摘要:

DALTON PERSPECTIVE J. Chem. Soc., Dalton Trans., 1998, Pages 2589–2596 2589 Redox selective reactions of organo-silicon and -tin compounds Jun-ichi Yoshida * and Keiji Nishiwaki Department of Synthetic Chemistry and Biological Chemistry, Graduate School of Engineering, Kyoto University, Kyoto, Japan The C]Si and C]Sn Û orbitals are higher in energy than C]H or C]C Û orbitals, and therefore can interact with neighboring � systems, non-bonding orbitals of heteroatoms, and other Û systems such as those of C]Si and C]Sn.Such interactions cause an increase of the HOMO level which in turn favors electron transfer. On the basis of this eVect various types of redox selective reactions of organo-silicon and -tin compounds have been developed. 1 Introduction In the past decades the chemistry of organo-silicon and -tin compounds has witnessed a steady advance in progress, revealing their unique properties and the development of their synthetic applications.1 In the last several years redox reactions of organo-silicon and -tin compounds have received significant research interest and the activity in this area has grown rapidly.2 This account will provide a brief outline of redox selective reactions of organo-silicon and -tin compounds using electrochemical methods, with special emphasis on their mechanistic principles. 2 Redox Reactions on the Surface of an Electrode: Molecular Orbital Considerations Before discussing the redox reactions of organo-silicon and -tin compounds, let us start oV by analyzing the redox reactions on the surface of the electrode3 from a view point of molecular orbital theory (Fig. 1). There is an energy band of electrons in the electrode, and this band is filled up to the Fermi level. In the solution phase there exist substrate molecules and they have molecular orbitals of discrete energies. The molecular orbitals are filled up to the HOMO (highest occupied molecular orbital). In the case of oxidation, an electron moves from the HOMO of the substrate molecule to the electrode to produce a cation radical species which undergoes subsequent reactions.In the case of reduction, an electron moves from the electrode to the LUMO (lowest unoccupied molecular orbital) of the substrate molecule to produce an anion radical species which undergoes subsequent reactions to give final products. Hereafter we only discuss oxidation because of the simplicity of the discussion, but we should keep in mind that a similar discussion can be applied to reduction, although the direction of the electron transfer is reversed.One of the major advantages of electrochemical reactions is that the reactivity of the electrode, i.e. the position of the Fermi level, can be easily adjusted by external tuning (i.e. tuning of the dial of a potentiostat). When the electrode potential is adjusted properly, the electron transfer from the HOMO of the substrate molecule to the electrode takes place smoothly, although there usually exists some energy barrier for this process.If there are two kinds of substrate having diVerent HOMO levels, selective Fig. 1 Schematic diagram of the electron transfer on the surface of the electrode Jun-ichi Yoshida After graduating from Kyoto University in 1975, Jun-ichi Yoshida became a research associate at Kyoto Institute of Technology in 1979. He obtained his Doctoral degree in 1981 at Kyoto University. He moved to Osaka City University in 1984 and became an associate professor there in 1992.In 1994 he joined the Department of Synthetic Chemistry and Biological Chemistry at Kyoto University as a full professor. His research has ranged from synthetic methodology, organosilicon chemistry, organotin chemistry, electroorganic chemistry, computational chemistry, and automated synthesis. Keiji Nishiwaki After receiving an M.S. from Osaka City University in 1994, Keiji Nishiwaki enrolled in a doctoral course at Kyoto University. In 1997 he joined Kinki University as a research associate.His current research interest is the development of new electron donors based on b-effects of Group 14 elements in electron transfer.2590 J. Chem. Soc., Dalton Trans., 1998, Pages 2589–2596 oxidation can be accomplished by tuning the electrode potential appropriately as shown in Fig. 2. If the HOMO levels are very close to each other, however, it is diYcult to oxidize one substrate without aVecting the other.When the substrate that we wish to oxidize has a lower HOMO level than that of the other species, it is, in principle, impossible to accomplish selective oxidation. However, we often use tricks to accomplish the desired reaction in spite of unfavorable orbital situations. For example, if the substrate that we wish to oxidize is adsorbed onto the electrode selectively and the other species are not, then electron transfer takes place selectively to accomplish the desired reaction.Therefore, the use of diVerent electrode material sometimes gives rise to diVerent reaction pathways. Although such external control of electron transfer, i.e. control by electrode materials, solvent, current density, and so on, is important, we have recently proposed the importance of the internal control approach, and demonstrated that silicon and tin are quite eVective as controlling groups in redox reactions. We call such a group ‘electroauxiliary’.Silicon and tin promote the electron transfer from b-situated p systems, heteroatoms having non-bonding p orbitals, and some s orbitals. The following sections focus on the principle of the action of silicon and tin as controlling groups and their applications to redox selective reactions. 3 Control of Redox Reaction by Orbital Interaction 3.1 Electron transfer reactions of allylsilanes and benzylsilanes (Û,�-interaction) The C]Si bonding s orbital is higher in energy than C]H or C]C bonding s orbitals, the energy match of the C]Si s orbital with the p orbital of a carbon–carbon double bond is better than that with C]H or C]C bonds.Therefore, considerable interaction between the C]Si s orbital with the neighboring p orbital is attained in allylsilanes and benzylsilanes 4 (Fig. 3). Such interaction produces two new molecular orbitals. The energy level of the orbital produced by a bonding interaction is lower than the original two5 and that of the orbital produced by an antibonding interaction is higher than the original two.As the original orbitals are both filled, the two new orbitals should also be filled, and the orbital of higher energy becomes the HOMO. Therefore, the interaction between the C]Si s orbital and the p orbital of the carbon–carbon double bond causes a significant increase of the HOMO level which in turn favors electron transfer. The magnitude of such an interaction depends upon the torsion angle Si]C]C]C.As shown in Fig. 4 the HOMO level is maximum when the torsion angle is 908, where the C]Si s orbital and the p orbital of the carbon– carbon double bond are in the same plane. Fig. 5 shows the computer generated picture of the HOMO of the allylsilane (torsion angle 908) which exhibits the antibonding interaction Fig. 2 Schematic diagram of the selective and non-selective electron transfer on the surface of the electrode Fig. 3 Energy diagram of the interaction of the C]Si s orbital and the p orbital of the C]] C bond of the allylsilane Fig. 4 Plots of HOMO level of H3SiCH2CH]] CH2 vs. torsion angle of Si]C]C]C (MP2/LANL2DZ) (eV ª 1.602 × 1029 J) Fig. 5 The HOMO of H3SiCH2CH]] CH2 (HYPERCHEM 5,6 HF/ 3-21G*) Table 1 Oxidation potential of allylsilane and related compounds Compound Si(CH3)3 (CH3)4Si Ep/V (vs. Ag–AgCl) * 1.30 1.85 >2.5 * Determined with cyclic voltammetry using a Pt working electrode in LiClO4–CH3CN.J. Chem. Soc., Dalton Trans., 1998, Pages 2589–2596 2591 Scheme 1 Si(CH3)3 –e CH3OH OCH3 OCH3 + 69% ( 68 : 32) Si(CH3)3 –e –'Si(CH3)3 +" between the C]Si s orbital and the p orbital of the carbon– carbon double bond.The oxidation potentials of allylsilanes are less positive than those of the corresponding alkenes as shown in Table 1, indicating that the silyl group at the allylosition acts as an activating group for alkenes toward oxidation.7 The selective electrochemical oxidation of the carbon–carbon double bond neighboring the C]Si bond in geranyltrimethylsilane is remarkable (Scheme 1).The other carbon–carbon double bond is not aVected at all. It is also noteworthy that the C]Si bond is cleaved selectively in the cation radical intermediate 8 to give the allyl radical which is further oxidized to the allyl cation. The oxygen nucleophiles such as methanol are introduced to the allyl cation intermediate to produce a mixture of two regioisomeric products.Chemical oxidation and photoelectron transfer oxidation of allylsilanes have also been studied extensively. Such reactions proceed by closely related mechanisms to cleave the C]Si bond selectively.9 3.2 Electron transfer from heteroatoms (Û,n-interactions) The introduction of a silyl group at the carbon adjacent to a heteroatom (e.g. oxygen,10 sulfur 11 or nitrogen 12) causes a significant decrease of the oxidation potential (Table 2). In the case of the ether, the eVect of silicon is especially remarkable.The oxidation potential decreases at least 0.8 V by the introduction of a silyl group on the carbon adjacent to the oxygen atom. The decrease of the oxidation potential by silyl-substitution is attributed to the interaction between the C]Si s orbital and the non-bonding p orbital of the heteroatom (i.e. oxygen) increasing the HOMO level 13 as shown in Fig. 6. The important role played by the orbital interaction is demonstrated by the geometric dependence of the HOMO level. The plots of the HOMO level of H3SiCH2OH (a model compound of silylsubstituted ethers) obtained by ab initio molecular orbital calculations vs.the torsion angle Si]C]O]H is shown in Fig. 7. When the C]Si s orbital and the non-bonding p orbital of the oxygen atom are in the same plane (torsion angle 908) the HOMO level is at maximum. The HOMO shown in Fig. 8 Table 2 Oxidation potential of a-heteroatom-substituted heteroatom compounds and related compounds Compound C7H15 OMe C7H15 OMe SiMe3 C7H15 SPh C8H17 SPh SiMe3 Ep/V (vs.Ag–AgCl) * >2.5 1.72 1.20 1.10 * Determined with cyclic voltammetry using a Pt working electrode in LiClO4–CH3CN. Fig. 6 Energy diagram of the interaction of the C]Si s orbital and the non-bonding p orbital of oxygen Fig. 7 Plots of HOMO levels of H3SiCH2OH and H3SiCH2SH vs. torsion angle of Si]C]O]H and Si]C]S]H, respectively (MP2/ LANL2DZ) Fig. 8 The HOMO of H3SiCH2OH (HYPERCHEM 5, HF/3-21G*)2592 J.Chem. Soc., Dalton Trans., 1998, Pages 2589–2596 Scheme 2 OCH3 Si(CH3)3 C7H15 OH OCH3 C7H15 OH OCH3 OCH3 C7H15 OH C7H15 OH OCH3 Si(CH3)3 Si(CH3)3 OCH3 Si(CH3)3 C7H15 OH OCH3 C7H15 OH C7H15 OH OCH3 OCH3 OCH3 C7H15 OH O OCH3 C7H15 SPh Si(CH3)3 C7H15 SPh OCH3 C7H15 OCH3 OCH3 –e CH3OH + 78% (b) –4e CH3OH –'Si(CH3)3 +' (a) H2O 92% (c) –2e CH3OH –2e CH3OH 64% –e exhibits the antibonding interaction between the C]Si s orbital and the non-bonding p orbital. When they are in a perpendicular orientation (torsion angle 0 and 1808) the HOMO level is at a minimum.It is noteworthy that we can design molecules having specific oxidation potentials by controlling their geometry.14 The magnitude of the orbital interaction strongly depends on the nature of the heteroatom. In the case of sulfide the eVect of silyl-substitution is not so large. The molecular orbital calculations of a model compound, H3SiCH2SH (Fig. 7) are consistent with the experimental oxidation potentials (Table 2).This is probably because the energy match between the C]Si s orbital and the non-bonding p orbital of the sulfur is not good, thus the eVective interaction between the two orbitals cannot be attained (see below). The silyl group also controls the reaction pathway in preparative electrochemical oxidation. The oxidation of silyl-substituted heteroatom compounds leads to the selective cleavage of the C]Si bond and the introduction of nucleophiles on the carbon.Weakening of the C]Si bond by the interaction of the C]Si s orbital with the half-vacant p orbital of the cation radical of the oxygen 15 seems to be responsible for the facile C]Si bond cleavage. Several examples of the use of silicon in electrochemical oxidation of heteroatom compounds are shown in Scheme 2.16 In the case of b-hydroxy-a-trimethylsilyl ether [Scheme 2(a)], the C]Si bond in the initially formed cation radical is cleaved selectively without aVecting other parts of the molecule to give the carbon radical which is further oxidized to the cationic intermediate. The attack of methanol gives b-hydroxyacetal as the final product. The free hydroxy group is not aVected at all during the course of the reaction. The anodic oxidation of compounds having two C]Si bonds is interesting.Two C]Si bonds are cleaved successively to give an ester probably via the hydrolysis of the initially formed orthoester [Scheme 2(b)]. The electrochemical oxidation of silyl-substituted sulfides also takes place smoothly.The cleavage Table 3 Oxidation potential of acylsilane and related compounds Compound C9H19 O Si(CH3)3 C9H19 O H C6H13 O CH3 Ed/V (vs. Ag–AgCl) * 1.23 >2.00 >2.00 * Decomposition potential determined with rotating-disc voltammetry using a Pt working electrode in LiClO4–CH3CN. of the C]Si bond followed by the introduction of methanol gives the O,S-acetal which is further oxidized under the same conditions to give the acetal.In the example shown in Scheme 2(c), the carbon–carbon double bond is not aVected at all during the course of the transformation. Oxidation reactions of silicon-substituted heteroatom compounds by chemical oxidizing agents and photoelectron transfer oxidation reactions have also been studied extensively.17 Silicon also activates carbonyl compounds toward electron transfer (Table 3). The introduction of a silyl group on the carbonyl carbon (acylsilanes) causes a significant decrease in the oxidation potential.18 The interaction of the C]Si s orbital with the non-bonding orbital of the oxygen is responsible and this interaction is facilitated by the fixed geometry; the rotation around the carbon–oxygen double bond is prohibited (Fig. 9). This interaction is also responsible for other unique properties of acylsilanes.19 The HOMO of H3SiC(]] O)H is shown in Fig. Fig. 9 Energy diagram of the interaction of the C]Si s orbital and the non-bonding p orbital of carbonyl oxygen Fig. 10 The HOMO of H3SiC(]] O)H (HYPERCHEM 5, HF/3-21G*)J.Chem. Soc., Dalton Trans., 1998, Pages 2589–2596 2593 10 to demonstrate the antibonding interaction between the C]Si s orbital and the non-bonding p orbital of the carbonyl oxygen. The electrochemical oxidation of silyl-substituted carbonyl compounds (acylsilanes) proceeds smoothly to cleave the C]Si bond, and various nucleophiles such as alcohols, water and amides can be introduced to the carbonyl carbon. In the example shown in Scheme 3, allyl alcohol is eVectively introduced as a nucleophile to give the corresponding allyl ester.The oxidation of acylsilanes by chemical oxidizing agents such as H2O2 also cleaves the C]Si bond.20 Tin is also eVective for the oxidation of heteroatom compounds. The introduction of a stannyl group on the carbon adjacent to the heteroatom decreases the oxidation potential significantly (Table 4).21 The eVect of tin is attributed to the interaction between the C]Sn s orbital and the non-bonding p orbital of the heteroatom, increasing the HOMO level which in turn favors electron transfer. The variation of the HOMO levels of the model compounds H3SnCH2OH and H3SnCH2SH with torsion angle is shown in Fig. 11. Whereas silyl substitution causes little decrease in the oxidation potentials of sulfides, stannyl substitution causes a significant decrease. Although the energy level of the p orbital of sulfur is much higher than the C]Si s orbital and they do not interact with each other eVectively, the energy level of the C]Sn s orbital is similar to that of the sulfur p orbital and they interact eVectively to increase the HOMO level (Fig. 12). Therefore, tin exhibits significant eVects for electron transfer from organosulfur compounds, whereas silicon has little eVect. The reaction pattern of the oxidation of stannyl-substituted heteroatom compounds is quite similar to that of silylsubstituted heteroatom compounds.21 Electrochemical oxidation leads to selective cleavage of the C]Sn bond and the introduction of nucleophiles on the carbon.The eVect of tin is especially important for oxidative carbon– Scheme 3 C10H21 O Si(CH3)3 –e C10H21 O Si(CH3)3 C10H21 O –'Si(CH3)3' C10H21 O OH C10H21 O O 92% –e Table 4 Oxidation potential of a-heteroatom-substituted organosilicon compounds Compound C8H17 SPh Sn(C4H9)3 C7H15 OCH3 Sn(C4H9)3 C7H15 OCO2CH3 Sn(C4H9)3 C7H15 OCOCH3 Sn(C4H9)3 (C4H9)4Sn E2� 1 /V (vs. Ag–AgCl) * 0.97 1.47 1.44 0.74 1.52 * Determined with rotating-disc voltammetry using a Pt working electrode in LiClO4–CH3CN.carbon bond formation.22 During oxidation of heteroatom compounds the carbocation, stabilized by the adjacent heteroatom, is generated via proton elimination under suitable conditions. The reaction of this carbocation with carbon nucleophiles should lead to eVective carbon–carbon bond formation (Scheme 4).There are, however, two major problems. The first is the oxidation of carbon nucleophiles. The oxidation potential of carbon nucleophiles is often lower than that of the starting heteroatom compounds, and therefore it is often diYcult to oxidize the heteroatom compounds without aVecting the carbon nucleophiles. The second problem is overoxidation. Since the carbon–carbon bond formation products are also heteroatom compounds and carbon is slightly more electron donating than hydrogen, the oxidation potentials of the products are Fig. 11 Plots of HOMO levels of H3SnCH2OH and H3SnCH2SH vs. torsion angle of Sn]C]O]H and Sn]C]S]H, respectively (MP2/ LANL2DZ) Fig. 12 Energy diagram of the interaction of the C]Si s orbital and the C]Sn or orbital with the non-bonding p orbital of sulfur Scheme 4 Y R H Y R Y R Nu Y R Sn Y R Y R Nu –2e –H+ Nu– –2e –H+ Nu– Y = heteroatom Nu = carbon nucleophile2594 J. Chem. Soc., Dalton Trans., 1998, Pages 2589–2596 Scheme 5 OCO2CH3 C7H15 Sn(C4H9)3 OCO2CH3 C7H15 Sn(C4H9)3 –(C4H9)3Sn• OCO2CH3 C7H15 Si(CH3)3 OCO2CH3 C7H15 68% (a) (b) O C7H15 Sn(C4H9)3 –e O C7H15 BF4 – O C7H15 F 83% ( cis/ trans = 2.8 : 1) O Sn(C4H9)3 –e O (CH3)3SiCN O NC 84% (c) –e slightly less positive than those of the starting material.Therefore, oxidation of the product is diYcult to avoid. The use of tin as an activating group (electroauxiliary) solves these two problems. The introduction of tin causes a dramatic decrease in the oxidation potentials of heteroatom compounds.The oxidation potential of tin-substituted heteroatom compounds is less positive than that of various carbon nucleophiles such as enol silyl ethers and allylsilanes. Therefore, tin-substituted heteroatom compounds can be oxidized selectively in the presence of these carbon nucleophiles. The carbocation intermediate generated by the cleavage of the C]Sn bond reacts with carbon nucleophiles to give the product. Since the product does not contain tin, its oxidation potential is higher than the starting tin-containing compounds.23 Therefore, there is no longer a problem with overoxidation.Several examples of electrooxidative carbon–carbon bond formation using tinsubstituted heteroatom compounds are shown in Scheme 5. Chemical oxidation reactions of a-heteroatom-substituted organotin compounds aimed at achieving carbon–carbon bond formation have also been developed.24 3.3 Electron transfer from Û systems (Û,Û-interaction) The interaction of a carbon–metal (metal = Si or Sn) s orbital with a neighboring carbon–metal s orbital is also eVective in raising the HOMO level.25 As a matter of fact, the oxidation potential of 1,2-bis(trimethylsilyl)ethane is much less positive Table 5 Oxidation potentials of 1,2-disilyl- and 1,2-distannylsubstituted ethanes and related compounds Compound Me3Si Me3Si SiMe3 Me3Si SiMe3 Bu3Sn SiMe3 Bu3Sn Me3Si SiMe3 Ed/V (vs.Ag–AgCl)* 2.19 2.20 1.74 2.02 1.45 1.22 Compound Bu3Sn SnBu3 SiMe3 SiMe3 SiMe3 SiMe3 Ed/V (vs.Ag–AgCl)* >2.5 1.65 1.41 0.66 * Decomposition potential determined with rotating-disc electrode voltammetry in LiClO4–CH3CN using a glassy carbon working electrode and a Ag–AgCl reference electrode. than that of (trimethylsilyl)ethane which has only one C]Si bond (Table 5).26 The lower oxidation potential of 1,2-bis(trimethylsilyl) ethane is attributed to the interaction between the C]Si s orbitals raising the HOMO level (Fig. 13). The HOMO of H3SiCH2CH2SiH3 is shown in Fig. 14, which indicates the antibonding interaction between two neighboring C]Si s orbitals. It is also noteworthy that the oxidation potential of 1,2-bis(trimethylsilyl)ethane is much lower than those of bis(trimethylsilyl)methane and 1,3-bis(trimethylsilyl)propane, indicating the importance of the interaction of the two neighboring orbitals. The geometric requirements of the present Fig. 13 Energy diagram of the interaction of two neighboring C]Si s orbitals Fig. 14 The HOMO of H3SiCH2CH2SiH3 (HYPERCHEM 5, HF/3- 21G*)J. Chem. Soc., Dalton Trans., 1998, Pages 2589–2596 2595 eVect are also interesting. The oxidation potential of cis- (exo,exo)-1,2-bis(trimethylsilyl)norbornane (Si]C]C]Si torsion angle ca. 18) is less positive than that of trans-1,2-bis(trimethylsilyl) norbornane (torsion angle ca. 1078). This is consistent with the molecular orbital calculations which indicate that the HOMO is at a minimum when the torsion angle is 908 (Fig. 15).The concept of s–s interaction is also applicable to organotin compounds.26 The oxidation potential of 1,2-bis(tributylstannyl) ethane is much lower than that of (tributylstannyl)- ethane, indicating significant interaction between the two C]Sn s orbitals raising the HOMO level. It is also interesting that the C]Si s orbital and the C]Sn s orbital interact with each other less eVectively, as demonstrated by the oxidation potential of 1-tributylstannyl-2-(trimethylsilyl)ethane. Probably the energy match between the two s orbitals is not good for eVective interaction because the energy level of the C]Sn s orbital is much higher than that of the C]Si s orbital.The preparative electrochemical oxidation of 1,2-bis(trimethylsilyl) norbornane results in the cleavage of two C]Si bonds to yield norbornene (Scheme 6).26 Presumably one of the C]Si bonds is cleaved in the cation radical intermediate to generate the carbon radical which is further oxidized to give the carbocation.The subsequent facile b-elimination of the silyl group produces a carbon–carbon double bond. 4 Combination of the Û,�-Interaction System and the Û,Û-Interaction System In the previous part of this account it was demonstrated that the C]Si and C]Sn s orbitals interact with neighboring p-systems, n-systems and s-systems to raise the HOMO level which in turn favors electron transfer. It is interesting to examine whether such interaction systems interact with each other. 1,2-Diphenyl-1,2-bis(trimethylsilyl)ethane is an example of a system which involves s,p-interaction systems and a s,s-inter- Fig. 15 Plots of HOMO level of H3SiCH2CH2SiH3 vs. torsion angle of Si]C]C]Si (MP2/LANL2DZ) Scheme 6 Si(CH3)3 Si(CH3)3 –e (3.0 F mol–1) CH3OH Si(CH3)3 Si(CH3)3 Si(CH3)3 Si(CH3)3 96% –e action system. The oxidation potential of this compound is less positive than that of benzyltrimethylsilane (s,p-interaction system) and that of 1,2-bis(trimethylsilyl)ethane (s,s-interaction system), indicating that the two systems interact with each other eVectively to raise the HOMO level (Table 6).27 The molecular orbital calculations also indicate that the HOMO is localized through two C]Si bonds and two benzene rings as shown in Fig. 16. This fact demonstrates the potential of the conjugation between the s,s-interaction systems and the s,p-interaction systems to construct various new electronic systems.Such eVve conjugation provides a new concept in material science. 5 Conclusion It is hoped that with the aid of molecular orbital calculations and computer graphics this Perspective helps to demonstrate to the reader that silicon and tin promote electron transfer from b-situated p systems, heteroatoms and s systems, and that this eVect is attributed to the interaction of C]Si or C]Sn s orbitals with neighboring p-type orbitals. Such s–p interaction, s–n interaction and s–s interaction enjoy versatile applications in the field of electron transfer.Several examples of the redox selective reactions of organo-silicon and -tin compounds based on these orbital interactions are demonstrated in this Perspective, although many other examples are omitted because of space limitations. Further work is in progress to explore the full scope of this concept and its application to synthetic transformations. 6 Acknowledgements We are pleased to acknowledge the Grant-in-Aid for Scientific Research from Monbusho, Nagase Science and Technology Fig. 16 The HOMO of PhCH(SiMe3)CH(SiMe3)Ph (HYPERCHEM 5, PM3) Table 6 Oxidation potential of 1,2-diphenyl-1,2-bis(trimethylsilyl)- ethane and related compounds Compound Me3Si SiMe3 SiMe3 SiMe3 SiMe3 meso DL Ed/V (vs. Ag–AgCl) * 1.23 1.32 1.39 1.74 * Decomposition potential determined with rotating-disc electrode voltammetry using a glassy carbon working electrode in LiClO4–CH3CN.2596 J.Chem. Soc., Dalton Trans., 1998, Pages 2589–2596 Foundation, Yamada Science Foundation, Nihon Noyaku, and Kanegafuchi Chemical Industry for partial financial support of this work. We also thank Shin-etsu Chemical for the gifts of some organosilicon compounds. We are also grateful to the many dedicated co-workers, whose names appear in the references. 7 References 1 Silicon: E. W. Colvin, Silicon in Organic Synthesis, Butterworths, London, 1981; W. P.Weber, Silicon Reagents for Organic Synthesis, Springer-Verlag, Berlin, 1983; tin: M. Pereyre, J.-P. Quintard and A. Rahm, Tin in Organic Synthesis, Butterworths, London, 1986. 2 J. Yoshida, in Top. Curr. Chem., 1994, 170, 39; J. Yoshida, T. Murata, S. Matsunaga, T. Maekawa, S. Shiozawa and S. Isoe, Rev. Heteroat. Chem., 1991, 5, 193; J. Yoshida, Yuki Gosei Kagaku Kyokaishi, 1996, 53, 53. 3 C. Amatore, in Organic Electrochemistry, eds. H. Lund and M. M. Baizer, Marcel Dekker, New York, 1991, p. 11. 4 H. Bock, Angew. Chem., Int. Ed. Engl., 1989, 28, 1627; J. C. Giordan, J. Am. Chem. Soc., 1983, 105, 6544; T. G. Traylor, W. Hanstein, H. J. Berwin, N. A. Clinton and R. S. Brown, J. Am. Chem. Soc., 1971, 93, 5715; R. S. Brown, D. F. Eaton, A. Hosomi, T. G. Traylor and J. M. Wright, J. Organomet. Chem., 1974, 66, 249. 5 I. Fleming, Frontier Orbitals and Organic Chemical Reactions, Wiley, New York, 1976; T. A. Alberight, J. K. Burdett and M.-H. Whangbo, Orbital Interactions in Chemistry, Wiley, New York, 1985. 6 HYPERCHEM, release 5.0, Hypercube Inc. 7 J. Yoshida, T. Murata and S. Isoe, Tetrahedron Lett., 1986, 27, 3373; J. Yoshida, T. Murta and S. Isoe, Tetrahedron Lett., 1987, 28, 211; H. Schäfer, D. Hermeling and K. H. Lange, Spring Meeting of the Electrochemical Society, Cincinnati, OH, Extend Abstracts, 1984, p. 441; T. Koizumi, T. Fuchigami and T. Nonaka, Chem. Express, 1986, 1, 355; T. Koizumi, T. Fuchigami and T. Nonaka, Bull. Chem.Soc. Jpn., 1989, 62, 219. 8 H. Bock and W. Kaim, J. Am. Chem. Soc., 1980, 102, 4429; M. Kira, H. Nakazawa and H. Sakurai, Chem. Lett., 1985, 1845. 9 M. Ochiai, M. Arimoto and E. Fujita, Tetrahedron Lett., 1981, 22, 4491; K. Ohga and P. S. Marianao, J. Am. Chem. Soc., 1982, 104, 617; T. Fujii, T. Hirao and Y. Ohshiro, Tetrahedron Lett., 1993, 34, 5601; K. Mizuno, M. Ikeda and Y. Otsuji, Tetrahedron Lett., 1985, 26, 461; K. Mizuno, K. Terasaka, M. Yasueda and Y. Otsuji, Chem.Lett., 1988, 145; K. Mizuno, M. Ysasueda and Y. Otsuji, Chem. Lett., 1988, 229; E. Baciocchi, T. Del Giacco, C. Rol and G. V. Sebastiani, Tetrahedron Lett., 1989, 30, 3573. 10 J. Yoshida, T. Murata and S. Isoe, J. Organomet. Chem., 1988, 345, C23; J. Yoshida, S. Matsunaga and S. Isoe, Tetrahedron Lett., 1989, 30, 219; J. Yoshida, S. Matsunaga, T. Murata and S. Isoe, Tetrahedron, 1991, 47, 615; J. Yoshida, H. Tsujishima, K. Nakano and S. Isoe, Inorg. Chim. Acta, 1994, 220, 129. 11 (a) J. Yoshida and S. Isoe, Chem. Lett., 1987, 631; (b) B. E. Cooper and W. J. Owen, J. Organomet. Chem., 1971, 29, 33. 12 J. Yoshida and S. Isoe, Tetrahedron Lett., 1987, 28, 6621, see also ref. 11(b). 13 J. Yoshida, T. Maekawa, T. Murata, S. Matsunaga and S. Isoe, J. Am. Chem. Soc., 1990, 112, 1962. 14 K. Nishiwaki and J. Yoshida, Chem. Lett., 1996, 171. 15 E. Block, A. J. Yencha, M. Aslam, V. Eswarakrishnan, J. Luo and A. Sano, J. Am. Chem. Soc., 1988, 110, 4748; M. Kira, H.Nakazawa and H. Sakurai, Chem. Lett., 1986, 497. 16 J. Yoshida, S. Matsunaga and S. Isoe, Tetrahedron Lett., 1989, 30, 219; J. Yoshida, S. Matsunaga, T. Murata and S. Isoe, Tetrahedron, 1991, 47, 615; J. Yoshida, T. Maekawa, Y. Morita and S. Isoe, J. Org. Chem., 1992, 57, 1321; K. Suda, K. Hotoda, J. Watanabe, K. Shiozawa and T. Takanami, J. Chem. Soc., Perkin Trans. 1, 1992, 1283; K. Suda, K. Hotoda, F. Iemuro and T. Takanami, J. Chem. Soc., Perkin Trans. 1, 1993, 1553. 17 M.A. Brumfield, S. L. Quillen, U. C. Yoon and P. S. Mariano, J. Am. Chem. Soc., 1984, 106, 6855; E. Hasegawa, W. Xu, P. S. Mariano, U.-C. Yoon and J.-U. Kim, J. Am. Chem. Soc., 1988, 110, 8099. 18 J. Yoshida, S. Matsunaga and S. Isoe, Tetrahedron Lett., 1989, 30, 5293; J. Yoshida, M. Itoh, S. Matsunaga and S. Isoe, J. Org. Chem., 1992, 57, 4877; K. Mochida, S. Okui, K. Ichikawa, O. Kanakubo, T. Tsuchiya and K. Yamamoto, Chem. Lett., 1986, 805. 19 P. C. B. Page, S. S. Klair and S. Rosenthal, Chem. Soc. Rev., 1990, 19, 147; A. G. Brook, Acc. Chem. Res., 1974, 7, 77; A. Ricci and A. Degl’Innocenti, Synthesis, 1989, 647. 20 J. A. Miller and G. Zweifel, Synthesis, 1981, 288; J. A. Miller and G. Zweifel, J. Am. Chem. Soc., 1981, 103, 6217. 21 J. Yoshida, Y. Ishichi, K. Nishiwaki, S. Shiozawa and S. Isoe, Tetrahedron Lett., 1992, 33, 2599; R. S. Glass, A. M. Radspinner and W. P. Singh, J. Am. Chem. Soc., 1992, 114, 4921. 22 H. J. Schafer, Angew. Chem., Int. Ed. Engl., 1981, 20, 911. 23 J. Yoshida, M. Itoh and S. Isoe, J. Chem. Soc., Chem. Commun., 1993, 547; J. Yoshida, Y. Morita, M. Itoh and S. Isoe, Tetrahedron Lett., 1994, 35, 5247; J. Yoshida, M. Itoh, Y. Morita and S. Isoe, J. Chem. Soc., Chem. Commun., 1994, 549; J. Yoshida, Y. Ishichi and S. Isoe, J. Am. Chem. Soc., 1992, 114, 7594; J. Yoshida, K. Takada, Y. Ishichi and S. Isoe, J. Chem. Soc., Chem. Commun., 1994, 2361. 24 K. Narasaka, T. Okauchi and N. Arai, Chem. Lett., 1992, 1229; J. Yamaguchi, Y. Takagi, A. Nakayama, T. Fujiwara and T. Takeda, Chem. Lett., 1991, 133. 25 A. Hosomi and T. G. Traylor, J. Am. Chem. Soc., 1975, 97, 3682. 26 J. Yoshida, T. Teramoto, K. Nishiwaki and S. Isoe, The Abstract of 16th Symposium on Electroorganic Chemistry, Tokyo, 1994, 83. 27 K. Nishiwaki and J. Yoshida, Chem. Lett., 1996, 787; J. M. Proter, X. Xuan, B. Blackman, D. Hsu and A. J. Fry, Tetrahedron Lett., 1997, 38, 7147. Received 5th May 1998; Paper 8/03343I

ISSN:1477-9226    

DOI:10.1039/a803343i

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, Pages 2597–2598 2597 pH-Induced switching of metal ion co-ordination: the structure of [Pd([18]aneN2S4?2H1)][BF4]4?2H2O from a twinned crystal ([18]aneN2S4 5 1,4,10,13-tetrathia-7,16-diazacyclooctadecane) Alexander J. Blake,a Robert O. Gould,b Gillian Reid *,c and Martin Schröder *,a a School of Chemistry, University of Nottingham, University Park, Nottingham, UK NG7 2RD b Department of Chemistry, University of Edinburgh, King’s Buildings, West Mains Road, Edinburgh, UK EH9 3JJ c Department of Chemistry, University of Southampton, Highfield, Southampton, UK SO17 1BJ Treatment of [Pd([18]aneN2S4)][BF4]2 ([18]aneN2S4 = 1,4,10,13- tetrathia-7,16-diazacyclooctadecane) with HBF4 led to a marked structural change at Pd(II) from distorted octahedral [N2S2 1 S2]-donation to a square-planar S4-donor coordination via protonation of the secondary amine functions; this pH-induced re-arrangement is accompanied by a reversible colour change from purple to yellow in MeCN solution.The chemistry of macrocyclic ligand complexes has been an area of considerable interest for some years.1 This is mainly due to the increased stability imparted by the macrocyclic eVect and also the conformational flexibility of saturated ring systems which can accommodate a range of metal stereochemistries and oxidation states.2 We have been interested in the interaction of transition metal centres with mixed thia/aza macrocycles since these ligands combine soft thioether donors of low basicity capable of binding transition metal ions, and harder, more basic amine functions within a cyclic configuration.3 We have reported previously the preparation, structural characterisation and redox properties of Pd(II) and Pt(II) complexes involving the macrocyclic ligand [18]aneN2S4 (1,4,10,13-tetrathia-7,16- diazacyclooctadecane) and its di-N-methylated derivative Me2[18]aneN2S4 (7,16-dimethyl-1,4,10,13-tetrathia-7,16-diazacyclooctadecane). 4 Metal(II) complexes of these ligands exhibit strikingly diVerent electrochemical properties, and these differences were traced back to the very diVerent structures adopted by the Pd(II) complexes of these two ligands.Thus [Pd([18]aneN2S4)]21 shows a distorted square planar N2S2-coordination with additional, long-range interactions to two apical S donor atoms, Pd ? ? ?S = 2.954(4), 3.000(3) Å, and exhibits a reversible Pd(II)/Pd(III) redox couple at 10.57 V vs.Fc/Fc1. In contrast, [Pd(Me2[18]aneN2S4)]21 shows a distorted square-planar S4-co-ordination with the tertiary amine functions directed away from the metal centre. This complex exhibits a reversible Pd(II)/Pd(I) redox couple at 20.74 V vs. Fc/Fc1, with no oxidative activity observed by cyclic voltammetry. In view of these unexpected diVerences in redox behaviour and their correlation with the diVering stereochemistries adopted by the complexes, we wished to establish whether stereo- S S N N S S R R R = Me: Me2[18]aneN2S4 R = H: [18]aneN2S4 chemical switching of co-ordination of [18]aneN2S4 to Pd(II) might be induced by changes in pH.Thus, addition of 40% aqueous HBF4 to a MeCN solution of [Pd([18]aneN2S4)][BF4]2 † Fig. 1 View of the structure of the [Pd([18]aneN2S4?2H1)]21 cation showing the numbering scheme adopted. The two weakly interacting BF4 2 anions and the N-based protons are also included while the other two BF4 2 anions, the methylene protons and H2O solvent molecules are omitted for clarity.Ellipsoids are shown at 40% probability. Selected bond lengths (Å) and angles (8): Pd]S(4) 2.3159(7), Pd]S(7) 2.3432(6), Pd ? ? ? F(7) 3.122(2); S(4)]Pd]S(7) 88.52(3), F(7) ? ? ? Pd]S(4) 92.85(4), F(7) ? ? ? Pd]S(7) 112.07. † Preparation of [Pd([18]aneN2S4?2H1)][BF4]4. An analytically pure sample of [Pd([18]aneN2S4)][BF4]2 (0.020 mg, 0.033 mmol) was dissolved in MeCN (2 cm3) giving a purple solution.Aqueous HBF4 (40%, 1 drop) was then added giving an immediate colour change to yellow. Upon standing for several days, yellow crystals were obtained in quantitative yield which were filtered and dried in vacuo. This product is very stable in the presence of HBF4, but the starting material is readily regenerated in the absence of acid, and this hampered our eVorts to obtain pure samples for microanalysis. Electrospray mass spectrum (MeCN): found m/z = 433; calculated for [106Pd([18]aneN2S4 1 H)]1 m/z = 433.UV/VIS spectrum [MeCN–HBF4(aq)]: lmax = 274 nm (emol ca. 11 300 dm3 mol21 cm21), 311 (sh) (ca. 3950). IR spectrum (CsI disk): 3400vs (br), 3220s, 2960w, 1635s, 1436m, 1386m, 1198m, 1072s (br), 884w, 795m, 646w, 549m, 527w, 502w cm21.2598 J. Chem. Soc., Dalton Trans., 1998, Pages 2597–2598 leads to an immediate colour change from red-purple of the parent 21 cation to yellow suggesting that indeed a significant stereochemical rearrangement is occurring.This is a totally reversible process with the red-purple solution being readily regenerated near neutral pH. Slow evaporation of a solution of the protonated complex in MeCN and aqueous HBF4 over several days furnished golden yellow, columnar crystals. A single crystal X-ray structure determination ‡ of this protonated complex confirms (Fig. 1) a centrosymmetric 41 cation in which the Pd(II) ion is co-ordinated to a distorted square-planar array of four thioether donors, Pd]S(4) = 2.3159(7), Pd]S(7) = 2.3432(6) Å, with each of the secondary amine centres protonated and directed away from the metal centre. In addition, there are long-range, apical interactions between the Pd(II) ion and one F ‡ X-Ray crystallography and crystal data for [Pd([18]aneN2S4?2H1)]- [BF4]4?2H2O.The selected crystal was coated with mineral oil, mounted on a glass fibre and immediately placed under a stream of cold nitrogen. Data collection used a Stoe Stadi-4 four-circle diVractometer equipped with an Oxford Cryosystems open-flow cryostat operating at 150 K, and graphite-monochromated Mo-Ka radiation (0.710 73 Å) using w–2q scans.C12H32B4F16N2O2PdS4, M = 818.28, monoclinic, space group P21/c, a = 10.068(3), b = 12.449(3), c = 11.179(3) Å, b = 106.20(2)8 (cell 1), U = 1345.5(6) Å3, Z = 2, Dc = 2.020 g cm23, m(Mo- Ka) = 11.30 cm21, F(000) = 816. Yellow block (0.50 × 0.35 × 0.25 mm). 4751 Reflections were measured to 2qmax = 508 and absorption corrected (y scans) on a monoclinic cell with a = 11.179(3), b = 12.449(3), c = 20.154(6) Å, b = 106.20(2)8 (cell 2).These data could not be sensibly assigned to a space group. There were no significant data for h,k,l with h = 2n, l = 2n 1 1, and while data with h, k and l all even had a mean value of E2 of 3.2, those with h, k and l all odd had ·E2Ò = 2.4 and other parity groups had mean values ranging between 0.2 and 0.9. The problem was resolved by treating the data as arising from a twinned crystal with dimensions of cell 1 above.This twinning arises because for this cell |a| ª |a 1 c/2|, and both make the same angle with c. The two unequal components have the twin matrix 21 0 2��� /0 21 0/0 0 1. In terms of the cell used for data collection, h,k,l combine the actual 2l9, k9, 22h9 for one component with l0, k0, 2h0 1 l0 for the other. Consequently, the first component contributes only to data with l = 2n and the second only to data with h and l either both even or both odd.The intensity distribution of the measured data clearly indicates that the structure contains a heavy atom on an inversion centre, and that the second component as described above is clearly predominant; the structure was solved roughly (direct methods) 5 by assuming that it was the entire structure and rejecting all other data. The accompanying computer program (deposited) was used to divide the data appropriately, giving 3272 unique reflections of which 2813 with F > 4s(F) were used in all calculations, and least-squares refinement proceeded smoothly using SHELXL,6 the final contribution of the first component being 0.2809(11) of that of the second.One half cation, two BF4 2 anions and one H2O solvent molecule were identified in the asymmetri. All non-H atoms were refined with anisotropic thermal parameters and H atoms associated with the C and N atoms were included in fixed, calculated positions, while those associated with the H2O molecules were located from the diVerence map, and included but not refined.At final convergence, R1 = 0.0292, wR2 = 0.0739 [F > 4s(F)] and R1 = 0.0408, wR2 = 0.0863 (all data) (based upon least-squares refinement on F2), S = 1.088 for 91 parameters and the final DF synthesis showed Dr in the range 0.53 to 20.55 e Å23. CCDC reference number 186/1080. atom of each of two BF4 2 anions, Pd ? ? ? F(7) = 3.122(2) Å.The other two BF4 2 anions are non-co-ordinating, although they are involved in significant H-bonding interactions with the protonated amine groups and H2O solvent molecules within the lattice. This H-bonding leads to an intricate network with N(1) ? ? ? F(1) = 2.868, N(1) ? ? ? O(1W) = 2.797, O(1W) ? ? ? F(2) = 2.976, O(1W) ? ? ? F(5) = 3.231, O(1W) ? ? ? F(6) = 2.918 Å. The primary co-ordination at the metal centre in this species is very similar to that in [Pd(Me2[18]aneN2S4)]21 with similar Pd]S bond distances [2.3261(22), 2.3239(21) Å].4 The electrospray mass spectrum of the yellow product (MeCN solution) shows peaks with the correct isotopic distribution at m/z = 433, consistent with [106Pd([18]aneN2S4 1 H)]1.The IR spectrum shows peaks associated with co-ordinated [18]aneN2S4 and BF4 2 anion, and strong absorptions due to H2O are also apparent and mask the N]H stretching region. The UV/VIS spectrum of [Pd([18]aneN2S4)][BF4]2 in MeCN solution shows transitions at lmax = 514 nm (emol 124 dm3 mol21 cm21), 322 (2815), 266 (9420) and 233 (12 140), and addition of a single drop of 40% aqueous HBF4 leads to the loss of these absorptions, with a new band appearing at 274 nm (ca. 11 300 dm3 mol21 cm21) together with a shoulder at 311 (ca. 3950) corresponding to [Pd([18]aneN2S4?2H1)][BF4]4?2H2O. These results further exemplify the considerable chemical and structural diversity associated with metal complexes involving thia/aza macrocycles and the complex reported represents the first example of pH dependence in these systems.It suggests also that mixed thioether–aza macrocycles might be avid potential metal-ion extractors at low pH with the corresponding anion associated with the protonated amine function. Acknowledgements We thank the EPSRC for support, and Johnson Matthey plc for generous loans of PdCl2. References 1 The Chemistry of Macrocyclic Ligand Complexes, ed. L. F. Lindoy, Cambridge University Press, Cambridge, 1989; Supramolecular Chemistry, ed. F. Vögtle, John Wiley and Sons, Chichester, 1991. 2 A. J. Blake and M. Schröder, Adv. Inorg. Chem., 1990, 35, 1; S. R. Cooper and S. R. Rawle, Struct. Bonding (Berlin), 1990, 71, 1 and refs. therein. 3 G. Reid and M. Schröder, Chem. Soc. Rev., 1990, 19, 239. 4 A. J. Blake, G. Reid and M. Schröder, J. Chem. Soc., Dalton Trans., 1990, 3363; G. Reid, A. J. Blake, T. I. Hyde and M. Schröder, J. Chem. Soc., Chem. Commun., 1988, 1397; J. P. Danks, N. R. Champness and M. Schröder, Coord. Chem. Rev., in the press. 5 G. M. Sheldrick, SHELXS 86, program for crystal structure solution, Acta Crystallogr., Sect. A, 1990, 46, 467. 6 G. M. Sheldrick, SHELXL 93, program for crystal structure refinement, University of Göttingen, 1993. Received 4th June 1998; Communication 8/04205E

ISSN:1477-9226    

DOI:10.1039/a804205e

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, Pages 2599–2600 2599 Synthesis and characterization of the organocopper–copper halide complex [CuMes*{Cu2Br2(SMe2)3}] (Mes* 5 C6H2But 3-2,4,6) Cheong-Soo Hwang, Marilyn M. Olmstead, Xiaoming He and Philip P. Power* Department of Chemistry, University of California, Davis, CA 95616, USA Treatment of 3 equivalents of CuBr with LiMes* (Mes* = C6H2But 3-2,4,6) in Et2O at ca. 278 8C aVorded, upon treatment with SMe2, the organocopper–copper halide complex [CuMes*{Cu2Br2(SMe2)3}] 1 which has the previously unobserved CuR:CuX ratio of 1 : 2; it is a very rare example of a structurally characterized CuR:CuX complex without chelating R groups.Organocopper–copper halide aggregates are an important and growing class of organocopper compounds.1 They are characterized by diVerent organocopper to copper halide ratios. To date, well defined, aggregates with the CuR/CuX (R = alkyl or aryl; X = halide) ratios of 2 : 1,2 2:2,2,3 2:3,4 4:25 have been characterized. In many instances chelating aryl ligands such as C6H4(CH2NMe2)-2 6 have played a key role in stabilizing these complexes.During investigations of the reaction of LiMes* (Mes* = C6H2But 3-2,4,6) 7 with copper bromide in diethyl ether solution, it was observed that most of the copper bromide appeared to have reacted when less than 0.5 equivalent of LiMes* was added. This observation suggested that an organocopper–copper halide, possibly of previously unobserved 1 : 2 stoichiometry, had formed.In this paper the synthesis and characterization of this new complex [CuMes*- {Cu2Br2(SMe2)3}] 1 are now reported. Compound 1 was synthesized † by the addition of LiMes* to CuBr in Et2O at ca. 278 8C. Warming to room temperature and the addition of dimethyl sulfide, followed by filtration and cooling in a 220 8C freezer, aVorded colorless crystals of 1 in moderate yield. Proton NMR spectroscopy of 1 in C6D6 solution indicated a 3: 1 ratio of SMe2 to Mes* groups.The Cu]Mes* bonding was indicated by the appearance of an ipso carbon resonance at d 167.50 in the 13C-{1H} NMR spectrum which is within the known range for ipso carbon shifts in arylcopper/arylcuprate solutions in SMe2.8 However, the exact † Under anaerobic and anhydrous conditions LiMes* (0.504 g, 2 mmol) in Et2O (20 mL) was added to a well-stirred suspension of CuBr (0.86 g, 6 mmol) in Et2O (20 mL) with cooling in a dry ice bath. After ca. 1 h, the mixture was allowed to come to room temperature whereupon ca. 2 mL of SMe2 was added. Stirring was continued for ca. 3 h and the pale yellow solution was filtered. The volume of the solution was reduced to ca. 10 mL and it was then stored in a 220 8C freezer for 24 h to aVord the product 1 as colorless crystals. Yield 0.75 g, 0.96 mmol, 48%; mp 127 8C (decomp.). It has not been possible to obtain a satisfactory combustion analysis of 1 owing to desolvation of the crystals.However, atomic absorption spectroscopy indicates an approximate Cu : Br ratio of 3 : 2. 1H NMR (C6D6, 25 8C): d 7.59 (br s, 2 H, m-C6H2); 1.92 (br s, 18 H, SMe3); 1.71 [s, 18 H, o-C(CH3)3]; 1.56 [s, 9 H, p-C(CH3)2]. 13C-{1H} NMR (C6D6, 25 8C): d 167.50 (i-C6H2), 150.33 (o-C6H3); 120.37 (p- C6H3); 119.66 (m-C6H3); 38.53 [o-C(CH3)3], 35.09 [p-C(CH3)3]; 33.38 [o- C(CH3)3]; 30.65 [p-C(CH3)3]; 18.78 (SMe2). The broad singlet obtained for the SMe2 signal is probably due to rapid exchange of SMe2 between the copper sites.Cooling the spectrum to 260 8C did not result in splitting of the signal. structure of 1 was established by X-ray crystallography.† The illustration in Fig. 1 indicates that one CuMes* unit is associated with two copper bromides to form a very unusual six-membered ring composed of three coppers, two bromides and an ipso carbon from the Mes* group. The Cu3Br2C array is almost planar (average deviation = 0.009 Å), but there are gross variations in the angles within the ring.There is an almost perpendicular angle of 88.28 between the plane of the Mes* ring and the Cu3Br2C core. The ring distances are also quite variable; the shortest involve the C(1)]Cu(2)]Br(2) unit where Cu(2)]C(1) and Cu(2)]Br(2) bond lengths of 1.972(14) and 2.295(3) Å, respectively are observed. These may be compared to the much longer Cu(1)]C(1) and Cu(3)]Br(2) distances of 2.09(2) Å and 2.558(4) Å. These structural data as well as the near linear co-ordination at Cu(2) [C(1)]Cu(2)]Br(2) 170.6(4)8] suggest that the compound can be viewed as a contact ion pair composed of a [BrCuMes*]2 anion and a [(Me2S)2Cu- (m-Br)Cu(SMe2)]1 cation.This view is further supported by the fact that the deviation of the Cu(2)]C(1) vector from the Mes* ring plane is 27.58 whereas the deviation for the Cu(1)]C(1) vector is 47.48. In the anion, the Cu]C and Cu]Br bond lengths are just slightly longer than those observed in the solvent separated anion [Cu(Br){CH(SiMe3)2}]2 in which Cu]C = 1.920(0) Å and Cu]Br = 2.267(2) Å.9 The longer distances in Fig. 1 Computer generated drawing of 1 with H atoms not shown. Selected bond distances (Å) and angles (8): Cu(1)]C(1) 2.09(2), Cu(1)]S(3) 2.306(6), Cu(1)]Br(1) 2.420(4), Cu(2)]C(1) 1.972(14), Cu(2)]Br(2) 2.295(3), Cu(3)]Br(1) 2.510(3), Cu(3)]Br(2) 2.558(4), Cu(3)]S(1) 2.278(5), Cu(3)]S(2) 2.267(5); C(1)]Cu(1)]S(3) 111.5(4), C(1)]Cu(1)]Br(1) 142.6(4), S(3)]Cu(1)]Br(1) 105.7(2), Cu(1)] C(1)]Cu(2) 50.4(4), C(1)]Cu(2)]Br(2) 170.6(4), Cu(2)]Br(2)]Cu(3) 88.95(11), Br(1)]Cu(3)]Br(2) 113.8(2), Cu(1)]Br(1)]Cu(3) 110.6(2), S(1)]Cu(3)]S(2) 119.8(2), C(2)]C(1)]C(6) 115.5(13).‡ Crystal data at 130 K with Mo-Ka (l = 0.710 73) radiation for C24H47- Br2Cu3S3: M = 782.24, a = 14.510(5), b = 18.993(5), c = 11.448(3) Å, U = 3155(2) Å3, orthorhombic, space group Pca21, m = 4.747 mm21, Z = 4, R1 = 0.085 for 2048 [I > 2(s)I] data, wR2 = 0.1874 for all 3831 data, full-matrix least squares based on F2.CCDC reference number 186/1079.2600 J. Chem. Soc., Dalton Trans., 1998, Pages 2599–2600 the anion of 1 can be attributed to the increased co-ordination numbers as a result of its association with the cation. The Cu]S distances observed in 1 lie in the middle of the currently known range [ca. 2.185(1) 10–2.383(2) Å 11] for SMe2 complexes of organocopper species. The Cu(1)]S(3) bond length in 1 is 2.306(6) Å, which is longer than the 2.267(5) and 2.278(5) Å Cu]S distances observed for Cu(3).The longer Cu(1)]S(3) distance is surprising in view of the fact that the Cu(1) coordination number is lower than that of Cu(3). Part of the explanation may be that, since Cu(1) is also bound to the Mes* ligand, its environment may be more crowded which leads to a lengthened Cu]S bond. The structural arrangement of 1 bears some resemblance to that of the chelated trimetallic species Cu3(Br){C6H4[CH2N- (Me)CH2CH2NMe2]-2}2 2.2 In this molecule the anion has the formula [Cu{C6H4[CH2N(Me)CH2CH2NMe2]}]2 with Cu]C distances that average 1.967(11) Å which is very similar to the Cu(2)]C(1) distance in 1.Two further Cu1 ions are complexed by the two N donors in each ‘arm’ of the ligand. These metals also interact with an ipso carbon on each aryl ring and have Cu]C distances that average 2.095(14) Å, which is identical to the Cu(1)]C(1) length in 1. The structure is completed by a Br2 ion which bridges the two amine complexed coppers to aVord Cu]Br distances that average 2.435(8) Å.This length is very like the 2.420(4) Å seen for the Cu(1)]Br(1) bond in 1. Longer bond lengths are seen for Cu(3) where distances of 2.510(3) and 2.558(4) Å were observed. This is probably a result of the higher co-ordination number at Cu(3). The structure of 1 is also characterized by a relatively short Cu(1) ? ? ? Cu(2) distance of 2.471(4) Å. This is longer than the corresponding average of 2.406(3) Å in 2 2 but close to the 2.443(1) Å seen in the unusual dimer [(Me2S)2Cu(m-C6H2Ph3- 2,4,6)CuC6H2Ph3-2,4,6].10 These Cu]Cu distances are indicative of a weak d10–d10 interaction between the metals.12 In summary, compound 1 represents the highest ratio (2 : 1) copper halide–organocopper complex that has been isolated to date.The structure of 1 provides further evidence that such complexes can be isolated in the absence of chelating ligands as well as underlining the importance of SMe2 as ligand or solvent in organocopper chemistry.13 Acknowledgements We are grateful to the National Science Foundation for financial support.References 1 G. van Koten, S. L. James and J. T. B. H. Jastrzebski, in Comprehensive Organometallic Chemistry II, eds. E. W. Abel, F. G. A. Stone and G. Wilkinson, Pergamon, Oxford, 1995, vol. 3, ch. 2. 2 For example: M. D. Janssen, M. A. Corsten, A. L. Spek, D. M. Grove and G. van Koten, Organometallics, 1996, 15, 2810 and refs.therein. 3 For example: E. Wehman, G. van Koten, C. J. M. Erkamp, D. M. Knotter, J. T. B. H. Jastrzebski and C. H. Stam, Organometallics, 1989, 8, 94 and refs. therein. 4 G. M. Kapteijn, I. C. M. Wehman-Ooyevaar, D. M. Grove, W. J. J. Smeets, A. L. Spek and G. van Koten, Angew. Chem., Int. Ed. Engl., 1993, 32, 72. 5 For example: E. Wehman, G. van Koten, J. T. B. H. Jastrzebski, M. A. Rotteveel and C. H. Stam, Organometallics, 1988, 7, 1477. 6 G. van Koten, J. Organomet. Chem., 1990, 400, 283. 7 M. Yoshifuji, I. Shima and N. Inamoto, Tetrahedron Lett., 1979, 41, 3963. 8 S. H. Bertz, G. Dabbagh, X. He and P. P. Power, J. Am. Chem. Soc., 1993, 115, 11 640; S. H. Bertz and G. Dabbagh, J. Am. Chem. Soc., 1988, 110, 3668. 9 H. Hope, M. M. Olmstead, P. P. Power, J. Sandell and X. Xu, J. Am. Chem. Soc., 1985, 107, 4337. 10 X. He, M. M. Olmstead and P. P. Power, J. Am. Chem. Soc., 1992, 114, 9668. 11 M. M. Olmstead and P. P. Power, J. Am. Chem. Soc., 1990, 112, 8008. 12 P. K. Mehrotra and R. HoVman, Inorg. Chem., 1978, 17, 2187; A. Dedieu and R. HoVman, J. Am. Chem. Soc., 1978, 100, 2074; K. M. Merz and R. HoVman, Inorg. Chem., 1988, 27, 2120. 13 S. H. Bertz and G. Dabbagh, Tetrahedron, 1989, 45, 425. Received 2nd July 1998; Communication 8/05108I

ISSN:1477-9226    

DOI:10.1039/a805108i

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, Pages 2601–2602 2601 EVects of non-molecular forces on molecular structure in tris(thiourea)- copper(I) Karin Johnson and Jonathan W. Steed * Department of Chemistry, King’s College London, Strand, London, UK WC2R 2LS Intermolecular-hydrogen bonding interactions have been shown to result in a shortening of the copper–copper distance by 0.274 Å to 2.5526(5) Å in [Cu2{m-SC(NH2)2}2- {SC(NH2)2}4]21 in comparison to three structures of the same cation in the absence of significant intermolecular interactions.We have recently begun a programme of research into the synergistic interaction between molecular and crystal structure, especially in systems exhibiting a symmetry mismatch.1,2 A number of complex, solid-state arrays have been generated as a result of non-complementary hydrogen bonding between metal aqua ions and crown ethers. In particular, the donor–acceptor system [UO2Cl2(H2O)3]?15-crown-5 exhibits a structure involving a total of sixteen unique metal complex–crown ether pairs.1 This kind of study is particularly rewarding in the case of labile systems because the overall crystal composition and indeed molecular stoichiometry is frequently governed by the imperative to maximise hydrogen-bonding interactions.Furthermore, multiple, strong hydrogen-bonding interactions may have a considerable eVect upon molecular geometry. For example, interaction of hydrated NiX2 (X = NO3, ClO4 or Br) with 18-crown-6 results in three very diVerent species; [Ni(H2O)5(NO3)]2[NO3]2?2(18-crown-6),3 [Ni(H2O)6]2[ClO4]4? 2(18-crown-6) and [Ni(H2O)6]3[NiBr2(H2O)4]Br6?4(18-crown- 6)?2H2O, the compositions being dependent entirely on the formation of strong hydrogen-bonding interactions within the crystalline solids.2 Significant distortions away from the ideal octahedral geometry are noted in each case.We now report the extension of these investigations to hydrogen-bond donors other than aqua complex ions.Along with [CuCl2]2, tris(thiourea)copper(I) chloride is one of the few water-soluble forms of CuI which do not disproportionate to CuII and Cu metal. Although labile in solution, [Cu{SC(NH2)2}3]Cl 1 exists as an infinite co-ordination polymer in the solid state 4 incorporating a single S-bound thiourea bridge from one metal centre to the next. Inter-chain interactions consist of NH ? ? ? Cl and NH ? ? ? S hydrogen bonds with N ? ? ? Cl/S distances in the range 3.25–3.40 Å.A survey of the Cambridge Structural Database 5 however, reveals that the [Cu{SC(NH2)2}n]1 moiety can adopt a variety of geometries depending on the identity of the counter anion. Thus the hydrated sulfate, perchlorate and tetrafluoroborate salts exist as [Cu2{m-SC(NH2)2}2{SC(NH2)2}4]21 2 metal–metal bonded dimers.6,7 Each dimer interacts with the anion via one or, at most, two hydrogen bonds NH ? ? ? F/O 2.92–3.03 Å. The hydrated [SiF6]2 salt exhibits a polymeric structure incorporating six-membered Cu3S3 rings.8 In the anhydrous form, [SiF6]22 gives rise to discrete, mononuclear [Cu{SC(NH2)2}4]1 3 tetrahedra,9 while the hydrogen phthalate anion results in units with NH ? ? ? O hydrogen bonds of 2.909 Å.10 Other related oligomeric and polymeric structures are also known.5 Given the preponderance of hydrogen-bond donors in the form of * E-Mail: jon.steed@kcl.ac.uk thiourea NH2 moieties in all these systems we decided to investigate the solid-state behaviour of 1 in the presence of crown ethers as hydrogen-bond acceptors.Accordingly, an equimolar mixture of [Cu{SC(NH2)2}3]Cl and 18-crown-6 was prepared in water.† Slow evaporation of this solution aVorded large, colourless blocks of a complex of formula [Cu2{m-SC- (NH2)2}2{SC(NH2)2}4]Cl2?2[SC(NH2)2]?2H2O?2(18-crown-6) 4 which was subjected to analysis by X-ray crystallography.‡ The dicopper(I) cation in 4 (Fig. 1) ostensibly resembles the binuclear species of type 2 isolated in the presence of tetrahedral anions such as ClO4 2, although contrasts sharply to the parent material 1.Interestingly, however, the Cu]Cu distance in complexes of type 2 is in the narrow range 2.827–2.862 Å, with Cu]S]Cu angles of ca. 71–748 suggesting a relatively weak Cu]Cu interaction. In the case of 4, in essentially the same cation, the Cu]Cu distance is compressed to a remarkable 2.5526(5) Å, with a resultant bond angle at the bridging sulfur of 63.646(14)8, clearly suggesting a much more significant metal–metal interaction.This eVect is also apparent in the Cu]S bond lengths which are rather longer to the bridging sulfur atoms in 4 compared to 2, and shorter to the terminal sulfurs. This marked diVerence in molecular structure is apparently due to the significantly greater number of intermolecular interactions in which the [Cu2{m-SC(NH2)2}2{SC(NH2)2}4]21 cation takes part in 4. Thus the dicopper cation is surrounded by a total of six crown ethers (which exist in two diVerent con- Fig. 1 The [Cu2{m-SC(NH2)2}2{SC(NH2)2}4]21 cation in 4.Selected bond lengths (Å) and angles (8): Cu(1)]S(3) 2.2752(5), Cu(1)]S(1) 2.2985(5), Cu(1)]S(2) 2.3988(5), Cu(1)]S(29) 2.4417(5), Cu(1)]Cu(19) 2.5526(5), Cu(1)]S(2)]Cu(19) 63.646(14) † The salt [Cu{SC(NH2)2}3]Cl (0.20 g, 0.61 mmol) was dissolved in distilled water (5 cm3) and added to a solution of 18-crown-6 (0.16 g, 0.61 mmol) in water (1 cm3) to give a colourless solution.The resulting mixture was allowed to stand in air for ca. 2 weeks resulting in the gradual deposition of the product as large, colourless blocks in near quantitative yield. ‡ Crystal data for 5: C32H84Cl2Cu2N16O14S8, M 1371.61, triclinic, space group P1� (no. 2), a = 10.1764(4), b = 11.1352(4), c = 13.5561(4) Å, a = 89.647(2), b = 93.263(2), g = 99.868(2)8, U = 1510.94(9) Å3, Z = 1, m = 11.36 cm21, T = 100 K. Reflections measured: 13 159, unique data: 5542 (Rint = 0.032), parameters: 407, R1 [F2 > 2s(F2)] 0.0303, wR2 (all data) 0.0783.CCDC reference number 186/1069. See http:// www.rsc.org/suppdata/dt/1998/2601/ for crystallographic files in .cif format.2602 J. Chem. Soc., Dalton Trans., 1998, Pages 2601–2602 formations) each of which form hydrogen bonds to four nitrogen atoms of the thiourea ligands, NH ? ? ? O 2.991(2)–3.107(2) Å, Fig. 2. In addition, the remaining two N-atoms hydrogen bond to the solvent water, N(2) ? ? ? O(1s) 2.986(2) Å, and the sulfur atom of the unco-ordinated thiourea molecule, N(3) ? ? ? S(4) 3.305(2) Å.This extensive solid-state network is completed by interactions from the water molecule to the chloride anion 3.258(2) and 3.301(2) Å and from the NH2 moieties of the unco-ordinated thiourea molecule to one of the crown ether ligands, N(7) ? ? ? O(1A), O(2A) 2.856(2) and 3.013(2) Å, respectively. The unco-ordinated thiourea also hydrogen bonds with the chloride ligands, N(8) ? ? ? Cl(1), Cl(10) 3.278(2) and 3.409(2) Å.Finally, it is noteworthy that the C]] S distance is very slightly shorter in the unco-ordinated thiourea than in the co-ordinated ligands; 1.718(2) vs. 1.724(2) Å (average). The CuI centre in 4 would be expected to exhibit a low Fig. 2 The crystal environment of the [Cu2{m-SC(NH2)2}2{SC- (NH2)2}4]21 cation in 4. Selected intermolecular contacts (Å): N(1) ? ? ? O(3a) 2.991(2), N(2) ? ? ? S(4) 3.350(2), N(3) ? ? ? S(4) 3.305(2), N(4) ? ? ? O(2a9) 3.017(2), N(5) ? ? ? O(2b) 2.937(2), N(6) ? ? ? O(1b) 2.915(2), N(7) ? ? ? O(1a) 2.856(2), N(7) ? ? ? O(2a) 3.013(2), N(8) ? ? ? Cl(1) 3.278(2), N(8) ? ? ? Cl(10) 3.409(2) ligand-field splitting energy and hence there is no reason why it should necessarily obey the 18-electron rule.As a result the electronic environment about the metal centres is highly malleable, allowing the complex to undergo very large distortions in order to accommodate and maximise stabilising hydrogenbonding interactions in the solid state.The sharp contrast between the co-ordination geometry of the [Cu2{m-SC(NH2)2}2- {SC(NH2)2}4]21 cation in 4 and the three closely related species 2 in which intermolecular hydrogen bonding is much less pralent highlights the importance of a consideration of the overall crystal, as well as molecular structure, on molecular properties in the solid state. Acknowledgements We thank the EPSRC and King’s College London for funding of the diVractometer system.Grateful acknowledgement is also given to the EPSRC Chemical Database Service at Daresbury and to the NuYeld Foundation for the provision of computing equipment. References 1 J. W. Steed, H. Hassaballa and P. C. Junk, Chem. Commun., 1998, 577. 2 J. W. Steed, B. J. McCool and P. C. Junk, J. Chem. Soc., Dalton Trans., submitted. 3 P. C. Junk, S. M. Lynch and B. J. McCool, Supramol. Chem., 1998, in the press. 4 Y. Okaya and C. Knobler, Acta Crystallogr., 1964, 17, 928. 5 April 1998 update, 181 309 entries: F. H. Allen and O. Kennard, Chem. Des. Autom. News, 1993, 8, 31. 6 M. van Meerssche, R. Kamara, J. P. Declerq and G. Germain, Bull. Soc. Chim. Belg., 1982, 91, 547. 7 F. Hanic and E. Durcanska, Inorg. Chim. Acta, 1969, 3, 293. 8 A. G. Gash, E. H. GriYth, W. A. SpoVord III and E. L. Amma, J. Chem. Soc., Chem. Commun., 1973, 256. 9 G. W. Hunt, N. W. Terry III and E. L. Amma, Acta Crystallogr., Sect. B, 1979, 35, 1235. 10 M. B. Cingi, A. M. M. Lanfredi, A. Tiripicchio and M. Tiripicchio Camellini, Acta Crystallogr., Sect. B, 1977, 33, 3772. Received 18th May 1998; Communication 8/03708F

ISSN:1477-9226    

DOI:10.1039/a803708f

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, Pages 2603–2605 2603 A sandwich complex of lithium oxide: {Li[BunC(NBut)2]}4?Li2O Tristram Chivers,*,†,a Andrew Downard a and Glenn P. A. Yap b a Department of Chemistry, University of Calgary, Calgary, Alberta, Canada T2N 1N4 b Department of Chemistry, University of Ottawa, Ottawa, Ontario, Canada K1N 6N5 Hydrolysis of {Li[BunC(NBut)2]}2, prepared from LiBun and ButNCNBut in hexanes, produced the nineteen atom cluster {Li[BunC(NBut)2]}4?Li2O; an X-ray structure determination revealed an Li2O molecule encapsulated by two eight-membered Li2N4C2 rings.Novel structural and/or reaction chemistry often results from ligands that provide unusual steric and/or electronic environments at metal centres. To this end N-silylated benzamidinates RC(NSiMe3)2 2 (R = aryl) 1,2 and, more recently, dialkylamidinates RC(NR9)2 2 (where R and R9 are diVerent alkyl groups) have been investigated extensively.1b,3–5 Although a wide range of both main-group and transition-metal complexes of these chelating ligands has been characterized,6 structural information for the lithium derivatives of these important reagents is limited to complexes in which the lithium ions are co-ordinated to Lewis bases such as RCN (R = aryl),7 THF,8 HMPA,9 N,N,N9,N9-tetramethylethylenediamine 9 or N,N,N9,N9,N0- pentamethyldiethylenetriamine.9 The complexes [4-MeC6- H4C(NSiMe3)2Li(THF)]2 8 and [PhC(NPh)2Li(HMPA)]2 9 form dimeric, step-shaped structures whereas chelating Lewis bases give rise to monomeric structures.9 We describe here the unexpected generation and crystal structure of the complex {Li[BunC(NBut)2]}4?Li2O 2 in which a molecule of lithium oxide is trapped between two twisted Li2N4C2 ladders of a dimeric lithium amidinate.Compared to other complexes of Li2O,10–14 complex 2 exhibits some novel features that result from the unique ligand behaviour of the Li2N4C2 ring. Amidinates Li[RC(NR9)2] are readily obtained by the nucleophilic addition of an organolithium reagent (RLi) to a carbodiimide R9NCNR9.3,5 In this work, the addition of a 2.5 M solution of LiBun in hexanes (3.7 mL) to a solution of 1,3-ditert- butylcarbodiimide (9.22 mmol) in hexane (10 mL) under argon at 23 8C produced a transparent, pale yellow solution.Removal of volatile materials in vacuo gave a viscous yellow oil, which was redissolved twice in diethyl ether (ª5 mL). Evaporation of the solvent in vacuo produced {Li[BunC(NBut)2]}n 1 as a fine yellow powder (8.63 mmol, 94%).‡ Recrystallization of 1 from a saturated toluene solution (4 d at 220 8C) produced a few X-ray quality crystals with NMR parameters significantly diVerent from those of 1.An X-ray structural determination revealed that the composition of these crystals is {Li[BunC(NBut)2]}4?Li2O 2 (Fig. 1).§ This nineteen atom cluster has a m6-OLi6 core. Six-fold coordination of O22 by metal cations in molecular compounds is rare and usually involves regular Oh symmetry.12–14 A major diVerence between the structure of 2 and those of other Li2O aggregates 11–14 is that the molecule of Li2O is readily identified in 2 because of the relatively low symmetry of this cluster.† E-Mail: chivers@acs.ucalgary.ca Thus 2 may be viewed as an almost linear Li2O molecule [Li(1)]O]Li(2) 175.8(2)8] sandwiched between two twisted Li2N4C2 ladders. The oxygen atom is tightly co-ordinated to all six lithium atoms, but the mean Li]O distance in the Li2O molecule [1.803(4) Å] is significantly shorter than that in the other Li]O bonds [mean value 1.869(4) Å], cf. 1.89(1) Å in [(cyclo-C5H9)N(H)Li]12?Li2O,13 1.81–1.90(2) Å in [Pri 2(Mes)- SiP]8Li16?Li2O14 (Mes = C6H2Me3-2,4,6). The distortion of the octahedral geometry of the OLi6 unit in 2 is reflected in the Li ? ? ? Li separations which range from 2.169(5) Å across the Li2N4C2 rings to 3.127(5) Å {cf. 2.63–2.67(2) Å in [Pri 2(Mes)- SiP]8Li16?Li2O14}.All six lithium atoms can be viewed as fourco- ordinate, but there is considerable variation in the Li]N bond distances. Those belonging to the Li2O moiety are bonded symmetrically to two nitrogen atoms of different Li2N4C2 rings [|d(Li]N)| 2.106(4) Å] and are also involved in a third, weaker Li ? ? ? N interaction [2.573(4) Å]. This results in a ‘pinching in’ of the Li2N4C2 rings as reflected from the values of |Li]O]Li| 70.9(2)8 and |N]C]N| = 115.3(9)8.The other four lithium atoms are bonded unsymmetrically to two nitrogen atoms of the same Li2N4C2 ring [|d(Li]N)| 2.03(2) and 2.36(2) Å]. As a result there are three four-co-ordinate and one five-co-ordinate nitrogen atom in each Li2N4C2 ring. The mean C]N bond distances are slightly longer for the four-coordinate compared to the five-co-ordinate N atoms [1.346(2) vs. 1.329(2) Å]. ‡ 1 Mp 51–54 8C. 1H NMR (25 8C, 200 MHz, C6D6): d 0.90 (t, 3 H, CH3CH2CH2CH2), 1.32 [s 1 m, 20 H, CH3CH2CH2CH2 and C(CH3)3], 1.85 (m, 2 H, CH3CH2CH2CH2), 2.50 (m, 2 H, CH3CH2CH2CH2). 13C NMR (25 8C, 50.288 MHz, C6D6): d 14.1 (s, CH2CH2CH2CH3), 24.1 (s, CH2CH2CH2CH3), 33.0 (s, CH2CH2CH2CH3), 33.2 (s, CH2CH2- CH2CH3), 33.7 [s, C(CH3)3], 51.6 [s, C(CH3)3], 178.4 [s, C(NBut)2Bun]; (25 8C, 75.432 MHz, solid state): d 14.2 (s, CH2CH2CH2CH3), 24.1 (s, CH2CH2CH2CH3), 34.1 [s br, CH2CH2CH2CH3, CH2CH2CH2CH3 and C(CH3)3], 51.3 [s, C(CH3)3], 175.3 [s, C(NBut)2Bun]. 7Li NMR (25 8C, 155.508 MHz, C6D6, relative to 1 M LiCl in D2O): d 20.62 (s); (25 8C, 116.54 MHz, solid state, relative to LiCl): d 1.46 (s). 2 Mp 132–134 8C (Found: C, 69.40; H, 12.78; N, 12.51. Calc. for C52H108Li6N8O: C, 69.16; H, 12.05; N, 12.41%). 1H NMR (25 8C, 200 MHz, C6D6): d 0.99 (t, 3 H, CH3CH2CH2CH2), 1.46 [s 1 m, 20 H, CH3CH2CH2CH2 and C(CH3)3], 1.95 (m, 2 H, CH3CH2CH2CH2), 2.60 (m, 2 H, CH3CH2CH2CH2). 13C NMR (25 8C, 50.288 MHz, C6D6): 15.5 (s, CH2CH2CH2CH3), 23.5 (s, CH2CH2CH2CH3), 32.9 (CH2CH2- CH2CH3), 34.5 (s, CH2CH2CH2CH3), 34.6 [s, C(CH3)3], 51.9 [s, C(CH3)3], 179.6 [s, C(NBut)2Bun]; (25 8C, 75.432 MHz, solid state): d 14.4 (s, CH2CH2CH2CH3), 24.5 (s, CH2CH2CH2CH3), 35.5 [s br, CH2CH2CH2CH3, CH2CH2CH2CH3 and C(CH3)3], 52.7 [s, C(CH3)3], 180.6 [s, C(NBut)2Bun]. 7Li NMR (25 8C, 155.508 MHz, C6D6, relative to 1 M LiCl in D2O): d 20.82 (s), 21.23 (s); (25 8C, 116.54 MHz, solid state, relative to LiCl): d 2.97 (s). § Crystal data: C52H108Li6N8O, M = 903.10, triclinic, space group P1� , a = 10.137(3), b = 14.205(4), c = 21.961(6) Å, a = 91.7718(5), b = 103.207(5), g = 101.442(5)8, U = 2008(1) Å3, Z = 2, m = 0.58 cm21, T = 213 K, 24 534 reflections collected, 13 710 independent reflections, Rint = 0.0701.The final R(F) and wR(F 2) values were 0.0545 and 0.0821, respectively. CCDC reference number 186/1071. See http:// www.rsc.org/suppdata/dt/1998/2603/ for crystallographic files in .cif format.2604 J. Chem. Soc., Dalton Trans., 1998, Pages 2603–2605 The most obvious explanation for the formation of 2 is the partial hydrolysis of 1 by trace amounts of water present in the solvent or flask used for recrystallization [reactions (1) and (2)].Li[BunC(NBut)2] 1 H2O æÆ LiOH 1 BunC(NBut)[N(H)But] (1) 1 3 5 1 1 LiOH æÆ {Li[C(NBut)2(Bun)]}4?Li2O 1 3 (2) 2 The adventitious presence of water has previously been identi- fied as the source of Li2O in aggregates with lithium amides.12,13 To test this hypothesis a stoichiometric amount of water was added, by syringe, to a 0.46 M solution of 1 in toluene (5 mL) at 23 8C.This produced an oily white solid, which was stirred for 1 h to give an opaque yellow solution. The volume of the solution was reduced by one-half and colourless crystals of 2 were obtained in 30% yield after 3 d at 214 8C. The analytical and spectroscopic characterization of 2 were completed on this product.‡ The observation of two resonances in the 7Li NMR spectrum (in C6D6) at d 20.82 and 21.23 (the latter is of lower relative intensity) suggests a higher average symmetry (D2) for 2 in solution compared to that observed (C2) in the solid state.The 7Li NMR specum of 1 in C6D6 exhibits a singlet at d 20.62. There are significant diVerences in the 13C NMR chemical shifts observed for 1 and 2.‡ In particular, d [C(NBut)2- (Bun)] provides a diagnostic distinction between 1 and 2 both in solution and, especially, in the solid state.The 1H NMR spectrum of the mother-liquor from reaction (1) showed it to consist of a mixture of unreacted 1 and the hydrolysis product 3. Thus hydrolysis of 1 is clearly established as a route to 2. Further support for this conclusion is provided by the observation that the direct reaction of 1 with LiOH in toluene at 23 8C for 48 h produces 2 in 41% yield, but 2 is not formed from the treatment of 1 with Li2O under similar conditions. A conceptual representation of the assembly of the nineteen atom cluster 2 from two Li2N4C2 dimers and a Li2O molecule is Fig. 1 Molecular structure and atomic numbering scheme for complex 2. Thermal ellipsoids are depicted at 30% probability. For clarity only the a-carbon atoms of Bun and But are shown. Selected bond distances (Å) and angles (8): O]Li(1) 1.805(4), O]Li(2) 1.801(4), O]Li(3), 1.880(4), O]Li(4) 1.883(4), O]Li(5) 1.852(4), O]Li(6) 1.862(4), Li(1)]N(1) 2.082(4), Li(1)]N(2) 2.664(4), Li(1)]N(8) 2.092(4), Li(2)]N(4) 2.102(4), Li(2)]N(5) 2.148(4), Li(2)]N(6) 2.481(4), Li(3)]N(1) 2.047(4), Li(3)]N(2) 2.505(4), Li(3)]N(3) 2.293(4), Li(4)]N(2) 2.316(4), Li(4)]N(3) 2.458(4), Li(4)]N(4) 2.050(4), Li(5)]N(6) 2.390(4), Li(5)]N(7) 2.443(4), Li(5)]N(8) 2.062(4), Li(6)]N(5) 2.048(4), Li(6)]N(6) 2.514(5), Li(6)]N(7) 2.376(4); |N]C]N| 115.1 [range 114.6(2)–115.9(2)] shown in Scheme 1, where the source of Li2O is LiOH produced by the hydrolysis of 1.An alternative source of LiOH and, hence, Li2O in the original formation of 2 is the commercial LiBun used for the preparation of 1.15 Indeed the 7Li NMR spectrum of fresh LiBun (2.5 M in hexanes, Aldrich) in C6D6 exhibited a small resonance at d 20.89 in addition to the dominant resonance at d 20.22 (vs. 1 M LiCl in D2O). The intensity of the former relative to that at d 20.22 increased upon addition of water to the solution, but not upon addition of solid LiOH. Although the identity of the d 20.89 species has not been established, we cannot rule out commercial LiBun as a source of Li2O in the formation of 2.Finally, we note that the co-ordination of Li2O does not aVect the use of 2 as a source of the chelating amidinate ligand BunC(NBut)2 2. For example, reaction of 2 (5.82 mmol) with PhBCl2 (5.29 mmol) in toluene (15 mL) produces PhB(Cl)[C(NBut)2Bun] 4 in 82% Yield.¶ The four-membered ring structure of 4 has been confirmed by X-ray crystallography and full details of this structure and those of related fourco- ordinate boron complexes will be reported in a separate publication.16 In summary, complex 2 provides the first demonstration of the ligand behaviour of a dimeric lithium amidinate.The entrapment of other alkali-metal chalcogenides, e.g. Li2S, Na2O, by lithium amidinates is an interesting possibility that will be pursued. Acknowledgements We thank NSERC (Canada) for financial support. References 1 (a) F. T. Edelmann, Coord. Chem. Rev., 1994, 137, 403; (b) J.Barker and M. Kilner, Coord. Chem. Rev., 1994, 133, 219. 2 R. Duchateau, A. Meetsma and J. H. Teuben, Chem. Commun., 1996, 223; R. Duchateau, C. T. Van Wee, A. Meetsma and J. H. Teuben, Organometallics, 1996, 15, 2291. 3 Y. Zhou and D. S. Richeson, Inorg. Chem., 1996, 35, 2448; Y. Zhou and D. S. Richeson, J. Am. Chem. Soc., 1996, 118, 10 850. 4 M. P. Coles and R. F. Jordan, J. Am. Chem. Soc., 1997, 119, 8125; M. P. Coles, D. C. Swenson and R. F. Jordan, Organometallics, 1997, 16, 5183.Scheme 1 Schematic representation of the formation of 2. (i) Dimerisation; (ii) partial hydrolysis ¶ Mp 76–79 8C (Found: C, 67.76; H, 9.97; N, 8.45. Calc. for C19H32BClN2: C, 68.16; H, 9.65; N, 8.37%). 1H NMR (25 8C, 200 MHz, C6D6): d 0.73 (t, 3 H), 1.11 (m, 2 H), 1.17 (s, 18 H), 1.60 (m, 2 H), 2.14 (m, 2 H) 7.2–8.1 (m, 5 H). 11B NMR (25 8C, 64.2 MHz, relative to BF3?OEt2): d 6.8 (s). EI-MS [70 eV (eV ª 1.602 × 10219 J)]: m/z 334 (M1, good agreement between calculated and observed isotopic distribution).J. Chem.Soc., Dalton Trans., 1998, Pages 2603–2605 2605 5 P. Berno, S. Hao, R. Minhas and S. Gambarotta, J. Am. Chem. Soc., 1994, 116, 7417; S. Hao, S. Gambarotta, C. Bensimon and J. J. H. Edema, Inorg. Chim. Acta, 1993, 213, 65. 6 H. H. Karsch, P. A. Schlüter and M. Reisky, Eur. J. Inorg. Chem., 1998, 433; J. Barker, N. C. Blacker, P. R. Phillips, N. W. Alcock, W. Errington and M. G. H. Wallbridge, J. Chem. Soc., Dalton Trans., 1996, 431. 7 T. Gebauer, K. Dehnicke, H. Goesmann and D. Fenske, Z. Naturforsch., Teil B, 1994, 49, 1444; M. S. Eisen and M. Kapon, J. Chem. Soc., Dalton Trans., 1994, 3507. 8 D. Stalke, M. Wedler and F. T. Edelmann, J. Organomet. Chem., 1992, 431, C1. 9 J. Barker, D. Barr, N. D. R. Barnett, W. Clegg, I. Cragg-Hine, M. G. Davidson, R. P. Davies, S. M. Hodgson, J. A. K. Howard, M. Kilner, C. W. Lehmann, I. Lopez-Solera, R. E. Mulvey, P. R. Raithby and R. Snaith, J. Chem. Soc., Dalton Trans., 1997, 951. 10 H. Dietrich and D. Rewicki, J. Organomet. Chem., 1981, 205, 281. 11 H.-J. Gais, J. Vollhardt, H. Günther, D. Moskau, H. J. Lindner and S. Braun, J. Am. Chem. Soc., 1988, 110, 978. 12 S. C. Ball, I. Cragg-Hine, M. G. Davidson, R. P. Davies, M. I. Lopez-Solera, P. R. Raithby, D. Reed, R. Snaith and E. M. Vogl, J. Chem. Soc., Chem. Commun., 1995, 2147. 13 W. Clegg, L. Horsburgh, P. R. Dennison, F. M. Mackenzie and R. E. Mulvey, Chem. Commun., 1996, 1065. 14 M. Driess, H.Pritzkow, S. Martin, S. Rell, D. Fenske and G. Baum, Angew. Chem., Int. Ed. Engl., 1996, 35, 986. 15 C. Lambert, F. Hampel, P. Rague von Schleyer, M. G. Davidson and R. Snaith, J. Organomet. Chem., 1995, 487, 139. 16 P. Blais, T. Chivers, A. Downard and M. Parvez, unpublished work. Received 29th June 1998; Communication 8/04955FJ. Chem. Soc., Dalton Trans., 1998, Pages 2603–2605 2605 5 P. Berno, S. Hao, R. Minhas and S. Gambarotta, J. Am. Chem. Soc., 1994, 116, 7417; S.Hao, S. Gambarotta, C. Bensimon and J. J. H. Edema, Inorg. Chim. Acta, 1993, 213, 65. 6 H. H. Karsch, P. A. Schlüter and M. Reisky, Eur. J. Inorg. Chem., 1998, 433; J. Barker, N. C. Blacker, P. R. Phillips, N. W. Alcock, W. Errington and M. G. H. Wallbridge, J. Chem. Soc., Dalton Trans., 1996, 431. 7 T. Gebauer, K. Dehnicke, H. Goesmann and D. Fenske, Z. Naturforsch., Teil B, 1994, 49, 1444; M. S. Eisen and M. Kapon, J. Chem. Soc., Dalton Trans., 1994, 3507. 8 D. Stalke, M. Wedler and F. T. Edelmann, J. Organomet. Chem., 1992, 431, C1. 9 J. Barker, D. Barr, N. D. R. Barnett, W. Clegg, I. Cragg-Hine, M. G. Davidson, R. P. Davies, S. M. Hodgson, J. A. K. Howard, M. Kilner, C. W. Lehmann, I. Lopez-Solera, R. E. Mulvey, P. R. Raithby and R. Snaith, J. Chem. Soc., Dalton Trans., 1997, 951. 10 H. Dietrich and D. Rewicki, J. Organomet. Chem., 1981, 205, 281. 11 H.-J. Gais, J. Vollhardt, H. Günther, D. Moskau, H. J. Lindner and S. Braun, J. Am. Chem. Soc., 1988, 110, 978. 12 S. C. Ball, I. Cragg-Hine, M. G. Davidson, R. P. Davies, M. I. Lopez-Solera, P. R. Raithby, D. Reed, R. Snaith and E. M. Vogl, J. Chem. Soc., Chem. Commun., 1995, 2147. 13 W. Clegg, L. Horsburgh, P. R. Dennison, F. M. Mackenzie and R. E. Mulvey, Chem. Commun., 1996, 1065. 14 M. Driess, H. Pritzkow, S. Martin, S. Rell, D. Fenske and G. Baum, Angew. Chem., Int. Ed. Engl., 1996, 35, 986. 15 C. Lambert, F. Hampel, P. Rague von Schleyer, M. G. Davidson and R. Snaith, J. Organomet. Chem., 1995, 487, 139. 16 P. Blais, T. Chivers, A. Downard and M. Parvez, unpublished work. Received 29th June 1998; Communication 8/04955F

ISSN:1477-9226    

DOI:10.1039/a804955f

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, Pages 2607–2609 2607 Desulfurisation of trithiocarbonates at a dimolybdenum centre: an unexpected insertion into a co-ordinated alkyne Harry Adams, Christopher Allott, Matthew N. Bancroft and Michael J. Morris*,† Department of Chemistry, University of Sheffield, Sheffield, UK S3 7HF On reaction with the dimolybdenum alkyne complex [Mo2- {m-C2(CO2Me)2}(CO)4(h-C5H5)2], dialkyl trithiocarbonates (RS)2C]] S were dismantled into sulfido (m-S), thiolate (m-SR) and CSR fragments; remarkably the last of these inserts into the middle of the alkyne to produce a dimetalla-allyl species with a m-C(CO2Me)C(SR)C(CO2Me) ligand.The desulfurisation of organic molecules is an important process which finds widespread application in organic synthesis and materials chemistry, in particular the formation of tetrathiafulvalenes by phosphite-induced coupling of two sulfur heterocycles. 1 Recently we described the unusual reaction of the dimolybdenum alkyne complex [Mo2{m-C2(CO2Me)2}(CO)4- (h-C5H5)2] 1 with cyclic trithiocarbonates (1,3-dithiole-2- thiones) S]] CS2C2R2 (R = CO2Me, SMe or SCOPh) to produce complexes 2, in which the C]] S bond had been cleaved to produce a m-sulfido ligand and a complex organic fragment derived by ring opening of the heterocycle and coupling with the alkyne ligand (Scheme 1).2 Subsequently we showed that further reaction of 2 with sulfur produced dithiolene complexes in which the C2R2 backbone of the original heterocycle was incorporated into the dithiolene ligand.3 Here we describe the reaction of 1 with acyclic dialkyl trithiocarbonates which sheds further light on the mechanism of the processes involved.Treatment of 1 with 1 equivalent of the trithiocarbonates (RS)2C]] S (R = Me, Pri or Bu) in refluxing toluene for 48 h followed by column chromatography aVorded green complexes 3 as the only isolable products (Scheme 1), though yields in general were lower than those of 2.Spectroscopic characterisation‡ indicated that 3 had a structure very similar to 2, though with certain diVerences. Crystals of 3a suitable for X-ray diVraction were obtained from dichloromethane and diethyl ether solution. § The structure is shown in Fig. 1, with selected bond lengths given in the caption. There are two independent molecules in the unit cell, but the only significant diVerence between † E-Mail: M.Morris@sheYeld.ac.uk ‡ Spectroscopic data (NMR in CDCl3, all signals are singlets unless otherwise stated). Satisfactory elemental analyses were obtained for all new compounds. 3a: green solid, 51% yield, m.p. 184–200 8C (decomp.). 1H NMR: d 5.94 (10 H, C5H5), 3.67, 3.53 (both 3 H, CO2Me), 2.18, 1.97 (both 3 H, SMe). 13C NMR (260 8C): d 176.3, 175.5 (both CO2Me), 141.2 (CSMe), 102.7 (CCO2Me), 99.6 (C5H5), 98.0 (CCO2Me), 52.8, 52.1 (both CO2Me), 46.3 (m-SMe), 13.7 (SMe). MS: m/z 603 (M1). 3b: green solid, 24% yield, m.p. 132–133 8C. 1H NMR: d 5.92 (10 H, C5H5), 3.66 (3 H, CO2Me), 3.52 (spt, 1 H, J = 6.6, CH), 3.49 (3 H, CO2Me), 1.19 (d, 6 H, J = 6.6, Me), 1.09 (d, 6 H, J = 6.7, Me), 0.49 (spt, 1 H, J = 6.7 Hz, CH). 13C NMR: d 175.3, 174.8 (both CO2Me), 140.9 (CSPri), 103.1 (CCO2Me), 99.2 (C5H5), 98.1 (CCO2Me), 69.3 (CH), 52.1, 51.6 (both CO2Me), 34.5 (CH), 28.4, 23.8 (both Me). MS: m/z 659 (M1). 3c: green oil, 26% yield. 1H NMR: d 5.92 (10 H, C5H5), 3.66, 3.52 (both 3 H, CO2Me), 2.70 (t, 2 H, J = 7.3, CH2), 1.81 (t, 2 H, J = 7.6, CH2), 1.55–1.08 (m, 8 H, CH2), 0.88 (t, 3 H, J = 7.2, Me), 0.78 (t, 3 H, J = 7.2 Hz, Me). 13C NMR: d 175.3, 174.9 (both CO2Me), 140.8 (CSBu), 102.9 (CCO2Me), 99.2 (C5H5), 98.7 (CCO2Me), 63.5 (CH2), 52.0, 51.5 (both CO2Me), 37.7, 31.2, 29.7, 21.9, 21.8 (all CH2), 13.7, 13.6 (both Me). MS: m/z 688 (M1). them lies in the position of C(13), which in one molecule is almost equidistant from the two metal atoms, but in the other is displaced towards one molybdenum; the former molecule is shown in Fig. 1 and the bond lengths quoted refer to this. The basic structure is indeed very similar to that of 2 and incorporates the now familiar quadruply-bridged MoIV motif. The two molybdenum atoms are joined by a bond of 2.5605(10) Å, even shorter than the 2.5825(7) Å observed in 2a, and are bridged symmetrically by the sulfido ligand S(3) and by the methanethiolate sulfur S(1). The methyl substituent on the bridging thiolate group is pointing away from the dimetallaallyl ligand, whereas in 2 it is constrained to point towards it by the linking chain.The main point of interest however lies in the dimetalla-allyl ligand itself. By analogy with 2 we had expected the carbon atom derived from the trithiocarbonate to be situated at the terminus of this ligand, i.e. m-C(SMe)C(CO2- Me)]] C(CO2Me). Instead, remarkably, it occupies the central position in a m-C(CO2Me)C(SMe)]] C(CO2Me) arrangement. Although in the molecule shown all three carbons of this ligand are equidistant from both metal atoms within experimental error, the bonds from the metals to the central carbon C(13) are much longer than those to the terminal carbons C(11) and C(12).The 13C NMR spectrum of 3a contains three peaks at d 141.2, 102.7 and 98.0 assigned to the carbons of the dimetalla-allyl ligand. This can be compared with 2a–2c, where all three peaks occur with very similar shifts in the region d 108–113.Since the two terminal carbons C(11) and C(12) are Scheme 1 C Mo C O OC Mo CO CO C MeO2C CO2Me S C C S C C C S Mo Mo R R MeO2C CO2Me S S S R R S R MeO2C C C C S Mo Mo RS CO2Me RS RS S 1 toluene, heat 3a: R = Me 3b: R = Pri 3c: R = Bu 2a: R = CO2Me 2b: R = SMe 2c: R = SCOPh toluene, heat § Crystal data for 3a: C19H22Mo2O4S3, M = 602.43, monoclinic, space group P21/n (a non-standard setting of P21/c, C5 2h, no. 14), a = 10.086(3), b = 13.077(3), c = 33.601(7) Å, b = 95.00(2)8, U = 4415(2) Å3, Z = 4, Mo-Ka radiation (l � = 0.710 73 Å), m(Mo-Ka) = 1.442 mm21, T = 293(2) K; 10 070 reflections measured, 7770 independent reflections (Rint = 0.0331), R1 = 0.0406 for 7765 unique data. CCDC reference number 186/1070.See http://www.rsc.org/suppdata/dt/1998/2607 for crystallographic files in .cif format.2608 J. Chem. Soc., Dalton Trans., 1998, Pages 2607–2609 Fig. 1 Molecular structure of one of the two independent molecues of complex 3a in the crystal (50% probability ellipsoids, H atoms omitted for clarity).Selected bond lengths (Å): Mo(1)]Mo(2) 2.5605(10), Mo(1)]S(1) 2.476(2), Mo(2)]S(1) 2.466(2), Mo(1)]S(3) 2.318(2), Mo(2)]S(3) 2.322(2), Mo(1)]C(11) 2.173(5), Mo(2)]C(11) 2.152(5), Mo(1)]C(12) 2.204(6), Mo(2)]C(12) 2.203(6), Mo(1)]C(13) 2.564(5), Mo(2)]C(13) 2.594(6), C(11)]C(13) 1.429(7), C(12)]C(13) 1.393(8) Scheme 2 Mo Mo SR S R S MeO2C MeO2C Mo Mo SR S MeO2C MeO2C Mo C MeO2C CO2 Me S C SR SR Mo C MeO2C CO2 Me C SR S R S SR or 3 alkyne bond breaks C–C(SR) bond breaks 2 C Mo C Mo in very similar environments, we assign the low field 13C NMR signal to the central CSMe feature.In the molecule shown C(13) is within bonding distance of both metals, but in the second molecule there is one short distance, Mo(1A)]C(13A) [2.509(6) Å] and one which is much longer, Mo(2A)]C(13A) [2.645 Å]. As in complex 2, the observation of equivalent h-C5H5 ligands in the NMR spectra of 3, even at low temperature, implies that a fluxional process is occurring in solution in which the central carbon of the dimetalla-allyl ligand is flipping back and forth between the two metal atoms, rendering them equivalent.The observation of two molecules in the unit cell which diVer only in the position of C(13) (and its associated substituent) provides additional evidence that this trajectory is plausible, since the energy diVerence between these two positions is evidently small. In our previous paper we hypothesized that the first step of the reaction mechanism was loss of a CO ligand, co-ordination of the thione group and cleavage of the C]] S bond to give a dithiocarbene.2 This was followed by coupling of the carbene carbon to the alkyne and cleavage of one of the C]S bonds.Obviously this mechanism is inadequate to explain the foation of 3, though it may still be correct for 2. We now propose that, after carbene formation, cleavage of the C]S bond occurs either before or after coupling with the alkyne, leading ultimately to the formation of a three-membered ring (Scheme 2).Cleavage of the C]C bond of the alkyne would then provide the observed product 3. Positioning of the CSR group in the centre of the dimetalla-allyl fragment is not possible in the case of 2 because it is anchored to the thiolate bridge through the spacer group, but this mechanism could still account for the formation of 2 by scission of one of the two C(CO2Me)]C(SR) bonds of the three-membered ring.The apparent insertion of carbyne ligands into the centre of an alkyne is not without precedent. For example, treatment of [WFe2(m3-CC6H4Me)(CO)8(h-C5H5)] with C2Ph2 gave two dimetalla-allyl complexes, one of which was [WFe- {m-CPhC(C6H4Me)CPh}(CO)5(h-C5H5)] with a rearranged chain.4 Moreover the compound [W2(m-CSiMe3)(m-CMe- CMeCSiMe3)(CH2SiMe3)4], reported by Chisholm et al., was found to undergo a fluxional process in which the substituents ofJ.Chem. Soc., Dalton Trans., 1998, Pages 2607–2609 2609 the bridging ligand changed places, i.e. CMeCMeCSiMe3 interconverted with CMeC(SiMe3)CMe, for which a similar threemembered ring intermediate was proposed.5 Other examples involving dimetalla-allyl ligands are known on trinuclear metal centres.6 Further studies on the reactivity of the unusual species 2 and 3 are currently in progress in our laboratory. Acknowledgements We thank the EPSRC for a studentship (to M.N. B.). References 1 M. Narita and C. U. Pittman, jun., Synthesis, 1976, 489; A. Krief, Tetrahedron, 1986, 42, 1209; J. M. Williams, J. R. Ferraro, R. J. Thorn, K. D. Carlson, U. Geiser, H. H. Wang, A. M. Kini and M.-H. Whangbo, Organic Superconductors: Synthesis, Structure, Properties and Theory, Prentice-Hall, Englewood CliVs, NJ, 1992. 2 H. Adams, M. N. Bancroft and M. J. Morris, Chem. Commun., 1997, 1445. 3 A. Abbott, M. N. Bancroft, M. J. Morris, G. Hogarth and S.P. Redmond, Chem. Commun., 1998, 389. 4 F. G. A. Stone, J. C. JeVery, K. A. Mead, H. Razay, M. J. Went and P. Woodward, J. Chem. Soc., Dalton Trans., 1984, 1383. 5 M. H. Chisholm, J. A. Heppert and J. C. HuVman, J. Am. Chem. Soc., 1984, 106, 1151. 6 E. Sappa, A. Tiripicchio and A. M. Manotti Lanfredi, J. Chem. Soc., Dalton Trans., 1978, 552; M. J. Morris, Ph.D. Thesis, University of Bristol, 1984. Received 2nd June 1998; Communication 8/04138EJ. Chem. Soc., Dalton Trans., 1998, Pages 2607–2609 2609 the bridging ligand changed places, i.e.CMeCMeCSiMe3 interconverted with CMeC(SiMe3)CMe, for which a similar threemembered ring intermediate was proposed.5 Other examples involving dimetalla-allyl ligands are known on trinuclear metal centres.6 Further studies on the reactivity of the unusual species 2 and 3 are currently in progress in our laboratory. Acknowledgements We thank the EPSRC for a studentship (to M. N. B.). References 1 M. Narita and C. U. Pittman, jun., Synthesis, 1976, 489; A. Krief, Tetrahedron, 1986, 42, 1209; J. M. Williams, J. R. Ferraro, R. J. Thorn, K. D. Carlson, U. Geiser, H. H. Wang, A. M. Kini and M.-H. Whangbo, Organic Superconductors: Synthesis, Structure, Properties and Theory, Prentice-Hall, Englewood CliVs, NJ, 1992. 2 H. Adams, M. N. Bancroft and M. J. Morris, Chem. Commun., 1997, 1445. 3 A. Abbott, M. N. Bancroft, M. J. Morris, G. Hogarth and S. P. Redmond, Chem. Commun., 1998, 389. 4 F. G. A. Stone, J. C. JeVery, K. A. Mead, H. Razay, M. J. Went and P. Woodward, J. Chem. Soc., Dalton Trans., 1984, 1383. 5 M. H. Chisholm, J. A. Heppert and J. C. HuVman, J. Am. Chem. Soc., 1984, 106, 1151. 6 E. Sappa, A. Tiripicchio and A. M. Manotti Lanfredi, J. Chem. Soc., Dalton Trans., 1978, 552; M. J. Morris, Ph.D. Thesis, University of Bristol, 1984. Received 2nd June 1998; Communication 8/04138E

ISSN:1477-9226    

DOI:10.1039/a804138e

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, Pages 2611–2613 2611 Structural parameters for the incomplete cuboidal cluster cation: [Nb3(Ï3-Cl)(Ï-O)3(OH2)9]41 in solutions of acid hydrolysed trivalent niobium from Nb K edge EXAFS David T. Richens * and Ian J. Shannon School of Chemistry, University of St. Andrews, North Haugh, St. Andrews, UK KY16 9ST A niobium K edge EXAFS study on the green solutions obtained following hydrolysis of NbCl3(dme) in aqueous non-oxidising strong acids has shown that the major species resulting in these solutions is the incomplete cuboidal chloridecapped triangular niobium cation, [Nb3(m3-Cl)(m-O)3(OH2)9]41.We have reported previously on the isolation and characterisation of the air-sensitive cationic species present within the green aqueous solutions that result following the aqueous acid hydrolysis of trivalent niobium complexes such as [NbCl3- (dme)] 1,2 and [Nb2Cl6(THT)3] (THT = tetrahydrothiophene).3 This species was first reported by Cotton and co-workers in the mid 1980s.3 Similar green solutions have also recently been reported to result from zinc reduction of NbCl5 in HCl–ethanol mixtures.4 A combination of redox titrations and oxygen-17 labelling NMR studies following cation-exchange column puri- fication has provided strong indications that the principal species present in these solutions is an incomplete triangular cluster ion of proposed structure [Nb3(m3-Cl)(m-O)3(OH2)9]41 [delocalised Nb(III,IV,IV)].1 Related work has shown that the putative chloride-capping ligand in the green ion can be readily replaced by e.g.sulfur, as in the complex [NH4]3[NMe4]3[Nb3- (m3-S)(m-O)3(NCS)9],3 or hydroxotrioxoborate, as in [Nb3- {m3-BO3(OH)}(m-O)3(HBpz3)3] (pz = pyrazol-1-yl).2 In the continuing absence of definitive X-ray structural data on the green cation we have carried out a niobium K edge EXAFS investigation on concentrated solutions of the green ion in CF3SO3H solutions. The findings have provided compelling evidence for the existence of the cluster [Nb3(m3-Cl)(m-O)3- (OH2)9]41 in these solutions following comparisons with a corresponding molybdenum K edge EXAFS study recently carried out on the structurally analogous incomplete cuboidal trinuclear cluster [Mo3(m3-S)(m-O)3(OH2)9]41 under similar conditions.5 A concentrated sample of the green niobium cation (0.06 mol dm23 per Nb3) in 3.0 mol dm23 CF3SO3H was prepared as described previously 1,2 following chromatographic elution with aqueous HCl, evaporation of the HCl eluates under vacuum and dissolution of the resulting chloride salt in aqueous 3.0 mol dm23 CF3SO3H.Samples were then loaded into a specially designed Perspex sample cell contructed with polyester (Mylar) windows for EXAFS measurements.5,6 The EXAFS spectra were collected on the Wiggler 1 beam line station 9.2 at the Synchrotron Radiation Source at the UK CLRC Daresbury Laboratory operating at 2 GeV (eV ª 1.602 × 10219 J) and 200 mA.The station was equipped with a water cooled harmonic rejecting double crystal Si(220) monochromator and mixtures of argon and helium gas ion chambers for measuring incident (Io) and transmitted (It) beam intensities respectively, and a 13 element Ge fluorescence detector manufactured by Canberra. Data were recorded at the metal K edge in fluorescence mode. For each sample four or five scans were recorded and combined to improve signal to noise.The sample temperature was 25.0 ± 0.5 8C. The suite of programs used to analyse the EXAFS data was that provided by the CLRC Daresbury Laboratory.7 The raw EXAFS data were processed using the program EXCALIB and the position of the absorption edge was determined from the derivative of the spectrum using EXBROOK. This program was also used to carry out background subtraction in order to extract the EXAFS function c(k). A k3 weighting was used to enlarge the oscillations at large k.These oscillations are then Fourier transformed to give a quasi-radial distribution function. Fitting of the incomplete cuboidal triangular structure was carried out with the EXCURV 92 8 program using curved wave theory and employing Hedin–Lindquist ground states and von Barth exchange potentials 9 to calculate appropriate phase shifts along with typical M]O, M]Cl, M]S and M]M distances from crystallographic data. The quality of fit is reported relative to the discrepancy index, R, and the goodness of fit relative to the fit index, Fl.10 The energy independent amplitude reduction factor (AFAC) 8 is a measure of the proportion of the electrons which contribute to an EXAFS-type scatter.It allows for the reduction in amplitude due to the presence of multiple excitations and is usually set to be in the range 0.7–0.9. During fitting the Debye–Waller factor (2s2) and the occupation number (N) were independently refined due to a high correlation. Comparative data taken from solutions of the green niobium cation and from similarly concentrated solutions of Mo3- (m3-S)(m-O)3 41(aq) (0.06 mol dm23) in 2.5 mol dm23 HClO4, prepared in a similar manner,5 is shown in Fig. 1. An excellent fit to the unfiltered metal K edge EXAFS is found in each case for an incomplete cuboidal triangular structure [M3(m3-X)- (m-O)3(OH2)9 2 nCln](4 2 n)1 (X = S for M = Mo; X = Cl for M = Nb). For both compounds the EXAFS data refines very well to a four shell model consisting of two oxygens atoms (m-O) at 1.91 (Mo), 2.04 Å (Nb); three terminal ligands (H2O or Cl2) at 2.16 (Mo), 2.22 Å (Nb); two heavy metal atoms at 2.60 (Mo), 2.78 Å (Nb) and one capping third period atom at 2.34 (S in the case of the Mo cluster), 2.49 Å (Cl in the case of the Nb cluster) with estimated errors ±0.03 Å, Table 1.The calculated bond distances 5 for the Mo3(m3-S)(m-O)3 41 core have been shown to be in excellent agreement with those obtained from two crystal structures of complexes of the Mo3(m3-S)(m-O)3 41 core with L-cysteinate and hydrogen nitrilotriacetate reported by Shibahara et al.11 and this gives weight to the reliability of the Hedin–Lindquist exchange potentials and von Barth ground states used in this work for calculating appropriate phase shifts for the refinement of the EXAFS data on these aqueous species.The similarity in the overall EXAFS patterns and Fourier transformed data apparent for both species in Fig. 1, provides compelling evidence for the presence of the Nb3(m3- Cl)(m-O)3 41 core in solutions of the green niobium cation. Niobium K edge EXAFS data were subsequently taken from a more diluted sample of the green cation (0.01 mol dm23 per2612 J. Chem. Soc., Dalton Trans., 1998, Pages 2611–2613 Table 1 Comparison of metal K edge EXAFS data on aqueous [M3(m3-X)(m-O)3(OH2)9]41 species Species M]O(m)/Å M]X(m3)/Å M]M/Å M]OH2/Å [Mo3(m3-S)(m-O)3(OH2)9]41 (0.06 mol dm23 per Mo3 in 2.5 mol dm23 HClO4 solution) Refined EXAFS distances (ca.±0.03 Å) Occupation numbers (N) Debye–Waller factors (2s2) 1.91 2.0 0.008 2.34(S) 1.0 0.009 2.60 2.0 0.004 2.16 3.0 0.009 AFAC = 0.77; R (%) = 23.0351;* Fl = 0.000 29.* Green niobium aqueous cation (0.06 mol dm23 per Nb3 in 3.0 mol dm23 CF3SO3H solution) Refined EXAFS distances (ca. ±0.03 Å) Occupation numbers (N) Debye–Waller factors (2s2) 2.04 2.0 0.005 2.49(Cl) 1.0 0.007 2.78 2.0 0.006 2.22 3.0 0.009 AFAC = 0.78; R (%) = 30.5547;* Fl = 0.000 55.* * R is the discrepancy index = Ú|cth(k) 2 cexp(k)|k3 dk/Ú|cexp(k)|k3 dk × 100%,10 which measures the quality of fit.Fl is the fit index = S(ki)n(ci th 2 ci exp) 10 which measures the goodness of fit where cth and cexp are the theoretical and experimental EXAFS respectively. Fig. 1 Experimental (——) and calculated (-----) metal K edge EXAFS spectra and Fourier transformed data from (a, b) the green niobium aqueous cation (0.06 mol dm23 per Nb3 in 3.0 mol dm23 CF3SO3H and (c, d) [Mo3(m3-S)(m-O)3(OH2)9]41 (0.06 mol dm23) in 2.5 mol dm23 HClO4 (phase shifts calculated using Hedin–Lindquist ground states and von Barth exchange potentials) Nb3) in 3.0 mol dm23 CF3SO3H.This was to check whether the rather large Debye–Waller factors, coupled with rather long distances to the terminal atoms particularly in the case of the niobium ion (Table 1), could be due to the presence of some coordinated Cl2 at the terminal sites as a result of incomplete aquation in solutions of the chloride salt at the higher concentrations (ª0.06 mol dm23).The refined EXAFS data are shown in Fig. 2 and Table 2. The poorer signal to noise (R = 34.8387, Fl = 0.000 66) reflected the lower niobium concentration. However the four backscattering shells around each niobium atom were more easily discernible fitting well to the fully aquated cluster [Nb3(m3-Cl)(m-O)3(OH2)9]41 with two oxygens (m-O) at 2.01 Å; three oxygens (H2O molecules) at 2.17 Å; two niobium atoms at 2.75 Å and the single capping chlorine atom at 2.49 Å, errors ±0.03 Å, Table 2.Significantly smaller Debye–Waller factors were found, increasing steadily with distance of backscatterer from the absorbing atom, consistent with the more well defined co-ordination sphere around each niobium atom in the now fully aquated cluster. The fitted distances for the Nb3(m3-Cl)(m-O)3 41 core are found to be in excellent agreement with those for the structurally analogous m3-sulfur-capped Nb3(m3-S)(m-O)3 41 core present in the crystal structure of the complex, [Nb3(m3-S)(m-O)3- (NCS)9]62,3 shown in Table 2 for comparison.Similarly the Nb]Cl distance to the single capping chlorine atom evaluated here (2.49 Å) is in excellent agreement with that found in theJ. Chem. Soc., Dalton Trans., 1998, Pages 2611–2613 2613 Table 2 Comparison of Nb K edge EXAFS data from the green niobium aqueous cation (0.01 mol dm23 per Nb3) in 3.0 mol dm23 CF3SO3H with crystal structure data on [Nb3(m3-X)(m-Y)3L9] species Species Nb]O(m)/Å Nb]X(m3)/Å Nb]Nb/Å Nb]L/Å Nb]O]Nb/8 Green niobium aqueous cation {[Nb3(m3-Cl)(m-O)3(OH2)9]41}: Refined EXAFS data (ca.±0.03 Å) Refined occupation number (N) Debye–Waller factors (2s2) 2.01 2.1 0.001 2.49(Cl) 1.0 0.003 2.75 1.9 0.004 2.17 2.9 0.002 87(1)* AFAC = 0.78; R (%) = 34.8387; Fl = 0.000 Crystal structural data on [Nb3(m3-X)(m-Y)3L9] species: [Nb3(m3-S)(m-O)3(NCS)9]623 [Nb3(m3-Cl)(m-Cl)3Cl6(PEt3)3]211 [Nb3(m3-Cl)(m-Cl)3Cl3(PEt2Ph)6]211 66. 2.03 —— 2.51(S) 2.51(Cl) 2.47(Cl) 2.76 2.97 2.83 2.14(N) 86(1) * Calculated value. crystal structures of two other known m3-Cl capped triangular niobium cluster complexes, [Nb3(m3-Cl)(m-Cl)3Cl6(PEt3)3]2 (2.51 Å) 12 and [Nb3(m3-Cl)(m-Cl)3Cl3(PEt2Ph)6]2 (2.47 Å).12 This excellent agreement in bond distance parameters provides a powerful vindication of the choice of calculated phase shifts.In conclusion, the success in extracting highly reliable structural parameters for the well characterised [Mo3(m3-S)(m-O)3- (OH2)9]41 ion, following fits to the refined metal K edge EXAFS pattern obtained from solutions of the above ion,5 has allowed meaningful evaluation of Nb K edge EXAFS data taken correspondingly from solutions of the green niobium cation. Fits to the latter EXAFS data, along with comparisons from known structural data on several triangular niobium complexes, provides compelling evidence that the green aqueous solutions Fig. 2 Experimental (——) and calculated (-----) Nb K edge EXAFS spectrum (a) and Fourier transformed data (b) from the green niobium aqueous cation (0.01 mol dm23 per Nb3) in 3.0 mol dm23 CF3SO3H obtained following acid hydrolysis of trivalent chloroniobium complexes,1–3 or via zinc reduction of NbCl5 in ethanolic HCl,4 contain the chlorine-capped incomplete cuboidal mixed valence cluster [Nb3(m3-Cl)(m-O)3(OH2)9]41 1.Acknowledgements We wish to thank the CLRC for an allocation of beamtime on the Synchrotron Radiation Source at Daresbury and also Dr. J. F. W. Mosselmans (Daresbury Laboratory) for assistance with the recording and initial processing of the EXAFS data. I. J. S. would like to thank the Royal Society of Edinburgh for the provision of a BP/RSE Research Fellowship. References 1 S. Minhas and D. T. Richens, J. Chem. Soc., Dalton Trans., 1996, 703. 2 S. Minhas, A.Devlin, D. T. Richens, A. C. Benyei and P. Lightfoot, J. Chem. Soc., Dalton Trans., 1998, 953. 3 F. A. Cotton, M. P. Diebold, R. Llusar and W. J. Roth, J. Chem. Soc., Chem. Commun., 1986, 1276. 4 B.-L. Ooi, T. Shibahara, G. Sakane and K.-F. Mok, Inorg. Chim. Acta, 1997, 266, 103. 5 G. Lente, A. M. Dobbing and D. T. Richens, Inorg. React. Mech., 1998, 1, 3. 6 J. R. Osman, D. T. Richens and J. A. Crayston, Inorg. Chem., 1998, 37, 1665. 7 EPSRC (UK) Daresbury Laboratory Program Library, 1991. 8 N. Binsted, J. W. Campbell, S. J. Gurman and P. L. Stephenson, EPSRC (UK) Daresbury Laboratory Program Library, 1992; N. Binsted, R. W. Strange and S. S. Hasnain, Biochemistry, 1992, 31, 12 117. 9 U. von Barth and L. Hedin, Int. J. Phys., 1972, C5, 1629. 10 N. Binsted, J. Evans, N. G. Greaves and R. J. Price, Organometallics, 1989, 8, 613. 11 T. Shibahara, H. Akashi, S. Nagahata, H. Hattori and H. Kuroya, Inorg. Chem., 1989, 28, 362. 12 F. A. Cotton, M. P. Diebold, X.Feng and W. J. Roth, Inorg. Chem., 1988, 27, 3414. Received 5th June 1998; Communication 8/04247KJ. Chem. Soc., Dalton Trans., 1998, Pages 2611–2613 2613 Table 2 Comparison of Nb K edge EXAFS data from the green niobium aqueous cation (0.01 mol dm23 per Nb3) in 3.0 mol dm23 CF3SO3H with crystal structure data on [Nb3(m3-X)(m-Y)3L9] species Species Nb]O(m)/Å Nb]X(m3)/Å Nb]Nb/Å Nb]L/Å Nb]O]Nb/8 Green niobium aqueous cation {[Nb3(m3-Cl)(m-O)3(OH2)9]41}: Refined EXAFS data (ca.±0.03 Å) Refined occupation number (N) Debye–Waller factors (2s2) 2.01 2.1 0.001 2.49(Cl) 1.0 0.003 2.75 1.9 0.004 2.17 2.9 0.002 87(1)* AFAC = 0.78; R (%) = 34.8387; Fl = 0.000 Crystal structural data on [Nb3(m3-X)(m-Y)3L9] species: [Nb3(m3-S)(m-O)3(NCS)9]623 [Nb3(m3-Cl)(m-Cl)3Cl6(PEt3)3]211 [Nb3(m3-Cl)(m-Cl)3Cl3(PEt2Ph)6]211 66. 2.03 —— 2.51(S) 2.51(Cl) 2.47(Cl) 2.76 2.97 2.83 2.14(N) 86(1) * Calculated value. crystal structures of two other known m3-Cl capped triangular niobium cluster complexes, [Nb3(m3-Cl)(m-Cl)3Cl6(PEt3)3]2 (2.51 Å) 12 and [Nb3(m3-Cl)(m-Cl)3Cl3(PEt2Ph)6]2 (2.47 Å).12 This excellent agreement in bond distance parameters provides a powerful vindication of the choice of calculated phase shifts.In conclusion, the success in extracting highly reliable structural parameters for the well characterised [Mo3(m3-S)(m-O)3- (OH2)9]41 ion, following fits to the refined metal K edge EXAFS pattern obtained from solutions of the above ion,5 has allowed meaningful evaluation of Nb K edge EXAFS data taken correspondingly from solutions of the green niobium cation.Fits to the latter EXAFS data, along with comparisons from known structural data on several triangular niobium complexes, provides compelling evidence that the green aqueous solutions Fig. 2 Experimental (——) and calculated (-----) Nb K edge EXAFS spectrum (a) and Fourier transformed data (b) from the green niobium aqueous cation (0.01 mol dm23 per Nb3) in 3.0 mol dm23 CF3SO3H obtained following acid hydrolysis of trivalent chloroniobium complexes,1–3 or via zinc reduction of NbCl5 in ethanolic HCl,4 contain the chlorine-capped incomplete cuboidal mixed valence cluster [Nb3(m3-Cl)(m-O)3(OH2)9]41 1.Acknowledgements We wish to thank the CLRC for an allocation of beamtime on the Synchrotron Radiation Source at Daresbury and also Dr. J. F. W. Mosselmans (Daresbury Laboratory) for assistance with the recording and initial processing of the EXAFS data. I. J. S. would like to thank the Royal Society of Edinburgh for the provision of a BP/RSE Research Fellowship. References 1 S. Minhas and D. T. Richens, J. Chem. Soc., Dalton Trans., 1996, 703. 2 S. Minhas, A. Devlin, D. T. Richens, A. C. Benyei and P. Lightfoot, J. Chem. Soc., Dalton Trans., 1998, 953. 3 F. A. Cotton, M. P. Diebold, R. Llusar and W. J. Roth, J. Chem. Soc., Chem. Commun., 1986, 1276. 4 B.-L. Ooi, T. Shibahara, G. Sakane and K.-F. Mok, Inorg. Chim. Acta, 1997, 266, 103. 5 G. Lente, A. M. Dobbing and D. T. Richens, Inorg. React. Mech., 1998, 1, 3. 6 J. R. Osman, D. T. Richens and J. A. Crayston, Inorg. Chem., 1998, 37, 1665. 7 EPSRC (UK) Daresbury Laboratory Program Library, 1991. 8 N. Binsted, J. W. Campbell, S. J. Gurman and P. L. Stephenson, EPSRC (UK) Daresbury Laboratory Program Library, 1992; N. Binsted, R. W. Strange and S. S. Hasnain, Biochemistry, 1992, 31, 12 117. 9 U. von Barth and L. Hedin, Int. J. Phys., 1972, C5, 1629. 10 N. Binsted, J. Evans, N. G. Greaves and R. J. Price, Organometallics, 1989, 8, 613. 11 T. Shibahara, H. Akashi, S. Nagahata, H. Hattori and H. Kuroya, Inorg. Chem., 1989, 28, 362. 12 F. A. Cotton, M. P. Diebold, X. Feng and W. J. Roth, Inorg. Chem., 1988, 27, 3414. Received 5th June 1998; Communication 8/04247K

ISSN:1477-9226    

DOI:10.1039/a804247k

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 2615–2623 2615 Luminescent and redox-active ruthenium(II) and osmium(II) complexes with a rigid allene-bridged polyphosphine † Bo Hong,* Steven R. Woodcock, Sylvia K. Saito, JeVrey V. Ortega Department of Chemistry, University of California, Irvine, CA 92697-2025, USA A series of monometallic and homo- and hetero-bimetallic RuII and OsII complexes with the polyphosphines 1,19- bis(diphenylphosphino)ethene (bppe) and 1,19,3,39-tetrakis(diphenylphosphino)allene (tppa) have been prepared and characterized by 31P-{1H} NMR spectroscopy, elemental analysis and fast atom bombardment mass spectral analysis.Their ground-state properties (including electrochemical behavior and electronic absorption) and excitedstate properties (including luminescence, quantum yields and MLCT excited state lifetimes) are reported herein. All complexes exhibit room-temperature luminescence, with long-lived 3MLCT excited states observed for monometallic and homobimetallic OsII complexes with both bppe and tppa ligands.In the heterobimetallic RuII–OsII complex [Ru(bpy)2(m-tppa)Os(bpy)2][PF6] (bpy = 2,29-bipyridine) the luminescence of the Ru(bpy)2-based unit is quenched by the connected Os(bpy)2-based unit via energy transfer across the tppa spacer containing the allene bridge. Two overlapping one-electron oxidations, corresponding to the two MII-based moieties (M = Ru or Os) bridged by the tppa spacer, are observed in the homobimetallic complexes.The construction of low-dimensional artificial supramolecular systems from a discrete number of molecular components has been an area of ongoing interest.1 Through the linkage of various prefabricated molecular components with light-related properties, photochemical molecular systems have been obtained to study the luminescence and redox properties, and intramolecular electron/energy transfer processes.1–4 These systems will include several representative features.Synthetically we must have control of the choice, orientation and spacing of auxiliary ligand, donor/acceptor and spacer. To ensure the directionality and also construct multicomponent supramolecular systems with well defined structures, rigid spacers must be used to aVord restricted conformational mobility and controllable distances between structural subunits. In addition, we must also have means to identify the mode and mechanism of energy or electron transfer as well as the electronic interaction between subunits spanned by spacers.The role of spacers in such systems can be multifaceted; spacers can serve as conducting components to promote long range electronic communication and photoinduced electron/ energy transfer, or as passive connecting components. In our search for suitable systems, we found that polyphosphines with rigid sp hybridized cumulenic carbon chains (Cn) are ideal candidates due to their rigidity, co-ordination versatility, and chemical/photochemical stability upon formation of various metal complexes.5 Despite their versatile applications as complexing ligands,5,6 polyphosphines have been much less incorporated in the construction of photochemical molecular systems when compared with N-containing ligands,7 especially polypyridines.The use of phosphine groups in spacers provides two advantages. First, they can serve as linkage components between the Cn chains and metal centers.Second, it has been reported 8–10 that the incorporation of phosphines in the polypyridyl –osmium(II) complexes can enhance the lifetimes of 3MLCT states. In this paper, we report the synthesis, characterization and photophysical studies of monometallic (2a and 2b) and bimetallic (2c–2e) ruthenium(II) and osmium(II) complexes of a specific tetratopic phosphine, namely, 1,19,3,39-tetrakis- (diphenylphosphino)allene (tppa) with an unsaturated C3 chain.Comparison of the redox chemistry and excited proper- † Non-SI unit employed: eV ª 1.602 × 10219 J. ties of these complexes with the analogous monometallic complexes (1a and 1b) of 1,19-bis(diphenylphosphino)ethene (bppe) will be carried out. Electronic absorption, steady-state and time-resolved emission spectroscopy have been used to study the ground-state absorption and redox properties, the excitedstate decay kinetics, and the energy-transfer process in this system. The 3MLCT excited-state quantum yields and lifetimes of bimetallic complexes 2c–2e are compared with the corresponding monometallic model complexes 2a and 2b, and the observed rate constant of energy transfer is estimated, from the MLCT excited-state lifetimes, for the structurally rigid and fixed bimetallic complex 2e.Results and Discussion Synthetic routes The preparation of monometallic and bimetallic complexes in this study reveals several interesting reactivity patterns. For any metal center, the polyphosphines bppe and tppa can coordinate in a bis(chelating) mode or serve as a monodentate ligand.An extended reaction time is required in order to obtain C C H H Ph2 P P Ph2 2+ M = Ru 1a, Os 1b C C PPh2 PPh2 Ph2 P P Ph2 2+ M = Ru 2a, Os 2b C P Ph2 Ph2 P Ph2 P P Ph2 4+ M, M¢ = Ru 2c, Os 2d M = Ru, M¢ = Os 2e C C M¢(bpy)2 C (bpy)2M (bpy)2M (bpy)2M2616 J. Chem. Soc., Dalton Trans., 1998, Pages 2615–2623 Scheme 1 N N Os N N Cl Cl C C C PPh2 PPh2 Ph2P Ph2P C C C PPh2 PPh2 Ph2P N N Os N N P Ph2 Cl C C C PPh2 PPh2 Ph2 P N N Os N N P Ph2 C C C Ph2 P N N Os N N P Ph2 2+ Ph2 P N N Ru N N P Ph2 + THF–ethylene glycol, reflux 1+ 2+ + + the desired complexes with bis(chelating) phosphines.This is partly due to the fact that a four-membered chelate ring is formed. Another reason is the extremely diVerent solubility of the metal-based precursors and phosphines; while [M(bpy)2Cl2] (M = Ru or Os, bpy = 2,29-bipyridine) is soluble in ethylene glycol, all ligands are intractable in this solvent.Hence, THF is used to introduce ligands into various reaction mixtures, but the refluxing temperature is significantly lower. In the preparation of monometallic complexes 1a, 1b, 2a or 2b, a higher ligand-to-metal ratio (2.2–3 : 1) is used. In all cases, the desired complexes are obtained as the dominant products mixed with minor amounts of monosubstituted complexes [M(bpy)2Cl(L)]PF6 (L = bppe or tppa; M = Ru or Os) and/or bimetallic complexes, Scheme 1.Chromatographic methods have been successfully applied in the separation of the major products. It is found that the combination of basic alumina and acetonitrile or acetonitrile–toluene mixtures provides satisfactory separations and aVords products with suitable purities for further analysis. The homobimetallic complexes 2c–2d are obtained using a suitable metal-to-ligand ratio of typically 2.2 : 1. The heterobimetallic complex 2e is prepared by the reaction between 2b and an excess amount of [Ru(bpy)2Cl2], Scheme 2.All complexes have been purified using chromatographic separations. Scheme 2 C C C PPh2 PPh2 Ph2 P N N Os N N P Ph2 2+ C C C Ph2 P N N Os N N P Ph2 4+ Ph2 P N N Ru N N P Ph2 N N Ru N N Cl Cl Ethylene glycol, reflux + FAB-MS analysis Several interesting features are observed in the FAB-MS study of complexes with bppe and tppa. First, it has been found that FAB-MS is a relatively soft ionization technique for these metal complexes. Many of the fragment ions observed only involved sequential loss of counter anions (PF6 2) and PPh2 units, Table 1.Similar observations have been reported in the FAB-MS analysis of ruthenium and osmium complexes with 2,3,5,6- tetrabis(2-pyridyl)pyrazine (TPPZ).11a The inner sphere metal– ligand co-ordination was left intact, making peak identification straightforward. Second, the FAB mass spectra of metal– polyphosphine complexes contain numerous informative peaks in which the observed isotope distributions are the same as the simulated ones.Fig. 1(a) shows the representative observed and simulated isotope patterns of {[Os(bpy)2(bppe)]PF6}1 (m/z 1045). In addition to the above features, it is also found that although all metal complexes with polyphosphines are very stable in air (no additional peaks corresponding to the oxidized PPh2 units are observed in the 31P NMR spectra of these complexes), oxidation of the PPh2 unit to OPPh2 in phosphines and their metal complexes are observed during the FAB-MS analysis, Fig. 1(b) and 1(c). This is supported by the control experiment using both EI-MS and FAB-MS analysis of polyphosphines. As shown in Fig. 1(c), the FAB-MS spectrum of the ligand tppa gives a complicated spectrum with peaks corresponding to tppa itself (M1, m/z 777) and the oxidized ligands (M1 1 O 793, M1 1 2O 809, M1 1 3O 825), while the EI-MS analysis of the same sample gives a clean spectrum without any peaks corresponding to OPPh2 units [inset in Fig. 1(c)]. Electrochemical analysis Cyclic voltammetry has been used to obtain the redox potentials of all complexes synthesized, and the results are given in Table 2. In this section, the redox properties of metal-based oxidations in 1a, 1b and 2a–2e will be discussed, followed by the study of ligand-based reductions in these complexes. The cyclic voltammogram of each of the complexes [M(bpy)2- (bppe)][PF6]2 (M = Ru 1a or Os 1b) is characteristic of a reversible metal-based one-electron process, where reversibility, as used here, implies that the ip a : ip c ratio is found to be approximately unity.1,8,10 The OsII–OsIII redox couple in 1b is found to be more positive, 11.37 V (vs.SCE), when compared with that observed in [Os(bpy)3]21 (10.81 V vs. SCE) and [Os(bpy)2- (dppm)]21 (11.27 V vs. SCE, dppm = Me2PCH2CH2PMe2).8b,10 Similarly, the RuII–RuIII redox couple in 1a is also shifted to 11.70 V, an increase of 0.41 V when compared with 11.29 V inJ.Chem. Soc., Dalton Trans., 1998, Pages 2615–2623 2617 Table 1 The FAB-MS analysis of ruthenium and osmium complexes of bppe and tppa 1a 1b 2a 2b 2c 2d 2e Complex [Ru(bpy)2(bppe)][PF6]2 [Os(bpy)2(bppe)][PF6]2 [Ru(bpy)2(tppa)][PF6]2 [Os(bpy)2(tppa)][PF6]2 [Ru(bpy)2(m-tppa)Ru(bpy)2][PF6]4 [Os(bpy)2(m-tppa)Os(bpy)2][PF6]4 [Os(bpy)2(m-tppa)Ru(bpy)2][PF6]4 m/z 955 809 1045 900 1167 1151 1037 1021 1005 820 1441 1425 1280 1273 1257 1241 1127 1111 1095 911 1765 1620 1435 1417 2088 1943 1797 1280 1273 1257 1241 1145 1127 1111 1095 1817 1798 1776 1709 1653 1508 Relative abundance (%) 100 11 100 50 22 17 13 100 27 79 44 26 26 39 81 25 40 100 40 100 17 10 15 6 100 39 67 67 50 17 18 28 41 25 19 7 18 28 100 85 Assignment {[Ru(bpy)2(bppe)]PF6}1 [Ru(bpy)2(bppe)]1 {[Os(bpy)2(bppe)]PF6}1 [Os(bpy)2(bppe)]1 {[Ru(bpy)2(tppa)]PF6}1 2 PPh2 1 O {[Ru(bpy)2(tppa)]PF6}1 2 PPh2 [Ru(bpy)2(tppa)]1 2 PPh2 1 2O [Ru(bpy)2(tppa)]1 2 PPh2 1 O [Ru(bpy)2(tppa)]1 2 PPh2 [Ru(bpy)2(tppa)]1 2 2PPh2 {[Os(bpy)2(tppa)]PF6}1 1 O {[Os(bpy)2(tppa)]PF6}1 [Os(bpy)2(tppa)]1 {[Os(bpy)2(tppa)]PF6}1 2 PPh2 1 2O {[Os(bpy)2(tppa)]PF6}1 2 PPh2 1 O {[Os(bpy)2(tppa)]PF6}1 2 PPh2 [Os(bpy)2(tppa)]1 2 PPh2 1 2O [Os(bpy)2(tppa)]1 2 PPh2 1 O [Os(bpy)2(tppa)]1 2 PPh2 [Os(bpy)2(tppa)]1 2 2PPh2 {[Ru(bpy)2(tppa)Ru(bpy)2]PF6}1 1 O [Ru(bpy)2(tppa)Ru(bpy)2]1 1 O [Ru(bpy)2(tppa)Ru(bpy)2]1 2 PPh2 1 O [Ru(bpy)2(tppa)Ru(bpy)2]1 2 PPh2 {[Os(bpy)2(tppa)Os(bpy)2][PF6]2}1 1 O {[Os(bpy)2(tppa)Os(bpy)2]PF6}1 1 O [Os(bpy)2(tppa)Os(bpy)2]1 1 O [Os(bpy)2(tppa)]1 {[Os(bpy)2(tppa)]PF6}1 2 PPh2 1 2O {[Os(bpy)2(tppa)]PF6}1 2 PPh2 1 O {[Os(bpy)2(tppa)]PF6}1 2 PPh2 [Os(bpy)2(tppa)]1 2 PPh2 1 3O [Os(bpy)2(tppa)]1 2 PPh2 1 2O [Os(bpy)2(tppa)]1 2 PPh2 1 O [Os(bpy)2(tppa)]1 2 PPh2 {[Os(bpy)2(tppa)Ru(bpy)2][PF6]2}1 1 F {[Os(bpy)2(tppa)Ru(bpy)2][PF6]2}1 {[Os(bpy)2(tppa)Ru(bpy)2][PF6]2}1 2 F [Os(bpy)2(tppa)Ru(bpy)2]1 1 O {[Os(bpy)2(tppa)Ru(bpy)2]PF6}1 2 PPh2 [Os(bpy)2(tppa)Ru(bpy)2]1 2 PPh2 Table 2 Formal potentials (vs.SCE) for Ru and Os complexes with bppe and tppa E/V vs. SCE [relative current intensity] (peak separation, DEp/mV) Complex 1a 1b 2a 2b 2c 2d 2e RuII/III 11.70 [1] (80) 11.55 [1] (70) 11.07 [2] (96) 11.63 [1] (73) OsII/III 11.37 [1] (70) 11.12 [1] (74) 10.77 [2] (100) 10.98 [1] (82) tppa0/2 21.09 [1] (72) 21.10 [1] (60) 21.05 [1] (66) 21.01 [1] (64) 21.14 [1] (72) First bpy0/2 21.24 [1] (60) 21.21 [1] (64) 21.39 [1] (62) 21.32 [1] (82) 21.35 [2] (94) 21.33 [2] (86) 21.45 [2] (123) DE2� 1 * 2.94 2.58 2.64 2.20 2.12 1.72 2.12 * DE2� 1 = E2� 1 (MII/III) 2 E2� 1 (first ligand reduction), M = Ru or Os.In compound 2e, the first metal oxidation is used in the calculation. [Ru(bpy)3][PF6]2.2,3 This indicates that the species with one chelate bppe ligand are much harder to oxidize. Presumably, this shift is due to the change from more electron donating bpy to the bppe ligand with one double bond.Hence, the MII centers in [M(bpy)3]21 are more electron rich. For each of the monometallic complexes with tppa, [M(bpy)2- (tppa)][PF6]2 (M = Ru 2a or Os 2b), a characteristic single metal-based one-electron redox wave is observed. The RuII– RuIII (11.55 V vs. SCE) and OsII–OsIII (11.12 V vs. SCE) redox potentials are found to significantly shift towards less positive potentials when compared with 1a and 1b, correspondingly.Each of the homobimetallic complexes [M(bpy)2(m-tppa)- M(bpy)2][PF6]4 (M = Ru 2c or Os 2d) features two overlapping one-electron processes, corresponding to two RuII centers at 11.07 V and two OsII centers at 10.77 V (vs. SCE), respectively (Fig. 2). No second metal-based oxidation is observed in the scanned region 21.6 V to 12.0 V. This indicates that the redox interaction between the two metal centers is rather weak, giving simultaneous one-electron oxidation of two metal centers.11c Presumably this is due to the orthogonality of the two allenic double bonds, which places the two co-ordination planes of the diphosphine chelates at 908 relative to each other.Here, an additional shift of 0.35–0.48 V towards the less positive potential is observed when compared with the corresponding monometallic complexes 2a and 2b. This shift may be due, partially, to the fact that two metal centers are now attached to the central tppa and, as a result, back bonding from each metal center to the polyphosphine tppa is reduced.The heterobimetallic complex [Os(bpy)2(m-tppa)Ru(bpy)2]- [PF6]4 2e features two one-electron processes at 11.63 and2618 J. Chem. Soc., Dalton Trans., 1998, Pages 2615–2623 10.98 V (vs. SCE). These two peaks are assigned to RuII–RuIII and OsII–OsIII redox couples, correspondingly, based on the observed redox potentials in monometallic complexes 2a and 2b. Upon scanning anodically, the osmium(II) center is first oxidized at 10.98 V followed by the oxidation of the ruthenium(II) center at 11.63 V, equation (1).RuII(tppa)OsII e2 RuII(tppa)OsIII e2 RuIII(tppa)OsIII (2) Several characteristic features are observed for the ligand reduction waves of 1a, 1b and 2a–2e. First, both monometallic complexes 1a and 1b exhibit consecutive one-electron reduction peaks, with the first peak at 21.24 V and 21.21 V for 1a and 1b, respectively (Table 2). Previously, the reported redox potential of the bpy0/2 couple was in the range of 21.27 V to 21.31 V for monometallic and bimetallic OsII complexes with 1,2,4,5- tetrakis(diphenylphosphino)benzene.10 In the electrochemical study of the [Os(bpy)2(PPh2CH]] CHPPh2)]21 complex, the first Fig. 1 (a) Computer simulated and observed isotope patterns of the FAB-MS peak of {[Os(bpy)2(bppe)]PF6}1 (m/z 1045). (b) Comparison of the EI-MS (inset) and FAB-MS analysis of the tppa ligand. (c) The FAB-MS analysis of 2b {M = [Os(bpy)2(tppa)][PF6]2} ligand reduction peak at 21.26 V (vs.SCE) was also assigned to bpy0/2.8c Hence, the first ligand reduction peaks in 1a and 1b are assigned to the bpy0/2 couple. The remaining second ligand reduction (21.50 V in 1a and 21.43 V in 1b) corresponds to the second bpy0/2 redox couple.9b No additional redox wave corresponding to bppe0/2 is observed in the scanned range from 22.0 to 12.0 V. Second, all of the first ligand reduction peaks of the complexes with tppa (2a–2e) are shifted towards the positive potentials when compared with those observed in complexes with bppe, Table 2.These redox potentials are much more positive than those found for the bpy0/2 couple.8c,10 It is possible that the allene-bridged tppa spacer is reduced before reduction of the auxiliary bpy ligands, and the first one-electron ligction in 2a–2e is assigned to the tppa0/2 redox couple. The second ligand reduction is assigned to the first bpy0/2 redox wave, Table 2.A similar observation has been reported for the bimetallic complex [Ru(tpy)(L)Os(tpy)] [tpy = 2,29:69,20-terpyridine, L = 1,4-bis(2,29:69,20-terpyridin-49-yl)buta-1,3-diyne].11b The buta- 1,3-diynyl carbon chain bridge is reduced at 21.04 V vs. SCE before reduction of the auxiliary tpy ligand at 21.34 V. Third, the observed peak current of the first ligand reduction (assigned to tppa0/2) is about half of the peak current of the second ligand reduction wave (assigned to the first bpy0/2 of each metal center) in the bimetallic complexes 2c–2e.This is consistent with a one-electron reduction of the tppa0/2 couple in the former and the two overlapping one-electron bpy0/2 reductions in the latter (one bpy per metal center). For monometallic complexes 1a, 1b, 2a and 2b, the observed relative peak currents of the ligand based reductions are approximately equivalent in each complex, corresponding to the one-electron reductions of tppa0/2 and bpy0/2 couples.Electronic absorption spectra The absorption maxima and the corresponding absorption coef- ficients of complexes 1a, 1b and 2a–2e are listed in Table 3 and their visible region spectra are compared in Fig. 3. The Ru monometallic complex 1a with bppe exhibits the lowest MLCT (Ru æÆ bpy) bands at 380 nm (predominantly triplet in character 8a,10) and the corresponding Os complex 1b exhibits the lowest MLCT (Os æÆ bpy) bands at 500 nm (triplet). When compared with the observed values of MLCT bands for [M(bpy)3][PF6]2 (lmax = 452 nm for M = Ru,12 and 640 nm for M = Os8b), the corresponding MLCT bands in 1a and 1b are Fig. 2 Cyclic voltammograms of [(bpy)2M(tppa)M(bpy)2][PF6]4 (M = Ru, a; Os, b): in MeCN, 0.1 mol dm23 Bu4NPF6, Pt disc working electrode, Pt wire counter electrode, Ag–AgCl reference electrode, scan rate 100 mV s21J. Chem. Soc., Dalton Trans., 1998, Pages 2615–2623 2619 Table 3 Summary of spectroscopic data labs/nm 1023F t a/ns (±10%) Complex 1a 1b 2a 2b 2c 2d 2e (e/dm3 mol21 cm21) 280 (39 840), 380 (11 300) 280 (32 600), 385 (8960), 500 (2840) 290 (40 700), 400 (9810), 450 (sh) (6040) 290 (44 050), 360 (13 050), 450 (8100), 550 (sh) (2800) 290 (57 200), 320 (17 900), 450 (7950) 295 (47 900), 370 (8950), 460 (6700), 605 (2050) 290 (71 200), 420 (13 200), 550 (sh) (3100) lem a/nm 530 600 535 570 560 580 590 (±10%)a,b 0.40 30 0.11 2.3 0.16 1.9 7.0 Ru-based 25 40 45 0.33 d 5.9 d Os-based 190 350 520 410 d hisckr c/s21 1.6 × 104 1.6 × 105 2.8 × 103 6.6 × 103 3.6 × 103 3.7 × 103 1.7 × 104 a lex = 470 nm, in acetonitrile.All solution samples were treated with three freeze–pump–thaw cycles before the measurements of emission and quantum yields. Luminescence lifetimes are obtained from the least-squares fit of single or double exponential decay. b Fem vs. [Ru(bpy)3][PF6]2 (Fem = 0.062). c hisckr = Femt21. In 2e, t2 is used in the calculation. d The short component in the lifetime of 2e was measured on a frequency domain fluorimeter (measurable range 50 ps–50 ns, excitation at 470 nm, monitored broadband with a cut-oV filter at 500 nm).The long component in the lifetime of 2e is obtained (as is the rest of the data in this table) on a nanosecond laser system equipped with a Continuum Surelite II-10 Nd:YAG laser (lowest measurable lifetime is ca. 10 ns with a 6 ns laser pulse width; excitation at 470 nm, monitoring emission wavelength at 590 nm). significantly blue-shifted.Similar blue-shift of the MLCT band has also been reported in the previous study of [Os(bpy)2(L)2]21 (L = phosphines such as PPh3, PMe2Ph, dppm, etc.) complexes. 8b,10,13a This blue-shift is ascribed to the stabilization of the ground state by the enhanced dp(M)–dp(L) back bonding and the destabilization of the excited state by the poorer s-donating phosphine ligand.10,13a The 3MLCT (M æÆ bpy) bands of complexes 2a and 2b bearing tppa are broad with lmax at approximately 450 nm for 2a and 550 nm for 2b, all of which are red-shifted from the observed values of 1a and 1b.The homobimetallic complexes 2c and 2d exhibit charge-transfer bands with lmax = 450 and 460 nm, respectively. An additional weak absorption band at 605 nm was also observed in 2d, Fig. 3. These values are further redshifted from the corresponding monometallic complex 2a and 2b. In the heterobimetallic complex 2e a broad absorption band with lmax at ca. 420 nm is observed, with a shoulder peak tailing into 550 nm.Emission and excited state lifetimes Steady-state and time-resolved emission spectroscopy have been used to study the excited-state properties, including roomtemperature luminescence, quantum yields and luminescence lifetimes. All data are summarized in Table 3. When compared with the luminescence properties of [M(bpy)3][PF6]2 (M = Ru, lem = 620 nm;14a M = Os, lem = 723 nm12a), blue-shifted emission maxima (shift of about 60–95 nm in Ru complexes 1a, 2a and 2c, and of about 120–150 nm for Os complexes 1b, 2b and 2d) have been observed in all mono- and Fig. 3 Comparison of the electronic absorption spectra of 1a, 1b and 2a–2e homobi-metallic complexes with bppe and tppa ligands. This indicates that the replacement of one bpy ligand by either bppe or tppa increases the energy gap between the ground state and the 3MLCT excited state. Such a blue-shift in emission is expected when a stronger ligand (such as a polyphosphine) is used to replace a relatively weak ligand (such as bpy).8 Time-resolved emission studies of complexes 1a, 1b and 2a– 2e have also been carried out at room temperature.The emission decay traces of all complexes, except 2e, fit to singleexponential decay curves, giving excited-state lifetimes as listed in Table 3. A representative decay trace and fit for 2d are shown in Fig. 4. The lifetimes of all OsII complexes are much longer than that of [Os(bpy)3][PF6]2 (t = 20 ns in acetonitrile 10,12b).Such a change in lifetime can be ascribed to the increased energy gap between the ground state and the emitting 3MLCT state, according to the energy gap law.8–10 Although phosphines do not absorb light, they will cause an increase of the 3MLCT excited states and, consequently, slow the non-radiative decay of OsII complexes and result in longer excited-state lifetimes. In contrast to OsII complexes, RuII compounds with bppe and tppa have much shorter lifetimes when compared with [Ru- (bpy)3][PF6]2 (t = 855 ns in acetonitrile 14a), presumably due to the mixing of the lowest dp(M)–p*(L) charge transfer band with higher energy excited states.10,12 While blue-shifted emis- Fig. 4 Excited state decay trace and fit of 2d: lex = 470, lem = 590 nm, in spectrograde acetonitrile at 295 K (freeze–pump–thaw three times prior to use)2620 J. Chem. Soc., Dalton Trans., 1998, Pages 2615–2623 sions are also observed in complexes 1a and 2a when compared with [Ru(bpy)3][PF6]2, the changes in their excited state lifetimes cannot be attributed solely to the energy gap law.The radiative decay rate constant, kr, can also be calculated from the emission quantum yield and the lifetime of emitting MLCT state (kr = Fem/t). This calculation is only correct if the eYciency of population of the emitting 3MLCT state is unity,14 since the observed radiative decay rate constant includes the intersystem crossing eYciency, hisc, to populate the 3MLCT states.When hisc is included, kr can be expressed as kr = Fem/ hisct or hisckr = Fem/t. All calculated data of hisckr are listed in Table 3. These hisckr values are found to be sensitive to the phosphines involved. Relatively high values of hisckr are found in complexes 1a and 1b with bppe (1.6 × 104 and 1.6 × 105 s21 respectively) when compared with much lower values in complexes 2a–2e with tppa (Table 3, with the lowest value at 2.8 × 103 s21).These low observed radiative decay constants may imply either rapid relaxation of the 1MLCT state back to the ground state or formation of another state.4c Ruthenium(II) diimine complexes have been found to have an eYcient intersystem crossing from the 1MLCT state to the 3MLCT state.4c,13b,c However, we cannot presently exclude or address the possibility of a diminished intersystem crossing eYciency in complexes with the tppa ligand. A similar observation has been reported by Schmehl and co-workers 4c in the study of mono- and bimetallic ruthenium(II) diimine complexes.Among a series of complexes, the lowest hisckr value of 6700 s21 was reported for the complex [Ru(dmb)2(bbdb)]21 while a higher hisckr value of 1.1 × 105 s21 was obtained for [Ru(dmb)3]21 [dmb = 4,49-dimethyl-2,29-bipyridine; bbdb = 1,4-bis(49-methyl-2,29- bipyridin-4-yl)buta-1,3-diene]. Upon excitation at 510–520 nm, the diVerential absorption spectra of OsII complexes with bppe and tppa are obtained and compared in Fig. 5. All spectra have a characteristic absorption peak around 375 nm (corresponding to the absorbing p æÆ p* of co-ordinated bipyridine) and bleaching of the MLCT absorption band in the region of 410–550 nm. Additional weak transient absorption bands are observed for 1b and 2b around 420–460 nm, which can be ascribed to the weakly absorbing charge-transfer bands. Several attempts have been made to detect the transient absorption signals of other systems with RuII centers but have failed to provide any observable signal.Energy transfer. In the heterobimetallic complex 2e, excitation in acetonitrile at 550 nm, where the OsII component is the dominant chromopore, results in the appearance of the lumi- Fig. 5 DiVerential absorption spectra of 1b (——, lex = 520), 2b (–––, lex = 520) and 2d (?–?–?, lex = 510 nm) in deoxygenated spectrograde MeCN at 25 8C nescence centered around 590 nm, characteristic of the emission from the OsII-based chromophore.Upon excitation at 470 nm (where the RuII-based chromophore absorbs approximately 50% of the incident photons), only emission at 590 nm is observed, also characteristic of the OsII component. Furthermore, the corrected luminescence excitation spectrum of 2e, when monitoring at the OsII emission wavelength of 590 nm, was found to have a close match with the corresponding absorption spectrum in the visible region of 400 to 570 nm. All the above results suggest the presence of energy transfer.Because of the well known energy diVerence between the lowest energy excited states of RuII and OsII polypyridine complexes, electronic energy transfer is expected to occur from the Rubased unit to the Os-based one.2,6,15–19 Following the conventional assumptions,15–19 the free energy change DG0 can be expressed as the diVerence between the spectroscopic energies of the energy donor (Ru-based) and acceptor (Os-based). The actual calculated value is ca. 1150 cm21 or 0.14 eV between Ruand Os-based units, estimated from the energy of the emission maxima of monometallic complexes with the tppa ligand. The energy transfer rate constant ken can be estimated by equation (2) 17 where tm and t represent the luminescence lifeken = 1/t 2 1/tm (2) times of the model complex and the system in question. By exciting the MLCT band at 470 nm, a single exponential decay is observed for the 3MLCT excited state of the model compound 2a, and the lifetime is found to be tm = 40 ns.For the heterobimetallic complex 2e, a long lifetime component of 410 ns is observed when measured using our nanosecond laser system (see Experimental section). This lifetime is comparable to that of the corresponding monomeric compound 2b within the range of experimental error, indicating an emission from the OsII part of 2e. In order to probe the lifetime of the quenched RuII excited state, measurement on a frequency domain fluorimeter was carried out (measurable range 50 ps–50 ns).Two additional short-lived components were revealed with t1 = 0.33 and t2 = 5.9 ns. Both components are much shorter than the 3MLCT lifetimes observed in 2a or 2c. Although it is expected to observe a quenched shorter lifetime from the RuII part of 2e as the result of intramolecular energy transfer from the Rubased unit to the OsII-based one, we are not clear about the other possible quenching mechanism that may cause the presence of two short components in the lifetime of 2e.From these two short components of lifetime, the rate constant of energy transfer in 2e can be estimated, using equation (2), to be higher than 1.4 × 108 but not exceeding 3.0 × 109 s21, assuming energy transfer is responsible for the decrease of the RuII excited-state lifetime. Mechanism of intramolecular energy transfer. In order to study the mechanism by which intramolecular energy transfer occurs in complex 2e with a tppa spacer, the spectroscopic overlap integral and the energy transfer rate constant are calculated here.11b–d,17 Previously, the energy transfer processes that take place in systems similar to 2e, but involve other types of flexible 18,20 (e.g.alkanes) or rigid spacers 11,12b,c,17 (e.g. polyphenylene), have been interpreted as occurring via a Föster- and/or Dexter-type mechanism. For energy transfer via a Föster-type dipole–dipole (i.e. through space) mechanism, the appropriate spectroscopic overlap integral (JF) can be expressed as equation (3).11b,c,21 Here JF = Ú F(n)e(n)n24dn Ú F(n)dn (3) F(n) is the luminescence intensity at wavelength n (in cm21), and e(n) is the molar absorption coeYcient (in dm3 mol21 cm21).11bJ. Chem.Soc., Dalton Trans., 1998, Pages 2615–2623 2621 The JF value thus calculated is 1.1 × 10213 cm6 mol21 (±10%) for the heterobimetallic complex 2e. The derived value here is of comparable magnitude to the estimated overlap integrals previously reported for systems with rigid alkyne-11d or phenylenebridged 17 Ru/Os complexes.The rate constant for triplet energy transfer occurring by the Föster mechanism can then be calculated using equation (4) 11b,c where K is the orientation factor kF = 8.8 × 10225 K2FLJF n4tLR6 (4) relating to the alignment of transition dipoles on donor and acceptor (K2 = 0.67 11d), FL and tL are the quantum yield and excited-state lifetime of the appropriate monomeric complex 2a, and n is the refractive index of the solvent acetonitrile. At 295 K the calculated kF value is (9 ± 1) × 107 s21 for an Ru]Os distance of ca. 9 Å.22 This calculated Föster energy-transfer rate constant is lower than the observed energy-transfer rate constant for 2e. The Dexter-type energy transfer can be described as a double exchange of electrons (i.e. via direct or superexchange-mediated electronic interaction) between donor and acceptor.1,11d,17,23 The rate constant of such an energy transfer can be expressed in the non-adiabatic limit as in equation (5).The electronic frequency ken = nen exp [2DG‡/RT] (5) nen and the free activation energy DG‡ can be evaluated according to equations (6) and (7).1,17 Here DG0 is the diVerence nen = (2Hen 2/h)(p3/lRT)� �� (6) DG‡ = (l/4)(l 1 DG0/l)2 (7) between the spectroscopic energies of donor and acceptor (ca. 1150 cm21 or 0.14 eV). The reorganization energy l can be estimated from the Stokes shift 17 of the acceptor (ca. 800 cm21 or 0.10 eV). The calculated value of exp[2DG‡/RT] is approximately 1 and, hence, ken is almost equal to nen. From equation (6) nen is estimated as (8.2 × 108)Hen 2 cm22. If the energy transfer occurs exclusively via an electron exchange mechanism, the calculated ken value would be equal to the observed one (1.4 × 108–3.0 × 109 s21), and the electronic coupling matrix element Hen can then be estimated to be 0.4–2 cm21 by the assumption of an exchange mechanism.Such a small Hen value corresponds to a situation in which the Ru and Os centers are almost in electronic isolation.11b The aforementioned electrochemical study has shown that the redox interaction across the tppa spacer with an allene bridge is rather weak. Based on all above calculations and observations, we conclude that the Dexter energy-transfer mem may be involved in the observed energy-transfer process in the heterobimetallic Ru–Os complex with the tppa spacer. Conclusion We have found that Ru and Os complexes with the polyphosphines bppe and tppa exhibit room-temperature luminescence, with long 3MLCT excited-state lifetimes of monometallic and bimetallic Os complexes.The two moieties spanned by the tppa spacer are in electronic isolation, however, the energy transfer across the same bridging ligand is found to be eYcient from the Ru-based unit to the Os-based one. We are currently exploring the preparation, electrochemical and spectroscopic properties of the luminescent and redox-active complexes containing polyphosphine spacers with longer cumulenic Cn bridges.Details regarding the eVect of various cumulenic Cn bridges on the redox interaction and energy transfer or electron transfer between two metal centers spanned by this type of rigid spacer will be investigated and reported later. Experimental General procedures All reactions pertaining to the preparation of metal– polyphosphine complexes were carried out under an N2 atmosphere and in the dark. Column chromatographic separation was performed in the dark using basic alumina (Brockman activity I, 60–325 mesh, from Fisher Scientific) and acetonitrile or toluene–acetonitrile (40 : 60, v/v) as eluent.Materials 1,19,3,39-Tetrakis(diphenylphosphino)allene (tppa),24,25 1,19- bis(diphenylphosphino)ethene (bppe) 25,26 and cis-[Os(bpy)2- Cl2] 27 were prepared according to the literature methods. The complex cis-[Ru(bpy)2Cl2] was purchased from Strem and used as received.Commercial grade solvents (acetonitrile, diethyl ether, toluene, methanol and ethanol) were dried over 4 Å molecular sieves prior to use. Tetrahydrofuran (THF) was dried and deoxygenated by heating to reflux under N2 for at least 24 h over sodium benzophenone ketyl and was freshly distilled prior to use. Commercial grade ethylene glycol was dried over 4 Å molecular sieves for at least 24 h and deoxygenated by degassing with dry N2 for 10 min or longer before it was used in reactions.Basic alumina was purchased from Fisher and directly used in chromatographic separations. All spectrograde solvents were also purchased from Fisher and used without further purification. Instrumentation The 31P-{1H} NMR spectra were obtained in CD3CN on an Omega 500 MHz spectrometer. Electron ionization and fast atom bombardment mass spectral analysis (EI-MS and FABMS) were recorded on a Fisions VG Autospec at the UCI Mass Spectral Laboratory.Absorbance spectra were recorded on a Hewlett-Packard 8453 diode array spectrophotometer. Steadystate emission spectra were obtained on an Hitachi F-4500 fluorescence spectrometer. Luminescence quantum yields of all complexes were measured in spectrograde acetonitrile relative to [Ru(bpy)3][PF6]2 (F= 0.062 14 in acetonitrile). All samples were treated with three freeze–pump–thaw cycles prior to measurements. The time-resolved emission spectroscopic studies were carried out on a nanosecond flash photolysis unit equipped with a Continuum Surelite II-10 Q-switched Nd:YAG laser and Surelite OPO (optical parametric oscillator) tunable visible source, a LeCroy 9350A oscilloscope, and a Spex 270 MIT-2x-FIX high performance scanning and imaging spectrometer.Syntheses [M(bpy)2(bppe)][PF6]2, M 5 Ru 1a or Os 1b. In a 100 mL three-necked flask (Ph2P)2C]] CH2 (155 mg, 0.39 mmol) was dissolved in 15 mL of dry THF and heated to reflux under N2.To this solution a 15 mL ethylene glycol solution of 0.13 mmol [M(bpy)2Cl2]?2H2O (M = Ru, 68 mg; M = Os, 75 mg), degassed with dry N2 for at least 10 min prior to use, was added dropwise using a pressure-equalizing funnel. The resulting mixture was refluxed for 2 h. An excess amount of NH4PF6 (200–300 mg) was added and the mixture was refluxed for up to 60 h to ensure the completion of reaction. The solution was cooled to room temperature, and the THF was removed on a rotary evaporator.The resulting ethylene glycol solution was added dropwise to 100 mL of a saturated aqueous KPF6 solution. The precipitate was collected by vacuum filtration, washed with 20 mL H2O2622 J. Chem. Soc., Dalton Trans., 1998, Pages 2615–2623 (2 ×) and 20 mL diethyl ether (3 ×), and vacuum dried overnight. The product thus obtained was chromatographically separated on basic alumina using acetonitrile as eluent to give the desired product as the second fraction.The first fraction, a minor product (5–20%) using toluene–acetonitrile (40 : 60, v/v) as eluent, was found to be [M(bpy)2Cl(bppe)]PF6. Complex 1a: Yield 135 mg (95%) (Found: C, 50.05; H, 5.03; N, 4.50. Calc. for C46H38F12N4P4Ru: C, 50.24; H, 3.48; N, 5.10%). 31P-{1H} NMR (CD3CN, 202 MHz, 22 8C): d 18.21. Complex 1b: Yield 96 mg (63%) (Found: C, 47.41; H, 3.85; N, 5.06. Calc. for C46H38F12N4OsP4: C, 46.47; H, 3.22; N, 4.71%). 31P-{1H} NMR (CD3CN, 202 MHz, 22 8C): d 223.32.[M(bpy)2(tppa)][PF6]2, M 5 Ru 2a or Os 2b. A procedure analogous to that for the preparation of 1a and 1b was followed with ligand-to-metal ratios of 2.2–3 : 1 and a refluxing period up to 60 h to ensure the completion of reaction. The product was separated on basic alumina using acetonitrile to give the desired product as the first fraction. A small amount (<20%) of the bimetallic complexes 2c or 2d could also be isolated as the second fraction using methanol as eluent.With a short refluxing period, a minor amount of the monosubstituted complex [M(bpy)2Cl(tppa)]PF6 could also be isolated as the first fraction (the yield of this product varied as the refluxing time was changed) using toluene–acetonitrile (40 : 60, v/v) as eluent, the desired compound was obtained as the second fraction using acetonitrile as eluent. Complex 2a: Yield 193 mg (80%) from 85 mg [Ru(bpy)2Cl2]?2H2O and 296 mg tppa (Found: C, 56.00; H, 4.52; N, 4.10. Calc.for C51H56F12N4P4Ru?2H2O: C, 56.25; H, 3.99; N, 3.70%). 31P-{1H} NMR (CD3CN, 202 MHz, 22 8C): d 2.34, 230.22. Complex 2b: Yield 230 mg (89%) from 100 mg [Os(bpy)2Cl2]?2H2O and 296 mg tppa (Found: C, 52.80; H, 3.94; N, 3.97. Calc. for C51H56F12N4OsP4?2H2O: C, 53.12; H, 3.77; N, 3.49%). 31P-{1H} NMR (CD3CN, 202 MHz, 22 8C): d 230.49, 237.37. [M(bpy)2(Ï-tppa)M(bpy)2][PF6]4, M 5 Ru 2c or Os 2d. These compounds could be obtained as side products in the preparation of 2a and 2b.They could also be prepared using the metal-to-ligand ratio of ca. 2.2 : 1. In a 100 mL three-necked flask a 15 mL ethylene glycol solution containing 0.164 mmol of [M(bpy)2Cl2]?2H2O was degassed with dry N2 for at least 10 min and then heated to above 120 8C. To this solution 0.075 mmol of tppa in 30 mL THF was added dropwise via a pressure-equalizing funnel. The resulting mixture was refluxed for 2 h. Excess NH4PF6 was then added and the mixture was refluxed for up to 60 h to ensure the completion of reaction.The solution was cooled to room temperature, and THF was removed on a rotary evaporator. The resulting ethylene glycol solution was added dropwise to 100 mL of saturated aqueous KPF6 solution. The precipitate was collected by vacuum filtration, washed with 20 mL H2O (2 ×) and 20 mL diethyl ether (3 ×), and vacuum dried overnight. The product was separated on a basic alumina column, using acetonitrile as eluent, to give the desired product as the second fraction.A small amount (10–20%) of the monometallic complexes 2a or 2b was obtained as the first fraction, using toluene–acetonitrile (40 : 60, v/v) as eluent. Complex 2c: Yield 61% (Found: C, 50.69; H, 3.72; N, 5.60. Calc. for C91H72F24N8P8Ru2: C, 50.06; H, 3.32; N, 5.13%). 31P-{1H} NMR (CD3CN, 202 MHz, 22 8C): d 140.87. Complex 2d: Yield 155 mg (85%) (Found: C, 46.45; H, 3.55; N, 4.98. Calc. for C71H72F24N8Os2P8: C, 46.28; H, 3.07; N, 4.75%). 31P-{1H} NMR (CD3CN, 202 MHz, 22 8C): d 83.64.[Ru(bpy)2(Ï-tppa)Os(bpy)2][PF6]4 2e. In a 100 mL threenecked flask 62 mg, 0.0384 mmol of complex 2b and 20 mg, 0.0384 mmol of [Ru(bpy)2Cl2]?2H2O were added together and purged with N2, 20 mL of degassed ethylene glycol was added and the resulting solution was heated at refluxing temperature for 6 h. The solution was then cooled to room temperature, and added dropwise to a 100 mL saturated aqueous solution of KPF6. The brown precipitate thus formed was collected using vacuum filtration, washed with 20 mL water (2 ×) and 20 mL diethyl ether (3 ×), and vacuum dried overnight.The product was separated on basic alumina using acetonitrile as eluent, and the desired product was collected as the second fraction. Yield 65 mg (74%) (Found: C, 46.93; H, 4.47; N, 4.83. Calc. for C71H72F24N8OsP8Ru?4H2O: C, 46.62; H, 3.44; N, 4.78%). 31P-{1H} NMR (CD3CN, 202 MHz, 22 8C): d 90.37, 231.14. Characterization of complexes by fast atom bombardment mass spectrometry In addition to elemental analysis and 31P-{1H} NMR spectral analysis, all the above complexes have also been studied using FAB-MS. 3-Nitrobenzyl alcohol was used as the matrix. A compilation of the peaks of higher mass in the spectrum of each complex, along with their assignment, is presented in Table 1. Electrochemical analysis Cyclic voltammetry measurements were performed in a standard three-electrode cell. A Ag–AgCl wire was used as a pseudoreference electrode, a platinum wire as the counter-electrode, and a 1.0 mm platinum disk electrode as the working electrode.A solution of the purified electrolyte in acetonitrile (0.1 mol dm23) was first scanned to ensure the absence of air, water and other impurities. Cyclic voltammograms were recorded with a CHI 630 electrochemical analyzer, with a scan rate of 100 mV s21. All experiments were referenced, after all scans had been taken, against an added ferrocene standard. No iR compensation was applied.Acknowledgements This work was supported by the UCI Startup Funds, the Faculty Research Grant from UCI Academic Senate Committee on Research, and National Science Foundation CAREER award (CHE-9733546). S. K. S. acknowledges financial support from the Undergraduate Research Opportunity Program at UCI. J. V. O. acknowledges support from the Graduate and Professional Opportunity Program (GPOP) from the US Department of Education. We thank Dr. Torbjoern Pascher from the California Institute of Technology and Dr.Wytze van der Veer at the UCI Laser Facility for their help in the set-up of a nanosecond laser system for the time-resolved emission study. We also thank Professor Wayne E. Jones, Jr. and his graduate student CliVord Murphy for the measurement of short lifetimes of compound 2e, and Dr. John Greaves at the UCI Mass Spectral Laboratory for his assistance in the FAB-MS analysis. References 1 J.-P. Sauvage, J.-P. Colin, J.-C.Chambron, S. Guillercz, C. Cudret, V. Balzani, F. Barigelletti, L. De Cola and L. Flamigni, Chem. Rev,, 1994, 94, 993; V. Balzani, A. Juris, M. Venturi, S. Campagna and S. Serroni, Chem. Rev., 1996, 96, 759. 2 M. Brady, W. Weng, Y. Zhou, J. W. Seyler, A. J. Amoroso, A. M. Arif, M. Böhme, G. Frenking and J. A. Gladysz, J. Am. Chem. Soc., 1997, 119, 775; K. Wärnmark, J. A. Thomas, O. Heyke and J.-M. Lehn, Chem. Commun., 1996, 701. 3 T. J. Meyer, Pure Appl. Chem., 1986, 58, 1193; A.Harriman and R. Ziessel, Chem. Commun., 1996, 1707. 4 (a) F. Scandola, R. Argazzi, C. A. Bignozzi, C. Chiorboli, M. T. Indelli and M. A. Rampi, Coord. Chem. Rev., 1993, 125, 283; (b) P. Belser, R. Dux, M. Baak, L. De Cola and V. Balzani, Angew. Chem., Int. Ed. Engl., 1995, 34, 595; (c) A. I. Baba, H. E. Ensley and R. H. Schmehl, Inorg. Chem., 1995, 34, 1198. 5 B.Hong and J. V. Ortega, Angew. Chem., Int. Ed. Engl., 1998, in the press; B. Hong, Comments Inorg. Chem., 1998, in the press; F.A. Cotton and B. Hong, Prog. Inorg. Chem., 1992, 40, 179. 6 F. A. Cotton, B. Hong, M. Shang and G. G. Stanley, Inorg. Chem., 1993, 32, 3620.J. Chem. Soc., Dalton Trans., 1998, Pages 2615–2623 2623 7 V. Balzani and F. Scandola, Supramolecular Photochemistry, Horwood, Chichester, 1991. 8 (a) E. M. Kober, J. V. Caspar, B. P. Sullivan and T. J. Meyer, Inorg. Chem., 1988, 27, 4587; (b) S. R. Johnson, T. D. Westmoreland, J. V. Caspar, K. R. Barqawi and T.J. Meyer, Inorg. Chem., 1988, 27, 3195; (c) E. M. Kober, B. P. Sullivan, W. J. Dressick, J. V. Caspar and T. J. Meyer, J. Am. Chem. Soc., 1980, 102, 7383; (d ) J. V. Caspar, E. M. Kober, B. P. Sullivan and T. J. Meyer, J. Am. Chem. Soc., 1982, 104, 630. 9 (a) R. G. Brewer, G. E. Jensen and K. J. Brewer, Inorg. Chem., 1994, 33, 124; (b) M. M. Richter and K. J. Brewer, Inorg. Chem., 1993, 32, 5762; (c) M. M. Richter and K. J. Brewer, Inorg. Chem., 1993, 32, 2827. 10 P.-W. Wang and M.A. Fox, Inorg. Chem., 1995, 34, 36. 11 (a) C. R. Arana and H. D. Abruna, Inorg. Chem., 1993, 32, 194; (b) V. Grosshenny, A. Harriman, M. Hissler and R. Ziessel, J. Chem. Soc., Faraday Trans., 1996, 92, 2223; (c) V. Grosshenny, A. Harriman, F. M. Romero and R. Ziessel, J. Phys. Chem., 1996, 100, 17 472; (d ) V. Grosshenny, A. Harriman, M. Hissler and R. Ziessel, Angew. Chem., Int. Ed. Engl., 1995, 34, 1100. 12 (a) C. Creutz, M. Chou, L. Netzel, M. Okumura and N. Sutin, J.Am. Chem. Soc., 1980, 102, 1309; (b) F. Barigelletti, A. Juris, V. Balzani, P. Belser and A. von Zelewsky, Inorg. Chem., 1983, 22, 3335; (c) A. Juris, P. Belser, F. Barigelletti, A. von Zelewsky and V. Balzani, Inorg. Chem., 1986, 25, 256; (d ) J. R. Winkler, T. L. Netzel, C. Creutz and N. Sutin, J. Am. Chem. Soc., 1987, 109, 2381; (e) Y. Fuchs, S. Lofters, T. Dieter, W. Shi, R. Morgan, T. C. Strekas, H. D. Gafney and A. D. Baker, J. Am. Chem. Soc., 1987, 109, 2691. 13 (a) J.V. Caspar, T. D. Westmoreland, G. H. Allen, P. G. Bradley, T. J. Meyer and W. H. WooddruV, J. Am. Chem. Soc., 1984, 106, 3492; (b) J. N. Demas and D. G. Taylor, Inorg. Chem., 1979, 18, 3177; (c) K. Mandal, T. D. L. Pearson, W. P. Krug and J. N. Demas, J. Am. Chem. Soc., 1983, 105, 701. 14 (a) J. V. Caspar and T. J. Meyer, J. Am. Chem. Soc., 1983, 105, 5583; (b) J. V. Caspar, B. P. Sullivan and T. J. Meyer, Inorg. Chem., 1984, 23, 2104. 15 V. Balzani, F. Bolletta and F. Scandola, J.Am. Chem. Soc., 1980, 102, 2552. 16 F. Scandola and V. Balzani, J. Chem. Educ., 1983, 60, 814. 17 F. Barigelletti, L. Flamigni, V. Balzani, J.-P. Collin, J.-P. Sauvage, A. Sour, E. C. Constable and A. M. W. C. Thompson, J. Am. Chem. Soc., 1994, 116, 7692. 18 C. K. Ryu and R. H. Schmehl, J. Phys. Chem., 1989, 93, 7961; L. De Cola, V. Balzani, F. Barigelletti, L. Flamigni, P. Belser, A. von Zelewsky, M. Frank and F. Vögtle, Inorg. Chem., 1993, 32, 5228. 19 N. Sutin, Acc.Chem. Res., 1982, 15, 275. 20 M. Furue, T. Yoshidzumi, S. Kinoshita, T. Kushida, S. Nozakura and M. Kamachi, Bull. Chem. Soc. Jpn., 1991, 64, 1632; R. H. Schmehl, R. A. Auerbach and W. F. Walcholtz, J. Phys. Chem., 1988, 92, 6202; R. H. Schmehl, R. A. Auerbach, W. F. Walcholtz and C. M. Elliott, Inorg. Chem., 1986, 25, 2440. 21 T. Föster, Discuss. Faraday Soc., 1959, 27, 7. 22 This distance, the midpoint between the Ru and Os centers, in 2e is estimated using the space filling model in Insight/Discover of Biosym/MSI. 23 D. L. Dexter, J. Chem. Phys., 1953, 21, 836. 24 H. Schmidbaur and T. Pollok, Angew. Chem., Int. Ed. Engl., 1986, 25, 348. 25 H. Schmidbaur, S. Manhart and A. Schier, Chem. Ber., 1993, 126, 2389. 26 I. J. Colquhoun and W. McFarlane, J. Chem. Soc., Dalton Trans., 1982, 1915. 27 D. A. Buckingham, F. P. Dwyer, H. A. Goodwin and A. M. Sargeson, Aust. J. Chem., 1964, 17, 325; E. M. Kober, J. V. Casper, B. P. Sullivan and T. J. Meyer, Inorg.Chem., 1988, 27, 4587. Received 6th April 1998; Paper 8/02809EJ. Chem. Soc., Dalton Trans., 1998, Pages 2615–2623 2623 7 V. Balzani and F. Scandola, Supramolecular Photochemistry, Horwood, Chichester, 1991. 8 (a) E. M. Kober, J. V. Caspar, B. P. Sullivan and T. J. Meyer, Inorg. Chem., 1988, 27, 4587; (b) S. R. Johnson, T. D. Westmoreland, J. V. Caspar, K. R. Barqawi and T. J. Meyer, Inorg. Chem., 1988, 27, 3195; (c) E. M. Kober, B. P. Sullivan, W. J. Dressick, J. V. Caspar and T.J. Meyer, J. Am. Chem. Soc., 1980, 102, 7383; (d ) J. V. Caspar, E. M. Kober, B. P. Sullivan and T. J. Meyer, J. Am. Chem. Soc., 1982, 104, 630. 9 (a) R. G. Brewer, G. E. Jensen and K. J. Brewer, Inorg. Chem., 1994, 33, 124; (b) M. M. Richter and K. J. Brewer, Inorg. Chem., 1993, 32, 5762; (c) M. M. Richter and K. J. Brewer, Inorg. Chem., 1993, 32, 2827. 10 P.-W. Wang and M. A. Fox, Inorg. Chem., 1995, 34, 36. 11 (a) C. R. Arana and H. D. Abruna, Inorg. Chem., 1993, 32, 194; (b) V.Grosshenny, A. Harriman, M. Hissler and R. Ziessel, J. Chem. Soc., Faraday Trans., 1996, 92, 2223; (c) V. Grosshenny, A. Harriman, F. M. Romero and R. Ziessel, J. Phys. Chem., 1996, 100, 17 472; (d ) V. Grosshenny, A. Harriman, M. Hissler and R. Ziessel, Angew. Chem., Int. Ed. Engl., 1995, 34, 1100. 12 (a) C. Creutz, M. Chou, L. Netzel, M. Okumura and N. Sutin, J. Am. Chem. Soc., 1980, 102, 1309; (b) F. Barigelletti, A. Juris, V. Balzani, P. Belser and A. von Zelewsky, Inorg. Chem., 1983, 22, 3335; (c) A. Juris, P. Belser, F. Barigelletti, A. von Zelewsky and V. Balzani, Inorg. Chem., 1986, 25, 256; (d ) J. R. Winkler, T. L. Netzel, C. Creutz and N. Sutin, J. Am. Chem. Soc., 1987, 109, 2381; (e) Y. Fuchs, S. Lofters, T. Dieter, W. Shi, R. Morgan, T. C. Strekas, H. D. Gafney and A. D. Baker, J. Am. Chem. Soc., 1987, 109, 2691. 13 (a) J. V. Caspar, T. D. Westmoreland, G. H. Allen, P. G. Bradley, T. J. Meyer and W. H. WooddruV, J. Am. Chem. Soc., 1984, 106, 3492; (b) J. N. Demas and D. G. Taylor, Inorg. Chem., 1979, 18, 3177; (c) K. Mandal, T. D. L. Pearson, W. P. Krug and J. N. Demas, J. Am. Chem. Soc., 1983, 105, 701. 14 (a) J. V. Caspar and T. J. Meyer, J. Am. Chem. Soc., 1983, 105, 5583; (b) J. V. Caspar, B. P. Sullivan and T. J. Meyer, Inorg. Chem., 1984, 23, 2104. 15 V. Balzani, F. Bolletta and F. Scandola, J. Am. Chem. Soc., 1980, 102, 2552. 16 F. Scandola and V. Balzani, J. Chem. Educ., 1983, 60, 814. 17 F. Barigelletti, L. Flamigni, V. Balzani, J.-P. Collin, J.-P. Sauvage, A. Sour, E. C. Constable and A. M. W. C. Thompson, J. Am. Chem. Soc., 1994, 116, 7692. 18 C. K. Ryu and R. H. Schmehl, J. Phys. Chem., 1989, 93, 7961; L. De Cola, V. Balzani, F. Barigelletti, L. Flamigni, P. Belser, A. von Zelewsky, M. Frank and F. Vögtle, Inorg. Chem., 1993, 32, 5228. 19 N. Sutin, Acc. Chem. Res., 1982, 15, 275. 20 M. Furue, T. Yoshidzumi, S. Kinoshita, T. Kushida, S. Nozakura and M. Kamachi, Bull. Chem. Soc. Jpn., 1991, 64, 1632; R. H. Schmehl, R. A. Auerbach and W. F. Walcholtz, J. Phys. Chem., 1988, 92, 6202; R. H. Schmehl, R. A. Auerbach, W. F. Walcholtz and C. M. Elliott, Inorg. Chem., 1986, 25, 2440. 21 T. Föster, Discuss. Faraday Soc., 1959, 27, 7. 22 This distance, the midpoint between the Ru and Os centers, in 2e is estimated using the space filling model in Insight/Discover of Biosym/MSI. 23 D. L. Dexter, J. Chem. Phys., 1953, 21, 836. 24 H. Schmidbaur and T. Pollok, Angew. Chem., Int. Ed. Engl., 1986, 25, 348. 25 H. Schmidbaur, S. Manhart and A. Schier, Chem. Ber., 1993, 126, 2389. 26 I. J. Colquhoun and W. McFarlane, J. Chem. Soc., Dalton Trans., 1982, 1915. 27 D. A. Buckingham, F. P. Dwyer, H. A. Goodwin and A. M. Sargeson, Aust. J. Chem., 1964, 17, 325; E. M. Kober, J. V. Casper, B. P. Sullivan and T. J. Meyer, Inorg. Chem., 1988, 27, 4587. Received 6th April 1998; Paper 8/02809E

ISSN:1477-9226    

DOI:10.1039/a802809e

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 2625.2633 2625 Syntheses, crystal structures and (spectro)electrochemical studies of novel clusters [Ru4(I-H)4(CO)10(L)] [L 5 2,29-bipyrimidine, 2,3-bis(pyridin-2-yl)pyrazine, 2,29-bipyridine] ¢Ó Jos NijhoV,a Maarten J. Bakker,a Frantis¢§ek Hartl,*,a Gideon Freeman,b Scott L. Ingham b and Brian F. G. Johnson *,c a Anorganisch Chemisch Laboratorium, Institute of Molecular Chemistry, Universiteit van Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, The Netherlands b Department of Chemistry, University of Edinburgh, West Mains Road, Edinburgh, UK EH9 3JJ c University Chemical Laboratory, University of Cambridge, Lensfield Road, Cambridge, UK CB2 1EW Three tetrahedral clusters [Ru4(m-H)4(CO)10(L)] [L = 2,29-bipyrimidine 1, 2,3-bis(pyridin-2-yl)pyrazine 2 and 2,29-bipyridine 3] were obtained from a substitution reaction of [Ru4(m-H)4(CO)12] with L, involving the monosubstituted intermediates [Ru4(m-H)4(CO)11(h1-L)].The solid-state structures of 1 and 3 have been elucidated by single-crystal X-ray analysis. In solution, dynamic behaviour of the hydride ligands in 1.3 was apparent from 1H NMR spectra. UV/VIS and resonance-Raman spectroscopy established a significant Ru(dp)�¡L(p*) charge transfer (MLCT) character of the lowest electronic transition of 1.3 in the visible region. Consistent with the dominantly L(p*)-localized LUMO, one-electron reduction of 1 and 2 produced corresponding radical anions 1b and 2b which could be characterized by IR, UV/VIS and EPR spectroscopy.Subsequent one-electron reduction of 1b and 2b yielded unstable dianions [Ru4(m-H)4(CO)10(L)]22 which were found to eliminate H2. The dihydrido dianionic products 1c, 2c were also formed via slow disproportionation of 1b and 2b. The radical anion 3b, containing the stronger s-donor 2,29-bipyridine anion, was detectable only on the subsecond time-scale of cyclic voltammetry.The electrochemically produced dihydrido dianion 3c is similar but diVerent from [Ru4(m-H)2(CO)10(bpy)]22 3c9 obtained by deprotonating 3 in reaction with NEt4OH. Tetranuclear hydridoruthenium carbonyl clusters have received considerable attention in homogeneous catalysis, in particular for their hydrogenation potential.1 Promising research opportunities exist in the field of their selective photo- and electrochemical activation under substantially milder conditions of temperature and pressure.The particularly well known example is hydrogenation of ethylene photoinduced by [Ru4- (m-H)4(CO)12] which proceeds without cluster fragmentation.2 Elimination of hydrogen from the latter cluster also occurs on its electrochemical reduction. The intimate mechanism of this process has recently been reported by Osella et al.3 Another approach to redox activation of low-nuclearity transition-metal clusters involves co-ordination of redox active ligands. Their protecting role as electron reservoirs oVers the possibility of (spectro)electrochemical characterization of otherwise short-lived reactive transients, in particular radicals, generated along the reduction path.This strategy has recently been applied for the clusters [Os3(CO)10(a-diimine)]. The introduction of the reducible a-diimine ligand permitted detailed investigation of the primary one-electron cathodic step followed by temperature-controlled splitting of an Os]Os- (a-diimine) bond in [Os3(CO)10(a-diimine)]~2 and concomitant electron transfer producing a reactive open-structure dianionic species.4 In consequence of the aforementioned investigations, we ¢Ó Supplementary data available: interpretation of the 1H NMR decoupling experiments carried out on complex 2 to assign the proton resonances of the dpp ligand: oV resonance irradiation at d 11.00, and irradiation at d 8.89, 8.67 and 7.34.For direct electronic access see http://www.rsc.org/suppdata/dt/1998/2625/, otherwise available from BLDSC (No.SUP 57401, 5 pp.) or the RSC Library. See Instructions for Authors, 1998, Issue 1 (http://www.rsc.org/dalton). Non-SI unit employed: G = 1024 T. became interested in the synthesis and electrochemistry of a-diimine-substituted derivatives of [Ru4(m-H)4(CO)12]. This cluster is known5,6 to undergo an eYcient thermal substitution of carbonyl ligands by tertiary phosphines, producing [Ru4- (m-H)4(CO)12 2 n(PR3)n] (n = 1.4), where n depends on the reaction conditions and the amount of phosphine added.A mixture of [Ru4(m-H)4(CO)12 2 n(PPh3)n] (n = 1.3) was also obtained under mild conditions via electrocatalytic (ETC) substitution reactions.3 In this paper we report on the targeted syntheses and characterization of the disubstituted derivatives [Ru4(m-H)4(CO)10- (L)] [L = 2,29-bipyrimidine (bpym), 2,3-bis(pyridin-2-yl)pyrazine (dpp) and 2,29-bipyridine (bpy)], with bpy as the strongest s-donor and bpym as the strongest p-acceptor in the series.Next, we have performed a (spectro)electrochemical study of their redox properties and reduction paths, aimed at comparison with the unsubstituted precursor [Ru4(m-H)4(CO)12] and with the related clusters [Os3(CO)10(a-diimine)]. The ligands L and the adopted numbering schemes are shown in Fig. 1. Experimental Materials The cluster [Ru4(m-H)4(CO)12] was synthesized from [Ru3- (CO)12] in a magnetically stirred BurghoV (250 ml) autoclave according to the literature method.7 2,29-Bipyrimidine (Alfa and Lancaster Chemicals), 2,3-bis(pyridin-2-yl)pyrazine and 2,29-bipyridine (both Aldrich) were used as purchased.Trimethylamine N-oxide dihydrate, Me3NO?2H2O (Alfa), was carefully dried by first refluxing the sample (15 g) in benzene (250 ml), typically overnight, to remove the water of crystallisation via Dean.Stark distillation. The benzene was then decanted oV2626 J. Chem. Soc., Dalton Trans., 1998, Pages 2625–2633 and the sample dried under vacuum on a Schlenk line, and sublimed prior to use.The supporting electrolyte NBu4PF6 (Aldrich) was dried in vacuo at 80 8C for 10 h. The salt NEt4OH was purchased from Fluka as a 25% solution in MeOH. Ferrocene (Fc) was used as received from BDH. Solvents (all Acros Chimica, analytical grade) were dried using a benzophenone– sodium mixture (THF, benzene), sodium wire (hexane), P2O5 (CH2Cl2) or CaH2 (MeCN), and freshly distilled under nitrogen.Silica gel (Kieselgel 60, Merck, 70–230 mesh) for column chromatography was dried and activated by heating overnight in vacuo at 160 8C. Preparative TLC was carried out using glass plates (20 × 20 cm) supplied by Merck, coated with a 0.25 cm layer of Kieselgel 60 F254. Syntheses of [Ru4(Ï-H)4(CO)10(L)] (L 5 bpym 1, dpp 2 or bpy 3) All clusters 1–3 were prepared according to the following general procedure. A solution of 200 mg [Ru4(m-H)4(CO)12] and 2.5 equivalents of L (L = bpym, dpp or bpy) in dichloromethane (150 ml) was cooled to 195 K in a dry ice–acetone bath.The reaction was initiated by the dropwise addition of 2.2 equivalents of Me3NO in dichloromethane (20 ml) over a 20 min period under stirring. After stirring for an additional 10 min the mixture was allowed to warm gradually to room temperature and was observed to change colour from yellow-orange to green (for L = bpym or bpy) and finally to deep red-brown. The solvent was evaporated and the crude product was separated from unreacted [Ru4(m-H)4(CO)12] by using either TLC with hexane–dichloromethane (3 : 7 v/v, for L = bpy or dpp) or 100% dichloromethane (for L = bpym) as eluent, or by column chromatography by gradual elution with hexane–dichloromethane mixtures (initially 1 : 9 v/v).All preparations and purifications were performed under a nitrogen atmosphere using Schlenk techniques. The purity of the cluster 1–3 was checked with elemental analysis or mass spectroscopy.All compounds were characterized by IR, 1H NMR (see Fig. 1 for the adopted numbering schemes) and UV/VIS spectroscopy. [Ru4(Ï-H)4(CO)10(bpym)] 1. IR [n(CO)/cm21; THF]: 2073m, 2042s, 2018s, 2000m, 1979m, 1945w. 1H NMR (CDCl3): d 9.16 (dd, J = 2.17, 4.74, 2 H, H2, H8), 9.07 (dd, J = 2.15, 5.66, 2 H, H4, H6), 7.62 (dd, J = 4.80, 5.61 Hz, 2 H, H3, H7), 215.9 (s, 2 H), 221.5s, 2 H). UV/VIS [lmax/nm (emax/M21 cm21); CH2Cl2]: 495 (3350), 359 (13 700) [Found (calc.for C18H10N4O10Ru4): C, 25.68 (25.66); H, 1.37 (1.20); N, 6.18 (6.65%)]. [Ru4(Ï-H)4(CO)10(dpp)] 2. IR [n(CO)/cm21; THF]: 2073m, 2041s, 2018vs, 2000m, 1979m, 1946w. 1H NMR (CDCl3): d 8.92 (d, J = 2.96, 1 H, H7), 8.89 (d, 1 H, H2), 8.67 (d, J = 2.9, 2 H, H6, H11), 8.04 (td, J = 1.75, 7.71, 1 H, H9), 7.94 (ddd, J = 1.11, 1.21, 7.90, 1 H, H8), 7.56 (m, 2 H, H4, H10), 7.34 (ddd, J = 1.45, 5.45, 7.62, 1 H, H3), 7.04 (d, J = 8.37 Hz, 1 H, H5), 215.8 (vw, br), 221.3 (w br).UV/VIS [lmax/nm (emax/M21 cm21); CH2Cl2]: 513 Fig. 1 Structures of the ligands L and adopted numbering schemes for clusters 1–3 N N H2 H3 H4 N N H8 H7 H6 Ru N H2 H3 H4 H5 N N H7 H6 N H8 H9 H10 H11 Ru N H2 H3 H4 H5 N H11 H10 H9 H8 Ru 1 L = bpym 3 L = bpy 2 L = dpp (3640), 357 (15 500) and 276 (26 700). Mass (FAB1): (m/z)1 925 (922 calc.) [M]1 [Found (calc. for C24H14N4O10Ru4): C, 31.12 (31.24); H, 1.62 (1.53); N, 5.91 (6.07%)]. [Ru4(Ï-H)4(CO)10(bpy)] 3. IR [n(CO)/cm21; THF]: 2071m, 2039vs, 2015vs, 1996s, 1974m, 1947sh. 1H NMR (CDCl3): d 8.92 (d, J = 5.78, 2 H, H2, H11), 8.11 (d, J = 6.6, 2 H, H5, H8), 7.96 (dt, J = 7.52 and 1.55, 2 H, H4, H9), 7.44 (ddd, J = 7.49, 5.61, 1.49 Hz, H3, H10), 215.68 (s, 2 H), 221.45 (s, 2 H). UV/ VIS [lmax/nm (emax/M21 cm21); CH2Cl2]: 476 (3210), 357 (12 100) and 301 (32 600). Mass (FAB1): (m/z)1 845 (844 calc.) [M]1 [Found (calc. for C20H12N2O10Ru4): C, 28.27 (28.44); H, 1.49 (1.43); N, 3.14 (3.32%)]. Attempted synthesis of [Ru4(Ï-H)4(CO)8(bpym)2] A solution of 155 mg [Ru4(m-H)4(CO)12] in 150 ml CH2Cl2 was cooled to 278 8C before addition of 2,29-bipyrimidine (91 mg, 2.8 equivalents, dissolved in 10 ml CH2Cl2). Trimethylamine N-oxide (66 mg, 4.2 equivalents, in 25 ml CH2Cl2) was then added dropwise to the reaction mixture over a 20 min period.The solution turned from dark green to deep red on gradual warming to room temperature. A large amount of unassigned insoluble material was formed in addition to a red product which was purified by TLC (100% CH2Cl2 as eluent) and identified by IR and mass spectroscopy as 1 and not [Ru4(m-H)4- (CO)8(bpym)2].X-Ray crystallography Crystals of the clusters 1 and 3 suitable for single-crystal X-ray analysis were grown from a dichloromethane–hexane solution at 220 8C (1) or 5 8C (3). Despite repeated attempts crystals of 2 have not been obtained. Data collection was performed on a Stoë Stadi four-circle diVractometer, equipped with an Oxford Cryosystems low-temperature device.The intensities were reduced to Fo 2 and an empirical absorption correction based on semiempirical y-scan data was applied. The structures were solved by direct methods, followed by Fourier-diVerence and full-matrix least-squares refinements on F 2 using the computer programs SHELXS 86 and SHELXL 93 and Figs. 2 and 3 generated using ORTEP.8 Crystal data for [Ru4(Ï-H)4(CO)10(bpym)] 1. C18H10N4O10Ru4? CH2Cl2, M = 929.49, monoclinic, a = 8.474(2), b = 18.897(4), c = 18.400(5) Å, b = 92.15(2)8, U = 2944.4(1) Å3, T = 150(2) K, space group P21/c, Z = 4, m(Mo-Ka) = 2.248 mm21, 4549 reflections collected, 3842 reflections independent (Rint = 0.0208) which were used in all calculations.The final wR(F 2) was 0.1024; R1 = 0.0431. Crystal data for [Ru4(Ï-H)4(CO)10(bpy)] 3. C20H12N2O10Ru4? 0.25 CH2Cl2, M = 865.83, triclinic, a = 8.6204(10), b = 9.9099(12), c = 16.431(2) Å, a = 73.625(10), b = 84.345(10), g = 88.371(9)8, U = 1340.1(3) Å3, T = 293(2) K, space group P1� , Z = 2, m(Mo-Ka) = 2.314 mm21, 4728 reflections collected, 4728 reflections independent which were used in all calculations.The final wR(F2) was 0.0994; R1 = 0.0347. CCDC reference number 186/1056. Spectroscopic measurements The UV/VIS absorption spectra were recorded on softwareupdated Perkin-Elmer Lambda 5 or Varian Cary 4E spectrophotometers, FTIR spectra on Bio-Rad FTS-7 or Perkin-Elmer 1600 spectrometers, and 1H NMR spectra on Bruker WH 250 MHz or AMX 300 spectrometers. X-Band EPR spectra were recorded on a Varian Century E-104A spectrometer. 2,29- Diphenyl-1-picrylhydrazyl (DPPH) was employed as an external ‘g mark’. Mass spectra (FAB1) were measured on a Kratos MS50TC spectrometer, calibrated with CsI.TheJ. Chem. Soc., Dalton Trans., 1998, Pages 2625–2633 2627 samples were run in a matrix solution (m-nitrobenzyl alcohol, 3-NOBA). Resonance-Raman measurements were performed using a Dilor Modular XY spectrometer which employs a backscattering geometry and a multichannel diode array detection system.Spectra were taken from KNO3 pellets at room temperature. Excitation lines of wavelengths 457.9, 476.5, 496.5 and 514.4 nm were obtained from an SP model 2016 Ar1 laser. (Spectro)electrochemical measurements Cyclic voltammograms were recorded in a gastight cell under a nitrogen atmosphere. The cell was equipped with Pt disc working (apparent surface area of 0.42 mm2), Pt gauze auxiliary, and Ag wire pseudo-reference electrodes.The working electrode was carefully polished with a 0.25 mm grain diamond paste. The scan rate was varied between 0.02 and 2 V s21. All redox potentials are reported against the ferrocene–ferrocenium (Fc/Fc1) redox couple used as an internal standard9 [E2� 1 (Fc/Fc1) = 10.58 V vs. SCE in THF]. The solutions for cyclic voltammetric experiments were typically 2 × 1023 M in the cluster compounds and 3 × 1021 M in NBu4PF6. The potential control was achieved with a PAR model 283 potentiostat equipped with positive feedback for ohmic-drop compensation.Bulk electrolyses were carried out in a gastight cell that consisted of three chambers separated at the bottom by S4 frits, with a Pt-flag working electrode (120 mm2 surface) in the middle, and Ag wire pseudo-reference and Pt gauze auxiliary electrodes in the lateral chambers. The concentrations used were 5 × 1023 M 1 and 2 and 3 × 1021 M NBu4PF6.Infrared and UV/VIS spectroelectrochemical experiments at room temperature were performed with an optically transparent thin layer electrode (OTTLE) cell,10 equipped with a Pt minigrid working electrode (32 wires per cm) and CaF2 optical windows. For IR and UV/VIS spectroelectrochemistry at low temperatures another OTTLE cell 11 was employed which fitted into a liquid-nitrogen cryostat. 12 The spectroelectrochemical samples were typically 1022 M in the cluster compounds.A PA4 potentiostat (EKOM, Czech Republic) was used to carry out the controlled-potential electrolyses. Results and Discussion Syntheses of [Ru4(Ï-H)4(CO)10(L)] In contrast to the availability of [Ru4(m-H)4(CO)12 2 n(PR3)n] (n = 1–4),5,6 the substitution reaction of [Ru4(m-H)4(CO)12] with an excess of the a-diimine ligand (L), initiated by an addition of 2.2 equivalents of Me3NO, resulted in the overall substitution of two carbonyl ligands, producing stable [Ru4(m-H)4(CO)10(L)] (L = bpym 1, dpp 2, bpy 3) in 10–20% yield. Attempts to synthesize the tetrasubstituted clusters [Ru4(m-H)4(CO)8(L)2] using 4.2 equivalents of Me3NO were unsuccessful.A similar situation applies for related [Os3(CO)12] which only aVords the disubstituted cluster [Os3(CO)10(L)].4,13 Irrespective of whether the reaction of [Ru4(m-H)4(CO)12] with L was carried out using 1.1 or 2.2 equivalents of Me3NO, only two carbonyl ligands were removed and substituted by the a-diimine.Addition of 2.2 equivalents of Me3NO only increased the yield of [Ru4(m-H)4(CO)10(L)]. For L = bpy or bpym, however, a green intermediate was observed during the reaction course at low temperatures. For L = bpym, this species was stable enough at T < 195 K to permit recording of its IR spectrum which showed n(CO) bands at 2080m, 2066s, 2053s, 2029m, 2022m and 2005w cm21 (in CH2Cl2). The spectrum is virtually identical to that of [Ru4(m-H)4(CO)11(py)] [n(CO) at 2087vw (sh), 2080m, 2066s, 2056m, 2029m, 2022m (sh), 2006w, 1989vw (br) cm21] 14 and closely resembles that of [Ru4- (m-H)4(CO)11{P(OMe)3}].15 The green intermediate is therefore most likely the monosubstituted cluster [Ru4(m-H)4(CO)11- (h1-L)] where L co-ordinates in a monodentate fashion.We therefore cude that the formation of [Ru4(m-H)4(CO)10(L)] is a stepwise process, as depicted in Scheme 1. It is probably the chelate eVect which favours this substitution pattern.The COextrusion reaction of the h1-bound a-diimine was discussed in detail by Lees and co-workers for [M(CO)n(a-diimine)] (n = 4 or 5; M = Cr, Mo or W).16 Solid-state structure of [Ru4(Ï-H)4(CO)10(L)] (L 5 bpym 1 or bpy 3) The isotropic solid-state structures of 1 and 3 are depicted in Fig. 2A and 2B, respectively. Selected bond lengths and angles of 1 and 3 are listed in Tables 1 and 2, respectively. The crystal structures of 1 and 3 confirm that the a-diimine ligand is bound in a bidentate fashion through the nitrogen lone pairs to a single ruthenium atom bearing only one terminal CO group.All ten carbonyl ligands are terminal, as could also be inferred from the IR spectra (see below). The bite angle N]Ru]N is 77.48 for both 1 and 3. Notably, the Ru]N bond lengths of 2.099(8) and 2.117(8) Å in 1 are considerably longer than those in 3 [2.095(5) and 2.090(5) Å]. This diVerence reflects the electron-withdrawing character of the unco-ordinated nitrogen atoms of the 2,29-bipyrimidine ligand, responsible for reduced NÆRu s-donation which results in weaker Ru]N bonds in 1.The apical bonds Ru(2)]C(21) in 1 and Ru(3)]C(31) in 3 are shorter than the other Ru]CO bonds in these clusters due to the increased electron density and hence stronger RuÆCO p-back bonding at the a-diimine-substituted site. The Ru4 core is edge-bridged by four hydride ligands, three of them spanning the Ru]Ru (L) bonds. The hydride positions, localized from Fourier-diVerence maps of the low-angle diVraction data, are also indicated by distortion of the tetrahedral cluster geometry due to lengthening of the four hydride- Scheme 1 Stepwise formation of the disubstituted clusters [Ru4(m-H)4- (CO)10(L)] Ru Ru Ru Ru C C C Ru Ru Ru Ru C N C Ru Ru Ru Ru N N C O N O O O O O + L –CO –CO N N = L = bpy, bpym or dpp Table 1 Selected bond lengths (Å) and angles (8) for [Ru4(m-H)4- (CO)10(bpym)] 1 Ru(1)]Ru(2) Ru(1)]Ru(3) Ru(1)]Ru(4) Ru(2)]Ru(3) Ru(2)]Ru(4) Ru(3)]Ru(4) Ru(2)]C(21) Ru(2)]N(1b) Ru(2)]N(4b) Ru(2)]H(1) 2.944(12) 2.787(13) 2.760(11) 3.035(12) 2.949(13) 2.931(12) 1.834(11) 2.099(8) 2.117(8) 1.74(3) Ru(2)]H(2) Ru(2)]H(3) Ru(3)]H(3) Ru(3)]H(4) Ru(4)]H(2) Ru(4)]H(4) mean Ru]C mean C]O N(1b)]Ru(3)]N(4b) 1.74(3) 1.75(3) 1.75(3) 1.75(3) 1.75(3) 1.74(3) 1.89 1.13 77.4(3) Table 2 Selected bond lengths (Å) and angles (8) for [Ru4(m-H)4(CO)10- (bpy)] 3 Ru(1)]Ru(2) Ru(1)]Ru(3) Ru(1)]Ru(4) Ru(2)]Ru(3) Ru(2)]Ru(4) Ru(3)]Ru(4) Ru(3)]C(31) Ru(3)]N(1) Ru(3)]N(2) Ru(1)]H(1) 2.925(7) 3.023(7) 2.795(8) 2.964(7) 2.780(8) 2.946(7) 1.844(7) 2.095(5) 2.090(5) 1.76(3) Ru(1)]H(4) Ru(2)]H(2) Ru(2)]H(4) Ru(3)]H(1) Ru(3)]H(2) Ru(3)]H(3) Ru(4)]H(3) mean Ru]C mean C]O N(1)]Ru(3)]N(2) 1.76(3) 1.77(3) 1.77(3) 1.76(3) 1.75(3) 1.76(3) 1.76(3) 1.90 1.13 77.4(2)2628 J.Chem. Soc., Dalton Trans., 1998, Pages 2625–2633 bridged Ru]Ru bonds [average Ru]Ru distance 2.964(6) Å in 1 and 2.964(5) Å in 3] compared to the two unbridged Ru]Ru bonds [average Ru]Ru distance 2.773(4) Å in 1 and 2.787(4) Å in 3].Similar deformation of the metal core has been reported for other tetrahedral Ru4 clusters with bridging hydride ligand(s).6,17 Fig. 2 A, An ORTEP drawing of the crystal structure of [Ru4(m-H)4- (CO)10(bpym)] 1. B, An ORTEP drawing of the crystal structure of [Ru4(m-H)4(CO)10(bpy)] 3 Table 3 Raman wavenumbers of co-ordinated 2,29-bipyrimidine from resonance-Raman spectra of [Ru4(m-H)4(CO)10(bpym)] 1 in KNO3 at room temperature, compared with those of related complexes [W(CO)4(bpym)] 19 and [Os3(CO)10(bpym)] 13 Compound n/cm21 1[ W(CO)4(bpym)] [Os3(CO)10(bpym)] 1575 1577 1578 1548 1548 1537 1466 1463 1463 1415 1417 1407 1335 1335 1340 1203 1198 1194 1025 1017 1018 Spectroscopic properties of [Ru4(Ï-H)4(CO)10(L)] IR and 1H NMR spectroscopy.The infrared spectra of 1–3 in the CO stretching region are very similar and correspond to terminal co-ordination of all carbonyl ligands.The slightly larger n(CO) wavenumbers of 1 and 2 relative to those of 3 reflect less pronounced RuÆCO p-back donation due to the larger p-acceptor capacity of the bpym and dpp ligands compared to that of the bpy ligand.18 The 1H NMR data of 1 and 3 (see Experimental section) confirm that the bpy and bpym ligands co-ordinate to Ru in a normal bidentate fashion. The two aromatic rings of these ligands remain magnetically equivalent upon co-ordination.For cluster 1, the D2h symmetry of the free bpym ligand is lost upon co-ordination. The 1H NMR resonances of the cluster 2 could be completely assigned with the aid of decoupling experiments described in detail in SUP 57401. The chemical shift of the ortho protons adjacent to the co-ordinated nitrogen atoms to lower values by 0.12–0.50 ppm in comparison with the resonances of the unco-ordinated ligands L is characteristic for sN, sN9-co-ordination of L to a low-valent metal centre.This eVect arises from a larger contribution of a resonance form in which the nitrogen atoms bear a partial negative charge and the adjacent carbon atoms a corresponding partial positive charge. Notably, the hydride resonances show a higher symmetry than expected from the solid-state structures. There are two hydride resonances around d 216 and 222, each accounting for two protons. The signals are rather weak and broad, indicating that a fluxional process may well be responsible for their pairwise equivalence. Electronic absorption and resonance-Raman spectra.The clusters 1–3 are orange to red-brown in colour. Their UV/VIS absorption maxima and corresponding molar absorption coe Ycients in CH2Cl2 are listed in the Experimental section. The position of the lowest energy absorption band depends on the electronic properties of the a-diimine ligands L and on the solvent polarity. For 3, the absorption maximum shifts from 486 nm in benzene to 463 nm in THF and 447 nm in acetonitrile.In contrast to this, the absorption maxima at higher energy hardly show any solvatochromism. The negative solvatochromism and the low-energy shift of lmax, resulting from replacement of 2,29-bipyridine by the stronger p-acceptor 2,29-bipyrimidine (from 463 to 484 nm in THF) point to a charge-transfer character of the lowest-energy electronic transition. Its nature was further investigated by resonance- Raman (rR) spectroscopy on visible excitation of 1 with four diVerent Ar1 laser lines (see Experimental section).The result is presented in Table 3. The main rR eVect is observed for bands in the 1000–1600 cm21 region which belong to internal stretching modes of the 2,29-bipyrimidine ligand. The presence of the Raman band at 1335 cm21, which is assigned to the inter-ring C]C stretching vibrations, indicates population of the lowest p*(b2u) orbital of 2,29-bipyrimidine in the charge-transfer excited state.19 Importantly, the peak due to the resonance enhanced ns(CO) vibration of 1 also showed up at 2070 cm21, indicating depopulation of the Ru(dp) orbitals involved in the RuÆCO p-back bonding.The combined resonance-Raman and UV/VIS features thus reveal that the lowest electronic transition of [Ru4(m-H)4(CO)10- (L)] is directed towards the a-diimine ligand L, having a signifi- cant Ru(dp)ÆL(p*) charge-transfer (MLCT) character. (Spectro)electrochemical studies of [Ru4(Ï-H)4(CO)10(L)] The redox properties of clusters 1–3 and their reduction paths were investigated by cyclic voltammetry, IR and UV/VIS spectroelectrochemistry, and by EPR spectroscopy.Redox potentials of 1–3 and their reduction products are presented in Table 4. Cyclic voltammograms of 2 and 3 are depicted in Fig. 3. Infrared n(CO) wavenumbers and UV/VIS spectroscopicJ. Chem. Soc., Dalton Trans., 1998, Pages 2625–2633 2629 Table 4 Redox potentials of the clusters 1–3 and their reduction products a Cluster 123 Epc (R1) b 21.54 21.57 21.88 DEp (R1/O1) 0.12 0.11 0.14 Epc (R2) c 22.23 22.23 22.35 Epc (R3) d 22.38 22.38 22.42 Epa (Om) e 10.39 10.41 10.35 Epa (O29) f 21.04 21.08 21.17 a Conditions: 2 × 1023 M 1–3 in THF–3 × 1021 M NBu4PF6, T = 298 K, Pt disc electrode, n = 100 mV s21; potentials given in V vs.E2� 1 (Fc/Fc1) (= 10.575 V vs. SCE); DEp(Fc/Fc1) = 0.11 V. b Reduction (1e) of 1–3. c Reduction (1e) of the corresponding radical anions 1b–3b. d Reduction (1e) of the dihydrido dianions 1c–3c. e Oxidation of 1–3.f Oxidation of the dihydrido dianions 1c–3c. Table 5 IR n(CO) wavenumbers of the clusters 1–3 and their electrochemical reduction products n/cm21 1 a,b 1b a,b 1c b 2 a,b 2 c 2b a,b 2b c 2c b 3 b 3c b 3e b 2073m 2065m 2006w 1770 (sh) 2073m 2074m 2063m 2066m 2008m 1762 (sh) 2071m 2025w 1879m 1985vw 1707w 2042s 2033s 1979m 1736m/w 2041s 2042s 2031s 2033s 1979m 1744m 2039vs 1990s 1862 (sh) 1948m 1680w 2018s 2008s 1956 (sh) 1712w 2018s 2017s 2006s 2008s 1957s 1724m/w 2015vs 1956 (sh) 1771 (sh) 1905s 2000m 1989m 1929s (br) 1677vw 2000m 1999m 1986m 1987m 1934s (br) 1678vw 1996s 1948vs 1735w 1892s 1979w 1970w 1889s 1979m 1979m 1966w 1966w 1892s 1974m 1923s 1712w 1848m 1945w 1943w 1867 (sh) 1946w 1942w 1942w 1941w 1875 (sh) 1947 (sh) 1910 (sh) 1670vw 1831m a In CH2Cl2–electrolyte at T = 298 K.b In THF–electrolyte at T = 298 K. c In CH2Cl2–electrolyte at T = 223 K. data of 1–3 and their reduction products are collected in Tables 5 and 6, respectively.Radical anions [Ru4(Ï-H)4(CO)10(L)]~2 (L 5 bpym or dpp). Reduction of the clusters 1 and 2 [cathodic peak R1 in Fig. Fig. 3 Cyclic voltammograms of 2 (A) and 3 (B). Oxidation of 3c generated via bulk electrolysis of 3 at E(R1) (B1). Conditions: Pt disc microelectrode (0.42 mm2 apparent surface), THF–NBu4PF6, T = 298 K, n = 0.1 V s21 3(A)] is both electrochemically and chemically reversible on the subsecond time-scale of cyclic voltammetry (defined by n > 0.02 V s21), satisfying the usual diagnostic tests.20 The product of this cathodic step was further studied in situ by IR and UV/VIS spectroelectrochemistry in a thin-layer cell.Fig. 4 shows the IR/UV/VIS spectral changes accompanying the reduction of 1. The n(CO) pattern of 1 remains preserved on reduction, being shifted by some 9–12 cm21 to smaller wavenumbers. The retention of isosbestic points indicates stability of the reduction product at room temperature over a period of ca. 10 min. After this period, however, a slow secondary reaction was evident from gradually increasing absorbances below 1900 cm21. This secondary reaction was not observed at T < 253 K, allowing characterization of the primary reduction product, denoted as 1b, by UV/VIS spectroscopy [see Fig. 4(a)]. The most prominent features of the UV/VIS spectrum of 1b are the two sharp absorption bands at 475 and 503 nm, accompanied by a lower-lying broad absorption band with lmax ª 850 nm.These bands can straightforwardly be attributed to intraligand (IL) electronic transitions of the co-ordinated radical anion [bpym]~2.19,21 From the cyclic voltammetric and spectroelectrochemical experiments it is concluded that the reduction of 1 is a one-electron process which produces the corresponding radical anion [Ru4(m-H)4(CO)10(bpym)]~2 1b with the odd electron dominantly localized on the lowest p* orbital of the 2,29-bipyrimidine ligand.This also applies for 2. In this case, however, the radical anion 2b (see Tables 5 and 6) partly decomposed in CH2Cl2 at room temperature in the course of the reduction of 2, i.e. more rapidly than 1b under the same conditions. The decomposition significantly slowed down in less polar THF. The UV/VIS spectra of 2b (see Table 6) were recorded in CH2Cl2 at T = 223 K where the radical anion remained inherently stable and could be fully reoxidized back to 2. The radical nature of 1b and 2b was unambiguously con- firmed by recording their EPR spectra (see Fig. 5). For this purpose 1b and 2b were generated in THF at T = 263 K by bulk electrolysis at the cathodic potential E(R1). The EPR spectra, found at g = 2.0015 for 1b and g = 2.0016 for 2b (i.e. close to the2630 J. Chem. Soc., Dalton Trans., 1998, Pages 2625–2633 Table 6 UV/VIS spectra of the clusters 1–3 and some of their reduction products lmax/nm (emax/M21 cm21) 1 a 1b a 2 b 2b b 3 c 3c c 256 (sh) (31 000) 278 (sh) (23 000) 274 (41 500) 266 (sh) (39 000) 301 (32 600) 336 (12 600) 358 (13 700) 362 (20 200) 357 (15 500) 349 (23 500) 354 (12 100) 484 (3350) 475 (7000) 486 (3700) 459 (sh) (9000) 463 (3200) ca. 500 (sh) 503 (7150) 763 (1500) 850 (650) 830 (1450) a In CH2Cl2 at T = 253 K. b In CH2Cl2 at T = 223 K. c In THF at T = 298 K. All solutions contained 3 × 1021 M NBu4PF6. free-electron value), are not well resolved, showing some hyper- fine structure which probably originates from splitting due to the 14N (I = 1, 99.63% abundance) and 1H (I = ��� , 99.98% abundance) nuclei of the a-diimine ligand. Additional splitting, partly responsible for the poorly resolved EPR signals, may arise from the 99Ru (I = 5 2 –, 12.7% abundance) and 101Ru (I = 5 2 –, 17.1% abundance) nuclei, and from the 1H nuclei bridging the cluster edges. Elimination of H2 from [Ru4(Ï-H)4(CO)10(L)]~2/22.Subsequent chemically irreversible one-electron reduction of the radical anions 1b and 2b at the cathodic potential E(R2) [see Table 4 and Fig. 3(A)] yielded 1c and 2c, respectively (see Tables 5 and 6). These compounds are identical with the species produced from 1b and 2b in a thermal, probably disproportionation 3 reaction (see above). The dianionic clusters 1c and 2c are also formed via the direct two-electron route, from unstable dianions [Ru4(m-H)4(CO)10(L)]22 (L = bpym or dpp) produced Fig. 4 Spectral changes in the IR n(CO) [(a), in THF at T = 298 K] and UV/VIS [(b), in CH2Cl2 at T = 253 K] regions, accompanying the reversible reduction of the cluster 1 producing the radical anion 1b during the one-electron reduction of 1b and 2b.The nature of the products 1c and 2c was elucidated by performing a (spectro)electrochemical study of the cluster 3. The reduction of 3 at E(R1) is shifted more negatively compared to the reduction potentials of 1 and 2 [see Fig. 3(B) and Table 4], being only partly chemically reversible at n = 0.1 V s21 (Ipa/Ipc ª 0.5).This result agrees with a predominantly bpy-localized one-electron cathodic step producing the unstable radical anion [Ru4(m-H)4(CO)10(bpy)]~2 3b. Obviously, the more basic bpy ligand with a higher-lying p* LUMO in comparison with the bpym and dpp ligands is less suited to accommodate the added electron in this case. The radical anion 3b, though still detectable on the reverse anodic scan due to its oxidation at E(O1) [see Fig. 3(B)], decomposes more rapidly than 1b and 2b and was not observed by IR, UV/VIS and EPR spectroscopy during spectroelectrochemical experiments at room temperature. Instead, the reduction of 3 on the time-scale of minutes yielded exclusively the dianion 3c, as was judged from the close similarity of IR spectra of 1c–3c (see Fig. 6, left, and Table 5). The UV/VIS spectrum of 3c is depicted in Fig. 6, right. The n(CO) bands between 1800 and 1670 cm21 in the IR spectra of 1c–3c are indicative of bridging CO ligands which may have replaced some of the edge-bridging hydrogen atoms on the reduction of 1b–3b.This behaviour is analogous to that reported for unsubstituted [Ru4(m-H)4(CO)12] whose irreversible two-electron reduction produces the dianion [Ru4(m-H)2- (CO)12]22 with three CO ligands bridging the basal Ru]Ru bonds.3,22 We could obtain the latter species neatly via stepwise deprotonation of [Ru4(m-H)4(CO)12] in THF on addition of 2 equivalents of NEt4OH in MeOH, in conformity with the Fig. 5 The EPR spectra of 1b (a) and 2b (b) in THF at 298 KJ. Chem. Soc., Dalton Trans., 1998, Pages 2625–2633 2631 recorded IR spectra attributed to the intermediate NEt4[Ru4- (m-H)3(CO)12] [n(CO) at 2074w, 2038s, 2032 (sh), 2015s, 1990s cm21] 23 and the yellow-orange final product [NEt4]2[Ru4(m-H)2- (CO)12] [n(CO) at 2031w, 1990s, 1951vs (br), 1904m, 1886m, 1814vw, 1761m, 1744m cm21] 22 (see Fig. 7). The salt NEt4OH was therefore chosen as a suitable reagent for the antideprotonation of 3.The reaction was performed in MeCN owing to poor solubility of the product(s) in THF. Initially a brown species 3c9 was formed whose IR spectrum closely resembled that of [NEt4]2[Ru4(m-H)2(CO)12] under the same conditions: n(CO) at 2004w, 1982m, 1941vs, 1900 (sh), 1887m, 1797vw, 1754m, 1737m cm21 (see Fig. 7). The smaller wavenumbers of 3c9 relative to those of [NEt4]2[Ru4(m-H)2(CO)12] (by ca. 10 cm21) point to co-ordination of the s-donor 2,29- bipyridine ligand, which was also evident from the 1H NMR spectrum of 3c9 in CD3CN.The signals of the bpy-protons of the parent cluster 3 [d 9.01 (d, 2 H), 8.44 (d, 2 H), 8.20 (dt, 2 H), 7.64 (dt, 2 H)] were replaced after addition of NEt4OH by a new set for 3c9 at d 8.80 (d, 2 H), 8.26 (d, 2 H), 8.07 (t, 2 H) and 7.56 (t, 2 H). These values deviate from those for the uncoordinated 2,29-bipyridine under identical conditions [d 8.80 (d), 8.52 (d), 7.99 (dt) and 7.50 (dt)].Notably, the cluster 3c9 could only be detected as an unstable intermediate. It transformed within a few minutes into another CO-bridged compound 3d Fig. 6 Left: infrared spectral changes [n(CO) region] accompanying the irreversible reduction of the cluster 3 producing the dianion 3c. Right: UV/VIS spectra of 3 (dot-dash line), 3c (full line) and the tetraanion 3e, the 2e reduction product of 3c (broken line). All spectra were recorded in THF at T = 298 K Fig. 7 The IR spectra [n(CO) region] of (a) [NEt4]2[Ru4(m-H)2(CO)12], (b) 3c9 in MeCN, and (c) 3d in MeCN. All spectra recorded at T = 298 K characterized by n(CO) bands at 2034 (sh), 2017m, 1992 (sh), 1967vs, 1955 (sh), 1926m, 1907m, 1821vw, 1770m cm21 (see Fig. 7) and, in the 1H NMR spectrum, by resonances due to the bpy ligand at d 8.68 (d, 2 H), 8.30 (d, 2 H), 8.09 (dt, 2 H), 7.56 (dt, 2 H) and broad hydride resonances at 216.8 (s, 1 H) and 223.7 (s, 1 H). Recall that the 1H NMR spectrum of [Ru4(m-H)2- (CO)12]22 exhibits at room temperature only a sharp hydride singlet at d 219.26.22 Importantly, IR spectroelectrochemical experiments revealed that 3d is identical to the product of electrochemical oxidation of 3c at the anodic potential E(O29) (see Fig. 3 and Table 4). The nature of the conversion of 3c9 into 3d, apparently demanding an oxidation step, was not investigated in detail. All attempts to crystallize 3d have been unsuccessful so far. In essence, the above IR and 1H NMR data show that deprotonation of 3 yields 3c9 which can be formulated as the tetrahedral dianion [Ru4(m-H)2(CO)10(bpy)]22, probably with three bridging CO ligands such as those established for the unsubstituted derivative [Ru4(m-H)2(CO)12]22.22 Successive electrochemical reduction of 3 and the radical anion 3b ultimately yields 3c exhibiting slightly lower n(CO) wavenumbers and a diVerent, more complex n(CO) pattern with regard to those of 3c9 (see Figs. 6 and 7).The question remains how much 3c corresponds to 3c9. According to the IR spectra, these species are not identical. There is also no evidence that 3c and 3c9 interconvert. The actual diVerence between them, however, is believed not to be significant as both identically produce 3d (see above). For comparison, both electrochemical reduction 3 and deprotonation 22 of [Ru4(m-H)4(CO)12] yielded the same product, viz. [Ru4(m-H)2(CO)12]22. At this stage of investigation we can conclude that, similarly to 3c9, the electrochemical reduction product 3c is also a dianionic Ru4-cluster with edgebridging hydride and carbonyl ligands, formed from the transient cluster [Ru4(m-H)4(CO)10(bpy)]22 via elimination of H2. The very similar averaged n(CO) wavenumbers and the almost a-diimine-independent oxidation potentials E(O29) of the dianions 1c–3c (see Tables 4 and 5) imply eVective delocalization of the negative charge over the cluster core, residing more at the CO-bridged basal Ru3(CO)9 moiety than at the apical Ru(a-diimine) fragment.The UV/VIS spectrum of 3c exhibits a broad band at 763 nm which may belong to Ru(bpy)- localized electronic transitions, for [Ru4(m-H)2(CO)12]22 does not absorb at such a low energy. Further reduction of 3c at the electrode potential E(R3), studied by IR/UV/VIS spectroelectrochemistry, yielded the cluster 3e (see Table 5). The CO-bridge absorptions in the IR spectrum of 3c shifted on the reduction to lower wavenumbers by approximately 30 cm21, which is slightly more than that found for [N(PPh3)2]2[Ru4(m-H)2(CO)12] 22 and [Ph4P]4[Ru4- (CO)12].24 The UV/VIS spectrum of 3e shows a broad structured band with maxima at 533 and 572 nm and shoulders at 505, 640, 700 and 760 nm (see Fig. 6).These features strongly resemble the visible absorption of the two-electron-reduced anion [Re(CO)3(bpy)]2 having the added two electrons strongly p-delocalized over the Re(bpy) chelate ring.25,26 We may thus assign 3e as [NBu4]4[Ru4(CO)10(bpy)], with the two extra electrons added to the dianion 3c predominantly residing on the p-system of the Ru(bpy) moiety.According to the cyclic voltammograms of 2 and 3, the reduction of 3c at E(R3) should be a one-electron process which aVords the radical [3c]~2. This transient species was indeed observed in the course of the UV/ VIS OTTLE experiment. Its UV/VIS spectrum exhibits a characteristic structured band with absorption maxima at 495 and 522 nm, belonging to an intraligand electronic transition of the one-electron-reduced ligand [bpy]~2.27 The radical [3c]~2 probably further disproportionated into 3c and 3e under liberation of H2.Recall that reduction of [Ru4(m-H)2(CO)12]22 is also initially a one-electron process; 3 although, deprotonation of the dianion with 1 equivalent of KH only produced a 1: 1 mixture of [Ru4(m-H)2(CO)12]22 and [Ru4(CO)12]42.22 In fact, the2632 J. Chem. Soc., Dalton Trans., 1998, Pages 2625.2633 only reported route to [Ru4(m-H)(CO)12]32 involved protonation of [Ru4(CO)12]42 with HBr.24 The reduction path of 1.3 is summarized in Scheme 2.Oxidation of [Ru4(I-H)4(CO)10(L)] 10. The cyclic voltammograms of 2 and 3 in Fig. 3 reveal that these clusters undergo chemically irreversible oxidation consuming more than one electron [Ia(Om) �£ 2Ic(R1)]. The very close values of the corresponding anodic peak potentials E(Om) (see Table 4) are indicative of an anodic process localized on the metal core, which may induce cleavage of Ru]Ru bonds.The oxidation of [Ru4- (m-H)4(CO)10(L)] was not further investigated. Comparison of the reduction paths of [Ru4(I-H)4(CO)10(L)] and [Os3(CO)10(L)]. Both [Ru4(m-H)4(CO)10(L)] and [Os3(CO)10- (L)] 4 (L = bpym, dpp or bpy) initially undergo one-electron reduction mainly localized on the lowest p* orbital of the adiimine ligand. The corresponding radical anionic products are therefore substantially more stable than those derived from the unsubstituted clusters [Ru4(m-H)4(CO)12] 3 and [Os3(CO)12],28 and can be detected by diverse spectroelectrochemical methods.The reactivity induced by reduction of these low-nuclearity clusters can conveniently be controlled by tuning the electronic properties of the a-diimine ligands L. Raising the s,p-donor character of the reduced a-diimine initiates secondary chemical/electron transfer reactions of the radical anions, whose nature is identical with the reactivity of the purely carbonyl precursors.In particular, the species [Os3(CO)12 2 2n- (L)n]~2 (n = 0 or 1) undergo cleavage of an Os]Os bond resulting ultimately in formation of open-structure dianionic clusters.4 Contrary to this, the species [Ru4(m-H)4(CO)12 2 2n- (L)n]~2 (n = 0 or 1) react via the loss of hydrogen to give the corresponding dianions [Ru4(m-H)2(CO)12 2 n(L)n]22. Notably, we have obtained no spectroscopic evidence for participation of the anion [Ru4(m-H)3(CO)10(L)]2 in the reduction path of [Ru4(m-H)4(CO)10(L)], as was reported 3 for the reduction of [Ru4(m-H)4(CO)12].The radical anions [Ru4(m-H)4(CO)10(L)]~2 are apparently more stable than the corresponding species [Os3(CO)10(L)]~2. In the latter case only [Os3(CO)10(bpym)]~2 could be characterized spectroscopically, becoming stable at T = 213 K. The radical [Os3(CO)10(dpp)]~2 is only stable on a subsecond time-scale of cyclic voltammetry while [Os3(CO)10(bpy)]~2 partly decomses at moderate scan rates even at T = 220 K and at room temperature it is not detectable at all.4 The order of increasing stability of the radical anions on co-ordination of stronger p-acceptor L, bpy ! dpp < bpym, applies also for [Ru4(m-H)4- (CO)10(L)]~2.All the latter species could be detected by conventional cyclic voltammetry at room temperature and, for L = dpp or bpym, they were also characterized by IR, UV/VIS Scheme 2 Reduction paths of the clusters [Ru4(m-H)4(CO)10(L)] [Ru4(m-H)4(CO)10(L)] 1.3 [Ru4(m-H)4(CO)10(L)].. 1b.3b [Ru4(m -H)4(CO)10(L)]2.+ e. (1b, 2b stable at T 253 K) .H2 1c, 2c 3c [Ru4(m-H)2(CO)10(L)]2. disprop. .H2 rapid (3b) slow (1b,2b) L = bpy 3d (.e. ?) 3e + e. 3c' .e. .H2 + 2e. + 2 OH. . 2 H2O and EPR spectroscopy. The large diVerence in the stability of [Ru4(m-H)4(CO)10(L)]~2 and the corresponding Os3 derivatives can be ascribed to a more delocalised bonding situation in the robust closely-packed tetrahedral Ru4(m-H)4 core.Conclusion The co-ordination of the reducible a-diimine ligand L in the novel clusters [Ru4(m-H)4(CO)10(L)] prevents rapid decomposition of the primary one-electron reduction products. The radical anions [Ru4(m-H)4(CO)10(L)]~2 (L = dpp or bpym) are stable on the time-scale of minutes and could be characterized by IR, UV/VIS and EPR spectroscopy. Regardless of this stabilizing influence, the overall reactivity remains unaVected.Either on uptake of another electron, or thermally via disproportionation, the radical anions [Ru4(m-H)4(CO)10(L)]~2 lose dihydrogen and transform to dianions [Ru4(m-H)2(CO)10(L)]22 which contain edge-bridging CO ligands. The reactivity is highest for [Ru4(m-H)4(CO)10(bpy)]~2 which could only be observed by cyclic voltammetry. In this respect, the reactivity of [Ru4- (m-H)4(CO)10(L)]~2 parallels that of unsubstituted [Ru4(m-H)4- (CO)12]~2. Acknowledgements We thank the Netherlands Foundation of Chemical Research (SON), the Netherlands Organization for Scientific Research (NWO) and the University of Edinburgh for financial assistance.Professor L. J. Yellowlees, Dr. N. Payne, Dr. S. Parsons and Dr. A. J. Blake (all from the University of Edinburgh) are jointly acknowledged for their helpful contributions to this work. References 1 L. N. Lewis, Chem. Rev., 1993, 93, 2693. 2 Y. Doi, S. Tamura and K. Koshizuka, J. Mol. Cat., 1983, 19, 213; Inorg.Chim. Acta, 1982, 65, L63. 3 D. Osella, C. Nervi, M. Ravera, J. Fiedler and V. V. Strelets, Organometallics, 1995, 14, 2501. 4 F. Hartl, J. W. M. van Outersterp and D. J. Stufkens, Organometallics, submitted for publication; J. W. M. van Outersterp, Ph.D. Thesis, University of Amsterdam, 1995. 5 F. Piacenti, M. Bianchi, P. Frediani and E. Benedetti, Inorg. Chem., 1971, 10, 2759. 6 M. Bianchi, P. Frediani, A. Salvini, L. Rosi, L. Pistolesi, F. Piacenti, S. Ianelli and M. Nardelli, Organometallics, 1997, 16, 482. 7 S.A. R. Knox, J. W. Koepke, M. A. Andrews and H. D. Kaesz, J. Am. Chem. Soc., 1975, 97, 3942. 8 G. M. Sheldrick, SHELXS 86, Program for the Solution of Crystal Structures, Acta Crystallogr., Sect. A, 1990, 46, 467; SHELXL 93, University of Gottingen, 1993; C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. 9 G. Gritzner and J. Kuta, Pure Appl. Chem., 1984, 56, 461. 10 M. Krejc¢§ik, M. Dane¢§k and F.Hartl, J. Electroanal. Chem. Interfacial Electrochem., 1991, 317, 179. 11 F. Hartl, H. Luyten, H. A. Nieuwenhuis and G. C. Schoemaker, Appl. Spectrosc., 1994, 48, 1522. 12 R. R. Andrea, H. Luyten, M. A. Vuurman, D. J. Stufkens and A. Oskam, Appl. Spectrosc., 1986, 40, 1184. 13 J. W. M. van Outersterp, M. T. Garriga Oostenbrink, H. A. Nieuwenhuis, D. J. Stufkens and F. Hartl, Inorg. Chem., 1995, 34, 6312. 14 G. Freeman, L. J. Yellowlees, S. I. Ingham and B. F. G. Johnson, unpublished work. 15 S. A. R. Knox and H. D. Kaesz, J. Am. Chem. Soc., 1971, 93, 4594. 16 M. J. Schadt, M. Gresalfi and A. J. Lees, Inorg. Chem., 1985, 24, 2942. 17 A. U. Harkonen, M. Ahlgren, T. A. Pakkanen and J. Pursiainen, J. Organomet. Chem., 1997, 530, 191. 18 J. W. M. van Outersterp, F. Hartl and D. J. Stufkens, Organometallics, 1995, 14, 3303. 19 W. Kaim, S. Kohlmann, A. J. Lees, T. L. Snoeck, D. J. Stufkens and M. M. Zulu, Inorg. Chim. Acta, 1993, 210, 159.J. Chem. Soc., Dalton Trans., 1998, Pages 2625–2633 2633 20 R. Greef, R. Peat, L. M. Peter, D. Pletcher and J. Robinson, in Instrumental Methods in Electrochemistry, ed. T. J. Kemp, Ellis Horwood, Chichester, 1985, ch. 6, p. 186. 21 P. S. Braterman, J.-I. Song, C. Vogler and W. Kaim, Inorg. Chem., 1992, 31, 222. 22 K. E. Inkrott and S. G. Shore, Inorg. Chem., 1979, 18, 2817. 23 J. W. Koepke, J. R. Johnson, S. A. R. Knox and H. D. Kaesz, J. Am. Chem. Soc., 1975, 97, 3947. 24 A. A. Bhattacharyya, C. C. Nagel and S. G. Shore, Organometallics, 1983, 2, 1187. 25 G. J. Stor, F. Hartl, J. W. M. van Outersterp and D. J. Stufkens, Organometallics, 1995, 14, 1115. 26 B. D. Rossenaar, F. Hartl and D. J. Stufkens, Inorg. Chem., 1996, 35, 6194. 27 M. Krejc¡ík and A. A. Vlc¡ek, J. Electroanal. Chem. Interfacial Electrochem., 1991, 313, 243. 28 A. J. Downard, B. H. Robinson, J. Simpson and A. M. Bond, J. Organomet. Chem., 1987, 320, 363. Received 16th April 1998; Paper 8/02849DJ. Chem. Soc., Dalton Trans., 1998, Pages 2625–2633 2633 20 R. Greef, R. Peat, L. M. Peter, D. Pletcher and J. Robinson, in Instrumental Methods in Electrochemistry, ed. T. J. Kemp, Ellis Horwood, Chichester, 1985, ch. 6, p. 186. 21 P. S. Braterman, J.-I. Song, C. Vogler and W. Kaim, Inorg. Chem., 1992, 31, 222. 22 K. E. Inkrott and S. G. Shore, Inorg. Chem., 1979, 18, 2817. 23 J. W. Koepke, J. R. Johnson, S. A. R. Knox and H. D. Kaesz, J. Am. Chem. Soc., 1975, 97, 3947. 24 A. A. Bhattacharyya, C. C. Nagel and S. G. Shore, Organometallics, 1983, 2, 1187. 25 G. J. Stor, F. Hartl, J. W. M. van Outersterp and D. J. Stufkens, Organometallics, 1995, 14, 1115. 26 B. D. Rossenaar, F. Hartl and D. J. Stufkens, Inorg. Chem., 1996, 35, 6194. 27 M. Krejc¡ík and A. A. Vlc¡ek, J. Electroanal. Chem. Interfacial Electrochem., 1991, 313, 243. 28 A. J. Downard, B. H. Robinson, J. Simpson and A. M. Bond, J. Organomet. Chem., 1987, 320, 363. Received 16th April 1998; Paper 8/02849D

ISSN:1477-9226    

DOI:10.1039/a802849d

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 2635–2641 2635 Binding, electrochemical and metal extraction properties of the new redox-active polyazacycloalkane 1,4,7,10,13,16-hexa- (ferrocenylmethyl)-1,4,7,10,13,16-hexaazacyclooctadecane José Manuel Lloris, Ramón Martínez-Máñez,*,† Teresa Pardo, Juan Soto and Miguel E. Padilla-Tosta Departamento de Química, Universidad Politécnica de Valencia, Camino de Vera s/n, 46071 Valencia, Spain The redox-functionalised polyazacycloalkane receptor 1,4,7,10,13,16-hexa(ferrocenylmethyl)-1,4,7,10,13,16- hexaazacyclooctadecane (L1) has been designed and synthesized and its binding, electrochemical and metal extraction properties studied.The results have been compared to those obtained for the parent redox-active polyazacycloalkane 1,4,8,11-tetra(ferrocenylmethyl)-1,4,8,11-tetraazacyclotetradecane, L3. The cadmium complex of L1 and the free receptor L3 have been crystallographically characterised. The structure of [CdL1][PF6]2?4CH3CN?4H2O consists of cationic [CdL1]21 units in which the cadmium ion exhibits a distorted octahedral environment.Solution studies by potentiometric methods have also been carried out in the presence of H1 and Cu21 in 1,4-dioxane–water (70 : 30 v/v, 25 8C, 0.1 mol dm23 KNO3) for L1 and thf–water (70 : 30 v/v, 25 8C, 0.1 mol dm23 [NBun 4][ClO4]) for L3. Electrochemical experiments as a function of the pH in the presence of Ni21, Cu21, Zn21, Cd21 or Pb21 have been carried out for L1 and L3.The solvent extraction properties of L1 have also been studied. Among recently developed functionalised molecules are redoxresponsive receptors which have proved to be suitable molecules for the electrochemical recognition of both cationic and anionic substrates.1 These new functionalised receptors comprise a well known framework able to bind substrates and reversible redox-active units. DiVerent electroactive moieties have been covalently anchored 2 near binding sites but probably the most widely used has been ferrocene.However most of the work has been devoted to the recognition of alkali- and alkaline-earth metal ions 3 and relatively less eVort has been carried out in the recognition of transition-metal ions and anions.4 Functionalisation of ferrocenyl groups with polyazacycloalkanes leads to new receptor molecules which (i) can bind transition-metal ions,5 (ii) can be water soluble 6 and (iii) are good candidates for anion binding due to the formation of highly charged species and/or hydrogen bonding networks.7 Apart from their electroactive character, these ferrocene-functionalised polyazaalkanes have other interesting properties which have probably not been fully studied.The lipophilic character of the ferrocenyl units and the presence of metal binding sites makes these molecules good candidates to be used in preconcentration of trace elements by solvent extraction for analytical purposes or as models for the elimination of toxic metals from water.Additionally the electroactive character of the ferrocenyl groups could make these molecules able to drive ions against a chemical concentration gradient by changing the co-ordination ability by oxidation or reduction of the electroactive groups.8 We have designed and synthesized the redox-functionalised polyazacycloalkane receptor 1,4,7,10,13,16-hexa(ferrocenylmethyl)- 1,4,7,10,13,16-hexaazacyclooctadecane (L1) and have studied its binding, electrochemical and metal extraction properties and the results have been compared to those obtained for the redox-active polyazacycloalkane 1,4,8,11-tetra(ferrocenemethyl)- 1,4,8,11-tetraazacyclotetradecane, L3.† E-Mail: rmaez@qim.upv.es Experimental Solvents and reagents Tetrahydrofuran and 1,4-dioxane were freshly distilled. 1,4,7,10,13,16-Hexaazacyclooctadecane (L2), L3 and (ferrocenylmethyl) trimethylammonium iodide were synthesized follow- N N N N N N N N N N NH NH HN HN NH NH HN HN NH HN L4 L1 Fe Fe Fe Fe L3 L2 Fe Fe Fe Fe Fe Fe2636 J.Chem. Soc., Dalton Trans., 1998, Pages 2635–2641 ing refs. 9, 10 and 11, respectively. Metal cations Cu21, Ni21, Zn21, Cd21 and Pb21 were used as their nitrate salts. Syntheses L1. 1,4,7,10,13,16-Hexaazacyclooctadecane (0.5 g, 1.94 mmol) and (ferrocenylmethyl)trimethylammonium iodide (7.45 g, 19.4 mmol) were heated to reflux in acetonitrile (250 cm3) for 4 d in the presence of sodium carbonate (7 g).The warm reaction mixture was filtered and the yellow solution evaporated to dryness. The resulting solid was dissolved in dichloromethane and chromatographed using dichloromethane–methanol (100 : 1) as eluent. Further recrystallisation in dichloromethane– hexane gave L1 as a yellow solid (600 mg, 20%) (Found: C, 61.79; H, 6.07; N, 5.70. C78H90Fe6N6?CH2Cl2 requires C, 61.96; H, 6.01; N, 5.50%). NMR (CDCl3): 1H d 4.09, 4.06 (two br resonances, 54 H, C5H5 and C5H4), 3.35 (br, 12 H, CH2) and 2.40 (br, 24 H, CH2); 13C-{1H} d 83.24 (Cipso, C5H4), 70.23, 68.44 (C5H5), 67.89 (C5H4), 54.09 (CH2) and 51.22 (CH2).Mass spectrum (FAB): m/z 1446 (M1). [CdL1][PF6]2. Compound L1 (70 mg, 0.05 mmol) and Cd(NO3)2?4H2O (15 mg, 0.05 mmol) were heated to reflux in ethanol (75 cm3) for 20 h. After cooling to room temperature the reaction mixture was filtered. An excess of [NH4][PF6] and water were added to the yellow solution to give a yellow precipitate which was filtered oV, washed with water and dried in vacuo to give [CdL1][PF6]2 (35 mg, 40%) (Found: C, 49.78; H, 4.76; N, 4.50.C78H90CdF12Fe6N6P2 requires C, 50.0; H, 4.87; N, 4.54%). NMR (CDCl3): 1H d 4.54 (t, 12 H, C5H4), 4.39 (t, 12 H, C5H4), 4.29 (s, 30 H, C5H5), 4.18 (s, 12 H, CH2), 3.09 (br, 6 H, CH2), 3.05 (br, 6 H, CH2), 2.71 (br, 6 H, CH2) and 2.68 (br, 6 H, CH2). Physical measurements The NMR spectra were measured on a Varian Gemini spectrometer operating at 300 K.Chemical shifts for 1H and 13C- {1H} are referenced to SiMe4 and CDCl3, respectively. Electrochemical data were obtained with a Tacussel IMT-1 programmable function generator, connected to a PJT 120-1 potentiostat. The working electrode was platinum with a saturated calomel reference electrode separated from the test solution by a salt bridge containing the solvent/supporting electrolyte. The auxiliary electrode was platinum wire. Potentiometric titrations were carried out in 1,4-dioxane–water (70 : 30 v/v, 0.1 mol dm23 KNO3) for L1 and thf–water (70 : 30, v/v, 0.1 mol dm23 [NBun 4][ClO4]) for L3 using a reaction vessel waterthermostatted at 25.0 ± 0.1 8C under nitrogen.Experimental potentiomeric details have been published previously.5 The concentrations of the metal ions were determined using standard methods. The computer program SUPERQUAD12 was used to calculate the protonation and stability constants. The titration curves for each system (ca. 250 experimental points corresponding to at least three titration curves, pH = 2log[H] range investigated 2.5–10, concentration of the ligand and metal ion ca. 1.2 × 1023 mol dm23) were treated either as a single set or as separated entities without significant variations in the values of the stability constants. Finally the sets of data were merged and treated simultaneously to give the stability constants. Metal extraction Metal extraction experiments were carried out using water and dichloromethane.The corresponding metal ion (nitrate salts) was dissolved in water (100 cm3, metal concentration ca. 3 × 1024 mol dm23), whereas the L1 receptor was dissolved in the organic phase (100 cm3, ligand concentration ca. 3 × 1024 mol dm23). Both phases were placed in a flask and the mixture stirred. Samples of the aqueous phase were taken as a function of time and the metal concentration determined using standard atomic absorption methods.Electrolysis of the dichloromethane solutions containing the L1 complex of Pb21 were carried out using a Tacussel PJT 120-1 potentiostat. The working electrode was graphite, whereas a saturated calomel electrode and platinum were used as reference and auxiliary electrodes, respectively. Crystallography [CdL1][PF6]2?4CH3CN?4H2O. Crystal data. C78H90CdF12- Fe6N6P2?4CH3CN?4H2O, M = 208.48, monoclinic, space group P21/n, a = 16.974(3), b = 15.987(3), c = 18.375(3) Å, b = 108.635(12)8, Z = 2, U = 4725(1) Å3, Dc = 1.632 g cm23, l(Mo- Ka) = 0.710 69 Å, T = 293(2) K, m(Mo-Ka) = 12.83 cm21.Well shaped crystals were obtained by slow diVusion of water into acetonitrile solutions of the complex. They were unstable owing to loss of solvent. A capillary containing a crystal with approximate dimensions 0.2 × 0.2 × 0.1 mm and the motherliquor was mounted on a Siemens P4 four-circle diVractometer. A total of 5092 reflections were measured of which 4852 were unique (Rint = 0.052) (3.6 < 2q < 45.08) using the 2q– w method.Lorentz-polarisation corrections were applied but no allowance was made for absorption. The structure was solved by direct methods (SHELXTL)13 and refined by fullmatrix least-squares analysis on F2. Disorder appears to aVect some atoms in unattached cyclopentadienyl rings and some atoms modelled highly anisotropically. However no well defined peaks were found near these atoms and no attempts were made to model the disorder.The refinement converged at R1 0.061 [F > 4s(F)] and R2 0.188 (all data). Largest peak and hole in the final diVerence map 10.50, 20.36 e Å23. L3. Crystal data. C54H64Fe4N4, M = 992.49, monoclinic, space group P21/c, a = 17.593(4), b = 10.915(2), c = 12.764(3) Å, b = 105.54(3)8, Z = 2, U = 2361.4(8) Å3, Dc = 1.396 g cm23, l(Mo-Ka) = 0.710 69 Å, T = 293(2) K, m(Mo-Ka) = 12.45 cm21. A well shaped orange crystal of L3 with approximate dimensions 0.21 × 0.19 × 0.14 mm was mounted on a Siemens P4 four-circle diVractometer equipped with Mo-Ka radiation.A total of 4644 reflections were measured of which 2203 were unique (Rint = 0.169) (4.4 < 2q < 40.08) using the 2q–w method. Corrections were applied and the structure solved as above. Disorder appears to aVect some atoms in unattached cyclopentadienyl rings but no attempts were made to model it. The refinement converged at R1 0.066 [F > 4s(F)] for 856 reflections] and R2 0.1127 (all data).Largest peak and hole in the final diVerence map 10.27, 20.27 e Å23. CCDC reference number 186/1022. Results and Discussion We have designed and synthesized the molecule 1,4,7,10,13,16- hexa(ferrocenylmethyl)-1,4,7,10,13,16-hexaazacyclooctadecane (L1) by reaction of 1,4,7,10,13,16-hexaazacyclooctadecane (L2) with (ferrocenylmethyl)trimethylammonium iodide in acetonitrile in the presence of sodium carbonate. The compound was isolated after column chromatography as a crystalline yellow solid in a 20% yield.The 1H NMR spectrum shows only four signals indicating that the molecule is highly symmetric in solution on the NMR timescale. Two signals at d 4.09 and 4.06 are attributed to the protons of the cyclopentadienyl rings whereas broad signals at d 3.35 and 2.40 are assigned to the CH2 protons from the ferrocenylmethyl groups and from the hexaazacyclooctadecane ring. The 13C-{1H} NMR and FAB mass spectra (m/z = 1446, M1) are also consistent with the proposed formulation.Compound L1 is a potential N-donor hexadentate ligand containing redox-active peripherally attached groups. It seemed to us interesting to check (i) its chelating ability towards metal ions and analyse, by comparison with the analogousJ. Chem. Soc., Dalton Trans., 1998, Pages 2635–2641 2637 non-functionalised receptor L2, the influence of the bulky ferrocenyl groups near the co-ordination site, (ii) the influence of the co-ordination of substrates (H1 and metal ions) in the redox properties of the electroactive groups and (iii) the solvent extraction properties of L1 further to study potential backextraction processes by oxidation of the redox-active moieties.Additional studies on the related redox-active polyazacycloalkane 1,4,8,11-tetra(ferrocenylmethyl)-1,4,8,11-tetraazacyclotetradecane have also been carried out. The cadmium complex of L1 has been obtained by reaction with cadmium nitrate in dichloromethane–ethanol mixtures and further addition of [NH4][PF6] and water.The 1H NMR spectrum of [CdL1][PF6]2 shows magnetically equivalent ferrocenylmethyl groups and only four signals [two pseudo-triplets d 4.54, 4.39 and two singlets d 4.29 (C5H5) and 4.18 (CH2)]. However protons in the polyazacycloalkane framework are not equivalent and four broad signals were observed. In order fully to characterise the co-ordination mode of L1 we have solved the crystal structure of the cadmium complex by using X-ray single crystal procedures.Fig. 1 (a) Molecular structure of the cation [CdL1]21 showing the atomic numbering scheme. (b) An alternative representation showing the outer lipophilic sphere Crystal structure of [CdL1][PF6]?4CH3CN?4H2O The crystallographic characterisation of L1 as its cadmium complex has been carried out. Table 1 lists selected bond distances and angles. Suitable crystals of [CdL1][PF6]2 were obtained from slow diVusion of water into an acetonitrile solution of the complex.The [CdL1]21 unit is depicted in Fig. 1. The structure of the [CdL1]21 cation consists of an inner binding domain of six nitrogen atoms co-ordinating the central metal ion giving it a distorted octahedral geometry. The six nitrogen atoms are also covalently attached to six ferrocenylmethyl groups making up an outer redox-active sphere. The hexaazacyclooctadecane moiety adopts the fac configuration (D3d symmetry). The Cd]N distances range from 2.338(8) to 2.535(10) Å.Some angles around the central atom are far from the ideal octahedral geometry such as N(3)]Cd]N(2) 73.7(3) or N(1)]Cd]N(2) 74.4(3)8, probably due to geometrical constraints imposed by the ferrocenyl moieties. However no important distortions are present in the ferrocenylmethyl groups. The ferrocenyl units show the typical sandwich conformation with the cyclopentadienyl rings parallel within the experimental error; Fe]Cp (centroid) distances range from 1.617 to 1.655, averaging 1.639 Å, whereas Fe]C (Cp) lengths are between 1.93(2) and 2.05(2) [average 2.01(2) Å].The three crystallographically non-equivalent Cd]Fe distances are Cd ? ? ?Fe(1) 6.326, Cd ? ? ?Fe(2) 6.593 and Cd ? ? ?Fe(3) 6.479 Å. Crystal structure of L3 A view of the L3 molecule is depicted in Fig. 2. Selected bond distances and angles are in Table 2. The structure consists of the polyazacycloalkane 1,4,8,11-tetraazacyclotetradecane (cyclam) containing ferrocenylmethyl groups covalently attached to the four nitrogen atoms.The ferrocenyl moieties show the typical sandwich configuration with parallel and planar cyclopentadienyl rings within the experimental error. The iron–cyclopentadienyl (centroid) distances range from 1.634 to 1.655 Å, Fig. 2 Molecular structure of compound L3 showing the atomic numbering scheme2638 J. Chem. Soc., Dalton Trans., 1998, Pages 2635–2641 averaging 1.643 Å. The corresponding Fe]C distances range from 1.97(2) to 2.05(2) Å [averaging 2.02(2) Å].Cyclic tetraamines such as cyclam and their derivatives can occur in several conformations. For instance in [14]aneN4 or R4[14]aneN4 macrocycles with alternating five- and sixmembered chelating rings a total of five combinations can be produced. Compound L3 shows a conformation with two ferrocenylmethyl groups above the N4 plane and the remaining two below the N4 plane (configuration IV as described in ref. 14). Important changes are produced in the configuration of the L3 receptor when it interacts with metal ions as we have recently reported.15 Protonation behaviour Data for L1 and L3 have been determined in 1,4-dioxane–water (70 : 30 v/v) and thf–water (70 : 30 v/v), respectively because of their insolubility in other solvents such as water or dmso–water mixtures in a wide pH range. Comparing the protonation constants of L2 in water 16 and L1 in 1,4-dioxane–water (70 : 30 v/v) it can be noted that there is a reduction of the basicity of the first three protonations and an enhancement of the basicity for the last three (see Table 3).A similar eVect can be observed when we compare the protonation constants of L4 in water with those obtained for L3 which behaves as a stronger acid than L4 for the first two protonation steps but as a strong base when the Table 1 Selected bond lengths (Å) and angles (8) for [CdL1][PF6]2? 4CH3CN?4H2O Cd]N(1) Cd]N(3) N(3)]C(9i) N(1)]C(7) N(1)]C(1) N(2)]C(6) N(1i)]Cd]N(1) N(1)]Cd]N(3) N(1)]Cd]N(2i) N(3)]Cd]N(2) C(7)]N(1)]C(1) C(7)]N(1)]Cd C(1)]N(1)]Cd C(4)]N(2)]C(2) C(4)]N(2)]Cd C(2)]N(2)]Cd C(5)]N(3)]Cd C(5)]N(3)]C(9i) C(9i)]N(3)]Cd 2.338(8) 2.488(9) 1.500(13) 1.465(14) 1.520(13) 1.491(13) 180.0 102.6(3) 105.6(3) 73.7(3) 110.1(9) 108.2(7) 113.6(6) 108.6(9) 105.6(7) 115.0(7) 107.9(6) 110.5(9) 105.3(6) Cd]N(2) N(3)]C(5) N(3)]C(3) N(1)]C(8) N(2)]C(4) N(2)]C(2) N(1i)]Cd]N(3) N(1)]Cd]N(2) N(3)]Cd]N(2i) C(7)]N(1)]C(8) C(8)]N(1)]C(1) C(8)]N(1)]Cd C(4)]N(2)]C(6) C(6)]N(2)]C(2) C(6)]N(2)]Cd C(5)]N(3)]C(3) C(3)]N(3)]Cd C(9i)]N(3)]C(3) 2.535(10) 1.490(14) 1.503(14) 1.511(13) 1.456(13) 1.491(13) 77.4(3) 74.4(3) 106.3(3) 110.9(9) 109.0(9) 104.9(6) 111.0(9) 109.6(8) 107.0(7) 111.1(9) 112.2(7) 109.7(9) Symmetry transformation used to generate equivalent atoms: i 2x 1 1, 2y 1 1, 2z 1 1.Table 2 Selected bond lengths (Å) and angles (8) for compound L3 Fe(1)]C(15) Fe(1)]C(16) Fe(1)]C(17) Fe(1)]C(13) Fe(1)]C(10) N(1)]C(6) N(1)]C(1) Fe(2)]C(25) Fe(2)]C(23) Fe(2)]C(24) Fe(2)]C(21) Fe(2)]C(22) N(2)]C(2) C(6)]N(1)]C(1) C(10)]C(1)]N(1) C(3)]N(2)]C(2) C(2)]N(2)]C(4) N(2)]C(3)]C(5) 1.98(2) 1.97(2) 1.975(14) 2.06(2) 2.08(2) 1.470(12) 1.473(14) 2.00(3) 2.02(2) 2.019(13) 2.035(14) 2.06(2) 1.481(13) 112.0(9) 112.6(10) 114.2(10) 111.0(10) 112.8(12) Fe(1)]C(19) Fe(1)]C(18) Fe(1)]C(14) Fe(1)]C(11) Fe(1)]C(12) N(1)]C(5) Fe(2)]C(26) Fe(2)]C(28) Fe(2)]C(29) Fe(2)]C(27) Fe(2)]C(20) N(2)]C(3) N(2)]C(4) C(6)]N(1)]C(5) C(5)]N(1)]N(1) C(3)]N(2)]C(4) N(2)]C(2)]C(20) N(1)]C(5)]C(3) 1.99(2) 1.990(14) 2.00(2) 2.050(14) 2.05(2) 1.491(14) 1.99(2) 1.99(2) 2.00(2) 2.03(2) 2.048(12) 1.46(2) 1.488(13) 110.9(10) 109.6(10) 109.9(12) 116.0(10) 113.6(11) last two protonations are compared (see Table 4).For instance, the diVerence between the logarithms of the first and the sixth protonation constants of L2 in water is 9.15, whereas this diVerence for L1 is only 5.77 in 1,4-dioxane–water.Two accumulated factors could have an eVect on the observed behaviour: (i) the use of diVerent solvents (water and mixtures of water–thf or water–1,4-dioxane) and (ii) the functionalisation with ferrocenyl groups which transform secondary amines into tertiary ones. To study the influence of the solvent the protonation constants of L4 have also been determined in thf–water (70 : 30 v/v), see Table 4. It can be noted that a reduction of the basicity behaviour for the first two protonations is also found when compared with the protonation constants obtained for L4 in water, whereas the last two protonation constants remain similar.Comparison between L1 and L2 in 1,4-dioxane–water (70 : 30 v/v) cannot be carried out owing to the fact that only four protonation constants can be determined for L2 (see Table 3). In contrast all six protonation constants were determined for L2 in water. This diVerent behaviour may be explained by taking into account that an important reduction of the relative permittivity, and therefore an enhancement of the electrostatic repulsion between ammonium groups, occurs in 1,4-dioxane– water (70 : 30 v/v) when compared with water or thf–water (70 : 30 v/v).Metal co-ordination Solution studies directed to the determination of the stability constants for the formation of complexes of L1 and L3 with Cu21 have been carried out in 1,4-dioxane–water (70 : 30 v/v, 0.1 mol dm23 KNO3) for L1 and in thf–water (70 : 30 v/v, 0.1 mol dm23 [NBun 4][ClO4]) for L3 (see Table 5).To evaluate the eVect that the bulky ferrocenyl groups and the solvent have on L1, potentiometric titrations of L2 with copper in 1,4-dioxane– water (70 : 30 v/v) have also been carried out. The use of 1,4-dioxane–water mixtures significantly reduces the stability constants with copper (see Table 5). For instance the formation constant of [CuL2]21 in 1,4-dioxane–water is about 108 times smaller than that in water (log K = 24.40).16 Additionally, the logarithms of the first and second protonation constants of [CuL2]21 in 1,4-dioxane–water {[CuL2]21 1 H1 [Cu(HL2)]31 and [Cu(HL2)]31 1 H1 [Cu(H2L2)]41} are Table 3 Stepwise protonation constants (log K) of L1 and L2 determined at 25 8C in 0.1 mol dm23 KNO3 in 1,4-dioxane–water (70 : 30 v/v) Reaction L 1 H HLc HL 1 H H2L H2L 1 H H3L H3L 1 H H4L H4L 1 H H5L H5L 1 H H6L L1 9.64(3) 8.33(2) 6.71(3) 6.01(2) 4.52(2) 3.87(2) L2 a 9.56(3) 8.97(3) 5.46(2) 1.76(3) —— L2 b 10.15 9.48 8.89 4.27 2.21 1.0 a This work, 1,4-dioxane–water (70 : 30 v/v), 25 8C, 0.1 mol dm23 KNO3.b Data from ref. 16, in water. c Charges have been omitted for clarity. d Values in parentheses are standard deviations of the last significant figure. Table 4 Stepwise protonation constants (log K) of L3 and L4 determined in thf–water (70 : 30 v/v) at 25 8C in 0.1 mol dm23 [NBun 4][ClO4] Reaction L 1 H HLc HL 1 H H2L H2L 1 H H3L H3L 1 H H4L L3 8.90(1) d 7.56(1) 6.24(1) 3.50(1) L4 a 10.54(1) 9.49(1) 2.01(2) 2.21(4) L4 b 11.59 10.62 1.61 2.42 a This work, thf–water (70 : 30 v/v), 25 8C, 0.1 mol dm23 [NBun 4][ClO4].b Data from ref. 16 in water at 25 8C, 0.5 mol dm23 KNO. c Charges have been omitted for clarity. d Values in parentheses are standard deviations of the last significant figure.J. Chem. Soc., Dalton Trans., 1998, Pages 2635–2641 2639 smaller than those of the second and third protonation constants of the free receptor suggesting that L2 is acting as hexadentate in the [CuL2]21 complex.In contrast the logarithms of the first and the second protonation constants of [CuL1]21 in 1,4-dioxane–water are higher than those of the third and fourth protonation constants of L1 suggesting that not all the nitrogen atoms are involved in the co-ordination to the copper(II) ion. This is also suggested by the smaller stability constant found for [CuL1]21 in relation to that of [CuL2]21 in 1,4-dioxane– water.This result appears to stand in contradiction with the crystal structure of L1 with Cd21 (see above) which shows L1 Fig. 3 Distribution diagram for the L1–H1–Cu21 system Fig. 4 Distribution diagram for the L3–H1–Cu21 system Fig. 5 Concentration of Pb21 and Fe21 in the aqueous phase using L1 as extractant and of Pb21 and Fe21 in the aqueous phase after oxidation of the ferrocenyl groups acting as hexadentate. However it has to be noted that both a diVerent solvent and a diVerent temperature have been used for the potentiometric studies and the synthesis of [CdL1][PF6]2.Compound L3 only forms mononuclear complexes with Cu21. Fig. 4 shows the distribution diagram for the L3– H1–Cu21 system. The logarithm of the constant of the first protonation step {[CuL3]21 1 H1 [Cu(HL3)]31} is smaller than that of the third ([H2L3]21 1 H1 [H3L3]31) of the free receptor in thf–water (70 : 30 v/v) suggesting a M]N cleavage upon protonation and indicating that the number of nitrogens co-ordinated in [CuL3]21 is four.In fact the crystal structure of [CuL3][ClO4]2 shows the L3 receptor acting as tetradentate.15 However the presence of four ferrocenyl groups imposes some constraints that makes the formation constant of [CuL3]21 108 times smaller than that of [CuL4]21 (log K = 27.2) 17 although it is quite similar to that of [Cu(Me4cyclam)]21 (log K = 18.3, Me4cyclam = 1,4,8,11-tetramethyl-1,4,8,11-tetraazacyclotetradecane). Solvent extraction Compound L1 is the analogous receptor to the well known L2 molecule but with an important change in the solubility properties because of the presence of the lipophilic ferrocenyl units.Many ligands with lipophilic substituents have been reported to be able to solubilise transition-metal ions in organic solvents but to our knowledge no work has been devoted to this property using ferrocene-functionalised receptors. To study the lipophilic properties of the ferrocenyl groups we have carried out solvent extraction experiments of the cations Ni21, Cu21, Zn21, Cd21 and Pb21 typically present in water, in dichloromethane using L1 as extractant. Experiments have been carried out with and without [NBun 4][PF6] in the organic phase.The results obtained are in Table 6. They show a selective preference for Cu21 and extraction eYciencies follow the order Cu21 > Pb21 > Cd21 > Ni21 > Zn21; Cu21 is quantitatively extracted with an extraction eYciency up to 99% followed by Pb21 which shows an eYciency of 85%.Additionally we have observed that the presence of PF6 2 anions in the organic phase enhances the distribution coeYcient. For instance extraction Table 5 Stability constants (log K) for the formation of copper(II) complexes of L1 and L2 in 1,4-dioxane–water (70 : 30 v/v) at 25 8C in 0.1 mol dm23 KNO3 and of L3 in thf–water (70 : 30 v/v) at 25 8C in 0.1 mol dm23 [NBun 4][ClO4] Reaction Cu 1 L 1 3H Cu(H3L) a Cu 1 L 1 2H Cu(H2L) Cu 1 L 1 H Cu(HL) Cu 1 L CuL Cu 1 L 1 H2O CuL(OH) 1 H H 1 CuL Cu(HL) H 1 Cu(HL) Cu(H2L) H 1 Cu(H2L) Cu(H3L) CuL 1 H2O CuL(OH) 1 H L1 31.60(3) b 26.6(3) 20.45(2) 13.46(2) 5.52(3) 6.99 6.15 5.00 27.94 L2 — 24.42(3) 21.67(2) 16.75(2) 9.63(2) 4.92 2.75 — 27.12 L3 —— 22.48(2) 19.06(2) 7.64(3) 3.42 —— 211.4 a Charges have been omitted for clarity.b Values in parentheses are standard deviations of the last significant figure. Table 6 Extraction eYciency (E) and distribution coeYcient (D) for extraction of Ni21, Cu21, Zn21, Cd21 and Pb21 metal ions from water using L1 as extractant in dichloromethane M21 Ni21 Cu21 Zn21 Cd21 Pb21 E(%) 44 99 15 64 85 D 0.8 99.0 0.8 1.8 5.32640 J.Chem. Soc., Dalton Trans., 1998, Pages 2635–2641 Table 7 Values of DE2� 1 /mV for receptors L1 and L3 in their interaction with metal ions at five diVerent pH values L1 a L3 b pH 3.0 4.5 6.0 7.5 9.0 Ni21 <5 <5 <5 <5 <5 Cu21 <5c <5 10 12 8 Zn21 10 12 11 16 14 Cd21 10 14 8 16 14 Pb21 <5 <5 <5 <5 <5 Ni21 10 10 <5 <5 <5 Cu21 38 41 58 87 90 Zn21 18 18 33 52 52 Cd21 <5 9 18 31 31 Pb21 <5 7 <5 <5 <5 a 1,4-Dioxane–water (70 : 30 v/v), 25 8C, 0.1 mol dm23 KNO3.b thf–water (70 : 30 v/v), 25 8C, 0.1 mol dm23 [NBun 4][ClO4]. eYciencies without PF6 2 in the organic phase for Cu21 and Pb21 were 85 and 47%, respectively. Extraction appears to be due to the presence of the lipophilic ferrocenyl sphere which encapsulates a hydrophilic inner binding domain [see Fig. 1(b)] as suggested by the fact that the lower extraction eYciencies were observed when the similar ferrocene-containing tetraazacycloalkane L3 (containing four ferrocenyl groups and four N-donor sites) is used as extractant under similar conditions (for example E values for L3 with Cu21 and Pb21 were 34 and 22%, respectively). However the smaller number of N-donor sites and the smaller size of the macrocyclic cavity can also be factors which could explain the diVerent extraction eYciency of L3 when compared with L1.Metals extracted in the organic phase can usually be backextracted using acidic water solutions. However the presence in L1 of redox units suggests that a new method to remove the metal from the organic solution to water can be used. As the oxidation of the ferrocenyl groups decreases the aYnity of the ligands for metal ions, due to electrostatic reasons, a switching mechanism could occur and the uptake or release of the central metal ion could be controlled by oxidising or reducing the ferrocenyl groups.Towards this goal we have performed oxidations on dichloromethane solutions containing the L1 complex of Pb21 in contact with an aqueous phase. Fig. 5 shows that after 0.5 h the extraction of lead into the organic phase is complete. Subsequent oxidation of the ferrocenyl groups shows, after 1 equivalent of electrons, an increase of the Pb21 concentration in water but further oxidation to six electrons makes [Pb21] decrease again.This behaviour could be explained by assuming that a chemical reaction is coupled to the electrochemical one as suggested by the fact that the iron concentration increases in the aqueous phase as the oxidation in the organic phase progresses, indicating that some kind of decomposition of L1 takes place preventing back extraction of Pb21 to water. This is in line with the chemical instability observed for the oxidised form of L1 obtained by electrolysis of dichloromethane solutions of L1.Further studies are being carried out using other redox-active systems. Electrochemical behaviour The main interest in the L1 and L3 receptors is the incorporation of redox centres near binding sites. This gives the ability to recognise substrates electrochemically. Part of the electrochemical behaviour of L3 has been published elsewhere.18 The electrochemical study has been performed in 1,4-dioxane–water (70 : 30 v/v) for L1 (25 8C, 0.1 mol dm23 KNO3), and thf–water (70 : 30 v/v) for L3 (25 8C, 0.1 mol dm23 [NBun 4][ClO4]).The redox potential of L1 and L3 is, as expected, pH-dependent. When the pH was decreased a steady displacement of E2� 1 to more anodic potentials was observed. The diVerence found between the oxidation potential at basic (12) and acidic pH (0) obtained by extrapolation of the curves E2� 1 versus pH was 67 and 68 mV for L1 and L3 respectively. We have also electrochemically studied the variation of E2� 1 versus pH for the L–H1–M21 systems (L = L1 or L3; M = Ni21, Cu21, Zn21, Cd21 or Pb21; M21 :L molar ratio = 1 : 1).In order to rationalise all these data the diVerence found between E2� 1 for the receptor–metal system and that for the free receptor [DE2� 1 = E2� 1 (receptor–metal) 2 E2� 1 (receptor)] has been monitored at five diVerent pH values from 3 to 9.To screen the electrochemical receptor response for each metal ion, DE2� 1 has been measured by rotating-disc electrode experiments (scan speed 10 mV s21, rotation speed 7000 revolutions min21); the values found at diVerent pH values for receptors L1 and L3 are shown in Table 7. Compound L1 shows a poor shift of E2� 1 (DE2� 1 < 20 mV), whereas L3 is selective for Cu21, Zn21 or Cd21 in a wide pH range over Ni21 and Pb21. One of the most promising applications of redox-active systems would be the development of modified electrodes and their use as amperometric sensors.The shift of E2� 1 in redox systems has been suggested as a method for considering that there is a recognition process, but another approach could be the monitoring of the current at a fixed potential. We have checked this possibility using receptor L3 and Cu21 and measured the current at the fixed potential of 455 mV vs. SCE (pH 7.0) when increasing amounts of Cu21 were added to a solution of L3.The current decreases when Cu21 is added until a 1 : 1 molar ratio is reached. A linear relationship is observed between the intensity and the Cu21 concentration from [Cu21]/[L3] = 0 to 0.9. Table 8 shows the quantitative determination of Cu21 from the decrease in current at a fixed potential in thf–water (70 : 30 v/v). The concentration is compared to that obtained from standard methods. The selectivity in Table 7 can be confirmed by the fact that Cu21 can also accurately be determined in the presence of Pb21.As was observed by one of the referees, there is no clear relation between the electrochemical receptor response for metal ions and the extraction power of L1 and L3. The receptor response of L3 is better than that of L1 but L1 is a better extractant. This has to be related with the better lipophilicity of L1 when compared with that of L3. In fact both processes, metal extraction and electrochemical response, are quite diVerent.The parameter DE2� 1 does not give necessarily a good evaluation of the receptor aYnity for cations. We have found, at least in aqueous solution, that a selective electrochemical response along with large DE2� 1 values can be related with (i) the presence of predominant receptor–metal complexes in a wide pH range (see below), (ii) the existence of pH ranges of selective complexation.15 The oxidation potential of L1 and L3 decreases when thy we can understand this behaviour by Table 8 Determination of the concentration of Cu21 in the presence of Pb21 with receptor L3 (pH 7.0) from current measurements at the fixed potential of 455 mV vs.SCE 104 [Cu21]/mol dm23 0.69(6) a [0.74] b 1.3(2) [1.3] 1.6(2) [1.4] 0.79(6) c [0.73] 1.3(2) [1.4] 1.6(2) [1.7] a Determined by electrochemical methods. Values in parentheses are the standard deviations of the last significant digit. b Determined by standard methods. c Determined in the presence of [Pb21] = 2.6 × 1024 mol dm23.J.Chem. Soc., Dalton Trans., 1998, Pages 2635–2641 2641 taking into account that a steady change from highly charged species (more diYcult to oxidise due to electrostatic repulsion between the positively charged electrode and the positively charged species) to neutral species is found (from HnLn1 to L). A similar steady change from 15 charged species [Cu(H3L)]51 to 11 species [CuL1(OH)]1 was observed in the L1–H1–Cu21 system (see the distribution diagram in Fig. 3). In contrast, L3 forms the complex [CuL3]21 with Cu21 which is predominant (see distribution diagram in Fig. 4). That behaviour is reflected in the electrochemistry of the L3–Cu21 system in which E2� 1 does not change when the pH changes over a wide range (see Fig. 6 where a comparison between the E2� 1 vs. pH curves for the L1– H1–Cu21 and L3–H1–Cu21 systems is shown). Taking into account that a steady shift of E2� 1 is observed for polyamines when the pH changes, the formation of predominant species with metal ions over a wide pH range leads to large DE2� 1 values.In fact the highest shift found in the presence of transition-metal ions was for the L3 receptor with Cu21 (DE2� 1 = 90 mV). In order to evaluate the role played by the water, the electrochemical recognition ability of receptors against transitionmetal ions has also been studied in aprotic solvents such as acetonitrile. Protonation of amine groups is not possible in this solvent and we have found that the oxidation potential shift, DE2� 1 , was approximately always the same for all metal ions.DiVerences of selectivity in aqueous solution have therefore to be attributed to the rich chemistry (with a large number of species in solution as described above in the potentiometric experiments) of these systems in the presence of water. Conclusion The new redox-active polyazacycloalkane L1 and the related ferrocene-functionalised L3 receptor have been characterised.Metal extraction into dichloromethane from water using L1 as extractant has been carried out. The idea of back extraction to Fig. 6 Plots of E2� 1 vs. pH for (a) L1–H1 and L1–H1–Cu21, (b) L3–H1 and L3–H1–Cu21 water by oxidation of the ferrocenyl groups has been tested, but the chemical instability of the oxidised form of L1 prevents back-extraction processes. However the idea of redoxswitchable extraction is an appealing one and further work is now in progress.Additionally L1 and L3 have the ability to recognise metal ions electrochemically. The combination of electrochemical and potentiometric techniques appears to indicate that the existence in solution of predominant receptor–metal complexes over a wide pH range (as was found for the [CuL3]21 complex) can lead to a large electrochemical response. Acknowledgements We should like to thank the Dirección General de Investigación Científica y Técnica (proyecto PB95-1121-C02-02) for support.References 1 P. D. Beer, D. Hesek, J. Hodacova and S. E. Stokes, J. Chem. Soc., Chem. Commun., 1992, 270; P. D. Beer, M. G. B. Drew, C. Haslewood, D. Hesek, J. Hodacova and S. E. Stokes, J. Chem. Soc., Chem. Commun., 1993, 229; B. Belavaux-Nicot, Y. Guari, B. Donziech and R. Mathieu, J. Chem. Soc., Chem. Commun., 1995, 585; P. D. Beer, Chem. Commun., 1996, 689; P. D. Beer, A. R. Graydon, A. O. M. Johnson and D. K. Smith, Inorg. Chem., 1997, 36, 2112. 2 R. E. Wolf and S. R. Cooper, J. Am. Chem. Soc., 1984, 106, 213; H. Bock, B. Hierholzer, F. Vogtle and G. Hollman, Angew. Chem., Int. Ed. Engl., 1984, 23, 57; D. A. Gustowski, M. Delgado, V. J. Gatto, L. Echegoyen and G. W. Gokel, J. Am. Chem. Soc., 1986, 108, 7553; A. Kaifer, L. Echegoyen, D. A. Gustowski, D. M. Goli and G. W. Gokel, J. Am. Chem. Soc., 1983, 105, 7168. 3 P. D. Beer, A. D. Keefe, H. Sikanyita, C. Blackburn and J. McAleer, J. Chem. Soc., Chem. Commun., 1990, 3289; P.D. Beer, Z. Chen, M. G. B. Drew and A. J. Pilgrin, Inorg. Chim. Acta, 1994, 25, 137. 4 C. Dusemund, K. R. A. S. Sandanayake and S. Shinkai, J. Chem. Soc., Chem. Commun., 1995, 333; H. Yamamoto, A. Ori, K. Ueda, C. Dusemund and S. Shinkai, Chem. Commun., 1996, 407; Z. Chen, A. R. Graydon and P. D. Beer, J. Chem. Soc., Faraday Trans., 1996, 92, 97. 5 M. J. L. Tendero, A. Benito, R. Martínez-Máñez, J. Soto, J. Paya, A. J. Edwards and P. R. Raithby, J. Chem. Soc., Dalton Trans., 1996, 343; M.J. L. Tendero, A. Benito, R. Martínez-Máñez, J. Soto, E. García-España, J. A. Ramirez, M. I. Burguete and S. V. Luis, J. Chem. Soc., Dalton Trans., 1996, 2923; M. J. L. Tendero, A. Benito, R. Martínez-Máñez and J. Soto, J. Chem. Soc., Dalton Trans., 1996, 4121. 6 J. M. Lloris, R. Martínez-Máñez, T. Pardo, J. Soto and M. E. Padilla-Tosta, Chem. Commun., 1998, 837; P. D. Beer, Z. Chen, M. G. B. Drew, A. O. M. Johnson, D. K. Smith and P. Spencer, Inorg.Chim. Acta, 1996, 143. 7 P. D. Beer, M. G. B. Drew and A. R. Graydon, J. Chem. Soc., Dalton Trans., 1996, 4129. 8 T. Saji and I. Kinoshita, J. Chem. Soc., Chem. Commun., 1986, 716. 9 T. J. Atkins, J. E. Richman and W. F. Oettle, Org. Synth., 1978, 58, 86. 10 P. D. Beer, J. E. Nation, S. L. W. McWhinnie, M. E. Harman, M. Hursthouse, M. I. Ogden and A. H. White, J. Chem. Soc., Dalton Trans., 1991, 2485. 11 D. Lednicer and C. R. Hauser, Org. Synth., 1960, 40, 31. 12 P. Gans, A.Sabatini and A. Vacca, J. Chem. Soc., Dalton Trans., 1985, 1195. 13 SHELXTL, version 5.03, Siemens Analytical X-Ray Instruments, Madison, WI, 1994. 14 L. F. Lydon, The Chemistry of Macrocyclic Ligand Complexes, Cambridge University Press, Cambridge, 1989. 15 M. J. L. Tendero, A. Benito, J. Cano, J. M. Lloris, R. Martínez- Máñez, J. Soto, A. Edwards, P. R. Raithby and A. Rennie, J. Chem. Soc., Chem. Commun., 1995, 1643. 16 A. E. Martell, R. M. Smith and R. M. Motekaitis, NIST Critical Stability Constants of Metals Complexes Database, Texas A & M University, College Station, 1993. 17 M. Kodama and E. Kimura, J. Chem. Soc., Chem. Commun., 1975, 891. 18 M. E. Padilla-Tosta, R. Martínez-Máñez, T. Pardo, J. Soto and M. J. L. Tendero, Chem. Commun., 1997, 887. Received 30th March 1998; Paper 8/02394HJ. Chem. Soc., Dalton Trans., 1998, Pages 2635–2641 2641 taking into account that a steady change from highly charged species (more diYcult to oxidise due to electrostatic repulsion between the positively charged electrode and the positively charged species) to neutral species is found (from HnLn1 to L).A similar steady change from 15 charged species [Cu(H3L)]51 to 11 species [CuL1(OH)]1 was observed in the L1–H1–Cu21 system (see the distribution diagram in Fig. 3). In contrast, L3 forms the complex [CuL3]21 with Cu21 which is predominant (see distribution diagram in Fig. 4). That behaviour is reflected in the electrochemistry of the L3–Cu21 system in which E2� 1 does not change when the pH changes over a wide range (see Fig. 6 where a comparison between the E2� 1 vs. pH curves for the L1– H1–Cu21 and L3–H1–Cu21 systems is shown). Taking into account that a steady shift of E2� 1 is observed for polyamines when the pH changes, the formation of predominant species with metal ions over a wide pH range leads to large DE2� 1 values. In fact the highest shift found in the presence of transition-metal ions was for the L3 receptor with Cu21 (DE2� 1 = 90 mV).In order to evaluate the role played by the water, the electrochemical recognition ability of receptors against trano been studied in aprotic solvents such as acetonitrile. Protonation of amine groups is not possible in this solvent and we have found that the oxidation potential shift, DE2� 1 , was approximately always the same for all metal ions. DiVerences of selectivity in aqueous solution have therefore to be attributed to the rich chemistry (with a large number of species in solution as described above in the potentiometric experiments) of these systems in the presence of water.Conclusion The new redox-active polyazacycloalkane L1 and the related ferrocene-functionalised L3 receptor have been characterised. Metal extraction into dichloromethane from water using L1 as extractant has been carried out. The idea of back extraction to Fig. 6 Plots of E2� 1 vs. pH for (a) L1–H1 and L1–H1–Cu21, (b) L3–H1 and L3–H1–Cu21 water by oxidation of the ferrocenyl groups has been tested, but the chemical instability of the oxidised form of L1 prevents back-extraction processes. However the idea of redoxswitchable extraction is an appealing one and further work is now in progress.Additionally L1 and L3 have the ability to recognise metal ions electrochemically. The combination of electrochemical and potentiometric techniques appears to indicate that the existence in solution of predominant receptor–metal complexes over a wide pH range (as was found for the [CuL3]21 complex) can lead to a large electrochemical response.Acknowledgements We should like to thank the Dirección General de Investigación Científica y Técnica (proyecto PB95-1121-C02-02) for support. References 1 P. D. Beer, D. Hesek, J. Hodacova and S. E. Stokes, J. Chem. Soc., Chem. Commun., 1992, 270; P. D. Beer, M. G. B. Drew, C. Haslewood, D. Hesek, J. Hodacova and S.E. Stokes, J. Chem. Soc., Chem. Commun., 1993, 229; B. Belavaux-Nicot, Y. Guari, B. Donziech and R. Mathieu, J. Chem. Soc., Chem. Commun., 1995, 585; P. D. Beer, Chem. Commun., 1996, 689; P. D. Beer, A. R. Graydon, A. O. M. Johnson and D. K. Smith, Inorg. Chem., 1997, 36, 2112. 2 R. E. Wolf and S. R. Cooper, J. Am. Chem. Soc., 1984, 106, 213; H. Bock, B. Hierholzer, F. Vogtle and G. Hollman, Angew. Chem., Int. Ed. Engl., 1984, 23, 57; D. A. Gustowski, M. Delgado, V. J. Gatto, L. Echegoyen and G. W. Gokel, J. Am. Chem. Soc., 1986, 108, 7553; A. Kaifer, L. Echegoyen, D. A. Gustowski, D. M. Goli and G. W. Gokel, J. Am. Chem. Soc., 1983, 105, 7168. 3 P. D. Beer, A. D. Keefe, H. Sikanyita, C. Blackburn and J. McAleer, J. Chem. Soc., Chem. Commun., 1990, 3289; P. D. Beer, Z. Chen, M. G. B. Drew and A. J. Pilgrin, Inorg. Chim. Acta, 1994, 25, 137. 4 C. Dusemund, K. R. A. S. Sandanayake and S. Shinkai, J. Chem. Soc., Chem. Commun., 1995, 333; H. Yamamoto, A. Ori, K. Ueda, C. Dusemund and S. Shinkai, Chem. Commun., 1996, 407; Z. Chen, A. R. Graydon and P. D. Beer, J. Chem. Soc., Faraday Trans., 1996, 92, 97. 5 M. J. L. Tendero, A. Benito, R. Martínez-Máñez, J. Soto, J. Paya, A. J. Edwards and P. R. Raithby, J. Chem. Soc., Dalton Trans., 1996, 343; M. J. L. Tendero, A. Benito, R. Martínez-Máñez, J. Soto, E. García-España, J. A. Ramirez, M. I. Burguete and S. V. Luis, J. Chem. Soc., Dalton Trans., 1996, 2923; M. J. L. Tendero, A. Benito, R. Martínez-Máñez and J. Soto, J. Chem. Soc., Dalton Trans., 1996, 4121. 6 J. M. Lloris, R. Martínez-Máñez, T. Pardo, J. Soto and M. E. Padilla-Tosta, Chem. Commun., 1998, 837; P. D. Beer, Z. Chen, M. G. B. Drew, A. O. M. Johnson, D. K. Smith and P. Spencer, Inorg. Chim. Acta, 1996, 143. 7 P. D. Beer, M. G. B. Drew and A. R. Graydon, J. Chem. Soc., Dalton Trans., 1996, 4129. 8 T. Saji and I. Kinoshita, J. Chem. Soc., Chem. Commun., 1986, 716. 9 T. J. Atkins, J. E. Richman and W. F. Oettle, Org. Synth., 1978, 58, 86. 10 P. D. Beer, J. E. Nation, S. L. W. McWhinnie, M. E. Harman, M. Hursthouse, M. I. Ogden and A. H. White, J. Chem. Soc., Dalton Trans., 1991, 2485. 11 D. Lednicer and C. R. Hauser, Org. Synth., 1960, 40, 31. 12 P. Gans, A. Sabatini and A. Vacca, J. Chem. Soc., Dalton Trans., 1985, 1195. 13 SHELXTL, version 5.03, Siemens Analytical X-Ray Instruments, Madison, WI, 1994. 14 L. F. Lydon, The Chemistry of Macrocyclic Ligand Complexes, Cambridge University Press, Cambridge, 1989. 15 M. J. L. Tendero, A. Benito, J. Cano, J. M. Lloris, R. Martínez- Máñez, J. Soto, A. Edwards, P. R. Raithby and A. Rennie, J. Chem. Soc., Chem. Commun., 1995, 1643. 16 A. E. Martell, R. M. Smith and R. M. Motekaitis, NIST Critical Stability Constants of Metals Complexes Database, Texas A & M University, College Station, 1993. 17 M. Kodama and E. Kimura, J. Chem. Soc., Chem. Commun., 1975, 891. 18 M. E. Padilla-Tosta, R. Martínez-Máñez, T. Pardo, J. Soto and M. J. L. Tendero, Chem. Commun., 1997, 887. Received 30th March 1998; Paper 8/02

ISSN:1477-9226    

DOI:10.1039/a802394h

出版商:RSC    

年代:1998

数据来源: RSC