state the Al2133 complex is a chiral helicate (racemic), while the Ga2133 complex is an achiral mesocate (Fig. 11). Although both complexes contain the same ligand, the structures are markedly diVerent: the Fig. 9 In a D3 symmetric triple helicate, the chelate planes are parallel. The spheres represent the pseudo-octahedral metal ions, the rods represent the ligands, and the arrows on the ligand rods indicate the coordinate vectors. Fig. 10 The crystal structure of the triple helicate [Ga273]62. distance between the two metal centers in Al2133 is 7.13 Å, while in Ga2133 this distance is 9.74 Å.The structures show that the helical cavity of Al2133 contains one encapsulated water molecule, while no encapsulated solvent was found in the Ga2133 mesocate (Fig. 11). The encapsulated water molecule in Al2133 is in close contact (2.9–3.0 Å) with six phenolic oxygen atoms that are pointing into the helicate cavity. The rather polar cavity of Al2133 forces the six hydrophobic methyl groups of the aliphatic linkers to point outwards. Only three methyl groups in Ga2133 are pointing outwards, while the remaining three methyl groups are directed into the cluster interior and are shielded from interaction with the solvent (Fig. 11). Other topologies Other topologies are also possible for M2L3 clusters. For example, the dihydroxamate siderophore alcaligin forms an M2L3 complex with Fe(III) in which each metal center is coordinated by one tetradentate and one bridging bis(bidentate) alcaligin ligand (H214, Fig. 12).78 A similar structure type has recently been reported by McCleverty and coworkers (15, Fig. 12).27 The alcaligin topology is more likely with the use of flexible ligands that can wrap around and coordinate a single metal ion, thus the use of rigid linkers can be used to avoid this topology. M4L6 Complexes Another cluster with the same ligand to metal ratio as the M2L3 triple helicate is the M4L6 tetrahedron, where the four metal ions act as the vertices and the six ligands act as the edges of the tetrahedron. Depending on the chiralities at the metal centers, the cluster can have either idealized C3 (DLLL/LDDD), S4 (DDLL) or T (LLLL/DDDD) symmetry.The first examples of M4L6 tetrahedral clusters were the surprises that Saalfrank and coworkers characterized as adamantoid-type clusters of both S4 and T symmetry (H216 and H217, Figs. 13 and 14).50,54,79 Remarkably, an ammonium cation was found to be encapsulated by one of these tetrahedral clusters.50 We have demonstrated the utility of the incommensurate symmetry interaction model in two approaches to the rational design of such clusters.Both approaches employ an ideally planar C2-symmetric bis(bidentate) ligand with a rigid backbone, but the orientation of the C2 axis of the cluster with respect to the plane of the ligand diVers. In the first design strategy, the 2-fold axis of the tetrahedron is intended to lie in the same plane as that defined by the ligand (Fig. 15). Since the chelate vectors must lie within the chelate planes at each of the1190 J.Chem. Soc., Dalton Trans., 1999, 1185–1200 Fig. 11 Ligand H213 forms both a chiral helicate (left) and an achiral mesocate (right).77 The pictures are based on the X-ray structure coordinates. The Al2133 helicate has a molecule of water in the cluster cavity similar to a previously reported iron(III) complex of a rhodotorulic acid analogue.72 four metal vertices, the angle between the chelate vectors within a given ligand must be 70.68.(This angle is simply the supplementary angle to 109.48, which is the angle between the 3-fold axes in a tetrahedron.) A 608 angle is formed for ligands H218 and H419 (Fig. 14); thus, the targeted structure can be achieved with only slight out of plane twisting by each of the chelating groups. Microcrystalline precipitates of Fe4186 or Ga4186 were obtained from the reaction of Fe(acac)3 or Ga(acac)3 with H218 and triethylamine in methanol.51 Both complexes show intense peaks for the M4L6 molecular ions in the FAB1 mass spectrum.In addition, the crystal structure of Ga4186 revealed that the tetrahedral cluster has S4 symmetry (two D and two L metal centers) in the solid state (Fig. 16). The ligand backbone is coplanar with the S4 axis, and there is a substantial cavity, which is partially open to the outside, in the cluster. Four Fig. 12 The dihydroxamate siderophore alcaligin 78 (H214) and the recently reported ligand (15) from McCleverty and coworkers 27 form M2L3 clusters of the above topology with Fe(III) and Ni(II), respectively.The lines represent the ligands and the spheres represent the metal ions. Fig. 13 As described by Saalfrank and coworkers, ligands H216 (left) 79 and H217 (right) 54 form M4L6 tetrahedral clusters with Fe(III).†† The cluster on the left has idealized S4 symmetry, while the cluster on the right has idealized T symmetry and encapsulates a molecule of NH4 1 (yellow). †† One of the four iron atoms in the cluster on the left is Fe(II).crystallographically identical DMF molecules partially fill the cavity. The average carbon to carbon distance between the DMF methyl groups within the cluster cavity is 4.0 Å. Ligand H419 also appears to form a tetrahedral cluster with Ga(III) (Fig. 16).80 The cluster K12Ga4196 precipitates from a methanol solution containing H419, Ga(acac)3 and KOH after six to eight hours. The 1H NMR spectrum (D2O) of the microcrystalline product shows only one set of ligand peaks, indicating a high symmetry solution structure on the NMR timescale.In addition, preparation of the complex in the presence of excess ligand does not disrupt the formation of the desired cluster: the 1H NMR spectrum of this mixture shows two sets of peaks, one for the free and one for the coordinated ligand. Unfortunately the high charge (212) of the cluster has made obtaining X-ray quality single crystals diYcult. The major ions observed in the electrospray mass spectrum (low resolution) match those expected for [K14Ga4196]21, [K15Ga4196]31 and [K16Ga4196]41.In the second design strategy, the 2-fold axis of the tetrahedron is designed to be perpendicular to the ligand plane (Fig. 17). The ideally planar ligand should have parallel coordinate vectors that point in opposite directions. To understand this design it helps to view the tetrahedral cluster as a truncated polyhedron. If the six ligands are to act as the six 2-fold symmetric faces (shown in blue) of the polyhedron, then the angle between the chelate planes (shown in red) is no longer important.The angle between the extended 2-fold plane (blue) and the C3 axis of the cluster is important, however, as this corresponds to the approach angle. This approach angle is 35.38 and corresponds to a perfect octahedral metal complex with a 608 twist angle. Clusters based on this design should be homochiral with idealized T symmetry (i.e., all D or all L metal centers).Ligands H420 17 and H421 81 were designed to form M4L6 tetrahedral clusters based on this strategy (Fig. 18). Molecular modeling 64 of the metal complexes [M = Ga(III), Fe(III)] of H420 indicated that the cluster would have a substantial cavity (250–350 Å3). Solution and solid state observations showed that one of the Et4N1 counterions is encapsulated within the [M4206]122 [M = Ga(III), Fe(III)] cluster interior. The 1H NMR spectrum (D2O) of K5(Et4N)7[Ga4206] showed two sets of Et4N1 resonances split in a 6 : 1 ratio.The larger set of Et4N1 peaks was shifted slightly upfield ‡‡ from free Et4NCl (d = 3.26, q; 1.27, t), while the smaller set was shifted substantially upfield, showing up at negative ppm (d = 20.70, m; 21.59, t)! Based on literature precedent,27,48,82–86 this extreme upfield shift was taken as an indication of the encapsulation of one Et4N1 cation by the tetrahedral cluster host. This assertion was further corro- ‡‡ The upfield shifts of the exterior Et4N1 resonances are attributed to a strong p-cation interaction between the aromatic naphthalene and catechol rings of the ligands and the Et4N1 counterions.J.Chem. Soc., Dalton Trans., 1999, 1185–1200 1191 Fig. 14 Ligands H216 and H217 formed M4L6 tetrahedral clusters as the result of fortuitous accidents. One approach to the rational design and synthesis of such clusters relies on the coordinate vectors (arrows) being 70.68 from each other.Ligands H218 51 and H419 80 form tetrahedral clusters based on this design. N O OH N OH O NH O OH OH NH O OH OH OH MeO2C OH MeO2C MeO O MeO O OH EtO2C EtO O OH EtO2C EtO O H218 H419 H216 H217 borated in the crystal structure of K5(Et4N)7[Fe4206] (Fig. 19), in which the naphthalene rings of the ligands are twisted around the arene–N bond so that they are in van der Waals contact with the encapsulated Et4N1. The distance between the iron atoms in the T symmetry cluster is 12.8 Å, bringing the cluster just into the nanometer regime.In an attempt to make a similar cluster with a larger cavity, ligand H421, based on a 2,7-diaminoanthracene backbone, was prepared (Fig. 18). This ligand also forms an M4L6 tetrahedral cluster, but only in the presence of an alkylammonium guest. The 1H NMR spectrum (D2O) of (Me4N)8[Ti4216] shows a single highly symmetric product with two Me4N1 resonances in a ratio of 7 : 1 (d = 3.97; 22.6).The presence of an extremely Fig. 15 One approach to the synthesis of M4L6 tetrahedral clusters relies on the plane of the ligand being coincident with the 2-fold axis of the tetrahedral cluster. As such, the coordinate vectors within a given ligand must be oriented 70.68 from each other. Fig. 16 Viewed down the crystallographic S4 axis, the structure of Ga4186 (left) revealed four DMF solvent molecules (yellow) pointing into the cluster cavity.51 The minimized 64,80 structure of the T symmetry isomer of the proposed [Ga4196]122 tetrahedron is shown (right).Notice that the highlighted ligands are nearly planar and that these planes are coincident with the 2-fold axes of the clusters. upfield-shifted Me4N1 resonance (22.6 ppm) in a ratio of one Me4N1 to six ligands can be interpreted as a direct indication of the encapsulation of one Me4N1 by the tetrahedral cluster [Ti4216]82.17,27,48,82–86 Again, this conclusion was corroborated in the crystal structure of (Me4N)8[Ti4216] (Fig. 20); one molecule of Me4N1 is located in the cavity of the T symmetry cluster. The distance between the titanium atoms averages 16.1 Å.However, in the absence of an alkylammonium guest molecule, H421 forms an M2L3 triple helicate with Ti(IV) (Fig. 20).81 Although the metal centers within a given complex have the same chirality, the overall structure is significantly distorted from idealized D3 geometry. The local pseudo-C3 axes at the two metal centers are not aligned, and two of the anthracene rings are oriented with their edges directed into the cluster interior, while the third anthracene ring is oriented approximately perpendicular to the plane bisecting the two former anthracene ring planes.This third ligand is substantially nonplanar. It is apparent that the greater bridge length and flexibility of the anthracene ligand allows for the formation of the M2L3 structure, but just barely. Other recently reported T symmetry M4L6 tetrahedral clusters illustrate the generality of the incommensurate symmetry interaction design strategy (Figs. 18 and 19). McCleverty and coworkers 27 synthesized ligand 15, which forms a tetrahedral cluster with Co(III). Solution and solid state evidence indicate that a molecule of BF4 2 is encapsulated in the cluster cavity. Although the ligand backbone is perpendicular to the ends of the ligand, the two chelating ends are essentially coplanar and perpendicular to the pseudo-two-fold axis of the cluster.In addition, the coordinate vectors are parallel and point in opposite directions. Another example for the utility of the design is the tetrahedral Ga(III) cluster by Stack and coworkers based on ligand H422.37 The ligand is essentially planar and Fig. 17 One can envision an M4L6 cluster in which the six ligands act as the six 2-fold symmetric faces (blue) of the truncated polyhedron. This design is ideally suited for a metal center with perfect octahedral coordination (i.e., approach angle = 35.38 or twist angle = 608).1192 J.Chem. Soc., Dalton Trans., 1999, 1185–1200 Fig. 18 Ligands H420 and H421 were designed to make M4L6 tetrahedral clusters. Based on this new design strategy, the coordinate vectors (arrows) in a ligand should be parallel and point in opposite directions. Ligands 1527 and H422 37 are recently reported examples of ligands that form tetrahedral M4L6 clusters and illustrate this design strategy.HN NH O O OH OH HO HO NH HN O O OH OH HO HO N N N N N N NH O OH OH HN O HO HO H421 H422 H420 15 perpendicular to the pseudo-two-fold axis of the T symmetry cluster. Again, the parallel coordinate vectors point in opposite directions. M4L4 Complexes An approach to the synthesis of M4L4 tetrahedral clusters has also been developed. In an M4L4 tetrahedral cluster the metal ions occupy the four vertices and the ligands occupy each of the four faces of the tetrahedron (Fig. 21). This implies that both the ligand and the metal ion must have 3-fold symmetry. As in Fig. 19 H420 forms an M4L6 tetrahedral cluster with Ga(III) and Fe(III). The crystal structure of the [Fe4206]122 cluster is shown at the top with the encapsulated Et4N1 shown in yellow. Ligands 15 and H422 also form M4L6 tetrahedral clusters with Co(II) and Ga(III), respectively. The crystal structures of BF4 2 à [Co4156]81 (bottom left) 27 and [Ga4- 226]122 (bottom right) 37 are shown.Note the similarity of the conformation of the highlighted ligand in each structure: the ligands are nearly planar, and the coordinate vectors are parallel and point in opposite directions. the previously described M2L3 helicates and M4L6 tetrahedra, three bidentate ligands coordinating a pseudo-octahedral metal ion can generate a 3-fold axis at the metal. Rather than using a C2-symmetric ligand, a C3-symmetric ligand can be utilized. This ligand must be rigid, however, so that no two chelating moieties on the ligand can coordinate a single metal ion.Ligand H623 satisfies this requirement (Fig. 21).18 If the ligand is ideally planar, as in the case of H623, then the approach angle for this type of cluster is 19.48 (Fig. 22). This ideal angle is less than four degrees from the approach angle of 238 (corresponding twist angle = 408) observed for tris- (catecholate) complexes of Ti(IV), Ga(III) and Fe(III);56,61,87 therefore, this design seems optimized for metal ions with significant distortions toward trigonal prismatic geometry.Ligand H623 reacted with Al(III), Fe(III) and Ga(III) under basic conditions to give precipitates whose 1H NMR and mass spectra indicated the expected M4L4 species.80 The high charge (122) of these clusters precluded the isolation of X-ray quality single crystals, however, because of the large number of countercations to be ordered in a crystalline lattice. The use of higher oxidation state metal ions [e.g., Ti(IV) and Sn(IV)], despite reducing the lability of the metal–ligand system, would lower the overall charge of the cluster to 82, thus reducing the number of countercations.With this in mind, the Ti(IV) and Sn(IV) complexes of H623 were prepared, and X-ray quality crystals were obtained of the (Et3NH)8[Ti4234] complex (Fig. 22).18 The cluster is a racemic mixture of homochiral tetrahedra (either all D or all L configuration within a given cluster).There is no evidence that the small cavity of the tetrahedron contains a guest, as observed in the previously described M4L6 clusters. The manganese(II) cluster reported by McCleverty and coworkers using ligand 24 further illustrates that a 3-fold symmetric tris(bidentate) ligand can be used to synthesize homochiral M4L4 tetrahedral clusters by self-assembly (Fig. 23). Two metal clusters We have recently demonstrated the rational design of a M2M93L6 mixed-metal cluster in which, rather than using a symmetric ligand to generate a symmetry element, two different metals generate the two incommensurate symmetry elements (Fig. 24).89 In principle, the ligand H225 forms part of an asymmetric unit of the cluster and must have two diVerent incommensurate symmetry interaction sites. As described earlier, a chiral triple helicate has idealized D3 symmetry, whileJ. Chem. Soc., Dalton Trans., 1999, 1185–1200 1193 Fig. 20 Ligand H421 forms an M2L3 helicate in the absence of a Me4N1 guest, but an M4L6 tetrahedron in the presence of Me4N1.The crystal structures of [Ti2213]42 (left) and Me4N1 à [Ti4216]82 (right) are shown.81 an achiral triple mesocate has C3h symmetry. Therefore, to synthesize a mixed-metal helicate (or mesocate) of stoichiometry M2M93L6, one must consider a three-fold interaction site and an orthogonal two-fold (mirror plane) interaction site (Fig. 24). As already illustrated, catechol ligands are relatively hard donors and generate a C3 axis when forming a tris-chelate with hard, trivalent or tetravalent metals [e.g., Al(III), Ga(III), Fe(III), Sn(IV), Ti(IV)].56,59,61,90,91 Phosphine ligands, on the other hand, are soft donors and can generate a two fold axis or mirror plane when coordinated to a square planar metal [e.g., Pd(II) or Pt(II)] in a trans fashion.92,93 A ligand containing both these functionalities arranged in the proper geometry can assemble a M2M93L6 cluster, because it is the smallest discrete species that would simultaneously fulfill the two orthogonal symmetry requirements.Fig. 21 An M4L4 tetrahedral cluster with the metals on the vertices and ligands on the faces of the tetrahedron can be synthesized using ligand H623. Fig. 22 If the ligand is ideally planar, as in the case of H623, then the angle that the 3-fold face (yellow) of the tetrahedron makes with the C3 axis is 19.48 and corresponds to the approach angle. The crystal structure of [Ti4234]82 is shown (right).18 The crystal structure of Cs4[Ti2256(PdBr2)3] shows that the complex has C3h symmetry (Fig. 25); the cluster is a mesocate with one of the titanium atoms having D- and the other having L-configuration.Significantly, three of the Cs1 counterions are located in clefts of the cluster (Fig. 25). Each is coordinated by four of the catecholate oxygens and two molecules of THF. The clefts of the cluster are so deep that the coordinating THF molecules can also be described as being buried. The palladium-coordinated bromine atoms are not in van der Waals contact with the caesium atoms, but they do shield the caesium atoms from other solvent molecules, helping to explain the low coordination number (6) of the Cs1 cations. Fig. 23 McCleverty and coworkers reported the M4L4 tetrahedron [Mn4244]41. The manganese atoms are ferromagnetically coupled.88 Fig. 24 A cluster with D3 (or C3h) symmetry can be designed using an asymmetric ligand H225.89 Interaction of the catechol moiety with an octahedral metal ion (blue spheres) can generate the necessary C3 axis, while interaction of the phosphine moiety with a square planar metal ion (red spheres) can generate the C2 axis (or mirror plane).Simultaneous satisfaction of these two symmetry requirements can lead to a cluster with D3 (or C3h) symmetry.1194 J. Chem. Soc., Dalton Trans., 1999, 1185–1200 Fig. 25 In the stereoview (top) of [Ti2256(PdBr2)3]42 the Pd atoms are colored purple and the titanium atoms are colored orange.Viewed down the crystallographic 3-fold axis, this space filling model shows the buried caesium cations (green) and their coordinated THF molecules. The hydrogen atoms are omitted for clarity.89 m(MxLy) versus n(MxLy) clusters Especially with high symmetry structures, it is imperative to have molecular weight data if structure assignment is to be certain, as there are several examples of ligands that make clusters of both m(MxLy) and n(MxLy) stoichiometries.McCleverty and coworkers reported that ligand 15 makes an M4L6 tetrahedron with Co(II) but an M2L3 alcaligin-type cluster with Ni(II) (Figs. 12, 18 and 19).27 We have reported that ligand H421 makes an M4L6 tetrahedron in the presence of an alkylammonium guest but an M2L3 triple helicate in the absence of guest (Fig. 20).81 Lehn and coworkers have reported both penta- and hexanuclear circular helicates of Fe(II) using ligand 26; in the absence of a Cl2 guest the hexanuclear circular helicate forms, while in the presence of Cl2 the pentanuclear helicate forms (Fig. 26).42 It has been similarly suggested that molecular “squares” (27) are in equilibrium in solution with molecular “triangles” (28) as shown in Fig. 27.31,94 In each of the above examples, the metal–ligand ratio is the same between the two structures, and therefore spectroscopic evidence or analytical data is of limited value. Only high resolution mass spectrometry or X-ray diVraction data could distinguish between an m(MxLy) or an n(MxLy) structure.Recently Stang and coworkers reported the synthesis of a nanoscale molecular hexagon (29, Fig. 27).39 Characterization of this compound included spectral and analytical results, but did not include mass spectrometry or crystallographic data. The formulation of the complex as a hexagon was based on the use of a linker with a 1208 bond angle. The discrepancy between the 1208 angle needed for a hexagon and the 1088 angle needed for an analogous pentagon, for example, is easily accounted for when one considers the size of the molecule.One edge of either of these macrocycles is over 23 Å. This includes seven angles that, when compared to analogous angles taken from the Cambridge Structural Database,95 would be expected to deviate from the ideal by two to three degrees each. This is more than enough to make up for the necessary 128 per edge.Even the previously reported “molecular square”, based on 908 angles, has corner angles around square planar platinum of 838.96 Independent molecular modeling 64 of both the hexagon (29) Fig. 26 Lehn and coworkers reported that the hexadentate bipyridine ligand 26 shown above makes a pentanuclear [Fe5265]101 circular helicate with Fe(II) in the presence of the guest Cl2, but a hexanuclear [Fe6266]121 circular helicate in the absence of Cl2.42J. Chem. Soc., Dalton Trans., 1999, 1185–1200 1195 Fig. 27 Fujita and coworkers have presented evidence that suggests the molecular “square” (27) and “triangle” (28) shown above are in equilibrium in solution.94 Stang and coworkers have suggested that the molecular “hexagon” (29) exclusively forms from the ligand and metal components based on the use of a linker with a 1208 angle.39 Independent molecular modeling 64 of both the proposed hexagon and an analogous pentagon reveal no obvious steric interactions that would cause the hexagon to be favored over the pentagon. and the pentagon (30) revealed no obvious steric interactions or strain around the metal centers in either proposed structure (Fig. 27). While NMR data can point to the existence of a symmetric structure, and both NMR and elemental analysis can determine metal–ligand or ligand–ligand9 ratios, only mass spectrometry or crystallographic data can give the full formula and structure. Dynamics of supramolecular clusters The geometric requirements for synthesizing clusters of various stoichiometries and symmetries are beginning to be understood. It is less clear, however, how these clusters assemble in solution from the ligand and metal components, and once assembled, how the clusters function.For example, how is1196 J. Chem. Soc., Dalton Trans., 1999, 1185–1200 geometric information transmitted between the multiple coordination sites of a given ligand? Is there coupling of the isomerization of chiral metal centers and, if so, what is the magnitude of this coupling as transmitted through the rigid ligand? For clusters that recognize and encapsulate guest molecules, what are the factors controlling the recognition process and how do guests enter and exit the cluster cavity? M2L3 Triple helicate stereo-isomerism dynamics When using rigid ligands to synthesize triple helicates, the chirality of the first metal center should induce the same chirality at the second metal center, so that only DD- and LLconfigured complexes are present.The magnitude of the mechanical coupling between the two metal centers and the mechanism of the inversion reaction have been investigated using the dinuclear Ga(III) complexes of ligands H48–H410 and similar mononuclear Ga(III) complexes based on simple bidentate catecholamide ligands.65,66,97 The methyl groups on the isopropyl substituents of H48 are magnetically equivalent in the free ligand, but become magnetically inequivalent around the chiral metal center.The activation parameters for this process were derived from an Eyring plot of the first-order rate constants calculated by line shape analysis. Consistent with an intramolecular mechanism, these parameters are not solvent dependent. The free energy inversion barrier (DD LL) for K6[Ga283] in DMSO-d6 (79.8 kJ mol21) or D2O solutions (78.7 kJ mol21, pD = 12.1) is only 1.2 times higher compared to the corresponding mononuclear complex.Two limiting cases for coupling of the two metal centers and their chirality can be considered: for weak coupling the barrier should remain essentially unchanged. However, for very strong coupling the two centers must move through the trigonal-prismatic transition state simultaneously (Fig. 28), and consequently the activation barrier would be expected to be eVectively twice the barrier for inversion of the mononuclear complex. The kinetic data show weak coupling of both metal centers that is about 22.6 kJ mol21.Thus, it is concluded that inversion of the L,L- and D,D-[Ga283]62 helicates involves the heterochiral L,D-[Ga283]62 anion as an intermediate, which is produced by a single twist event along the reaction pathway. At lower pD a second mechanism becomes dominant in D2O. In contrast to the mononuclear complex, the dinuclear Fig. 28 Potential energy diagrams and stereochemical courses for intramolecular inversion of (a) the mononuclear complex and (b) the L,L-[Ga283]62 and D,D-[Ga283]62 dinuclear complexes involving the heterochiral L,D-[Ga283]62 as an intermediate (solid lines).Inversion with a hypothetical, concerted twisting of both metal centers is indicated by a dashed line. K6[Ga283] helicate shows a second order proton dependence below pD = 7. This constitutes a remarkable confirmation of the mechanism outlined in Fig. 28: inversion of one center, which occurs rapidly because of the single protonation, does not change the overall chirality owing to the higher energy of the heterochiral intermediate and its consequent short lifetime.Only when the second metal center is also protonated can the overall inversion of the helicate occur. In the absence of mechanical coupling of the metal centers only a single proton dependence would be expected because the heterochiral intermediate would have the same energy as the homochiral anions and, consequently, a long lifetime. rac-(ƒƒ/ÀÀ)-M2L3 Helicate to Àƒ-M2L3 mesocate interconversion dynamics As noted earlier, in the solid state H213 forms a helicate with Al(III) but a mesocate with Ga(III) (Fig. 11).77 The methyl substituents in the H213 backbone serve as markers for following the solution structure of the metal complexes by 1H NMR; in the helicate these two methyl groups are equivalent, while in the mesocate the methyl groups are diastereotopic.As expected for the mesocate, the 1H NMR spectrum of Ga2133 in DMSO-d6 shows two singlets for the methyl groups in the ligand spacer; however, the presence of an additional singlet indicates that the helicate form of this complex is also present in solution.Variable temperature 1H NMR experiments reveal that these two structures are in thermodynamic equilibrium, with the helicate being preferred at high temperatures. Additional investigations revealed that the spontaneous meso-to-helix conversion is an entropy-driven process, which must be a consequence of diVerent numbers of solvent molecules associated with the two forms of the complex.77 While Al2133 also displays dynamic behavior, it is considerably slower than the corresponding Ga(III) complex. It has been previously shown that inversion of Ga(III) helicates is fast on the NMR time scale and proceeds through an intramolecular Bailar twist.65,66,97 Inversion of configuration in mononuclear Al(III) and Ga(III) complexes proceeds through the same mechanism,98,99 and the isomerization or exchange rates for Ga(III)-trischelates are consistently faster than for Al(III)-trischelates.100 Stereoisomerism in M4L6 tetrahedral clusters As described, the M4L6 tetrahedral cluster based on ligand H218 crystallizes as the S4 isomer (DDLL chiralities at the four metal vertices).51 Low temperature 1H NMR experiments reveal, however, that Ga4186 is a mixture of T (DDDD/LLLL), C3 (DLLL/LDDD) and S4 (DDLL) isomers in solution (CDCl3).20 With decreasing temperature the broad resonance of one of the ligand protons, which is pointing into the cavity, splits into five distinct peaks, representing the three isomers.The integration of the peaks at 220 K yields a ratio of C3 :S4 :T isomers of 58 : 38 : 4. If the distribution between the isomers were purely statistical, one would expect the ratio of C3 :S4 :T isomers to be 50 : 37.5 : 12.5. Although the isomers are not present in an exact statistical distribution, the distribution shows that the stabilities of the three isomers are very similar, and, therefore, the mechanical coupling between the metal centers is negligible.Ligand exchange in hydroxamate iron(III) complexes has been previously studied,58 but isomerization of a simple trishydroxamate iron(III) or gallium(III) complex is certainly too fast to follow by NMR. The slower rate of interconversion detected here can be attributed to the geometric properties of the ligand and the cluster.In order for a metal center to change its chirality it is necessary to pass through a trigonal prismatic transition state. Since four coordination centers are tethered in the tetrahedron, the Bailar twist is the only mechanically possible rearrangement. To do this, the ligands in contact with the active metal must pass through a conformation in which theJ. Chem. Soc., Dalton Trans., 1999, 1185–1200 1197 ligand’s two coordinate vectors can not coexist in the chelate planes of each metal center.In eVect, because the ligand maintains an angle of only 608 in its planar form it forces a very distorted trigonal prismatic intermediate. The stereochemical courses and potential energy diagram for the isomerization of DDDD to LLLL cluster are drawn in Fig. 29. In order to isomerize from DDDD to LLLL, the cluster has to go through all intermediate stereoisomers. Since both NMR observations and MM2 calculations suggest that all these isomers are very close in energy, they are drawn here at the same energy level. The potential energy diagram can be then simplified if we assume that isomerization of DDLL to DDDL will have the same energy barrier as its isomerization to DLLL, since both processes require inversion of configuration at only one metal center.The inversion from DDDL to DDDD must have the same energy barrier as inversion of DLLL to LLLL, since these are mirror image processes. Coalescence of the 1H NMR resonances is observed at 300 K, corresponding to an activation barrier of 58 kJ mol21.Self-recognition in M2L3 triple helicates A diVerent issue of designed order was addressed in a family of helicate complexes of varying, but fixed, metal–metal distance. It was intended that the information stored in rigid bis(catecholamide) ligands (H44–H46, Fig. 7) be used to overcome the intrinsic disorder of mixtures to produce a highly ordered system of complexes in solution. These ligands are unique in that, because of the rigidity and varying distances between the catecholamide functionalities, it is geometrically impossible to form a mixed ligand (M2L2L91)62 complex.There are few synthetic examples of self-recognition, despite nature’s ability to perform this feat in many ways. Lehn and coworkers have demonstrated the self-recognition of double stranded copper(I) helicates that diVer in the number (2–5) of metal coordination sites.101 More recently Stack and coworkers have shown that racemic mixtures of chiral ligands stereoselectively form complexes in which all ligands are of the same chirality.37,102 When mixtures of any two or all three of the ligands shown in Fig. 7 are equilibrated at room temperature with Ga(acac)3 in basic methanol, both 1H NMR spectroscopy and electrospray mass spectrometry indicate that only the individual complexes form (Fig. 30). Remarkably, no oligomeric or mixed-ligand species are observed in solution.Selective encapsulation of alkylammonium guests by a tetrahedral cluster host The tetrahedral cluster [Ga4206]122 shows remarkable discrimination between alkylammonium guests.17 There are orders of magnitude diVerences between the association equilibrium Fig. 29 Potential energy diagram and stereochemical courses for intramolecular inversion of the T symmetry (LLLL/DDDD), C3 symmetry (LLLD/DLLL) and S4 symmetry (LLDD) isomers of Ga4186. constants, Keq, for Me4N1, Et4N1 and Pr4N1, and these differences allow for the quantitative step-wise exchange of one guest for another (Fig. 31). If Pr4N1 is added to a solution of K6(Me4N)6[Ga4206], the Pr4N1 quickly (<1 min) and quantitatively is incorporated into the cluster cavity, displacing Me4N1. In turn, if Et4N1 is added to this same solution, the Et4N1 displaces the Pr4N1 rapidly (<1 min) and quantitatively! In the presence of either Me4N1, Et4N1 or Pr4N1, the tetrahedral cluster selectively encapsulates Et4N1.No mixtures are observed by 1H NMR. The thermodynamic parameters for the inclusion reaction in water have been determined by measuring the temperature dependence of the association equilibrium constants (Keq).82 Since the exchange between free and encapsulated guests is slow on the NMR time scale, their relative concentrations can be determined by integration of the corresponding 1H NMR resonances. In the absence of any other guests the cavity of the [Ga4206]122 host will most likely be filled with solvent molecules. Both 39K NMR spectroscopy and the lack of a K1 concentration dependence on the equilibrium values suggest that the K1 cations are located outside the cavity and do not interact with the interior of the cluster.The van’t HoV plots for the encapsulation of Me2Pr2N1, Pr4N1 and N,N,N9,N9-tetramethyl-1,3-propanediammonium by the host [Ga4206]122 anion show that encapsulation of the cationic guests into this dodecaanion is an endothermic process.The enthalpies and entropies are both positive; the encapsulation is an entropy-driven process. Encapsulation of a cation by a dodecaanion is endothermic due to the very large, and dominant, solvation enthalpies of the ions (Fig. 32). The free energy of hydration is predicted by the Born equation to be 2162 z2 r21 kcal mol21, where z are units of charge and r the diameter of the ion in Å.103 The corresponding entropy of hydration is 22.8 z2 r21 kcal mol21 at 298 K, predicting a DH of hydration of 2165 z2 r21 at 298 K.Because DH of hydration is z2 dependent, solvation of the 212 anion is the largest term. This and the cation solvation override the enthalpy gained on partial charge neutralization. This model also makes a clear prediction that higher charge cations will not be encapsulated and that highly solvated, singly charged cations (i.e., K1) should be poor guests. These predictions are con- firmed by the observed behavior.The major entropy terms in the host–guest complexation reaction are all positive. A large entropy gain upon host–guest complexation is predicted based both on desolvation of the ions and release of encapsulated water by the host (Fig. 33). This model is consistent with other examples of ion pairing or com- Fig. 30 Schematic representation of self-recognition in gallium(III) triple helicates. The diVerent sized rods represent the diVerent length ligands. Spheres represent the gallium ions.Fig. 31 Schematic representation of stepwise guest exchange from the cavity of the tetrahedral cluster [Ga4206]122. The red spheres represent Me4N1, the green spheres Pr4N1 and the blue spheres Et4N1.1198 J. Chem. Soc., Dalton Trans., 1999, 1185–1200 plexation of metallic ions with anionic ligands, which are also entropically driven processes.104 We have further investigated the role of solvent by measuring the thermodynamic parameters of the encapsulation of alkylammonium cations by the tetrahedral cluster [Ga4206]122 in various solvents.105 The DG values for the encapsulation event show excellent correlation with several empirical scales that describe the polarity of the solvent, suggesting that the selectivity and extent of inclusion depends on the solvation of the dodecaanionic host and the cationic guest.Guest-induced M2L3 helicate to M4L6 tetrahedron conversion We have shown that two diVerent clusters,81 a triple helicate and a tetrahedron, can be prepared using identical ligand (H421) and metal components (Fig. 20).27,37,42,43,106–109 Simply the addition of an appropriate guest is enough to shift the equilibrium from the entropically preferred helicate to the tetrahedron. Since the only diVerence in the two systems described above is the presence or absence of Me4N1, it should be possible to transform a triple helicate into a tetrahedral cluster simply upon addition of Me4N1. In order to test this hypothesis, the Ga(III) analogues were prepared because of the greater lability of Ga(III) compared to Ti(IV). The addition of 20 equivalents of Me4NCl to a K6[Ga2213] solution in D2O revealed that this was indeed possible.Complete transformation of the helicate into the tetrahedral cluster was observed via 1H NMR spectroscopy over the course of 5 days (pD 6.5, T 40 8C). Similar studies of this transformation at higher pD values (pD 7.5) and lower temperatures (room temperature) showed lower conversion rates due to the slower kinetics of metal–ligand rearrangement under these conditions.Fig. 32 Born–Haber cycle for guest encapsulation by [Ga4206]122. Solvation of the negative twelve anion is the largest enthalpy term (DH4) because of the z2 dependence of solvation enthalpy. The enthalpy of the encapsulation event in aqueous solution (denoted by a red X) was determined experimentally. Fig. 33 Upon encapsulation of the guest, the “frozen” solvent molecules in the cluster cavity are released, resulting in a favorable entropy gain. Summary In this review we have illustrated the utility and generality of an approach to the designed synthesis of supramolecular clusters based on metal–ligand interactions.An analysis of the high symmetry seen in the natural protein clusters (e.g., ferritin and viral protein coats) is based on the incommensurate symmetry numbers of the interaction sites and the fixed relative angles between these symmetry axes.The use of this model in the successful design of several metal–ligand clusters is illustrated. Rigid ligand geometries, while chosen to accommodate the targeted cluster geometry, preclude the formation of alternative structures. This process is greatly facilitated by molecular modeling in the early stages of design. Triple helicates of M2L3 stoichiometry are based on bis(bidentate) ligands with C2 symmetry interacting with “octahedral” metal centers that generate C3 axes when coordinated by three bidentate chelators. When the angle between the C3 and C2 axes is rigidly fixed at 908 by the use of a rigid linker between the two coordinating ends of the ligand, a chiral triple helicate is the most likely structure.When less rigid linkers are used, achiral mesocates and even alcaligin-type topologies are increasingly possible. Tetrahedral clusters of M4L6 stoichiometry are similarly based on bis(bidentate) C2-symmetric ligands and “octahedral” metal centers.In contrast to the M2L3 helicates, when the angle between the C3 and C2 axes is rigidly fixed at 57.48, a tetrahedral cluster results. Two strategies for achieving this geometry and the resulting M4L6 tetrahedral clusters are presented. Alternatively, M4L4 tetrahedra can be synthesized by imposing C3 symmetry on the ligand, which is designed to span the face of the tetrahedron and bridge three metal vertices. In addition to illustrating these design approaches towards tetrahedral structures prospectively in the synthesis of numerous clusters, we have also shown retrospectively that several clusters reported by others conform to the geometric parameters called for by our model.The initial investigation of the dynamic behavior of these synthetic supramolecular clusters has begun. We are beginning to understand the mechanical coupling (or lack thereof) between chiral metal centers in M2L3 and M4L6 clusters, the kinetics and host-guest chemistry of multi-metal complexes, the self-recognition properties in pre-designed rigid systems and the dramatic role that guest molecules can play in the formation of clusters of n(MxLy) (n = 1, 2, 3, .. .) stoichiometries. The host–guest chemistry of these clusters oVers the first promise of achieving synthetically what nature accomplishes in the supramolecular clusters of ferritin and viral protein coats. In both cases the natural clusters protect valuable guest molecules by providing a nanometer scale environment that is significantly diVerent from the surrounding solution.We have described the first indication that we can significantly alter the properties of the guest molecules in the host clusters we have prepared. The further development of the reaction chemistry of the encapsulated guests is an exciting prospect. Acknowledgements This research is supported by the National Science Foundation through Grant No. CHE-9709621 and exchange grants from NATO (SRG 951516) and NSF (INT-9603212).We thank Darren W. Johnson for his help with manuscript preparation. References 1 J.-M. Lehn, Supramolecular Chemistry: Concepts and Perspectives, VCH, Weinheim, 1995. 2 P. M. Harrison and P. Arosio, Biochim. Biophys. Acta, 1996, 1275, 161. 3 P. M. Proulxcurry and N. D. Chasteen, Coord. Chem. Rev., 1995, 144, 347.J. Chem. Soc., Dalton Trans., 1999, 1185–1200 1199 4 S. Stefanini, P. Vecchini and E. Chiancone, Biochem., 1987, 26, 1831. 5 E. Arnold and M. G. Rossmann, Acta Crystallogr., Sect. A, 1988, 44, 270. 6 D. L. D. Caspar and A. Klug, Physical Principles in the Construction of Regular Viruses, in Proceedings of Cold Spring Harbor Symposia on Quantitative Biology, Long Island Biological Association, New York, 1962, vol. 27. 7 A. Hadfield, J. Hajdu, M. S. Chapman and M. G. Rossmann, Acta Crystallogr., Sect. D, Biol. Crystallogr., 1995, 51, 859. 8 D. M. Lawson, P. J. Artymiuk, S.J. Yewdall, J. M. A. Smith, J. C. Livingstone, A. TreVry, A. Luzzago, S. Levi, P. Arosio, G. Cesarini, C. D. Thomas, W. V. Shaw and P. M. Harrison, Nature, 1991, 349, 541. 9 E. J. W. Whittaker, Acta Crystallogr., 1965, 9, 855; 862; 865. 10 J. Lu, T. Paliwala, S. C. Lim, C. Yu, T. Niu and A. J. Jacobson, Inorg. Chem., 1997, 36, 923. 11 O. M. Yaghi, H. Li and T. L. Groy, Inorg. Chem., 1997, 36, 4292. 12 O. Crespo, M. C. Gimeno, P. G. Jones, A. Laguna and C. Sarroca, Chem. Commun., 1998, 1481. 13 R. K. Kumar, S. Balasubramanian and I. Goldberg, Chem. Commun., 1998, 1435. 14 T. F. Magnera, L. M. Peslherbe, E. Körblová and J. Michl, J. Organomet. Chem., 1997, 548, 83. 15 B. F. Abrahams, B. F. Hoskins, J. Liu and R. Robson, J. Am. Chem. Soc., 1991, 113, 3045. 16 M. J. Zaworotko, Chem. Soc. Rev., 1994, 283. 17 D. L. Caulder, R. E. Powers, T. Parac and K. N. Raymond, Angew. Chem., Int. Ed. Engl., 1998, 37, 1840. 18 C. Brückner, R. E. Powers and K. N. Raymond, Angew.Chem., Int. Ed. Engl., 1998, 37, 1837. 19 D. L. Caulder and K. N. Raymond, Angew. Chem., Int. Ed. Engl., 1997, 36, 1439. 20 T. Beissel, R. E. Powers, T. N. Parac and K. N. Raymond, J. Am. Chem. Soc., 1999, in the press. 21 K. N. Raymond, D. L. Caulder, R. E. Powers, T. Beissel, M. Meyer and B. Kersting, Proc. of the 40th Robert A. Welch Found. on Chem. Res., 1996, 40, 115. 22 L. Carlucci, G. Ciani, P. Macchi and D. M. Proserpio, Chem. Commun., 1998, 1837. 23 J. L. Heinrich, P.A. Berseth and J. R. Long, Chem. Commun., 1998, 1231. 24 K. K. Klausmeyer, T. B. Rauchfuss and S. R. Wilson, Angew. Chem., Int. Ed. Engl., 1998, 37, 1694. 25 M. Fujita, S.-Y. Yu, T. Kusukawa, H. Funaki, K. Ogura and K. Yamaguchi, Angew. Chem., Int. Ed. Engl., 1998, 37, 2082. 26 S. Roche, C. Haslam, H. Adams, S. L. Heath and J. A. Thomas, Chem. Commun., 1998, 1681. 27 J. S. Fleming, K. L. V. Mann, C.-A. Carraz, E. Psillakis, J. C. JeVery, J. A. McCleverty and M. D. Ward, Angew.Chem., Int. Ed. Engl., 1998, 37, 1279. 28 R. Schneider, M. W. Hosseini, J.-M. Planeix, A. DeCian and J. Fischer, Chem. Commun., 1998, 1625. 29 T. Konno, K. Tokuda and K. Okamoto, Chem. Commun., 1998, 1697. 30 J.-P. Sauvage, Acc. Chem. Res., 1998, 37, 611. 31 S. B. Lee, S. Hwang, D. S. Chung, H. Yun and J.-I. Hong, Tetrahedron Lett., 1998, 39, 873. 32 M. Albrecht, Chem. Soc. Rev., 1998, 27, 281. 33 C. J. Jones, Chem. Soc. Rev., 1998, 27, 289. 34 E. C. Constable, M. Neuburger, L.A. Whall and M. Zehnder, New J. Chem., 1998, 219. 35 O. D. Fox, N. K. Dalley and R. G. Harrison, J. Am. Chem. Soc., 1998, 120, 7111. 36 B. Olenyuk, A. Fechtenkötter and P. J. Stang, J. Chem. Soc., Dalton Trans., 1998, 1707. 37 E. J. Enemark and T. D. P. Stack, Angew. Chem., Int. Ed. Engl., 1998, 37, 932. 38 P. J. Stang, Chem. Eur. J., 1998, 4, 19. 39 P. J. Stang, N. E. Persky and J. Manna, J. Am. Chem. Soc., 1997, 119, 4777. 40 P. J. Stang and B. Olenyuk, Acc. Chem.Res., 1997, 30, 502. 41 C. Piguet, G. Bernardinelli and G. Hopfgartner, Chem. Rev., 1997, 97, 2005. 42 B. Hasenknopf, J.-M. Lehn, N. Boumediene, A. Dupont-Gervais, A. VanDorsselaer, B. Kneisel and D. Fenske, J. Am. Chem. Soc., 1997, 119, 10956. 43 B. Hasenknopf, J.-M. Lehn, B. O. Kneisel, G. Baum and D. Fenske, Angew. Chem., Int. Ed. Engl., 1996, 35, 1838. 44 E. J. Enemark and T. D. P. Stack, Inorg. Chem., 1996, 35, 2719. 45 E. J. Enemark and T. D. P. Stack, Angew. Chem., Int.Ed. Engl., 1995, 34, 996. 46 A. F. Williams, Pure Appl. Chem., 1996, 68, 1285. 47 D. S. Lawrence, T. Jiang and M. Levett, Chem. Rev., 1995, 95, 2229. 48 M. Fujita, D. Oguro, M. Miyazawa, H. Oka, K. Yamaguchi and K. Ogura, Nature, 1995, 378, 469. 49 R. W. Saalfrank, R. Burak, S. Reihs, N. Löw, F. Hampel, H.-D. Stachel, J. Lentmaier, K. Peters, E.-M. Peters and H. G. vonSchnering, Angew. Chem., Int. Ed. Engl., 1995, 34, 993. 50 R. W. Saalfrank, R. Burak, A. Breit, D. Stalke, R.Herbst-Irmer, J. Daub, M. Porsch, E. Bill, M. Müther and A. X. Trautwein, Angew. Chem., Int. Ed. Engl., 1994, 33, 1621. 51 T. Beissel, R. E. Powers and K. N. Raymond, Angew. Chem., Int. Ed. Engl., 1996, 35, 1084. 52 M. Fujita, J. Yazaki and K. Ogura, J. Am. Chem. Soc., 1990, 112, 5645. 53 M. Fujita, S. Nagao, M. Iida and K. Ogura, J. Am. Chem. Soc., 1993, 115, 1574. 54 R. W. Saalfrank, A. Stark, M. Bremer and H. Hummel, Angew. Chem., Int. Ed. Engl., 1990, 29, 311. 55 F. A.Cotton and G. Wilkinson, Advanced Inorganic Chemistry, 5th edn., John Wiley and Sons, New York, 1988. 56 B. A. Borgias, S. J. Barclay and K. N. Raymond, J. Coord. Chem., 1986, 15, 109. 57 B. Borgias, A. D. Hugi and K. N. Raymond, Inorg. Chem., 1989, 28, 3538. 58 M. T. Caudle and A. L. Crumbliss, Inorg. Chem., 1994, 33, 4077. 59 T. M. Garrett, P. W. Miller and K. N. Raymond, Inorg. Chem., 1989, 28, 128. 60 M. J. Kappel, V. L. Pecoraro and K. N. Raymond, Inorg. Chem., 1985, 24, 2447. 61 T. B. Karpishin, T. D. P. Stack and K. N. Raymond, J. Am. Chem. Soc., 1993, 115, 6115. 62 J. Xu and K. N. Raymond, 1999, manuscript in preparation. 63 J. Xu, D. W. Johnson and K. N. Raymond, 1998, unpublished results. 64 CAChe 4.0, Oxford Molecular Group, Inc., USA, 1997. 65 B. Kersting, M. Meyer, R. E. Powers and K. N. Raymond, J. Am. Chem. Soc., 1996, 118, 7221. 66 M. Meyer, B. Kersting, R. E. Powers and K. N. Raymond, Inorg. Chem., 1997, 36, 5179. 67 M. Albrecht and S.Kotila, Angew. Chem., Int. Ed. Engl., 1996, 1208. 68 L. J. Charbonniere, G. Bernardinelli, C. Piguet, A. M. Sargeson and A. F. Williams, J. Chem. Soc., Chem. Commun., 1994, 1419. 69 L. J. Charbonniere, M.-F. Gilet, K. Bernauer and A. F. Williams, Chem. Commun., 1996, 39. 70 C. J. Carrano and K. N. Raymond, J. Chem. Soc., Chem. Commun., 1978, 501. 71 C. J. Carrano, S. R. Cooper and K. N. Raymond, J. Am. Chem. Soc., 1979, 101, 599. 72 R. C. Scarrow, D. L. White and K. N. Raymond, J.Am. Chem. Soc., 1985, 107, 6540. 73 M. Albrecht and S. Kotila, Angew. Chem., Int. Ed. Engl., 1995, 34, 2134. 74 M. Albrecht and S. Kotilla, Chem. Commun., 1996, 2309. 75 M. Albrecht, H. Rottele and P. Burger, Chem. Eur. J., 1996, 2, 1264. 76 M. Albrecht and C. Riether, Chem. Ber., 1996, 129, 829. 77 J. Xu, T. Parac and K. N. Raymond, Angew. Chem., Int. Ed. Engl., 1999, submitted for publication. 78 Z. Hou, C. J. Sunderland, T. Nishio and K. N. Raymond, J. Am. Chem. Soc., 1996, 118, 5148. 79 R. W. Saalfrank, B. Horner, D. Stalke and J. Salbeck, Angew. Chem., Int. Ed. Engl., 1993, 32, 1179. 80 R. E. Powers, The Rational Design of Supramolecular Assemblies, Ph.D. Thesis, University of California, Berkeley, CA, 1997. 81 M. Scherer, D. L. Caulder, D. W. Johnson and K. N. Raymond, Angew. Chem., Int. Ed. Engl., 1999, submitted for publication. 82 T. Parac, D. L. Caulder and K. N. Raymond, J. Am. Chem. Soc., 1998, 120, 8003. 83 P. Jacopozzi and E. Dalcanale, Angew. Chem., Int. Ed. Engl., 1997, 36, 613. 84 S. Mann, G. Huttner, L. Zsolnai and K. Heinze, Angew. Chem., Int. Ed. Engl., 1996, 35, 2808. 85 J. Bryant, M. T. Blanda, M. Vincenti and D. J. Cram, J. Am. Chem. Soc., 1991, 113, 2167. 86 P. Timmerman, W. Verboom, F. C. J. M. vanVeggel, J. P. M. vanDuynhoven and D. N. Reinhoudt, Angew. Chem., Int. Ed. Engl., 1994, 33, 2345. 87 D. L. Kepert, Inorganic Stereochemistry, Springer Verlag, Heidelberg, 1982. 88 A. J. Amoroso, J. C. JeVery, P. L. Jones, J. A. McCleverty, P. Thornton and M. D. Ward, Angew. Chem., Int. Ed. Engl., 1995, 34, 1443.1200 J. Chem. Soc., Dalton Trans., 1999, 1185–1200 89 X. Sun, D. W. Johnson, D. L. Caulder, R. E. Powers, K. N. Raymond and E. H. Wong, Angew. Chem., Int. Ed. Engl., 1999, in the press. 90 C. G. Pierpoint and R. M. Buchanan, Coord. Chem. Rev., 1981, 38, 45. 91 C. G. Pierpoint and C. W. Lange, Inorg. Chem., 1994, 41, 331. 92 C. A. McAuliVe, in Comprehensive Coordination Chemistry, vol. 2, ed. G. Wilkinson, F. G. A. Stone and F. W. Abel, Pergamon, Oxford, 1987, ch. 14. 93 W. Levason, in The Chemistry of Organophosphorus Compounds, vol. 1, ed. F. R. Hartley, Wiley, New York, 1990, ch. 16. 94 M. Fujita, O. Sasaki, T. Mitsuhashi, T. Fujita, J. Yazaki, K. Yamaguchi and K. Ogura, Chem. Commun., 1996, 1535. 95 F. H. Allen and O. Kennard, Chem. Des. Automat. News, 1993, 8, 31. 96 P. J. Stang and D. H. Cao, J. Am. Chem. Soc., 1994, 116, 4981. 97 B. Kersting, J. R. Telford, M. Meyer and K. N. Raymond, J. Am. Chem. Soc., 1996, 118, 5712. 98 J. R. Hutchison, J. G. Gordon and R. H. Holm, Inorg. Chem., 1971, 10, 1004. 99 R. C. Fay and T. S. Piper, Inorg. Chem., 1964, 3, 348. 100 S. S. Eaton, G. R. Eaton, R. H. Holm and E. L. Muetterties, J. Am. Chem. Soc., 1973, 95, 116. 101 R. Krämer, J.-M. Lehn and A. Marquis-Rigault, Proc. Natl. Acad. Sci., USA, 1993, 90, 5394. 102 M. A. Masood, E. J. Enemark and T. D. P. Stack, Angew. Chem., Int. Ed. Engl., 1998, 37, 928. 103 C. S. G. Phillips and R. J. P. Williams, Inorganic Chemistry, 1st edn., vol. 1, Oxford University Press, New York, 1965. 104 I. M. Klotz, Ligand-Receptor Energetics, 1st edn., John Wiley & Sons, Inc., New York, 1997. 105 T. N. Parac, D. L. Caulder and K. N. Raymond, 1999, manuscript in preparation. 106 R. Krämer, J.-M. Lehn, A. DeCian and J. Fischer, Angew. Chem., Int. Ed. Engl., 1993, 32, 703. 107 C. Provent, S. Hewage, G. Brand, G. Bernardinelli, L. J. Charbonniére and A. F. Williams, Angew. Chem., Int. Ed. Engl., 1997, 36, 1287. 108 F. M. Romero, R. Ziessel, A. Dupont-Gervais and A. VanDorsselaer, Chem. Commun., 1996, 551. 109 P. N. W. Baxter and J.-M. Lehn, Chem. Commun., 1997, 1323. Paper 8/08370C

ISSN:1477-9226    

DOI:10.1039/a808370c

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1999, 1201–1202 1201 Solvent extraction of strontium nitrate by a crown ether using room-temperature ionic liquids † Sheng Dai,*a Y. H. Ju b and C. E. Barnes b a Chemical Technology Division, Oak Ridge National Laboratory, PO Box 2008, Oak Ridge, TN 37831-6181, USA. E-mail: i9d@ornl.gov b Department of Chemistry, University of Tennessee, Knoxville, TN 37996-1600, USA Received 10th December 1998, Accepted 2nd February 1999 The preliminary results described here show that unprecedentedly large distribution coefficient (D) values can be achieved using ionic liquids as extraction solvents for the separation of metal ions by crown ethers.This work highlights the vast opportunities in separation applications for ionic liquids with crown ethers. The discovery of crown ethers has led to a new class of ligands for alkaline and alkaline earth metal cations.1 They have found wide applications in designing novel solvent extraction systems that are selective for certain metal ions based on the sizes of the crown-ether rings.2 These systems utilize organic phases containing crown ethers to extract target metal ions from aqueous solutions through the complexation of the targeted radionuclides and the specific crown ethers.The eYciency of such extraction processes is strongly dependent not only on cations but also counter anions.3 The alkaline and alkaline earth cations which coexist with hydrophobic counter anions in aqueous solutions are more readily extracted into the organic phases than those with hydrophilic anions.3 In fact, the distribution coeYcient (D) for the extraction of strontium nitrate into organic phases from its aqueous solutions by crown ethers is always less than one, even though the thermodynamic driving force for the complexation of Sr21 with a number of crown ethers is very favorable.3 This distribution coeYcient is low because the solvation free energies for simple inorganic anions such as nitrate by organic solvents are not favored thermodynamically. 2 Accordingly, diYculty in increasing solvent extraction eYciency using crown ethers as extractants lies in the unfavorable transport of anions from organic phases to inorganic phases. Limited solubilities of ionic species in nonionic organic solvents are the main problem associated with conventional solvent extractions. A number of strategies have been proposed to address this problem including the addition of hydrophobic anions to aqueous solutions.2,3 The obvious drawback to this approach is that more chemicals would be added to the system thereby possibly increasing the toxicities and complexity of the original aqueous solutions.We have been interested in developing ambient-temperature and hightemperature ionic liquids as new separation media for actinides and fission products.4 In this communication, we report the use of a number of room-temperature ionic liquids as extraction solvents to remove strontium nitrate from aqueous phases into ionic liquids by a crown ether.The distribution coeYcients for these ionic liquid extraction systems are several orders of magnitude better than those observed for current extraction systems based on organic solvents. The use of ionic liquids to separate toxic metal ions and † The submitted manuscript has been authored by a contractor of the U.S. Government under contract No.DE-AC05-96OR22464. Accordingly, the U.S. Government retains a paid-up, nonexclusive, irrevocable, worldwide license to publish or reproduce the published form of this contribution, prepare derivative works, distribute copies to the public, and perform publicly and display publicly, or allow others to do so, for U.S. Government purposes. organic molecules has been previously investigated by us 4 and others.5 Notably, Rogers and co-workers have reported the successful extraction of organic acids from aqueous solutions into an ionic liquid (1-butyl-3-methylimidazolium hexafluorophosphate). 5 However, the distribution values for a number of organic acids are normally similar or less than those of the corresponding extraction systems based on non-ionic organic solvents.In this case,5 the main advantage of using this particular room-temperature ionic liquid as an extraction solvent is the low vapor pressure of the ionic liquid as compared to most conventional organic solvents. The unique solvation capability of the room-temperature ionic liquid has not been used to full advantage.The very intrinsic property of the ionic liquid is that it consists only of ions, and that it can be made hydrophobic.6 The novel dual properties of these new ionic liquids make them eYcient solvents for the extraction of ionic species from aqueous solutions. From a thermodynamic perspective, the solvation of ionic species, such as crown-ether complexes, NO3 2, and SO4 22, in the ionic liquids, should be much more favored thermodynamically than those of conventional solvent extractions.This is one of the key advantages of using ionic liquids in separations involving ionic species. This prompted us to consider using these ionic liquids as extraction solvents for ionic species. In this report, we describe the eVects ionic liquids have on improving the ability of crown ethers to remove metal ions from aqueous solutions. To our knowledge, no solventextraction system based on mixtures of ionic liquids and crown ethers has been reported in the literature.The room-temperature ionic liquids (see below) used in this work were 1-R1-2-R2-3-methylimidazolium bis[(trifluoromethyl) sulfonyl]amide (R1R2MeIm1Tf2N2), where R1 = ethyl, propyl, or butyl and R2 = H, or methyl, and 1-R1-2-R2-3-methylimidazolium hexafluorophosphate (R1R2MeIm1 PF6 2). These ionic liquids were synthesized as described in the literature and are known to be immiscible in water.4,5,7,8 The concentration of water in the ionic liquids was too low to be measurable by FTIR, as reported previously.8 The metal compound used in this work was strontium nitrate, because 90Sr is a fission product and there is no eYcient extraction methodology available for its removal from radioactive waste sites.The crown ether chosen was dicyclohexyl-18-crown-6 (2,3,11,12-dicyclohexano-1,4,7, 10,13,16-hexaoxacyclooctadecane), which is known to form a strong complex with Sr21.2 The extraction experiments were conducted by contacting 1 mL of an ionic liquid with a 10 mL aqueous solution of Sr(NO3)2 (1.5 × 1023 M) for about 2 hours in a vibrating mixer.The pH value of the initial aqueous solution was 4.10. The concentration of the crown ether in the ionic liquids was 0.15 M. For comparison purposes, identical extraction experiments were conducted in which strontium nitrate was extracted from N N R1 Me R2 X– (Anion) +1202 J.Chem. Soc., Dalton Trans., 1999, 1201–1202 an aqueous phase into conventional organic extraction solvents such as toluene and chloroform containing 0.15 M of the same crown ether. Uptake of strontium nitrate by the ionic liquid or organic phases was measured by determining the remaining Sr21 in the aqueous solutions with inductively coupled plasma atomic emission (ICP-AE). Table 1 gives extraction results for the six ionic liquids and two conventional extraction solvents tested.It is clear that the distribution coeYcients of the ionic liquids using the crown ether as an extractant are all greater that those of the organic solvents. Without the crown ether, the partitioning of Sr(NO3)2 into the ionic liquid phases is very small indicating that the crown ether plays an important role as a complexing extractant. This is consistent with our assertion that the solvation free energies of ion pairs by the ionic liquids are very favorable, thereby increasing the extraction coeYcients.The highest distribution coeYcient value is over 10,000, which is four orders of magnitude greater than those of the conventional extraction systems.2 The concentration ratio of Sr21 in the ionic liquid phase versus the aqueous phase D, is strongly dependent on the R1 and R2 groups. The D values of the ionic liquids with R2 = H are always larger than those with R2 = Me. This is consistent with our previous observation that the solubility of ionic uranium compounds in ionic liquids 4 is highly influenced by the R2 group through the ionic interaction between the solvent cationic molecules and ionic solutes.When R2 = H, we expect that hydrogen bonding interactions will be much stronger than when R2 = Me. Accordingly, higher solubility and distribution coeYcients result for the ionic liquids with R2 = H. The D values of the ionic liquids with bis[(trifluoromethyl)- sulfonyl]amide anion are much larger than those with hexa- fluorophosphate anion.This indicates that the anion also plays a key role in solvation of the crown ether complex. This diVerence, which is induced by the solvent anions, can be rationalized by considering the diVerence in the anion size of two ionic liquid systems. The cationic complex of Sr21 and the crown ether form a large cation, which is stabilized when solvated by a big anionic species. Since the bis[(trifluoromethyl)sulfonyl]- amide anion is larger in size than that of the hexafluorophosphate anion, the D values of the former extraction systems are larger than those of the latter.Support for this rationale can be found by comparing the corresponding distribution values of Sr21 for the ionic liquids without the crown ether. The D values of the plain ionic liquids with Tf2N2 anion are very close to those of the plain ionic liquids with PF6 2 anion. Under this condition, the strontium cation, whose size is much smaller Table 1 Comparison of the extraction results obtained using ionic liquids and conventional solvents Extract phase BuMe2ImPF6 BuMeImPF6 EtMe2ImTf2N EtMeImTf2N PrMe2ImTf2N PrMeImTf2N C6H5CH3 CHCl3 Da (with crown ether in melts) 4.2 2.4 × 101 4.5 × 103 1.1 × 104 1.8 × 103 5.4 × 103 7.6 × 1021 7.7 × 1021 D(without crown ether in melts) 0.67 0.89 0.81 0.64 0.47 0.35 nmb nm a The D value is defined as D = [Molten salt concentration of Sr21]/ [aqueous solution concentration of Sr21].b nm = not measurable. than that of its complex with the crown ether, will be directly solvated by the solvent anions. Accordingly, there are no solvation advantages in this case. The same argument has been employed in explaining the diVerences in clathrate formation induced by crown ethers.9 In conclusion, the preliminary results described here show that unprecedentedly large D values can be achieved using ionic liquids as extraction solvents for the separation of metal ions by crown ethers.Furthermore, the distribution coeYcient values can be easily tailored by varying the substituting groups in the imidazolium cation and the counter anions. This work highlights the vast opportunities in separation applications for ionic liquids with crown ethers. Based on the same principle, we also envision the potential future uses of the ionic liquids containing crown ethers as a selective liquid membrane and liquid ionselective electrodes. Acknowledgements The authors wish to thank Mr M.C. Burleigh for his help to run ICP-AE analyses of the samples and Mr Y. S. Shin for his assistance with synthesizing ionic liquid precursors. The authors also wish to thank Dr P. V. Bonnesen and Dr B. A. Moyer for helpful discussions. This work was supported by the Division of Chemical Sciences, OYce of Basic Energy Sciences, U.S. Department of Energy, under contract No. DE-AC05- 960R22464 with Lockheed Martin Energy Research Corp.References 1 C. J. Pedersen, J. Am. Chem. Soc., 1967, 89, 7017; C. J. Pedersen, Angew. Chem., Int. Ed. Engl., 1988, 27, 1021; D. J. Cram, Angew. Chem., Int. Ed. Engl., 1988, 27, 1009; J. Lehn, Angew. Chem., Int. Ed. Engl., 1988, 27, 89. 2 M. Hiraoka, Crown Compounds: Their Characteristics and Applications, Elsevier, Amsterdam, 1982; L. F. Lindoy, The Chemistry of Macrocyclic Ligand Complexes, Cambridge University Press, Cambridge, 1990; B. A. Moyer, Complexation and Transport, in Molecular Recognition: Receptors for Cationic Guests, ed.G. W. Gokel, in Comprehensive Supramolecular Chemistry, ed. J. L. Atwood, J. E. D. Davies, D. D. MacNicol, F. Vögtle and J.-M. Lehn, Pergamon, Elsevier, Oxford, 1996, vol. 1, ch. 10, pp. 377–416. 3 W. J. McDowell, Sep. Sci. Technol., 1988, 23, 1251; W. J. McDowell and R. R. Shoun, Proceedings of ISEC’77, 1977, 1, 95. 4 S. Dai, Y. Shin, L. M. Toth and C. E. Barnes, Inorg. Chem., 1997, 36, 4900; S. Dai, L. M. Toth, G. D. Del Cul and D. H. Metcalf, J. Phys. Chem., 1996, 100, 220; S. Dai, L. M. Toth, G. D. Del Cul and D. H. Metcalf, Inorg. Chem., 1995, 34, 412. 5 J. G. Huddleston, H. D. Willauer, R. P. Swatloski, A. E. Visser and R. D. Rogers, Chem. Commun., 1998, 1765. 6 C. L. Hussey, Adv. Molten Salt Chem., 1983, 5, 185; C. L. Hussey, in Chemistry of Nonaqueous Solvents, ed. A. Popov and G. Mamantov, VXH Publishers, New York, 1994, ch. 4; R. T. Carlin and J. S. Wilkes, in Chemistry of Nonaqueous Solvents, ed. A. Popov and G. Mamantov, VXH Publishers, New York, 1994, ch. 5; K. R. Seddon, J. Chem. Technol. Biotechnol., 1997, 68, 351; Y. Chauvin and H. Olivier-Bourbigou, CHEMTECH, 1995, 25, 26. 7 J. S. Wilkes and M. J. Zaworotko, J. Chem. Soc., Chem. Commun., 1992, 965. 8 P. Bonhote, A. P. Dias, N. Papageorgiou, K. Kalyanasundaram and M. Gratzel, Inorg. Chem., 1996, 35, 1168; N. Papageorgiou, Y. Athanassov, M. Armand, P. Bonhote, H. Pettersson, A. Azam and M. Gratzel, J. Electrochem. Soc., 1996, 143, 3099. 9 H. Hassaballa, J. Steed, P. C. Junk, M. R. J. Elsegood, Inorg. Chem., 1998, 37, 4666. Communication 8/09672D

ISSN:1477-9226    

DOI:10.1039/a809672d

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1999, 1203–1204 1203 Heterogeneous photocatalysis for synthetic purposes: oxygenation of cyclohexane with H3PW12O40 and (nBu4N)4W10O32 supported on silica Alessandra Molinari, Rossanno Amadelli, Leonardo Andreotti and Andrea Maldotti * Dipartimento di Chimica, Centro di Studio su Fotoreattività e Catalisi del C.N.R., Università degli Studi di Ferrara, Via L. Borsari 46, 44100, Ferrara, Italy. E-mail: mla@dns.unife.it Received 8th February 1999, Accepted 26th February 1999 Heterogenisation of H3PW12O40 and (nBu4N)4W10O32 with silica provides new photocatalytic systems able to oxygenate cyclohexane with O2 without causing the oxidative mineralization processes which typically occur on the surface of the photosensitive semiconductor TiO2.Catalytic oxidation of unactivated C–H bonds under mild conditions (room temperature and atmospheric pressure), using environmentally friendly reagents such as O2, is a field of ever growing interest.1 In this context, a number of authors are investigating photoexcited polyoxometalates, which exhibit a noticeable activity in the oxidation of numerous organic compounds, including saturated hydrocarbons.2 Some of our recent contributions in this research area deal with the oxygenation of alkanes photocatalysed by (nBu4N)4W10O32 and by H3PW12O40 in homogeneous solution.3 In view of the interest in the heterogenisation of polyoxometalates, 4 the photocatalytic activity for cyclohexane oxidation of the two mentioned polyoxotungstates supported on silica is here reported for the first time.We demonstrate that both supported polyoxotungstates present the important advantage of being more easily handled than in the homogeneous phase. An important aspect is that their photocatalytic activity can be investigated in media where they are insoluble. Specifically, we could investigate both of them in pure cyclohexane and observe a quite good selectivity for the ketone formation in the case of H3PW12O40/SiO2.Since we describe heterogeneous photosensible systems, the comparison with a dispersed semiconductor such as TiO2 becomes natural. Contrary to this well investigated semiconductor,5 the (nBu4N)4- W10O32/SiO2 photocatalyst does not cause any mineralisation of the substrate, while maintaining a comparable eYciency for the reaction under investigation. H3PW12O40 was heterogenised on silica following the previously described † ‘impregnation’ procedure which results in the fixation of the polyoxoanion to the support, possibly through electrostatic interactions with the protonated silica surface.6 We successfully applied the same method to support (nBu4N)4W10O32.† Since this compound is present in the form of cation–anion aggregates in organic solvent,7 it is likely that it is adsorbed on silica as an ionic couple, with the tetraalkylammonium cations acting as a bridge between the surface and the decatungstate anion.This interpretation is supported by previous investigations on the modes of polyoxomolybdate adsorption on silica,8 as well as by infrared spectra, which reveal the presence of (nBu4N)1 cations and W10O32 42 on the surface of the support.‡ Table 1 reports the photocatalytic properties of the so obtained heterogeneous systems in the oxygenation of cyclohexane at 20 8C and 760 Torr of O2, in diVerent dispersing media and at excitation wavelengths chosen on the basis of the absorption spectra of the two polyoxotungstates § (entries 1,2,6,7).Table 1 also shows some data previously obtained in homogeneous solution (entries 3 and 8), where the cyclohexane photooxidation quantum yields are 0.35 at 325 nm and 0.03 at 254 nm for (nBu4N)4W10O32 and H3PW12O40, respectively.3 Finally, entries 4 and 5 report the results obtained using TiO2 powder dispersions to induce cyclohexane photooxidation, according to previously reported investigations.9 Table 1 Photocatalytic oxidation of cyclohexane by O2 (760 Torr) Entry 1 b 2 b 3 c 4 d 5 d 6 b 7 b 8 c Photocatalytic system (nBu4N)4W10O32/SiO2 (nBu4N)4W10O32/SiO2 (nBu4N)4W10O32 3 TiO2 9 TiO2 9 H3PW12O40/SiO2 H3PW12O40/SiO2 H3PW12O40 3 Solvent or dispersing medium C6H12 C6H12/CH2Cl2 1/1 CH2Cl2/C6H12/CH3CN 6/3/1 C6H12 C6H12/CH2Cl2 1/1 C6H12 C6H12/CH2Cl2 1/1 CH2Cl2/C6H12/CH3CN 6/3/1 Excitation wavelengths/ nm (irradiation time/min) l > 280 (90) l > 280 (90) l > 280 (90) l > 280 (90) l > 280 (90) l = 254 (360) l = 254 (360) l = 254 (360) Product ratio a cyclohexanone: cyclohexanol 11 1 5.7 1.5 2.3 1.1 1 CO2 a/ mmol <0.5 <0.5 <0.5 35 <0.5 <0.5 <0.5 Alcohol 1 ketone/ mmol 22 36 25 23 42 44 6 a Analysis of cyclohexanol and cyclohexanone was carried out by gas chromatography and carbon dioxide was determined by a turbidimetric method using an absorbing solution of Ba(OH)2 in glycerol.Reported values are ±10%. b 3 ml of dispersing medium containing 15 g l21 of supported polyoxotungstate. 1.45 mmol of irradiated polyoxotungstate.c 3 ml of solution 2 × 1024 mol dm23. 0.6 mmol of irradiated polyoxotungstate. CH3CN is necessary for the dissolution of the catalyst. d 3 ml of dispersing medium containing 3 g l21 of TiO2.1204 J. Chem. Soc., Dalton Trans., 1999, 1203–1204 In every case, photoexcitation of the two polyoxotungstates leads to the oxidation of cyclohexane to cyclohexanol and cyclohexanone as the main stable products (more than 90% of the overall oxidised alkane).The ketone to alcohol concentration ratio is always close to one except in entry 6 where the H3PW12O40/SiO2 presents a good selectivity for the ketone formation. Iodometric analysis indicates that hydroperoxides, which are proposed to be the primary products during the oxygenation of alkanes by illuminated (nBu4N)4W10O32,4,10 are present only in negligible amounts (less than 5% of the overall oxidized cyclohexane).As far as the stability of the photocatalysts is concerned, it is noteworthy that they can be used for subsequent cycles of oxidation without any release of the polyoxotungstates during the experiments, and without any loss of their photocatalytic activity. It is to be noted that entries 1–5 in Table 1 allow a semiquantitative comparison of the photocatalytic properties of (nBu4N)4- W10O32, (nBu4N)4W10O32/SiO2 and TiO2 because (i) excitation wavelength range and irradiation time are the same; (ii) the amounts of the photocatalysts have been chosen so as the maximum absorption of the incident light is the same for all the systems (see notes b, c and d in Table 1).On this basis, we can state that heterogenisation of (nBu4N)4W10O32 inhibits only partially its photocatalytic eYciency expressed as the ratio between the value of the mmoles of oxidized cyclohexane, reported in the last column of Table 1 (entries 2, 3), and the mmoles of irradiated decatungstates: 25 in the heterogeneous system and 41 mmol in the homogeneous one (see notes b, c and d in Table 1).Despite a diVerent product distribution, both in cyclohexane (entries 1, 4) and in mixed solvent (entries 2, 5), the mmoles of oxidised cyclohexane with (nBu4N)4W10O32/SiO2 are very close to those obtained with TiO2. Although a quantitative comparison between the two heterogeneous systems is diYcult, it is reasonable to speculate that the photoactive centres on the heterogeneous polyoxotungstate catalyst are sensibly lower than for an equal amount of TiO2.This makes the (nBu4N)4- W10O32/SiO2 system even more interesting from an eYciency point of view. Another reason that makes the heterogenised decatungstate very promising for applied synthetic purposes is that, in contrast to TiO2, it does not induce any mineralisation process of the substrate. In fact, it photocatalyses the oxygenation of cyclohexane to cyclohexanone and cyclohexanol without the formation of carbon dioxide.Acknowledgements This research was supported by M.U.R.S.T and C.N.R. (project 95/95-5%). Notes and references † 0.1 g of catalyst H3PW12O40 or (nBu4N)4W10O32 was dissolved in a suitable solvent (H2O and CH3CN, respectively) and then 1 g of colloidal silica (0.012 mm, Strem Chemicals) was added. After stirring and evaporation of the solvent, the obtained powder contained about 10% (w/w) of polyoxotungstate. ‡ Infrared spectra of (nBu4N)4W10O32 /SiO2 were recorded in KBr using a diVuse reflectance accessory.nBu4N1: stretching of –CH3 and –CH2 (2963 and 2875 cm21), and bending of C–H (1481 and 1383 cm21). W10O32 4–: 958, 890, 802 cm21. §Irradiation of (nBu4N)4W10O32 /SiO2 and of TiO2 was performed with a Hanau Q 400 Hg lamp, using a suitable cut-oV filter (l > 280 nm; 15 mW cm22), while H3PW12O40/SiO2 system was irradiated with a Hg low pressure lamp (l = 254 nm; 3 mW cm22).All the experiments were carried out at 20 ± 1 8C under oxygen at 760 Torr. 1 A. Bielanski and J. Haber, Oxygen in Catalysis, Marcel Dekker, New York, Basel, Hong Kong, 1991; A. Kroty and J. P. Kingsley, Chemtech, 1996, 39. 2 M. T. Pope, Heteropoly and Isopoly Oxometalates, Springer Verlag, Berlin, Heidelberg, New York, Tokyo, 1993; C. L. Hill and C. M. Prosser-McCartha, Photosensitization and Photocatalysis using Inorganic and Organometallic Compounds, Kluwer Academic Publishers, Dordrecht, 1993. 3 A. Maldotti, A. Molinari, P. Bergamini, R. Amadelli, P. Battioni and D. Mansuy, J. Mol. Catal., 1996, 113, 147; A. Maldotti, A. Molinari, R. Argazzi, R. Amadelli, P. Battioni and D. Mansuy, J. Mol. Catal., 1996, 114, 141; A. Maldotti, R. Amadelli, V. Carassiti and A. Molinari, Inorg. Chim. Acta, 1997, 256, 309; A. Molinari, A. Maldotti, R. Amadelli, A. Sgobino and V. Carassiti, Inorg. Chim. Acta, 1998, 272, 197. 4 N. Mizuno and M. Misono, Chem. Rev., 1998, 98, 199; L. K. Volkova, E. S. Rudakov and V. P. Tretyakov, Kinet. Katal., 1996, 37, 540; 1995, 36, 373. 5 M. R. HoVmann, S. T. Martin, W. Choi and D. W. Bahnemann, Chem Rev., 1995, 95, 69; P. Pichat, Catal. Today, 1994, 19, 313. 6 Y. Wu, X. Ye, X. Yang, X. Wang, W. Chu and Y. Hu, Ind. Eng. Chem. Res., 1996, 35, 2546. 7 M. Fournier, R. Thouvenot and C. Rocchiccioli-DeltcheV, J. Chem. Soc., Faraday Trans., 1991, 87, 349. 8 C. Rocchiccioli-DeltcheV, M. Amirouche, M. Che, J. M. Tatibout and M. Fournier, J. Catal., 1990, 125, 292; A. Chemseddine, C. Sanchez, J. Livage, J. P. Launay and M. Fournier, Inorg. Chem., 1984, 23, 2609. 9 P. Boarini, V. Carassiti, A. Maldotti and R. Amadelli, Langmuir, 1998, 14, 2080; R. Amadelli, M. Bregola, E. Polo, V. Carassiti and A. Maldotti, J. Chem. Soc., Chem. Commun., 1992, 1355; W. Mu and J. M. Pichat, Catal. Lett., 1989, 3, 73. 10 L. P. Ermolenko and C. Giannotti, J. Chem. Soc., Perkin Trans. 2, 1996, 1205; L. P. Ermolenko, J. A. Delaire and C. Giannotti, J. Chem. Soc., Perkin Trans. 2, 1997, 25. Communication 9/01051C

ISSN:1477-9226    

DOI:10.1039/a901051c

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 1205–1207 1205 On the mechanism of carboxylate ligand scrambling at Mo2 41 centers: evidence for a catalyzed mechanism Malcolm H. Chisholm * and Ann M. Macintosh Department of Chemistry, Indiana University, Bloomington, IN 47405, USA Received 13th January 1999, Accepted 1st March 1999 The reaction between Mo2(O2CBut)4 and Mo2(O2CCF3)4 has been studied by 1H and 19F NMR spectroscopy in the solvents C6D6, CD2Cl2, CD3CN, pyridine-d5 and THF-d8.In each solvent ligand exchange is observed with the formation of Mo2(O2CBut)3(O2CCF3) 1, cis- and trans-Mo2(O2CtBu)2(O2CCF3)2 (cis-2, trans-2) and Mo2(O2CBut)- (O2CCF3)3 3. The approach to equilibrium is solvent dependent with the rate being C6D6 ~ CD2Cl2 > CD3CN and THF-d8. Attempts to quench the ligand exchange by the addition of proton and carboxylate anion traps such as BaCO3, Cs2CO3, proton sponge, 2,6-di-tert-butylpyridine (2,6-But 2-py) and [Mo2(O2CBut)2(CH3CN)4]21[BF4 2]2 all failed.In the presence of [Mo2(PhNCHNPh)2(CH3CN)4]21[BF4 2]2 and 2,6-But 2-py the ligand exchange reaction is halted. These data are used to argue for a catalyzed ligand exchange reaction involving free carboxylate anion or carboxylic acid in the reaction between Mo2(O2CBut)4 and Mo2(O2CCF3)4 to give Mo2O2(CBut)n(O2CCF3)4 2 n, where n = 0–4. Similarly, the reaction between Mo2(O2CBut)4 and [Mo2(O2CBut)2(CH3CN)4]21[BF4 2]2 to give [Mo2(O2CBut)3(CH3CN)2]1[BF4 2] is suppressed in the presence of 2,6-di-tert-butylpyridine.Introduction In our attempts to prepare parallel and perpendicularly linked polymers of M–M quadruply bonded complexes supported by carboxylate ligands 1 we have been thwarted by deleterious side reactions leading to polymer/oligomer degradation. The chief cause of this polymer degradation seems to be due to carboxylate group scrambling reactions. Such reactions have been noted before by Cotton 2 in studies of eqn.(1), where n = 0 to 4 and Mo2(O2CH)4 1 Mo2(O2CCF3)4 Mo2(O2CH)n(O2CCF3)4 2 n (1) by us in a similar reaction between Mo2(O2CBut)4 and Mo2- (O2CH2But)4.3 In the present work we describe studies of the closely related reaction shown in eqn. (2) as a function of various solvents and Mo2(O2CBut)4 1 Mo2(O2CCF3)4 Mo2(O2CBut)n(O2CCF3)4 2 n (2) additives aimed at elucidating the facility and mechanism of this ligand exchange process (n = 0–4). Although it is known that certain compounds with M–M quadruple and triple bonds may associate to give tetranuclear species, e.g. 2W2(OiPr)6 W4(OPri)12 4 and 2Mo2Cl4L2- (HOMe)2 æÆ Mo4Cl8L4 1 4MeOH,5 it is hard to see how an associative mechanism leading to a [Mo4(O2CR)8] complex could achieve this facile exchange. We therefore suspected that the mechanism for ligand exchange might involve fortuitous RCO2H which was present in the solution since we have previously studied the fluxionality of the Mo2(m-O2CBut)4(h1-O2- CBut)2 anion6 in benzene-d6 and a well known route to Mo2(O2CR)4 compounds involves the reaction shown in eqn.(3).7 Mo2(O2CMe)4 1 RCO2H (excess) heat solvent Mo2(O2CR)4 1 4MeCO2H (3) Results and discussion Reaction (2) is readily followed by both 1H and 19F NMR spectroscopies. The compounds Mo2(O2CBut)3(O2CCF3) 1, and Mo2(O2CBut)(O2CCF3)3 3, are formed concurrently. Subsequently the formation of Mo2(O2CBut)2(O2CCF3)2 2, which exists in cis and trans isomers, is seen. In a typical reaction ca. 5 mg of each of the pivalate and trifluoroacetate were mixed (in as close to a 1 : 1 mole ratio as possible, given the small quantities involved) and dissolved in the deuteriated NMR solvent ca. 0.7 mL. The results obtained by monitoring reaction (2) with 19F NMR are qualitatively similar to those obtained using 1H NMR. However, the interpretation of the 19F NMR spectra was complicated by the presence of one or more overlapping resonances. In addition, chemical shift values in 19F NMR spectra are highly sensitive to changes in temperature and solvent.Therefore, we report only the results based on the 1H NMR spectra. The reactions were monitored at room temperature in a Varian Gemini-2000 spectrometer. Within 1 h in benzene-d6 and CD2Cl2 there is essentially no evidence for the homoleptic carboxylates and after ca. 3 h an equilibrium mixture of 1, 2, and 3 is formed. Rather interestingly, we see that after 1 h the formation of 2 is in a 1 : 1 ratio of cis and trans isomers but with time the ratio of the cis isomer increases to ca. 2.5 : 1 in benzene-d6. Thus the trans isomer of 2 must be formed kinetically at essentially the same rate as the cis even though the cis isomer is favored thermodynamically (statistics alone give a 2 : 1 cis : trans preference). A series of 1H spectra showing the formation of 1 and 3 and 2 (cis 1 trans) are shown in Fig. 1. Qualitatively the same results are seen in CD2Cl2 but in CD3CN and THF-d8, the formation of 1, 3 and 2 are notably slower such that equilibrium is not attained until ca. 48 h. Also in the more polar CD3CN the thermodynamic preference for cis-2 : trans-2 is ca. 5 : 1. The initial formation of 2, however, has a smaller cis to trans ratio, once again indicating a kinetic preference for trans ligand exchange. Attempts to follow 2 in pyridine-d5 are thwarted by a direct reaction between Mo2- (O2CCF3)4 and pyridine which causes formation of h1-O2CCF3 ligands.81206 J.Chem. Soc., Dalton Trans., 1999, 1205–1207 The rate of ligand scrambling, which followed the order C6D6 ª CD2Cl2 > CD3CN ª THF-d8, led us to question the role of axial ligation to the Mo2 41 centers. We therefore investigated the scrambling in C6D6 in the presence of added PPh3 (ca. 60 equiv). [This tertiary phosphine is known to bind axially to Mo2(O2CCF3)4 without displacing any Mo–O bonds.8] However, the added PPh3 failed to yield any significant decrease in the rate of carboxylate group scrambling. We also investigated the influence of dilution on the rate of carboxylate scrambling and found that with dilution the rate of scrambling was decreased. While this could be viewed as evidence for a bimolecular reaction pathway involving an [Mo2]2 species it could also result from a bimolecular pathway involving [Mo2] and an adventitious reagent.In order to investigate the possible role of adventitious carboxylic acid in promoting a catalyzed ligand exchange reaction we studied the reaction (2) in the presence of various additives intended to capture any mischevious free RCO2H.Finely divided BaCO3 and Cs2CO3 were introduced since they could act as scavengers for carboxylic acids. However, these failed to suppress reaction (2). Since it is necessary to trap both a proton source and any free carboxylate anion, we studied reaction (2) in the presence of [Mo2(O2CBut)2(CH3- CN)4]21[BF4 2]2 and proton sponge [1,8-bis(dimethylamino)- naphthalene].However, here a reaction occurred between [Mo2(O2CBut)2(CH3CN)4]21[BF4 2] and the added proton sponge. This led us to investigate reactions involving the use of 2,6-di-tert-butylpyridine (2,6-But 2-py) as a proton trap and [Mo2(O2CBut)2(CH3CN)4]21[BF4 2]2 as a carboxylate anion trap. In these studies of reaction (2), we observed ligand scrambling to [Mo2(O2CBut)2(CH3CN)4]21[BF4 2]2. [These reactions have to be carried out in CD3CN because the Mo2(O2CBut)2 21 cationic complex is not soluble in benzene-d6 or CD2Cl2.] Given the lability of the coordinated CH3CN ligands in [Mo2(O2CBut)2- (CH3CN)4]21 it is possible that the Mo2 41 center is suYciently electrophilic to remove a carboxylate ligand from a Mo2- (O2CR)4 complex.In fact a reaction occurs between [Mo2- (O2CBut)2(CH3CN)4]21[BF4 2]2 and Mo2(O2CCF3)4 virtually instantaneously even in the presence of 2,6-di-tert-butylpyridine. Likewise, the fully solvated cationic complex [Mo2- (CH3CN)10]41[BF4 2]4 9 is capable of removing a pivalate ligand from Mo2(O2CBut)4 in the presence of added 2,6-di-tertbutylpyridine. We have, however, noted that the pivalate exchange reaction between [Mo2(O2CBut)2(CD3CN)4]21[BF4 2]2 Fig. 1 1H NMR spectra of the carboxylate scrambling reaction involving Mo2(O2CtBu)4 and Mo2(O2CCF3)4 recorded in benzene-d6, 500 MHz at 22 8C showing the disappearance of Mo2(O2CtBu)4 (purple) and the concomitant formation of Mo2(O2CtBu)3(O2CCF3) 1 (blue) and Mo2(O2CCF3)3(O2CtBu) 3 (black) followed by formation of Mo2(O2CtBu)2(O2CCF3)2 2, which occurs in both trans (green) and cis (red) isomers.The formation of trans-2 is seen to be kinetically favored. and Mo2(O2CBut)4 in CD3CN to give [Mo2(O2CBut)3(CD3- CN)2]1[BF4 2] is greatly suppressed by added 2,6-But 2-py. This suggested that a suitable protic trap such as 2,6-But 2-py in the presence of a more eYcient carboxylate anion scavenger might completely suppress reaction (2).In this context we turned to the use of the [Mo2(PhNCHNPh)2(CH3CN)4]21[BF4 2]2 salt which we have found to have kinetically inert formamidinato ligands with respect to ligand exchange.10 Mo2(O2CBut)4 and Mo2(O2CCF3)4 in CD3CN showed no evidence for ligand exchange, reaction (2), after 12 h in the presence of 2,6-But 2-py and [Mo2(PhNCHNPh)2(CH3CN)4]21- [BF4 2]. The ability of [Mo2(PhNCHNPh)2(CH3CN)4]21[BF4 2]2 to function as a carboxylate trap was confirmed independently by reacting the dicationic complex with 2 equivalents of Na(O2CCBut) in CD3CN.From this we can conclude that carboxylate scrambling in reaction (2) is suppressed in the presence of appropriate protic and carboxylate traps. It is therefore unnecessary to invoke [M2]2 activated complexes in order to achieve carboxylate exchange and it would seem that carboxylate supported dimers of “dimers” and higher oligomers may be chemically persistent in the presence of appropriate traps such as 2,6-But 2-py and [Mo2(PhNCHNPh)2(CH3CN)4]21[BF4 2] which only need to be present in low concentrations.Experimental All manipulations were carried out by using standard Schlenkline and glove-box techniques under an atmosphere of argon or nitrogen. The deuteriated solvents benzene-d6, CD2Cl2, CD3CN, THF-d8 and pyridine-d5 were degassed with argon and stored over molecular sieves (3 Å or 4 Å) prior to use. The 1H and 19F NMR spectra were recorded on a 300 Varian Gemini- 2000 NMR spectrometer at 300 and 288.2 MHz, respectively. Higher resolution 1H NMR spectra were recorded on a 500 Varian Inova NMR spectrometer. 1H NMR spectra were referenced to residual protio impurities of the deuteriated solvents. 19F NMR spectra were referenced externally relative to CF3- CO2H. The dimolybdenum complexes Mo2(O2CBut)4,11 Mo2(O2CCF3)4,12 [Mo2(O2CBut)2(CH3CN)4 21][BF4 2]2 1 and [Mo2(PhNCHNPh)2(CH3CN)4]21[BF4 2]2 10 were synthesized according to published procedures.Proton sponge“, Cs2CO3 and BaCO3 were purchased from Aldrich and were dried under vacuum for 12 h. The 2,6-di-tert-butylpyridine was purchased from Aldrich, freeze–pump–thaw degassed and stored under Ar over 4 Å sieves prior to use. The samples were prepared using Young“ NMR tubes in a helium glove-box. The Mo2(O2CR)4 complexes were weighed using an analytical balance accurate to 0.1 mg. A 1 : 1 solution of the Mo2(O2CR)4 complexes was prepared by weighing out equimolar amounts of Mo2(O2CBut)4 and Mo2(O2CCF3)4 and adding the deuteriated solvent. 1H NMR (d, C6D6): Mo2(O2CBut)4 1.42 (s); 1 1.37 (s), 1.33 (s) (2H: 1H); trans-2 1.34 (s); cis-2 1.28 (s); 3 1.24 (s). (d, CD3CN): Mo2(O2CBut)4 1.38 (s); 1 1.39 (s), 1.42 (s) (2H : 1H); cis-2 1.43 (s); trans-2 1.47 (s); 3 1.47 (s). (d, THF-d8): Mo2- (O2CBut)4 1.41 (s); 1 1.41 (s), 1.45 (s) (2H : 1H); cis-2 1.46 (s); trans-2 1.50 (s); 3 1.50 (s). Acknowledgements We thank the National Science Foundation for financial support and Drs Marty Pagel and Uli Werner for assistance in the production of Fig. 1. References 1 R. H. Cayton, M. H. Chisholm, J. C. HuVman and E. B. Lubkovsky, J. Am. Chem. Soc., 1991, 113, 8709. 2 H. Chen and F. A. Cotton, Polyhedron, 1995, 14, 2221.J. Chem. Soc., Dalton Trans., 1999, 1205–1207 1207 3 J. M. Casas, R. H. Cayton and M. H. Chisholm, Inorg. Chem., 1991, 30, 358. 4 M. H. Chisholm, D. L. Clark and M. J. Hampden-Smith, J. Am. Chem. Soc., 1989, 111, 574. 5 T. R. Ryan and R. E. McCarley, Inorg. Chem., 1982, 21, 2072. 6 R. H. Cayton, S. T. Chacon, M. H. Chisholm and K. Folting, Polyhedron, 1993, 12, 415. 7 F. A. Cotton and R. A. Walton, in Multiple Bonds Between Metal Atoms, Oxford University Press, New York, 2nd edn., 1993. 8 The lability of Mo2(O2CCF3)4 toward donor ligands in the formation of Mo2–h1-O2CCF3 complexes has been previously documented: T. R. Webb and T. Y. Dong, Inorg. Chem., 1982, 21, 114; G. S. Girolami, V. V. Mainz and R. A. Andersen, Inorg Chem., 1980, 19, 805. 9 F. A. Cotton and K. J. Wiesenger, Inorg. Synth., 1992, 29, 134. 10 M. H. Chisholm, K. Folting, J. C. HuVman, A. M. Macintosh and S. I. Iyer, preceding paper. 11 A. G. Brignole and F. A. Cotton, Inorg. Synth., 1972, 13, 81. 12 F. A. Cotton and J. G. Norman, Jr., J. Coord. Chem., 1971, 1, 161. Paper 9/00388F

ISSN:1477-9226    

DOI:10.1039/a900388f

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 1209–1212 1209 Reduction of the anti-cancer drug analogue cis,trans,cis- [PtCl2(OCOCH3)2(NH3)2] by L-cysteine and L-methionine and its crystal structure † Lie Chen, Peng Foo Lee, John D. Ranford,* Jagadese J. Vittal and Siew Ying Wong Department of Chemistry, National University of Singapore, Lower Kent Ridge Road, 119270 Singapore. E-mail: chmjdr@nus.edu.sg Received 15th January 1999, Accepted 3rd March 1999 The complex cis,trans,cis-[PtCl2(OCOCH3)2(NH3)2] 1 has been synthesized as a simplified and more soluble model of the anticancer drug cis,trans,cis-[PtCl2(OCOCH3)2(NH3)(C6H11NH2)] (JM216).The crystal structure of 1 shows an octahedral co-ordination sphere around the PtIV with strong intramolecular and weak intermolecular hydrogen bonding. The kinetics of reduction of 1 by the divalent sulfur amino acids L-cysteine and L-methionine has been determined over a range of pH values by multinuclear NMR.The reduction is strongly pH dependent, being related to the protonation state of the amino acid and the basicity of the sulfur. Reduction rates are dramatically slower than for previous models of platinum(IV) drug systems. Cisplatin, cis-diamminedichloroplatinum(II), is an anticancer drug widely used to treat a variety of tumours, especially those of the testes, ovaries, head, and neck.1 Its intensive use has been restricted by several drawbacks such as sidee Vects, the development of resistance during therapy and the need for intravenous administration.2 This has therefore been the impetus for a world-wide search for new platinumcontaining drugs.The eVorts have resulted in the development of “second-generation” cisplatin analogues, such as carboplatin [diammine(cyclobutane-1,1-dicarboxylato)platinum(II)], and “third generation” analogues, e.g. cis,trans,cis-[PtCl2- (OCOCH3)2(NH3)(C6H11NH2)] [ 2, JM216; bis(acetato)amminedichloro( cyclohexylamine)platinum(IV)].3 Platinum(IV) complexes undergo ligand substitution reactions much more slowly than their platinum(II) analogues, and are therefore usually seen as prodrugs for platinum(II) compounds.Currently the compounds tetraplatin [tetrachloro( 1,2-diaminocyclohexane)platinum(IV)], iproplatin [cis, trans,cis-dichlorodihydroxobis(isopropylamine)platinum(IV)] and JM216 are undergoing clinical trials 4 with the latter being the first orally administered platinum complex.5 The platinum(IV) species is supposedly activated by reduction to its platinum(II) analogue before hydrolysis and substitution reactions can occur.6 Thiol and thioether containing biomolecules and ascorbic acid have been proposed as the major cellular components responsible for that reduction.7 Therefore L-cysteine (L-H2cys) and L-methionine (L-Hmet), either as free amino acids or as components of intracellular peptides and proteins, are considered to be potential reductants for PtIV.Studies on the platinum(IV) complexes trans-[Pt(CN)4X2]22 (where X is Cl2 or Br2) as stable models for anticancer drugs with L-Hmet8 and four thiols 7 including L-H2cys have been conducted. Stop-flow techniques had to be employed because of the rapid reactions. The reactions are pH dependent and for the thiols were limited to pH values below 5 because of this. The proposed mechanism involves association of the reductant through the chloro ligands in a transition state and subsequent electron transfer.† Supplementary data available: 15N NMR spectra of 1 with L-cysteine. Available from BLDSC (No. SUP 57512, 2 pp.). See instructions for Authors, 1999, Issue 1 (http://www.rsc.org/dalton). In the present study, we use cis,trans,cis-[PtCl2(OCOCH3)2- (NH3)2] 1 as a simplified model of 2, as its aqueous solubility is significantly higher than that of 2, its synthesis more straightforward, especially for incorporation of 15N labels, and it eliminates possible isomerism in reaction products due to the inequivalent amines in 2.We report here the crystal structure and characterization of 1 and kinetic studies of its reduction by the sulfur containing amino acids L-H2cys and L-Hmet using multinuclear NMR. Experimental Preparations Cisplatin was synthesized following the method of Dhara 9 and cis,trans,cis-[PtCl2(OH)2(NH3)2] 3 was prepared by oxidation of cisplatin with H2O2.10 15N-Enriched complex 1 was synthesized using cis-[PtCl2(15NH3)2] prepared from 15NH4Cl.cis,trans,cis-[PtCl2(OCOCH3)2(NH3)2] 1. This was synthesized by carboxylation of its hydroxide analogue 3 with acetic anhydride. A solution of 3 (0.830 g, 2.49 mmol) was stirred in acetic anhydride (20 cm3, 0.212 mol) for 2 h. The product was precipitated with hexane and the solution cooled at 0 8C for 24 h. The solid was filtered oV and washed with ethanol, diethyl ether and dried in vacuo. Yield: 0.60 g (60%) (Found: C, 11.64; H, 2.94; N, 6.65.Calc. for C4H12Cl2N2O4Pt: C, 11.48; H, 2.87; N, 6.70%). 1H NMR (D2O): d 2.10 (s). NMR spectroscopy The NMR spectra were recorded on Bruker ACF300 (1H 300 MHz) and AMX500 (15N 50.70 MHz) spectrometers using 5 mm NMR tubes. The chemical shift references were as follows: (internal) 1H, 4,4-dimethyl-4-silapentane-1-sulfonate, DSS; 15N, 1.5 M 15NH4Cl in 1 M HCl containing 10% D2O in a capillary. Samples for 1H NMR were made up in deuteriated phosphate1210 J. Chem.Soc., Dalton Trans., 1999, 1209–1212 buVer (0.1 M, pH* 7.0) which was prepared by freeze-drying solutions in water and redissolving in D2O, with pH* readjustment as necessary; pH* refers to the pH meter reading in D2O solution. Samples for 15N NMR were prepared in 90% water– 10% D2O phosphate buVer. Solubility determination Samples were suspended in water then ultrasonicated for 2 min. The excess of solid was filtered oV, the volume of supernatant was measured prior to lyophilization, then the solid was weighed.Cyclic voltammetry The cyclic voltammetry was performed in a BAS 100B (Bioanalytical Systems Inc.) cell. The electrodes were: working, glassy carbon; counter, platinum foil; reference Ag–AgCl–Cl. All solutions were degassed by bubbling with N2 before measurements. Crystallography Crystal data. C4H14Cl2N2O5Pt 1, M = 436.16, monoclinic, space group C2/c, a 14.9973(3), b 8.57220(10), c 11.1352(2) Å, b 126.7690(10)8, U = 1146.74(3) Å3, l = 0.71073 Å, Z = 4, Dc = 2.526 Mg m23, m(Mo-Ka) 12.701 mm21, 293 K, crystal dimensions 0.38 × 0.32 × 0.27 mm.Data collection and structural analysis. Data were collected on a Siemens SMART CCD diVractometer with the crystal sealed in a glass capillary tube. Preliminary cell constants were obtained from 45 frames of data (width 0.38 in w). Final cell parameters were obtained by global refinements of reflections obtained from integration of all the frame data. A total of 3597 reflections were collected in the q range 2.92–29.078 (220 £ h £ 19, 211 £ k £ 5, 214 £ l £ 13), 1421 independent reflections (Rint 0.0547) with a frame width of 0.38 in w and a counting time of 20 s per frame. The collected frames were integrated using the preliminary cell-orientation matrix.The software SMART11 was used for collecting frames of data, indexing reflections and determination of lattice parameters, SAINT for integration of reflection intensity and scaling, SADABS12 for absorption correction and SHELXTL13 for space group and structure determination, refinements (fullmatrix least squares on F2), graphics and structure reporting.All the non-hydrogen atoms in the molecules were refined anisotropically. The hydrogen atoms were placed in the ideal positions using riding models. Maximum and minimum transmission 0.1159 and 0.0305 respectively with goodness of fit on F2 1.114 and extinction coeYcient 0.0062(5). Final R indices [I > 2s(I)] were R1 0.0386, wR2 0.0965 for the 70 parameters refined.The corresponding values for all data after merging were R1 0.0391 and wR2 0.0967 respectively; F(000) 816. The largest diVerence peak 2.226 and hole 22.765 e Å23 were associated with Pt. CCDC reference number 186/1374. See http://www.rsc.org/suppdata/dt/1999/1209/ for crystallographic files in .cif format. Results and discussion The mechanism of action of platinum(IV) anticancer drugs appears first to involve reduction to the kinetically more labile platinum(II) analogue.Reduction of the antitumour agent cis,trans,cis-[PtCl2(OCOCH3)2(NH3)(C6H11NH2)] 2 is expected to give the cisplatin analogue cis-[PtCl2(NH3)(C6H11NH2)]. As the aqueous solubility of the cyclohexylamine compounds is lower than that of the ammine derivatives and, as the latter are easier to prepare (especially as 15N labelled materials for multinuclear NMR studies), we have synthesized cis,trans,cis- [PtCl2(OCOCH3)2(NH3)2] 1 as a simplified version of 2.This has allowed us to study the reduction kinetics of 1 with the divalent sulfur amino acids L-H2cys and L-Hmet as well as characterize the solid state structure of the complex. Incorporation of 15NH3 permits multinuclear NMR investigation of reactions. The solubilities of 1 (18.3 mg cm23) and 2 (0.4 mg cm23) were determined, demonstrating the significant improvement in solubility on replacing the hydrophobic cyclohexylamine with ammine. Crystal structure of cis,trans,cis-[PtCl2(OCOCH3)2(NH3)2] 1 A thermal ellipsoid diagram for complex 1 is shown in Fig. 1. Selected bond distance and angle data are given in Table 1. The co-ordination geometry about the PtIV is octahedral. The two Cl atoms [Pt–Cl1 2.318(2) Å] and two ammine groups [Pt–N1 2.049(6) Å] are in a square-planar arrangement around the Pt atom similar to that of cisplatin.14 The two acetato groups [Pt– O1 2.030(6) Å] are axial to this plane [O1–Pt–O1A 176.7(3)8].Intramolecular hydrogen bonds between the ammine and nonco- ordinated carboxyl oxygen (N1 ? ? ? O2 2.723 Å) and solvent water (N1 ? ? ? O1S 2.894 Å) are observed. It is likely that in solution a solvent molecule is associated with the two ammine groups as observed in the solid state structure. The remaining ammine hydrogen has a very weak intermolecular interaction to a co-ordinated carboxylato O [N1 ? ? ? O1 3.25 Å (N1: x, 1 2 y, z 2 0.5)] and the water hydrogens have weak hydrogenbonding interactions to the chloro moieties in an adjacent molecule [O1SA ? ? ? Cl1 3.44 Å (O1SA: x, y 2 1, z)].Excluding the interactions to water, the hydrogen-bonding is very similar to that seen for cis,trans-[Pt(en)Cl2(OCOCH3)2] 15 (where en is 1,2-diaminoethane) and [Pt(en)(OCOCH3)2(cbdca)] (where H2cbdca is cyclobutane-1,1-dicarboxylic acid).16 Reactions of complex 1 L-Cysteine. A time-course series of 1H NMR spectra of a solution of excess of L-H2cys and complex 1 shows a decrease in the acetato singlet (d 2.11) and L-H2cys resonances with concomitant increases in free acetate (d 1.90) and cystine.The Fig. 1 Molecular structure of complex 1 showing intra- and intermolecular hydrogen bonds. Table 1 Selected bond lengths (Å) and angles (8) for cis,trans,cis- [PtCl2(OCOCH3)2(NH3)2] 1 Pt–O1 Pt–N1 Pt–Cl1 O1–Pt–O1A O1–Pt–N1 O1–Pt–N1A O1–Pt–Cl1 O1–Pt–Cl1A 2.030(6) 2.049(6) 2.318(2) 176.7(3) 97.6(2) 84.8(2) 85.3(2) 92.3(2) C1–O1 C1–O2 C1–C2 N1–Pt–Cl1 N1–Pt–Cl1A N1–Pt–N1A Cl1–Pt–Cl1A 1.303(11) 1.213(11) 1.502(14) 177.1(2) 89.1(2) 90.7(4) 91.23(10)J.Chem. Soc., Dalton Trans., 1999, 1209–1212 1211 formation of free acetate shows the PtIV has been reduced, the reductant being L-H2cys as seen by the formation of cystine [d 4.11(m), aH; 3.40(m), bH; 3.16(m), bH]. The 15N NMR spectra (SUP 57512) of the reaction show a decrease in the peak from 1 (d 241.74) and increase in a single resonance attributable to cisplatin (d 268.62).17 Accordingly, the reaction is formulated as in eqn. (1).Under the conditions of pseudo-first cis,trans,cis-[PtCl2(OCOCH3)2(NH3)2] 1 2HSCH2CH(NH3 1)CO2 2 æÆ cis-[PtCl2(NH3)2] 1 [–SCH2CH(NH3 1)CO2 2]2 1 2CH3CO2 2 1 2H1 (1) order kinetics, excess of L-H2cys is present and this then reacts further with the cisplatin produced. The 15N NMR spectra of the reaction of 15N-1 with 15N-L-H2cys (1 :5) were recorded. Initial resonances for free L-H2cys and 1 present decreased with time and a small peak for cisplatin was detectable. After 24 h no cisplatin or platinum products were detectable in solution. Solid cystine had precipitated from solution, with the remaining L-H2cys having formed oligomeric species, as described previously.18 The resonance for 15NH4 1 formed from released 15NH3 overlapped with that of the reference.L-Methionine. Reaction of complex 1 with L-Hmet (1: 2) also produced free acetate and cisplatin as seen in the 1H and 15N NMR spectra respectively.The methyl singlet of L-Hmet was used as a convenient handle on the reaction progress as well as product identification. As the reaction proceeded resonances for free L-Hmet decreased [d 2.12(s), CH3S] with a concomitant increase in a new set of signals [d 2.74 (s)]. Separate spectra of L-Hmet and L-methionine S-oxide were recorded which showed that methionine was consumed and the new sharp singlet observed is due to L-methionine S-oxide. The reaction is formulated as in eqn. (2).cis,trans,cis-[PtCl2(OCOCH3)2(NH3)2] 1 CH3SCH2CH2CH(NH3 1)CO2 2 1 H2O æÆ cis-[PtCl2(NH3)2] 1 CH3S(O)CH2CH2CH(NH3 1)CO2 2 1 2CH3CO2 2 1 2H1 (2) Incubation of complex 1 with excess of L-Hmet (1 : 4) for a week showed from 15N NMR spectra that all cisplatin had reacted and new resonances were present. These have been assigned to cis-[Pt(L-met-S,N)2] (d 220.06, 220.24, 221.10) and trans-[Pt(L-met-S,N)2] (d 238.31, 238.60, 238.81).17 A weak singlet at d 18.64 is near to the resonance for free L-Hmet but could also be due to a monodentate S-only bonded species such as cis-[Pt(L-met-S,N)(L-met-S)2]2.The resonances from chelated L-Hmet typically appear in the same range as for cis- [Pt(L-met-S,N)2] and would be obscured by these. Kinetics Reduction of complex 1 was studied under pseudo-first-order conditions with L-H2cys and L-Hmet in a 20-fold excess. 1H NMR was used to monitor the decrease of the acetato singlet at d 2.10 or the increase of free acetate at d 1.90.A plot of ln CA or ln(CA 2 CZ) against t was linear (where CA is the peak intensity for bound and CZ free acetate) and the first-order rate constants can be obtained from the slope. The possibility of initial hydrolysis to give an activated aqua intermediate was ruled out by measuring the hydrolysis rate for 1. After 20 d in 0.1 M deuteriated pH* 7.0 phosphate buVer no new peaks were observable by 1H NMR. The kinetic data were fitted assuming a mechanism whereby all protonation states of the reductant react in parallel, as suggested by Elding and co-workers 7,8 and depicted in Scheme 1.The observed rate constants, kobs, are proportional to the excess concentration of reductant. Overall second-order kinetics are observed according to eqn. (3), where k9 denotes the second- 2d[PtCl2(OCOCH3)2(NH3)2]/dt = d[PtCl2(NH3)2]/dt = d[CH3CO2 2]/2dt = k9[reductant][PtCl2(OCOCH3)2(NH3)2] (3) = kobs [PtCl2(OCOCH3)2(NH3)2] where kobs = k9[reductant] (4) order rate constants for diVerent reductants.Data were fitted‡ using all kn and Kan as variables and k9 are given in Table 2 along with the activation parameters, DH‡ and DS‡. For both L-H2cys and L-Hmet the rates increase as the pH increases, consistent with the deprotonated species being better reductants, Fig. 2. The fitted values for pKan derived from these plots agree well with the literature values (in parentheses): 19 L-H2cys, pKa1 1.8 (1.9), pKa2 7.6 (8.1), pKa3 10.5 (10.1); L-Hmet, pKa1 2.5 (2.22), pKa2 8.5 (9.02).For L-H2cys k9 is ca. 2000 times bigger than for L-Hmet, and may be attributable to the greater basicity of the thiol as compared to the thioether.8 However, values of k9 are orders of magnitude lower for the Scheme 1 Table 2 Rate constants as a function of temperature and activation parameters for the reduction of cis,cis,trans-[PtCl2(OCOCH3)2(NH3)2] 1 by L-cysteine and L-methionine at pH* 7.0 k9 (290 K) k9 (295 K) k9 (300 K) k9 (305 K) k9 (310 K) k1 b k2 k3 k4 DH‡ c DS‡ d L-Cysteine a (2.9 ± 0.4) × 1023 (5.5 ± 0.5) × 1023 (7.0 ± 0.4) × 1023 (12.5 ± 0.6) × 1023 — (3.3 ± 4.0) × 1024 (2.9 ± 1.8) × 1023 (2.2 ± 0.4) × 1022 (6.2 ± 0.3) × 1022 65.7 ± 3.0 266.5 ± 2.5 L-Methionine — (2.1 ± 0.2) × 1026 (3.9 ± 0.1) × 1026 (6.0 ± 0.3) × 1026 (11.1 ± 0.4) × 1026 (1.7 ± 0.5) × 1026 (3.4 ± 0.3) × 1026 (1.26 ± 0.03) × 1025 — 80.2 ± 2.7 283.2 ± 4.4 a k9/dm3 mol21 s21.b kn/dm3 mol21 s21.c kJ mol21. d J K21 mol21. ‡ Cysteine: k9 = k1[H1]3 1 k2Ka1[H1]2 1 k3Ka1Ka2[H1] 1 k4Ka1Ka2Ka3 [H1]3 1 Ka1[H1]2 1 Ka1Ka2[H1] 1 Ka1Ka2Ka3 Methionine: k9 = k1[H1]2 1 k2Ka1[H1] 1 k3Ka1Ka2 [H1]2 1 Ka1[H1] 1 Ka1Ka21212 J. Chem. Soc., Dalton Trans., 1999, 1209–1212 antitumour analogue 1 than for trans-[Pt(CN)4X2]22 (where X is Cl2 or Br2) 7,8 which was employed as a stable antitumour model complex. This has allowed us to determine all values of k9, which was not possible for trans-[Pt(CN)4X2]22 at higher pH values using stop-flow techniques, due to rapid reaction.The dramatically greater reactivity of trans-[Pt(CN)4X2]22 over 1 is a direct result of the in-plane cyano ligands which stabilize the platinum(II) oxidation state. As a measure of this we investigated the redox behaviour of complex 1 and trans- [Pt(CN)4X2]22. As is typical for platinum(IV) species, a reduction step is seen but the corresponding oxidation is not observed as reduction results in loss of the axial ligands.The cathodic potential for 1 of 2689 mV is more negative than for [Pt(en)Cl2{OC(O)R}2] (where R is alkyl) 15 which ranged from 2493 to 2546 mV and may be a reflection of the greater substitutional stability of the en chelate. All these complexes have significantly greater values than does trans-[Pt(CN)4X2]22 of 2399 mV. Our system is a close analogue to the drug JM216 (2) and aVords a comparison of the kinetics expected for this.In biological media the major S-containing reductants are glutathione (g-glutamylcysteinylglycine), L-H2cys and L-Hmet. Given that glutathione and L-H2cys have similar kn values for equivalent deprotonated species, the significantly greater rate of reduction of 1 by L-H2cys and the typical concentrations of divalent sulfur compounds present in blood plasma and cells,20 it is predicted that the predominant reductant will be glutathione > L-H2cys > L-Hmet. Platinum(IV) complexes prefer a low-spin, substitution-inert, d6 octahedral geometry, so initial complex formation between reductant and PtIV is highly unlikely.The reduction of PtIV involves a two-electron transfer process and halide-mediated reductive-elimination reactions of platinum(IV)–halogen com- Fig. 2 Plots of k9 vs. [H1] for (a) L-cysteine and (b) L-methionine where the solid lines are those fitted using the equations in the text. plexes involving various reductants have been suggested to take place via an attack by reductant on co-ordinated halide.21 By analogy, the reductive elimination reaction of 1 may be interpreted in terms of oxygen-bridge electron transfer, with the transition state formulated as follows (where R = H, for L-H2cys, CH3 for L-Hmet).Acknowledgements Support for this work by the National University of Singapore (Grant RP 950651) and Miss Geok Kheng Tan for assistance with the crystal structure is greatly appreciated. References 1 K.M. Comess and S. J. Lippard, Molecular Aspects of Anticancer Drug-DNA Interactions, eds. S. Neidle and M. Waring, MacMillan, London, 1993, vol. 1, p. 134. 2 R. B. Weiss and M. C. Christian, Drugs, 1993, 46, 360. 3 L. R. Kelland, B. A. Murrer, G. Abel, C. M. Giandomenico, P. Mistry and K. R. Harrap, Cancer Res., 1992, 52, 822. 4 R. M. Roat and J. Reedijk, J. Inorg. Biochem., 1993, 52, 263. 5 F. I. Raynaud, D. E. Odell and L. R. Kelland, Br. J. Cancer, 1996, 74, 380. 6 G.R. Gibbons, S. D. Wyrick and S. G. Chaney, Cancer Res., 1989, 49, 1402; M. Laverick, A. H. W. Nias, P. J. Sadler and I. M. Ismail, Br. J. Cancer, 1981, 43, 732; J. L. Van der Veer, A. R. Peters and J. Reedijk, J. Inorg. Biochem., 1986, 26, 137. 7 T. S. Shi, J. Berglund and L. I. Elding, Inorg. Chem., 1996, 35, 3498. 8 T. S. Shi, J. Berglund and L. I. Elding, J. Chem. Soc., Dalton Trans., 1997, 2073. 9 S. Dhara, Indian J. Chem., 1970, 8, 193. 10 A. V. Babaeva, Dokl. Akad. Nauk SSSR, 1939, 23, 653. 11 SMART & SAINT Software Reference Manuals, Version 4.0, Siemens Energy & Automation, Inc., Analytical Instrumentation, Madison, WI, 1996. 12 G. M. Sheldrick, SADABS, software for empirical absorption correction, University of Göttingen, 1996. 13 SHELXTL Reference Manual, Version 5.03, Siemens Energy & Automation, Inc., Analytical Instrumentation, Madison, WI, 1996. 14 G. H. W. Milburn and M. R. Truter, J. Chem. Soc. A, 1966, 1609. 15 L. T. Ellis, H. M. Er and T. W. Hambley, Aust. J. Chem., 1995, 48, 793. 16 Y. Deng and A. R. Khokar, Inorg. Chim. Acta, 1993, 204, 35. 17 R. E. Norman, J. D. Ranford and P. J. Sadler, Inorg. Chem., 1992, 31, 877. 18 R. N. Bose, S. K. Ghosh and S. Moghaddas, J. Inorg. Biochem., 1997, 65, 199. 19 R. M. Smith and A. E. Martell, Critical Stability Constants, Plenum, New York, 1989, vol. 6, suppl. 2, pp. 20–24. 20 H. C Potgieter, J. B. Ubbink, S. Bissbort, M. J. Bester, J. H. Spies and W. J. H. Vermaak, Anal. Biochem., 1997, 248, 86; A. Pastore, R. Massoud, C. Motti, A. Lo Russo, G. Fucci, C. Cortese and G. Federici, Clin. Chem., 1998, 44, 825. 21 J. Berglund, R. Voigt, S. Fronaeus and L. I. Elding, Inorg. Chem., 1994, 33, 3346; P. Chandayot and Y.-T. Fanchiang, Inorg. Chem., 1985, 24, 3532, 3535. Paper 9/00441F

ISSN:1477-9226    

DOI:10.1039/a900441f

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 1221–1226 1221 The eVect of counter cations on second-order asymmetric transformations in fac-ƒ- and À-tris(R-cysteinato-N,S)cobaltate(III) complexes and the kinetics of mutarotation Masakazu Kita *a and Kazuaki Yamanari b a Chemistry Department, Naruto University of Education, Takashima, Naruto 772-8502, Japan. E-mail: kitam@naruto-u.ac.jp b Department of Chemistry, Graduate School of Science, Osaka University, Toyonaka, Osaka 560-0043, Japan Received 19th October 1998, Accepted 18th February 1999 A rapid reversible inversion (epimerization) between the D-(R,R,R) and L-(R,R,R) diastereomers for fac-D- and L-tris(R-cysteinato-N,S)cobaltate(III) occurs at room temperature in an aqueous solution of each salt.The half-lives for the mutarotation were measured (k = kL 1 kD): 80 and 6 minutes at 20 and 40 8C, respectively. The equilibrium constants, K = [L-(R,R,R)]/[D-(R,R,R)], are 1.2 at 20 8C and 1.0 at 40 8C, which means that the cobalt inversion rate constant kL nearly equals kD.Eight kinds of metal salts, M3[Co(R-cys-N,S)3] [M = Li1, Na1, K1, Rb1, Cs1; M1/2 = Ca21, Sr21, Ba21], were prepared. It was found that all these salts show a second-order asymmetric transformation under a solid-solution equilibrium at room temperature of the kind first demonstrated for the Ba21 salt by Arnold and Jackson. The phenomenon led to the D-(R,R,R) diastereomer for the salts of Li1, Na1, K1, Rb1, Cs1, Ca21 and Sr21, but to the L-(R,R,R) diastereomer for Ba21.These results could be clearly explained by solubility measurements of each diastereomeric salt pair. Asymmetric transformation between diastereomers causes the change of a diastereomeric ratio from unity. The equilibrium shift in homogeneous solution is known as a first-order asymmetric transformation and the subsequent selective crystallization of one of the diastereomers under conditions of relatively rapid isomer equilibration in solution is called a second-order asymmetric transformation.1 The extent of the former phenomenon can depend on the solvent and has been reported in the neutral chromium(III) complexes such as (2)cyclo-O,O9-(1R, 2R-dimethylethylene)dithiophosphate(12) 2 and O,O9-bis{(1)- S-2-methylbutyl}dithiophosphate(12).3 The latter one has been applied for the asymmetric synthesis of a L-tris(ethylenediamine) cobalt(III)(31) complex 4 and the optical resolution of a tris(oxalato)chromate(III)(32) complex.5 The synthesis of the green tris(R-cysteinato)cobaltate(III) complex [Co(R-cys-N,S)3]32 was first reported by Schubert in 1933.6 From many studies concerning this complex, it has been claimed that only the fac isomer seems to be formed and that the formation of fac-[Co(R-cys-N,S)3]32 is stereospecific,7 and that the D-(R,R,R) diastereomer is formed exclusively over the L-(R,R,R) diastereomer.8 Recently, Arnold and Jackson have reported some important findings in this fac-[Co(R-cys- N,S)3]32 system: two diastereomers, the D-(R,R,R) and L-(R, R,R) diastereomers, were confirmed to exist in solution by 13C NMR spectroscopy and the previously unknown L-(R,R,R) diastereomer was successfully isolated as the barium salt through a novel second-order asymmetric transformation.9 They also isolated a metastable Ba21 double salt containing a 1 : 1 ratio of the two diastereomers through crystallisation at ca. 60 8C where this salt is the least soluble amongst the three alternatives (L, D or LD).Our work has been concerned with cobalt(III)-thiolato complexes containing (R)-cysteinato and their oxidation derivatives, which show unique behaviour, i.e., stereospecificity,10 trans influence,11 linkage isomerism,12,13 and sulfur chirality induced through metal coordination.10–13 Therefore, we became interested in the above second-order asymmetric transformation phenomenon for fac-[Co(R-cys-N,S)3]32 and started a preliminary investigation. We discovered new results, some of which relate to the diVering views of the two groups.14,15 Herein we report the following: eight kinds of metal salts in fac-Dand L-tris(R-cysteinato-N,S)cobaltate(III), M3[Co(R-cys-N,S)3] [M = Li1, Na1, K1, Rb1, Cs1 and M1/2 = Ca21, Sr21, Ba21] were prepared and the solution kinetics of mutarotation were investigated in these salts.Experimental ƒ-Li3[Co(R-cys-N,S)3] and ƒ-M3[Co(R-cys-N,S)3] (M 5 Na1, K1, Rb1 and Cs1; M1/2 = Ca21, Sr21 and Ba21) An aqueous solution (20 cm3) of 1.64 g (9.5 mmol) of (R)- H2cys?HCl?H2O and 1.20 g (28.5 mmol) of LiOH?H2O was added to an aqueous solution (50 cm3) of 0.50 g (1.9 mmol) of [Co(NH3)6]Cl3.The mixture was stirred for 2 h at 70 8C until no further evolution of ammonia could be detected. The resulting deep green solution showed a CD (circular dichroism) spectrum with the ca. 3 : 2 mixture of the D-(R,R,R) and L-(R,R,R) diastereomers. Addition of 20 cm3 of ethanol to the solution gave a gray-blue precipitate and then the suspension was stirred at 20 8C for 12 h to yield blue needles of the pure D(R,R,R) diastereomer.It should be noted that the immediate addition of an excess of ethanol (100 cm3) resulted in the precipitation of the ca. 3 : 1 mixture of the D-(R,R,R) and L-(R,R,R) diastereomers. Yield = 0.90 g (82%). Found: C, 18.74; H, 5.01; N, 7.40. Calc. for D-Li3[Co(R-cys)3]?7H2O = C9H29N3O13S3- CoLi3: C, 19.19; H, 5.19; N, 7.46%.The other alkali-metal salts of the pure D-diastereomer were obtained by using the appropriate alkali-metal hydroxide instead of LiOH. The yields were ca. 80–90%. Found for the sodium salt: C, 15.55; H, 5.10; N, 6.02. Calc. for D-Na3[Co- (R-cys)3]?11H2O = C9H37N3O17S3CoNa3: C, 15.82; H, 5.46; N, 6.15%. Found for the potassium salt: C, 16.62; H, 4.45; N, 6.46. Calc. for D-K3[Co(R-cys)3]?6.5H2O = C9H28N3O12.5S3CoK3: C,1222 J. Chem. Soc., Dalton Trans., 1999, 1221–1226 16.61; H, 4.34; N, 6.46%.Found for the rubidium salt: C, 12.88; H, 3.62; N, 5.00. Calc. for D-Rb3[Co(R-cys)3]?9H2O = C9H33- N3O15S3CoRb3: C, 12.95; H, 3.98; N, 5.03%. Found for the caesium salt: C, 11.19; H, 2.92; N, 4.37. Calc. for D-Cs3- [Co(R-cys)3]?8H2O = C9H31N3O14S3CoCs3: C, 11.27; H, 3.26; N, 4.38%. The Ca21 and Sr21 salts of the pure D-(R,R,R) diastereomer were obtained by a similar asymmetric transformation, but the experiments were carried out under Ar to prevent the formation of insoluble MCO3. The pure D-(R,R,R) diastereomer of the lithium salt could be converted to Ca21, Sr21 or Ba21 salt by rapid addition of concentrated aqueous CaCl2, SrCl2 or BaCl2.Found for the calcium salt: C, 16.77; H, 4.71; N, 6.45. Calc. for D-Ca1.5[Co(R-cys)3]?9H2O = C9H33N3O15S3CoCa1.5: C, 16.93; H, 5.21; N, 6.58%. Found for the strontium salt: C, 15.24; H, 3.47; N, 5.63. Calc. for D-Sr1.5[Co(R-cys)3]?9H2O = C9H33N3- O15S3CoSr1.5: C, 15.23; H, 4.69; N, 5.92%.Found for the barium salt: C, 13.76; H, 3.38; N, 5.15. Calc. for D-Ba1.5[Co(R-cys)3]? 9H2O = C9H33N3O15S3CoBa1.5: C, 13.78; H, 4.24; N, 5.36%. À-Ba1.5[Co(R-cys-N,S)3] and À-M3[Co(R-cys-N,S)3] (M 5 Na1, K1, Rb1, and Cs1; M1/2 5 Ca21 and Sr21) All experiments were performed under an Ar atmosphere. An aqueous solution (20 cm3) of 1.64 g (9.5 mmol) of (R)- H2cys?HCl?H2O and 4.45 g (14.3 mmol) of Ba(OH)2?8H2O was added to a hot aqueous solution (50 cm3) of 0.50 g (1.9 mmol) of [Co(NH3)6]Cl3. A green precipitate began to deposit before long and the mixture was stirred for 2 h at 708C.This crystalline precipitate was composed of the pure L-(R,R,R) salt. Yield = 2.30 g (96%). Found: C, 15.09; H, 3.29; N, 5.82. Calc. for L-Ba1.5[Co(R-cys)3]?5H2O = C9H25N3O11S3CoBa1.5: C, 15.17; H, 3.54; N, 5.90%. The pure L-(R,R,R) barium salt could be converted to the Li1, Na1, K1, Rb1 or Cs1 salt by adding a stoichiometric amount of an aqueous Li2SO4, Na2SO4, K2SO4, Rb2SO4 or Cs2SO4 solution.Then the resultant precipitate of BaSO4 was rapidly removed by filtration and the addition of an excess of ethanol to the filtrate gave the desired salt: all yields were ca. 50%. The lithium salt could be further converted into the Ca21 and Sr21 salts by adding stoichiometric amounts of aqueous CaCl2 and SrCl2 solutions, respectively, and a small amount of methanol. Each yield was ca. 30%. Found for the lithium salt: C, 19.64; H, 5.19; N, 7.52.Calc. for L-Li3[Co(R-cys)3]? 6.5H2O = C9H28N3O12.5S3CoLi3: C, 19.50; H, 5.09; N, 7.58%. Found for the sodium salt: C, 17.64; H, 4.51; N, 6.47. Calc. for L-Na3[Co(R-cys)3]?7H2O = C9H29N3O13S3CoNa3: C, 17.68; H, 4.78; N, 6.87%. Found for the potassium salt: C, 16.58; H, 4.40; N, 6.36. Calc. for L-K3[Co(R-cys)3]?7H2O = C9H29N3O13- S3CoK3: C, 16.38; H, 4.43; N, 6.37%. Found for the rubidium salt: C, 13.19; H, 3.65; N, 5.01. Calc. for L-Rb3[Co(Rcys) 3]?8H2O = C9H31N3O14S3CoRb3: C, 13.23; H, 3.83; N, 5.14%.Found for the caesium salt: C, 11.34; H, 2.99; N 4.41. Calc. for L-Cs3[Co(R-cys)3]?7H2O = C9H29N3O13S3CoCs3: C, 11.49; H, 3.11; N, 4.46%. Found for the calcium salt: C, 14.37; H, 4.37; N, 5.51. Calc. for L-Ca1.5[Co(R-cys)3]?7.5H2O = C9H30- N3O13.5S3CoCa1.5: C, 17.67; H, 4.85; N, 6.64%. Found for the strontium salt: C, 17.67; H, 4.94; N, 6.87. Calc. for L-Sr1.5- [Co(R-cys)3]?10H2O = C9H35N3O16S3CoSr1.5: C, 14.85; H, 4.85; N, 5.77%. The purity of each diastereomer was checked by 1H NMR measurements as shown in Figs. 1 and 2. Determination of kinetic data The exact values of De and e for the D-(R,R,R) and L-(R,R,R) diastereomers could be determined by extrapolating to time zero: e = 376 ± 1 dm3 mol21 cm21 at 582 ± 1 nm and De = 29.90 ± 0.05 dm3 mol21 cm21 at 598 ± 1 nm for the D-(R,R,R) diastereomer and 355 ± 1 dm3 mol21 cm21 at 595 ± 1 nm and De = 110.00 ± 0.05 dm3 mol21 cm21 at 585 ± 1 nm for the L-(R,R,R) diastereomer.In the absence of an excess of (R)- cysteinate, the mononuclear complex fac-[Co(R-cys)3]32 is known to be readily converted to the trinuclear complex Fig. 1 1H NMR spectra of fac-D- and fac-L-Li3[Co(R-cys-N,S)3] in D2O measured within 5 minutes after dissolving at 30 8C. Fig. 2 1H NMR spectral change with time at 30 8C: fac-D-Li3[Co- (R-cys-N,S)3] in D2O as a starting complex.J. Chem. Soc., Dalton Trans., 1999, 1221–1226 1223 [Co{Co(R-cys)3}2]32 in aqueous solution.16,17 In fact, the absorption spectral change in aqueous solution had no isosbestic points as shown in Fig. 3(a). Therefore, the equilibrium state for the D-(R,R,R) L-(R,R,R) inversion could not be realised under such conditions. In order to prevent the polymerization reaction we treated [Co(R-cys)3]32 in a 0.1 M M2{(R)-cysteinate} aqueous solution. After this treatment the absorption spectral change with time exhibited isosbestic points as shown in Fig. 3(b). The concentration of the free (R)-cysteinate may aVect the inversion rates.Within 3 h, however, the CD change rate of the lithium salt of the L-(R,R,R) (or D-(R,R,R)) diastereomer in aqueous solution is similar to that in a 0.1 M Li2(R-cys) aqueous solution, where the isodichroic point at 503 nm in the aqueous solution is coincident with that in a 0.1 M Li2{(R)-cysteinate} aqueous solution. Thus, the CD contribution due to the trinuclear complex formation is negligible small. The equilibrium constants and inversion rate constants (Tables 1 and 2) were obtained at diVerent temperatures: 20, 30 and 40 8C, with a complex concentration of ca. 2 mmol dm23. The manner of typical exponential CD change and the equilibrium state (23 h after dissolution) of the L complex agreed with those of the D one as shown in Fig. 4. The kinetic measurements by CD spectra were recorded at a fixed wavelength (598 nm). The data treatment was carried out according to the scheme of general reversible unimolecular reactions.18 The plots of ln(Deeq 2 Det)D against time are shown in Fig. 5, where the subscript “t” or “eq” means at time t or at equilibrium, respectively, and the subscript D means an initial complex configuration. Det is expressed as DeL (1 2 xt) 1 DeDxt where the character “x” represents the mole fraction (x = [D-(R,R,R)]/{[D-(R,R,R)] 1 [L-(R,R,R)]}. The slopes of linear regressions, 2(kD 1 kL), of ln(Deeq 2 Det)D versus time in Fig. 5 were in fair agreement of the mean values for the D and L complexes as starting materials, and hence they were used for evaluating the individual rate constants.Solubility measurement Solubilities of the diastereomeric salts were measured at 20 8C Fig. 3 Absorption spectral changes with time at 20 8C: (a) the D-(R,R,R) diastereomer in water and (b) the D-(R,R,R) one in a 0.1 M Li2(R)-cysteinate solution. Table 1 Equilibrium constants and thermodynamic parameters of fac-Li3[Co(R-cys)3] in 0.1 M Li2(R-cys) aqueous solutions a T/8C Deeq/dm3 mol21 cm21 K = [L-(R,R,R)]/[D-(R,R,R)] DH8/kJ mol21 DG8/kJ mol21 DS8/J mol21 K21 20 10.85 1.17 25.6 ± 0.2 20.35 ± 0.1 217.6 ± 0.5 30 10.35 1.08 40 10.14 1.01 a The solutions were prepared with complex concentrations of 2 mmol dm23.in water for the Ca21, Sr21 and Ba21 salts and in water–ethanol (1 : 3) for the Li1, Na1, K1, Rb1 and Cs1 salts with use of the molar absorption coeYcients at 582 nm for D-(R,R,R) (e = 376 dm3 mol21 cm21) and at 595 nm for L-(R,R,R) (e = 355 dm3 mol21 cm21).Since both the diastereomers rapidly epimerize as shown in Fig. 2, we measured the solubility of the solution within ten minutes after mixing a sample with the solvent. These solubility data are listed in Table 3 and the solubility ratios were used in the discussion. Measurements UV/Visible absorption spectra were recorded on a Hitachi U-3400 spectrophotometer, 1H NMR spectra on a Bruker ARX300 spectrometer, and CD spectra on a JASCO J-600 spectropolarimeter.Results and discussion Kinetics of mutarotation The 1H NMR spectra of fac-(S)-D- and L-[Co(R-cys-N,S)3]32 in D2O are shown in Fig. 1; they were recorded within 5 minutes after dissolution. The diastereomeric purity could be checked based on these 1H NMR spectra. Fig. 2 shows the spectral changes of the D-(R,R,R) diastereomer during the first 90 minutes after dissolution and clearly indicates rapid inversion around a cobalt center. The signals of the amino protons in the coordinated R-cysteinate gradually disappeared with time because of H–D exchange.The two amino protons at 4.3(doublet) and 3.75(triplet) ppm can be assigned to the axial sites (which are noted at H^ in Fig. 6) and to the perpendicular sites (which are noted as H||) to the C3 molecular axis, respectively. This assignment was confirmed by the following NMR Fig. 4 CD spectra of the D- and L-diastereomers in a 0.1 M Li2- (R)-cysteinate solution and at an equilibrium state at 20 8C.Fig. 5 The relationship between time and ln(Deeq 2 Det)D. The subscript D denotes the configuration of the starting complex.1224 J. Chem. Soc., Dalton Trans., 1999, 1221–1226 Fig. 6 SPARTAN molecular models of the L-(R,R,R) (right) and D-(R,R,R) (left) diastereomers. The axial sites are noted as H^ and the perpendicular sites are noted as H|| to the C3 molecular axis. Table 2 Rate constants and activation parameters for mutarotation of fac-D L-Li3[Co(R-cys)3] in 0.1 M Li2(R-cys) aqueous solutions a T/8C k = kD 1 kL/s21 kD/s21 kL/s21 Ea/kJ mol21 ln A DH�/kJ mol21 DG�/kJ mol21 DS�/J mol21 K21 20 (1.45 ± 0.08) × 1024 (7.8 ± 0.08) × 1025 (6.7 ± 0.08) × 1025 D(R,R,R) æÆ L(R,R,R) 101 ± 2 32 ± 2 98.5 ± 0.6 94 ± 7 16 ± 20 30 (5.77 ± 0.17) × 1024 (3.00 ± 0.17) × 1024 (2.77 ± 0.17) × 1024 L(R,R,R) æÆ D(R,R,R) 107 ± 2 32 ± 2 104.5 ± 0.6 94 ± 6 35 ± 19 40 (2.00 ± 0.09) × 1023 (1.00 ± 0.09) × 1023 (1.00 ± 0.09) × 1024 a The solutions were prepared with complex concentrations of 2 mmol dm23.experiment. Addition of Li2SO4 to the sample solution caused a lower-field shift from 4.3 to 5.0 ppm, whereas no change was observed for the signal at 3.75 ppm. It is well known that the axial amino protons along the C3 axis preferentially interact Table 3 Solubilities (1023 mol dm23) of fac-L- and D-M3[Co(R-cys)3] in H2O–EtOH (1: 3) or in H2O at 20 8C and their solubility ratios M SL SD SL/SD fac-D- and D-M3[Co(R-cys)3] in H2O–EtOH (1: 3) Li Na K Rb Cs 4.8 3.1 7.7 3.5 1.6 14 0.18 1.3 0.74 1.8 22 43 2.7 2.2 fac-L- and D-M1.5[Co(R-cys)3] in H2O Ca Sr Ba >100 >70 1.1 0.23 5.2 9.7 ca. 430 ca. 13 0.11 with sulfate anion and form intermolecular hydrogen bonds,19 which leads to the lower field shift of the axial site protons. Furthermore, the signal at 4.3 ppm is more quickly reduced by the H–D exchange than that at 3.75 ppm as shown in Fig. 2. The result can be interpreted in terms of the site diVerences: the H|| amino protons at 4.3 ppm are located at uncovered sites but the H^ protons at 3.75 ppm are blocked by the carboxyl groups as shown in Fig. 6. On the other hand, a similar NMR experiment of Li2SO4 addition to L-Li3[Co(R-cys)3] solution did not change the chemical shifts of the amino protons. The multiplets at 2.5 and 3.4 ppm gradually appeared with time. These signals are due to the trinuclear complex formation, [Co{Co(R-cys)3}2]32. Upon addition of an excess of the (R)- cysteinate salts, however, the appearance of the signals at 2.5 and 3.4 ppm is clearly suppressed.Therefore, the same conditions were applied for the kinetic experiments. Fig. 3(b) shows absorption spectral changes of the D-(R,R,R) diastereomer with time in a 0.1 M Li2{(R)-cysteinate} aqueous solution. The first d–d absorption band at 582 nm decreases with time, which corresponds to the increase of the L-(R,R,R)J. Chem. Soc., Dalton Trans., 1999, 1221–1226 1225 diastereomer.Two isosbestic points were observed at 625 and 531 nm. The CD spectra changes in Fig. 4 also indicate that the interconversion of two components occurs in 0.1 M (R)- cysteinate aqueous solution. The relationships between time and ln(Det 2 Deeq)L [or ln(Deeq 2 Det)D] at three diVerent temperatures are shown in Fig. 5. The good linear relationships confirm that the present system can be treated as a reversible first order reaction in a 0.1 M (R)-cysteinate solution.This reaction corresponds to the configuration inversion of the cobalt centre as shown in the following scheme: fac-D-(R,R,R) kD kL fac-L-(R,R,R) The equilibrium constants at 20, 30 and 40 8C are listed in Table 1. The Deeq values gave the mole fractions of the two diastereomers in solution. The equilibrium constant K = [L- (R,R,R)]/[D-(R,R,R)] gradually approaches unity with an increase in temperature, which indicates that the D-(R,R,R) diastereomer is more favourable at higher temperatures.The calculated standard free energy and entropy at 25 8C are also listed in Table 1. The negative enthalpy change (DH8 = 25.6 ± 0.2 kJ mol21) indicates the higher stability of the L-(R,R,R) ion over the D-(R,R,R) ion. The negative entropy change (DS8 = 217.6 ± 0.5 J K21 mol21) suggests that the L-(R,R,R) complex has less degrees of freedom (less flexibility) or occupies less space in solution than the D-(R,R,R) one. This fact is supported by SPARTAN calculations 20 for the present complexes as shown in Fig. 6. The D-(R,R,R) diastereomer adopts the (lel)3 conformation and has the carboxyl groups at equatorial positions. On the other hand, the L-(R,R,R) diastereomer can also take the (lel)3 conformation if the carboxyl groups are directed axially along the C3 axis. The electrostatic interactions between the anionic carboxyl groups and the partially positive amino protons may relate to the unexpected stability of this L-(R,R,R) diastereomer.Thus, there is almost no energy diVerence between these two diastereomers in the [Co(R-cys-N,S)3]32 system. The present system provides a striking contrast to [Co(Rpn) 3]31. In the [Co(R-pn)3]31 system, there is a large energy diVerence between two diastereomers and the D-[Co(R-pn)3]31 diastereomer is more stable than L-[Co(R-pn)3]31.21 The inversion rate constants and activation parameters are listed in Table 2. The half-lives for epimerisation were calculated from the observed rate constants (k = kL 1 kD): 80 and 6 minutes at 20 and 40 8C respectively.This means that higher temperatures can readily accelerate the inversion rate and the equilibrium state is rapidly reached. Activation parameters for inversion of L-[Co{(2)-N-(a-phenylethyl)dithiocarbamato)}3] in CH2Cl2 are DH� = 80.3 kJ mol21 and DS� = 239 J mol21 K21. The inversion in these dithiocarbamato 22 or acetylacetonato 23 Co(III) complexes is usually thought to occur by a trigonal twist mechanism, though the inversion of the tris(diphenyldithiocarbamato) cobalt(III) complex is considered to occur via a bond rupture mechanism from an NMR study on DV�.24 On the other hand, the activation parameters for inversion of L-tris{(2)cyclo-O,O9-(1R,2R-dimethylethylene)dithiophosphato} chromium(III) 2 in CHCl3 are DH� = 101 kJ mol21 and DS� = 38 J mol21 K21: a one-ended dissociative mechanism through a five-coordinated intermediate has been proposed in this system.2 In our system, the higher DH� (ca. 100 kJ mol21) and the positive DS� values (ca. 25 J mol21 K21) in Table 2 are indicative of the one-ended dissociation mechanism rather than the trigonal twist one. The inversion rate around a cobalt centre has been suggested to be catalysed by Co21/(R)-cysteinate ions 9 but this is unlikely. We examined whether the presence of 1, 5 or 10 ppm of Co21 in 0.1 M Li2(R)-cysteinate can aVect the inversion rate of 2 mmol dm23 of D-[Co(R-cys)3]32 solution: the inversion rates are unchanged within k = (1.4 ± 0.1) × 1024 s21.On the other hand, the presence of Co21 alone caused the formation of a trinuclear complex, [Co{Co(R-cys)3}2]32.16,17 In fact, addition of a few ppm of Co21 to a 1 mmol dm23 aqueous solution of D-[Co- (R-cys)3]32 without free (R)-cysteinate showed instantaneously the characteristic absorption and CD spectra of the (L,L)- trinuclear complex. This results agrees well with the trinuclear complex study.16 All of the D- or L-fac-tris(R-cysteinato-N,S)cobaltate(III) complexes with alkali and alkaline earth metal cations gave the same 1H NMR, absorption and CD spectra in dilute aqueous solution.Furthermore, the inversion rates and equilibrium states of L-Ba1.5[Co(R-cys-N,S)3] in 0.1 M Ba(R-cys) and D- and L-Na3[Co(R-cys-N,S)3] in 0.1 M Na2(R-cys) are in fair agreement with those of the corresponding lithium complexes in 0.1 M Li2(R-cys). Thus, the type of counter cation does not aVect the mutarotation in such a dilute aqueous solution, which is in contrast to the following second-order asymmetric transformation.Second-order asymmetric transformations in the present salts When [Co(NH3)6]31 (ca. 2 mmol dm23) and an excess of M2(R)- cysteinate (9.5 mmol) in 50 cm3 of water were stirred at 70 8C for 2 h, all solution compositions of Li1, Na1, K1, Rb1 and Cs1 salts became D-(R,R,R): L=(R,R,R) = ca. 57 : 43 as shown in Fig. 7(a). On addition of an equal volume of ethanol, the obtained green precipitate (0.8 g, yield 71%) was composed of D-(R,R,R): L-(R,R,R) = ca. 71 : 29 (Fig. 7(b)) and its filtrate was of D-(R,R,R): L-(R,R,R) = ca. 33 : 67 (Fig. 7(c)). Furthermore, the subsequent stirring of the resultant suspension of (b) and (c) at 20 8C for 12 h led to the almost optically pure D-(R,R,R) diastereomer as a solid in these systems. Fig. 7(d) exhibits the CD spectrum of the D-(R,R,R) diastereomer finally obtained, whose diastereomeric ratio was D-(R,R,R): L-(R,R,R) = ca. 98 : 2. Such a change should be due to a second-order asymmetric transformation. On the other hand, the reaction mixtures containing Ca21, Sr21 and Ba21 salts under an Ar atmosphere immediately resulted in less-soluble precipitates. The precipitates were not optically pure. However, stirring the suspension at 70 8C for 2 h gave the pure D-(R,R,R) diastereomer for Ca21 and Sr21 and the L-(R,R,R) diastereomer for Ba21 as a solid. Arnold and Jackson have reported the second-order asymmetric transformation of the barium salt and obtained the L- (R,R,R) diastereomer as the final solid in the systems starting Fig. 7 CD spectra due to (a) preparative reaction solution ([Co- (NH3)6]Cl3 and 5K2(R)-cysteinate) at 70 8C for 2 h; (b) precipitate obtained by adding an equal volume of ethanol to (a); (c) its filtrate; (d) the precipitate after the suspension of (b) and (c) was stirred at 20 8C for 12 h.1226 J. Chon Trans., 1999, 1221–1226 from Na3[Co(CO3)3]?3H2O, trans-[CoCl2(py)4]Cl?6H2O or K[Co(edta)]?2H2O.9 In the present systems, we found that the second-order asymmetric transformation exists in all salts and the opposite D-(R,R,R) diastereomer is obtainable as the final solid in all salts except for the barium salt.It is quite interesting that the absolute configurations of the final product dramatically change depending upon the counter cations. Solubilities were determined for each salt of the diastereomeric pair (in Table 3).Though the fairly rapid mutarotation in each system prevents the accurate determination of solubility, the solubility ratio becomes a useful index. The diastereomeric solubility ratio (SL/SD) is larger than unit in the salts of Li1, Na1, K1, Rb1, Cs1, Ca21 and Sr21, whereas the ratio (SL/SD) is smaller than unity in the Ba21 salt. The potassium salt shows the largest solubility ratio. In alkaline earth metal salts, the solubility ratios decrease with increasing atomic number (or ionic radii) of the cation.The solubility ratio becomes less than unity only in the barium salt. It seems quite reasonable to consider that these solubility relationships determine which diastereomer would be obtained as the final solid. When a 0.1 M Li2(R)-cysteinate aqueous solution of [Co(Rcys) 3]32 (ca. 1 mmol dm23) was kept standing for 1 day, the resultant equilibrium ratio became ca. 1 : 1 of the D-(R,R,R) and L-(R,R,R) diastereomers. When [Co(NH3)6]31 (ca. 30 mmol dm23) and an excess of M2(R)-cysteinate were reacted at 70 8C for 2 h, the obtained solution composition was fairly D-rich, D-(R,R,R) :L-(R,R,R) = ca. 3 : 2. The diVerence of these diastereomeric ratios may be ascribed to the diVerent solution concentrations. The solution concentration in the former system is very dilute for both species but the concentration of the latter is very high. In fact, the dilution of the concentrated solution with water led to a diastereomeric ratio of ca. 1 : 1 in several hours. In addition, since the preparative solution is highly viscous, it may contain some microcrystals and the second-order asymmetric transformation to the D- (R,R,R) diastereomer may be induced by the microcrystals. This speculation is supported by the observation that the diastereomeric ratio of the concentrated solution increased to D-(R,R,R) :L-(R,R,R) = ca. 2 : 1 by standing overnight in a refrigerator. After Arnold and Jackson’s work, there were some disputes between Gillard14 and Arnold and Jackson.15 The main point raised was a concern with the optical purity of potassium fac-Dtris( R-cysteinato-N,S)cobaltate(III).25 Arnold and Jackson were correct in noting that if crystallisation of M3[Co(R-cys)3] is carried out too rapidly, a D/L mixture results, as confirmed in this work.However, as clearly shown by the present work, optically pure fac-D-K3[Co(R-cys)3] can be obtained if the experimental conditions are carefully chosen to induce a second-order asymmetric transformation even though there is a reversible inversion between the D-(R,R,R) and L-(R,R,R) diastereomers in solution.Acknowledgements We are grateful for a Grant-in-Aid for Scientific Research No. 09640669 (for M. K.) from the Ministry of Education, Science, Sports and Culture. References 1 E. E. Turner and M. M. Harris, Quart. Rev. (London), 1947, 1, 299. 2 P. Biscarini, Inorg. Chim. Acta, 1985, 99, 183, 189. 3 P.Biscarini, R. Franca and R. Kuroda, Inorg. Chem., 1995, 34, 4618. 4 J. A. Broomhead, F. P. Dwyer and J. W. Hogarth, Inorg. Synth., 1960, 6, 183. 5 G. B. KauVman, N. Sugisaki and I. K. Reid, Inorg. Synth., 1989, 25, 139. 6 M. P. Schubert, J. Am. Chem. Soc., 1933, 55, 3336. 7 R. G. Neville and G. Gorin, J. Am. Chem. Soc., 1956, 78, 4891, 4893; G. Gorin, J. E. Spessard and G. A. Wessler, J. Am. Chem. Soc., 1959, 81, 3193. 8 R. D. Gillard and R. Maskill, Chem. Commun., 1968, 160. 9 A. P. Arnold and W. G. Jackson, Inorg. Chem., 1990, 29, 3618. 10 (a) M. Kita, K. Yamanari and Y. Shimura, Bull. Chem. Jpn., 1982, 55, 2873; (b) W. G. Jackson, A. M. Sargeson and P. O.Whimp, J. Chem. Soc., Chem. Commun., 1976, 934. 11 (a) M. Kita, K. Yamanari, K. Kitahama and Y. Shimura, Bull. Chem. Soc. Jpn., 1981, 54, 2995; (b) M. J. Root and E. Deutsch, Inorg. Chem., 1981, 20, 4017. 12 M. Kita, K. Yamanari and Y. Shimura, Bull. Chem. Soc. Jpn., 1989, 62, 3081. 13 M. Murata, M. Kojima, M. Kita, S. Kashino and Y. Yoshikawa, Chem. Lett., 1996, 675. 14 R. D. Gillard, Polyhedron, 1991, 10, 1453. 15 A. P. Arnold and W. G. Jackson, Polyhedron, 1991, 10, 2847. 16 K. Okamoto, S. Aizawa, T. Konno, H. Einaga and J. Hidaka, Bull. Chem. Soc. Jpn., 1986, 59, 3859. 17 S. Aizawa, K. Okamoto, T. Konno, H. Einaga and J. Hidaka, Bull. Chem. Soc. Jpn., 1988, 61, 1601. 18 J. C. Kendrew and E. A. Moelwyn-Hughes, Proc. R. Soc. London, A, 1940, 176, 353. 19 R. Larsson, S. F. Mason and B. J. Norman, J. Chem. Soc. A, 1966, 30. 20 SPARTAN (version 5.0), Wavefunction, inc., 1997. The semiempirical PM3(tm) method was applied to the geometrical optimization of the present diastereomers. The last convergence energy diVerence was below 1.0 × 1024 kJ mol21. The calculated heat of formation of fac-D-[Co(R-cys-N,S)3]32 (28329.7 kJ mol21) is quite similar to that of fac-L-[Co(R-cys-N,S)3]32 (28325.2 kJ mol21), which is consistent with the experimental result (see Table 1). 21 E. J. Corey and J. C. Bailar, J. Am. Chem. Soc., 1959, 81, 2620. 22 R. A. Haines and S. M. F. Chan, Inorg. Chem., 1979, 18, 1495. 23 D. Katakis and G. Gordon, Mechanisms of Inorganic Reactions, Wiley, New York, 1987, p. 218. 24 G. A. Lawrance, M. J. O’Connor, S. Suvachittanont, D. R. Stranks and P. A. Tregloan, Inorg. Chem., 1980, 19, 3443. 25 L. S. Dollimore and R. D. Gillard, J. Chem. Soc., Dalton Trans., 1973, 933. Paper 8/08097F

ISSN:1477-9226    

DOI:10.1039/a808097f

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 1227–1234 1227 Interaction of zinc(II) with the cyclic octapeptides, cyclo[Ile(Oxn)- D-Val(Thz)]2 and ascidiacyclamide, a cyclic peptide from Lissoclinum patella Lisbeth Grøndahl,a Nikolai Sokolenko,b Giovanni Abbenante,b David P. Fairlie,b Graeme R. Hanson c and Lawrence R. Gahan *a a Department of Chemistry, The University of Queensland, Brisbane, QLD 4072, Australia b 3D Centre, The University of Queensland, Brisbane, QLD 4072, Australia c Centre for Magnetic Resonance, The University of Queensland, Brisbane, QLD 4072, Australia Received 12th November 1998, Accepted, 23rd February 1999 The interactions between zinc salts and the naturally occurring cyclic octapeptide ascidiacyclamide in methanol, as well as a synthetic analogue cyclo[Ile(Oxn)-D-Val(Thz)]2, were monitored by 1H NMR and CD spectroscopy. Three zinc complexes were identified, their relative amounts depending on the nature of the anion (perchlorate, triflate or chloride) and the presence or absence of base.Binding constants for two of the zinc species were calculated from CD or 1H NMR spectra, [Zn(L 2 H)]1 (KZn(L2H) = [Zn(L 2 H)1]/[Zn21][(L 2 H)2] = 107 ± 2 M21; 95% methanol/ 5% water, 298.0 K, NEt3/HClO4 buVer 0.04 M) and [ZnLCl]1 (KZnClL = [ZnClL1]/[Zn21][Cl2][L] = 107.2 ± 0.1 M22; d3-methanol, 301 K). Introduction Cyclic peptides having the 24-azacrown-8 macrocyclic structure have been isolated from the aplousobranch ascidian Lissoclinum patella.1–4 Many examples of these cyclic peptides have been isolated and characterised, a common structural feature being the presence of oxazoline and thiazole rings.The patellamides (Chart 1) which exhibit these features are also diVerentiated by the presence of various substituents and by diVerences in the amino acids comprising the structure. A common and intriguing feature of these cyclic peptides is that their structures define a macrocyclic cavity and we,5–7 and others,8 have sought to exploit the macrocyclic properties of these compounds by studying their complexation reactions.That the marine environment from which these macrocyclic peptides originate contains inorganic salts suitable for complexation is unquestioned 8 although the role of these ligands in metal ion complexation and transport within their environment has not been fully elucidated. The solid state structures of ascidiacyclamide and patellamide A have been reported to exhibit a saddle shaped arrangement9–13 whereas patellamide D exhibits a twisted figure of eight configuration.14 The saddle shaped arrangement is retained in the structure of the copper(II) complex of ascidiacyclamide 5 and the same structure was predicted for the copper(II) complex of patellamide D.6 The two copper(II) ions in [Cu2(ascid)2(m-CO3)(H2O)2]?2H2O are bonded through thiazole and oxazoline nitrogens and through a deprotonated isoleucine residue, each metal atom being five coordinate with the remaining coordination sites occupied by a water molecule and the bridging carbonate moiety.5 There is evidence that in solution monomeric and dimeric species (e.g.[CuL] and [Cu2L]; L = deprotonated ascidiacyclamide and patellamide D) are present 5–7 and for patellamide A, B and E species such as ML and M2L [M = Cu(II), Zn(II)] have been predicted.8 In the case of the complexes with patellamide A, B and E binding constants for the formation of the copper and zinc complexes, in the absence of base, have been reported.8 As part of a larger program aimed at investigating (i) the complexation properties of these cyclic peptides, and (ii) the eVects on complexation of structural modification to the peptidic backbone,15 we now report a study of the complexation reactions of two patellamide ligands, ascidiacyclamide (1) and an analogue cyclo[Ile(Oxn)-D-Val(Thz)]2 2 with zinc(II) in methanol.Ascidiacyclamide diVers from cyclo[Ile(Oxn)-DVal( Thz)]2 by the presence of a methyl substituent on each of its oxazoline rings. This study, employing 1H NMR and CD spectroscopy, investigated the role of base and anions in the complexation reactions of zinc(II) with 1 and 2.The eVects on complexation of the structural modifications will be reported in a subsequent publication.15 Experimental All reagents were of analytical grade and used without further purification (methoxycarbonylsulfamoyl)triethylammonium hydroxide (Burgess Reagent) was from Aldrich.Ascidiacyclamide was isolated and purified from L. patella collected from Heron Island on the Great Barrier Reef, Australia.4 Cyclo[Ile(Oxn)-D-Val(Thz)]2 was prepared as described below. The peptides were found to be pure by HPLC and 1H NMR and the water content was determined from CHN-analysis. Mass spectra were obtained on a triple quadrupole mass spectrometer (PE SCIEX API III) equipped with an Ionspray (pneumatically assisted electrospray) atmospheric pressure ionisation source (ISMS).Solutions of compounds in 9 : 1 acetonitrile/0.1% aqueous trifluoroacetic acid for preparative work and methanol for analytical work were injected by syringe infusion pump at mM–pM concentrations and flow rates of 2–5 ml min21 into the spectrometer. Molecular ions, {[M 1 nH]n1}/n, were generated by ion evaporation and focussed into the analyser of the spectrometer through a 100 mm sampling orifice. Full scan data was acquired by scanning quadrupole-1 from m/z 100–900 with a scan step of 0.1 Dalton and a dwell time of 2 ms.For preparative work, 1H NMR spectra were recorded with either a Varian Gemini 300 or a Bruker ARX 500 spectrometer. Proton assignments were determined by pre-saturation experiments or 2D NMR experiments (DFCOSY, TOCSY). Preparative scale reversed phase HPLC separations were performed with Waters Delta-Pak PrepPak C18 40 mm × 100 mm1228 J.Chem. Soc., Dalton Trans., 1999, 1227–1234 cartridges (100 Å); analytical reversed phase HPLC was performed with a Vydac 218TP5415 C18 column using gradient mixtures of water/0.1% TFA (solvent system A) and water 10%/ acetonitrile 90%/TFA 0.1% (solvent system B). The purified compounds were analysed by HPLC starting at (65% A, 35% B) using a linear gradient of 2% min21. For analytical work, 1D and 2D NMR spectra were recorded Chart 1 Cyclic peptides. O N NH N S HN N O HN N S NH O O O O O N NH N S HN N O HN N S NH O O O O O N NH N S HN N O HN N S NH O O O O O N NH N S HN N O HN N S NH O O O O Ascidiacyclamide (1) Patellamide A cyclo[Ile(Oxn)-D-Val(Thz)]2 (2) 3 5 6 8 7 10 11 12 13 14 15 16 17 3 5 6 7 8 10 11 17 18 16 15 14 (1) (2) (2) (1) Patellamide D with a Bruker 400 MHz instrument at 301 K or with a Bruker 500 MHz instrument at variable temperature.The CD spectra were recorded with a JASCO J-710 spectrometer equipped with a NESLAB temperature controller maintaining the cell at 25.00 ± 0.02 8C.The spectra were recorded on a 0.1 cm JASCO cell in the wavelength interval 240–300 nm and with a scan speed of 20 or 50 nm min21. For the binding constant determinations all metal solutions were diluted with Milli Q water and analysed by ICP-AAS or ICP-MS. For the CD titrations in 95% methanol/5% water 0.1 or 0.2 equivalents of zinc (2 mL of a 0.01 M solution measured with a 10 mL Hamilton syringe) was added to 200 mL of a 1 mM peptide solution.A total of at least 14 spectra were used in binding constant determinations. No supporting electrolyte other than a NEt3/HClO4 buVer (0.04 M) was used. The program Specfit 16 (a program for global least squares fitting of equilibrium and kinetic systems using factor analysis and Marquardt minimization) was used to extract binding constants from the CD spectra by employing a model including free peptide, free metal and the 1 : 1 metal–peptide complex. 1H NMR titrations were done in d3- or d4-methanol on 2 to 6 mM peptide solutions by adding zinc solutions with a 10 mL Hamilton syringe. No supporting electrolyte was used and the solutions were kept at ambient temperature. The binding constant was determined from the relative ratio of peptide and complex, obtained from the 1H NMR data, by a semi-manual iterative procedure using the program KaleidaGraph.17 The model included the three equilibria (eqns. (8), (9) and (10)) together with the concentration relations CL = [L] 1 [ZnClL], CZn = [Zn21] 1 [ZnCl1] 1 [ZnCl2] 1 [ZnClL] and CCl = 2CZn = [Cl2] 1 [ZnCl1] 1 2[ZnCl2] 1 [ZnClL].The binding constant was obtained from an average of three values calculated from final peptide:complex ratios of 7 : 4, 6 : 5 and 5 : 11. Synthesis of cyclo[Ile(Oxn)-D-Val(Thz)]2 Abbreviations: DIPEA = diisopropylethylamine; DMF = N,Ndimethylformamide; BOP = [benzotriazol-1-yloxy-tris(dimethylamino) phosphonium] hexafluorophosphate; TFA = tri- fluoroacetic acid; KOBt = the potassium salt of hydroxybenzotriazole; Thz = thiazole; Oxn = oxazoline.Boc-Ile-Ser-D-Val(Thz)-OEt. Boc-D-Val(Thz)-OEt 18 (2.9 g, 8.85 mmol) was stirred in TFA (30 ml) for 30 minutes. The solution was evaporated in vacuo and the oily residue neutralised with 5% NaHCO3 (200 ml) and extracted with dichloromethane (2 × 50 ml). The organic extract was dried over Na2SO4, filtered and evaporated to give H-D-Val(Thz)-OEt as an oil.H-D-Val(Thz)-OEt was coupled to Boc-Ser-OH (2.5 g, 10 mmol) by dissolving both in DMF (50 ml) together with KOBt (1.7 g, 10 mmol) and BOP reagent (4.42 g, 10 mmol). After stirring at room temperature for 1 hour dimethylaminopropylamine (1.3 ml, 10 mmol) was added and the resulting mixture evaporated to dryness and redissolved in ethyl acetate (100 ml). The solution was washed with 10% NaHSO4 (2 × 50 ml), 5% NaHCO3 (2 × 50 ml) and brine (2 × 50 ml), dried over Na2SO4, filtered and evaporated to give Boc-Ser-D-Val(Thz)- OEt as an oil.The oil was deprotected with TFA and coupled to Boc-Ile-OH (2.40 g, 10 mmol) by dissolving both in DMF (30 ml) together with BOP (4.42 g, 10 mmol) and DIPEA (4.3 ml, 25 mmol) and stirring the mixture at room temperature for 1.5 hours. Dimethylaminopropylamine (1 ml, 8 mmol) was then added and the solution evaporated. The oily residue was redissolved in ethyl acetate (200 ml) and the solution was washed with 10% NaHSO4, 5% NaHCO3 and brine solutions, dried over Na2SO4, filtered and evaporated.The residue was purified by column chromatography on silica gel using ethyl acetate as eluent to give the title compound as an oil (yield: 3.2 g, 69% based on initial Boc-D-Val(Thz)-OEt). 1H NMR (CDCl3) d 8.08, s, 1H, Thz-H; 7.51, d, 1H, JValNH-Vala = 8.7 Hz, Val-NH;J. Chem. Soc., Dalton Trans., 1999, 1227–1234 1229 7.32, d, 1H, JSerNH-Sera = 7.0 Hz, Ser-NH; 5.28, dd, 1H, JValNH-Vala = 8.7 Hz, JVala-Valb = 5.4 Hz, Val-aH; 5.05, d, 1H, JIleNH-Ilea = 5.9 Hz, Ile-NH; 4.46, m, 1H, Ser-a; 4.38, q, 2H, JCH2–CH3 = 7.0 Hz, OEt-CH2; 4.16, m, 1H, Ser-bH; 4.03, m, 1H, Ile-aH; 3.72, m, 1H, Ser-bH; 3.36, s, 1H, Ser-OH; 2.42, m, 1H, Val-b; 2.00, m, 1H, Ile-bH; 1.50–1.59, m, 1H, Ile-g(CH2); 1.38, m, 3H, OEt-CH3; 1.38, m, 9H, Boc-(CH3); 1.10–1.18, m, 1H, Ile-g(CH2); 1.04, d, 3H, JValb-Valg = 6.6 Hz, Val-g(CH3); 0.85–1.01, m, 9H, Val-g(CH3), Ile-g(CH3), Ile-d(CH3).HPLC; retention time rt = 14.8 min.ISMS: M 1 H = 529.2. Boc-Ile1-Ser1-Val1(Thz)-Ile2-Ser2-Val2(Thz)-OEt. Boc-Ile-Ser- D-Val(Thz)-OEt (450 mg, 0.852 mmol) was deprotected with 4 M HCl/dioxane solution (5 ml) for 30 minutes. The solution was evaporated and the residue redissolved in a mixture of CH2Cl2 (50 ml) and 5% NaHCO3 (100 ml). The CH2Cl2 solution was separated, the water solution was washed with CH2Cl2 (1 × 50 ml) and ethyl acetate (1 × 50 ml). The combined organic layers were dried over Na2SO4 and evaporated to give H-Ile- Ser-D-Val(Thz)-OEt as an oil.Boc-Ile-Ser-D-Val(Thz)-OEt (450 mg, 0.852 mmol) was dissolved in a mixture of 1 M LiOH (1 ml) and ethanol (5 ml). The reaction mixture was stirred for 6 h at room temperature and acidified to pH ~ 9 with CO2 (dry ice) and HOBt (14 mg, 0.1 mmol) was added. The solution was filtered, evaporated and redissolved in DMF (10 ml) and BOP reagent (443 mg, 1 mmol) added to H-Ile-Ser-D-Val(Thz)-OEt. The reaction was stirred for 4 h at room temperature and dimethylaminopropylamine (0.1 ml, 0.8 mmol) added.Ethyl acetate (100 ml) was added and the solution washed with 10% NaHSO4 (2 × 50 ml), 5% NaHCO3 (2 × 50 ml) and brine (1 × 50 ml), dried over Na2SO4, filtered and evaporated. The residue was dissolved in 15 ml of acetonitrile/water mixture (1 : 1) and purified using preparative HPLC to give the title compound as a white powder (669 mg, 86.3%). 1H NMR (d6- DMSO, 293 K): d 8.59, d, 1H, J = 9.2 Hz, Val2-NH; 8.54, d, 1H, J = 9.2 Hz,Val1-NH; 8.42, s, 1H, Thz-H; 8.30, d, 1H, J = 8.7 Hz, Ser2-NH; 8.19, s, 1H, Thz-H; 7.89, d, 1H, J = 8.7 Hz, Ile2-NH; 7.79, d, 1H, J = 8.5 Hz, Ser1-NH; 6.79, d, 1H, J = 8.8 Hz, Ile1- NH; 4.95, m, 4H, Val1-aH,Val2-aH, Ser1-bH, Ser2-bH; 4.89, m, 3H, Ser1-bH, Ser2-aH, Ile2-aH; 4.29, q, 2H, J = 7.9 Hz, -OEt; 3.84, m, 1H, Ile1-aH; 3.60, m, 2H, Ser1-aH, Ser2-aH; 2.28, m, 2H, Val1-bH, Val2-bH; 1.79, m, 1H, Ile2-bH; 1.68, m, 1H, Ile1- bH; 1.42, m, 2H, Ile1-gCH, Ile2-gCH; 1.34, s, 9H, Boc; 1.29, t, 3H, J = 7.9 Hz, -OEt; 1.06, m, 2H, Ile1-gCH, Ile2-gCH; 0.94– 0.80, m, 12H, Val1-gCH3, Val2-gCH3; 0.80–0.73, m, 12H, Ile1- gCH3, Ile2-gCH3, Ile1-dCH3, Ile2-dCH3.HPLC; rt = 17.1 min. ISMS: M 1 H = 911.7. Cyclo-[-Ile-Ser-D-Val(Thz)-]2. Boc-[Ile-Ser-D-Val(Thz)-]2- OEt (650 mg, 0.714 mmol) was dissolved in 4 M HCl/dioxane solution (10 ml) for 30 min at room temperature. The solution was evaporated and the residue was redissolved in a mixture of 1 M LiOH (2 ml) and ethanol (10 ml).The reaction mixture was stirred for 6 h at room temperature, then acidified to pH ~ 9 with CO2 (dry ice) and HOBt (14 mg, 0.1 mmol) was added. The solution was filtered, evaporated and redissolved in DMF (20 ml) and the solution added via syringe pump to a solution of BOP reagent (885 mg, 2 mmol) in DMF (200 ml) over 5 h. The reaction mixture was stirred overnight and evaporated. The residue was dissolved in 15 ml of acetonitrile/water mixture (1 : 1) and purified using preparative HPLC to give the title compound as a white powder (209 mg, 38.4%). 1H NMR (d6- acetone) d 9.52, d, 1H, J = 5.5 Hz, Ile-NH; 8.67, d, 1H, J = 6.0 Hz, Ser-NH; 7.79, d, 1H, J = 9.5 Hz, Val-NH; 7.47, s, 1H, Thz-H; 5.34, m, 1H, Val-aH; 4.47, m, 1H, Ser-aH; 4.12, m, 1H, Ser-bH; 3.99, m, 1H, Ser-bH; 3.93, dd, JIlea-IleNH = 5.5 Hz, JIlea-Ileb = 6.0 Hz, 1H, Ile-aH; 2.42, m, 1H, Val-b; 2.39, m, 1H, Ile-b; 1.81, m, 1H, Ile-gCH2; 1.29, m, 1H, Ile-gCH2; 1.14, d, 3H, JIleb-Ileg = 7 Hz, Ile-gCH3; 1.07, d, 3H, JValb-Valg = 6.5 Hz, Val-gCH3; 1.03, d, 3H, Jvalb-Valg = 6.5 Hz, Val-gCH3; 0.89, t, 3H, JIled-Ileg = 7.3 Hz, Ile-dCH3.ISMS: M 1 H = 765.4. HPLC; rt = 13.0 min. Cyclo[-Ile(Oxn)-D-Val(Thz)-]2. Cyclo[-Ile-Ser-D-Val(Thz)-]2 (200 mg, 0.262 mmol) and Burgess Reagent (120 mg, 0.5 mmol) were dissolved in dry THF (5 ml). The reaction mixture was heated and stirred for 5 h at reflux temperature and evaporated. The residue was dissolved in 15 ml of acetonitrile/water mixture (1 : 1) and purified using preparative HPLC to give the title compound as a white powder (25 mg, 13.2%). 1H 500 MHz NMR (CDCl3) d 8.02, s, 2H, Thz-H; 8.02, d, 2H, Ile-NH; 7.30, d, 2H, JValNH-Vala = 10 Hz, Val-NH; 5.21, m, 2H, Val-a; 4.82, m, 2H, Oxn-H; 4.74, m, 2H, Ile-a; 4.66, m, 2H, Oxn-H; 4.58, m, 2H, Oxn-H; 2.29–2.34, m, 2H, Val-b; 1.95, m, 2H, Ile-b; 1.30, m, 2H, Ile-gCH2; 1.17, m, 2H, Ile-gCH2; 1.13, d, 6H, JValb-Valg = 6.7 Hz, Val-gCH3; 1.05, d, 6H, JValb-Valg = 6.6 Hz, Val-gCH3; 0.80, d, 6H, JIleb-Ileg = 6.7 Hz, Ile-gCH3; 0.73, t, 6H, JIleg-Iled = 7.2 Hz, Ile-dCH3.HPLC; rt = 17.4 min. ISMS: M 1 H = 729.2. Results and discussion NMR characterisation of cyclo[Ile(Oxn)-D-Val(Thz)]2 2 The 1H NMR spectrum of 2 in d3-methanol was assigned on the basis of 2D COSY and TOCSY NMR spectra (Table 1). Both amide protons of 2 had small temperature coeYcients (Ile NH 0.2 ppb/deg, Val NH 1.1 ppb/deg) indicating relatively little interaction of these protons with the solvent.19 The proton/ deuterium exchange rate of both amide protons of 2 was very slow, i.e.in d3-methanol with 12% d4-methanol no proton exchange could be detected over a month. This implies that the NH groups have very limited exposure to the solvent.19 However, the proton/deuterium exchange rate of the Ile N(1)H protons in d4-methanol is faster (within 5 minutes) than that of the Val N(2)H proton (within 10 minutes). Similarly, the proton/deuterium exchange rate in 1 in d3-acetonitrile with 3% d4-methanol is faster for the Ile N(1)H proton (t1/2 ~ 14 days) than the Val N(2)H proton (t1/2 > one month).Since small temperature coeYcients as well as slow proton exchange are observed for both amide protons of 2 it seems likely that the peptide takes one of the conformations found for the other patellamides,9,14 type II and type III (Chart 2). In both of these two conformations the amide protons are shielded from the solvent.In type II the amide protons are buried within the cavity of the saddle, whilst in type III all amide protons take part in intramolecular hydrogen bonding. Thus, slow proton exchange as well as small temperature coeYcients are expected for both conformations and cannot therefore be used to distinguish between them. It has been found from solid state structures of 1 that in the saddle shape conformation, type II, the peptide is rather flexible in that the C(O)–N–Ca–(Thia/Oxn) torsion angles vary for Table 1 1H NMR chemical shifts for 2, 3 and 5 (CD3OH) Proton(s) 35678 10 11 13 14 15 16 17 N(1) N(2) 2 7.93 (s) 5.17 (dd) 2.38 (m) 1.06 (d) 1.15 (d) ——— 2.00 (m) 1.4 (m), 1.2 (m) 0.80 (t) 0.82 (d) 7.99 (J = 7.8 ± 0.4) 7.53 (J = 10.1 ± 0.3) 3 7.94 (s) 4.9 (d) 2.6 (m) 1.21 (d) 0.8–0.9 (d) ——— 1.75 (m) 1.4 (m), 1.2 (m) 0.8–0.9 (t) 0.21 (d) —— 5 7.97 (s), 8.06 (s) 5.1 2.65, 3.00 (m) 1.12, 1.14 (d) 0.85, 0.88 (d) —— 4.9 1.65 (m), 1.9 (m) 1.1 (m), 1.2 (m) 0.70 (t), 0.82 (t) 0.34 (d), 0.34 (d) 8.79 8.76, 8.831230 J.Chem. Soc., Dalton Trans., 1999, 1227–1234 diVerent solvates of the peptide.10 Thus, the Ile torsion angle varies from 2127 to 21608 and the Val torsion angle varies from 98 to 1348.9–12 Translating these torsion angles into 3JHNCaH coupling constants using the possible range of parameters in the Bystrov–Karplus equation 20 yields 3JHNCaH (Ile) of 6.8 to 11.6 Hz, and 3JHNCaH(Val) 8.5 to 11.5 Hz.Patellamide D takes the type III conformation in the solid state and translating torsion angles from this structure 14 into coupling constants yields 3JHNCaH(Ile) 5.4 to 7.8 Hz and 3JHNCaH(Val) 7.3 to 9.8 Hz. The observed coupling constants for 2 (Table 1, 3JHNCaH(Ile) 7.8 ± 0.4 Hz and 3JHNCaH(Val) 10.1 ± 0.3 Hz) fall within either of these ranges, thus, the coupling constants for the patellamides do not provide information on the conformation in solution.A 2D NOESY NMR spectrum of 2 in d3-methanol yielded a number of non-sequential cross peaks (3–5, 3–8, N2–7, N2–8, N2–16, N2–17, N1–16, N1–17, 13–16 and 13–17) which implies type II conformation of the peptide. In particular the N2–17 cross peak points towards type II; the equivalent distance in 1 is 3 Å 9–12 and in patellamide D it is >5 Å,14 thus a type II conformation seems most likely and would be in agreement with the result of a study by CD spectroscopy on the structurally similar peptide patellamide A where it was found that at room temperature the peptide took the saddle shape, type II conformation.8 Zn21–peptide interactions Zinc binding to the cyclic octapeptides 1 and 2 was followed by CD and 1H NMR spectroscopy.In d3-acetonitrile broadening Chart 2 Conformations of the patellamides (taken from ref. 10). N S HN O O HN O N S N O O N HN O N S N O O N NH O H H HN N O O HN O N S NH O S N O NH O N .......... ........... N S NH O O NH O ............... ....... ...... Type II Type I Type III and changes in the chemical shifts of the NMR signals, as well as the development of new broad signals, were observed when a zinc solution was added to a peptide solution and this solvent was therefore not suitable for a study of zinc peptide interactions. In d3-methanol it was found that the zinc ion was not substitutionally labile on the NMR time scale. Similar slow exchange behavior has been reported for calcium complexes of bicyclic peptides.21,22 With the peptides 1 and 2, as the zinc concentration increased a new set of signals appeared and the original peptide signals disappeared.Therefore, d3-methanol was chosen for a detailed investigation of zinc interactions with the peptides. It was found that up to three zinc complexes could be observed, their relative amounts depending on the type of anion and the presence of base. In the following text we will focus on the conditions under which we observe simple reaction stoichiometry, that is when using triflate or perchlorate salts in the presence of base, as well as when using the chloride salt in the absence of base.Having chloride ion as well as base present leads to very complicated systems which will also be described briefly. In all studies with the peptides 1 and 2, we have not observed any diVerences in their behaviour. Therefore, in the following text we assume that the reactivities of 1 and 2 are identical.However, we do point out on which peptide a specific experiment was done. Peptide interactions with zinc triflate or perchlorate In the absence of base, adding zinc triflate or zinc perchlorate to a solution of 2 in d3-methanol (up to 8 equivalents of zinc to peptide) gave no change in the 1H NMR spectrum. However, in the presence of a base, complex formation was observed and could be followed by 1H NMR without precipitation of Zn(OH)2 when 1,8-diazabicyclo[5.4.0]undec-7-ene was used as base (3 equivalents to the peptide).Titration with up to one equivalent of zinc triflate led to shifts in the base signals but no changes were observed for the peptide signals. The final spectrum after addition of a total of 2 equivalents of zinc triflate showed the presence of one new species, complex 3, and no signals from the free peptide remained (Fig. 1). We interpret this as the base interacting more strongly with zinc than does the peptide; however, after saturating the base the second equivalent of zinc reacts with the peptide forming a 1 : 1 complex.Using 1 in place of 2 gives similar changes upon zinc addition. Investigation of the zinc complexation to the peptide in the presence of triethylamine by CD-spectroscopy showed a gradual spectral change (Fig. 2). The plot (at 265 nm) of q Fig. 1 1H NMR spectra in d4-methanol of: (a) a solution containing cyclo[Ile(Oxn)-D-Val(Thz)]2 (1.94 mg, 0.00248 mmol) and 1,8-diazabicyclo[ 5.4.0]undec-7-ene (1 ml, 0.0067 mmol) in 0.7 ml; (b) added zinc triflate solution (0.00306 mmol); (c) added zinc triflate (a total of 0.0051 mmol). [B = signals arising from 1,8-diazabicyclo[5.4.0]undec-7-ene; numbers refer to assignments of the resonances of peptide 2 in (a) and (b) and of complex 3 in (c) (see Table 1)].J.Chem. Soc., Dalton Trans., 1999, 1227–1234 1231 (mdeg) versus equivalents of added zinc(II) reached a plateau after approximately two equivalents of zinc perchlorate.Further addition of zinc caused additional changes in the spectrum but no second plateau was observed after adding seven equivalents of zinc. Employing the data up to the addition of two equivalents of zinc for the determination of a binding constant resulted in a model including a 1 : 1 complex giving a significantly better fit to the data (half as large residuals) than a model including a 2 : 1 complex. The mass spectrum of 2 and zinc triflate indicated the presence of a [ZnL(CF3SO3)]1 species but no dimeric zinc peptide complex could be detected (Table 2).However, the mass spectrum of a solution of 2 and zinc chloride indicate the presence of both mononuclear and dinuclear complexes containing chloride, i.e. [ ZnLCl]1 and [Zn2(L 2 2H)Cl(H2O)]1 (Table 2). Structure of the zinc–peptide complex 3 The zinc–peptide complex 3 formed in solutions containing cyclo[Ile(Oxn)-D-Val(Thz)]2 (or ascidiacyclamide), base and Fig. 2 CD titration of a 95% methanolic solution containing cyclo- [Ile(Oxn)-D-Val(Thz)]2 (1.00 mM, 0.2 mmol) and triethylamine/ perchloric acid buVer, [NEt3] = 0.04 M, [HClO4] = 0.02 M, added (from positive to negative Cotton eVect) 0, 0.02, 0.04, 0.06, 0.08, 0.10, 0.12, 0.14, 0.16, 0.20, 0.24, 0.28, 0.32 and 0.40 mmol zinc perchlorate. Table 2 Electrospray mass spectral data for cyclo[Ile(Oxn)-DVal( Thz)]2 and its zinc complexes in methanol m/z Species Exptl.a Calc.b Cyclo[Ile(Oxn)-D-Val(Thz)]2 LH1 L(H2O)H1 L(CF3SO3H2)1 729.2 747.6 879.3 729.9 747.9 879.9 Mononuclear zinc cyclo[Ile(Oxn)-D-Val(Thz)]2 complexes Zn(L 2 H)1c ZnL(CF3SO3)1c Zn(L 2 2H)Na1 Zn(L 2 H)(H2O)(CH3OH)1 ZnLCl1c 791.2 941.3 813.4 843.5 827.3 791.3 941.3 814.3 841.3 827.8 Dinuclear zinc cyclo[Ile(Oxn)-D-Val(Thz)]2 complexes Zn2(L 2 2H)Cl(H2O)1 908.3 908.2 a Cone voltage 40 V.b Calculated mass over charge ratios were determined using naturally abundant isotopes. c Cone voltage 120 V. zinc triflate (or perchlorate). 1H NMR titrations in the presence of the base 1,8-diazabicyclo[5.4.0]undec-7-ene together with CD and mass spectroscopy gave evidence for 3 being a 1: 1 (metal : peptide) complex (see above). Since the complex only forms in the presence of base we suggest that the peptide is partly deprotonated. On the basis of proton/ deuterium exchange rates of the free peptide as well as comparison with the copper structure,5 the most likely site for deprotonation is the isoleucine amide nitrogen.The 1H NMR spectrum of complex 3 was assigned on the basis of 2D COSY and TOCSY NMR spectra, and chemical shifts of the peptide 2 and of the zinc complex 3 are given in Table 1. In 3 the number of signals in the 1H NMR spectra implies that the molecules have C2 symmetry. In order to form a C2- symmetric zinc complex in which one zinc is bound to a partly deprotonated peptide, we must assume deprotonation at two identical amide nitrogens to form a mononuclear complex, i.e.[Zn(L 2 2H)]. However, it is hard to justify that zinc chooses to form a C2-symmetric complex when it requires deprotonation of two amide protons. It is possible that the apparent C2- symmetry arises from fast exchange or fast molecular motion which, on average, yields a C2-symmetric molecule. A low temperature 1H NMR study (down to 290 8C) of complex 3 was not able to confirm this as slow exchange or freezing out of one dominant conformation was not achieved.In conclusion complex 3 is proposed to be a [Zn(L 2 H)]1 complex in which the zinc ion is coordinated to deprotonated isoleucine amide nitrogen and probably also to one or two other donor atoms from the peptide. Kinetic eVects in zinc–peptide interactions in the presence of base Although the Zn21 ion is generally considered labile, a close- fitting ligand can alter the reaction rate significantly as has been observed in a macrobicyclic complex.23 Furthermore, slow kinetics in barium ion binding to cyclic tetrapeptides in 95% methanol have been observed.24,25 Formation of complex 3 was monitored by CD spectroscopy.The spectrum of a solution of 1, triethylamine (4 equivalents to peptide) and zinc perchlorate (1.5 equivalents to peptide) was indicative of partial zinc–peptide complex formation. The solution was monitored over one hour and no change in the spectrum could be observed. Addition of perchloric acid (4 equivalents to peptide) immediately reformed the spectrum of the free peptide.Thus, there seems to be no kinetic eVects in the formation of complex 3 and reversibility in the system is evident (eqn. (1)). Therefore we chose to quantify this by a binding constant determination employing CD spectroscopy (see below). Investigation of the formation of complex 3 by NMR spectroscopy over a longer time period resulted in the emergence of a more complicated picture. When adding one equivalent of zinc triflate to a solution of 1 mM 2 and three equivalents of base (triethylamine or 1,8-diazabicyclo[5.4.0]undec-7-ene), partial formation of complex 3 was observed instantly. However, after three days another species, complex 4, was present as well (Fig. 3). Looking at the sample over months it was clear that the formation of 4 was indeed a very slow process. Thus, in the presence of base an initial equilibrium is established forming complex 3 (eqn. (1)) followed by an additional slow process forming complex 4 (eqn.(2)). Zn21 1 L fast [Zn(L 2 H)]1 1 H1 (1) [Zn(L 2 H)]1 slow [Zn(L 2 H)*]1 (2) Due to co-precipitation of zinc hydroxide in aged basic solutions it was not possible to extract quantitative information from these NMR experiments and a binding constant for1232 J. Chem. Soc., Dalton Trans., 1999, 1227–1234 complex 4 could not be obtained. In solutions of 2, base and zinc triflate the ratio between 2 and complex 3 and 4 indicates that 4 is also a 1 : 1 (metal : peptide) complex.Since it (like 3) only forms in the presence of base, we suggest that 4 is also a [Zn(L 2 H)]1 complex in which the zinc ion binds diVerently to the peptide than in complex 3. Rearrangements in zinc peptide complexes have been observed for other systems.26 Binding constant for the Zn21–peptide complex 3 The spectral changes associated with zinc addition (up to 2.4 equivalents, Fig. 2) in basic solutions were used to determine the binding constant for complex 3.Measurements were performed in 95% methanol solutions either containing four equivalents of NEt3, or in buVered solutions with 0.04 M NEt3 and 0.02 M HClO4. No supporting electrolyte other than the buVer was used. The results (Table 3) show that NEt3 solutions and the buVered solutions yield the same value for the binding constant for a 1 : 1 complex. Furthermore, there is no eVect of the peptide (ascidiacyclamide 1 versus cyclo[Ile(Oxn)-DVal( Thz)]2 (2)) or of the anion (perchlorate versus triflate) within experimental error.Peptide solutions containing the base 1,8-diazabicyclo- [5.4.0]undec-7-ene instead of NEt3 show similar changes upon addition of zinc. However, the data did not indicate simple formation of a 1 : 1 complex which is probably due to the zinc ion interacting with the base, as was also concluded from the 1H NMR experiments. The perchlorate salts of lithium, sodium, magnesium, calcium and barium, all metals that prefer oxygen donor ligands, did not give rise to spectral changes when performing titrations similar to the one described above.Fig. 3 Methyl region of a 1H NMR spectrum of cyclo[Ile(Oxn)-DVal( Thz)]2 (1.14 mg, 0.0016 mmol) in 0.7 ml d4-methanol with added 1,8-diazabicyclo[5.4.0]undec-7-ene (0.005 mmol) and zinc triflate (0.0019 mmol); spectrum recorded after 4 months. (Numbers refer to resonances assigned to peptide 2, complex 3 and complex 4.) Table 3 Binding constants for base chloride assisted formation of Zn21–peptide complexes in 95% methanol Ligand 12221 Base NEt3–buVer a NEt3–buVer a NEt3 b NEt3 b — Metal salt Zn(ClO4)2?6H2O Zn(ClO4)2?6H2O Zn(ClO4)2?6H2O Zn(CF3SO3)2 ZnCl2 log K 3.0 ± 0.3 c,d 2.7 ± 0.3 c,d 2.9 ± 0.3 c,d 3.4 ± 0.3 c,d 7.2 ± 0.1 e,f a 0.04 M NEt3 and 0.02 M HClO4. b 4 equivalents.c Kobs in eqn. (3). d Determined from CD titration. e KZnClL in eqn. (8), an average of three values calculated from final peptide :complex ratios of 7 : 4, 6 : 5 and 5 : 11.f Determined from batch NMR titration in d3-methanol. The calculated binding constant obtained in titrations using zinc perchlorate or zinc triflate (log K = 3.0 ± 0.3) describes the overall equilibrium of eqn. (3) (L is 1 or 2), in which triethyl- Zn21 1 L 1 NEt3 Kobs [Zn(L 2 H)]1 1 HNEt3 1 (3) amine assists deprotonation of the peptide. We have chosen to describe zinc perchlorate and zinc triflate as being totally dissociated as there is evidence that this is the correct description in a methanol solution when the concentration of the salt is less than 0.2 M.27 Eqn.(3) can be described as a sum of eqn. (4)–(6). Zn21 1 (L 2 H)2 KZn(L 2 H) [Zn(L 2 H)]1 (4) HNEt3 1 KH NEt3 NEt3 1 H1 (5) L KHL (L 2 H)2 1 H1 (6) The binding constant KZn(L 2 H) (eqn. (4)) which describes the binding of zinc to the deprotonated peptide is the formation constant for the complex. Unfortunately, the protonation constants for amides are not known, and therefore the formation constant can only be estimated.By assuming pKH L = 15 (amide deprotonation),28 and pKH NEt3 = 10.9,29 we estimate the value to be KZn(L 2 H) = [Zn(L 2 H)1]/[Zn21][(L 2 H)2] = 107 ± 2 M21. The binding constant for the non-base assisted formation of the complex can be calculated using eqn. (3) and (5) and yields K = [Zn(L 2 H)1][H1]/[Zn21][L] = 1027.9 ± 0.3 M21. Clearly, the assistance from a moderately strong base (pKa > 9) is needed for the formation of complex 3 in which the zinc ion stabilises the deprotonated peptide.Peptide interactions with zinc chloride in the absence of base When zinc chloride is added to a solution of 2 in d3-methanol changes in the 1H NMR spectrum are observed (Fig. 4) and similar changes are observed when 1 is used in place of 2. These changes are assigned to formation of a product, complex 5, which lacks C2 symmetry. When a total of 6 equivalents of zinc chloride has been added to the peptide solution approximately 80% of peptide still remains unreacted.When a total of 45 equivalents of zinc chloride has been added 10% remains. Thus, in the course of adding 45 equivalents of zinc chloride a gradual formation of a new product is evident. By CD spectroscopy, Fig. 4 Selected regions of the 1H NMR spectra of: (a) a solution containing cyclo[Ile(Oxn)-D-Val(Thz)]2 (1.3 mg, 0.0017 mmol) in 0.7 ml d3- methanol; (b) added zinc chloride solution (0.009 mmol zinc); (c) added zinc chloride solution (a total of 0.075 mmol).[Numbers refer to assignments of the resonances of complex 5 (see Table 1).]J. Chem. Soc., Dalton Trans., 1999, 1227–1234 1233 the spectral changes associated with addition of zinc chloride seem to reach a plateau after addition of 4 equivalents (Fig. 5). Further addition of zinc chloride results in further changes. Applying a simple model similar to that used by Freeman et al.8 (eqn. (7)) for formation of a 1 : 1 complex for the first part of Zn21 1 L KZnL [ZnL]21 (7) this titration yields a binding constant which approximates that reported for patellamide A (log K = 4.5).8 In that work the implication was that the equilibrium is shifted totally to formation of the 1 : 1 complex when 4 equivalents of zinc chloride is added.However, in our work the 1H NMR titration showed that a simple 1 : 1 complex is not totally formed at this point in the titration. It is clear that a diVerent model is needed to explain the observed changes in the CD spectra upon addition of zinc chloride.Structure of the zinc–peptide complex 5 Complex 5 formed in solutions of 1 or 2 and zinc chloride. The 1H NMR spectrum of 5 in d3-methanol shows only three amide signals, thus one amide signal is missing. From 2D TOCSY NMR experiments two valine and one isoleucine amide protons could be observed. One explanation for the absence of one amide signal is that deprotonation at an amide nitrogen has occurred as a consequence of zinc binding to the amide nitrogen in a metal assisted deprotonation reaction.However, if that was the case, one would expect this complex to dominate in basic solution which is not the case as complex 3 (which has a very low binding constant) is dominating at low metal to peptide ratios (see later). Another explanation for the absence of one of the amide signals in the 1H NMR spectrum is that zinc is binding to the neutral amide linkage, thereby enhancing the proton exchange rate of the amide proton leading to broadening of the signal.Enhanced proton exchange rates upon zinc binding has been observed for other peptides.30 Binding a metal ion to a neutral amide linkage can occur through the amide oxygen of the amide tautomer (a in Chart 3) or the amide nitrogen of the imidol tautomer (b in Chart 3).31 Of the two potential coordination sites the carbonyl oxygen is considered the most likely on the basis of protonation, infrared and crystallographic studies28 and to the best of our knowledge there is no Fig. 5 CD titration of a 95% methanolic solution containing cyclo- [Ile(Oxn)-D-Val(Thz)]2 (1.1 mM, 0.21 mmol) with added (from positive to negative Cotton eVect), 0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.5, 4.5, 5.5 and 7.5 equiv. zinc chloride. crystallographic evidence for a zinc ion coordinated to the nitrogen of the imidol tautomer of a neutral amide linkage. The 1H NMR spectra of complexes 3 and 5 are very similar, both containing highly shielded Ile methyl groups (Table 1), thus suggesting that the peptide conformation is similar for the two complexes, but not necessarily that the zinc ion is coordinated to the peptide in a similar manner.On this basis we propose that the zinc ion in complex 5 is coordinated to the isoleucine amide oxygen of the neutral amide linkage and it is likely that the metal also binds to one or two other donor atoms of the peptide. Since complex 5 only can be detected in the presence of chloride ion it is assumed that chloride is also coordinated to zinc in the complex.Kinetic eVects in zinc–peptide interactions in the absence of base In our studies we found that zinc binding, particularly in the absence of base and presence of chloride ion, is a slow process. For example, adding 20 equivalents of zinc chloride to a solution of 2 and leaving it for three days changes the extent of reaction from 54 to 65%. After an additional week, no further reaction had taken place.Due to the slow kinetics we decided to obtain a binding constant for complex 5 by monitoring the 1H NMR spectra of equilibrated solutions of peptide 1 and zinc chloride in diVerent ratios. Binding constant for the zinc–peptide–chloride complex 5 Solutions of 1 and zinc chloride in d3-methanol were kept at ambient temperature and the spectra recorded frequently over two weeks to ensure that equilibrium had been reached.The relative intensities of the thiazole signals (obtained from the thiazole integrals) provided the relative ratio of starting material 1 and complex 5 which were used to calculate the binding constant.32 The model used to calculate a binding constant for complex 5 took into account that zinc chloride is not dissociated in methanol under the conditions studied (log K1 = 3.9 and log K2 = 4.2; eqn. (9) and (10)).33 Considering each of the Zn21 1 L 1 Cl2 KZnClL [ZnClL]1 (8) Zn21 1 Cl2 K1 ZnCl1 (9) ZnCl1 1 Cl2 K2 ZnCl2 (10) four complexes (ZnxClyL, x and y = 1 or 2) as possible candidates for complex 5, it was found that a complex with the stoichiometric formula ZnClL gave the best fit to the data (agreement between binding constants at diVerent points in the titration was more than 5 times better than any other considered stoichiometric formula).Thus, the model included eqn. (8) to (10) and yielded a binding constant of KZnClL = [ZnClL1]/ [Zn21][Cl2][L] = 107.2 ± 0.1 M22 (Table 3).Despite the high binding constant for complex 5 a large excess of zinc chloride is needed to completely form the complex because of the competing equilibria (eqn. (9) and (10)) forming the ZnCl1 and ZnCl2 complexes. Peptide interactions with zinc chloride in the presence of base In 1H NMR titrations of 2 with zinc chloride in the presence of base (triethylamine or 1,8-diazabicyclo[5.4.0]undec-7-ene) the three complexes 3, 4 and 5 could all be observed.The ratio Chart 3 Metal binding modes to a neutral amide linkage. O R NHR¢ M N R OH M R¢ a b1234 J. Chem. Soc., Dalton Trans., 1999, 1227–1234 between the three complexes changed throughout the titration; complex 3 was initially the major product but after two equivalents of zinc chloride had been added complex 5 dominated. In solutions which had been left for several days, the ratio between the complexes had changed further. Obviously the product distribution at any time in the titration is controlled by thermodynamic as well as kinetic factors.Due to the complexity of this system these titrations were not taken any further. Conclusion It was possible to obtain reliable binding constants for the [Zn(L 2 H)]1 complex (in the presence of base) and for the [ZnClL]1 complex (in the absence of base). In our study of the patellamides 1 and 2 we have shown the importance of taking into account kinetic factors. Also we found that no simple ZnL21 complex formed which illustrates the importance of examining the species of the reaction mixture before calculating binding constants.The binding constant for a zinc patellamide A complex (K = 104.5) was determined previously assuming a simple ZnL21 complex, thus, employing the model described by eqn. (7).8 Although similar constants can be calculated for the structurally similar 1 and 2 (Chart 1), we have shown by NMR that the model described by eqn. (7) does not adequately describe the species in solution, instead a model described by eqn.(8)–(10) was employed for the system. Acknowledgements The authors acknowledge financial support from the Australian Research Council and from the Danish Natural Science Research Council. We thank a referee for helpful comments. References 1 C. M. Ireland, A. R. Durso, Jr., R. A. Newman and M. P. Hacker, J. Org. Chem., 1982, 47, 1807. 2 Y. Hamamoto, M. Endo, M. Nakagawa, T. Nakanishi and K. Mizukawa, J.Chem. Soc., Chem. Commun., 1983, 323. 3 D. F. Sesin, S. J. Gaskell and C. M. Ireland, Bull. Soc. Chim. Belg., 1986, 95, 853. 4 B. M. Degnan, C. J. Hawkins, M. F. Lavin, E. J. McCaVrey, D. L. Parry, A. L. van den Brenk and D. J. Watters, J. Med. Chem., 1989, 32, 1349. 5 A. L. van den Brenk, K. A. Bryiel, D. P. Fairlie, L. R. Gahan, G. R. Hanson, C. J. Hawkins, A. Jones, C. H. L. Kennard, B. Moubaraki and K. S. Murray, Inorg. Chem., 1994, 33, 3549. 6 A. L. van den Brenk, D.P. Fairlie, G. R. Hanson, L. R. Gahan, C. J. Hawkins and A. Jones, Inorg. Chem., 1994, 33, 2280. 7 A. L. van den Brenk, Ph.D. Thesis, The University of Queensland, 1994. 8 D. J. Freeman, G. Pattenden, A. F. Drake and G. Siligardi, J. Chem. Soc., Perkin Trans. 2, 1998, 129. 9 T. Ishida, M. Tanaka, M. Nabae and M. Inoue, J. Org. Chem., 1988, 53, 107. 10 T. Ishida, Y. In, M. Doi, M. Inoue, Y. Hamada and T. Shioiri, Biopolymers, 1992, 32, 131. 11 Y. In, M. Doi, M. Inoue and T. Ishida, Acta Crystallogr., Sect. A, 1994, 50, 2015. 12 Y. In, M. Doi, M. Inoue and T. Ishida, Acta Crystallogr., Sect. A, 1994, 50, 432. 13 Y. In, M. Doi, M. Inoue, T. Ishida, Y. Hamada and T. Shioiri, Chem. Pharm. Bull., 1993, 41, 1686. 14 F. J. Schmitz, M. B. Ksebati, J. S. Chang, J. L. Wang, M. B. Hossain and D. D. van der Helm, J. Org. Chem., 1989, 54, 3463. 15 R. Cusack, L. Grøndahl, D. P. Fairlie, G. R. Hanson, T. W. Hambley and L. R. Gahan, manuscript in preparation. 16 R. A. Binstead and A. D. Zuberbühler, SPECFIT, A Program for Global Least Squares Fitting of Equilibria and Kinetic Systems Using Factor Analysis and Marquardt Minimization, Version 2.09, © 1993–1995 Spectrum Software Associates, Chapel Hill, NC, USA. 17 KaleidaGraph (version 3.0.4), graph and fitting program software for Macintosh. 18 E. Aguilar and A. I. Meyers, Tetrahedron Lett., 1994, 35, 2473. 19 H. Kessler, Angew. Chem., Int. Ed. Engl., 1982, 21, 512. 20 V. F. Bystrov, Prog. NMR Spectrosc., 1976, 10, 41. 21 B. E. Campbell, K. R. K. Easwaran, G. C. Zanotti, M. A. Staples, E. T. Fossel and E. R. Blout, Biopolymers, 1986, 25, S47. 22 G. Zanotti, F. Pinnen, G. Lucente, L. Paolillo, G. D’Auria, E. Trivellone and M. D’Alagni, Biopolymers, 1989, 28, 305. 23 L. Grøndahl, A. Hammershøi, A. M. Sargeson and V. J. Thöm, Inorg. Chem., 1997, 36, 5396. 24 Y. Fusaoka, E. Ozeki, S. Kimura and Y. Imanishi, Int. J. Peptide Res., 1989, 43, 104. 25 S. Kimura and Y. Imanishi, Biopolymers, 1983, 22, 2383. 26 E. A. Lance and R. Nakon, Inorg. Chim. Acta, 1981, 55, L1. 27 S. A. Al-Baldawim and T. E. Gough, Can. J. Chem., 1969, 47, 1417. 28 H. Sigel and R. B. Martin, Chem. Rev., 1982, 82, 385. 29 C. D. Richie, J. Am. Chem. Soc., 1969, 91, 6749. 30 D. L. Rabenstein and S. Libich, Inorg. Chem., 1972, 11, 2960. 31 T. C. Woon and D. P. Fairlie, Inorg. Chem., 1992, 31, 4069. 32 R. S. Macomber, J. Chem. Educ., 1992, 69, 375. 33 S. Ahrland, The Chemistry of Non-aqueous solvents, Ed. J. J. Lagowski, Vol. VA, Academic Press, N.Y., 1978, p. 1. Paper 8/08836E

ISSN:1477-9226    

DOI:10.1039/a808836e

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 1235–1240 1235 Tetrahedral d0 and d10 transition metal ions sharing edges in the solid state: electronic structure and bonding Pere Alemany,a Jaime Llanos b and Santiago Alvarez c a Departament de Química Física, Universitat de Barcelona, Diagonal 647, 08028 Barcelona, Spain b Departamento de Química, Facultad de Ciencias, Universidad Católica del Norte, Avda. Angamos 0610, Casilla 1280, Antofagasta, Chile c Departament de Química Inorgànica, Universitat de Barcelona, Diagonal 647, 08028 Barcelona, Spain Received 1st December 1998, Accepted 17th February 1999 The bonding and electronic structure in one- and two-dimensional copper–vanadium sulfides, based on tight binding band calculations, has been studied.Theoretical evidence for the existence of weak donor–acceptor interactions from the d10 copper(I) ion toward the empty d orbitals of VV is discussed. Maximization of the number of d10–d0 interactions explains the structural choice among alternative distributions of the transition metal ions within the sulfide matrix.The double chains of Ba2Cu3VS6 and the layers of KCu2VS4 have an electron deficiency associated with the presence of hypervalent sulfur bridges that slightly enhances the Cu ? ? ? V interaction. The metal–metal interactions between formally closed shell linear and square planar transition metal complexes have been the subject of increasing theoretical interest in recent years.1 In particular, the linear d10 ML2 and square planar d8 ML4 complexes give intermolecular M ? ? ? M contacts through donor–acceptor interactions between a filled d orbital of one metal atom and the empty p orbital of another one.2–5 Although a large number of compounds have been synthesized in which a d10 and a d0 ion are put in close proximity by bridging ligands, a theoretical study of the nature of the d10–d0 interactions has not been reported so far to the best of our knowledge. In this paper we present a semiempirical theoretical study of the electronic structure and bonding in extended structures with edge-sharing tetrahedral copper(I) and vanadium(V) ions 1.The nature of the d10 ? ? ?d0 interactions in these compounds is not expected to be very diVerent from that found in other molecular 6 or extended structures in which the d10 ion can be CuI, AgI or AuI, and the d0 ion MoVI, WVI or VV. The structural motif 1 can be found in tetranuclear 7 2 or pentanuclear 8,9 complexes.A variety of other d10–d0 combinations can be found in polynuclear compounds.6,10 Combinations of d10 and d0 ions also appear forming chains 11–13 in Ba2Cu3S2VS4, K2CuVS4 and Ag2CuVS4, or two-dimensional networks14–16 in KCu2VS4, Na2Cu3VS4 and KCu2NbSe4. Bonding in the binuclear model compound [S2V(Ï-S)2- CuS2]62 As a first approximation to the electronic structure of the chains and layers of edge-sharing CuS4 and VS4 tetrahedra we consider the bonding in the simple molecular anion [S2V(m-S)2- CuS2]62, with structural motif 1, and study the Cu ? ? ? V interaction through an analysis of its molecular orbitals, whose energies are represented in Fig. 1.The s and p orbitals of Cu V Cu V S S S S Cu PPh3 Cu Ph3P Cu Ph3P PPh3 1 2 and V (not represented in Fig. 1) appear at high energy due to their strong antibonding M–S character consistent with sp3 hybrid metal orbitals acting as acceptors toward the sulfido ligands.Among the low lying occupied levels one can identify Fig. 1 Molecular orbital diagram for a model binuclear compound [S2V(m-S)2CuS2]62 1 with the structural parameters from the structure of Ba2Cu3VS6.1236 J. Chem. Soc., Dalton Trans., 1999, 1235–1240 the sulfur 3s and 3p orbitals, from which we single out those incorporating bonding character between the metal atoms and the bridging sulfur atoms, referred to here as framework orbitals. All copper 3d orbitals appear occupied at low energies, whereas empty vanadium 3d orbitals appear at high energies, in keeping with a formal description of their oxidation states as CuI and VV.A small splitting of the d orbitals in t2-like and e-like sets is also observed, as expected for an approximately tetrahedral ligand field. The splitting is inverted for the case of Cu (i.e., the t2 below the e set) because of their lower energy compared to that of the sulfur lone pair orbitals. A small positive Cu–V overlap population (0.0046) is found in our calculations, indicative of a weak bonding interaction that could be attributed to a d10 to d0 electron donation.However, as the Cu and V atoms are held together by the bridging sulfur atoms, the positive overlap population might result from through-bond interaction. Therefore, we need to analyse the bonding relationships in the VS2Cu ring to find out whether the positive overlap population should be attributed to the delocalized bonding in that ring or to a direct Cu–V interaction, or to both.For an idealized structure of the binuclear compound with an M2S2 ring, symmetry labels corresponding to the D2h point group can be used. In that case the four molecular orbitals with M–Sb bonding character (where Sb is a bridging sulfur atom) can be described schematically as in 3. We refer to these as the framework bonding orbitals and use the symbol f to identify them and f* for their antibonding counterparts. Whenever these four orbitals are occupied and their antibonding counterparts empty (i.e.a framework electron count 17,18 of 8, or FEC = 8), they account for the four two-electron bonds of the M2S2 ring. Since one of these orbitals (ag) is metal–metal bonding, another (b1u) metal–metal antibonding across the ring, no net metal–metal bonding interaction exists when FEC = 8. However, for compounds with less electrons (FEC = 6 or 4), a short through-ring metal–metal distance would be favoured, stabilizing the occupied ag and destabilizing the empty b1u orbital.The outcome in that case is a net metal–metal bonding interaction.17,18 Therefore, when analysing the possible existence of a bonding Cu ? ? ? V interaction in the compounds under study, the possibility of electron deficiency cannot be disregarded. For that reason we have explicitly included in the MO diagram (Fig. 1) the framework orbitals 3. In our model heterobinuclear compound all the framework bonding orbitals are occupied (FEC = 8), hence no significant V? ? ? Cu bonding can be expected to arise from the framework interactions.Such qualitative reasoning is confirmed by an analysis of the contribution of each MO to the V–Cu overlap population. The largest contribution comes from the two bonding MOs composed mainly of the copper and vanadium dz2 and dxz orbitals, not from the framework bonding orbitals. Furthermore, if either the copper or the vanadium 3d orbitals are removed from the basis set, the V–Cu overlap 3 ag b3u b1u b2g population becomes negative.In summary, the weak V–Cu bonding interaction results from donor–acceptor interactions between the occupied d orbitals of Cu and the empty d orbitals of V (4). A small additional contribution to V–Cu bonding comes from the b1u framework bonding orbital, which appears to be slightly metal–metal bonding due to mixing of the empty d orbitals (5, allowed by the low symmetry of the Cu–V system 1) while a large part of its metal–sulfur bonding character is preserved.This behaviour diVers from that previously found for similar interactions between two copper(I) ions, in which only copper sp3 hybrids participate in the b1u framework orbital that has the through-ring antibonding character depicted in 3. Electronic structure and bonding in the chain compound K2CuVS4 We look now at the single chains found in the crystal structure 12 of K2CuVS4, which can be derived by fusing together [S2V- (m-S)2CuS2]62 anions.The calculated density of states (DOS) diagram nicely reproduces the level ordering found for the model binuclear compound. In Fig. 2 we show the DOS in the energy window that contains the copper 3d bands (between 213 and 214 eV), the sulfur 3p bands, including the framework bonding orbitals (around 212 eV), and the empty vanadium 3d bands (above 28 eV). Other than the formation of electronic bands, no important changes appear in the Fig. 2 Density of states (DOS) diagram around the Fermi level (horizontal dashed line) for the CuVS4 22 chain in K2CuVS4, calculated with the experimental structural data. The shaded areas represent the contribution of the Cu (left) and V atoms (right) to the total DOS. The integral of such contributions is also represented (vertical dashed lines). 4 5 + .. ..J. Chem. Soc., Dalton Trans., 1999, 1235–1240 1237 electronic structure when moving from the binuclear to the one-dimensional compound.Again, a small positive overlap population (0.0059) is found for each Cu ? ? ? V contact. If calculations are performed for the same chain but replacing the vanadium atoms by copper, a negative Cu ? ? ? Cu overlap population is found, confirming that the positive Cu ? ? ?V overlap population should be attributed to a weak d10–d0 donor–acceptor interaction. A corollary of the additional stability gained by the chain through the Cu ? ? ? V interactions is that the preferred structure of K2CuVS4 should be the one with the maximum number of Cu ? ? ? V contacts.To confirm this, we repeated the calculations for an idealized chain (all M–S distances 2.30 Å) with diVerent distributions of the metal atoms. The case in which two neighbouring Cu atoms are followed by two V atoms in the unit cell (i.e. ? ? ? CuCuVVCuCuV ? ? ? ) is found to be significantly less stable (10 kcal mol21 per formula unit) than the regular chain with alternating Cu and V atoms ( ? ? ? CuVCuVCu ? ? ? ) experimentally found in K2CuVS4.It is thus clear that replacing Cu ? ? ? Cu or V ? ? ? V contacts by Cu ? ? ? V ones enhances the stability of the chain. As in the model binuclear compound studied above, the anionic chains in K2CuVS4 are electron precise, in the sense that all the M–S linkages can be described as two-electron bonds. This means that the electron count for each VS2Cu ring (FEC = 8) precludes a significant degree of Cu ? ? ? V bonding attributable to interaction through the bridges.This can be verified by looking at the COOP curves (Crystal Orbital Overlap Population,19 Fig. 3). There, the Cu–V bonding peak at approximately 213.7 eV [Fig. 3(a)] coincides with bonding dz2(Cu)–dz2(V) and dxz(Cu)–dxz(V) peaks [Fig. 3(b)]. The bonding peak at 212.7 eV corresponds to the ag framework orbital 3, compensated by the antibonding peak of b1u at 211.4 eV. The dz2 metal orbitals interact both through space, as indicated by the bonding peak corresponding to the dz2(Cu) band (213.7 eV), and through the bonds, as seen in the bonding region coincident with the framework bonding orbitals (around 212 eV).Furthermore, omitting the vanadium 3d orbitals from the basis set yields a COOP curve [Fig. 3(c)] in which the bonding characteristics of the copper dz2 and dxz orbitals have been annihilated, thus supporting the existence of a bonding donor– acceptor interaction between the d orbitals of the two metal atoms.Fig. 3 Crystal orbital overlap population (COOP) curve in the region of the framework antibonding orbitals for V ? ? ? Cu contacts (a) and its s and p orbital components (b), calculated for the CuVS4 22 chain in K2CuVS4 using the experimental structural data. Also shown is the COOP curve for the V ? ? ? Cu contacts calculated without the vanadium 3d orbitals (c). The scale for each COOP curve (×102) is indicated at the bottom.Electronic structure and bonding in the double chains of Ba2Cu3VS6 Another one-dimensional compound with edge-sharing vanadium and copper ions is Ba2Cu3VS6, which features double chains shown in Fig. 4, with an ordered distribution of vanadium and copper ions that we schematically represent for the subsequent discussion in 6a. Notice that in such chains there are two diVerent types of sulfur atoms. Those in the centre of the chain, labelled Sc, act as bridges to four metal atoms (m4 co-ordination mode) in a square-pyramidal arrangement 7.In contrast, the external sulfur atoms, Se, are shared by two metal atoms each, in a m co-ordination mode. According to the DOS diagram (Fig. 5), the valence and conduction bands can be described from low to high energy in terms of copper 3d, sulfur 3p (including the framework bonding orbitals), and vanadium 3d orbitals. Although diVerences in bonding between this double chain and the single chain studied above (Fig. 3) are not apparent at first sight, a simple excercise in electron counting indicates that the double and single chains present an important diVerence. As already noticed, the Sc atoms present a square pyramidal environment 7. With such geometry, one of the sulfur sp3 lone pair orbitals is pointing away from the metal atoms, leaving only three lone pairs available to bond to four metal atoms. Since there are four Sc atoms per unit cell, each bearing one non-bonding lone-pair Fig. 4 Structure of the Cu3VS6 42 double chain in Ba2Cu3VS6. The tetrahedra represent the co-ordination sphere of the V atoms, the spheres represent the Cu atoms. Cu V Cu Cu V Cu Cu Cu Cu V Cu V Cu Cu Cu Cu Cu V Cu Cu Cu V Cu Cu V V Cu Cu Cu Cu Cu Cu Cu V V Cu Cu Cu Cu Cu 6d 6a 6c 6e 6b1238 J. Chem. Soc., Dalton Trans., 1999, 1235–1240 orbital, and eight external Se atoms with two outward reaching lone pairs each, there is a total of 28 sulfur valence orbitals per unit cell to construct the framework bonding orbitals (f).But there are a total of 32 M–S linkages per unit cell, four more than the number of available electron pairs. In other words, this compound has an electron deficiency of eight electrons per unit cell with twelve metal–metal contacts. In correspondence, formally only 28 f* orbitals can be formed, while a total of 32 4s and 4p metal orbitals are available for that. As a result, four combinations of the metal s and p orbitals are not allowed by symmetry to interact with the sulfur lone pairs.For instance, the combination of the metal 4s orbitals depicted in 8 cannot mix with the inner sulfur atoms (Sc) at the centre of the Brillouin zone. How is such electron deficiency reflected in the band calculations? For the analysis of the M–S bonding we have built a symmetrized Cu2S3 42 double chain derived from the structure of Cu3VS6 42, replacing all the V atoms by Cu. The density of states (DOS) calculated in the region of the f* levels is presented in Fig. 6(b), together with the M–S COOP curves for the Se and central Sc sulfur atoms. Comparison with the analogous curves for a single chain [Fig. 6(a)] shows two relevant diVerences: on the one hand the number of f* bands is larger due to the doubling of the unit cell, and on the other hand there are two bands that appear at lower energy in the double than in the single chain. The lower energy of these bands is due to their non-bonding character with respect to the M–Sc bonds and little antibonding character with respect to the M–Se bonds, as seen in the COOP curve, in contrast with the clear M–S anti- Fig. 5 The DOS diagram for the valence and conduction bands of the Cu3VS6 42 double chain in Ba2Cu3VS6, calculated with the experimental structure. The shaded areas represent the contribution of the Cu (left) and V atoms (right) to the total DOS. Se Se Sc Se Sc Se Se Sc Se Se Se Sc 8 bonding character of all the f* bands in the single chain.The negligible Sc contribution and the M–Sc non-bonding character of those two bands is in accord with the schematic description in 8. On the other hand, the M–S antibonding character of these bands in the single chain implies some degree of electron transfer from the sulfur lone pairs to the metal atoms, hence partial population of metal–metal antibonding levels. In the double chain, the M–Sc non-bonding nature of the lowest two f* bands implies a lesser population of M–M antibonding orbitals.Consequently, the bonding character of the Cu ? ? ?V interaction is enhanced in the double chain (notice the higher positive value of the overlap population in the double chain compared to that in the single chain, Table 1), and even the Cu ? ? ? Cu interactions become slightly bonding. We must conclude that Ba2Cu3VS6 is analogous to K2CuVS4 in the sense that d10–d0 bonding interactions appear in both compounds, but the electron deficiency present in the former case slightly enhances those bonding interactions. Although the asymmetry of the unit cell in the double chains of the barium salt makes a discussion of the bond distances somewhat cumbersome, the M–Sc distances are slightly longer than in the single chain of K2CuVS4, whereas the M–Se ones are approximately unchanged or even slightly shorter (Table 2), as expected from the qualitative discussion above.We turn now to the colouring problem in Ba2Cu3VS6.In a unit cell containing six copper and two vanadium atoms several distributions of the cations can be foreseen, as sketched in 6. With the simple qualitative idea that Cu ? ? ?V, but not Cu ? ? ? Cu or V? ? ?V, contacts contribute to the stability of the structure, one would easily predict that those arrangements that maximize the Fig. 6 The DOS diagram and COOP curve for the M–S bonds in the region of the framework antibonding orbitals, calculated for a single chain Cu2S4 62 (a) and for a double chain Cu4S6 82 (b).For these calculations the M–S bond distances and S–M–S bond angles were taken as 2.30 Å and 109.58, respectively. Table 1 Calculated overlap populations between the metal atoms for compounds of diVerent dimensionalities in the experimental structures (the corresponding metal–metal distances, in Å, are given in parentheses), except for the model molecular anion [S2V(m-S)2CuS2]62 for which the structure of a portion of the [Cu3VS6]42 chain was adopted Compound [S2V(m-S)2CuS2]62 [CuVS4]22 [Cu3VS6]42 [Cu2VS4]2 Dimensionality Molecule Single chain Double chain (parallel) (perpendicular) Layers Cu ? ? ?V 0.0046 (2.747) 0.0059 (2.719) 0.0128 (2.655) 0.0140 (2.737) 0.0158 (2.704) (2.691) Cu ? ? ? Cu 0.0033 (2.732) 0.0070 (2.979) 20.0033 (3.668)J.Chem. Soc., Dalton Trans., 1999, 1235–1240 1239 Table 2 Structural data a for compounds of diVerent dimensionalities with edge-sharing tetrahedra of V and M (M = Cu or Ag) Compound K3VS4 TlVS4 [Cu3(PPh3)4VS4] (2) [Cu4(SPh)3(dtc)VS4]2 K2CuVS4 K2AgVS4 Rb2AgVS4 Ba2Cu3S2VS4 KCu2VS4 Cu3VS4 Dimensions 00001111 23 M–V 2.626–2.790 2.599–2.653 2.719 2.904 2.810 2.655 2.735 b 2.692–2.704 2.697 M–Sc 2.318–2.525 2.298–2.300 2.299 M–Se 2.211–2.349 2.268–2.276 2.313 2.515 2.513 2.291–2.334 2.292–2.300 V–Sc 2.195 2.233 2.219 V–Se 2.053–2.163 2.171 2.148–2.226 2.186–2.216 2.177 2.178 2.177 2.164 2.146–2.192 Ref. 12, 20 21 78 12 13 13 11 14,22 23,24 a All distances in Å; Sc indicates a m4 atom, Se a non-bridging, m or m3 atom. b Connected through two m4 bridges.number of Cu ? ? ?V, contacts, i.e. 6a–6c will be the most stable ones. Such qualitative expectations are supported by calculations with a symmetrized structure (Table 3). Even if one cannot fully rely on the small energy diVerences given by the EHTB (Extended Hückel Tight Binding) calculations on a model system with fixed geometry to predict the relative stability of structures 6a–6c, it is clear that structures 6d and 6e should be expected to be significantly less stable.Structure 6b, which seems a reasonable alternative to the experimental structure 6a on energetic grounds, is likely to be less stable due to the strain introduced in a linear chain by the shorter V–S bonds compared to the Cu–S ones, a fact that has been disregarded in our model calculations by assuming identical Cu–S and V–S bond distances. Band electronic structure of the layered compound KCu2VS4 The structure of KCu2VS4 can be described as formed by perpendicular chains of edge-sharing copper and vanadium tetrahedra (Fig. 7). The vanadium ions are connected to two Fig. 7 Structure of the Cu2VS4 2 layers in KCu2VS4. The tetrahedra represent the coordination sphere of the V atoms, the spheres represent the Cu atoms. Table 3 Relative energies (kcal mol21) and sum of all M ? ? ? M and M–S (M = Cu or V) overlap populations in the unit cell for diVerent arrangements of the metal atoms in the lattice of Ba2Cu3VS4 Structure 6a 6b 6c 6d 6e Energy 0.0 21.1 2.4 17.2 17.2 M? ? ?M 0.1310 0.1344 0.1304 0.1006 0.1207 M–S 10.4292 10.4487 10.4193 10.3642 10.2583 Cu1 ones along the c axis through opposite edges, forming linear chains, and to the Cu2 ions along the a axis through neighbouring edges, resulting in a zigzag chain.Each Cu atom is connected to two V atoms through opposite edges of the CuS4 tetrahedron. There are three types of sulfide ions: S2 atoms are bridging one V and one Cu1; S3 atoms are bridging one V, one Cu1 and one Cu2 atom; S1 presents an unusual umbrella-shaped co-ordination (9), with one V atom in the handle, two Cu2 and one Cu1 atom in the ribs.We note in passing that the same geometry is found for the sulfide ions in the three-dimensional structure of Cu3VS4. In what follows we will refer to the m and m3 sulfur atoms (S2 and S3, respectively) as Se, and to the m4 atoms (S1) as Sc. Again, the rule that the number of Cu ? ? ? V contacts is maximized seems to apply to this compound, as would be expected if weakly attractive d10–d0 interactions exist between the two types of metal atoms.On the other hand, some degree of electron deficiency exists in this compound also. Half of the sulfur atoms in this structure (Sc) are bridging one vanadium and three copper ions in an umbrella-like fashion, thus leaving one unshared sulfur lone pair directed away from the layer.With two formula units per unit cell, two of the f* bands are non-bonding with respect to the umbrella sulfur atoms for a total of 24 linkages and 8 metal–metal contacts, resulting in an eVective FEC of 7.5. It is interesting that in this structure all the V–S–Cu bond angles are rather small (ª738), whereas the noncyclic Cu–S–Cu bond angles are ª1078. Such small angles might be taken as indication that there is an electronic preference for a short V ? ? ? Cu contact, also reflected in slightly shorter Cu ? ? ? V distances (2.692 and 2.704 Å) than in the single chains of K2CuVS4 (2.719 Å).The DOS diagram for the valence and conduction bands of this layered compound (Fig. 8) appears to be quite similar to those of the single and double chain compounds (Figs. 2 and 5, respectively). The COOP curve for the Cu ? ? ? V contacts (not shown here) presents the same features previously found for the one-dimensional chains indicating the existence of donor–acceptor interaction between the d orbitals of the two metals.Concluding remarks In the model molecular anion [S2V(m-S)2CuS2]62 and in the single chains [V(m-S)2Cu(m-S)2]• 22 of K2CuVS4 the overlap populations and orbital analyses indicate the existence of1240 J. Chem. Soc., Dalton Trans., 1999, 1235–1240 weak donor-acceptor d10–d0 interactions between the Cu and V atoms. In the double chains of Ba2Cu3S2VS4 the d10–d0 interactions coexist with electron deficiency in the framework bonding bands that contributes to stronger Cu ? ? ? V and weaker M–S interactions.Ab initio theoretical studies on related molecular systems with MoVI and CuI, AgI or AuI seem to con- firm the existence of weakly bonding d10–d0 interactions.10 An indirect evidence of the existence of stabilizing Cu ? ? ?V interactions is provided by the higher calculated stability of those structures that maximize the number of Cu ? ? ?V contacts, as experimentally found in the chain and layered compounds studied in this paper as well as for related extended structures with d10–d0 contacts, including compounds of different structural dimensionalities, from molecular species to the three-dimensional network of Cu3VS4. Appendix Molecular orbital and bond structure calculations presented in this work have been made with the extended Hückel method25–27 using the modified Wolfsberg–Helmholz formula 28 for the evaluation of the oV-diagonal elements of the Hamiltonian matrix.The atomic parameters adopted in these calculations are shown in Table 4. For extended systems, numerical integrations over the irreducible wedge of the Brillouin zone have been performed using a 101 k-point mesh for 1-D chains in K2CuVS4 and in Ba2Cu3VS6, and a 100 k-point mesh for the layers in KCu2VS4. Fig. 8 The DOS diagram for the valence and conduction bands of the Cu3VS6 42 layer in Ba2Cu3VS6. The shaded areas represent the contribution of the Cu (left) and V atoms (right) to the total DOS, and the dashed line indicates the Fermi level.Table 4 Valence shell ionization potentials (Hii), orbital exponents (zij), and combination coeYcients (cj) used for the extended Hückel calculations Atom S Cu V Orbital 3s 3p 4s 4p 3da 4s 4p 3db Hii/eV 222.761 212.081 28.345 24.216 213.162 26.850 23.910 28.181 zii 2.122 1.827 2.200 2.200 5.950 1.300 1.300 4.750 a c1 = 0.5933, zi2 = 2.300, c2 = 0.5744. b c1 = 0.4755, zi2 = 1.700, c2 = 0.7050.The geometry for the binuclear model compound CuVS6 discussed in the first section has been taken from the experimental structure data found for Ba2Cu3VS6. The geometries for chains, double chains, and 2-D layers have been taken from the experimentally determined structures of K2CuVS4, Ba2Cu3VS6, and KCu2VS4, respectively. For the analysis of the coloring problem in the chain compounds we have considered idealized structures in which all metal–sulfur distances are set to 2.3 Å and all metal atoms are supposed to have a perfect tetrahedral co-ordination environment.Acknowledgements Financial support to this work has been provided by Dirección General de Enseñanza Superior (Spain), project PB95-0848- C02-01 and Fonds Desarrollo Cientifico y Tecnológica (Chile), grant 1960372. S. Alvarez is grateful for a Visiting Professorship funded by Comisión Nacional de Investigación Cientifica y Tecnológica (Chile). The computing resources at the Centre de Supercomputació de Catalunya (CESCA) were funded in part through a grant by Fundació Catalana per a la Recerca and Universitat de Barcelona.References 1 P. Pyykkö, Chem. Rev., 1997, 97, 597 and refs. therein. 2 J. J. Novoa, G. Aullón, P. Alemany and S. Alvarez, J. Am. Chem. Soc., 1995, 117, 7169. 3 G. Aullón, P. Alemany and S. Alvarez, Inorg. Chem., 1996, 35, 5061. 4 G. Aullón and S. Alvarez, Chem. Eur. J., 1997, 3, 655. 5 X.-Y. Liu, F. Mota, P. Alemany, J.J. Novoa and S. Alvarez, Chem. Commun., 1998, 1149. 6 W.-W. Hou, X.-Q. Xin and S. Shi, Coord. Chem. Rev., 1996, 153, 26. 7 A. Müller, J. Schimanski and H. Bogge, Z. Anorg. Allg. Chem., 1987, 544, 107. 8 Y. Yang, Q. Liu, L. Huang, D. Wu, B. Kang and J. Lu, Inorg. Chem., 1993, 32, 5431. 9 Q. Liu, Y. Yang, L. Huang, D. Wu, B. Kang, C. Chen, Y. Deng and J. Lu, Inorg. Chem., 1995, 34, 1884. 10 F. Mota, X.-Y. Liu, C. Gimeno, A. Laguna and S. Alvarez, to be submitted. 11 C. Mujica, C. Ulloa, J. Llanos, K. Peters, E.-M. Peters and H. G. von Schnering, J. Alloys Compds., 1997, 255, 227. 12 P. Dürichen and W. Bensch, Eur. J. Inorg. Solid State Chem., 1996, 33, 309. 13 W. Bensch and P. Dürichen, Chem. Ber., 1996, 129, 1207. 14 K. Peters, E.-M. Peters, H. G. von Schnering, C. Mujica, G. Carvajal and J. Llanos, Z. Kristallogr., 1996, 211, 812. 15 C. Mujica, J. Llanos, G. Carvajal and O. Wittke, Eur. J. Solid State Inorg. Chem., 1996, 33, 987. 16 Y. Lu and J. Ibers, J. Solid State Chem., 1991, 94, 381. 17 P. Alemany and S. Alvarez, Inorg. Chem., 1992, 31, 4266. 18 G. Aullón, P. Alemany and S. Alvarez, J. Organomet. Chem., 1994, 478, 75. 19 T. Hughbanks and R. HoVmann, J. Am. Chem. Soc., 1983, 105, 3528. 20 J. M. van den Berg and R. de Vries, K. Ned. Akad. Wet., Ser. B: Phys. Sci.: Proc., 1964, 67, 178. 21 M. Vlasse and L. Fournes, C.R.H. Acad. Sci., Ser. C, 1978, 287, 47. 22 W. Bensch, P. Dürichen and C. Weilich, Z. Kristallogr., 1996, 211, 933. 23 F. E. Riedel, W. Paterno and W. Erb, Z. Anorg. Allg. Chem., 1977, 437, 127. 24 C. Mujica, G. Carvajal, J. Llanos and O. Wittke, Z. Kristallogr., 1998, 213, 12. 25 R. HoVmann and W. N. Lipscomb, J. Chem. Phys., 1962, 36, 2179. 26 R. HoVmann, J. Chem. Phys., 1963, 39, 1397. 27 M.-H. Whangbo, R. HoVmann, J. Am. Chem. Soc., 1978, 100, 6093. 28 J. H. Ammeter, H.-B. Bürgi, J. C. Thibeault and R. HoVmann, J. Am. Chem. Soc., 1978, 100, 3686. Paper 8/09369E

ISSN:1477-9226    

DOI:10.1039/a809369e

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 1241–1248 1241 Relative basicities of the oxygen atoms of the Linquist polyoxometalate [Mo6O19]22 and their recognition by hydroxyl groups in radical cation salts based on functionalized tetrathiafulvalene � donors Anne Dolbecq,a Aude Guirauden,a Marc Fourmigué,a Kamal Boubekeur,a Patrick Batail,*a Marie-Madeleine Rohmer,b Marc Bénard,b Claude Coulon,c Marc Sallé d and Philippe Blanchard d a Institut des Matériaux de Nantes, UMR 6502 CNRS-Université de Nantes, BP 32229, 2, rue de la Houssinière F-44322 Nantes Cedex 03, France.E-mail: batail@cnrs-imn.fr b Laboratoire de Chimie Quantique, UMR 7551 CNRS-Université Louis Pasteur, F-67000 Strasbourg, France c Centre de Recherches P. Pascal (CRPP), CNRS, av. Dr. Schweitzer, F-33600 Pessac, France d IMMO, UMR CNRS-Université d’Angers, 2, bd. Lavoisier, 49045 Angers, France Received 3rd December 1998, Accepted 10th February 1999 Electrocrystallization of three hydroxylated donor molecules derived from tetrathiafulvalene (TTF) or ethylenedithiotetrathiafulvalene (EDT-TTF), i.e.Me3TTF-CH2OH, EDT-TTF(CH2OH)2 and TTF(CH2OH)4, in the presence of [n-Bu4N1]2[Mo6O19]22 aVorded 2 : 1 cation radical salts, [donor1?]2[Mo6O19]22, whose crystal structures have been solved by X-ray diVraction. In the three diVerent salts complex hydrogen bond networks develop in the solid state where the oxygen atoms of both the hydroxyl groups and the [Mo6O19]22 anions act as hydrogen bond acceptors.The observed hydrogen bonding directed toward one surface, bridging oxygen atom of [Mo6O19]22 is rationalized by an analysis of ab initio calculations of the distribution of electrostatic potentials. Introduction Electroactive molecular materials based on polyoxometalates with a variety of sizes, shapes and spin states have been actively investigated in the past ten years and this topic has been recently reviewed.1 For example, following our report 2 on [BEDT-TTF]8[SiW12O40], Coronado and co-workers 3 described a series of salts, isostructural to the former, and where the delocalized electrons of the conducting mixed-valence BEDTTTF slabs coexist with the localized magnetic moments of the paramagnetic anions [XW12O40]n2, X = CuII (n = 6), CoII (n = 6) or FeIII (n = 5).3 In the latter the magnetic atom hidden within the core of the heteropolyanion does not interact with the conducting p slabs.In the series 4 [BEDT-TTF]8[Xn1Zm1- (H2O)M11O39](12 2 n 2 m)2 (X = PV or SiIV; M = MoVI or WVI; Z = FeIII, CrIII, MnII, .. ., etc.), where the magnetic center is introduced on the outer shell of the polyanion, polymerized polyanion networks or positional disorder of the paramagnetic Z center are observed with no evidence for any sizeable magnetic interaction between the cationic slabs and the anions. This appears to be a general feature of all such organic– inorganic constructions associating “classical” organic p-donor molecules, i.e.TTF, TMTTF, TMTSF, BEDT-TTF, BEDSTTF or BET-TTF† and polyoxoanions such as [Mo6O19]22.5,6 Several tetrathiafulvalene-based molecules bearing competing hydrogen bond donor and/or acceptor capabilities were recently described and engaged in cation radical salts. Examples include alcohols 7,8 in Me3TTF(CHMeOH) and † TTF = Tetrathiafulvalene; TMTTF = tetramethyltetrathiafulvalene; TMTSF = tetramethyltetraselenafulvalene; BEDT-TTF = bis(ethylenedithio) tetrathiafulvalene; BEDS-TTF = bis(ethylenedithio)tetraselenafulvalene and BET-TTF = bis(ethylenethio)tetrathiafulvalene. EDT-TTF(CH2OH), phosphonate anions 9 in Me3TTF-PO3H2 and the zwitterionic [Me3TTF-PO3H2]1? and thioamides 10 in TTF-CSNHMe.The architectures of the salts based on these multifunctional redox precursors result from a delicate yet eVective balance between the requirements of the mode of overlap of the open-shell conjugated p systems and those of the (transverse) hydrogen bonds directed towards the inorganic anions.Such principles are further illustrated and rationalized in this paper where novel constructions based on the hydroxymethyl- functionalized TTF and EDT-TTF cores and the Linquist anion [Mo6O19]22 are discussed. Electrocrystallization 11 of the mono-, di- and tetrahydroxylated donor molecules Me3TTF(CH2OH), EDTTTF( CH2OH)2 and TTF(CH2OH)4 in the presence of [n- Bu4N1]2[Mo6O19]22 led to 2 : 1 salts, [donor]2[Mo6O19], whose crystal structures, hydrogen bonding pattern and magnetic properties are described and analysed in this paper.The capacity of [Mo6O19]22 to act as a hydrogen bond acceptor will be demonstrated by electrostatic potential distribution (ESP) calculations, conducted on [Mo6O19]22 and its hypothetical protonated form, [HMo6O19]2. S S S S OH S S S S OH HO S S S S S S OH HO OH HO Me3TTF(CH2OH) EDT-TTF(CH2OH)2 TTF(CH2OH)41242 J. Chem. Soc., Dalton Trans., 1999, 1241–1248 Results The parent hydroxymethyl substituted tetrathiafulvalene, TTF-CH2OH, was first described by Green 12 as an unstable compound prepared by reduction of the corresponding aldehyde, TTF-CHO.We reported that methyl substituted tetrathiafulvalenes such as Me3TTF aVorded more stable derivatives and could easily be lithiated as TTF itself for further derivatization. 7 Accordingly, Me3TTF was lithiated with lithium diisopropyl amide and treated with N-methylformanilide to aVord the corresponding aldehyde in good yield.Reduction with NaBH4 in EtOH gives Me3TTF(CH2OH).13 The 3,4-(ethylenedithio)- 3,4-bis(hydroxymethyl)tetrathiafulvalene EDT-TTF- (CH2OH)2 and the tetrakis(hydroxymethyl)tetrathiafulvalene TTF(CH2OH)4 were analogously obtained by reduction of the corresponding di- and tetra-aldehyde, as previously described.14,15 Owing to the limited solubility of those hydroxylmethylated donor molecules in common organic solvents, electrocrystallizations had to be conducted in specific solvents or solvent mixtures (PhCN, MeCN–EtOH or DMF) at higher temperatures (see Experimental section). [Me3TTF-CH2OH]2[Mo6O19] As shown in Fig. 1, the [Mo6O19]22 anions are located on inversion centers at the edges of the triclinic unit cell while an inversion centered organic dimer sits at the center of the unit cell, a structural organization which can be described as CsCl type, observed in cation radical salts of organic donors with similar large all-inorganic anions such as [Re6S5Cl9]2 or [Mo6Cl8(NCS)6]22, in [TMTSF]2[Re6S6Cl8] 16 or [TTF(SMe)4]2- [Mo6Cl8(NCS)6].17 One should note that the bridging oxygen atoms in [Mo6O19]22 are not equidistant from the flanking Mo atoms and O–Mo bond distances cluster into two groups, around 1.85–1.90 and 1.96–1.99 Å, respectively.This distortion, reported in the first structural determination of the [Mo6O19]22 anion,18 is observed here and in the following two structures.Within the organic dimer, each molecule is oxidized to the cation radical and adopts a characteristic boat conformation with a limited folding of the dithiole rings along the S ? ? ? S axis, by 5.1(1)8 along S1 ? ? ? S2 and 2.0(1)8 along S3 ? ? ? S4, a short plane-to-plane intermolecular distance, 3.272(4) Å, and a nearly eclipsed conformation. The oxygen atom of the hydroxyl group is found to be disordered on four diVerent positions associated with three carbon atoms (C7, C8, C9).As a consequence, the hydroxyl hydrogen atom could not be located on Fig. 1 A view of the unit cell of [Me3TTF(CH2OH)]2[Mo6O19]. The four positions occupied by the disordered O atom have been darkened. Fourier diVerence maps and the analysis of hydrogen bonding in this salt will only rely on short intermolecular O ? ? ?O distances. Indeed, as shown in Table 1, each of the four partially occupied hydroxyl sites is found to be engaged in a short Odonor ? ? ?Oanion contact, between 2.78 and 3.05 Å, with the terminal and bridging oxygen atoms of the [Mo6O19]22 polyanion only, namely O3 and O13, O32, respectively while no short Odonor ? ? ?Odonor short contacts could be identified.Those donor–anion close contacts can be compared for example with a similar hydrogen bond interaction identified between the droxyl group of EDT-TTF(CH2OH) and the perrhenate anion ReO4 2 where a single O ? ? ? O contact was observed at the longer 3.33 Å distance.8 The [Me3TTF-CH2OH1?]2 dimers are not fully isolated from each other but form weakly interacting stacks along the [100] direction.Indeed, calculated interaction energies 19 (from extended Hückel calculations with doublez orbitals for C, S and O) between two neighboring Me3TTFCH2OH1? cation radicals amount to 0.89 eV for the interaction within the dimer and 0.21 eV for the interaction between dimers. Therefore, the two radical species are associated in the strongly stabilized bonding combination of the HOMO of each Me3TTF-CH2OH molecule and one expects the salt to be diamagnetic.A unique resonance line centered at g = 2.0006 is however observed by EPR spectroscopy on a single crystal, a value characteristic of TTF-based organic radicals, and attributed to magnetic defects. Its linewidth (DH) decreases from 7 G at room temperature to 4 G at 250 K and is constant down to 4 K, while the temperature dependence of the integrated spin susceptibility shows a Curie–Weiss behavior with a small deviation at higher temperatures, as confirmed by the SQUID susceptibility temperature dependence, well represented by the Curie–Weiss law, c = c0 1 [C/(T 2 q)], with C = 0.055 cm3 K mol21 and q = 22 K (Fig. 2). [EDT-TTF(CH2OH)2]2[Mo6O19] This salt crystallizes in the monoclinic system, space group P21/c, and the asymmetric unit consists of one [Mo6O19]22 polyanion, in general position, and two independent (A and B) EDT-TTF(CH2OH)2 molecules.Note that one of every two hydroxyl groups is found to be disordered on two positions. Thus, and despite the low-temperature X-ray data collection, hydrogen atoms could not be identified on the Fourier difference map. The two organic molecules organize into dimers Fig. 2 Temperature dependence of the magnetic susceptibility of [Me3TTF(CH2OH)]2[Mo6O19]. Table 1 Shortest intermolecular Odonor ? ? ?Oanion distances (Å) in [Me3TTF(CH2OH)]2[Mo6O19] O7A? ? ? O13I O7B ? ? ? O1III O9 ? ? ? O32 2.78(2) 3.05(2) 2.90(1) O7B ? ? ? O3II O8 ? ? ? O3II 2.92(2) 2.85(1) I 11x, y, z; II x, y, 21 1 z; III 1 2 x, 2 2 y, 1 2 z.J.Chem. Soc., Dalton Trans., 1999, 1241–1248 1243 (Fig. 3), with a short plane-to-plane distance (3.31 Å) and a folding [along S1A ? ? ? S2A 6.6(3), S3A ? ? ? S4A 7.5(1), S1B ? ? ? S2B 5.6(3) and S3B ? ? ? S4B 5.2(1)8] of the dithiole rings away from each other (boat conformation). This structure, observed above in [Me3TTF(CH2OH)]2[Mo6O19], is characteristic of fully oxidized, dicationic TTF dimers, as confirmed by the calculated interaction energy between the HOMOs of the two molecules, to 0.8 eV.No EPR signal was observed for this salt, a consequence of the strongly dimerized structure and the absence of any sizeable interdimer overlap interaction. In the solid state the segregation of the hydrophilic CH2OH and hydrophobic SCH2CH2S outer ends of the organic donor molecules leads to a pattern of segregated slabs, a motif already encountered with phosphonate-substituted tetrathiafulvalenes.9 A complex hydrogen bond network develops within the hydrophilic layer and involves the hydroxyl groups of the donor molecules as well as oxygen atoms of the [Mo6O19]22 anion.Short O ? ? ? O contacts are indeed identified within dimers, with the neighboring ones and with oxygen atoms of the polyanion (Table 2). Note that the shortest Odonor ? ? ?Oanion distances Fig. 3 Projection along the b axis of [EDT-TTF(CH2OH)2]2[Mo6O19] showing the segregation of the disordered hydroxyl groups into layers.Fig. 4 A view of the mixed organic–inorganic plane in [EDTTTF( CH2OH)2]2[Mo6O19]. involve the bridging O35 atom, in contrast with the longer distances observed with the terminal O3 and O5 atoms. Within a layer (Fig. 4), [Mo6O19]22 dianions and orthogonal dicationic organic dimers alternate and the whole structure can be described as belonging to the NaCl structural type.[TTF(CH2OH)4]2[Mo6O19] Considering that the single hydroxyl group appears disordered on three diVerent methyl carbon atoms in [Me3TTF- (CH2OH)]2[Mo6O19], it was anticipated that an ordered, eventually isostructural salt could be obtained by engaging the tetrakis(hydroxymethyl)tetrathiafulvalene TTF(CH2OH)4 with the same counter anion. Although electrocrystallization experiments with this tetrahydroxylated donor molecule proved diYcult because of its very limited solubility in common solvents, a few crystals were obtained in DMF.The salt crystallizes in the triclinic system, space group P1� . One [Mo6O19]22 anion and two crystallographically independent radical cations are located on inversion centers (Fig. 5). The hydroxyl groups are now fully ordered and hydrogen atoms were unambiguously identified in Fourier diVerence maps, allowing for a precise analysis of the hydrogen bonding pattern. Indeed, as shown in Table 3, each of the four independent H atoms of the hydroxyl groups is engaged in one and only one hydrogen bond.Note also that the O ? ? ? O distances are comparable to those revealed in the two former structures, confirming the hydrogen bond character of O ? ? ? O interactions identified above in [Me3- Fig. 5 A projection of the structure of [TTF(CH2OH)4]2[Mo6O19] along the a axis. Both [Mo6O19]22 and TTF(CH2OH)4 1? moieties are located on inversion centers, at the corners of the cell and the center of the edges respectively.Note the hydrogen bond H4B ? ? ? O4B cyclic motif as well as the H5B ? ? ? O21 donor–polyanion hydrogen bond. Table 2 Shortest intermolecular O ? ? ? O distances (Å) in [EDTTTF( CH2OH)2]2[Mo6O19] Intradimer interactions O1A1 ? ? ?O1B1 2.60(2) O1A2 ? ? ?O1B1 2.85(2) Interdimer interactions O2A? ? ? O2BI O1B2 ? ? ? O2BII 2.66(1) 2.69(2) O2A? ? ?O2AI O1A1 ? ? ? O2BIII 2.84(2) 2.86(2) Dimer–Mo6O19 interactions O1B2 ? ? ? O35 O1A1 ? ? ? O3 2.70(2) 2.89(2) O1B1 ? ? ? O35 O1B2 ? ? ? O5 2.82(2) 2.96(2) I 12x, 21 2 y, 2z; II 1 2 x, 2y, 2z; III x, 2��� 2 y, ��� 1 z.1244 J.Chem. Soc., Dalton Trans., 1999, 1241–1248 TTF(CH2OH)]2[Mo6O19] and [EDT-TTF(CH2OH)2]2[Mo6O19]. Three hydrogen atoms are engaged in a bond with oxygen atoms of the CH2OH groups while one and only one out of the four becomes bonded to a bridging oxygen atom (O21) of the polyanion. As shown in Fig. 5, molecules B are engaged in an inversion centered O4B–H4B ? ? ? O4B9 motif which develops along the c axis while hydrogen atom H5B is engaged in a hydrogen bond with the bridging O21 atom of the [Mo6O19]22 anion.Fig. 6 shows the organic slab which develops perpendicular to the b 1 c direction where molecules A are connected to each other along the a axis through the O4A– H4A? ? ? O5A hydrogen bond while the H5A atom links those columns with B molecules. Of particular note is the cooperative eVect associated with all those hydrogen bonds as observed here in two extended motifs, a cyclic one ? ? ? O4B– H4 ? ? ? O4B9–H4B9 ? ? ? (Fig. 5) and a linear one O4A– H4A? ? ? O5A–H5A ? ? ? O5B–H5B ? ? ? O21anion (Fig. 6). This co-operativity, often encountered with hydroxyl groups and particularly in carbohydrates,20 has been shown substantially to increase the stability of such arrays. Discussion It is striking that in both [Me3TTF(CH2OH)]2[Mo6O19] and [EDT-TTF(CH2OH)2]2[Mo6O19] the organic p donor radical cations are strongly dimerized with intermolecular S ? ? ? S distances as short as 3.32(2) Å and the hydroxyl groups disordered onto several positions with no clear hydrogen bond pattern.By contrast, in [TTF(CH2OH)4][Mo6O19] the tetrahydroxylated cation radicals are orthogonal to each other and do not overlap. Instead, a precise, ordered, O–H ? ? ? O hydrogen bond network is identified and connects the alternating organic and inorganic ions. Thus, from two principal competing interactions, namely (i) the overlap interaction of open-shell cation radicals which Fig. 6 A view along the b 1 c direction of the donor plane showing the H4A ? ? ? O5A and H5A ? ? ? O5B hydrogen bond network. Table 3 Hydrogen bond geometrical characteristic OH)4]2[Mo6O19] O5A–H5A ? ? ? O5BI O4A–H4A ? ? ?O5AII O5B–H5B ? ? ? O21III O4B–H4B ? ? ? O4BIV O? ? ? O/Å 2.738(5) 2.778(5) 2.873(6) 2.895(5) H? ? ? O/Å 1.934 1.977 2.058 2.137 O–H? ? ? O/8 166.4 165.3 173.1 153.7 I 11x, y, z; II 1 2 x, 2y, 1 2 z; III 21 2 x, 1 2 y, 1 2 z; IV 2x, 1 2 y, 1 2 z.stabilizes the formation of strongly associated, eclipsed donor dimers, and (ii) the directional O–H ? ? ? O hydrogen bonds, the electronic stabilization, and dimer formation, clearly prevails over the hydrogen-bond interactions when solely one and two hydroxyl groups are present on the TTF core. A similar dimer formation was also observed in [TMTTF]2[Mo6O19],6 where no hydroxyl groups are present. This behaviour can be tentatively rationalized by a simple estimation of the energies involved in those two competing interactions.Indeed the O–H ? ? ?O hydrogen bond energy21 is estimated between 15 and 20 kJ mol21, i.e. between 0.15 and 0.20 eV, to be compared with the overlap stabilization of the radical cation within a dimer, which amounts to 0.35 to 0.5 eV per molecule. It therefore appears that the stabilization energy provided by one or two hydrogen bonds cannot compete favorably with the dimer formation.Conversely, the set of four ordered hydrogen bonds per donor molecule identified in [TTF(CH2OH)4]2[Mo6O19] provides suYcient intermolecular interaction energy to prevent the formation of discrete dimers. In two of the three structures described above the oxygen atoms of the hydroxyl groups were found to be disordered on several positions. Beside the subsequent practical diYculty in locating the hydrogen atoms in Fourier diVerence maps, the occurrence of significant disorder indicates that the oxygen atoms of the [Mo6O19]22 anion exhibit no particularly strong hydrogen bond acceptor character.This was deduced earlier by Barcza and Pope22 from their investigation of the stability constants in solution of 1 : 1 adducts of several polyanions with the chloral hydrate [Cl3CCH(OH)2], a strong hydrogen bond donor. In [TTF(CH2OH)4]2[Mo6O19] where no disorder is encountered and where hydrogen atoms were unambiguously identified, one of the four independent H atoms was however found to be engaged in a hydrogen bond with [Mo6O19]22 while the three others are hydrogen-bonded to oxygen atoms of the CH2OH groups, as observed in the crystal structures of a variety of alcohols.In the competition between hydroxylic and [Mo6O19]22 oxygen atoms the former are clearly better hydrogen-bond acceptors. Nevertheless, polyoxoanions prove to be able also to engage in hydrogen bonds.23 The two diVerent kinds of oxygen atoms at the surface of the [Mo6O19]22 polyanion, namely the bridging and terminal ones, do not behave similarly since the shortest hydrogen bonds identified in the three structures involve almost exclusively bridging oxygen atoms.Indeed as reported by Barcza and Pope,22 “the terminal oxygen atoms are strongly polarized toward the interior of the anion by p bonding to the metal atoms”. Similar observations on the relative basicities of oxygen sites in polyanions were derived from ab initio studies 24 of a variety of polyoxometalates such as [V10O28]62 and its protonated derivative [H3V10O28]32.25 In order to rationalize the above experimental observations, the distribution of electrostatic potentials (ESP) was determined by ab initio calculations on the [Mo6O19]22 anion.19 The electrostatic potential distribution has been displayed in two diVerent sections of the anion, one containing four Mo, four bridging, four terminal as well as the central O atoms (Fig. 7, top) and one containing three bridging and three terminal O atoms (Fig. 7, bottom). It is clearly seen that the deepest minima in any of these planes are located in the neighborhood of the bridging O atoms, at an altitude of about 1.2 Å from those atoms in the first plane (Fig. 7, top). Therefore, one expects that the protonation of [Mo6O19]22 would specifically occur on those bridging O atoms. If we now consider the possible localization of a hydrogen bond donor at the proximity of the anion surface, the situation is much less clear-cut.Considering that the observed (O)H ? ? ?Oanion distances amount to ª2 Å, the former deep potential minima are probably not so pertinent anymore. The electrostatic potential distribution has thus been calculated in a plane parallel to the triangular face of the [Mo6O19]22 anion, at respectively 1.1 (Fig. 8, top) and 2.2 Å (Fig. 8, bottom) awayJ. Chem. Soc., Dalton Trans., 1999, 1241–1248 1245 from the surface.While the three deep minima are indeed found at a short distance (1.1 Å) from the surface on the three bridging oxygen atoms, they merge into a deep shallow zone delocalized between the three bridging O atoms at larger distances (2.2 Å) from the surface. As a consequence, we expect weakly acidic hydrogen atoms approaching a face of the anion to be mainly attracted by an extended array rather than by a specific O bridging atom. It is also of interest to evaluate the eVect of an approaching H atom on the electrostatic potential distribution of the anion.Similar calculations were conducted on the hypothetical [HMo6O19]2 anion with the H atom co-ordinated to a bridging O atom at a 1.1 Å distance and located in a plane containing four Mo atoms. As shown in Fig. 9, protonation of O7 strongly modifies the electrostatic potentials of the neighboring O atoms. It particularly decreases the minima of the bridging O6 and O8 atoms (Fig. 9, top) while the O5 and O13 atoms Fig. 7 Sections of the electrostatic potentials for [Mo6O19]22, in the plane containing four Mo, four terminal and four bridging O atoms (top) and in the plane containing three terminal and three bridging O atoms (bottom). Lowest contour: 20.27835 hartree (top), 20.26646 hartree (bottom). First contour interval: 5 × 1025 hartree (top), 4 × 1025 hartree (bottom). Successive contour intervals increased by a factor of 2 in both sections. pertaining to the same triangular face are much less aVected (Fig. 9, bottom). Note also that a secondary minimum now appears on the terminal O18 atom also pertaining to the same triangular face (Fig. 9, bottom). This eVect is more clearly seen while going away from the triangular face of the anion as described above for [Mo6O19]22. Indeed, as shown in Fig. 10 when moving away from the same triangular face described in Fig. 9 (bottom), the minima associated with the bridging O5, O13 atoms and with the terminal O18 atom are both clearly identified at 1.1 Å from the surface (Fig. 10, top) while at 2.2 Å the strongest minimum is associated with the terminal O18 atom (Fig. 10, bottom). We can therefore conclude from those calculations that (i) the bridging O atoms are definitely the most basic ones in [Mo6O19]22, (ii) hydrogen bond donors experience at 2–2.2 Å from the oxide surface a shallow negatively charged zone located at the center of the triangular faces of the anion and (iii) protonation of the anion strongly modifies the potential Fig. 8 Sections of the electrostatic potentials for [Mo6O19]22, away from the triangular anion face containing three terminal and three bridging O atoms at z = 1.1 Å (top) and 2.2 Å (bottom) respectively. Lowest contour: 20.285 hartree (top), 20.220 hartree (bottom). All contour intervals equal to 0.003 hartree (top), 0.001 hartree (bottom).1246 J. Chem. Soc., Dalton Trans., 1999, 1241–1248 distribution and eventually makes terminal O atoms available for further interaction with electrophilic moieties.In their interaction with hydrogen bond donors such as hydroxyl groups, surface O atoms of [Mo6O19]22 are in competition with the O atoms of the hydroxyl groups which exhibit stronger hydrogen-bond acceptor capabilities. Further work along those lines could therefore involve H-donor molecules with no competing acceptor atoms such as for example tetrathiafulvalenes bearing ammonium groups. Indeed, crystalline alkylammonium (RNH3 1, R2NH2 1) polyoxometalates have been shown to exhibit photochromism upon UV irradiation (l £ 400 nm), a process involving the transfer of a proton from a hydrogen-bonded alkylammonium nitrogen to a bridging oxygen atom at the photoreducible site in the edge-shared MoO6 octahedral lattice.26 Modifications of the [Mo6O19]22 anion itself could also be considered in order to increase the total charge and/or the charge density on the surface O atoms as in [Mo6O18(NO)]32 for example.27 Fig. 9 Sections of the electrostatic potentials for the hypothetical [HMo6O19]2, in the plane containing four Mo, four terminal and four bridging O atoms (top) and in the plane containing three terminal and three bridging O atoms (bottom); the H atom is linked to O7 which pertains to both planes. Lowest contour: 20.18092 hartree (top), 20.15167 hartree (bottom). First contour interval: 4 × 1025 hartree. Successive contour intervals increased by a factor of 2.Experimental Synthesis of 3-formyl-4,39,49-trimethyltetrathiafulvalene To a solution of Me3TTF7 (1 g, 4.06 mmol) in dry Et2O (100 mL) at 278 8C under nitrogen was added NH(i-Pr)2 (0.64 mL, 4.46 mmol) followed by BuLi (2.5 M in hexanes, 1.8 mL, 4.46 mmol). The yellow suspension was stirred for 1 h and Nmethylformanilide (0.54 mL, 4.46 mmol) added. The reaction mixture was slowly warmed to room temperature and hydrolysed with acidified water (0.01 M HCl, 50 mL) giving rise to a dark red solution. This was extracted with toluene, washed with 0.01 M HCl, water and dried with MgSO4.Chromatography on SiO2 (toluene) aVorded the required compound as a red product, recrystallized from acetone, yield 0.7 g (63%) mp 170– 171 8C [Calc. for C10H10OS4 (Found): C, 43.75 (44.47); H, 3.68 (3.82); O, 5.83 (5.81); S, 46.72 (45.93)%]. 1H NMR (CDCl3, TMS): d 1.93 (s, 6 H, Me), 2.15 (s, 3 H, Me) and 9.70 (s, 1 H, CHO). UV(DMF): lmax = 446 nm. Fig. 10 Sections of the electrostatic potentials for the hypothetical [HMo6O19]2, away from the triangular anion face containing three terminal and three bridging O atoms at z = 1.1 Å (top) and 2.2 Å (bottom) respectively.Lowest contour: 20.160 hartree (top), 20.11- hartree (bottom). All contour intervals equal to 0.003 hartree (top)- 0.001 hartree (bottom)J. Chem. Soc., Dalton Trans., 1999, 1241–1248 1247 Table 4 Crystallographic data for the three 2 : 1 salts with [Mo6O19]22 Donor molecule Formula Formula weight Crystal dimensions/mm Crystal system Space group a/Å b/Å c/Å a/8 b/8 g/8 V/Å3 Z Dc/g cm23 m/mm21 Data collected q Range/8 Independent reflections Obs.reflections [I > 2s(I)] Parameters refined Absorption correction Tmax, Tmin R(F) [I > 2s(I)] wR(F2) (all) Goodness of fit Residual density/e Å23 Me3TTF-CH2OH C20H24Mo6O21S8 1432.51 0.4 × 0.2 × 0.1 Triclinic P1� 7.830(1) 11.288(2) 11.375(2) 109.08(2) 95.29(2) 103.41(2) 908.8(3) 1 2.617 2.551 9813 2.0, 27 3642 3204 257 Empirical 0.793, 0.580 0.0260 0.0797 1.051 1.28, 20.55 EDT-TTF(CH2OH)2 C20H20Mo6O23S12 1588.72 0.21 × 0.11 × 0.02 Monoclinic P21/c 20.654(2) 11.5411(8) 17.110(1) 101.342(9) 3998.8(5) 4 2.639 2.538 34134 2.0, 26.9 8212 6918 543 Numerical 0.915, 0.640 0.0637 0.1335 1.292 2.7, 21.5 TTF(CH2OH)4 C20H24Mo6O27S8 1528.51 0.28 × 0.05 × 0.02 Triclinic P1� 8.668(2) 10.862(2) 11.478(2) 69.58(3) 83.28(3) 77.00(3) 985.9(3) 1 2.574 2.371 10552 2.04, 26.95 3914 2948 277 Numerical 0.991, 0.827 0.0247 0.0403 0.89 0.57, 20.69 Synthesis of 3-(hydroxymethyl)-4,39,49-trimethyltetrathiafulvalene The compound Me3TTF-CHO (0.7 g, 2.55 mmol) was suspended in EtOH (50 mL) and NaBH4 (0.24 g, 6.37 mmol) added by fractions.The mixture was warmed to reflux for 30 min, cooled to room temperature and the solvent evaporated. Trituration of the solid residue with water and filtration aVorded an orange solid, recrystallized from acetone, yield 0.66 g (94%), mp 203–204 8C [Calc.for C10H12OS4 (Found): C, 43.44 (43.25); H, 4.38 (4.48); O, 5.79 (6.04); S, 46.38 (46.13)%]. 1H NMR (CDCl3, TMS): d 1.93 (br s, 9 H, Me) and 4.23 (s, 2 H, CH2). UV (DMF): lmax = 476 nm. Electrocrystallization conditions The electrolyte, [n-Bu4N]2[Mo6O19], was prepared as reported elsewhere 28 and recrystallized twice before use. The donor molecules were galvanostatically electrooxidized in 2 mM solutions of [n-Bu4N]2[Mo6O19], in PhCN at 40 8C for EDTTTF( CH2OH)2, in MeCN–EtOH (1: 1) at 20 8C for Me3TTF- (CH2OH) or in DMF at 30 8C for TTF(CH2OH)4.Crystals were harvested on the electrode after typically one week. Crystallography Crystallographic data for the three salts are summarized in Table 4. Data were collected on a Stoe Imaging Plate Diffraction System (IPDS) at 150 K with Mo-Ka radiation (l = 0.71073 Å). Structures were solved by direct methods using SHELXS 8629 and refined by full-matrix least-square procedures using SHELXL 93.30 All non-hydrogen atoms were refined anisotropically except the disordered oxygen atoms found in [Me3TTF-CH2OH]2[Mo6O19] and in [EDTTTF( CH2OH)2]2[Mo6O19].In the former the O atom is disordered on four positions with occupation parameters refined to values close to 1/3 on C8 and C9 and to 1/6 for the two partially occupied O sites linked to C7. These values were thus fixed and the isotropic thermal parameters of O8, O9 and O7A, O7B were then refined. In the EDT(CH2OH)2 salt one of every two O atoms of each independent donor molecules A and B was found disordered on two positions.Refinement of their occupation parameters converged to 0.6 and 0.4 for the two O positions linked to C1A, to 0.5 and 0.5 for the two O positions linked to C1B. These values were thus fixed and the isotropic thermal parameters of O1A1, O1A2, O1B1 and O1B2 refined. Furthermore, in this salt, atom C7A could not be refined anisotropically and was therefore refined with an isotropic thermal parameter.The strong residual peak in the Fourier diVerence map is located at 0.9 Å from atom Mo3. Hydrogen atoms of the methylene groups were introduced at calculated positions in [EDT-TTF(CH2OH)2]2[Mo6O19] and [TTF(CH2OH)4]2- [Mo6O19], included in structure factor calculations and not refined (riding model); they were not introduced in [Me3- TTF(CH2OH)]2[Mo6O19]. Hydrogen atoms of the hydroxyl groups could not be found in Fourier diVerence maps for [Me3TTF(CH2OH)]2[Mo6O19] and [EDT-TTF(CH2OH)2]2- [Mo6O19] but they were unambiguously identified in [TTF- (CH2OH)4]2[Mo6O19] but not refined.CCDC reference number 186/1353. See http://www.rsc.org/suppdata/dt/1999/1241/ for crystallographic files in .cif format. Magnetic measurements The ESR spectra were recorded on a Varian X-band spectrometer operating at 9.3 GHz and equipped with an OXFORD ESR 900 helium cryostat. Magnetic susceptibility measurements for [Me3TTF(CH2OH)]2[Mo6O19] were performed on a Quantum Design MPMS-2 SQUID magnetometer operating in the range 1.7–300 K.Acknowledgements This work was supported by the Centre National de la Recherche Scientifique and the Région Pays de Loire. References 1 E. Coronado and C. J. Gómez-García, Chem. Rev., 1998, 98, 273. 2 A. Davidson, K. Boubekeur, A. Pénicaud, P. Auban, C. Lenoir, P. Batail and G. Hervé, J. Chem. Soc., Chem. Commun., 1989, 1373. 3 C. J. Gómez-García, L. Ouahab, C. Giménez-Saiz, S. Triki, E.Coronado and P. Delhaes, Angew. Chem., Int. Ed. Engl., 1994, 33, 223; C. J. Gómez-García, C. Giménez-Saiz, S. Triki,1248 J. Chem. Soc., Dalton Trans., 1999, 1241–1248 E. Coronado, P. Le Maguéres, L. Ouahab, L. Ducasse, L. C. Sourisseau and P. Delhaes, Inorg. Chem., 1995, 34, 4139. 4 J. R. Galán-Mascarós, C. Giménez-Saiz, S. Triki, C. J. Gómez- García, E. Coronado and L. Ouahab, Angew. Chem., Int. Ed. Engl., 1995, 34, 1460. 5 D. Attanasio, C. Bellitto, M. Bonamico, V.Fares and P. Imperatori, Gazz. Chim. Ital., 1991, 121, 155; S. Triki, L. Ouahab, J. Padiou and D. Grandjean J. Chem. Soc., Chem. Commun., 1989, 1068. 6 S. Triki, L. Ouahab, D. Grandjean and J.-M. Fabre, Acta Crystallogr., Sect. C, 1991, 47, 1371. 7 A. Dolbecq, M. Fourmigué, P. Batail and C. Coulon, Chem. Mater., 1994, 6, 1413. 8 Panchard, K. Boubekeur, M. Sallé, G. Duguay, M. Jubault, A. Gorgues, J. D. Martin, E. Canadell, P. Auban-Senzier, D. Jérome and P. Batail, Adv.Mater., 1992, 4, 579. 9 A. Dolbecq, M. Fourmigué, F. C. Krebs, P. Batail, E. Canadell, R. Clérac and C. Coulon, Chem. Eur. J., 1996, 2, 1275. 10 A. J. Moore, M. R. Bryce, A. S. Batsanov, J. N. Heaton, C. W. Lehmann, J. A. K. Howard, N. Robertson, A. E. Underhill and I. F. Perepichka, J. Mater. Chem., 1998, 8, 1541. 11 P. Batail, K. Boubekeur, M. Fourmigué and J.-C. P. Gabriel, Chem. Mater., 1998, 10, 3005. 12 D. C. Green, J. Org. Chem., 1979, 44, 1477. 13 See also A. J. Moore, M. R. Bryce, A. S. Batsanov, J. C. Cole and J. A. K. Howard, Synthesis, 1995, 675. 14 P. Blanchard, M. Sallé, G. Duguay, M. Jubault and A. Gorgues, Tetrahedron Lett., 1992, 33, 2685. 15 M. Sallé, A. Gorgues, M. Jubault, K. Boubekeur and P. Batail, Tetrahedron, 1992, 48, 3081. 16 K. Boubekeur, Ph.D. Thesis, University of Rennes, 1989. 17 A. Guirauden, I. Johannsen, P. Batail and C. Coulon, Inorg. Chem., 1993, 32, 2446. 18 H. R. Allcock, E. C. Bissell and E. T. Shawl, Inorg. Chem., 1973, 12, 2963. 19 R. HoVmann, J. Chem. Phys., 1963, 39, 1397. 20 G. A. JeVrey, M. E. Gress and S. Takagi, J. Am. Chem. Soc., 1977, 99, 609; G. A. JeVrey and L. Lewis, Carbohydr. Res., 1978, 60, 179. 21 G. A. JeVrey and W. Saenger, Hydrogen Bonding in Biological Structures, Springer, Berlin, Heidelberg, 1991. 22 L. Barcza and M. T. Pope, J. Phys. Chem., 1975, 79, 92. 23 See also C. Rimbaud, L. Ouahab, J. P. Sutter and O. Kahn, Mol. Cryst. Liq. Cryst., 1997, 306, 67. 24 M.-M. Rohmer, M. Bénard, J.-P. Blaudeau, J.-M. Maestre and J.-M. Poblet, Coord. Chem. Rev., 1998, 178–180, 1019. 25 J.-Y. Kempf, M.-M. Rohmer, J.-M. Poblet, C. Bo and M. Bénard, J. Am. Chem. Soc., 1992, 114, 1136. 26 T. Yamase, Chem. Rev., 1998, 98, 307. 27 A. Proust, R. Thouvenot, F. Robert and P. Gouzerh, Inorg. Chem., 1993, 32, 5299; A. Proust, R. Thouvenot, S.-G. Roh, J.-K. Yoo and P. Gouzerh, Inorg. Chem., 1995, 34, 4106. 28 M. Che, M. Fournier and J.-P. Launay, J. Chem. Phys., 1979, 71, 1954. 29 G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467. 30 G. M. Sheldrick, SHELXL 93, Program for the Refinement of Crystal Structures, University of Göttingen, 1993. Paper 8/09442J

ISSN:1477-9226    

DOI:10.1039/a809442j

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 1249–1255 1249 Synthesis and characterization of a new class of chelating bis(amidinate) ligands Glenn D. Whitener, John R. Hagadorn and John Arnold * Department of Chemistry, University of California and Chemical Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720-1460, USA Received 25th November 1998, Accepted 2nd February 1999 A series of diamidines based on trans-1,2-diaminocyclohexane have been prepared.Reactions of trans-1,2- diaminocyclohexane with a variety of aryl acid chlorides yields the corresponding diamides in very high (91–95%) yield. Conversion of the diamides to the diimine chloride is carried out by treatment with PCl5 in CH2Cl2. Reaction of the diimine chloride with aniline in CH2Cl2 finally yields the diamidine in good yields on multi-gram scales. Alternatively, a one-pot reaction between the diamide, PCl5 and aniline gives the products, although yields are generally lower.The solid-state structure of the unsubstituted diamidine shows localized C–N single and double bonds in the amidine moiety; intra- and inter-molecular hydrogen bonding is also observed between amidine groups. Alkali metal derivatives (M = Li, Na, K) were prepared by reaction of the diamidines with either nBuLi or MN(SiMe3)2. Isolated yields of these alkali metal derivatives, which crystallize as thf adducts, are moderate (34–64%). X-Ray crystallography shows that in all cases the alkali metal atoms bridge the two amidinate groups within the same molecule, forming a C2 symmetric eight-membered ring.Nonetheless, there are marked diVerences in coordination geometries of the series. Introduction Benzamidinate ligands [PhC(NSiMe3)2 2] are useful ancillary ligands that form complexes with a variety of metals from across the periodic table.1–9 In particular, group 4 amidinates are known to support a variety of substitution and insertion reactions,1,2,10–12 comparable to those of related metallocene complexes (Cp2M, M = Sc, Ti, Zr).13–15 Activity in this area has been intense recently, as interest grows in finding useful and robust ligands to support olefin polymerization reactions.2,3,6,7,13,14,16–22 We are interested in extending bisamidinate chemistry to systems where the two amidines are tethered by short linkers in a constrained manner similar to ansa metallocenes.23–31 These systems diVer considerably from their non-constrained relatives, most notably in the influence on polymer microstructure when C2 symmetric catalysts are used in propylene polymerization.26,32 Our entry point to linked amidinate systems is based largely on work carried out in the 1950s,33 where linked amidines were being investigated for their use as local anesthetics, with the hope that the linked functionality would oVer lesser toxicity than that of related compounds available at the time. A highly attractive feature of ligands derived from this synthetic methodology is the extremely facile resolution that is available for chiral diamines.34–36 Thus, the synthesis of C2 symmetric ligands can proceed on frameworks that are already resolved and enantiomerically pure.Also appealing about this methodology is the ready commercial availability of starting diamines. Here we describe the synthesis and characterization of the first examples of chiral C2-symmetric bisamidines on a multigram scale and their conversion to alkali metal derivatives.The latter complexes are, in turn, useful reagents for the preparation of transition metal complexes.37 Results and discussion Synthesis of diamidines Resolution of trans-1,2-diaminocyclohexane can be performed at the outset of derivatization by preferential salt formation with either D- or L-tartaric acid, depending on whether the R,Ror S,S- form, respectively, is desired.34,35 It should be noted that while all compounds reported herein were prepared from the unresolved diamine backbone, further reactivity studies will employ the resolved compound.Synthesis of diamides 1–3 proceeded easily by reaction of trans-1,2-diaminocyclohexane with an appropriate aryl acid chloride in the presence of triethylamine as shown in eqn. (1). The reaction was nearly quantitative in all cases, regardless of the aryl acid chloride employed, and mechanical loss accounted for nearly all loss of final product.p-Toluoyl chloride is useful in this regard as the methyl group provided a good 1H NMR handle for characterization of the diamide 3, and the eventual diamidine. Solubility of the diamides was greatly dependent on the substitution of the aryl acid chloride. As might be expected, the di-tert-butyl diamide 2 was the most soluble, dissolving readily in CH2Cl2 and CHCl3 and sparingly in benzene, ether (Et2O), and acetone. For comparison, the unsubstituted diamide 1 was soluble only in hot dimethyl sulfoxide (DMSO).1250 J.Chem. Soc., Dalton Trans., 1999, 1249–1255 Preparation of diamidines from diamides is best accomplished in two steps. First, reaction of the diamide with two equiv. of PCl5 in methylene chloride yields the corresponding diimine chloride (eqn. (2)). The diimine chlorides are crystalline compounds soluble in Et2O, benzene, tetrahydrofuran (thf) and methylene chloride, and can be prepared on large scales (~40 g). In the second step, reaction of the diimine chloride with two equiv.of the desired aniline derivative gave the diamidine in high yield (89%). Alternatively, the whole reaction sequence can also be carried out as a one-pot reaction by formation of the diimine chloride in situ, followed by addition of aniline, although the scale of the reaction, and product purification appears to suVer somewhat in the process. We abbreviate the completely unsubstituted diamidine as CDA-H2. Substituted diamidines are abbreviated according to their substitution pattern. Thus, a diamidine with a tert-butyl group on the phenyl ring attached to the ipso carbon is abbreviated (But)CDA-H2 (6), and one with a methyl group on the ipso carbon phenyl ring and a methoxy group on the phenyl ring attached to the terminal amidine nitrogen is abbreviated (Me, OMe)CDA-H2.A variety of diamidine compounds have been prepared using the above procedures. As mentioned above, substitution of the aryl rings in the para positions greatly aVects solubility and characterization of the family of diamidines.Proton NMR spectra of the series of diamidines are informative due to the number of diVerent functionalities in the molecule. Resonances from the aryl ring attached to the C atom of the amidine functionality are consistently upfield of signals from the aryl ring attached to the N atom of the amidine. Protons on the amidine moieties appear at ca. 6 ppm, considerably farther upfield than common amines, but within the range of common amides.38 Most resonances from the cyclohexyl ring are found between 2.5 ppm and 1.4 ppm, with that of the protons on the carbon atoms to which the amidines are attached near 4 ppm.IR spectra of the diamidines all showed a characteristic nN–H at approximately 3200 cm21 along with a number of strong absorbances in the region from 1650 to 1500 cm21 due to nC]] N. Synthesis of alkali metal diamidinates The lithium, sodium, and potassium derivatives were synthesized by reaction of a diamidine with the appropriate metal (bistrimethylsilyl)amide, MN(SiMe3)2 (M = Li, Na, K; eqn.(3)). All the alkali metal derivatives crystallize nicely from thf as bright yellow solids, with coordinated thf clearly apparent in the 1H NMR spectrum. The lithium and sodium derivatives contain two thf’s per alkali metal, while the potassium derivative shows the presence of 2.5 thf molecules per K atom. Relative to the starting diamidine, the protons on the cyclohexyl ring a to the amidinate functionalities in the lithium derivative shift upfield considerably (~0.3 ppm), although this is not observed in the sodium and potassium derivatives.The strongest absorbance in the IR spectrum of the diamidine is the C]] N stretch at 1505 cm21. All alkali metal derivatives show strong absorptions at lower energy to the diamidines from which they were synthesized, indicative of the change from a localized double bond to a delocalized functionality.The stretches also decrease in energy from the lithium to the potassium derivative, with the lithium salt showing a strong absorbance at 1487 cm21, the sodium salt at 1475 cm21, and the potassium salt at 1470 cm21. We have already shown that these derivatives function as useful synthons in salt metathesis reactions with group 4 metal halides 37 and related work is continuing. Structural studies CDA-H2. A view of the molecular structure is shown in Fig. 1 with relevant crystallographic details in Table 1 and selected metrical parameters in Table 2. The compound crystallizes racemically in the centrosymmetric space group P21/n. The cyclohexyl backbone is in the thermodynamically preferred chair conformation. All hydrogen atoms were found and refined. There is an intramolecular hydrogen bond at 2.00 Å between H61 on N1 and N4 on the other amidinate group, with an angle of 1538. Additionally, intermolecular hydrogen bonding is also observed between H62 on N3 and N2 on a neighboring diamidine molecule in the unit cell.The C–N bond lengths in each amidine functionality diVer by approximately 0.07 Å, in accord with the expected localized structure.39 The double bond present a to the cyclohexyl ring in the imine chloride precursor has migrated to a position between the two neighboring aryl groups.J. Chem. Soc., Dalton Trans., 1999, 1249–1255 1251 Table 1 Crystal data and collection parameters Formula Formula weight Space group Temperature/8C a/Å b/Å c/Å V/Å3 a/8 b/8 g/8 Zd calc/g cm23 Scan type, deg Frame collection time/s Reflections measured 2q range/8 m/cm21 Tmin, Tmax Crystal dimensions/mm No.of reflections measured No. of unique reflections No. of observations [I > 3s(I)] No. of variables Rint (%) R (%) Rw (%) GOF CDA-H2 C32H32N4 472.62 P21/n 2115 13.941(5) 10.115(0) 18.986(9) 2673.5(3) 93.125(0) 4 1.174 w, 0.3 15 Hemisphere 3–51.3 0.697 0.694, 0.774 0.18 × 0.45 × 0.49 12320 4781 4847 325 2.1 5.0 6.4 2.175 (Me,Me)CDA-Li2(thf)4 C52H70N4O4Li2 829.00 P2/n 2119 29.797(6) 11.613 30.359(1) 10482.1(5) 93.862(1) 8 1.051 w, 0.3 30 Hemisphere 3–52.2 0.652 0.723, 0.923 0.18 × 0.22 × 0.30 49068 19443 7544 1203 3.6 7.0 8.8 2.513 (MeO,tBu)CDA-Na2(thf)5 C62H90N4O7Na2 1049.40 P1� 2125 11.352(9) 13.831(8) 20.648(1) 3003.0(1) 73.041(1) 89.784(1) 76.114(1) 2 1.160 w, 0.3 30 Hemisphere 3–46.5 0.869 0.685, 0.977 0.27 × 0.17 × 0.10 12485 8360 5499 694 2.67 6.00 7.63 2.229 (Me,Me)CDA-K2(thf)5 C56H78N4O5K2 965.44 C2/c 2136 to 2133 20.753(1) 13.637(7) 19.852(1) 5516.0(3) 97.965(1) 4 1.153 w, 0.3 20 Hemisphere 3–52.2 2.198 0.826, 0.904 0.13 × 0.16 × 0.22 25320 10142 2596 300 2.6 5.4 6.8 2.191 Alkali metal derivatives.A few common features of all the alkali metal derivatives will be discussed first. Firstly, they all crystallize racemically in centrosymmetric space groups (Table 1). The C–N bonds in the amidinates are approximately equal at an average of 1.34 Å, in the range of other known main group and transition metal amidinate complexes (see Tables 3–5).1–3,6–8 This value is intermediate to those observed in 5, which showed bond lengths of 1.30 and 1.37 Å, and is indicative of the delocalized nature of the amidinate moiety.The alkali metal atoms bridge the separate amidinate groups of the Fig. 1 ORTEP49 plot of CDA-H2 5 with thermal ellipsoids at 50% probability level.Table 2 Selected distances (Å) and angles (8) for CDA-H2 5 C7–N1 C7–N2 C20–N3 C20–N4 N1–C7–N2 N3–C20–N4 C1–N1–C7 C6–N3–C20 1.352(2) 1.298(2) 1.368(2) 1.298(2) 121.2(2) 118.9(2) 122.4(1) 123.5(1) N1–H61 N3–H62 H61–N4 H62–N29 C7–N2–C14 C20–N4–C27 C6–C1–N1 C1–C6–N3 0.97 1.01 2.00 2.40 120.6(1) 120.6(2) 111.2(1) 113.3(1) same molecule, in a well-established bonding mode for these types of compounds.40 Each alkali metal atom binds either two (Li,Na) or three (K) thf molecules to complete its coordination sphere.The structure of (Me,Me)CDA-Li2(thf)4 10, shown in Fig. 2, confirms that the lithiums are bound in monodentate fashion to each of the nitrogen atoms, as evidenced by a typical C7–N2– Li1 angle of 1138, and that on average they lie only 1.38 Å from the planes defined by the amidinate groups. Each lithium is pseudo-tetrahedral, with two coordinated thf molecules and Table 3 Selected bond distances (Å) and angles (8) for (Me,Me)CDALi2( thf)4 10 C7–N1 C7–N2 C22–N3 C22–N4 C60–N5 C60–N6 C75–N7 C75–N8 Li1–N2 Li1–N3 Li1–O1 N1–C7–N2 N3–C22–N4 N5–C60–N6 N7–C75–N8 N1–C1–C6 N3–C6–C1 N2–Li1–N3 N1–Li2–N4 N6–Li3–N7 N5–Li4–N8 O1–Li1–O2 O3–Li2–O4 O5–Li3–O6 O7–Li4–O8 O1–Li1–N2 1.303(8) 1.366(8) 1.314(8) 1.358(8) 1.317(8) 1.361(7) 1.334(7) 1.355(8) 2.06(1) 2.15(1) 2.00(1) 125.1(6) 124.7(6) 125.3(6) 123.9(6) 111.0(5) 110.9(6) 108.2(5) 108.1(6) 110.0(5) 110.6(5) 100.1(5) 95.5(5) 96.4(5) 101.3(5) 109.8(6) Li1–O2 Li2–N1 Li2–N4 Li2–O3 Li2–O4 Li3–N6 Li3–N7 Li3–O5 Li3–O6 Li4–N5 Li4–N8 Li4–O7 Li4–O8 O1–Li1–N3 O2–Li1–N2 O2–Li1–N3 O3–Li2–N1 O3–Li2–N4 O4–Li2–N1 O4–Li2–N4 O5–Li3–N6 O5–Li3–N7 O6–Li3–N6 O6–Li3–N7 O7–Li4–N5 O7–Li4–N8 O8–Li4–N5 O8–Li4–N8 2.01(1) 2.10(1) 2.10(1) 2.02(2) 1.97(1) 2.06(1) 2.13(1) 2.02(1) 1.95(1) 2.07(1) 1.98(1) 1.98(1) 1.97(1) 112.2(6) 106.0(6) 120.0(6) 123.2(6) 106.0(6) 102.8(6) 122.2(6) 107.6(6) 122.6(6) 111.0(6) 108.3(6) 119.2(6) 105.0(6) 108.3(6) 112.2(6)1252 J.Chem. Soc., Dalton Trans., 1999, 1249–1255 has average Li–N bond lengths of 2.09 Å, well in the range of similar alkali metal amidinate compounds.The environment of the sodium atoms in the (But,OMe)- CDA-Na2(thf)4 (11, Fig. 3), parallels that of 10 with the main diVerence being that the alkali metal in the former compound sits slightly above the plane defined by the N–C–N of the amidinate, with a typical C8–N1–Na1 angle of 958. This gives the complex a greater degree of h3, allyl-like binding.As expected, the average Na–N bond length of 2.40 Å is longer than that observed in 10, but is within the range of similar compounds. In contrast to the lithium and sodium derivatives, the potassium atoms in (Me,Me)CDA-K2(thf)5 (12, Fig. 4), sit well above the plane defined by the amidinate N–C–N plane, with an angle of just less than 908, and are bound primarily through the delocalized p system of the amidinates. Each potassium has a close contact with the central carbon of an amidinate; the bond Fig. 2 ORTEP49 plot of the core structure of (Me,Me)CDA-Li2(thf)4 10 with thermal ellipsoids at 50% probability level. All hydrogens, thf ring carbons and aryl ring carbons (other than ipso) omitted for clarity. Table 4 Selected bond distances (Å) and angles (8) for (tBu,OMe)- CDA-Na2(thf)4 11 C8–N1 C8–N2 C25–N3 C25–N4 Na1–N1 Na1–N3 N1–C8–N2 N3–C25–N4 N1–Na1–N3 N2–Na2–N4 O3–Na1–O4 O5–Na2–O6 N1–Na1–O3 1.339(5) 1.318(5) 1.315(6) 1.360(5) 2.428(4) 2.426(4) 125.2(3) 124.8(4) 107.4(1) 104.8(1) 94.3(1) 88.9(1) 99.2(1) Na1–O3 Na1–O4 Na2–N2 Na2–N4 Na2–O5 Na2–O6 N1–Na1–O4 N3–Na1–O3 N3–Na1–O4 N2–Na2–O5 N2–Na2–O6 N4–Na2–O5 N4–Na2–O6 2.336(4) 2.297(3) 2.367(4) 2.396(3) 2.332(4) 2.310(4) 114.4(1) 129.4(1) 111.6(1) 140.0(1) 105.1(2) 102(1) 111.7(1) Table 5 Selected bond distances (Å) and angles (8) for (Me,Me)CDAK2( thf)5 12 C4–N1 C4–N2 K1–N1 K1–N2 N1–C4–N2 N1–K1–N2 N1–C3–C3* O1–K1–O2 O1–K1–O3 O2–K1–O3 1.319(5) 1.346(5) 2.737(3) 2.781(4) 124.4(4) 99.4(1) 108.5(3) 88.4(1) 156.8(2) 76.7(2) K1–O1 K1–O2 K1–O3 O1–K1–N1 O1–K1–N2 O2–K1–N1 O2–K1–N2 O3–K1–N1 O3–K1–N2 2.707(4) 2.653(3) 3.052(9) 97.6(1) 96.7(1) 158.5(1) 100.4(1) 90.6(2) 103.4(2) length of 3.08 Å is very close to the bond length of the oxygen atom of a coordinated thf (3.05 Å).A close contact with one of the hydrogen atoms of one of a neighboring tolyl group at 2.73 Å suggests that some amount of electron density is donated to the potassium atom. All of the K–N bonds are approximately equal, with an average bond length of 2.76 Å.As might be expected, this value is much longer than the other alkali metal derivatives. Conclusions We have synthesized a new class of chelating bis(amidinate) ligands based on trans-1,2-diaminocyclohexane. The ready availability and resolution of this backbone make it ideal as a foundation upon which to build new C2 symmetric ligands. Deprotonation by alkali metal amides provides an entry to synthetically useful derivatives for salt metathesis reactions with metal halides; preliminary studies along these lines have already appeared and further eVorts are currently in progress with a wide range of transition metals.Fig. 3 ORTEP49 plot of the core structure of (But,OMe)CDANa2( thf )4 11 with thermal ellipsoids at 50% probability level. All hydrogens, thf ring carbons and aryl ring carbons (other than ipso) omitted for clarity. Fig. 4 ORTEP49 plot of the core structure of (Me,Me)CDA-K2(thf)5 12 with thermal ellipsoids at 50% probability level.All hydrogens, thf ring carbons and aryl ring carbons (other than ipso) omitted for clarity.J. Chem. Soc., Dalton Trans., 1999, 1249–1255 1253 Experimental General considerations All manipulation of air sensitive compounds was performed using standard inert atmosphere glove box and Schlenk techniques. 41 Tetrahydrofuran, diethyl ether and hexanes were either distilled from sodium/benzophenone under nitrogen or obtained as anhydrous materials from Aldrich that were passed through a column of activated alumina, then degassed with nitrogen or argon.42 CH2Cl2 was distilled from CaH2 under nitrogen prior to use.Toluene was distilled from sodium under nitrogen prior to use. Deuteriated solvents were pre-dried over 4 Å molecular sieves. C6D6 was vacuum transferred from sodium/ benzophenone. CDCl3 was vacuum transferred from CaH2. Solutions of nBuLi in hexanes were purchased from commercial suppliers and were used without further purification.All 1H and 13C{1H} NMR spectra were recorded at ambient temperature in CDCl3 unless otherwise specified. Chemical shifts (d) are reported relative to tetramethylsilane at 0.00 ppm. IR samples were prepared as mineral oil mulls between KBr plates. Melting points were determined in sealed capillary tubes under nitrogen (where appropriate) and are uncorrected. General conditions for synthesis of linked diamides 2 Equiv.of the acyl chloride (ArC(O)Cl) and 2.4 equiv. of NEt3 were dissolved in methylene chloride or chloroform and stirred for about 10 min. To the resulting orange solution was added a solution of 1 equiv. of trans-1,2-diaminocyclohexane in the same solvent, forming a white precipitate. Addition was carried out slowly due to the exothermic nature of the reaction. A condenser was fitted to the flask and the reaction mixture was heated to reflux for 14 h. The solution was cooled to room temperature and filtered on a medium porosity glass frit.The resulting oV-white solid was washed with water (2×), saturated sodium bicarbonate solution (1×), then water again (2×). The solid was dried with benzene using a Dean–Stark apparatus; benzene was then removed under reduced pressure to yield a colorless solid. Trans-1,2-diphenylamidocyclohexane 1. Yield: 37 g, 91%. Mp: >300 8C. 1H NMR (d6-DMSO): d 8.23 (m, 2 H, aryl), 7.70 (d, 4 H, J 7.2 Hz, aryl), 7.44 (m, 2 H, aryl), 7.39 (m, 4 H, aryl), 3.93 (m, 2 H, amide), 1.92 (d, 2 H, J 12 Hz, cyclohexyl), 1.74 (m, 2 H, cyclohexyl), 1.50 (m, 2 H, cyclohexyl), 1.29 (m, 2H, cyclohexyl). 13C{1H} NMR (d6-DMSO): d 166, 134, 130, 128, 127, 115, 52, 31, 24.IR/cm21: 3311 (s), 1635 (s), 1551 (s), 1493 (w), 1455 (w), 1332 (w), 695 (w), 667 (w). Trans-1,2-di(4-(tert-butyl)phenyl)amidocyclohexane? (benzene)0.5 2. Yield: 18 g, 95%. The isolated product contains half an equivalent of benzene that remained even after extended drying under vacuum.Mp: 238–240 8C. 1H NMR: d 7.67 (d, 4 H, J 8.6 Hz, aryl), 7.30 (d, 4 H, J 8.5 Hz, aryl), 7.13 (s, 3 H, benzene), 4.03 (m, 2 H, amide), 2.22 (d, 2 H, J 10.9 Hz, cyclohexyl), 1.82 (m, 2 H, cyclohexyl), 1.25 (s, 18 H, tert-butyl). 13C{1H} NMR: d 168, 154, 131, 128, 126, 125, 54, 34, 32, 31, 24. IR/cm21: 3325 (s), 1637 (s), 1634 (s), 1611 (s), 1559 (s), 1557 (s), 1553 (s), 1549 (s), 1544 (s), 1541 (s), 1509 (s), 1456 (w), 1331 (w), 1272 (w), 667 (s).Trans-1,2-di(p-tolyl)amidocyclohexane 3. Yield: 95 g, 95%. Mp: >300 8C. 1H NMR: d 7.64 (d, 4 H, J 8.2 Hz, aryl), 7.14 (d, 4 H, J 7.9 Hz, aryl), 7.01 (m, 2 H, amide), 3.92 (m, 2 H, cyclohexyl), 2.32 (s, 6 H, methyl), 2.15 (d, 2 H, J 10.4 Hz, cyclohexyl), 1.72 (d, 2 H, J 6.4 Hz, cyclohexyl), 1.31 (m, 4 H, cyclohexyl). 13C{1H} NMR: d 168, 142, 131, 129, 127, 55, 32, 25, 21. IR/cm21: 3281 (m), 2924 (s), 2854 (s), 1640 (vs), 1614 (w, sh), 1571 (m), 1553 (m), 1508 (w), 1462 (m), 1452 (m), 1377 (w), 1333 (m), 834 (w), 757 (w), 712 (w), 685 (w).General conditions for conversion of diamides to diimine chlorides To a stirred suspension of PCl5 in CH2Cl2 was added a suspension of the diamide in CH2Cl2. A condenser was fitted to the flask and the reaction mixture refluxed for 14 h. All volatile material was removed under reduced pressure and the resulting solid was extracted with Et2O. Filtration through Celite followed by concentration resulted in the formation of crystalline material, which was isolated by filtration.Cooling the filtrate to 230 8C for 22 h gave the product as colorless crystals. Conversion of 3 to diimine chloride 4. Yield: 17 g, 65%. Mp: 143–145 8C. 1H NMR: d 7.82 (d, 4 H, J 8.3 Hz, aryl), 7.12 (d, 4 H, J 8.0 Hz, aryl), 4.18 (m, 2 H, cyclohexyl), 2.32 (s, 6 H, tolyl), 1.93 (m. 4 H, cyclohexyl), 1.52 (m, 4 H, cyclohexyl). 13C{1H} NMR: d 141.6, 141.5, 133.3, 129, 128.8, 67.1, 30.3, 24, 21.3.IR/cm21: 1676 (s), 1611 (w), 1508 (w), 1456 (m), 1377 (w), 1352 (w), 1311 (w), 1247 (w), 1223 (w), 1181 (w), 1134 (w), 1072 (w), 1024 (w), 932 (w), 886 (w), 861 (w), 837 (w), 821 (w), 780 (w), 714 (w), 628 (w), 616 (w), 535 (w). General conditions for synthesis of linked diamidines (a) PCl5 was suspended in CH2Cl2. The chosen aniline was then added dropwise with immediate formation of a colorless solid in a very exothermic reaction. A condenser was fitted to the flask and the solution was refluxed for 16–23 h.Removal of all volatile material gave a solid that was washed successively with water and saturated sodium bicarbonate until the washes remained clear. The solid was then taken up in CH2Cl2 and the solution washed with 2 M KOH (2×) and brine (2×), followed by drying over MgSO4. Removal of all volatile material under reduced pressure produced a yellow foam that was Soxhlet extracted with hexanes for 12 h. Removal of hexane from the filtrate under reduced pressure yielded a colorless solid as the final product.(b) It is found that separating the steps in (a), above, gives comparable yields while allowing the reaction to be carried out on a much larger scale. The diimine chloride was dissolved in CH2Cl2 and the solution cooled to 5 8C. Aniline was added dropwise, then the reaction warmed to room temperature. A condenser was fitted to the flask and the solution refluxed for 27 h. Removal of all volatile material under reduced pressure gave a white solid that was taken up in CH2Cl2. The solution was washed with 1 M KOH/H2O (3×) and brine (1×), then dried over Na2SO4.Filtration from the Na2SO4 on a medium porosity glass frit followed by removal of all volatile material under reduced pressure produced a white solid which was washed with Et2O (1×) to aVord the desired diamidine. CDA-H2 5. (a) Yield: 9.9 g, 67%. Mp: 171–173 8C. 1H NMR: d 7.25 (m, 10 H, aryl), 6.94 (t, 4 H, J 7.10 Hz, aryl), 6.74 (t, 2 H, J 7.16 Hz, aryl), 6.31 (d, 4 H, J 6.97 Hz, aryl), 5.99 (s, br, 2 H, aryl), 4.15 (s, br, 2 H, amidine), 2.44 (m, 2 H, cyclohexyl), 1.81 (s, br, 2 H, cyclohexyl), 1.44 (s, br, 4 H, cyclohexyl). 13C{1H} NMR: d 158.3, 150.7, 135.5, 129.0, 128.6, 128.2, 128.1, 123, 120.9, 56.1, 32.7, 24.9.IR/cm21: 3219 (w, broad), 1608 (w), 1590 (w), 1542 (w), 1526 (w), 1462 (m), 1377 (w), 695 (w). Anal. Calcd. for C32H32N4: C, 81.32; H, 6.82; N, 11.85; Found: C, 81.13; H, 6.90; N, 11.95%.(tBu)CDA-H2 6. (a) Yield: 2.7 g, 67%. Mp: 212–215 8C. 1H NMR: d 7.22 (d, 4 H, J 9.1 Hz, aryl), 7.13 (d, 4 H, J 7.9 Hz, aryl), 6.93 (t, 4 H, J 7.1 Hz, aryl), 6.72 (t, 2 H, J 6.9 Hz, aryl), 6.34 (d, 4 H, J 7.1 Hz, aryl), 5.97 (s, 2 H, aryl), 5.27 (s, 2 H, amidine), 4.10 (m, 2 H, cyclohexyl), 2.43 (m, 2 H, cyclohexyl), 1.78 (m, 2 H, cyclohexyl), 1.42 (m, 6 H, cyclohexyl), 1.26 (s, 18 H, tert-butyl). 13C{1H} NMR: d 158, 152, 151, 132, 128.4, 128.1, 125, 123, 120, 56, 53, 34, 32, 31, 24.IR/cm21: 3420 (w),1254 J. Chem. Soc., Dalton Trans., 1999, 1249–1255 3306 (w, broad), 3227 (w, broad), 1605 (w), 1587 (w), 1525 (w), 1486 (w), 1462 (m), 1377 (w), 836 (w), 753 (w), 696 (w). (tBu,OMe)CDA-H2 7. (a) Yield: 3.3 g, 44%. Mp: 206–211 8C. 1H NMR: d 7.23 (d, 4 H, J 8.9 Hz, aryl), 7.11 (d, 4 H, J 8.2 Hz, aryl), 6.51 (d, 4 H, J 8.6 Hz, aryl), 6.26 (d, 4 H, J 8.2 Hz, aryl), 5.29 (s, 2 H, amidine), 4.10 (s, br, 2 H, cyclohexyl), 3.64 (s, 6 H, methoxy), 2.42 (m, 2 H, cyclohexyl), 1.78 (s, br, 2 H, cyclohexyl), 1.41 (m, 4 H, cyclohexyl), 1.27 (s, 18 H, tert-butyl). 13C{1H} NMR: d 158.34, 154, 151.98, 132.82, 128.44, 125.07, 123.74, 114.65, 113.49, 56.06, 55.29, 34.67, 32.77, 31.24, 24.99. IR/cm21: 3225 (w, broad), 1599 (s), 1502 (s), 1462 (s, broad), 1237 (s), 1038 (w), 832 (m). (Me,Me)CDA-H2 8. (a) Yield: 26 g, 76%. Mp: 112–115 8C. 1H NMR: d 7.11 (d, 4 H, J 7.8 Hz, aryl), 7.03 (d, 4 H, J 7.8 Hz, aryl), 6.75 (d, 4 H, J 7.7 Hz, aryl), 6.21 (d, 4 H, J 7.5 Hz, aryl), 5.90 (s, br, 2 H, amidine), 4.12 (s, br, 2 H, cyclohexyl), 2.42 (d, 2 H, J 10.8 Hz, cyclohexyl, 2.32 (s, 6 H, methyl), 2.17 (s, 6 H, methyl), 1.8 (s, br, 2 H, cyclohexyl), 1.02 (m, 4 H, cyclohexyl). 13C{1H} NMR: d 158, 148, 139, 133, 129.8, 129.7, 129.1, 128.8, 128.7, 128.6, 115, 56, 33, 25, 21, 20.IR/cm21: 3210 (m), 2922 (vs), 2856 (vs), 1615 (vs), 1600 (vs), 1528 (vs), 1505 (vs), 1463 (s), 1377 (m), 1340 (m), 1325 (m), 1259 (m), 824 (s).(Me)CDA-H2 9. (a) Yield: 25 g, 78%. (b) Yield: 59 g, 89%. Mp: 191–194 8C. 1H NMR: d 7.08 (d, 4 H, J 8.0 Hz, aryl), 6.99 (d, 4 H, J 11.8 Hz, aryl), 6.92 (t, 4 H, J 7.6 Hz, aryl), 6.71 (t, 4 H, J 7.3 Hz, aryl), 6.29 (d, 2 H, J 7.6 Hz, aryl), 5.92 (s, br, amidine), 4.10 (s, br, 2 H, cyclohexyl), 2.41 (m, 2 H, cyclohexyl), 2.28 (s, 6 H, methyl), 1.78 (s, br, 2 H, cyclohexyl), 1.41 (m, 4 H, cyclohexyl). 13C{1H} NMR: d 158.4, 151, 139, 132.7, 128.9, 128.6, 128.1, 123, 120.8, 56.1, 32.8, 25, 21.3.IR/cm21: 3226 (w, br), 1606 (s), 1590 (s), 1543 (s), 1526 (s), 1487 (m), 1462 (vs), 1377 (m), 1350 (w), 1339 (w), 1327 (w), 1308 (w), 1260 (w), 1252 (w), 1230 (w), 1211 (w), 1183 (w), 1139 (w), 1071 (w), 1023 (w), 926 (w), 914 (w), 898 (w), 863 (w), 844 (w), 823 (w), 802 (w), 767 (w), 722 (w), 696 (w), 637 (w), 593 (w). (Me,Me)CDA-Li2(thf)4 10. (a) A solution of nBuLi in hexanes (2.56 M solution, 7.58 mmol) was added dropwise to a cold (278 8C) solution of compound 8 (2.00 g, 3.79 mmol) in toluene (50 mL).The cold bath was removed and the solution allowed to warm to room temperature. After 16 h all volatile material was removed under reduced pressure to yield a pale yellow solid. The solid was extracted with thf (50 mL), filtered, and concentrated to 20 mL under reduced pressure. Cooling to 230 8C yielded bright yellow crystals (1.8 g, 56%) which were isolated by filtration and dried under vacuum. On removal of all volatile material under reduced pressure the crystals were observed to become significantly more opaque, indicating some degree of coordinated solvent loss.(b) Deprotonation can be performed by addition of toluene to a solid mixture of LiN- (SiMe3)2 (3.17 g, 18.9 mmol) and 8 (5.00 g, 9.47 mmol) at room temperature. Similar isolation of the product yielded a yellow crystalline solid (5.02 g, 64%). Mp: >300 8C. 1H NMR (CD3- CN): d 7.15 (d, 4 H, J 6.9 Hz, aryl), 6.99 (d, 4 H, J 7.7 Hz, aryl), 6.47 (d, 4 H, J 8.1 Hz, aryl), 5.86 (d, 4 H, J 8.0 Hz, aryl), 3.81 (d, 2 H, J 8.5 Hz, cyclohexyl), 3.65 (m, 16 H, thf), 2.29 (s, 6 H, methyl), 2.00 (s, 6 H, methyl), 1.81 (m, 16 H, thf), 1.77 (m, 2 H, cyclohexyl), 1.70 (m, 4 H, cyclohexyl). 13C{1H} NMR: d 157, 154, 152, 144, 128, 125, 123, 113, 57, 55, 34, 32, 31, 25. IR/cm21: 1573 (w), 1514 (vs), 1487 (vs), 1467 (s), 1410 (m), 1365 (vs), 1337 (s), 1256 (s), 1044 (s), 823 (s).Anal. Calcd. for C26H35LiN2O2: C, 75.34; H, 8.51; Found: C, 74.98; H, 8.32%.(tBu,OMe)CDA-Na2(thf)5 11. A stirred suspension of 7 (5.00 g, 7.75 mmol) in 50 mL of ether was cooled to 220 8C. To this was added a cold (220 8C) solution of Na[N(SiMe3)2] (2.84 g, 15.5 mmol) in ether (30 mL). The resulting yellow suspension was gradually warmed to room temperature and stirred for 14 h. All volatile material was removed under reduced pressure to yield a yellow-orange solid which was extracted with 60 mL of thf. The solution was filtered and concentrated to 12 mL.Crystallization at 5 8C for 2 days aVorded 5.2 g (64%) of yellow, crystalline product. Mp: 122–127 8C. 1H NMR (C6D6): d 7.54 (m, 4 H, aryl), 7.20 (d, 4 H, J 8.0 Hz, aryl), 6.62 (d, 4 H, J 8.5 Hz, aryl), 6.31 (m, 4 H, aryl), 4.38 (m, 2 H, cyclohexyl), 3.49 (m, 18 H, thf), 3.30 (s, 6 H, methyl), 2.43 (m, 2 H, cyclohexyl), 2.13 (m, 2 H, cyclohexyl), 1.84 (m, 4 H, cyclohexyl), 1.36 (m, 18 H, thf), 1.17 (s, 16 H, tert-butyl). 13C{1H} NMR: d 168, 151, 149, 142, 129.3, 129, 128, 125, 123, 114, 67, 62, 55, 36, 34, 31, 26, 25.IR/cm21: 1606 (w, br), 1564 (w, br), 1414 (w, br), 1376 (w, br), 1361 (w, br), 1268 (w, br), 1252 (w, br), 1228 (m, sh), 1046 (w, sh), 904 (w, br), 840 (w, br), 824 (w, sh). Anal. Calcd. for C62H90N4Na2O7: C, 70.96; H, 8.64; N, 5.34; Found: C, 70.61; H, 8.56; N, 5.41%. (Me,Me)CDA-K2(thf)5 12. A solution of KN(TMS)2 (3.78 g, 18.9 mmol) in ether was added to a cold (278 8C) suspension of 8 (5.00 g, 9.47 mmol) in ether.The suspension was allowed to warm to room temperature and stirred for 16 h. Removal of all volatile material yielded a slightly yellow solid, which was extracted with 220 mL of thf to yield a bright yellow solution. The solution was filtered through Celite and concentrated to 125 mL. Cooling to 230 8C yielded 2.8 g (34%) of crystalline material in two crops. Mp: 218–221 8C. 1H NMR (CD3CN): d 7.16 (d, 4 H, J 7.9 Hz, aryl), 6.96 (d, 4 H, J 7.7 Hz, aryl), 6.55 (d, 4 H, J 7.4 Hz, aryl), 5.97 (s, 4 H, broad, aryl), 3.84 (s, 2 H, broad, cyclohexyl), 3.64 (m, 16 H, thf), 2.26 (s, 6 H, methyl), 2.05 (s, 6 H, methyl), 1.80 (m, 16 H, thf), 1.70 (s, 2 H, broad, cyclohexyl), 1.36 (s, 4 H, broad, cyclohexyl). 13C{1H} NMR: d 163, 153, 146, 139, 138, 134, 130.5, 130.3, 129, 128, 123, 116, 68, 61, 35, 27, 26, 21, 20.6, 20.5. IR/cm21: 1606 (m), 1572 (m), 1482 (vs), 1470 (vs), 1410 (s), 1366 (vs), 1341(s), 1301 (m), 1282(m), 1253 (m), 1172 (m), 1054 (s), 910 (m), 820 (s).Anal. Calcd. for C56H78K2N4O5: C, 69.67; H, 8.14; N, 5.80; Found: C, 69.39; H, 8.18; N, 5.82%. General procedures for X-ray crystallography of 5, 8, 9, and 10 Pertinent details for the individual compounds can be found in Table 1, and below. A crystal of appropriate size was mounted on a glass capillary using Paratone-N hydrocarbon oil. The crystal was transferred to a Siemens SMART diVractometer/ CCD area detector,43 centered in the beam (Mo-Ka), and cooled by a nitrogen flow low-temperature apparatus which had been previously calibrated by a thermocouple placed at the same position as the crystal.Preliminary orientation matrix and cell constants were determined by collection of 60 ten second frames, followed by spot integration and least squares refinement. A hemisphere of data was collected then the raw data were integrated (XY spot spread = 1.608; Z spot spread = 0.608) using SAINT.44 Cell dimensions reported in Table 1 were calculated from all reflections with I > 10s.Data analysis and absorption correction were performed using Siemens XPREP45 and SADABS. The data were corrected for Lorentz and polarization eVects, but no correction for crystal decay was applied. The reflections measured were averaged. The structure was solved and refined with the teXsan software package 46 using direct methods 47 and expanded using Fourier techniques.48 All non-hydrogen atoms were refined anisotropically, unless stated otherwise.Hydrogen atoms were assigned idealized positions and were included in structure factor calculations, but were not refined, unless stated otherwise. The final residuals were refined against the data for which F 2 > 3s(F2). The quantity minimized by the least squares program was Sw(|Fo| 2 |Fc|)2, where w is the weight of a given observation. The p factor, used to reduce the weight of intense reflections, was set to 0.03 throughout the refinement.The analytical forms of the scattering factor tablesJ. Chem. Soc., Dalton Trans., 1999, 1249–1255 1255 for the neutral atoms were used and all scattering factors were corrected for both the real and imaginary components of anomalous dispersion. CCDC reference number 186/1347. CDA-H2, 5. All hydrogen atoms were found in the diVerence Fourier map, and were not assigned idealized positions. (Me,Me)CDA-Li2(thf)4, 8. O10, C116, and C117 were modelled isotropically at 1/2 occupancy as part of a disordered thf molecule on an inversion center.C110, C111, C112, C113, and C114 were modelled isotropically at 1/2 occupancy as a completely disordered thf molecule located in the crystal lattice. C51, and C52 were modelled anisotropically at 1/2 occupancy as a carbon atom disordered over two positions. (tBu,OMe)CDA-Na2(thf)4, 9. C59, C60, C61, C62, C63, C64, and C65 were modelled anisotropically as part of a completely disordered thf molecule; C62, C63, C64, and C65 were modelled at 1/2 occupancy; C59, C60, and C61 were modelled at full occupancy. (Me,Me)CDA-K2(thf)5, 10.C28, C29, C30, C31, and O3 were modelled isotropically at 1/2 occupancy as a thf molecule disordered between coordination to K1 and K1*. Acknowledgements We thank the NSF for the award of a pre-doctoral fellowship to G. D. W. and the DOE for support of this work. References 1 J. R. Hagadorn and J. Arnold, Organometallics, 1996, 15, 984. 2 J. R. Hagadorn and J.Arnold, J. Chem. Soc., Dalton Trans., 1997, 3087. 3 S. Hao, P. Berno, R. K. Minhas and S. Gambarotta, Inorg. Chim. Acta, 1996, 244, 37. 4 J. R. Hagadorn and J. Arnold, Inorg. Chem., 1997, 36, 132. 5 D. G. Dick, R. Duchateau, J. J. H. Edema and S. Gambarotta, Inorg. Chem., 1993, 32, 1959. 6 C. C. Chang, C. S. Hsiung, H. L. Su, B. Srinivas, M. Y. Chiang, G. H. Lee and Y. Wang, Organometallics, 1998, 17, 1595. 7 S. R. Foley, C. Bensimon and D. S. Richeson, J. Am. Chem. Soc., 1997, 119, 10359. 8 Y. Zhou and D. S. Richeson, Inorg. Chem., 1996, 35, 2448. 9 H. H. Karsch, P. A. Schluter and M. Reisky, Eur. J. Inorg. Chem., 1998, 433. 10 J. R. Hagadorn and J. Arnold, J. Am. Chem. Soc., 1996, 118, 893. 11 J. R. Hagadorn and J. Arnold, Inorg. Chem., 1997, 36, 2928. 12 J. R. Hagadorn and J. Arnold, Organometallics, 1998, 17, 1355. 13 J. Richter, F. T. Edelmann, M. Noltemeyer, H.-G. Schmidt, M. Shmulinson and M. S. Eisen, J. Mol. Catal. A, 1998, 130, 149. 14 D.Herskovics-Korine and M. S. Eisen, J. Organomet. Chem., 1995, 503, 307. 15 R. Duchateau, C. T. van Wee, A. Meetsma, P. T. van Duijnen and J. H. Teuben, Organometallics, 1996, 15, 2279. 16 E. Muller, J. Muller, F. Olbrich, W. Bruser, W. Knapp, D. Abeln and F. T. Edelmann, Eur. J. Inorg. Chem., 1998, 87. 17 P. J. Stewart, A. J. Blake and P. Mountford, Inorg. Chem., 1997, 36, 3616. 18 M. P. Coles and R. F. Jordan, J. Am. Chem. Soc., 1997, 119, 8125. 19 M. P. Coles, D. C. Swenson, R.F. Jordan and V. G. Young, Organometallics, 1998, 17, 4042. 20 E. A. C. Brussee, A. Meetsma, B. Hessen and J. H. Teuben, Organometallics, 1998, 17, 4090. 21 A. T. Vlahos, T. A. Kabanos, C. P. Raptopoulou and A. Terzis, Chem. Commun., 1997, 269. 22 P. J. Stewart, A. J. Blake and P. Mountford, Inorg. Chem., 1997, 36, 1982. 23 A. Dormond, J. Tirouflet and F. LeMoigne, J. Organomet. Chem., 1974, 69, C7. 24 F. R. W. P. Wild, L. Zsolnai, G. Huttner and H. H. Brintzinger, J. Organomet.Chem., 1982, 232, 233. 25 F. R. W. P. Wild, M. Wasiucionek, G. Huttner and H. H. Brintzinger, J. Organomet. Chem., 1985, 288, 63. 26 V. C. Gibson, B. S. Kimberley, A. J. P. White, D. J. Williams and P. Howard, Chem. Commun., 1998, 313. 27 V. I. Tararov, D. E. Hibbs, M. B. Hursthouse, N. S. Ikonnikov, K. M. A. Malik, M. North, C. Orizu and Y. N. Belokon, Chem. Commun., 1998, 387. 28 K. Mashima, Y. Nakayama and A. Nakamura, Adv. Polym. Sci., 1997, 133, 1. 29 L. Schwink and P. Knochel, Chem. Eur. J., 1998, 4, 950. 30 M. Tokunaga, J. F. Larrow, F. Kakiuchi and E. N. Jacobsen, Science, 1997, 277, 936. 31 E. J. Corey and A. Guzman-Perez, Angew. Chem., Int. Ed., 1998, 37, 388. 32 H. H. Brintzinger, D. Fischer, R. Mulhaupt, B. Rieger and R. M. Waymouth, Angew. Chem., Int. Ed. Engl., 1995, 34, 1143. 33 A. J. Hill and J. V. Johnston, J. Org. Chem., 1954, 76, 920. 34 T. A. Whitney, J. Org. Chem., 1980, 45, 4214. 35 J. F. Larrow, E. N. Jacobsen, Y. Gao, Y. Hong, N. Xiaoyi and C. Zepp, J. Org. Chem., 1994, 59, 1939. 36 P. J. Walsh, D. K. Smith and C. Castello, J. Chem. Educ., 1998, 75, 1459. 37 J. R. Hagadorn and J. Arnold, Angew. Chem., Int. Ed., 1998, 37, 1729. 38 R. M. Silverstein, G. C. Bassler and T. C. Morrill, Spectrometric Identification of Organic Compounds, John Wiley & Sons, Inc., New York, 1991. 39 J. D. Korp and I. Bernal, Chem. Scr., 1982, 60. 40 A. Steiner and D. Stalke, Inorg. Chem., 1993, 32, 1977. 41 E. F. Shriver and M. A. Bredzon, The Manipulation of Air Sensitive Compounds, John Wiley & Sons, Inc., New York, 1986. 42 A. B. Pangborn, M. A. Giardello, R. H. Grubbs, R. K. Rosen and F. J. Timmers, Organometallics, 1996, 15, 1518. 43 SMART Area-Detector Software Package, Madison, WI, 1993. 44 SAINT: SAX Area-Detector Integration Program, Madison, WI, 1995. 45 XPREP: Part of the SHELXTL Crystal Structure Determination Package, Madison, WI, 1994. 46 TeXsan: Crystal Structure Analysis Package, 1992. 47 SAPI91: Structure Analysis Programs with Intelligent Control, Tokyo, 1991. 48 DIRDIF92: The DIRDIF Program System; Technical Report of the Crystallography Laboratory, Nijmegen, The Netherlands, 1992. 49 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. Paper 8/09211G

ISSN:1477-9226    

DOI:10.1039/a809211g

出版商:RSC    

年代:1999

数据来源: RSC