The Synthesis, Electrochemistry and Reactivity of Binuclear Copper(I) Complexes as Mimics of Protein Active Sites Citation Kreh, Robert Paul (1981) The Synthesis, Electrochemistry and Reactivity of Binuclear Copper(I) Complexes as Mimics of Protein Active Sites. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/ffc7-y238. https://resolver.caltech.edu/CaltechTHESIS:10292025-224718353 Abstract A series of binuclear copper(I) complexes has been synthesized and characterized as potential models for binuclear copper protein sites. The first group of compounds was prepared by the condensation of 2-hydraxy-5-methylisophthalaldehyde with a variety of primary amines (RNH 2 ) followed by the addition of Cu(CH 3 CN) 4 BF 4 and pyrazolate (pz), 3, 5-dimethyl pyrazolate (Me 2 pz) and 7-azaindolate (aza) anions (X) to yield [Equation. See abstract in scanned thesis for details]. A crystal and molecular structural analysis of [Equation. See abstract in scanned thesis for details] (R = 2-(2'-pyridyl)ethyl) showed each copper bound to the phenoxide oxygen, as well as to one imine and one pyrazolate nitrogen. The pyridine nitrogens were not coordinated to copper, but an intermolecular Cu-Cu interaction (2.97 Å) was observed. Overall coordination about copper(I) is best described as highly distorted pyramidal, with long axial (Cu-Cu) coordination. The copper-copper interaction appears to be associated with a ca. 600 nm (solid state) absorption which is not present in solution or in the solid state for compounds with bulky "side arms" (R). All compounds are proposed to be three-coordinate in solution, but they have exhibited no tendency to bind additional ligands such as CO or pyridine. Most of the binuclear copper(I) complexes gave quasi-reversible electrochemical behavior, with two distinct one-electron processes. The reduction potentials were varied over a wide range by modifications of the sidearms (R) and the bridge (X), with the most positive potentials observed for R = t-butyl and X = 3, 5-dimethylpyrazolate (E f 1 = +O.239, E f 2 = + O.80 V vs. nhe). Biological implications of the observed reactivities and redox properties of these compounds are discussed. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Gagné, Robert R. Thesis Committee: Unknown, Unknown Defense Date: 2 September 1980 Record Number: CaltechTHESIS:10292025-224718353 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:10292025-224718353 DOI: 10.7907/ffc7-y238 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 17737 Collection: CaltechTHESIS Deposited By: Ben Maggio Deposited On: 13 Nov 2025 22:16 Last Modified: 13 Nov 2025 22:25 Thesis Files PDF - Final Version See Usage Policy. 26MB Repository Staff Only: item control page

THE SYNTHESIS, ELECTROCHEMI5TRY AND REACTIVITY OF BINUCLEAR COPPER(!) COMPLEXES AS M™1CS OF PROTEIN ACTIVE SITES Thesis by Robert Paul Kreh In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1981 (Submitted September 2, 1980) ii ACKNOWLEDGMENTS Many people made contributions toward the work presented in this thesis. I will no doubt forget to mention a few, but some who come to mind are listed below: My interest in chemistry was first generated by Lillian Abramson at Montgomery Blair High School (Silver Spring, Maryland). This interest was kindled by Professors Gynith Giffin, Thomas McGrath, Neil Potter and Robert Nylund at Susquehanna University (Selinsgrove, Pennsylvania). At Caltech, I have enjoyed the guidance of my research advisor, Robert Gagne, and I appreciate the freedom he has allowed me in making many decisions on my own. I also thank Dave Evans, Harry Gray and Sunney Chan for many helpful discussions. The Gagne research group has been great. In the early days, there were Judy Allison with her laugh, Carl Koval with his electrochemistry and Tom Sm~th with his great knowledge of inorganic chemistry. More recently, trials and tribulations have been shared with Cliff Spiro, George Lisensky, Mike Ingle, Bill Marritt, Dave Marks, Larry Henling, and Rich Durand. John Dodge and Mike McCool deserve a special word of thanks for their work on the crystal structures presented in this thesis. I am especially grateful to Pat Anderson, the group secretary, for her help and cooperation whenever it was needed. I would also like to thank Henriette Wymar for an excellent job of typing this thesis. iii The chemistry staff here at Caltech has been very helpful and understanding, and in particular, special thanks go to Jan Mitchell and Sally Muir of the analytical lab. I have enjoyed the support and companionship of many friends here at Caltech. Jan and Mark Mitchell have been the best of friends for the past three years. Their support was invaluable. Various members of the Chan research group, including Randy Morse, Tom Stevens and Gary Brudvig, have helped with social diversions ( ?) along the way. I also have enjoyed the friendships of Jim and Patty Eisenach and Sridhar. This thesis is dedicated to my parents, Dave and Louise, because they gave me the support and freedom to pursue whatever I chose in life and taught me to do it with the best of my ability. iv ABSTRACT A series of binuclear copper(!) complexes has been synthesized and characterized as potential models for binuclear copper protein sites. The first group of compounds was prepared by the condensation of 2-hydraxy-5-methylisophthalaldehyde with a variety of primary amines (RNH 2 ) followed by the addition of Cu(CH 3 CN) 4 BF4 and pyrazolate (pz), 3, 5-dimethyl pyrazolate (Me 2 pz) and 7-azaindolate (aza) anions (X) to yield C~ ISOIM(R) 2 (X). A crystal and molecular structural analysis of C~ ISOIM(Etpy) 2 (pz) (R = 2-(2' -pyridyl)ethyl) showed each copper bound to the phenaxide oxygen, as well as to one imine and one pyrazolate nitrogen. The pyridine nitrogens were not coordinated to copper, but an intermolecular Cu-Cu interaction (2. 97 A) was observed. Overall coordination about copper(!) is best described as highly distorted pyramidal, with long axial (Cu-Cu) coordination. The copper-copper interaction appears to be associated with a ca. 600 nm (solid state) absorption which is not present in solution or in the solid state for compounds with bulky "side arms" (R). All compounds are proposed to be three-coordinate in solution, but they have exhibited no tendency to bind additional ligands such as CO or pyridine. Most of the binuclear copper(!) complexes gave quasi-reversible electrochemical behavior, with two distinct one-electron processes. The reduction potentials were varied over a wide range by modifications of the sidearms (R) and the bridge (X), with the most positive potentials observed for R =!_-butyl and X = 3, 5-dimethylpyrazolate (E! = +O. 239, V E; = + O. 80 V vs. nhe). Biological implications of the observed reactivities and redox properties of these compounds are discussed. The reaction of N, N, N', N' -tetrakis-(2-pyridylmethyl)ethylenediamine (TPEN) with Cu(CH 3 CN) 4 BF4 yielded ClJi(TPEN)(BF4 ) 2 • A complete crystal and molecular structural analysis showed each copper to be bound to two pyridine nitrogens, and an intramolecular Cu-Cu interaction (2. 78 A) was found. The tertiary nitrogens do not appear to be coordinated to the coppers. C~ (TPEN)(BF4 ) 2 reacts with CO to yield C~ (TPEN)(C0) 2 (BF4 ) 2 , and the crystal and molecular structure of this dicarbonyl adduct was also determined. Each copper was found in similar distorted tetrahedral geometries, but the bond lengths and bond angles around each copper are slightly different. This is manifested in two carbonyl stretching frequencies. Investigation of the dioxygen reactions of C~ ISOIM(R) 2 (pz), where R = a variety of aJkyl, aromatic and heterocyclic groups, indicated ligand oxidation and pyrazolate dissociation upon oxygenation. As a result, the synthesis of new binuclear three-coordinate copper(!) complexes was proposed for investigation of their dioxygen reactivities. The proposed ligands were synthesized by the condensation of 2-hydroxy-5-methylisophthalaldehyde with 2-(2' -aminoethyl)pyridine and 2-(2' -aminoethyl)-1-methylimidazole, but the desired cuprous complexes were mt isolated in pure form. vi TABLE OF CONTENTS Page ~Pleil Introduction 1 ~ 16 Unusual Structural and Reactivity Types for Copper(!): Synthesis, . Structural and Redox Properties of Binuclear Copper(!) Complexes which are Probably Three-Coordinate in Solution and Experience Intermolecular Metal-Metal Interactions in Solid State ~ 67 The Synthesis and Structures of (N, N, N', N' -Tetrakis(2-Pyridylmethyl) Ethylenediamine) Di copper(!) and its Dicarbonyl Adduct ~ 112 Dioxygen Reactions of Three Coordinate Binuclear Copper(!) Complexes, Leading to the Design and Attempted Synthesis of Improved Models for Copper Proteins ~ Summary and Conclusions 139 Appendix I 144 ~ Investigation of a Reported One-Step Two-Electron Reduction vii Page ~~ An Approach to the Synthesis of Binuclear, 156 Tetrahedral Copper(!) Complexes ~ Abstracts 163 1. Metal Complexes Derived from Face-to-Face 164 1, 10-Phenanthroline Macrocycles 2. Gold Metallocycles 172 3. Models for Blue Copper Proteins 189 4. Macrocyclic Cuprous Complexes as Photosensitizers 198 5. Transition Metal Formyl Chemistry 208 CHAPTER I Introduction Hemocyanin, tyrosinase and Ia.cease are proteins which utilize a binuclear copper active site for a direct interaction with molecular oxygen. Hemocyanin is responsible for reversible oxygenation in the blue blood of various molluscs and arthropods. 1 Tyrosinase, found in a variety of organisms, catalyzes the oxidation of substrates by dioxygen (e.g., Scheme 1). 2 The Ia.ceases of Asian lacquer trees and Scheme 1 ~ H - 0 - ~ CH-i-CCX:>H ~ 2 I NH 2 2 H O p - ~CH-i-COOH + H 20 2 I NH H / 2 H 0 I H-C-COOH 2 I NH2 white rot fungi catalyze the reduction of oxygen to water (equation 1) (1) while oxidizing a variety of substrates .. 3 In addition to the binuclear copper site, laccase contains two other isolated copper ions ("type I" and "type II") which are also necessary for enzymatic activity. 2 These binuclear copper protein sites are especially intriguing because they consist of two cupric ions in close proximity which do not exhibit the paramagnetism typical of copper(II) ions. 4 - 6 This is seen in the diamagnetism and lack of EPR signals at room temperature. (The total E PR intensity of laccase accounts for only 50% of the copper, presumably due to the type I and type II copper present in the protein.) 6 Temperature-dependent magnetic studies indicate that the cupric ions of the binuclear sites in oxyhemocyanin and laccase are antiferromagnetically coupled. 4 The binuclear cupric site in laccase acts as a two-electron acceptor. 7 Oxidation-reduction titrations indicate that this site is reduced in a one-step, two-electron transfer (n = 2) but reoxidized in two one-electron steps (n = 1). This is rather novel behavior, since the electrochemistry of polynuclear complexes generally shows distinct one-electron processes _for interacting metals. 8 In fact, it appears that closer proximity and larger interactions lead to greater separations of the one-electron processes. 9 The potentials found for the reduction of the laccases were +O. 43 V for Rhus laccase and +O. 78 V for Polyporus lac case. 7 (Potentials are versus the normal hydrogen electrode, nhe). The reduction of both cupric ions of binuclear site 5 in tyrosinase also occurs at a single potential (0. 36 V vs. nhe). The reoxidation of tyrosinase has not been reported,nor have electrochemical measurements on hemocyanin been reported. The reduction potentials measured for these binuclear sites are very positive for copper(II) ions, indicating the stabilization of the 3 cuprous state relative to the cupric state. On the contrary, the reduction potentials for synthetic Cu(II) complexes with nitrogen and oxygen ligands tend to be quite low, usually well below the potential for the normal hydrogen electrode. 1O Soft donors such as sulfur ligands may be proposed to account for the high reduction potentials exhibited by these copper proteins. l l, 12 However, spectroscopic and titration studies indicate the absence of sulfur ligands and the presence of nitrogen ligands such as imidazole. 13 The above discussion presents several interesting questions about these binuclear copper sites. In addition to the direct study of the proteins, the investigation of appropriate model compounds can yield useful information about the questions posed below. 1) What are the ligands bound to the copper centers? Protein studies have indicated that the majority of ligands may be imidazole nitrogens from histidine. 13 The fallowing structure. and ligand environment, 1, was proposed for oxyhemocyanin by Bosnich and -- 4 coworkers on the basis of electronic spectra. 11 The phenolate oxygen atom, the peroxide ligand, the thioether ligand and an imidazole nitrogen occupy an approximately square-planar array around each copper. This configuration was arrived at by the comparison of the electronic absorption spectra of several synthetic Cu(II) complexes to that of axyhemocyanin. This model, however, contradicts other evidence against the presence of sulfur atoms. 13 Similar models have been recently proposed by T. G. Spiro and coworkers for oxyhemocyanin and deoxyhemocyanin (shown below). 14 N I 1 cu N Oxyhemocyanin. \. N Deoxyhemocyanin The models are based upon extended X-ray absorption fine structure (EXAFS) and resonance Raman (RR) spectroscopy of hemocyanin. In these models, the nitrogen ligands are from imidazole rings (histidine), and X is suggested to be a phenolate oxygen from tyrosine. The resonance Raman studies imply bridging peroxide, i.e. , the transfer of one electron from each Cu(I) upon oxygenation. The above model for oxyhemocyanin is essentially the same as an earlier proposal by Solomon and coworkers, based upon electron paramagnetic resonance 5 (EPR) and electronic absorption spectroscopy of a series of hemocyanin derivatives. 15 2) What is the nature of the Cu-Cu interaction? The strong antiferromagnetic coupling between the two cupric ions of oxyhemocyanin is presumably due to superexchange through an endogenous ligand (tyrosine oxygen) and/ or the bridging peroxide. 4, 15 Strong antiferromagnetic interactions have been observed in a mnnber of binuclear copper(II) complexes. 8, 16 For example, compound 2 is diamagnetic, - 2 exhibiting sharp NMR absorptions and weak EPR signals. 16 Hence, it is not unreasonable to propose strong interactions through a bridging phenolate oxygen in oxyhemocyanin. 3) What is the origin of the unusual electrochemical behavior, including the very high reduction potentials? While the majority of synthetic complexes containing closely interacting metal centers exhibit stepwise reduction and oxidation of the metal centers, compound~ (above) appears to be an exception. Both coppers in i are reported to be reduced and oxidized at a single potential. 16 An investigation of this system is presented in Appendix I. The very 6 positive reduction potentials exhibited by the binuclear cupric sites in la.cease and tyrosinase are still curious. The potentials observed for synthetic copper(II) complexes employing nitrogen and oxygen ligands are usually quite negative, due to the relative stabilization of the oxidized Cu(II) state by these "hard" ligands. 8 , lO This dilemma will be addressed in this thesis. That is, what type and geometry of oxygen and nitrogen ligands could lead to the stabilization of the cuprous state and generate these high reduction potentials? 4) What factors are responsible for reversible 0 2 binding and 0 2 activation? Marginal success has been achieved in obtaining reversible oxygenation of synthetic cuprous complexes. The fallowing complex, ~' isolated by Osborn et al. , was found to react with oxygen, 2 + 3 possibly leading to reversible oxygenation in the solid state. 17 This reaction has yet to be well characterized by the authors. Reversible oxygenation (in solution) of the mononuclear Cu(I) complex, 1, has been reported by Wilson and coworkers. 18 Only preliminary results 7 have been communicated, but stoichiometry measurements imply a bridging 0 2 between two coppers in the oxygenated species. While a number of reports have appeared describing copper catalyzed autoxidation of organic substrates, the relationships between oxygen activation and coordination environment around the copper ions remains unclear. 19 For this purpose, cuprous complexes containing a well defined ligand environment are desired. Polydentate (including macrocyclic) ligands are useful in this regard. The autoxidation of macrocyclic ligands bound to copper(!) has recently been investigated in our laboratory. 20 An understanding of oxygen activation is relevant to fuel cell technology. 21 The reduction of oxygen to water (equation 1) is the cathodic reaction involved in a fuel cell. For a fuel cell to become an efficient device for energy conversion, a catalyst must be found for this reduction. By adequately "mimicking" the binuclear copper protein sites, a synthetic catalyst for the reduction of oxygen to water 8 may eventually emerge. The work presented in this thesis involves the synthesis, characterization and reactivity of a series of binuclear copper complexes as potential models for the binuclear copper protein sites. Although a large number of binuclear copper complexes are known, most of these have involved the higher (+2) valence state of copper. To understand the inherent stability and diaxygen reactions of the reduced protein sites, binuclear Cu(I) complexes should be investigated. A few binuclear Cu(I) complexes (e.g., 31! and 522 ) have been recently - - .2..+ 5 reported, but these have employed sulfur ligands for the stabilization of the cuprous centers. Since sulfur ligands are most likely not present in the binuclear copper protein sites, 13 it is desirable to use only nitrogen and oxygen ligands in model compounds. The problem is, then, to stabilize the +1 oxidation state of copper while utilizing nitrogen and oxygen ligands. The two main factors which appear to be involved in the relative stabilization of Cu(II) and Cu(I) are summarized in Table 1. 8 , lO, 23 9 Table 1. Preferences of Copper Valence states ~ Cull (cf) a) octahedral (distorted), square-pyramidal or square-planar geometry tetrahedral, trigonal or digonal geometry b) "hard" sigma donors (amines, H2O, OH-, NH 3 ••• ) "soft" donors, particularly those capable of 1T backbonding (CO, CN-, SI¼, P~, aromatic nitrogen ... ) Most of the binucleating ligands employed to date are much more conducive to Cu(II) than Cu(I). Inorganic chemists appear to be enamored with amines, imines, alcohols, ethers, ketones and aldehydes for use as ligands. These ligands are to be avoided if stabilization of Cu(l) is desired. Also, macrocyclic ligands are often used because they can promote unusual structural types while minimizing substitution reactions around the metal when the complexes are in solution. 8 , 24 Yet these macrocycles tend to be unfavorable for Cu(I) due to the square-planar environment of ligands often dictated by the macrocycle. - These factors are demonstrated by the binuclear compound, 6, which has received attention in several laboratories, including ours. 8 This binuclear complex exhibits a two-step reduction at -0. 52 V and -0. 91 V versus nhe. These very negative potentials indicate the destabilization of Cu(I) relative to Cu(II) in this ligand environment. 1.0 6 In an effort to circumvent the above problems of Cu(I) destabilization, binucleating ligands which contain "softer" aromatic nitrogen ligands in a more flexible framework (to allow tetrahedral geometry around both coppers) were designed. (Also, aromatic type nitrogens begin to approach the ligands (i.e., histidine) suggested for the binuclear copper protein sites.) These modifications on the macrocyclic ligand system of§_ have yielded a very stable copper(!) complex, - 7. A tetrahedral array' of oxygen and nitrogen ligands around each 7 11 copper was expected, but X-ray structural analysis showed only three ligands bound to each copper and in addition, an intermolecular Cu-Cu interaction was observed. This Cu-Cu interaction appears to be associated with an electronic absorption band at 600 nm. The reduction potentials were varied systematically over a wide range by modifications in the ligand system, and the more positive values approach the high potentials exhibited by the copper proteins. These results along with biological implications are presented in Chapter II. The investigation of binuclear copper(!) complexes was extended - - to compound 8 and its dicarbonyl adduct 9. Chapter III describes the 8 9 synthesis, characterization and X-ray structural analysis of each of these compounds. Again, an unexpected coordination environment was observed for the binuclear cuprous system, !!• . Each copper 12 appears to bind only two nitrogen ligands (from the pyridine rings) in addition to the observed Cu-Cu interaction. The "expected" pseudo- - tetrahedral geometry of ligands was observed for 9, but the two copper centers are not equivalent, and this is manifested in two different carbonyl stretching frequencies. Chapter IV presents an investigation into the dioxygen reactions of binuclear cuprous compounds such as 7. The results led to the design and attempted synthesis of new binuclear three-coordinate -- -- copper systems, 10 and 11. Although the ligand syntheses were 10 11 accomplished, the copper(!) complexes were not isolated in pure form. All of the work presented in this thesis was done by this author with the exception of the X-ray structural analyses, and these were - - - performed by John Dodge (7 and 8) and Mike McCool (9). 13 References and Notes (1) Lontie, R. in ''Inorganic Biochemistry", Eichhorn, G. I. ed.; Elsevier: New York, N. Y., 1973; p. 344. (2) (a) Mason, H. S. Ann. Rev. Biochem. ~ ' 35, 595-634. (b) Vanneste, W. H. and Zuberbilhler, A. in "Molecular Mech_a nisms of Oxygen Activation", Hayaishi, 0. ed. ; Academic Press: 1974; p. 371. (3) (a) Fee, J. A. Structure and Bonding llTI_, 23, 1-60. (b) Vanngard, T. I. in "Biological Applications of Electron Spin Resonance Spectroscopy", Swartz, H. M. ; Bolton, J. R. ; and Borg, D. C., eds.: J. Wiley: New York, N. Y., 1972, p. 411. (4) Solomon, E. I.; Dooley, D. M.; Wang, R.; Gray, Ho B.; Cerdonio, M. ; Magno, F.; Romani, G. L. J. Amer. Chem. Soc. 1976, -98, 1029-1031. ~ (5) Makino, N. ; McMahill, P. ; Mason, H. S. ; Moss, T. H. J. Biol. Chem. ll.11, 249, 6062-6066. (6) Malmstrom, B.G.; Reinhammer, B. R.; Vanngard, T. Biochem. Biophys. Acta~' 205, 48. (7) . 0 (a) Reinhammar, B. R. M.; Vanngard, T. Eur. J. Biochem . .!171, 18, 463-468. Reinhammar, B. R. M . . Biochem. Biophys. Acta -1972, ~275, - 245-259. (b) Farver, O. Goldberg, M.; Lancet, D.; Pecht, I. Biochem. Biophys. Res. Commun. ~ ' 73, 494-500. (8) Gagne, R. R.; Koval, C. A.; Smith, T. J. J. Amer. Chem. Soc. - 1977, 99, 8367-8368. Gagne, R.R.; Koval, C. A.; Smith, T. J.; ~ 14 References (continued) ~ Cimolino, M. C. J. Amer. Chem. Soc. ~ ' 101, 4571-4580. (9) (a) Creutz, C.;.Taube, H. J. Amer. Chem. Soc. ~ ' 91, 3988-3989. Creutz, C.; Taube, H. J. Amer. Chem. Soc. ~ ' 95, 1086-1094. (b) Weaver, T. R.; Meyer, T. J.; Adeyemi, S. A.; Brown, G. M.; Eckberg, R. P.; Hatfield, W. E.; Johnson, E. C.; Murray, R. W.; Untereker, D. J. Amer. Chem. ---- - Soc. 1975, 97, 3039-3048. - (c) Callahan, R. W.; Keene, F. R.; Meyer, T. J.; Salmon, D. J. J. Amer. Chem. Soc. llTI_, 99, 1064-1073. (10) Patterson, G. S.; Holm, R. H. Bioinorg. Chem. ~ ' !, 257-275. (11) Admundsen, A. R.; Whelan, J.; Bosnich, B. J. Amer. Chem. Soc. 1977, 99, 6730-6739. ----- (12) Potentials in the range of +0. 3 to +0. 9 V (vs. nhe) have been reported for copper(II) complexes of polythiamacrocycles: Dockal, E. R.; Jones, T. E.; Sokol, W. F.; Engerer, R. J.; Rorabacher, D. B.; Ochrymowycz, L. A. J. Amer. Chem. Soc. 1976, -98, 4322-4324. ~ (13) See Wurzbach, J. A.; Grunthaner, P. J.; Dooley, D. M.; Gray, H. B. ; Grunthaner, F. J. ; Gay, R. R. ; Solomon, E. I. ---- - J. Amer. Chem. Soc. 1977, 99, 1257-1258, and references contained therein. (14) Brown, J. M.; Powers, L.; Kincaid, B.; Larrabee, J. A.; ---- Spiro, T. G. J. Amer. Chem. Soc. 1980, 102, 4210-4216. - 15 References (continued) Larrabee, J. A.; Spiro, T. G. J. Amer. Chem. Soc. ~ 1980, 102, 4217-4223. (15) Eickman, N. C.; Himmelwright, R. S.; Solomon, E. I. Proc. Natl. Acad. USA -- -Sci. --1979, ~ -76, 2094-2098. (16) Lintvedt, R. L.; Tomlonovic, B.; Fenton, D. E.; Glick, M. D. Adv. Chem. Ser. ~ ' 150, 407-425. (17) Bulkowski, J. E.; Burk, P. L.; Ludmann, M. F.; Osborn , J. A. J. Chem. Soc., Chem. Commun. fil'I, 498-499. (18) Simmons, M. G.;Wilson, L. J. J. Chem. Soc.,Chemo Commun. 1978, 634-636. ~ (19) For example, see Brackman, W.; Havinga, E. Recueil, l.955, ~ 74, 1021-1039. Jameson, R. F.; Blackburn, N. J. J. Inorg. Nucl. Chem. ~ ' 37, 809-814. Urbach, F. L.; Knopp, U ..; Zuberbuhler, A. D. Helv. Chim. Acta~' 61, 1097-1106. Munakata, M., Nishibayashi, S.; Sakamoto, H. J. Chem. Soc., Chem. Commun. 1980, 219-220. ~ (20) Lisensky, G. C. , Ph.D. Dissertation, California Institute of Technology, 1981. (21) Bockris, J.; Shrinivasan, S. "Fuel Cells: Their Electrochemistry", McGraw Hill, New York, 1969. (22) Alberts, A. H.; Annunziata, R.; Lehn, J. M. J. Amer. Chem. Soc . .!J!.ll, 99, 8502-8504. (23) James, B. R.; Williams, R. J. J. Chem. Soc. 1961, 2007-2019. (24) Gagne, R. R.; Allison, J. L.; Gall, R. S.; Koval, C. A. ~ J. Amer. Chem. Soc. 1977, _, 99 7170-7178. ~ 16 CHAPTER II Unusual structural and Reactivity Types for Copper(!): Synthesis, structural and Redax Properties of Binuclear Copper(!) Complexes which are Probably Three-Coordinate in Solution and Experience Intermolecular Metal-Metal Interactions in the Solid state. Robert R. Gagne, Robert P. Kreh and John A. Dodge Contribution No. 5645 from the Division of Chemistry and Chemical Engeineering, California Institute of Technology, Pasadena, California 91125. 17 Introduction ~ Protein binuclear copper sites effect remarkable reactions with dioxygen including reversible binding (hemocyanin), 1 activation (tyrosinase) 2 and reduction (laccase). 3 structural information contrasting these active sites is limited but similarities are notable: sulfur ligands have been proposed but most studies suggest only nitrogen and/or oxygen coordination; 4 in the oxidized forms all three binuclear sites are strongly antiferromagnetically coupled; 5 ' 6 the tyrosinase and laccase binuclear sites exhibit two electron reductions at potentials which are rather high for the proposed all nitrogen/oxygen copper coordination. 6 , 7 Model studies have addressed ligand environment(s), redox properties, magnetic interactions and dioxygen binding in these protein active sites. B-l 2 To help define protein structure/reactivity relationships we are endeavouring to catalogue fundamental copper(!) coordination chemistry in relatively simple mononuclear and binuclear complexes. As discussed elsewhere, polydentate ligands, including macrocycles, can be utilized to minimize problems associated with both copper(II) and copper(!) substitution !ability. 13 , 14 This approach has resulted in unusual structural and reactivity types for copper including four- and five-coordinate copper(!) species. 13 - 15 The - binuclear complex 1 provided an opportunity for meauring intramolecular electron transfer rates between copper centers, but the - reduction potentials observed for 1 (equations 1 and 2) are significantly more negative than those exhibited by the proteins (e.g., +0.36 V for 18 n=0.1.2 2+ CuII CuII L + e -0. 52 V (1) -0. 91 V (2) tyrosinase). 7 ' 14 Modification of the macrocyclic complex, 1, has - - yielded a series of binuclear complexes 2, in which the R and X groups X= 2 \ I N-N V 19 were varied systematically, resulting in a wide range of reduction potentials. In this manner, utilizing only nitr~en and oxygen ligands, high reduction potentials comparable to that of protein binuclear sites have been achieved. The new binuclear copper(!) complexes appear to be only three-coordinate in solution but in some instances experience significant intermolecular copper(I)-copper(I) interactions in the solid state. Synthesis and Characterization of Complexes -- Three copper(II) complexes, 3-5, were prepared by simple condensation of 2-hydraxy-5-methylisophthalaldehyde with primary amines followed by addition of copper(II), Scheme I. The new compounds are blue-green due to weak absorption in the 630-645 nm region (probably ligand field transitions) as listed in Scheme I for methanol solutions. Solid state magnetic susceptibility measurements at 25 ° C, also listed in Scheme I, suggest that the two copper(II) centers are antiferromagnetically coupled. Analogous air sensitive copper(!) complexes were prepared in a helium atmosphere in a fashion similar to the synthesis of the copper(II) species. The dialdehyde was condensed with two equivalents of a primary amine, RNH 2 , then treated with a bridging bidentate ligand, XH, in the presence of mse, followed by addition of Cu(CH 3 CN) 4 BF4 (Scheme II). The compounds synthesized, along with the abbreviations to be used, are listed in Table I. All of the compounds listed gave satisfactory C, Hand N analyses. Selected compounds were also analyzed for copper, giving the expected values. The infrared spectra 20 Scheme I. Preparation of Binuclear CuIIcuII. Complexes. (Spectra ~ were recorded in methanol solution at 25 ° C. Magnetic susceptibilities are given in B. M. , measured at 25 ° C, and corrected for diamagnetism, but not for T. I. P.) 2+ 0 0 0 H R 3 4 -o~ -CH 2 A (E) JJ-eff/Cu 630(160) 1.4 645 (120) 1.2 630 (100) 0.96 N- -CH 2CH-0~ N- NH 5 -CH2CH2 -I:) N Table Compound Nuinber ~ I. Binuclear Cu(I)-Cu(I) Complexes, Abbreviations Utilized and NMR Data. Sldeana (R) Bridge (X) Nl1R Ass lgr-.ients1 Abbreviation al Q2 a 3 81 Bz PY1 PZz PZ3 Z-(pyrldyl )-methyl pyrazolate Cu(I)2IS0IM(Mepy)2(pz) 2.05 6 . 66 7. 56 4.60 8.46 8.02 6 . 23 2-(pyrldyl )-aethyl 3-5 dl111ethylpyrazolate Cu(l)2IS0JM(Mepy)z(Me2pz) 2 .06,s 6.69,s 7.72,s 4.67,s 8.37,d 2.13,s/b 6 . 21 7-aulndolate Cu(I ) 1SO!M(Mepy) 2 (aza) 2 2 .09,s 6 . 72 7. 54 4. 70,s 2-(2'-pyrldyl)-ethyl pyrazolate Cu( I) z I S0IM( Etpy) z( pz) 1.92,s 6 . 47 1. 50 3.82,t 3.29,t 8.42,d 7 .87 6.47 10 2-(2' - pyrldyl )-ethyl 3-5 dimethylpyrazolate Cu(l) JS0JM(Etpy) 2 (Me 2pz) 2 1 . 94,s 6.48,s 1. 52,s 3.88,t 3.37,t 8.40,d 2.46,s/b 6.17 11 2- ( 2' -pyridy !)-ethyl 7-azalndolate Cu( I) 1s0111( Etpy) 2 (aza) 2 1 . 96,s 7 . 60 3 . 44, 3 . 90,t { 3.20 12 pheny IIlle thy I pyrazolate Cu( l)2IS0ll1(MePh)2(pz) 2. 10,s 6.72 7 .54 4. 26 8.00 6.26 13 phenylinethyl 3-5 dimethylpyrazohte Cu(I) 1S0111(MePh) (He pz) 2 2 2 2.10,s 6.70 7.57 4.37 2.06,s/b 9 14 phenyl111ethyl 7- azalndolate 15 2-phenylethyl pyrazolate 16 2-phenylethyl 3-5 dimethylpyrazolate 17 2-phenylethyl 7-azalndolate 18 n-propyl pyrazolate 19 1-propyl pyrazolate 20 t-butyl pyrazolate 21 t-butyl 3-5 dimethyl pyr1zolate Cu(I) IS0IH(MePh) (aza) 2 2 Cu( I) JS0IM ( EtPh) (pz) 2 2 2 . 14,s 8.39 6 . 77 6.51 7.37 3.45,t 3.07,t 7.93,d 6 . 51 Cu(I) IS0IM(EtPh) {He pz) 2 2 2 1.97,s 6.52,s 7.38,s 3.50,t 3.10,t 2.61,s/b 6.17 Cu( I) IS0IM(EtPh) (aza) 2 2 Cu( I )2 1S0111( 1-Pr) (pz) 2 1. 98, s 6. 53,s 7. 46,s 3.18, 3.52,t { 2.97 2 . 09,s 6.70,s 7.57,s 3. 15,t 7 . 70,s 2.90,m 7 .87,d 1.20,d 7 .87 ,d 6 . 45,t 2 . 10,s 2.13,s 6 .85,s 7. 91,s 1. 26,s 7 . 93,d 6.54,t Cu(I)zlS0IM(!.-Bu)2(He2pz) 2.10,s 6.81,s 7 . 85,s 1 . 28,s 2.57,s/b 6.10,s 7. 75,d 6 . 33 23 phenyl pyrazolate 24 p-(d1methylam1no)phenyl pyrazolate Cu( I) I S0111( PhNMe2) 2( pz) 2 25 p-acetyl phenyl pyrazolate Cu( Il21S0111(PhCOMe)2(pz) 7-azalndolate 2. 14,s 6.85,s 7.98,s 2.04,s 6.67 7 .52 2.07 8 . 39 e.U,d 8.36,d t,.) 6 . 43, t 1.78,m Cu( I) JSOIM(2 - Pr) (pz) 2 2 Cu( I) I S0IM(!_-Bu) 2 ( pz) 2 Cu(I) IS0IM(!_-8u) 2(aza) 2 Cu( I ) IS0IM(Ph) (pz) 2 2 t-butyl 8.10 ~ l.34,. 22 IZl6 4.40, [ 4 . 32 1.96,s 6.70,s IU3 8.43,d 2-(pyrldyl ) -111ethyl ~ IZlz 8. U,4 { 1.23, s 6.75,d 8.45,d 7. 55 ~All spectra were obtained in d 6 benzene solution at 34°C. The values listed are given in ppm, 6, relative to TMS. The assignments for the resonances are given in Figure 2 and as follows: a1 = protons from the methyl group on the aromatic ring arising from 2-hydroxy-5-methylisophthalaldehyde; a2 = aromatic protons from the 2-hydroxy-5-methylisophthalaldehyde ring; a3 = the imine protons; 81 = methylene protons (or methine proton) adjacent to the imine nitrogen; 82 = methylene (or methyl) protons two carbons away from the imine nitrogen; py1 • the protons on the carbon adjacent to the pyridine nitrogen; pz2 = protons in the 2 and 4 positions on the pyrazole ring; pz3 m proton in the 3 position on the pyrazole ring; aza2, aza3, aza6 = the protons on the 2,3 and 6 positions, respectively, on the 7-azaindole bridge. S = singlet, d = doublet, t = triplet, m = multiplet. If not listed in the Table, the multiplicity could not be determined due to broadness or overlapping of the peaks. E..These peaks are assigned to the methyl groups attached to the pyrazole ring. £These peaks could not be discerned because of overlapping peaks and/or limited solubility. 22 Scheme II. ~ A 0 0 H l)CH3CN, fl 0 + 2RNH2 2)XH,2No0CH3 3) 2Cu(CH3CN)4 BF4 of all compounds showed the absence of aldehydic carbonyl and amine N-H stretches, consistent with complete imine condensation. The multitude of peaks in the region 700 - 1640 cm - i made absolute assignments difficult. Mass spectra were obtained for compounds "16, 19, "20, 21, and -" "-" -" "-" 22, each of which exhibits parent ion peaks in agreement with the "-" formulation given above.. The presence of two coppers in these compounds is indicated by three parent peaks due to the presence of both 63 Cu and 65 Cu. The relative intensities of these three peaks, averaged over all the compounds analyzed were 1 :0. 90:0. 24 for 63 63 65 Cl.½L: Cu65 CuL: Cl.½L (theoretical ratio is 1 :0. 89:0. 20). The electronic absorption spectrum of cu!rsOIM(EtPh) 2 (pz), 1.§., shown in Figure 1 (curve B), is typical of the solution spectrum of all the binuclear copper(!) complexes examined. Spectra obtained in the solid state (nujol mulls) varied. The solid state spectrum of cu!ISOIM(EtPh) 2 (pz), ll, shown in Figure 1 (curve A), is typical of 23 B w u z <! (I) ~ Cf) C (I) <! 400 500 600 700 WAVELENGTH (nm) 1 Figure 1. Electronic absorption spectra: (A), Cu 2 ISOIM(EtPh) 2 (pz), 15, in the solid state; (B), Cu1 ISOIM(EtPh) (pz), •15, in a hexane ~ .,.._,._ solution; and (C), 2 2 -- .,.._,._ cuJrsoIM(t-Bu) 2 (pz), 20, in the solid state. - 24 that for most compounds examined.· Ncte that the absorption present in solution at 444 nm (€ = 8600) shifts to lower energies, ca. 464 nm, in the solid state. In addition a new absorption was found in the solid state spectrum at ca. 600 nm. The 600 run absorption was net found in solution even in spectra of concentrated THF solutions, in which these compounds are considerably soluble. Not all species examined exhibit the solid state spectrum described above. Compounds .,..._,.._ 19, ,..._ 20, .,..._,.._ 21, and ,..._ 22 exhibit solid state spectra as shown for C~ISOIM{!_-Bu) 2 (pz), ~' in Figure 1 (curve C), which can be compared to the solution spectra of all complexes. Under no conditions have compounds ,..._ 19-22 ,..._ shown notable absorption in the 600 nm region. All cuprous compounds were shown to be diamagnetic by the presence of very sharp NMR resonances (see Figure 2). Table I lists the peaks observed for each complex. The assignments given were made by cross-compari~ons of all the spectra. In each case, the identity of the organic entity is confirmed. Integration values were in agreement with the proposed structures. ~ All compounds presented here are unreactive towards carbon monoxide. This was shown definitively for CO saturated solutions (CH 2 C¼) of C~ISOIM(Etpy) 2 (pz), i, and C~1iSOIM(l-Pr) 2 (pz), 18, ,..._ both of which show infrared absorptions in the region 1500-1650cm-i, - characteristic of the original complex, 9 or 18, but neither show any 1 ,..._ oonds in the region 1700-2200 cm- which would be attributable to a carbonyl complex. Also, evaporation of THF solutions of these compounds with a stream of CO gave the original compounds, as identified 25 N O N TMS a, /32 /'v'/'/~ Cu /3 Cu N-N benzene PZ2 ♦ a2 I /3 3 VPZ2 pz3 13, a3 /32 PZ3 J 10 ..... ....., -- - ~ 6 8 4 . 2 CHEMICAL SHIFT (ppm,8) ~ NMR spectrum of d6 benzene at 34 ° C. I Cu 2 ISO™(l-Pr) 2 (pz), ~' in J . ~ 0 26 by m spectra. All copper{!) compounds react with oxygen at the ambient temperature to form green or brown products from orange DMF solutions. This reaction could not be reversed by subsequent bubbling of argon through the solutions. The air sensitivity of these complexes prevented reliable solution molecular weight measurements. Electrochemistr~. ~ All electrochemical results presented here were obtained in DMF solutions at a platinum indicating electrode. Ferrocene was used as an internal standard and all potentials have been converted to versus nhe using Ef = +O. 40 mV vs. nhe for ferrocene. 16 The cyclic voltammograms for the Cu(II) compounds, 3-5, -- were irreversible as shown in Figure 3 for c~1ISOIM:(Mepy) 2 (0H)2+, ~. Addition of one equivalent of pyrazole to the same solution, however, gave the quasi-reversible electrochemistry also shown in Figure 3. 17 Similar results were observed for cuJ1sOIM:(Etpy) 2 {0H/+, 1· In con- trast cuJISOIM:(histh(OH/+, .§_, gave very irreversible electrochemistry even with added pyrazole. - Constant potential electrolysis (cpe) of 3 plus an equivalent amount of pyrazole at -0. 35 V and -0. 70 V indicated that each wave corresponds to a one-electron process. No attempt was made to isolate a mixed valence species, but the binuclear cuprous complex, cuJISOIM:(Mepy) 2 (pz), 2, was synthesized by cpe of an acetonitrile solution containing Cl½IIISOIM(Mepy) 2 (OH) +, i, and an equivalent 2 amount of pyrazole at -0. 70 V (n = 2. 0). The cyclic voltammogram for cuJrsOIM:(Mepy) 2 (pz), .§_, was, of - course, identical to that shown in Figure 3 for 3 plus one equivalent 27 _______ / I -1.0 -0.8 -0.6 -0.4 -0.2 0.0 I I 0.4 0.2 VOLTS (nhe) ~ Cyclic voltammograms of DMF solutions: CuII2_ ISOIM(Mepy)2(OH) 2+, 3, (----); cuII2ISOIM(Mepy)2(0H) 2+, 3, - containing one equivalent of pyrazole (--). - Both voltammograms were observed with a platinum electrode, using 0. lM TBAP as the electrolyte at a scan rate of 200 m V/ sec. 28 of pyrazole. All binuclear copper(!) compounds presented in Table II exhibited cyclic voltammetry similar to the solid line in Figure 3; however, large peak separations (Epa-EPc = 100-150 mVat 50 mV/sec) were observed for many compounds at a platinum electrode. 18 Cyclic voltammetry at a hanging mercury drop electrode (HMDE) gave smaller peak separations (70-90 mV) for most compounds, and the formal potentials (Ef) were identical to those observed using a platinum .e lectrode. The potentials recorded in Table II were determined at a platinum electrode because this electrode allowed direct comparison with the internal standard, ferrocene, on the same scan. It was found that the potentials could be read more accurately using differential pulse voltammetry, and this technique was employed to obtain the values presented in Table II. In all cases, the formal potentials were within 10 mV of the potentials observed by cyclic voltammetry. All compounds containing the 7-azaindole anion as the bridge exhibited irreversible electrochemistry, often showing two anodic waves but only ill-defined cathodic waves, and hence the electrochemistry of these species is not included here. Constant potential electrolysis at a potential 200 mV more pos'itive than the second oxidation peak was performed on compounds ~' 1, !Q, _!1, and ~' as typical Cu(I) species, and each gave a twoelectron oxidation (n = 2. 0 ± 0.1). The average separation of the two formal potentials (Ei-E;) is 194 ± 40 mV for all new compounds listed in Table II. From this value for E!-E;, the average comproportionation constant, K, is calculated to be 1. 9 x 103 for the equilibrium in equation 3, where L represents the entire ligand system. 14 Note that all complexes containing a 29 Table n. Reduction Potentials for Binuclear Copper Complexes.! Ef1 Cu(Il)Cu(Il)L +a + e- Cu(Il) Cu (I) L + Ef Cu(Il)Cu(I)L -+2 + eNo. 1 "' Compound Cu(Il) 2 (ISO™) 2 (Pr )2 +2 2 Cu(I)Cu(I) L Ef Q. Ef Q. d -0. 52- d -0.91- 2 l 2.0 6.,._ Cu(I\ISOm (Mepy)2 (pz) -0. 211 -0. 452 7 Cu(I) 2 ISOIM (Mepy) 2 (Me2 pz) -0.190 -0.374 9 Cu(I) 2 ISOIM (Etpy) 2 (pz) -0.110 -0.344 2.0 10. - Cu(I) 2 ISOIM (Etpy) 2 (Me2 pz) -0.113 -0.267 2.0 12 ,_,.,._ Cu(I) 2 ISOIM (Me Ph) 2 (pz) 0 146 -0.081 13 .,._.,._ Cu(I) 2 ISOIM (Me Ph )2 (M e 2 pz) 0.206 0.005 15 Cu(I) 2 ISOIM(EtPh) 2 (pz) 0.128 -0.078 16 Cu(I) 2 ISOIM (EtPh) 2 (Me2 pz) 0.205 0.009 18 .......,._ Cu(I) 2 ISOIM(l-Pr) 2 (pz) 0.146 -0. 076 19 Cu(I) 2 ISOIM (2-Pr )2 (pz) 0.193 0.001 1. 9 20 .......,._ Cu(I) 2 ISOIM(t-Bu) 2 (pz) 0.240 0.053 2.0 21 .,._...... Cu(I) 2 ISOIM (t-Bu) 2 (Me2 pz) 0.239 0. 080 23 .......,._ Cu(I) 2 ISOIM (Ph) 2 (pz) 0.144 -0.032 24 ............ Cu(I) 2 ISOIM (PhNMe2 \ (pz) 0.146 -0.048 25 ,_,.,._ Cu(I)2 ISOIM (PhCOMe )2 (pz) 0.152 0.008 "" ~hese values were measured by differential pulse polarography in DMF using a platinum indicating electrode. Q.potentials are given in volts~nhe. Potentials were measured vs. f errocene as an internal redox couple then corrected to vs. nhe using a value of 0. 40V for ferrocene vs. nhe. £These values for n were determined by constant potential electrolysis at a potential 200 mV more positive than Values are given for all complexes actually examined by cpe. ~Reference 13. Ei• 30 ·I I II II 2+ Cu Cu L + Cu Cu L (3) 3, 5-dimethylpyrazolate bridge have smaller formal potential separations (Ef - Ei) and correspondingly smaller comproportionation constants than their pyrazolate bridged analogues. Addition of excess pyridine (up to one hundred-fold excess) to the - - electrochemical solutions of 6, 12, and 13 did not cause a significant • ,.._ shift ( < 10 mV) in the potentials, but merely broadened the peaks. --- - Cyclic voltammetry of 6, 7, 9, and 10 in other solvents such as CH2 C~, CH3 CN and THF, gave similar behavior to that observed in DMF, while the electrochemistry for all ether compounds was irreversible in these alternative solvents. The crystal and molecular structure of ctJiISOTh1:(Etpy) 2 (pz), ~' was determined. Basic crystal data are summarized in Table III. Tables IV-VII present the atomic parameters and interatomic distances and angles. The atom labeling scheme is illustrated in Figure 4. The bonds to the copper atoms are shown in Figure 4. Each copper is bound to three ligand atoms in what is almost a tee geometry, the largest angles being 167. 6° (Nl-Cul-N3) and 170. 7° (N2-Cu2-N4). The intramolecular copper-copper distance, 3. 304 A, is sufficiently long that any direct interaction is unlikely. There still remains, however, the possibility of direct intermolecular interaction between copper atoms. The molecules form an 31 - Basic Crystal Data for Cu(I)JSOIM(Etpy)ipy), 9. FW = 565. 6 V = 2502. 3 (10) A.3 space group P2 1 /c (No. 14) Z =4 'A p calc = 1. 50 g cm- 3 b = 4. 929 (1) A = 23. 8 cm- 1 A(CuKJ = 1. 5418 A a = 20. 980 (4) 0 0 c = 24. 773 ( 6) A f3 = 102. 38 (2) 0 µ 0 9 '"0 J2C20 ~ C 2 6 ~C2 1, -,C25 c1s: 'I. ,w I ~ -.- L ··• .., t.'l'J I I I ~ c21 ~ C23 ~ ORTEP drawing of Cu12 ISOIM(Etpy) 2 (pz), ~' including atomic labeling scheme. Each hydrogen is given the same number as the carbon to which it is bound. ~ N) Table IV. ~ C-.1 t..J 01 NI Nl NJ N4 NS N6 ti u ti C4 u u JlG y & U717(Jl IIHOlS) l 2iS♦JU ♦) J7111U) 4Z920(3} 4401(1) J90~(1) 4994111 Cl t• CI0 Cll Cll en ..,o,,, 102 JIU]) 162411) ll 16(2) ZZZ9(2> Ul4(Z) 106J(7) J41lln lOSU(I) IUZCI> 1)1111) 141)(10, Hll HU 1 Nl4 -J(l) lUJ 5' U) 113(5) 94(3) 41Ul Jtt(J) ll llO) J0IIIUU )219()) lZlJUJ 5219(1) 55D1U) 5789(1.) S593t1l 5111 (J) 6l19U) 1499(3) 101,> B,A 1 4942(3) 113312) 1249(1) 211-tfl) IH(l) Jl98U> h801Zl 2955(3) JD54U) )610(4' 4166U) IIU(Z) 44)(♦) ]15(1) 291(1) 69(2) J8(1) 125(1) )62(2) ZHU) HICZ) 119(,, ..JZS2(9J 1 JSl(&J 1116(8) 1441(10) 411U s-t()) om SllJ(l) 320(Z) 2106(Z> 2176(1) 3531 (l) 2991(1) 2591 (1) 1257(2) IH412J JJ(J) 1111(1) 62](1) 595(1) 621(1) 592 ll) 56S(Z) 338(2) 2SJ(Z) 261(1) )11(2.) H6(1) UJUl 61 Ul 41(1) S-t U) J<t(41 ,o<i> 67(4J 44(3) !HU> 5237(2) OlfzJ 51J(z.) HIZ HU ism 58(.I) 6 l(J) l6(J) um 1Hl5J ,,,3) 16U> 11(4) H(3) 15('tJ 17(-t) ll-tl5J JZ (i) 141 (6) 92('1) S6U) 11 (3) 15(4) 50 (3) 63 (3) 6ZU) 106 (5) 147 (7) 114 ,,, 94 ,,, ,01,s U(J) 11(4) 136(6} -•nu -•m SO(Z.) 45(1) 45(2> 15]0(8) 1324(8> lllOUot 1008(11) 16400) 69J(8) 506(8} SH") HID IIU.) om 13(3) 4113(1) 4160(1) 030(1) 5Ul(2J 5624Ul 6012(3) 5645121 401(2) 451(2) 501CU 529 (2) 495(2) H9 -7(2) -4(1) -J (Z.) Sl514) um um 10001(8) 11901(8\ 11920(9) 11154(9) lOOHUO) 10164 (16) IJUOOl 1111(8) 6101(9) 139501) 1962(11) lL H6• . HJ Hl+) 2,,..) 6(1) 90 ('4) •\&U) 44(3) 46(5) 54(3) 6SU) JHl(IJ 1667(1) JUJU) 4lll 11) y H6 111()) 191JJ ZHSC1> 6244(1) )( H6' -53(-t) 62 (1) 64 (3) 1070(1' HJ H4 u,., 601'4'> 5851'1) 5 I (Z) 5'(3) "(3) Ui CH Cl9 ClO u,. !5HC-t> U(ll 13170) 1030(2' J9Zll) •atJ, Cl7 u.. Sm+) 02(4) !H(2) J6J On> C21 t22 ClJ C24 t2S CIS CH u._ 51(2.) 16S14f9J 15J9J(l1) 1 ns,no> ll162(U) 20JCJ7U3l 21U6UZ> 20198(Ul JUJ(9) 4646(11) 2604 (10) 2021(1'> 119(1 J) -I0511U) -l661lit tH Uu 41(2J SIOJ 11214 <•> uum u~ JI (2) 5111) uum 4090(.Z\ 016(2) Cl Atomic Parameters and ESD's for Cu(I)JSOIM(Etpy):a<pz), ~- U(3) -4(J) -1 (J) -1 Cl) -'4 (Z) 49()) 5012) 5lUl 65(3) -9 (2.) -4(2.) -1 (Z) -5 (3) 1(2.) H(3) 11l't) 48(3) 46(2) 49 (J) -J(Z) l (Z) -1 J (3) I U) 2 (5) H(I) 66(3) 67 (3) SJ fl> J6fiJ S 1 (5) 1Z(4) 53 (3) 60U) 80(4) 62 (3) 99 (+) H13) • (3) (2.) _, - • (J) -1 l fl) 16 (4) -1 (4) -20('1) - , (2J Hl4 1 Hl6 HU Hll Hl9 HZO H20' H2l H21 1 H2l H24 H25 H26 ,.oo ,.10 6.10 6.60 ,.10 S.40 5.40 1.20 em "(2) 2112> ll(S) 14(1) 16(1) ZZU.) 16(1) 10 (Z) zo (2.) sm -6l)) .22(5) 2 (J) 0(4) l7 (3) -l4UJ 4(2) 46 ("\) 85 ,,, 44 (+) z. 013) ·lli> 7 (3J •5(3) 0(4) -3616) -11 (5) 14(4) e,A 1 1.20 1.00 1.10 ,,tc,) 511(9) '"", 548(2.l 49J<ll 522(1) ,nm JZlm snm -J6(1) -UCJO) Z,(1) •UZUO) 6'1(1) Ut<Zl -Jl](ICJ} Hllll 3 2 _ , (3) uom Uij has been multiplied by Hf for the copper atoms and 10 for the remaining nonhydrogen 2 213) 228 (1) 166(1) 166(21 2H(l) 5 atoms. 0(1) 11]9(9) ~Fractional coordinates have been multiplied by factors of 10 for the copper atoms, 104 for the remaining nonhydrogen atoms, and to' for the hydrogen atoms. 6(]) Hl6UO> U lO(CJ) U02<9J 2057(9) 2U(81 216UI) IS (2) HU -4 UI 23 UJ u -4 (4) .,156(1) ·7 (3) -7(1) O(J) 22 (1) (3) -45(5) 378(2) 256(1) 261 (z.) JJ6(2) 460(2) 2 UW 16~(2) 101(2.) - JOU) •61)) _, (J) -J9 (Jl y -,m •)(l) 4(4) 17 (+) O(S) Ill) -J(I) 5 ()) l (2) · l (2J um 119 (6) 11 l+J - 7 UI J(t) 9(1) U(Z) 131 ,., 6(3) zm -2 (3) 2Z(Z) 4S (J) JS (3) 27UJ X 5.00 5.10 1.00 1.00 1.00 ll 113) ~The form of the thermal ellipsoid is exp(-211 (U 11 h ah+ • • • +2Uuklb*c*)] for the anisotropic thermal parameters. 1.10 l.40 ,.,o ' • JO 6.l0 6.30 1.00 9.10 1.50 ,.10 C-" C-" 34 Table V. ~ - Bond Distances (A) for Cu(I)JSOIM(Etpy)ipz), 9. Cul-O1 Cul-Nl Cul-N3 Cu2-Ol Cu2-N2 Cu2-N4 O1-Cl Nl-C3 Nl-C13 N2-C9 N2-C20 N3-N4 N3-C12 N4-Cl0 N5-Cl5 N5-C19 N6-C22 N6-C26 Cl-C2 Cl-CB · C2-C3 C2-C4 C4-C5 C5-C6 C5-C7 C7-C8 C8-C9 Cl0-Cll Cll-C12 C13-Cl4 Cl4-C15 Cl5-C16 • 2.089 1. 901 1.882 2.097 2) 1. 894 3) 1. 878 1.300 1. 281 5) 1. 482 5) 1.296 5) 1. 469 5 1.372 5 1. 355 6 1. 349 6 1.330 6 1. 353 7) 1. 331 6) 1. 354 8) 1. 424 1. 421 1.449 ll 1.399 5) 1. 378 (6) 1. 513 7) 1.366 6) 1. 413 1. 453 ~l 1. 355 7) 1.348 7) 1. 507 7l 1. 501 7 1. 363 (7 Il !~ C16-C17 C17-Cl8 Cl8-Cl9 C20-C21 C21-C22 C22-C23 C23-C24 C24-C25 C25-C26 C3-H3 C4-H4 C6-H6 C6-H6' C6-H6" C7-H7 C9-H9 Cl0-Hl0 Cll-Hll C12-H12 C13-Hl3 C13-H13' C14-H14 Cl4-H14' C16-Hl6 Cl 7-Hl 7 Cl8-Hl8 Cl9-H19 C20-H20 C20-H20' C21-H21 C21-H21' C23-H23 C24-H24 C25-H25 C26-H26 1. 391 1.344 1.345 1. 508 7) 1. 499 1. 354 il 1. 385 9) 1. 322 11) 1. 339 10) 1. 03 4) 0.90 4) 0.90 0. 85 0.90 0.96 0. 96 0. 95 !l 0.88 0.96 0.89 1.05 1. 01 5) o.0. 99 87 ~5l 4 0. 98 5) a. 91 98 !4l 0. 5 o. 99 4) 0. 97 !4) o. 95 4l 0. 90 4 o. 97 (4) 0. 99 !5) o. 97 5l 0. 96 5 il !I i Table VI. ~ - Bond Angles (deg.) for Cu(l)JSOIM(Etpy) 2 (pz), 9. O1-Cul-Nl O1-Cu1-N3 Nl-Cu1-N3 O1-Cu2-N2 O1-Cu2-N4 N2-Cu2-N4 Cul-O1-Cu2 Cul-O1-Cl Cu2-Ol-Cl Cul-N1-C3 Cu1-N1-C13 C3-Nl-C13 Cu2-N2-C9 Cu2-N2-C20 C9-N2-C20 Cu1-N3-N4 Cu1-N3-C12 N4-N3-C12 Cu2-N4-N3 Cu2-N4-C10 N3-N4-C10 C15-N5-C19 C22-N6-C26 O1-C1-C2 O1-C1-C8 C2-Cl-C8 C1-C2-C3 C1-C2-C4 C3-C2-C4 N1-C3-C2 C2-C4-C5 91. 7 1) 97. 1 1 167. 6 2 91. 5 1 96. 8 1 170. 7 2 104. 2 127. 5 127. 7 3) 126. 1 117. 3 116. 6 125. 5 120. 4 3 114. 1 4 120. 6 3 133. 1 3 106. 3 4 121. 2 3 131. 8 107. 0 116. 4 115. 9 121. 5 120. 2 4 118.4 4 124. 0 4 118. 6 4 117. 4 4 129. 2 4l 124. 0 (4 ~l· i! ii !i C4-C5-C6 C4-C5-C7 C6-C5-C7 C5-C7-C8 Cl-C8-C7 C1-C8-C9 C7-C8-C9 N2-C9-C8 N4-C10-C11 C10-C11-C12 N3-C12-C11 N1-C13-C14 C13-C14-Cl5 N5-C15-C14 N5-C15-C16 C14-C15-C16 C15-C16-C17 C16-C17-C18 C 17-C18-C 19 N5-C19-C18 N2-C20-C21 C20-C21-C22 N6-C22-C21 N6-C22-C23 C21-C22-C23 C22-C23-C24 C23-C24-C25 C24-C25-C26 N6-C26-C25 121. 8 (4) 116.9 4 121. 3 4 123. 4 4 118.8 4 124. 8 4 116.4 4 129. 0 4 110. 6 5 105. 2 5 110.0 5 109. 8 4) 112. 1 5 116. 2 4 122. 2 5 121. 5 5 119.6 5 118.4 6 119. 1 124. 3 110. 1 4 112. 8 4 116. 0 4 123. 0 5 121. 1 5 118.4 6 119. 5 7 119.3 7 123. 8 7 ~~ w t1I Table Vll. ~ Nonbonding Distances and Associated Angles for - Cu(I)JSOIM(Etpy)ipz), 9. Cul-Cu2 =I= Cu1-H13' O1-Cu1-Cu2++ N1-Cu1-Cu2+ N3-Cu1-Cu2 3. 3036 ( 4) 2. 64 (4) 95. 83. 72 il~ 1 104. 5 1 Cul-Cu2+ Cu2-H20',._ 2. 9685( 4) 3. 09 (4) * 105. 5 ~1~ 92. 1 1 O1-Cu2-Cul=t: N2-Cu2-Cu1:t= N4-Cu2-Cu1 + Translated by one unit cell in positive y direction * Translated by one unit cell in negative y direction 89. 6 (1) ~ 0) 37 infinite stack in the direction of the b axis, a section of which is shown in Figure 5. Although the separation between molecules in the direction of the b axis is 4. 929 A (just the length of the b axis), the molecule is far from perpendicular to this axis, the angle between the a.xis and the mean molecular plane being 41 °. Thus, "overlapping" portions of neighboring molecules come rather close to one another. The average distance of atoms 01, Cl, CS, C9, and N2 from the mean plane of atoms Nl, C3, C2, Cl, and 01 in the neighboring molecule just below is 3. 20 A. This is somewhat shorter than the separation usually observed between stacked 1r-delocalized molecules. 19 The intermolecular copper-copper distance is shorter yet, at 2. 968 A. A least squares plane calculation including the atoms of the benzene ring, the methyl carbon, and the phenolic oxygen shows these atoms to be coplanar, with no deviation from the mean plane exceeding 0. 009 A. The deviations of the copper atoms from this plane are 0. 04 A for Cul and 0. 29 A for Cu2. In each case the direction of the out-of-plane separation is that which reduces the intermolecular copper-copper distance. Each of the two C=N groups is also twisted slightly out of plane in the direction of the copper atom to which it is bound (deviations: O. 04 A for C3, 0. 05 A for Nl, O. 04 A for C9 and 0.10 A for N2). The atoms of the pyrazolate group are coplanar with no deviation from the mean plane exceeding 0. 006 A. The benzene and pyrazole rings are not quite coplanar, the dihedral angle being 170(3) 0 • There is also a slight deviation of each copper atom from the plane defined by the three atoms to which it is bound. These deviations ~ Crystal packing of CuI2 ISOIM(Etpy) 2 (pz), ~, showing a section of an infinite stack of molecules which forms in the solid state. The three molecules are corresponding ones from three different unit cells, translated along the b axis (which lies in the vertical direction in the orientation of this figure). All copper atoms lie in the same plane, parallel to the page. The pyrazolate groups are directed out of the page, toward the reader. The shortest copper-copper distance, 2. 97 A, occurs between atoms in neighboring molecules, connected by dashed lines in the figure. The two dotted lines correspond to two different intermolecular Cu-H distances, 2. 64 A (Cul-Hl3') and 3. 09 A (Cu2-H20' ). 0 0 0 . 39 ~ Stereoview illustrating crystal packing of Cu12 ISOIM(Etpy\(pz), - 9. The molecules depicted comprise more than one unit cell, although the edges of only one cell are shown. 40 are 0.14 A for Cul and 0. 06 A for Cu2. Again, in each case the direction of the out-of-plane separation is that which reduces the intermolecular copper-copper distance. It should be noted that a possible attraction for another copper atom is not the only factor favoring this direction for the out-of-plane separation. As shown in Figure 5, there is a hydrogen atom on the opposite side of each copper atom, and in the case of Cul this hydrogen is at a distance of 2. 64 A. (This is also the copper that is farther out of its donor-atom plane.) Thus, at least for Cul, copper-hydrogen steric interactions may influence the out-of-plane displacement of the copper ion. A final point which should be noted concerning the coordination about copper is that the copper-nitrogen bonds are shorter than most which have been reported for Cu(I) complexes with nitrogen donor ligands. 22 The average lengths are 1. 88 A for the copper-pyrazole nitrogen bonds and 1. 90_A for the copper-imine nitrogen bonds. The copper-oxygen distances are substantially longer than this, averaging 2. 09 A. Also, as stated previously, the N-Cu-N bond angles approach 180° (167. 6° and 170. 7°). These factors may suggest some degree of two-coordinate character, which may be related to the unusual chemical behavior of the compound (e.g. , unreactivity toward CO). Discussion ~ - The rather negative reduction potentials observed for complex 1 (equations 1 and 2) likely result from a square-planar arrangement of "hard" oxygen and nitrogen ligands. 25 The complexes reported here 41 were designed to stabilize copper(!) in an attempt to better model binuclear copper protein sites. For example, the condensation of 2-hydroxy-5-methylisopthalaldehyde with 2-aminomethy Ipyridine - resulted in the binucleating ligand in 3, which is more flexible than - the ligand system in 1 and also presents relatively soft pyridines for coordination. Two additional binucleating ligands providing even greater flexibility were · obtained using 2-(2' -aminoethyl)pyridine and histamine, and these ligands reacted readily with copper(II) saH:s to yield the binuclear copper(II) complexes 1, and 1· Complexes similar to 1 and 1 were reported during the course of the present study. 26 In all three copper(II) complexes, both copper atoms are presumably bound to an aromatic nitrogen. This has been demonstrated by an X-ray structure determination of c~11ISOI1\,1(hist) (OH) 2 +, 1· 27 2 Each copper(!!) is actually five-coordinate with one of the copper atoms bound to the oxygen of a water molecule and the other copper bound to the oxygen of the hydroxy bridge of an adjacent molecule. -- The electrochemistry of these non-macrocyclic complexes, 3-5, 2 is irreversible as shown in Figure 3 for c~11ISOIM(Mepy) (OH) +, ~. 2 This may be a result of the relatively labile hydroxy bridge, which is a poor ligand for copper(!). The introduction of pyrazole did produce reversible electrochemical behavior, also shown in Figure 3, leading to the isolation of a stable binuclear copper(!) complex, C~ISOIM(Mepy)ipz), 1, by constant potential electrolysis (n = 2). 17 The reduction potentials observed for C~ISOIM(Mepy) 2 (pz), 1, (Ef = -0. 21 V, E; = -0. 45 V) were appreciably more positive than those observed for - the macrocyclic complex, 1, indicating that the binucleating ligand in 42 - 6 does indeed provide a better environment for copper(!) relative to copper(II). The copper(!) compound ctJiISOIM:(Mepy)ipz), ~' was also syn- thesized directly from cuprous starting materials. Similar reactions with 2-aminomethylpyridine or 2-(2' -aminoethyl)pyridine as sidearms and with pyrazole, 3, 5-dimethylpyrazole or 7-azaindole as bridging - ligands gave complexes 7-11. The coordination environment around "'"""" each copper(!) of these complexes was expected to be a.tetrahedral arrangement of one oxygen and three nitrogen ligands. The crystallographic results presented herein for ctJi1 SOIM(Etpy) 2 (pz), ~' did not yield the expected structure. The sidearm pyridine nitrogen atoms are not coordinated. The overall coordination about copper approximates trigonal pyramidal, with a long, axial, copper-copper interaction with Cu1-cu1 = 2. 97 A. Copper(I)-copper(I) interactions with metal separations as short as 2. 45 A are known, but in all previously characterized species h~ving proposed Cu1-cu1 interactions there is at least one bridging ligand between the two interacting copper atoms. 28 It is reasonable to assume that the structures of the analogous species, 1-ll, are similar to this structure found for ctJiISOIM(Etpy) 2 (pz), ~The structure of ctJiISOIM(Etpy) 2 (pz), ~' suggested the synthesis of a series of compounds which had no donor atoms on the sidearms (R), i.e., complexes 12-25, Table I. These compounds also proved to - "'""""""' be crystalline compounds, stable in the absence of dioxygen. Since the polydentate ligand systems employed provide only three-coordination for each copper, these compounds are presumed to be three-coordinate in solution. 43 All compounds which have non-bulky sidearms are brown in the solid state, this color being due, in part, to a 600 nm band observed in the solid state nujol mull spectrum (Curve A in Figure 1). This band is not present in the spectrum of the complex with t-butyl sidearms, cu!ISOIM(t_-Bu)ipz), ~' which is red in the solid state as well as in solution (Curve C, in Figure 1). This 600 nm band may be attributed to the intermolecular copper-copper interaction which may be inhibited by the presence of large t-butyl sidearms. This copper-copper interaction appears to be only a solid state phenomenon, since the 600 nm band was not found in solution spectra, even in very concentrated solutions. This novel dimerization of a1° copper(!) centers may be closely related to the dimerization observed for rhodium(!), d8 , isocyanide complexes. 29 The intermolecular bonding of Rh(CNPh): monomers results from the mixing of higher energy orbitals with the lower energy (filled) d orbitals, resulting in overall stabilization. These bonding interactions give rise to an electronic absorption at lower energy than . the original monomer transition. Similarly, the 600 nm absorption observed in the solid state for C~ISOIM(EtPh) 2 (pz), ll, is lower in energy, compared to the 440 nm absorption observed for the same compound in solution (Figure 1). Note that in the solid state, the higher energy band is at 460 nm. This shift from solutionto solid state and the asymmetric nature of the 440 nm solution band suggest that this solution absorption is due to several transitions, one of which may be shifted to lower energy (600 nm) upon interaction with another copper(!). Complexes with apparent copper(I)-copper(I) interactions 44 have not been reported to exhibit notable electronic absorption spectra. Exact assignments of the electronic absorption spectral bands in the copper(!) complexes reported herein await further studies. The solution structure of cu!ISOIM(Etpy) 2 (pz), i, appears to be similar to that found in the solid state with the absence of the intermolecular Cu-Cu interaction. The pyridine nitrogens do not appear to be bound in·benzene solution, since the NMR resonance for the proton on the carbon adjacent to the pyridine nitrogen (PYi , Table I) of cu;rsOIM(Etpy) 2 (pz), i, occurs at the same position, within experimental error, as the corresponding proton from 2-(2' -aminoethyl)pyridine (i.e., 8. 44 ± 0. 02 ppm o from TMS). In fact, all compounds which contain pyridine on the sidearms exhibit this resonance at the same position within experimental error (i.e., 8. 42 ± 0. 04 ppm o from TMS). The resonance of this proton would be expected to shift downfield upon binding to a copper(!) ion. This effect has been observed for the proton in the 2 posit~on of an imidazole ring, which shifts by O. 4 O. 7 ppm downfield upon binding copper(!) to the imidazole ring. 30 Sharp resonances were observed for the pyrazole protons of these compounds (for instance, the doublet at 7. 87 ppm and the triplet at 6. 43 ppm in Figure 2). This indicates that there is no equilibrium between bound and unbound pyrazole, and hence the pyrazolate bridge must be totally bound to the copper(!) ions or totally dissociated in benzene solutions (unlikely, at best). The 7-azaindolate bridge also appears to be completely boW1d to the copper ions in benzene solutions, since the resonances for the {3 2 aliphatic hydrogens (on the sidearms) are split by the unsymmetric bridge (for instance, compound~ in 45 Table I). Also, the resonances found for the 7-azaindolate prcxons are very sharp peaks, indicating that no dissociation equilibrium is occurring. The new copper(!) compounds serve as an interesting series for electrochemical comparisons. Several trends can be observed in Table II. In the following discussion, the formal potentials for the second reduction, (E!), equation 4, will be used since these processes (4) were more reversible and slightly more systematic than the E! pcxentials. The general trends, however, are the same with both processes. The first point of interest is the effect of possible binding sites (i.e., pyridine) in the sidearms. The following series of potentials (E;) was observed: cu;1soIM(EtPh) 2 (pz), 1.§_, (-0. 078 V) ~ cu!ISOIM(MePh) 2 (pz), _ll, (-0. 081 V) » C~ISOIM(Etpy) 2 (pz), ~' (-0. 344 V) > cu;lsoIM(Mepy)ipz), §_, (-0. 452 V). This indicates that the complexes containing the pyridine rings provide a more favorable environment for copper(II) relative to copper(!) when compared to their benzene ring analogues. Since the pyridine nitrogens do not appear to be bound to the copper(!) ions (in solution or in the solid state), the effect of these pyridine rings must be largely on the oxidized copper(II) sites. Hence, it appears that the pyridines bind to the copper ions upon oxidation, .,.._ 26, and dissociate from copper(!) upon - reduction, 9, Scheme m. Recall that in the similar cupric complex, 46 2 c~11ISOIM(hist) 2 (OH) +, ~' the imidazole nitrogens were shown to be coordinated in the solid state X-ray structure. 27 n is reasonable that sidearm pyridine nitrogens would also co.ordinate to copper(II). The above series also indicates that the methyl pyridine sidearm provides a better environment for copper(II} than does the ethyl pyridine sidearm. This may be a result of the geometry of the sidearms and their relative ability to bind to the copper(II) centers. Scheme m. ~ A N O N \ 't~~ Cu/Cu N \ I N-N V 9 N- 26 The introduction of methyl substituents on the pyrazolate bridges causes an increase in the reduction pctential (E!) as follows: C~ISOIM(Mepy) 2 (Me 2 pz), 1, (-0. 374 V) > C~ISOIM:(Mepy) 2 (pz), ~' (-0. 452 V) and cu!1sOIM(MePh) 2 (Me 2pz), ll, (+0. 005 V) > C~ISOIM(MePh) 2 (pz), _ll, (-0. 081 V). An inductive effect of the methyl groups would stabilize copper(II) relative to copper(!), but the reverse trend is actually observed. The steric bulk of the methyl 47 groups may be responsible for the relative destabilization of copper(II). These methyl groups may inhibit the binding of both pyridine (from the . sidearms) and DMF (solvent) to the oxidized copper{II) ions by partially blocking the fourth, square-planar binding site around each copper(II). Also, shielding the copper centers from the solvent molecules with hydrophobic groups (such as these methyls) may not allow the polar DMF molecules to efficiently solvate the charged copper(II) species. Relative stabilization of copper(!) also results from more bulky sidearms, as reflected in the following series (E!): C~ISOIM:(!_-Bu)ipz), ~' (+0. 053 V) > C~ISOIM(2-Pr) 2 (pz), }&_, (+0. 001 V) > cu!rsoIM(l-Pr) 2 {pz), _!&, (-0. 076 V). Finally, an electronic effect on E! is observed in the series of compounds which have phenyl rings bonded directly to the imine nitrogens: C~ISOIM:(PhCOMe) 2 (pz), ~' (+0. 008 V) > C~ISOIM(Ph)ipz), ~' {-0. 032 V) > C~ISOIM:(PhNM~) 2 (pz), M_, (-0. 048 V). Here, the electron withdrawing carbonyl substituent results in relative copper(II) stabilization. Similar effects of remote substituents have been reported recently for a series of mononuclear copper(!!) complexes. lO The oxidation/reduction of these compounds in two one-electron steps was the expected behavior for two interacting metal centers. 14 , 32 This sequential behavior has been observed for the series of binuclear complexes, 27, recently investigated in our laboratories. 33 Stepwise .,...,..... oxidation/ reduction was also found for many ruthenium(II) dimers, in which closer proximity and greater interactions between the ruthenium centers correlates with a greater separation of the two redox 48 M = Mn , Fe, Co, N i , Cu 27 processes (and larger comproportionation constants). 34 That all the new complexes exhibit two one-electron redox waves further suggests that the binuclear compounds are monomeric in solution. Oligomerization would probably result in intermolecular copper-copper interactions leading to more complex electrochemical behavior. Even the complexes with !_-butyl sidearms, which were designed to inhibit intermolecular interactions, show electrochemical behavior similar to all other compounds. It thus seems likely that all new complexes have comparable solution structures, the most probable having threecoordinate copper(!). The average comproportionation constant found for the binuclear 3 copper complexes in this study (from Table II) was K = 1. 9 x 10 , which is smaller than that found for the macrocyclic binuclear copper complex, 6 1 (K = 4 x 10 ). 14 This is reasonable since the lower value was - observed for the present copper(!) compounds which contain one 49 pyrazolate type bridge and one phenoxide bridge, compared to two - phenoxide bridges between the coppers in 1. The pyrazolate bridge is expected to result in a larger Cu-Cu distance and smaller in.tramolecular copper interactions. The oxidized binuclear copper(II) complexes with 26, Scheme m, were not isolated in the bridging pyrazolate, such as ,.,_ present study, but the relative effects of alkaxide vs. pyrazolate bridges on metal interactions have been demonstrated in similar binuclear copper(II) compounds,and the trend was the same as found here. 35 The non-reactivity of the copper(!) complexes in the present study toward carbon monoxide is an addition to the confusing phenomenon of carbon monoxide binding to copper(!) complexes. In general, copper(!) complexes have been found to bind CO if a fourth coordination site is available. 12 , 31 , 36 On the other hand, most four-coordinate copper(!) compounds do not bind CO, with the exception of several square-planar, four-co~rdinate complexes which bind CO as a fifth ligand. 13 - 15 Obviously, the geometry and type of ligands around copper(!} affect the CO binding ability for copper(!). The nature of these effects, however, remains obscure. It may not be valid to compare exact reduction potentials of simple metal complexes with those of corresponding protein systems due to varying solvation effects. 37 Nonetheless, analysis of trends or comparisons of approximate potentials may be useful. Within this context the high reduction potentials observed for binuclear copper 50 protein sites have been "mimicked" by the binuclear copper complexes presented in this study. Hence, it is not unreasonable to believe that these proteins utilize only oxygen and/or nitrogen ligands around each copper. Indeed, a three-coordinate copper-ligand environment can be considered for the reduced form of the binuclear site. The nonreactivity of the new compounds towards CO is in contrast to hemocyanin. This does not preclude a three-coordinate reduced binuclear site, since some three-coordinate copper(!) complexes have been shown to bind C0. 12 , 31 , 36 The oxidation/reduction of these compounds in well separated one-electron steps, contrasts with the available electrochemical information on the binuclear protein sites. While no electrochemical data are available on hemocyanin, for both la.cease and tyrosinase, a single potential has been associated with the overall ~wo-electron reduction of the binuclear copper site. 6, 7 ' 38 Correlations between - the electrochemical behavior of the new compounds, 6-25 and of the ,....,... proteins is difficult due to salvation effects, the different techniques employed, possible non-equilibria in the proteins, etc. 37 ' 38 Nonetheless, the electrochemical behavior reported for the new compounds suggests that the two-electron reduction of laccase and of tyrosinase is probably not a simple reduction of two equivalent copper ions strongly interacting through bridging ligands. 51 Conclusions ~ stable binuclear copper(!) complexes have been prepared using binucleating ligands. Several of the new complexes exhibit coppercopper interactions in the solid state, as suggested by a peculiar visible absorption spectrum. This was confirmed for C~ISOIM:(Etpy) 2 (pz), ~' via a complete molecular structure determination. In addition to the copper-copper interaction each copper(!) is bound to one oxygen and two nitrogen ligands. In solution all the new cuprous complexes likely contain three-coordinate copper, which appears to be a quite stable environment for copper(!). All new complexes exhibit two one-electron redax processes, at well defined potentials. The highest reduction potentials were observed for complex .,..,.,..,. 21 in which the copper(!) centers are somewhat buried within the hydrophobic substituents. With reduction potentials of E! = 0. 239 V and E; = 0. 080 V, the ligand environment in ll represents a substantial improvement on the original square-planar, four- - coordinate ligand environment provided by the starting model, 1. The reason for this high (relative) stability of copper(!) within the new molecules is not obvious. While many three-coordinate mononuclear copper(!) complexes have been isolated employing very soft phosphorous or sulfur donors, very few three-coordinate mononuclear (or binuclear) copper(!) complexes have been isolated with nitrogen or oxygen donors. 23 , 28 , 39 The binuclear nature of the compounds in this study may be responsible for their stability. Perhaps the second copper serves to anchor the binucleating ligands, promoting a favorable three- 52 coordinate environment around each copper(!}. 1 Relatively high reduction potentials on the order exhibited by these complexes (and by the binuclear prctein site) may be necessary for reversible oxygen binding. 40 Unfortunately these complexes have thus far reacted irreversibly towards oxygen. We are presently investigating modifications on the sidearms and their effect on the nature of the oxygenation reaction. ~ We appreciate helpful discussions from T. J. Smith and financial assistance from the National Institutes of Health (Grant No. PHS AM 18319-04) and the International Copper Research association. 53 I Materials. All chemicals were reagent grade and were used as ~ received unless otherwise noted. Copper(II) tetrafluoroborate, ground to a powder then dried for several days in vacuo (25 ° C), was used as Cu(BF 4 ) 2 • 6H 2O. Tetrabutylammonium perchlorate, TBAP, (Southwestern Analytical Chemicals) was dried exhaustively in vacuo (25 ° C) before use. Reagent grade N,N-dimethylformamide, DMF, was dried over MgS04 and 4A molecul.3:r sieves for 24 hours and then vacuum distilled. 2-Hydroxy-5-methylisophthalaldehyde was prepared by a modification of the ltterature method. 41 Tetra(acetonitr.ile) copper(!) tetrafluoroborate was also prepared by the published method. 42 C~IIISOIM(Mepy) 2OH(BF4 ) 2 • CH 3OH • H2 O, 1· Aminomethylpyridine (0. 44 ml, 4 mmoles) was added to a solution of 2-hydraxy-5methylisophthalaldehyde (0. 33 g, 2 mmoles) in methanol (50 ml). The resulting yellow solution was boiled at reflux for 1 hour, after which, Cu(BF4 ) 2 • 6H 2 O (1. 38 g; 4 mmoles) was added. This resulting bluegreen solution was refluxed for 45 minutes. Removal of the solvent (methanol) using a rotary evaporator gave a blue-green solid, which - upon recrystallization from methanol yielded blue-green crystals of 3. These crystals were washed with methanol followed by diethylether and dried under vacuum for several hours. Anal. (C 22H26N4O4 Gl~}½F8 ), C, H, N, Cu. C~IIISOIM(Etpy) OH(BF ) 2 4 2 • CH 3 CH2 OH, 1· 2-(2'-Aminoethyl)- pyridine (0. 50 ml, 4 mmoles) was added to a solution of 2-hydraxy-5methylisophthalaldehyde (0. 33 g, 2 mmoles) in ethanol (25 ml). The 54 resulting yellow solution was boiled at reflux for 5 min.J and then Cu(BF4) 2 • 6H 2O (1. 38 g, 4 mmoles) was added to give a green solution. After boiling for 5 min. , this solution was allowed to cool slowly to 4 ° C, and the resulting blue-green solid was collected and recrystal- - lized from ethanol to give blue-green crystals of 4. The product was washed with ethanol, followed by diethyl ether and dried under vacuum for several-hours. Anal. (C 25H30 N4O3 C~B2 F 8 ) C, H, N; Cu: Calcd, 17. 3; Found 18. 0. cu!11soIM:(hist) 2 OH(BF4) 2 • H2O, .§_. Histamine (0. 44 g, 4 mmoles) was added to a solution of 2-hydroxy~5-methylisophthalaldehyde (0. 33 g, 2 mmoles) in methanol (50 ml). The resulting yellow-orange solution was boiled at reflux for 15 min., and then Cu(BF4 ) 2 • 6H 2O (1. 38 g, 4 mmoles) was added to give a green solution. The solution was reduced to a small volume ("'10 ml) using a rotary evaporator, and the resulting green solid was collected by vacuum filtration. Recrystallization of this solid fro~ ethanol gave light blue-green needles of .§_. The product was collected and washed with ethanol followed by diethylether, and then dried in vacuo for several hours. Anal. (C 19H24N6O3 C~~F8 ) C, H, N; Cu; Calcd 18. 6; Found 19. 3. Cu rous Com ounds. The series of copper(!) compounds _,_.,... 6-25 were synthesized in the manner described below for cu;ISOThf (Etpy) 2 (pz), §!_, with the following modifications. The appropriate amine (RNH 2) and bridge (XH) were used to provide the R and X groups listed in Table I. The initial imine condensation is slow for the aromatic amines, and the amine and dialdehyde were allowed to react overnight, in refluxing 55 - acetonitrile, in the synthesis of .,.._,..,_ 23-25. Compounds 7, .,.._,..,_ 10, .,.._,..,_ 13, .,.._,..,_ 16, and .,.._,..,_ 23-25 ~.,.._,..,_ were recrystallized with DMF instead of CH 3 CN. Compounds 20-22 were recrystallized from heptane instead of acetonitrile. ~.,.._,..,_ C 1ISOTh1(Etpy) 2 (pz), 9.. The following was performed under helium in a Vacuum Atmospheres Dri-Lab inert atmosphere chamber. 2-(2' -Aminoethyl)pyridine (0. 24 ml, 2 mmoles) was added to a solution of 2-hydroxy-5-methylisophthalaldehyde (0.16 g, 1 mmole) in acetonitrile (15 ml). The resulting yellow solution was heated at reflux for 15 min., and then cooled to the ambient temperature. Pyrazole (0. 07 g, 1 mmole) and sodium methoxide (0.11 g, 2 mmole) were added to the solution, and after stirring briefly Cu(CH 3 CN) 4 BF4 (0. 63 g, 2 mmoles) was added to give a brown solution and a brown precipitate. This mixture was boiled at reflux for i 5 min. , and then the acetonitrile was removed by evaporation under vacuum. The brown solid was dissolved in hot toluene (30 ml), leaving behind a white solid (NaBF4 ) upon vacuum filtration. The toluene was then removed from the filtrate by evaporation under vacuum to yield a brown solid. - Recrystallization from acetonitrile gave brown needles, 9. The product was collected by vacuum filtration, washed with acetonitrile and dried for several hours in vacuo. Anal. (C 2 JI 26N6OCl½) C, H, N, Cu. sical Measurements. Sample preparation for physical studies on the air-sensitive materials was accomplished in a Vacuum Atmospheres Dri-Lab inert atmosphere chamber, under a helium atmosphere. Helium-saturated spectroquality solvents were used for solution studies. 56 Magnetic susceptibility measurements were done on powdered samples at the ambient temperature using a Cahn Instruments Faraday balance. The calibrant utilized was HgCo(SCN) 4 , and diamagnetic corrections were made using Pascal's constants. Electronic spectra were recorded on a Cary 14 spectrophotometer. Solid state spectra were obtained from Nujol mulls on filter paper against a Nujol-saturated filter paper as a blank. Infrared spectra were recorded on a Beckman ffi-12 Infrared Spectrophotometer. Solid state spectra were obtained from Nujol mulls pressed between KBr plates. Solution spectra were obtained using calcium fluoride s_o lution cells (path length .of 1 mm). Proton magnetic resonance spectra were recorded on a Varian EM390 spectrophotometer at 90 MHz (34 ° C). The solvent utilized was d6 -benzene containing TMS as the reference. Mass spectra and elemental analyses were performed by the Caltech Microanalytical_ Laboratory. ~ A Princeton Applied Research (PAR) Model 174A polarographic analyzer was used for cyclic voltammetry and differential pulse voltammetry. For display purposes, a HewlettPackard 7004B X-Y recorder was utilized. The apparatus used for constant potential electrolysis (cpe) consisted of a PAR Model 173 potentiostat-galvanostat coupled with a Model 179 digital coulometer. Cyclic voltammetry was done in a single compartment cell with a volume of ca. 5 ml. In all solvents the supporting electrolyte was 0.1 M TBAP. The working electrode consisted of a platinum button electrode or a: hanging mercµry drop electrode. In all cases, the 57 auxiliary electrode was a coiled platinum wire, and the reference electrode consisted of a silver wire immersed in an acetonitrile solution containing AgNO3 (0.01 M) and TBAP (0.1 M), all contained in a 9-mm glass tube fitted on the bottom with a fine porosity sintered glass frit. The apparatus employed for constant potential electrolysis was similar, except a two compartment H-cell was used to isolate the auxiliary electrode from the working compartment. A platinum gauze was used as the working electrode for cpe. All potentials are reported versus the normal hydrogen electrode, nhe. This was accomplished by the use of an internal reference redox couple, namely ferrocene, for which the formal potential is reported to be +0. 400 V versus nhe in water. 16 It has been proposed that the ferrocene reduction potential changes very little in different solvents, and hence it is a good solvent independent redox couple. 16 Experimentally, small amounts ( ca. 10-3 M) of ferrocene were added to solutions containing tp.e compounds of interest and formal potentials for both couples were measured under the same conditions. - X-Ra Data Collection and Reduction. Acicular crystals of 9 were grown by slow evaporation of an acetonttrile solution under helium. Preliminary oscillation and Weissenberg photographs showed the space group to be P2 1 /c (No. 14) uniquely defined by its extinctions. A crystal of dimensions 0. 04 mm x 0. 08 mm x 0. 33 mm (which had been cut from a larger needle) was mounted on a Syntex P21 fourcircle diffractometer for data collection. No attempt was made to protect the crystal from atmospheric oxygen as no air sensitivity has been exhibited in the solid state. Cell parameters were determined 58 by a least-squares fit to fifteen automatically centered reflections with 28 > 56°. The resulting parameters are given in Table m. Intensity data were collected from two octants using 0 -20 scans to a maximum 28 of 130°. The scan range extended from 1 ° below the Cu.Ka 2 0 value to 1 ° above the CuKrv value, and the scan rate was 1 • u.2 1 ° /min, with the total background counting time equal to the total scan time. Three check reflections were measured after every fifty reflections to monitor the crystal and instrument stability. Because the check reflections showed an overall drop in intensity of 5% during the course of data collection, the data were scaled in eleven groups according to the average intensity in each group of one of the check reflections (002). Lorentz and polarization corrections were then applied. standard deviations of intensities were calculated using the formula 2 2 2 a (I) = [S + (Bl + B2) + (dS) ] /(Lp) , where S, Bl, and B2 are the scan and two background counts and d was taken as 0. 02. 43 Absorption corrections were calculated by the method of gaussian quadrature. 44 After symmetry-extinct reflections were deleted and equivalent reflections averaged, there remained 4269 unique data (3390 > 0; a rather high percentage of the high 2 e data was "unobserved"). Solution and Refinment of the structure. With the exception of C. K. Johnson's ORTEP program, all computer programs were from the CRYM system of crystallographic programs. Literature values were used for the scattering tactors for cu+, O, N, and C, 45 the real 59 part of the anomalous dispersion correction for copper 45 and for the H scattering factors. 46 The function minimized in the least-squares 2 2 2 refinement was ~w (IF O 1 - !k' F c 1 ·) , where the weight w = 1/a2(F:), and F O and F c are the observed and calculated structure factors. The positions of all nonhydrogen atoms were found using standard Patterson-Fourier techniques. Least-squares refinement with these atoms and isotropic temperature factors led to R = ~ 11 kF O 1-1 F c 11 /Ti IkF O I = O. 087 (for those reflections with F~ > 2a). Use of anisotropic temperature factors for all nonhydrogen atoms lcmered R to O. 062. The 26 hydrogen atoms were located by difference map techniques. Each was assigned a fixed isotropic temperature factor equal to 1. 0 A greater than that for the carbon 2 atom to which it was bound. Hydrogen positional parameters were refined. The final refinement was by blocked-matrix least-squares with all positional parameters in one nn trix and the anisotropic thermal parameters an~ scale factor in another. No data were omitted from this refinement. The final R was O. 042 for the 2620 reflections 2 with F O > 2a; R for all data was O. 063. The final goodness of fit, . LJW(k2 F 2O - F 2c) 2/(n-p), was 2. 85, where n = 4269 1s the number of observations and p = 394 is the number of parameters. The largest feature in the final difference Fourier map was a peak of height 3 O. 96 eA - , located approximately 0. 7 A from Cu2 in the positive y direction. The second highest peak (0. 87 eA - 3 ) was not in a position of chemical significance (between the pyrazole groups of neighboring molecules). Final parameters and interatomic distances and angles are presented in Tables IV-VIl. 60 References ~ (1) Lontie, R. in "Inorganic Biochemistry'", Eichhorn, G. I. ed.; Elsevier: New York, N. Y., 1973; p. 344. (2) (a) Mason, H. S. Ann. Rev. Biochem. ~ ' 35, 595-634. (b) Vanneste, W. H. and Zuberbuhler, A. in "Molecular Mechanisms of Oxygen Activation", Hayaishi, 0. ed.; Academic Press: 1974; p. 371. (3) (a) Fee, J. A. structure and Bonding~' 23, 1-60. (b} Vanngard, T. I. in "Biological Applications of Electron Spin Resonance Spectroscopy", Swartz, H. M.; Bolton, J. R.; and Borg, D. C., eds.: J. Wiley: New York, N. Y., 1972; p. 411. (4) See Wurzbach, J. A.; Grunthaner, P. J.; Dooley, D. M.; Gray, H. B.; Grunthaner, F. J.; Gay, R. R.; Solomon, E. I. J. Amer. Chem. Soc. llll, 99, 1257-1258 and references contained therein. (5) Solomon, E. I.; Dooley, D. M.; Wang, R.; Gray, H.B.; Cerdonio, M.; Mogno, F.; Romani, G. L. J. Amer. Chem. Soc. 1976, 98, 1029-1031. ~ (6) Makino, N.; McMahill, P.; Mason, H. S.; Moss, T. H. J. Biol. Chem . .!111, 249, 6062-6066. (7) (a) Reinhammar, B. R. M"; Vanngard, T.I. Eur. J. Biochem. !JU!, 18, 463-468. Reinhammar, B. R. M. Biochem. Biophys. Acta 1.fil, 275, 245-259. (b) Farver, 0.; Goldberg, M.; Lancet, D.; Pecht, I. Biochem. Biophys. Res. Commun. ~ ' 73, 494-500. 61 References (continued) (8) For example see: Amundsen, A. R.; Whelan, J.; Bosnich, B. J. Amer. Chem. Soc. ~ 1977, 99, 6730-6739. Simmons, M. G.; Wilson, L. J. Chem. Commun. ~ ' 634-636. Bulkowski, J. E.; Burk, P. L.; Ludmann, M. F.; Osborn , J. A. Chem. Commun. !m, 498-499. Lehn, J. M . .; Pine, S. H.; Watanabe, E.; Willark, A. K. J. Amer. Chem. Soc. !m, 99, 6766-6768. Alberts, A. H. ; Annunziata, R. ; Lehn, J. M. J. Amer. Chem. Soc. !m, 99, 8502-8504. (9) Arcus, C. S.; Wilkinson, J. L.; Mealli, C.; Marks, T. J.; --- - lbers, J. A.; J. Amer. Chem. Soc.; 1974, 96, 7564-7565. Mealli, C.; Arcus, C. S.; Wilkinson, J. L.; Marks, T. J. ; Ibers, J. A. J. Amer. Chem. Soc. 1976, 98, 711-718. --------- (10) Yokoi, H.; Addison, A. W. Inorg. Chem. !m, 16, 1341-1349. (11) Fenton, D. E.; Lintvedt, R. L. J. Amer. Chem. Soc. 1978, ~ ------------ 100, 6367-6375. • (12) Gagne, R. R.; Gall, R. S.; Lisensky, G. C.; Marsh, R. E.; Speltz, L. M. Inorg. Chem. ~ ' 18, 771-781. (13) Gagne, R. R. J. Amer. Chem. Soc. ~ ' 98, 6709-6710. Gagne, R. R.; Allison, J. L.; Gall, R. S.; Koval, C. A. --- - J. Amer. Chem. Soc. 1977, 99, 7170-7178. (14) Gagne, R. R. ; Koval, C. A.; Smith, T. J. J. Amer. Chem. Soc. !m, 99, 8367-8368. Gagne, R. R.; Koval, C. A.; Smith, T. J.; Cimolino, M. C. J. Amer. Chem. Soc. ~ ' 10, 4571-4580. 62 References (continued) (15) Gagne, R. R.; Allison, J. L.; Lisensky, G. C. Inorg. Chem. 1978, -17, 3563-3571. Gagne, R.R.; Allison, J. L.; Ingle, ~ D. M. J. Amer. Chem. Soc. fil!t, 18, 2767-2774. (16) Bauer, D.; Breant, M. in "Electroanalytical Chemistry", Vol. 8, Bard, A. J. ed., Marcel Dekker, Inc.: New York, N. Y., 1975, pp. 282-344. (17) Attempts to isolate cu!1ISOTh1(Mepy) 2 (pz) + directly from Cun 2 starting materials led only to cuPISOTh1:(Mepy)(OH/+, ~(18) All electrochemical measurements were made without the use of iR compensation. The cyclic voltammetric peak to peak separa- tions are comparable to those we have observed for other chemically reversible redox couples involving polydentate ligands. 13 , 14 For comparison purposes, under the cyclic voltammetric conditions utilized ferrocene was observed to give peak to peak separations of 70 ~ 80 mV. (19) The separation between planes of carbon atoms in graphite is 3. 35 A, for example. 20 In a recent series of crystal structures of Schiff bases akin to the ligand in the present structure, interplanar distances of 3. 42 to 3. 50 A were found. 21 See also: Herbstein, F. H. in "Perspectives in structural Chemistry", Dunitz, J. D. and Ibers, J. A. eds., Vol. 4, John Wiley and Sons: New York, N. Y., 1972, p. 166. (20) Cotton, F. A.; Wilkinson, G. "Advanced Inorganic Chemistry", Interscience Publishers: New York, N. Y., 1972, p. 288. 63 References (continued) (21) Moustakali-Mavridis, I.; Hadjoudis, E.; Mavridis, A. Acta (22) Cryst. ~ ' B34, 3709-3715. Some of the shorter cu1-N distances that have been reported are as follows: 1.937 Ato 1.943 Ain [l,1-difluoro-4,5,11,12-tetramethyl-l-bora-3, 6, 10, 13-tetraaza-2, 14-dioxacyclotetradeca3, 5, 1O, 12-tetraenato] copper(!), which contains copper in a distorted square-planar geometry;1 5 1. 946 A and 1. 948 .A for the non-bridging pyrazole groups in bis [(hydrotris(l-pyrazolyl)borato)copper(I)], which contains copper in a distorted tetrahedral geometry; 9 1. 97 Ato 2. 02 A in tris-(2-picoline)copper(I)perchlorate, in which copper is three-coordinate, with a distorted trigonal-planar geometry. 23 One structure that does have a Cu 1-N bond about as short as some of those in the present structure is the dimer of diazoaminobenzene copper(!). This contains two-coordinate copper, with Cu-N distances of 1. 898 A and 1. 939 A (also a Cu-Cu distance of 2. 451 A). 24 (23) Lewin, A. H. ; Michl, R. J. ; Gani, P. ; Lepore, U. Chem. ---- Commun. 1972, 661-662. (24) Brown, I. D.; Dunitz, J. D. Acta Cryst., fil!., 14, 480-485. (25) Structural and electronic factors influencing copper reduction potentials have been documented: Patterson, G. S.; Holm, R. H. Bioinorg. Chem., ~ ' 4, 257-275. (26) Grzybowski, J. J. ; Merrell, P. H. ; Urbach, F. L. Inorg. Chem. ~ ' 17, 3078-3082. 64 References (continued) (27) Gagne, R. R.; McCool, M.; Marsh, R. E. Acta Cryst., in press. (28) For example see: Camus, A.; Marsich, N.; Nardin, G.; Randaccio, L. Inorg. Chim. Acta !m 23, 131-144. Jardine, F. Adv. Inorg. Chem. Radiochem. !.ill, 17, 115-163. Mehrotr.a, P. K.; Hoffmann, R. Inorg. Chem . .!_!!ll, 17, 21872189, and references contained therein. (29) Mann, K. R.; Gordon, J. G.; Gray, H. B. J. Amer. Chem. Soc. !.ill, 97, 3533-3555. Lewis, N. S.; Mann, K. R.; Gordon, J. G.; Gray, H. B. J. Amer. Chem. Soc. ~ ' 98, 7461-7463. Mann, K. R.; Lewis, N. S.; Williams, R. M.; Gray, H. B.; Gordon, J. G. Inorg. Chem.!.!!]!, 17, 828-834. (30) Suguira, Y. Inorg. Chem. !.!!]!, 17, 2177-2182. We have also observed a 0.13 ppm downfield shift of the proton on the carbon adjacent to the pyridine nitrogen of N, N, N', N' -tetrakis-(2pyridylmethyl)-ethylenediamine (TPEN) upon binding of Cu 1 to . this pyridine ring. 31 (31) Gagne, R. R.; McCool, M.; Marsh, R. E.; Dodge, J. A.; Kreh, R. P., unpublished results. (32) A curious exception to the stepwise electrochemistry exhibited by interacting metals is that reported for a series of binuclear triketonato copper(II) complexes. 11 (33) Spiro, C. L. , Ph.D. Dissertation, The California Institute of Technology, 1981. 65 References (continued) (34) (a) Creutz, C. and Taube, H. J. Amer. Chem. Soc. ~ ' 91, 3988-3989. Creutz, C. and Taube, H. J. Amer. Chem. Soc. !Jt:G, 95, 1086-1094. (b) Weaver, T. R.; Meyer, T. J.; Adeyemi, S. A.; Brown, G. M.; Eckberg, R. P.; Hatfield, W. E. ; Johnson, E. C. ; Murray, R. W. ; Untereker, D. J. Amer. Chem. Soc. ~ ' 97, 3039-3048. (c) Callahan, R. W.; Keene, F. R. ; Meyer, T. J. ; Salmon, D. J. J. Amer. Chem. Soc. 1977, 99, 1064-1073. ----(35) Robson, R. Inorg. Nucl. Chem. Lett. IDQ, .2., 125-128. (36) Bruce, M. I.; Ostazewski, A. P. J.C. S. Dalton 1973, 2433-2436. ---- Churchill, M. R.; DeBoer, B. G.; Rotella, F. J.; Abu Salah, O. M.; Bruce, M. I. Inorg. Chem. ~ ' 14, 2051-2056. (37) Hill, C. L.; Renaud, J.; Holm, R. H.; Mortenson, L. E. J. Amer. Chem. Soc. !.m, 99, 2549-2557. (38) Farver, O.; Goldberg, M.; Wherland, S.; Pecht, I. Proc. Natl. ---- - Acad. Sci. U.S.A. 1978, 75, 5245-5249. (39) Ellen, P.; Bradley, D.; Hursthouse, M.; Meek, D. Coord. Chem. Rev. !.m, 24, 1-95. (40) Reversibility of oxygenation in mononuclear cobalt complexes was shown to be related to reduction potentials: Carter, M. J.; Rillema, D. P.; Basolo, F. J. Amer. Chem. Soc. !ill, 96, 392-400. (41) Ullmann, F.; Brittner, K. Chem. Ber. ~ ' 42, 2539-2548. (42) Hathaway, B. I.; Holah, D. G.; Postlehtwaite, J. D. J. Chem. Soc.~' 3215-3218. 66 References (continued) {43) Peterson, S. W.; Levy, H. A. Acta Cryst . .!Jt§1, 10, 70-76. (44) Busing, W. R.; Levy, H. A. Acta Cryst . .!Jt§1, 10, 180-182. (45) "International Tables for X-Ray Crystallography," Vol. ID; Kynoch Press: Birmingham, England, 1962. (46) Stewart, R. F. ; Davidson, E. R. ; Simpson, W. T. J. Chem. Phys. ~ ' 42, 3175-3187. 67 CHAPTER ID The Synthesis and Structures of (N, N, N', N' -Tetrakis-(2-Pyridylmethyl)Ethylenecliamine) Di.copper(!) and its Di.carbonyl Adduct Robert R. Gagne, Robert P. Kreh, and John A. Dodge Richard E. Marsh and Michael McCool Contribution No. 6294 from the Di.vision of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91125. 68 Introduction ~ Copper(!) complexes show a variety of cc;>0rdination environments ranging in coordination number from two to five, l, 2 The number of ligands bound to copper(!) seems to be influenced greatly by the imposed geometry and type of ligands. We have recently - investigated a series of binuclear copper(!) compounds, including 1, ~ A - c; c,l- r~~Y, 0 ~ -.../\.~~ N-N V 1 which appear to be three coordinate in solution while experiencing significant intermolecular copper-copper interactions in the solid state. 3 Although these compounds contain an "available" (fourth) · coordination site on the metal, they are unreactive towards ligands such as pyridine and carbon monoxide. On the other hand, a stable carbon monoxide adduct is formed by the binuclear Cu(I) complex derived from the ligand, N, N, N', N' tetrakis(2-pyridylmethyl)ethylenediamine (TPEN). 69 In view of the unusual structure and lack of carbon monoxide reactivity - for system 1, a complete crystal and molecular structural analysis - - of both CuiTPEN)(BF4 ) 2 , 2, and its carbonyl adduct 3 have been undertaken for comparative purposes. Synthesis and Characterization of Com The synthesis of the ligand and the cuprous complexes - are presented in Scheme 1. Complex 2 appears slightly green in the solid state, but no absorptions above 450 nm were observed - in the visible spectrum of the solid, 2 (Nujol mull, under He). - The crystal and molecular structure of 2 has been determined, and the results are presented below. - The binuclear copper(I) complex, 2, is diamagnetic, as expected, showing very sharp signals in the proton Nl\ffi. (Figure 1). This spectrum confirms the identity of the organic entity. Table I lists ihe peaks observed for this complex along with the chemical shifts for the free ligand in the same solvent (CD 3CN). The binding of copper(I) to the ligand in complex~ is evidenced by the rather large changes in chemical shifts, Scheme 1 HCI 4 ~ + H2 N~NH • I 2 NaOH H20 ► ~p d ~ TPEN I) Cu(CH3 CN)4 BF4 , CH 3 CN 2)CH 3 0H oc-Cu-~-cu-co I -l I. ·7 .. co CH30H •2BF4 Cu 2 (TPEN)(CO~(BF4)2 ,·~c,, · 'CJ 6 eu N" (3) •2BF4 Cu 2 (TPEN)(BF4) 2 (2) -::s 0 71 a ocetonitrile TMS ocetonitrite TMS 2 0 /3 PY6 10 I 8 6 1 4 CHEMICAL SHIFT (ppm, 8) ~ NMR spectrum of the TPEN ligand (top) and C11i(TPEN)(BF4 ) 2 , - 2 (bottom) in CD 3CN at 34°C. Table I. ~ Proton NMR Frequencies for TPEN and Ct1i(TPEN)(BF4 ) 2 a a ~ PY·a PY4 ~ ~ TPEN 3.68 2.63 7.39 7.62 7.11 8. 44 Ct1i (TP EN) (BF4 ) 2 3.72 2.98 7.23 7.77 7.34 8. 55 ComEound aAll spectra were obtained in CD 3CN solution at 34°C under an atmosphere of helium. The values listed are given in parts per million, o, relative to Me4 Si. resonances are given in Figure 1. The assignments for the -1 l\l 73 relative to the free ligand. The resonances due to three of the protons on the pyridine ring ( PY4 , PY5 , py J move downfield upon binding to the positive copper(!) ion, 4 while the py resonance shows an anomalous 3 upfield shift. The methylene protons from the ethylenediamine bridge (/3} also show a surprising upfield shift in the copper(I) complex. These anomalies could be due to ring current effects from the pyridine rings. - The binuclear complex, 2, is insoluble in most organic solvents (e.g., CH2 Cl.z, CH 3OH), but it is very soluble in acetonitrile, resulting - in a deep yellow solution. The solid complex, C~(TPEN)(BF4 ) 2 , 2, is a very faint green color, and this color change upon dissolution is probably indicative of CH 3CN coordination to the cuprous ions. This - compound, 2, also dissolved in N, N-dimethylformamide, but the solution turned immediately to blue, and copper metal plated out of solution. However, the disproportionation reaction was reversed entirely by purging the solution with carbon monoxide, presumably resulting in a colorless solution of a copper(!) carbonyl adduct (vida infra). - While complex 2 is only slightly soluble in methanol, saturating - a suspension of 2 in methanol (ca. 100 mg of solid in 20 ml of solvent) • with carbon monoxide resulted in the dissolution of all of the solid to give a colorless solution. Purging the solution with nitrogen re-precipitated - the solid, 2 (identified by I. R. ). This process was totally reversible and could be repeated at will. Evaporation of the methanol solvent with a stream of CO resulted in off-white crystals of the dicarbonyl, - C~(TPEN)(CO) 2 (BF4 ) 2 , 3. 74 - - While the solid complexes, 2 and 3, show no apparent reaction - with dioxygen, solutions of the dicarbonyl (3 in C~C~pH 3OH or DMF) exhibit a color change from colorless to blue. This reaction could not be reversed by subsequent bubbling of argon through the solutions, and these autoxidation products have not been characterized. - The carbonyl complex, 3, in methylene chloride solution exhibits a single carbonyl stretching frequency, vCO' at 2110 cm- 1 • The frequency of this CO stretch is very high compared to that found in other copper(I) carbonyl complexes. 2 , 5, 6 The cationic character and - "soft'' pyridine ligands of complex 3 may result in less .1T donation from copper(I) into the 1r* orbi ta.I of CO, thereby raising the frequency of the CO stretch. The solid state I. R. spectrum (Nujol mull) 1 of~ exhibits vCO as a strong, sharp doublet at 2097 and 2107 cm- • To help explain the origin of this phenomenon, a complete structural analysis was performed which confirms the presence of two distinct carbonyls in the solid state. Table II summarizes the basis crystal data for the compound. 2 The atomic labeling scheme for C~(TPEN) + is shown in Figure 2. The molecule lies on a crystallographic two-fold rotation axis, and primes are used to designate symmetry related atoms within the same molecule. The atomic parameters and interatomic distances and angles are given in Tables ID- VI. As seen in Figures 2 - 4, the binuclear complex is folded into a skewed "face-to-face'' configuration. Probably the most significant 75 Table II. ~ - Crystal Data for Cl½(TPEN)(BF4 ) 2 , 2, and - Cl½(TPEN)(CO) 2 (BF4 ) 2 , 3. Compound Cl½ (TP EN) (BF4 ) 2 Cl½ (TP EN) (CO)2 (BF4 ) 2 Formula (per asymmetric unit) BC13CUF4H14N3 B2 C28 Ct½F8 ~ Formula w_e ight 362.63 781. 27 Space group C2/ c (no. 15) P2 1/ c (no. 14) 0 13. 099( 4) 17. 706(5) 0 15. 974(6) 10. 372(3) c, A 0 14. 472(5) 19. 601(5) f3, deg 102. 74(2) 112. 77(2) V, A o3 2953. 7(17) 3319. 2(16) z 8 4 a, A b, A Pealed' g/cm µ, cm- 1 radiation 0 :>t, A 3 1. 63 8 N6O2 1. 56 15.8 23. 7 MoK0 CuKa 0.71069 1. 5418 __J 76 C4 C14 ~ ORTEP drawing showing the nonhydrogen atoms of 2 C~(TPEN) + as seen along the twofold axis. Forty percent probability ellipsoids are shown. The atomic labeling scheme is given, and each hydrogen is assigned the same number as the carbon to which it is bound. In the discussion of this molecule in the text, primes are used to designate atoms derived by rotation about the axis. Table III. Atomic Parameters and eds's for the Nonhydrogen Atoms in C~(TPEN)(BF4) 2 , 2. - ~ xa Cu N1 C2 C3 C4 C5 C6 C7 NB C9 C10 C11 C12 C13 C14 C15 N16 F1A F2A F3A F4A y 3694 8 3656 9 191 5 1769 5 2163 8 -550 6 3216 9 -660 7 3879 8 19 3508 9 750 7 2448 8 839 6 1664 6 1956 8 1124 5 1554 4 451 7 2323 5 1601 7 1426 6 525 6 1864 7 2888 7 248 6 3100 7 -594 6 2245 8 -1129 6 1246 8 -793 6 14 5 1045 4136 2337 5 5526 2097 8 1686 6 5387 i! 5484 10) 3076 7 z 16687J9) 1486 1290 6 1271 7 1469 8 1679 7 1698 6 1873 7 2418 5 2247 7 3447 6 3654 5 3909 6 4080 6 4034 7 3802 7 3621 5 390 76! 1446 -118 6 218(8 b Uu- U22 470 87 68} 83 6 95 10) 85 9 67 8 87 9 95 9 76 6 91 10) 96 8 71 7 59 7 82 8 88 9 79 8 67 6 66 6 114 7 149 9 220 12) 592 8l 691 62 7 72 i} 93 77 13 . 97 150 10! 10~ 194 17 99 10 166 14 94 10 90 9 59 81 9 82 58 6 58 56 7 121 11) 57 7 71 64 8 46 81 8 86 93 10) 88 i! 69 119 10) 69 103 9 49 122 8 90 i! 176 9 305 13) 66 8 111 8) 102 8 133 10) 186 11) U33 !~ U12 U13 U23 -81 10) -18 5 -3 8 33 9 15 9 -23 9 -11 8 -29 7 -8 5 -10 6 -5 6 87 -11 7 16 7 19 7 67 55 -4 5 19 51 7 116(9) 59 6 10 5 16 7 46 9 28 8 22 8 67 15 7 85 -18 8 -15 7 76 36 17 7 32 8 25 7 26 5 20 6 16 6 20 6 108 9 28 11) -17 1 15 36 10) 33 9· 97 18 7 7 5) -:J 97 -:J 67 10 6 -2 8 20 8 22 8 -1 7 13 6 -14 6 -39 8 -35 6 -37 8 II - Table III. Atomic Parameters and esd's for the Nonhydrogen Atoms in Cl.½(TPEN)(BF4 ) 2 , 2. (cont'd) ~ FlB F2B F3B F4B BA BB 02 X y z B,.A 473 2 587 2 522 2 627 2 513 0 552 0 283 2) 251 1 157 2 234 2 230 0 231 0 14 2 136 2 53 2 17 2 48 0 55 0 13. 8 1. 2) 10.3 0.9 14.1 1.1 11. 6 0.9 8. 5 0.7 6.2 1. 4 5 aFractional coordinates have been multiplied by 10 for the copper atoms and 104 otherwise. 4 3 Uij has been multiplied by 10 for the copper atoms and 10 otherwise. b The form of the thermal 2 ellipsoid is exp( -211'2 (U11 h a*2 + • • • + 2 U2 J{lb*c*)] for the anisotropic thermal parameters. -:J (X) 79 Table IV. Hydrogen Atomic Parameters for Cl.½(TPEN)(BF4 ) 2 , 2. - ~ H2 H3 H4 H5 H7A H7B H9A H9B Hl0A Hl0B H12 Hl3 H14 H15 xa y z B,A 170 348 461 397 165 250 16 89 110 223 346 381 235 66 -103 -120 -6 123 192 203 237 281 160 176 64 -80 -172 -117 116 112 146 181 12'7 222 157 246 382 362 397 423 417 377 7.8 7.8 8.2 8.0 7.2 7.2 8.0 8.0 7.3 7.3 6. 6 7.4 6.7 7.1 02 3 aFractional coordinates have been multiplied by 10 • 80 - Table V. Interatomic Distances (A) for Cl½(TPEN)(BF4 ) 2 , 2. a ~ Cu-Nl Cu-N16' Nl-C2 N1-C6 C2-C3 C3-C4 C4-C5 C5-C6 C6-C7 C7-N8 N8-C9 N8-Cl0 C9-C9' Cl0-Cll Cll-C12 Cll-N16 Cl2-Cl3 Cl3-Cl4 Cl4-C15 Cl5-N16 1. 936(7) 1.890 8) 1. 346 12 1. 355 12 1. 398 15 1.356 16 1. 353 16 1.402 15 1. 513 14 1.490 12 1.501 11 1.496 11 1. 521 13 1.495 13 1. 383 12 1. 339 11 1. 384 13 1.399 14 1. 386 14~ 1. 330 12 BA-FlA BA-F2A BA-F3A BA-F4A BB-FlB BB-F2B BB-F3B BB-F4B 1. 284 1. 411 1. 398 1.406 1. 36 1. 20 1. 24 1.23 Nonbonding DJstances Cu-Cu' Cu-NS Cu-Nl6 BA-BB 2. 779(2) 2. 295(7~ 2. 827(8 0.50 aPrimes are used to designate atoms operated upon by the twofold axis (i.e., in the equivalent position x, y, ½-z). 81 Table VI. ~ - Interatomic Angles (deg) for Cl½(TPEN)(BF4 ) 2 , 2. ~ N16' -Cu-Nl C2-Nl-Cu C6-Nl-Cu C6-Nl-C2 C3-C2-Nl C4-C3-C2 C5-C4-C3 C6-C5-C4 C5-C6-Nl • C7-C6-N1 C7-C6-C5 N8-C7-C6 C9-N8-C7 Cl0-N8-C7 C10-N8-C9 C9' -C9-N8 Cll-Cl0-N8 Cl2-Cl1-Cl0 Nl6-Cl1-Cl0 Nl6-C11-C12 Cl3-Cl2-C11 C14-Cl3-Cl2 Cl5-C14-Cl3 N16-C15-Cl4 C11-Nl6-Cu' C15-Nl6-Cu' C15-Nl6 Cll F2A-BA~lA 146. 8 3 125. 3 6 116. 9 6 117.28 122. 7 9 119.2 10~ 119. 5 11 120. 0 10 121. 5 9 114. 9 8 123. 5 9 112. 0 8 106. 7 7 110. 5 7 112. 1 6 114. 3 7 111. 0 7 121. 9 8 115. 6 8 122. 5 8 120. 2 8 117. 3 9 118. 6 9 123. 9 9 124. 1 6 118. 4(6 117.5(8 104.9 ~ee footnote a for Table V. F3A-BA-FlA F4A-BA-FlA F3A-BA-F2A F4A-BA-F2A F4A-BA-F3A F2B-BB-F1B F3B-BB-F1B F4B-BB-F1B F3B-BB-F2B F4B-BB-F2B F4B-BB-F3B 110.0 108.4 111. 6 113.7 108.1 110.9 110.8 113.2 109.1 105.0 107.6 Nonbonding Angles Nl-Cu-Cu' NB-Cu-Cu' N16' -Cu-Cu' N8-Cu-N1 N16'-Cu-N8 C7-N8-Cu C9-N8-Cu C10-N8-Cu 129. 6 2 77. 7 2 71. 7 2 82. 0 3 130. 8 3 97. 9i5~ 114. 85 113. 6 5 82 ~ ORTEP drawing showing the nonhydrogen atoms of 2 Cl½(TPEN) + as viewed perpendicularly to the twofold axis. Forty percent probability ellipsoids are shown. 83 ~ Stereo drawing of C~(TPEN) +. 2 84 aspect of this structure is the close intramolecular copper-copper contact which it allows. The separation is only 2. 780 A, even though there are no short ligand bridges as in most other known structures containing close contacts between Cu(n atoms. 1 , 7 The c_o ordination about copper (identical for each of the two atoms) is difficult to place in a specific geometric category. The two bonds to pyridine nitrogens are short (Cu-Nl = 1. 94 A, Cu-N16' = 1. 89 A), with an N1-Cu-N16' angle of 146. 6° .8 There is also a relatively close contact between each copper atom and a tertiary amine nitrogen, although this is quite long for a bond (Cu-NB = 2. 30 A). Finally, there is the close contact between copper atoms themselves. The angles about copper for these various interactions are given in Table VI. (It might also be noted that the distance between Cu and Nl 6 of the pyridine ring below, 2. 83 A, is sufficiently long to preclude anything but a very weak interaction. ) As is suggested by Figui-es 2-4, each copper atom lies almost in the plane defined by the two bound pyridine nitrogens and the nearest amine nitrogen. The deviation of copper from this plane is 0. 07 A, in the direction that gives the shorter Cu-Cu' distance. Each copper atom is also nearly coplanar with each of the two pyridine rings to which it is bound. The deviation of Cu from the plane of the ring containing N16' is only 0. 04 A; for the Nl ring the deviation of Cu is larger (0. 20 A) and in the direction that gives the shorter Cu-Cu' distance. 85 Each pyridine ring is nearly planar, the largest deviations from the respective mean planes being 0. 014 Afor each of Nl and C6 (in opposite directions) and 0. 019 Afor each of Cll' and C12' (again in opposite directions). The dihedral angle for these two ring planes is 169°. The reason for the skewed nature of the TPEN ligand is not clear, but some possible explanations can be suggested. First, with the molecule twisted as it is, the lone electron pair of each amine nitrogen is pointed in the direction of the nearest copper atom, as indicated by the following angles about N8: Cu-N8-C7 = 98. 0°; Cu-N8-C9 = 114. 6°; Cu-N8-Cl0 = 113. 3°. Thus, an affinity of copper for a third ligand could be a contributing factor in the stabilization of this conformation. However, in view of the length of the Cu-N8 separation (2. 30 A) this attraction could not be very strong, and the interaction with another copper atom would seem to be of greater importance. An examination of CPK molecular models suggests that other steric factors could be responsible for the observed conformation, and that the Cu-N8 proximity could be coincidental. Of course, this structure does prevent the pyridine rings from being perfectly eclipsed. (See Figure 3.) Also, some degree of rotation about the C9-C9' bond, away from an eclipsed conformation of the ethylenediamine moiety, appears to be necessary to allow the pyridine rings to move sufficiently far apart to accommodate the copper atoms. Rotation about the N8-C9 86 bond (and the equivalent N8'-C9' bond), which results in pointing each nitrogen lone pair away from the center of the molecule and toward copper, relieves van der Waals repulsions between hydrogen atoms (e.g., between H7A, H9A, and HlOA' ). Molecular models also indicate that, in order for the pyridine rings to be_close enough to bind copper, the pyridine nitrogen lone pairs cannot be collinear. This, plus an attraction between copper atoms, would explain the deviation of the Nl-Cu-Nl6' angle from 180°. Finally, some mention should be made of the BF4 - anions. These are distributed uniformly throughout the crystal and are separated by approximately half a unit cell in the direction of each of the crystal axes. The BF4 - groups are disordered between two positions which differ primarily by rotation of the group with a small degree of translation. The populations of these sites are approximately 0. 65 and 0. 35, respectively. No fluorine atom lies within 3. 5 Aof a copper atom. The molecular structure and numbering scheme for the cation Cl½(TPEN)(CO)/+ are shown in Figure 5. Basic crystal data are listed in Table II. Atomic parameters and interatomic distances and angles are presented in Tables VII-X. - Although the binuclear complex 3 contains two copper ions in identical chemical environments, a center of symmetry was not found C13' C4 N16' C9' ex:, -1 C2' C3' C14 ~ ORTEP drawing showing the nonhydrogen atoms of C~(TPEN)(CO) 2 +. The atomic 2 labeling scheme is given, and each hydrogen is assigned the same number as the carbon to which it is bound. Although primes are used here, this molecule does not contain a twofold rotation axis. - Table VII. Atomic Parameters and esd's for the Nonhydrogen Atoms in CUa(TPEN)(CO) 2 (BF4 ) 2 , 3 ~ Cu Cu' 018 018' Bl B2 Fl F2 F3 F4 F5 F6 F7 FSA F8B N1 N16 N8 N16' Nl' NB' C17 C6 C5 C4 C3 C2 Cll C12 C13 b a x- y z Uu- 64488 5 81118 5 5329 3 7923 4 3857 5 553 5 4497 3 3888 3 3918 3 3155 3 -203 4 471 3 1061 3 296 8 960 8 6808 3 6321 3 7728 3 9140 3 7137 3 8276 3 5776 4 7581 4 7967 4 7542 5 -6756 5 6407 4 7005 4 7049 5 6346 6 12678 27758 1361 5020 2529 10~ 2391 11 2591 7 3461 6 1372 6 2608 5 2925 7 1287 5 3182 5 2600 11~ 2029 16 3015 5 -39 5 813 5 2526 5 2159 5 730 4 1289 7 3044 6 4201 7 5315 7 5299 7 4138 7 -660 6 -1400 7 -1533 7 30151 5 17656 5 1472 3 2558 4 3087 6 4120 6 2889 4 3576 3 3433 3 2490 2 3967 6 4428 3 4591 3 3442 5 3690 7 3516 3 37·34 3 3639 2 1556 2 866 2 1979 2 2047 4 4018 3 4325 4 4115 4 3602 4 3320 4 4159 3 4755 4 4905 5 569 5 574 5 103 4 142 5 74 6 56 6 128 4 131 4 111 4 103 3 116 5 147 4 107 4 205 14~ 152 11 63 3 57 3 43 3 44 3 44 3 44 3 66 5 59 4 73 5 108 7 107 6 72 5 81 5 126 7 189 10) i! U22 U33 U12 U13 U23 670 6 604 6 138 5 107 5 86 8 95 9 246 7 178 6 139 5 117 4 140 6 85 3 158 5 148 10~ 336 20 54 4 53 4 55 3 62 4 54 4 54 3 75 5 69 5 82 6 71 6 63 5 60 5 41 4 55 5 49 5 633 6 655 6 69 4 166 6 96 8 122 9 190 6 136 4 157 5 101 3 452 14) 142 158 46 124 10) 54 3 78 4 44 3 46 3 45 3 42 3 66 5 47 4 72 5 96 6 84 6 73 5 66 5 97 6 141 8 -7 5 45 25 4 14 4 16 6 56 16 5 21 4 29 3 37 3 -19 4 -1 3 -59 4 -58 10~ -18 12 -1 3 -8 3 -4 3 -8 3 33 33 12 4 -7 4 -17 4 -23 5 95 54 24 12 5 -6 6 114 5 208 5 -13 3 62 5 16 6 -2 6 92 4 84 34 4 -15 3 25 7 45 4 73 13 8 97 9 23 3 40 3 20 3 19 3 14 3 18 3 94 30 4 40 4 64 5 61 5 28 4 46 4 72 5 123 8 -10 5 -64 5 -3 4 -62 4 -16 6 88 35 5 -84 4 43 4 -21 3 78 8 15 ·3) ~-51 4 46 -31 12) -2 3-8 3 53 13 23 23 -3 5 -3 4 -22 5 -30 5 65 -5 4 54 64 75 i~ Table VII. ~ C14 C15 C7 C10 C9 C17' C6' C5' C4' C3' C2' Cll' C12' C13' C14' C15' C7' Cl0' C9' - Atomic Parameters and esd's for the Nonhydrogen Atoms in C~(TPEN)(CO) 2 (BF4 ) 2 , 3 (cont'd) 5651 6 5637 4 7977 4 7719 4 8280 3 7984 4 9486 3 10103 4 10375 4 10044 5 9423 4 7083 3 6405 4 5767 4 5825 4 6518 4 9167 3 7827 3 7987 3 -956 8 -208 7 1766 7 -495 6 914 6 4132 7 1372 6 979 7 1827 9 3015 8 3332 6 886 6 313 6 1070 7 2353 7 2875 6 539 6 129 6 166 6 4449 5 3869 4 4251 3 3935 3 3246 3 2268 4 1741 3 1525 4 1132 5 965 4 1186 3 780 3 262 3 -202 3 -122 3 417 4 2192 3 1242 3 2526 3 138 8 77 5 53 4 63 4 49 4 72 5 38 4 64 5 69 5 90 6 63 5 51 4 76 5 55 4 55 4 61 4 49 4 55 4 51 4 63 6 62 5 84 5 65 5 71 5 77 6 76~~ 88 6 141 8 104 7 78 5 52 4 52 5 86 6 78 5 49 4 69 5 50 4 61 4 170 9 117 6 45 4 58 4 54 4 91 6 51 4 93 6 113 7 80 6 51 4 39 4 61 5 50 4 55 4 71 5 59 4 54 4 58 4 -27~5 ·- 18 4 -11 4 11 4 -3 3 74 24 -1 4 -19 5 -42 5 -21 4 -4 4 -8 4 -12 4 74 -3 4 16 4 2 3 -5 3 119 8 54 5 17 3 27 4 17 3 26 4 20 3 40 4 63 5 41 5 17 4 21 3 31 4 84 18 4 25 4 ·20 3 23 3 26 3 5 -21 6 -29 5 -2 4 12 4 34 -28 5 54 -2 5 -12 6 -3 5 -1 4 -1 3 -5 4,· (X) -7 4 (0 15 4 64 84 -2 3 0 4) aFractional coordinates have been multiplied by 10 for the copper atoms and 104 otherwise. u1).. has been multiplied by 104 for the copper atoms and 10 otherwise. bThe form of the thermal 3 ellipsoid is exp[-211'2 (U11 h2 a*2 + .•. + 2 U23Klb*c*)] for the anisotropic thermal parameters. 90 Table VIII. Hydropeters for Cui(TPEN)(CO) 2 (BF4 ) 2 , ~ ~ H2 H3 H4 H5 H7A H7B H9A H9A Hl0A Hl0B H12 H13 H14 H15 H2' H3' H4' H5' H7A' H7B' H9A' H9B' Hl0A' Hl0B' H12' H13' H14' H15' y z B,A02 414 605 607 422 146 187 175 62 -109 -63 -179 -198 -108 17 374 286 70 -56 6 -68 14 -65 72 -30 18 159 358 413 299 345 431 466 462 443 315 354 359 433 505 530 453 357 47 -41 -55 22 100 131 233 262 268 213 164 98 72 108 6. 1 6.8 7.0 6.4 5. 5 5. 5 5.4 5.4 5. 6 5. 6 7.7 8.6 8. 1 7.3 5. 8 6.0 6. 3 5. 8 5. 2 5.2 5. 1 5. 1 5.6 5.6 7. 3 8.4 7.6 6.0 ~The x, Y attonal coordinates have been multiplied by 10 3• 91 - Table IX. Interatomic Distances for Cu2 (TPEN)(CO) 2 (BF4 ) 2 , 3. ~ ' Bonding Distances Cu-Nl Cu-Nl6 Cu-NB Cu-C17 Nl-C6 Nl-C2 Nl6-Cll Nl6-Cl5 N8-C7 N8-Cl0 N8-C9 C17-O18 C6-C5 C6-C7 C5-C4 C4-C3 C3-C2 Cll-C12 Cll-Cl0 C12-C13 Cl3-C14 C14-C15 C9-C9' Bl-Fl B1-F2 Bl-F3 B1-F4 2. 042(5) 2.030!5! 169 5 1. 809 7 1. 341 8 1. 341 9) 1.340 1. 346 1. 482 1.479 i! 1. 465 1.100 1. 396 1. 487 i! 1. 353 11~ 1. 366 11 1. 369!11 1. 375 9) 1. 500~9) 1. 391 lll 1. 350 12 1.369 11) 1. 515 9) 1. 333 13~ 1. 347 13 1. 362 13 1.340 14) Cu'-Nl' Cu' -Nl6' Cu' -NB' Cu' -Cl 7' Nl' -C6' Nl' -C2' Nl6'-Cll' Nl6'-C15' NB' -C7' NB' -Cl0' NB' -C9' C17'-O18' C6' -C5' C6' -C7' C5' -C4' C4' -C3' C3' -C2' Cll'-C12' Cll '-Cl0' C12'-C13' Cl3' -C14' Cl4'-C15' B2-F5 B2-F6 B2-F7 B2-F8A B2-F8B 2. 036(5) 2. 032!5! 160 5 1. 781 7 1. 330 8 1. 334(9 1. 329 8 1. 326 8 1. 489 8 1. 481 8 1. 476 8 1.110 10) 1. 372 1.499 1. 387 1.340 10) 1. 383 10) 1. 378 9) 1. 494 8) 1. 374 11) 1. 350 lll 1.368 10 1. 369 14) 1. 327 1. 301 12! l. 247 16 1. 356 19 !~ Nonbonding Contacts Nl-N16 Nl-NB Nl6-N8 Cu-Cu' 3. 760 353!8! 2. 7 2. 717 7 4. 764 1 Nl' -N16' Nl' -NB' N16'-N8' 3. 760 293~7~ 2. 7 2. 735 7 92 - ~al_?!~~-- Interatomic Angles (deg) for Ct½(TPEN)(CO) 2 (BF4 ) 2 , 3. ..----~ N16-Cu-Nl NB-Cu-Nl Cl7-Cu-Nl N8-Cu-Nl6 Cl7-Cu-Nl6 Cl7-Cu-N8 C6-Nl-Cu C2-Nl-Cu C2-N1-C6 Cll-N16-Cu C15-N16-Cu C15-N16-C11 C7-N8-Cu C10-N8-Cu C9-N8-Cu Cl0-N8-C7 C9-N8-C7 C9-N8-Cl0 O18-C17-Cu C5-C6-Nl C7-C6-Nl C7-C6-C5 C6-C5-C4 C3-C4-C5 C2-C3-C4 C3-C2-Nl C12-Cll-Nl6 C10-C11-N16 Cl0-C11-C12 C13-C12-C11 C14-Cl3-C12 C15-Cl4-Cl3 C14-Cl5-N16 C6-C7-N8 C11-C10-N8 C9' -C9-N8 F2-Bl-F1 F3-B1-F1 F4-Bl-Fl F3-Bl-F2 110. 9 4 81. 9 4 116. 7 4 80. 6 4 122. 1 4 135. 6 4 114. 4 5 126. 9 5 117. 9 6 115. 8 5 125. 0 5 118. 7 6 103. 3 5 104. 7 5 116. 8 5 109. 9 6 109. 1 6 112. 5 6 174. 7 6 121. 7 6 115. 6 6 122. 7 6 118. 8 7 120. 1 8 118. 7 8 122. 8 7 122. 5 6 115. 1 6 122. 4 6 118. 3 7 118. 5 8 121. 3 8 120. 6 7 112. 9 6 110. 9 6 113. 4 6 111. 5 9 105. 9 9 110. 5 9 107. 7 9 Nl6' -Cu' -Nl' NB' -Cu' -Nl' Cl 7' -Cu' -Nl' N8'-Cu' -Nl6' Cl 7' -Cu' -N16' Cl 7' -Cu'-N8' C6'-Nl' -Cu' C2' -Nl' -Cu' C2' -Nl' -C6 Cll' -Nl6' -Cu' C15' -C16' -Cu' C15' -C16' -Cll' C7' -NB' -Cu' Cl0' -NB'-Cu' C9'-N8' -Cu' Cl0' -NB' -C7' C9'-N8' -C7' C9' -N8-' ~C10' O18'-Cl 7'-Cu' C 5 '-C 6' -Nl' C7' -C6' -Nl' C7' -C6'-.C5' C6' -C5' -C4' C3' -C4' -C5' C2' -C3' -C4' C3' -C2' -Nl' Cl2' -C11' -Nl6' Cl0' -Cll' -Nl6' Cl0' -Cll' -C12' Cl3' -C12' -Cll' Cl4' -C13' -Cl2' C15' -C14' -Cl3' Cl4' -C15'-Nl6' C6' -C7' -NB' C 11' -C 10' - NB' C9-C9' -NB' F6-B2-F5 F7-B2-F5 F8A-B2-F5 F8B-B2-F5 93 - Table X. Interatomic Angles (deg) for Cl½(TPEN)(CO) 2 (BF4 ) 2 , 3. (cont'd) ~ F4-Bl-Fl F4-Bl-F3 110. 3(9) 110. 9(9) Nonbonding Angles Nl6-Nl-N8 Nl-N8-Nl6 N8-Nl6-Nl Nl6' -Nl' -NS' Nl '-NS' -Nl6' NS' -Nl6' -Nl' F7-B2-F6 F8A-B2-F6 F8B-B2-F6 F8A-B2-F7 F8B-B2-F7 F8B-B2-F8A F8B-F8A-B2 F8A-F8B-B2 112. 8!8) 125. 1 10) 102. 7 11) 121. 0 10) 103. 16(11) 56. 5~12i 66. 3 26 57. 3 27 94 in the crystal structure. This is evidenced by the slightly different bond lengths and angles for each half of the molecule, which are listed in Tables IX and X. The origin of this asymmetry is not obvious, especially since no intermolecular interactions less than 3. 6 Awere observed. - The stereoview of 3 shown in Figure 6 depicts a distorted tetrahedral geometry around the copper ions. Each copper is bound to three ligand nitrogens, with two short Cu-N distances of ca. 2. 04 A (between copper and the pyridine nitrogens) and one long Cu-N distance of ca. 2. 17 A (between copper and the tertiary amine nitrogen). While the shorter Cu-N bond lengths are within the expected range for cu1-N bonds, 2 , 3,-5, 6 the Cu1-N (tertiary amine) is rather long, possibly indicating a lower affinity between the cuprous ion and the "hard" tertiary amine ligand. The Cu-C bond lengths average 1. 80 A, and the C-O bond lengths average 1. 11 A with a Cu-C-O angle of 176°. These values are in agreement with other previously reported four- and five-coordinate copper(!) carbonyls. 2 , 5, 6 The distorted environment around each copper is also evident in the observed bond angles. Two of the N-Cu-N bond angles around each copper average 82° (between either short Cu-N bond and the long Cu-N bond), and the remaining N-Cu-N angle averages 110° (between the short Cu-N bonds). Each copper atom is ca. 1. 14 A above the basal plane defined by the three nitrogen ligands, with carbon monoxide occupying an apical position. The three dihedral angles between the planes formed by the atoms of the inner coordination sphere (Nl, N8, Nl6, Cl 7 and the copper atom) are listed in Table XI. 95 ~ Stereoview illustrating the distorted tetrahedral ligand environment around each copper in Cl½(TPEN)(CO)/+. Table XI. Plane Calculations for CUa(TPEN)(BF4 ) 2 , 3. ~ Plane 1 2 3 4 5 6 7 8 9 10 11 12 13 _14 Atoms in plane Dihedral angle equation of plane (Ax + By + Cz = D) between planes A B C D 0.0266 0.3610 0.2111 0.9675 0.9247 0.2914 0.5637 0.7530 0.8486 0.0968 0.0956 0.9050 0.2081 o. 5373 0.1148 -0.6378 0.7365 -0.0104 -0.2665 0.9545 0.3041 -0.6254 -0.0441 -0.5098 0.9194 0. 1286 0.1868 0.6085 0.9054 -0.7671 0.6743 0.6075 -0.1073 -0.1718 -0.9263 -0.4801 0.8146 0.7508 -0.3779 0.0237 0.8046 -0.7465 6.919 L 833 2.542 7.446 9.576 3.567 1. 364 4.941 9.245 2.521 1. 994 6.018 6. 312 6.888 Nl, N8, N16 Nl' N8' N16' ' Cu, ' N8, N16 } Cu, Nl, C17 Cu, Nl, N16 } Cu, N8, Cl 7 Cu, Nl, N8 } Cu, N16, Cl 7 Cu' Nl' N8' Cu' ' N16'' Cl 7' } Cu' ' N16' ' Nl' } Cu' ' NB' ' C17' Cu'' N16 1 NB' } Cu' ' Nl' ' C17' ' 68.0° 89.9° 67. 0° -· 69.7° 90. 1° 70.4° ' Deviations from Plane Plane 1 2 Cu -1. 1141 C17 -2.8618 Cu' C17' -1. 1377 -2. 8714 co 0) 97 Deviations from 90° (expected for a perfect tetrahedron ) indicate a distorted environment. These distortions are probably the result of constraints from the polydentate ligand, and these constraints are due to the presence of only two carbon atoms between each ligating nitrogen. A lower affinity of these cuprous ions toward the "hard" tertiary amine nitrogens may also contribute to the unusual geometry. Only the outer atoms (oxygen of carbon monoxide and the outer carbons of the pyridines) of the dication are undergoing any significant thermal vibrations, but both tetrafluoroborate anions have significant anisotropic motions. One anion is well ordered, the boron atom (Bl) remaining positionally fixed (as indicated by its small temperature parameters) with the four fluorines (Fl- F4) undergoing thermal vibrations. The second anion is positionally disordered at the boron (B2) and at two fluorines. One of these fluorines, (F8) is clearly disordered with a 50% occupancy at each of two positions separated by 1. 3 A. Another fluorine (F7) is similarly disordered with a 0. 5 A positional separation. The center boron atom ( B2) also occupies two positions, but the separation is less than 0. 5 A. As a result, anisotropic temperature factors adequately described the electron disti:-ibution of the boron atom (B2) and the small positional disorder of one fluorine (F7), while two positions (FSA and F8B) with anisotropic temperature parameters were required for the other disordered fluorine. The two remaining fluorines of this anion are positionally settled and are not significantly involved in the disorder of the other atoms. 98 Discussion ~ The apparent preference of cuprous ions for lower coordination numbers (i.e., two, three, and four) 1 ' 3 suggested the possibility of the hexadentate TPEN ligand performing as a bridging ligand between two copper(!) ions. Indeed, a binuclear cuprous complex employing the TPEN ligand was isolated. The structure of Cl.½(TPEN)(BF4 ) 2 - (2), however, was somewhat surprising. Rather than the expected three or four (including BF4 as a ligand) coordination, each copper appears to bind only two nitrogen ligands. In fact, the nitrogen ligands originate from pyridine rings on opposite ends of the ethylenediamine bridge (Figures 1 and 2). Apparently the molecule adopts this conformation in order to maximize the Cu-Cu interaction and minimize the Cu-N (tertiary amine) interaction. (Molecular models corroborate this explanation. ) - The structure presented herein for 2 as well as the structure reported for 1 3 indicate a preference of copper(I) toward a somewhat - linear arrangement of two nitrogen ligands. 9, 12 (This was observed - in 1 by very short Cu-N distances and rather long Cu-O distances.) Note, however, that in both cases, the potential third ligand (i.e., tertiary nitrogen or phenolate oxygen) is rather "hard" and hence unfavorable for copper(I). Therefore, the observed two coordinate character of these compounds may simply be an avoidance of undesirable - - ligands. Both systems (1 and 2) apparently foster a direct Cu-Cu interaction. 9 , 13 - Cl.½(TPEN)(BF4 \ (2) undergoes an extensive ligand rearrange- 99 - ment in the presence of carbon monoxide to form the dicarbonyl, 3. The crystal structure of this complex exhibited the "expected" - pseudotetrahedral geometry. In 3, all nitrogen ligands originate from the same side of the ethylenediamine bridge. In the present system - - (i. e. , 2 and 3) the preferred coordination about copper depends on the presence or absence of a potential fourth ligand (CO or CH 3CN). - On the contrary, the previously examined system (1) showed no tendency to bind CO or pyridine as a fourth ligand. The reason for this difference is not clear since molecular models indicate that - complexes such as 1 could form a pseudotetrahedral geometry with the addition of a fourth ligand. Thus, there appears to be a special - stabilization for cuprous ions in 1 such that significant energy would 12 be lost upon the binding of a fourth ligand. The origin of such stabilization, however, remains obscure. The structures reported in this paper, as well as the structures 3 found for other cuprous systems, l- demonstrate that a wide variety of coordination environments can be assumed by copper(!) ions. Obviously, the number and geometry of ligands pref erred by copper(!) are extremely sensitive to minor ligand alterations. These effects have not been well defined, and as a result, one must exercise caution in the prediction of structures for cuprous compounds. A few final points about the structure of Cl.½(TPEN)(CO) 2 2+ deserve mention. This structure bears a close resemblance to the 100 structure recently reported for Cl.½(en)s{CO)/+ (en= ethylenediamine). 5, 14 In this latter complex, the two centrosymmetric copper ions are bridged by an ethylenediamine, and a chelating en and a CO complete a pseudotetrahedral geometry around each copper. Although the atoms of the carbonyl group were found to be distributed over two positions (with C-O bond distances differing by 0. 01 Aand the Cu-C bond distances differing by 0. 02 A), only one CO stretching frequency 15 This is most likely due t.o a was observed (vco = 2078 cm- 1). dynamic equilibrium between the two carbonyl sites, with a rate faster than the LR. ti.me scale (-10 12 sec- 1 for~ vco of 10 cm- 1 ). This contrasts with ClJ..?(TPEN)(CO) 2 (BF4 ) 2 which exhibits two CO stretching 1 frequencies (vco = 2097, 2107 cm- ), which apparently arise from two distinctly different copper centers (not in equilibrium in the solid - state). This system (3) demonstrates that minor alterations in bond lengths and angles of metal carbonyls can be manifested by different CO stretching frequencies. 16 101 Materials. All chemicals were reagent grade and were used ~ as received unless otherwise noted. N, N, N', N' -tetrakis-(2-pyridylmethyl}-ethylenediamine (TPEN) was prepared by a modification of the literature method. 17 Tetra(acetonitrile) copper(!') tetrafluoroborate was also prepared by the published procedure. 18 The following was performed under helium in a Vacuum Atmospheres Dri-Lab inert atmosphere chamber. Cu(CH 3CN) 4 BF4 (O. 315 g, 1 mmol) was added to a solution of TPEN (0. 21 g, 0. 5 mmoles) in 25 ml of acetonitrile. After stirring for 10 min at ambient temperature, the solvent was removed under vacuum. Addition of methanol (25 ml) to the resulting orange oil yielded slightly yellow-green crystalline product. The product was · collected by vacuum filtration, washed with methanol, and dried for several hours in vacuo. Anal. (C 2 Jl28N6C~B2 F8 ) C, H, N. - A suspension of 2 ( ca. 100 mg) in 20 ml of methanol (under heliUllV was placed in an Erlenmeyer flask and covered with a rubber septum. Purging the system with CO caused dissolution of the solid. Evaporation of the solvent with a - slow stream of CO yielded off-white crystals of 3. Anal. (C 28H28 N60 2 C~~F8 ) C, H, N; Cu: calcd, 16. 3; found, 16. 8. Ph sical Measurements. Sample preparation for physical studies on the air-sensitive materials was accomplished in a Vacuum Atmospheres Dri-Lab inert atmosphere chamber, under a helium 102 atmosphere. Helium saturated spectroquality solvents were used for solution studies. Electronic spectra were recorded on a Cary 14 spectrophotometer. Solid state spectra were obtained from Nujol mulls on filter paper ('under He) against a Nujol-saturated filter paper as a blank. Infrared spectra were recorded on a Beckman m-12 Infrared Spectrophotometer. Solid state spectra were obtained from Nujol mulls pressed between KBr plates. Solution spectra were obtained using calcium fluoride solution cells (path length of 1 mm). Proton magnetic resonance spectra were recorded on a Varian EM390 spectrophotometer at 90 MHz (34°C). The solvent utilized was CD 3CN containing Tl\ifS as the reference. Elemental analyses were performed by the Caltech Microanalyti cal Laboratory. Collection and Reduction of X-Ray Intensity Data for ~ . Yellow needles of Cl½(TPEN)(BF4 ) 2 , ~' were grown by slow evaporation (under helium) of a solution of ,.. 2 in acetonitrile/bis(2-methoxyethyl)ether (ca. 1:1). Preliminary Weissenberg and precession photographs showed the crystals to be monoclinic with the z axis parallel to the needle axis. Two alternatives for the space group, Cc (no. 9) or C2/ c (no. 15) were indicated by systematic absence of the following reflections: hkl, h + k = 2n + 1; hOI, 1 =2n + 1. The space group was shown to be C2/ c by satisfactory solution and refinement of the structure. 103 A crystal of dimensions 0. 31 mm x 0. 12 mm x 0. 07 mm which had been cut from a longer needle, was mounted for data collection on a Syntex P2 1 four-circle diffractometer using graphitemonochromatized MoKa radiation. Cell parameters were determined by a least-squares fit to 15 automatically centered reflections with 18° < 20 <25°. The resulting parameters are given in Table II. Intensity data were collected from four octants (excluding reflections for which k + I= 2n + 1) using 0- 20 scans. The maximum 20 value was 45° for data in the hkl and -hkl octants, and 25° for data in the h-kl and - h-kl octants. The scan range extended from 1° below the MoK al 20 value to 1° above the MoK~ value, and the scan rate was 1° /min., with the total background counting time equal to the total scan time. Three check reflections were measured after every 60 reflections to monitor the crystal and instrument stability. These showed no systematic variations throughout the course of data collection. The data were prepared for solution of the structure by first applying Lorentz and polarization corrections. Standard deviations of intensities were calculated using the formula where S, Bl, and B2 are the scan and two background counts and d was taken as 0. 02. 19 Absorption corrections were calculated by the method of Gaussian (Jladrature. 20 After symmetry-extinct reflections were deleted, data from the hkl and -hkl octants were averaged with 104 those from the h-kl and -h-kl octants, respectively. There remained 1941 unique data, with 1598 of these greater than zero. Solution and Refinement of the Structure for C With the exception of C. K. Johnson's ORTEP program, all computer programs were from the CRYM system. All scattering factors (Cu+, F, N, C, B) with the exception of those for hydrogen21 were taken from the International Tables, 22 as was the real part of the anomalous dispersion correction for copper. The function minimized in the least-squares refinement was "z:iw(F0 2 - F c2 )2, where the weight w = l/0-2 (F0 2 ). The position of the copper atom was determined by means of a three-dimensional Patterson synthesis. Under the assumption that the correct space group was C2/ c, this atom was used as a phasing model for a three-dimensional Fourier map. This map revealed the positions of most of the carbon and nitrogen atoms. A cycle of least-squares, followed by another Fourier map, gave the positions of the remaining nonhydrogen atoms. Full matrix least-squares refinement including all of these atoms with isotropic temperature factors led to R = L 11 FO I - I F c 11 /L I F O I = 0. 111, for those reflections with F O 2 > 3a. Use of anisotropic temperature factors lowered R to 0. 067. Hydrogen atom positions were calculated assuming a carbon-hydrogen distance of 0. 97 A. All of these positions corresponded to regions of positive electron density in a difference Fourier map. Hydrogen atoms were included in the least-squares with the isotropic temperature factor for 105 2 each fixed at 1. 0 A. greater than that previously determined for the bound carbon atom. The hydrogen positional parameters were not refined but were recalculated periodically. The large and highly anisotropic temperature factors which had been determined at this point for the fluorine and boron atom~ indicated a probability of disordering of the BF4 - anion. A detailed Fourier map of the anion showed that the fluorine atoms occupied two different sets of positions, although one of these sets had a much lower occupancy than the other. Consequently, two partial BF4 groups were included in the least- squares along with a single population factor. The fluorine atoms in the more and less heavily occupied positions were refined with anisotropic and isotropic temperature factors, respectively. The boron atom for each of these BF4 - groups was placed at coordinates calculated by averaging the corresponding fluorine coordinates. These coordinates were not refined but were ·r ecal~lated periodically (final boron-boron separation = 0. 50 X). The isotropic temperature factors for the boron atoms were refined. The final, refined populations of the more and less heavily occupied positions were 0. 71 and 0. 29, respectively. The final refinement of the structure was done by blockdiagonal least-squares with two matrices. All nonhydrogen positional parameters were included in one matrix; the second contained all nonhydrogen thermal parameters, the BF4 - population factor, and the scale factor. No reflections were omitted from this refinement. The 106 final R was 0. 051 for the 665 reflections with F O2 > 3a, and 0. 143 for 2 .!. all data. The goodness of fit, [ ~w(FO - F /) 2 /(m-s) J2 was 1. 18, where m = 1941 is the number of observations and s = 209 is the number of parameters. Features on a final difference map were between -1. 4 e/ A3 ( at the posi ti.on of Cu) and 1. 2 e/ A3 ( near F4A). Final atomic parameters are given in Table III and IV. Collection and Reduction of X-Ray Intensity Data for An approximately cubic crystal, 0. 2 mm on an edge, was sealed in epoxy to prevent decomposition and was used for all X-ray diffraction measurements. Preliminary oscillation and Weissenberg photographs showed Laue symmetry 2/m with systematic absences h0l with 1 odd and 0k0 with k odd, indicating the monoclinic space group P2/ c; the crystal was then transferred to a Syntex P2 1 diffractometer operating with Ni-filtered CuKa radiation. Unit-cell dimensions were obtained by centering 15 well-scattered reflections with 20 values ranging from 23° to 48°. Intensities were measured using a 0-20 scan at a rate of 2° /minute with an equal amount of time spent on backgrounds; three check reflections, monitored every 100 measurements, showed no intensity decay nor abnormal fluctuations. All reflections out to 2 0 = 100° were measured in two different quadrants (h, k, 1 and h, -k, -1) and were averaged to yield 3538 net intensities, of which 2998 were greater than zero. The initial attempt at solving the structure was by manual l application of direct methods based on 95 reflections with IE greater 107 than 1. 85, three origin-fixing signs and two symbolic signs. The resulting E map was not structurally reasonable because, as it turned out, one of the early sign entries was reversed. A three-dimensional Patterson map was then calculated, and quickly led to the positions of the Cu atoms. The C, N, 0, B and F atoms were recovered from F O and clifference maps; the positions of the hydrogen atoms were calculated using a C-H bond length of 0. 9 .A. In the least-squares refinement the same quantity was minimized as in the preceding structure, with weights calculated in the same manner. A late clifference map indicated rather severe disorder for one of the BF4 - groups: while three of the fluorine atoms could be adequately represented by large, anisotropic B's, the fourth could not and accordingly was split into two half-occupied sites separated by about 1. 2 A. In the final least-squares cycles, three matrices were collected, each of order between 140 and 150; the coordinates of the 49 nonhydrogen atoms in one, anisotropic B's for one-half the cation and one of the BF4 groups in a second, and anisotropic B's of the remaining atoms,' a scale factor, and an extinction parameter in the third. The final value of the goodness-of-fit was 1. 76 for the 3538 measured reflections; the R factors were 0. 062 for the 2998 reflections with net intensity greater than zero and 0. 042 for the 2271 reflections with intensities greater than 3 esd's above background. Features on a difference map were between -0. 8 and 0. 6 e/ A3, in generally uninteresting places. 108 All calculations were carried out on the CRYM system of -programs. Final atomic parameters are given in Tables VII and VITI, based on the numbering scheme shown in Figure 5. ~ We appreciate helpful discussions with Grant A. Mauk and financial assistance from the National Institutes of Health (Grant No. PHS AM18319), and the International Copper Research Association. 109 References and Notes · (1) Jardine, F. H. Adv. Inorg. Chem. Radiochem. ~ ' 17, 115163; Eller, P. G. ; Bradley, D. C. ; Hurs thouse, M. B. ; Meek, D. W. Coord. Chem. Rev. ~ ' !!, 1-95; Camus, A. ; Marsich, N.; Nardin, G.; Randaccio, L. Inorg. Chim. Acta 1977, -23, 131-144. ~ (2) Gagne, R. R. ; Allison, J. L. ; Gall, R. S. ; Koval, C. A. J. Amer. Chem. Soc. ~ ' ~ 7170-7178. (3) Gagne, R. R.; Kreh, R. P.; Dodge, J. A. J. Amer. Chem. Soc. ~ ' 101, 6917-6927. (4) Similar downfield shifts have been observed for imidazole protons upon binding of the imidazole nitrogen to a cuprous ion: Auguira, Y. Inorg. Chem. ~ ' 17, 2177-2182. (5) Pasquali, M.; Floriani, C.; Gaetani-Manfredotti, A. Inorg. Chem. 1980, 19, 1191-1197. ~- (6) Bruce, M. I.; Ostazewski, A. P. J. Chem. Soc. Dalton Trans. 1973, 2433-2436; Churchill, M. R.; DeBoer, .B. G.; Rotella, ~ F. J. ; Abu Salah, 0. M. ; Bruce, M. I. Inorg. Chem. ~ ' 14, 2051-2056. (7) Mehrotra, P. K.; Hoffmann, R. Inorg. Chem. ~ ' !1, 2187- (8) 2189, and references cited therein. The short cu1-N bond lengths found here are re minis cent of the short cu1-N bond lengths (1. 88-1. 90) which were found for 1. These are shorter than most Cu1-N bond lengths which have - 110 been reported and may be characteristic of two coordinate copper(!}. 3, 9 {9} Somewhat similar structures were observed for tetrakis[ 1, 3-dimethyltriazenocopper(I)] lO and bis- [diazoaminobenzenecopper(I)] 11 in which each copper is bound linearly to two nitrogens, and in addition, close Cu-Cu contacts were observed. {10) O'Connor, J. E. ; Janusonis, G. A. ; Corey, E. R. J. Chem. Soc. Chem. Commun. .,.._,,..,__ 1968, 445-446 . (11) (12) ----In general, two coordinate copper(!) complexes are found with Brown, I. D.; Dunitz, J. D. Acta Crysl 1961, 14, 480-485. strong a-donor nitrogen ligands (e.g., Cu(NHJ/) while four coordination appears to be favored with 1r-acceptor nitrogen ligands (e.g., Cu(pyridine)/): James, B. R.; Williams, R. J.P. - J. Chem. Soc. 1961, 2007-2019. (13) Mehrotra and Hoffman have reported molecular orbital calculations which indicate a positive attraction between cuI (d10'} centers through the mixing in of metal s and p orbitals. 7 (14) A similar structure was also found for Cl½(histamine)iCO)/+, but here each copper has a cliff erent ligand environment, resulting in different bond lengths and angles for each carbonyl, and two CO stretching frequencies were observed (vco = 2055, 1 2066 cm- ): Pasquali, M; Floriani, C. ; Gaetani-Manfredotti, A. ; - Guastini, C. J. Chem. Soc., Chem. Commun. 1979, 197-199. 111 (15) It should be noted, however, that the differences in bond lengths do not exceed the standard deviations given. The differences in bond angles for the two carbonyl sites are more significant {16) Two dlff erent molecules were found in the crystal structure of [hydrotris(l-pyrazolyl)borato] copper(!) carbonyl, and while the Cu-C bond lengths differ by 0. 02, the C-O bond lengths are equivalent, and only one CO stretching frequency was observed 1 6 (v CO = 2083 cm - ). (17) Anderegg, G.; Wenk, F. Helv. Chim. Acta -.,..,..,...,.. 1967, -50, 23302332. (18) Hathaway, B. I.; Holah, D. G.; Postlethwaite, J. D. J. Chem. Soc. 1961, 3215-3218. --.,..,..,...,.. (19) Peterson, S. W.; Levy, H. A. Acta Cryst. ~ ' 10, 70-76. (20) Busing, W. R.; Levy, H. A. Acta Cryst ~ , 10, 180-182. (21) Stewart, R. F.; Davidson, E. R.; Simpson, W. T. J. Chem. Phys. ~ , 42, 3175-3187. (22) ''International Tables for X-ray Crystallography", Vol. m; Kynoch Press: Birmingham, England, 1962. 112 CHAPTER IV · Dioxygen Reactions of Three-Coordinate Binuclear Copper(!) Complexes, Leading to the Design and Attempted Synthesis of Improved Models for Copper Proteins The synthesis and characterization of an extensive series of - binuclear copper(!) complexes, 1, are presented in Chapter II (Table I). R= , 1 X= \ etc. I I / N-N , ' fNIIN\ V ~ The compounds were synthesized as potential models for the binuclear copper protein sites in hemocyanin, tyrosinase and laccase. These proteins are responsible for reversible dioxygen binding (hemocyanin), dioxygen activation (tyrosinase) 2 and the reduction of dioxygen to water (Iaccase). 3 Hence, it was of interest to investigate the dioxygen 1 113 Table I. Binuclear eu1cu1 Complexes and Abbreviations Utilized. Compound Nwt:iber Sidearm (R) Bridge 2 ..... 2-(pyridyl)-methyl pyrazolate 3 ,.. 2-(pyridyl)-methyl 3-5 dimethylpyrazolate 4 2-(pyridyl) -methyl 7-azaindolate 92 2-(2' -pyridyl)-ethyl pyrazolate 2-(2' -pyridyl)-ethyl 3-5 dimethylpyrazola.te ! 2-(2' -pyridyl)-ethyl 7-azaindolate cu! IOOIM(Etpy) (aza) 8 ..... phenylmethyl pyrazolate ~ ISOIM{MePh) 2 {pz) i phenylmethyl 3-5 dimethylpyrazolate !Q phenylmethyl 7-azaindolate 11 2-phenylethyl pyrazolate 12 2-phenylethyl 3-5 dimethylpyrazolate 13 2-phenylethyl 7-azaindolate 14 1-propyl pyrazolate cu! ISOIM(EtPh)2(aza) cu! ISOIM(l-Pr)2(pz) 15 2-propyl pyrazolate cu! IOOIM(2-Pr) {pz) ~ ISOIM(t_-Bu) 2 (pz) " """ "" "" "" "" Abbreviation (X) cu! OOIM(Mepy) (pz) cu! ISOIM(Mepy) (Me2pz) 2 2 cu! OOIM(Mepy) (aza) 2 ~ ISOIM(EtpyMpz) / cu! OOIM(Etpy) (Me pz) 2 2 2 cu! OOIM(MePh)2(Me2PZ) cu! ISOIM(MePh)2(aza) cu! IroIM(EtPh) (pz) 2 cu! OOIM(EtPh) (Me2Pz) 2 2 !i ll !_-butyl pyrazolate !_-butyl 3-5 dimethylpyrazolate 18 !_-butyl 7-azaindolate cu! ISOIM(t-Bu'2(aza) ~ phenyl pyrazolate ~ ISOIM(Ph'2(pz) ~ p-( dimethylamino)phenyl pyrazolate p-acetylphenyl pyrazolate bis-(2-pyridyl)methyl pyrazolate cu! ISOIM(PhNM~) (pz) cu! ISOIM(PhCOMe) (pz) cu! IroIM(Mepy (pz) 1-adarnantyl pyrazolate C~ ISOIM(ada) 2 (pz) "" -ll 21 23 "" cu! OOIM(!_-Bu) 2(Me pz) 2 2 2 2) 2 114 - reactivity of these synthetic binuclear cuprous systems, 1. All compounds listed in Table I are unreactive towards diaxygen in the solid state, but they are oxidized by dioxygen in DMF solution (as evidenced by an orange _. green color change). The lack of solid state reactivity may be due to the necessity for pyrazole dissociation prior to reaction / with 0 2 (vide infra). The rate of reaction is (qualitatively) much slower for compounds which do not contain potential ligands (i.e. , pyridine nitrogens) on the "sidearms" (R). This is not surprising, since the presence of these pyridine nitrogens was found to stabilize the oxidized Cu(II) species (see Chapter II). Attempts to reverse the 0 2 reaction by purging with argon were unsuccessful at ambient temperature and at -50° C. In most cases the products from the oxygen reactions have not been characterized, but oxygen uptake measurements were performed on a series of the cuprous complexes and the results are shown in Figure 1. The reaction~ were run at both -79°C and at ambient temperatures (20-25 ° C), but the temperature did not affect the amount of dioxygen consumed. A stoichiometry of 0. 5 O2 /C~L (where L is the entire ligand system) corresponds to the complete reduction of diaxygen to water (or hydroxide or oxide) by four copper(!) ions from two - binuclear species. This uptake was found for 5, which contains 2-(2' -pyridyl)ethyl groups on the "sidearms" (R). In contrast, compounds i and ll exhibit a stoichiometry of 1. 5 O2 /Cl½L. This high value is indicative of ligand oxidation, and the likely targets are the - methylene and methine groups of the ligands in 2 and .,._..._ 22, respectively. The activation of these positions by both the imine nitrogen and the 115 Cu(I)Cu(I)L + o ? 2 0 2 : Cu (I)Cu(I)L A r~Y'\.Y o N... A~~ 'Lf',.. c~c;tt ~ V N-N 0.5 V 5 1.5 N A{ ·N,...I ; 0 N ::-. . \ / ' / N Cu Cu N-N - V N 1 ~ 1.5 ~ 22 ~ Oxygen uptake measurements in CH2 C~/ solutions. The values given here were observed at both -79°C and ambient temperatures. _J 116 pyridine rings may lead to oxidation by dioxygen. 4 The binding of dioxygen to the cuprous ions is apparently necessary for this ligand oxidation since no ligand oxidation is observed for the analogous copper(II) complex, .,...,.,..,. 24, even in the presence of excess 2+ / '- L uJI Cu 'OH ' / 24 pyrazole. The mechanisms and products of these reactions have not been investigated, but hydroxylations of the carbons adjacent to the imine nitrogens are plausible. 5 This would account for the uptake of 1 o 0 O2 /Cl¼L in addition to the 0. 5 O2 /C~L required for the fourelectron reduction of di~ygen by four Cu1 ions (as was observed for 5) . .,..._ The formation of a ketone on one side of .,..._2 is also consistent with the stoichiometry, 4 , 6 but this is not possible for the methine carbons in 22 . .,...,.,..,. Other explanations for these results are certainly possible, and further work would be necessary to advance any explanation beyond mere speculation. The dioxygen reactions were not investigated further, since ligand oxidation was not desired in the present study, nor was the four-electron reduction of dioxygen by two binuclear cuprous compounds of particular interest. However, these results 117 suggested the avoidance of "activated" positions in the ligand, and this factor was taken into account in the design of additional copper(!) comp~xes (vide infra). The large adamantyl groups on the "sidearms" of 23 were intro"" duced in an effort to inhibit a bimolecular reaction between ,...,..,. 23 and a dioxygen adduct of ,...,..,. 23, which could ultimately lead to the (undesired) four-electron reduction of dioxygen. In this manner, the format ion of a stable diaxygen adduct was envisioned. This rationale proved unsuccessful, however, as shown in equation 1. The isolated product ) 23 + 24 other products (1) 118 indicates major ligand dissociation during the dioxygen reaction. - Although other (uncharacterized) products were present, 25 appeared - to .be a major product. The identity of 25 was confirmed by direct synthesis from copper(II) starting materials. These results are discouraging, and they suggest that the pyrazole may have dissociated during many of these diaxygen reactions. The preliminary results on the dioxygen reactivity of the -- binuclear cuprous systems, 2-23 (Table 1) indicate that the processes were not simple binding of dioxygen to copper. Possible problems appear to be activated ligand positions and dissociable ligands. These - - factors were taken into account in the design of 26 and 27, which were 26 27 desired for investfation of their dioxygen reactivity. The three- - - coordinate environments of nitrogen and oxygen ligands in 26 and 27 were predicted to be stable configurations for Cu1 on the basis of the results presented in Chapter I on three-coordinate copper(!). - 8omplex 27 was of particular interest since it closely approximates :he proposed ligands in hemocyanin. 7 119 .._ The synthesis of 26 was unsuccessful, leading instead to a ""' product with a Cu:ligand ratio of 3:2 ~ - , ~' equation 2). The A 0 0 H +. 0 (2) (CH 3 CH 2 ) 3 N O--NH1 28 actual structure of _... 28 is not known, but the structure proposed in equation 2 is in agreement with the elemental analysis and allows for four ligands around each copper(!). A mononuclear Cu1 complex, 29, ""' was isolated in the absence of base (equation 3). Here again, the A 0 0 H 0 ~ ·BF4 N,Ju~BN / 'N 5 '/ 29 ~ - (3) 120 - actual structure of 29 is unknown, but the ligand can provide four nitrogen ligands for one copper(!). These results suggest that the - desired three-coordinate environment of these ligands in 26 may not be most favorable for Cu1, and hence alternative structures such as - - 28 and 29 are formed. - It was hoped, however, that the use of imidazole ligands in 27 (rather than the pyridine ligands in 26) would stabilize Cu1 in this three- - coordinate geometry. Synthesis of 27 required the prior synthesis of 2-(2'aminoethyl)-1-methylimidazole dihydrochloride, 30. The litera""' 8 ture synthesis of this compound proved unsuccessful, , 9 but the procedure shown in Scheme 1 yielded the desired product in fair yield. The primary amine, .,....,... 30, condensed readily with 2-hydroxy-5methyl-isophthalaldehyde and yielded the binuclear copper(II) complex, - - 31 (equation 4). However, the desired copper(!) complex (27), proved A O O 0 H (4) 31 30 elusive, despite numerous attempts (equation 5). A variety of solvents were employed including acetonitrile, methanol, ethanol and acetone. 121 Scheme 1 ~ N [~ N CN picric acid HCl \ N -2HCI [~~NH1 30 122 A O O 0 H + (5) 27 30 Triethylamine, sodium methoxide and tetrabutylammonium hydroxide were tried as bases, and the reactions were run at both reflux and ambient temperatures. In all ~ases, impure products (based on I. R. and elemental analysis) were isolated, with most reactions leading to some disproportionation of Cu1 to cu 0 and Curr. This was observed by a copper mirror on the reaction flask, and the prcton magnetic resonance spectra exhibited only very broad peaks. (Pure copper(!) complexes should be diamagnetic.) Similar results were obtained using 1-benzyl-2-(2'-aminoethyl)imidazole, 32, which was synthesized in a manner analogous to 30. 9 -- -- -- This amine (32) was investigated because of possible solubility enhancement from the benzyl groups. Again, copper(II) species were straightforwardly synthesized (equations 6 and 7), but attempts to synthesize analogous copper(!) compounds led to disproportionation and impure products. -- - -- The synthesis of cuprous analogues of 31, 33 and 34 could, in 123 A 0 0 H 0 + Cu (H,. 0)6 ( B~)2 CH 3 0H ) 32 (7) theory, be accomplished by electrochemical reduction of these Curr compounds. Unfortunately, only very broad, irreversible reduction waves were observed in the cyclic voltammograms of these compounds (at pt or Hg electrodes). Hence, clean electrochemical reduction of these compounds does not seem tenable. Addition of excess pyrazole to a solution of .,..,_ 31 in DMF, gave the quasi-reversible electrochemical behavior shown in Figure 2. This 124 -0.9 -0.7 -0.5 -0.3 -0.1 0.1 VOLTS (nhe) 2 ~ - Cyclic voltammogram of CuI\ISOIM(EtMimMOH) + (31, containing excess pyrazole) in DMF solution with 0. 1 M """' TBAP as the electrolyte. A platinum indicating electrode was used, and the scan rate was 200 mV/sec. 125 cyclic voltammogram is similar to those observed for the threecoordinate cuprous compounds presented in Table 1 (see Chapter II). This suggested the synthesis of the binuclear cu1 complex, 35 -- -- (equation 8). By analogy to compounds 2-7, each Cu1 in complex 35 -- (8) 30 is assumed to be three-coordinate in solution (with the imidazole nitrogens not bound to the copper centers). An intermolecular Cu-Cu interaction is likely in the solid state, and indeed, a visible absorption -- at ca. 580 nm is observed for 35 in the solid state (again, see Chapter II). -- The cyclic voltammogram of 35 is essentially identical to that shown in Figure 2. The reduction potentials are E! = -0. 34 V {for 2 CuIIcuIIL + ;= Cullcu1L1 and E; = -0. 53 V (for Cullcu1L+ ~ Cu1Cu1L). These potentials comiru-e to E; = -0.11 V and E; = -0. 34 V for the - analogous compound with pyridines on the sidearms (5). The more negative potentials of ~ indicate a relative stabilization of cull vs. Cu1, and this is probably due to a stranger binding of the imidazole nitrogens to the oxidized cupric ions (compared to the pyridine nitrogens). 126 This is in agreement with the stronger a -donating ability of the imidazole nitrogens. Presumably, the imidazole nitrogens bind to the cupric ions upon oxidation and dissociate from the cuprous ions upon - reduction (equation 9). {9) The results presented herein would seem to suggest that com- - plexes 2-23 and .,...,,.._ 35 provide a stable three-coordinate geometry for Cu1 .,...,,.._ . while the proposed complexes .,...,,.._ 26 and .,...,,.._ 27 do not. Of course, this is not a valid assumption, since the failure to produce .,...,,.._ 26 and .,...,,.._ 27 may be a . result of the synthesis methods employed. For example, unstable intermediates may have led to disproportionation and impure products. Also, a mixture of unwanted products may have been preferentially insoluble. However, the paucity of reported three-coordinate cuprous complexes with nitrogen ligands 10 suggests that this is not the most favorable environment for copper(!). Recall that the crystal structure found for cu!ISOIM(Etpy)(pz), §_, indicated some degree of two coordinate character due to the short Cu-N distances and long Cu-0 distances (see Chapter II). The possible stabilization of ~ and _ll by the addition of a fourth 127 ligand was considered. Attempts were made to introduce a bridging chloride, but no pure compounds were obtained. An effort was also made to synthesize a carbonyl complex by repeating the synthesis shown in equation 5 under an atmosphere of CO. Once again, disproportionation and impure products resulted. Thus, we have not yet been able to stabilize the system by the addition of these potential ligands, but further work in this direction may be desirable. Toward this end, an approach to the synthesis of polydentate ligands which can encompass two copper ions in tetrahedral geometries is presented in Appendix II. 128 Experimental Materials. All chemicals were reagent grade and were used as ~ received unless otherwise noted. Copper(II)tetrafluoroborate, ground to a powder then dried for several days in vacuo (25 ° C), was used as Cu(BF4 ) 2 • 6H 20. Tetrabutylammonium perchlorate, TBAP, (South- western Analytical Chemicals) was dried exhaustively in vacuo (25 ° C) before use. · Reagent grade N,N-dimethylformamide, DMF, was dried over MgS04 and 4A molecular sieves for 24 hours and then vacuum distilled. 2-Hydroxy-5-methylisophthalaldehyde was prepared by a modification of the literature method. ll, 12 Tetra(acetonitrile) copper(!) tetrafluoroborate 13 and bis-(2-pyridyl) methylamine 14 were also prepared by published methods. The preparation of compounds - - - 2-21 (which are numbered 6-25 in Chapter II) is described in Chapter II ,,..._,..._ ,,..._,..._ and ,,..._,..._ 22 and ,,..._,..._ 23 were synthesized analogously. The syntheses of ,all cuprous compounds were performed under helium in a Vacuum Atmospheres Dri-Lab inert atmosphere chamber. The synthesis of - the cupric complex, 24 (given the number 3 in Chapter II), is also ,,..._,..._ presented in Chapter II. Dioxygen Reaction of cu; ISOIM(ada) (pz), ll• A solution of 23 2 (O. 079 g) in 10 ml of DMF was exposes to dioxygen for several hours, yielding green crystalline product. The product, Cun [ISOTh1(ada) 2 ] 2 ; (25), was collected by filtration, washed with acetone and dried under ,,..._,..._ vacuum for several hours. Anal. Calcd. for C 58H74 N4 0 2 Cu: C, 75. 53; H, 8. 03; N, 6. 07. Found: C, 73. 93; H, 7. 95; N, 6. 46. CuII[ISOIM(ada) 2] 2 (~). 1-aminoadamantane (0. 60 g, 4 mmoles) was added to a soluti~n of 2-hydraxy-5-methylisophthalaldehyde 129 (0. 33 g, 2 mmoles) in acetonitrile (20 ml). This mixture was heated at reflux for 15 min., after which, a hot solution of cupric acetate (0. 20 g, 1 mmole) in 10 ml of acetonitrile was added. The resulting mixture (brown-green) was heated at reflux for 15 min. , and the solution volume was reduced to 10 ml by rotary evaporation. The solid product was collected by filtration and washed with acetonitrile and acet.one -. Dissolution of this solid in CH 2 C¼ (20 ml), followed by slow addition of acetonitrile, gave brown-green crystalline product, -- 25, which was collected by filtration, washed with acetone and dried under vacuum for several hours. Anal. Calcd. for C 58H74 N4 0 2 Cu: C, 75. 53; H, 8. 03; N, 6. 07; Cu, 6. 89. Found: C, 73. 73; H, 8. 43; N, 5. 87; Cu, 7. 3. cu; [ISOIM(Etpy) BF (~). 2-(2' -aminoethyl)pyridine (0. 24 ml, ] 2 2 2 mmoles) was added to a mixture of 2-hydroxy-5-methylisophthalaldehyde (0.16 g, 1 mmole) in 10 ml of acetonitrile in an oxygen-free (helium) glove box. The mixture was heated briefly to dissolve all of the dialdehyde. Addition of triethylamine (0.14 ml, 1 mmole) followed by Cu(CH 3 CN) 4 BF 4 (0. 62 g, 2 mmoles) yielded a crystalline orange solid. The mixture was heated to reflux and filtered. The orange product was washed twice with ethanol, then acetone and dried for 3 hrs under vacuum. Anal. Calcd. for C4 JI 46N8 O2 C'll:3BF4 : C, 54.15; H, 4. 51; N, 10. 98; Cu, 18. 69. Found: C, 53. 68; H, 4. 54; N, 11. 46; Cu, 18. 6. 1 Cu [(H)ISOTh1(Etpy) 2 ] BF 4 (W. 2-(2' -aminoethyl)pyridine (0. 24 ml, 2mmoles) was added ., to a solution of 2-hydroxy-5-methylisophthalaldehyde (0.16 g, 1 mmole) in 10 ml of methanol in an oxygen 130 free glove box. This solution was heated to reflux, and a hot solution of Cu(CH 3 CN) 4 BF4 (0. 31 g, 1 mmole) in 10 ml of methanol was added, resulting in an orange solution. Upon cooling to the ambient temperature, orange crystals of 29 formed. The product was collected by --- filtration, washed with ethanol and diethylether and dried under vacuum for several hours. Anal. Calcd. for C23H24N40CuBF4 : C, 52. 84; H, 4. 60; N,- 10. 72; Cu, 12.16. Found: C, 52. 22; H, 4. 64; N, 10. 45; Cu, 12. 4. . . . a, 9 1-meth y1- imidazole (26 g, 0. 34 moles) was mixed with 47 g of 37% formaldehyde in water and sealed in a glass, high-pressure reaction vessel. The solution was heated to 115 ° C for 15 hrs, then allowed to cool to room temperature. The residual pressure was released, and the solvent was removed by rotary evaporation. To the resulting oil was added 100 ml of ethanol and 25 ml of cone. HCl. The solvent was again removed by rotary e¥aporation to give an oil, which crystallized upon cooling with an ice bath. Recrystallization from 2-propanol yielded white crystals of 2-hydroxymethyl-1-methylimidazole (28 g, 51% yield). Anal. Calcd. for C 5H9 N2OCl: C, 40. 42; H, 6. 06; N, 18. 86. Found: C, 40. 50; H, 6.17; N, 18.97. 2-chloromethyl-l-methylimidazole hydrochloride. 9 2-hydroxymethyl-1-methylimidazole (20 g, 0.135 moles) was added over a 10 min. period to 20 ml of thionyl chloride which was cooled by an external ice bath. The solution was heated at reflux for 20 min. Removal of the solvent using a rotary evaporator (cleaned thoroughly afterwards) gave a white, crystall~e solid which was washed with diethyl ether and 131 recrystallized from isopropanol. The product was collected by filtration, washed with diethyl ether and dried under vacuum overnight to yield 19 g (84%} of 2-chloromethyl-1-methylimidazole hydrochloride. Anal. Calcd. for C 5H8N2 C~: C, 35. 93; H, 4. 79; N, 16. 77. Found: C, 36.21; H, 4.68; N, 17.040 • • drochloride. 9 Dry, powdered NaCN (8 g, ·o. 16 moles} was added to a stirred solution of 150 ml of dimethylsulfaxide. To this solution was added 2-chloromethyl-1methylimidazole (13. 4 g, 0. 08 moles) over a 5 min period while cooling with an external ice bath to keep the temperature below 45 ° C. The mixture was stirred and heated to keep the temperature at ca. 40° C for one hour. The dark-brown suspension was then added dropwise, I with stirring, to a hot (85 ° C) mixture of 24. 4 g (Oo 08 moles) of 75% picric acid (25% water) suspended in 300 ml of water. The resulting brown solution was cooled to 5 ° C, yielding a brownish-yellow precipitate which was collected by filtration and washed twice with water. The crude picrate from above was suspended in 80 ml of nitrobenzene, and 20 ml of cone. HCl was added. After mixing in a separatory funnel, the nitrobenzene (lower) layer was drawn off, and the aqueous phase was washed with 2 x 20 ml of chloroform. These washes were combined with the nitrobenzene solution, and this combined organic phase was extracted with 20 ml of cone. HCl. The combined aqueous extracts were washed with 20 ml of CHCls. The solvent was then removed from the aqueous solution by rotary evaporation. The resulting brown solid was recrystallized twice from ethanol to yield slightly brown crystals of 2-cyanomethyl-1-methylimidazole hydro- 132 chloride (5g, 40% overall yield). Anal. Calcd. for CJI 8 N3 Cl: C, 45. 71; H, 5.08; N, 26.67. Found: C, 45.64; H, 5.26; N, 26.80. 2-(2'. -aminoethyl)-1-methylimidazole dihydrochloride (1Q). 15 A mixture of 2-cyanomethyl-1-methylimidazole hydrochloride (4 g, 0. 025 moles), platinum oxide (0. 25 g), cone HCl (10 ml) and methanol (250 ml) was subjected to 3 atm of hydrogen pressure for 12 hrs. The catalyst was removed by filtration, and the solvent was removed by rotary evaporation. Ethanol (200 ml) was added, and the solvent was again removed by rctary evaporation. The resulting oil was immediately transferred into the glove box (under He) due to the hygroscopic nature of the product. The oil soon crystallized and was recrystallized I from ethanol to yield slightly brown crystals of 2-(2' -aminoethyl)-1methylimidazole (3. 2 g, 64% yield). Anal. Calcd. for CJI 13 N3 C¼: C, 36. 36; H, 6. 56; N, 21. 21. Found: C, 36. 37; H, 6. 47; N, 20. 84. c~1ISOTh1:(EtMim)2OH(BF~ Tetrabutylammonium hydroxide (2-. 08 g of a 25% methanol solution, 2 mmoles) was added to a suspension of 2-(2' -aminoethyl)-1-methylimidazole dihydrochloride (0. 2 g, 1 mmole) in 25 ml of ethanol. To the resulting solution was added 2-hydroxy-5-methylisophthalaldehyde (0. 08 g, 0. 5 mmoles) to give an orange-yellow solution. This solution was heated at reflux for 15 min. then allowed to cool to the ambient temperature. Tetrabutylammonium hydroxide (0. 52 g of a 25% methanol solution, 0. 5 mmole) was added with stirring, followed by the addition of Cu(H2 O)a(BF 4) 2 • The mixture was heated at reflux for 30 min. and filtered while hot. Allowing the filtrate to cool slowly to 133 0° C gave microcrystalline blue-green solid (31), ,..._,..._ which was collected by filtration, washed with ethanol and dried under vacuum overnight. Anal. Calcd. for C23H32N6O2 Cl½B2 F 8 : C, 38.09; H, 4.42; N, 11.59. Found: C, 37.70; H, 4.49; N, 11.89. 9 The literature method 1-benzyl-2-(2'aminoethyl)imidazole (32). _,.._ was employed for the synthesis of ,..._,..._ 32, but the dihydrochloride of ,..._,..._ 32 could not be crystallized, and therefore, the free base was isolated as an oil and used directly in further syntheses. c~1ISOIM(EtBim) OH(BF ) (ll). 1-benzyl-2-(2' -aminoethyl)2 4 2 ' imidazole (0. 2 g, 1 mmole) was added to a suspension of 2-hydroxy-5methylisophthalaldehyde in 15 ml of methanol. The resulting orangeyellow solution was heated at reflux for 10 min. , followed by the addition of Cu(H 2 O) 6 (BF4 ) 2 (0. 35 g, 1 mmole), which yielded a bluegreen solution. After heating this solution at reflux for 15 min., a small amount of white solid was removed by filtration, and the solvent was removed from the f_iltrate by rotary evaporation. The resulting green solid was dissolved in 25 ml of acetone, and a light-blue, microcyrstalline solid (33) _,.._ came out of solution. This product was collected by filtration, washed with boiling acetone and dried under vacuum for several hours. Anal. Calcd. for C33 H34N6O2 Cl.½B2 F 8 : C, 46. 78; H, 4. 02, N, 9. 92. Found: C, 46.11; H, 3. 95; N, 9. 81. cu;1ISOIM(EtBim) 2 (Cl) 3 (ll). 1-benzyl-2-(2' -aminoethyl)- imidazole (0. 2 g, 1 mmole) was added to a suspension of 2-hydroxy5-methylisophthalaldehyde in 15 ml of methanol. The resulting orangeyellow solution was heated at reflux for 10 min., followed by the addition of CuC¼ • 2H 2 O (0.17 g, 1 mmole) which yielded a green ' 134 solution. After heating this solution at reflux for 15 min. , the solvent was removed by rotary evaporation. Recrystallization of the resulting -- solid from ethanol gave an olive-green microcrystalline product, 34, which was collected by filtration, washed with ethanol and acetone, and dried overnight under vacuum. Anal. Calcd. for C33 H33 N6OCl¼C1s: C, 51.93; H, 4.33; N, 11.02. Found: C, 51.60; H, 4.23; N, 11.01. 1 C ISOTh1(EtMim) 2 (pz) (35). 2-(2'-aminoethyl)-1-methylimidazole clihyd.rochloride (0. 2 g, 1 mmole) was added to a solution of 2-hydraxy5-methylisophthalaldehyde in 20 ml of methanol (in an oxygen-free glove box). To this solution was added sodium methoxide (0. 11 g, 2 mmoles), which resulted in a yellow-orange solution. This solution was heated at reflux for 15 min. , then allowed to come to ambient temperature over the next 15 min. Pyrazole (0 ~034 g, 0. 5 mmole) and sodium methoxide (0. 055 g, 1 mmole) were added and the solution was stirred for 5 min. Addition of a solution of Cu(CH 3 CN)4 BF4 in 15 ml of hot methanol yielded _an immediate orange-brown solid. The mixture was cooled to ambient temperature, and the solid was collected by filtration. The orange-brown solid was dissolved in hot toluene and a white precipitate was removed by filtration. Removal of the solvent from the filtrate by evaporation under vacuum yielded solid product, which was recrystallized from methanol to give orange-brown needles of ~. This product was collected by filtration, washed with ethanol and dried under vacuum for several hours. Anal. Calcd. for C24H28 N8 OCl¼:C, 50.44;H, 4o90;N, 19.61. Found:C, 49.01; H, 4.85; N, 19.43. 135 en U take Measurements. An accurately weighed sample (ca. 0.1 mmole) was placed in a 10 ml round-bottom flask along with a magnetic stir bar. This flask was connected to an 8 ml calibrated volume apparatus on the vacuum line, and the system was evacuated. Dry methylene chloride (stored over CaH 2 ) was vacuum transferred onto the sample to give a solution volume of ca. 7 ml. The sample chamber was isolated, and the calibrated volume was filled with a measured pressure of dioxygen (ca. 600 Torr) and then closed off. The temperature was also recorded. For the low temperature reactions, a dry ice-acetone bath was then placed around the reaction flask to keep the temperature at -79 ° C. The calibrated volume of 0 2 was then exposed to the sample, and the solution was stirred for 6 hrs. The remaining dioxygen was then collected with a Toepler pump, with the methylene chloride being caught by three intervening ·liquid nitrogen traps. The gas was collected in a calibrated volume (13. 76 ml), and the pressure and temperature were measured. The amount of dioxygen consumed was then calculated by difference. Ph sical Measurements. Sample preparation for physical studies on the air-sensitive materials was accomplished in a Vacuum Atmospheres Dri-Lab inert atmosphere chamber, under a helium atmosphere. Helium saturated spectroquality solvents were~used for solution studies. Electronic spectra were recorded on a Cary 14 spectrophotometer. Solid state spectra were obtained from Nujol mulls on filter paper against a Nujol saturated filter paper as a blank. Infrared spectra were recorded on a Beckman ffi-12 Infrared spectrophotometer. S'olid 136 state spectra were ottained from Nujol mulls pressed between KBr plates. Proton magnetic resonance spectra were recorded on a Varian EM390 spectrophotometer at 90 MHz (34 ° C). The solvent utilized was CD3 CN containing TMS as the reference. Elemental analyses were performed by the Caltech Microanalytical Laboratory. ~ A Princeton Applied Research (PAR) Model 174A polarographic analyzer was used for cyclic voltammetry and differential pulse voltammetry. For display purposes, a HewlettPackard 7004B X-Y recorder was utilized. Cyclic voltammetry was done in a single compartment cell with a volume of ca. 5 ml. of DMF. The supporting electrolyte was 0.1 M TBAP. The working electrode consisted of a platinum button electrode or a hanging mercury drop electrode. In all cases, the auxiliary electrode was a coiled platinum wire., and the reference electrode consisted of a silver wire immersed in an acetoni trile solution containing AgNO3 (0. 01 M) and TBAP (0.1 M), all contained in a 9 mm glass tube fitted on the bottom with a fine porosity sintered glass frit. All potentials are reported versus the normal hydrogen electrode, nhe. This was accomplished by the use of an internal reference red ax couple, namely ferrocene, for which the formal potential is reported to be +0. 400 V versus nhe in water. 16 It has been proposed that the ferrocene reduction potential changes very little in different solvents, and hence it is a good solvent independent redox couple. 16 , 17 3 Experimentally, small amounb:(cao 10- M) of ferrocene were added to solutions containing the compounds of interest and formal potentials for both couples were measured under the same conditions. 137 References and Notes (1) Lontie, R. in "Inorganic Biochemistry", Eichhorn, G. I. ed.; Elsevier: New York, N. Y., 1973; p. 344. (2) (a) Mason, H. S. Anne Rev, Biochem. ~ ' 35, 595-634. (b) Vanneste, W. H. and Zuberbuhler, A. in "Molecular Mechanisms of Oxygen Activation", Hayaishi, O. ed.; Academic Press: 1974; p. 371. (3) (a) Fee, J. A. Structure and Bonding~' 23, 1-60. (b) Vanngard, T. I. in "Biological Applications of Electron Spin Resonance Spectroscopy", Swartz, H. M. ; Bolton, J. R. ; and Borg, D. C., eds.: J. Wiley: New York, N. Y., 1972; p. 411. (4) Metal-catalyzed autoxidations of similar activated positions have been reported. See: Sprecher, C. A.; Zuberbuhler, A. D. Angew. Chem. Int~ Ed. Engl. !m, !.2_, 189. Urbach, F. L.; Knopp, U.; Zuberblihler, A. D. Helv. Chim. Acta ,!1.ll, &!_, 1097-1106. (5) There appears to be little, if any, precedent for this alkyl carbon hydroxylation. For a discussion of metal-catalyzed autoxidation of organic substrates, see: Sheldon, R. A.; Kochi, J. K. Adv. Cat. 1976, -25, 272-413. ~ (6) For a discussion of related ligand "ketonations" see: Lisensky, G. C. , Ph.D. Dissertation, California Institute of Technology, 1981, and references cited therein. (7) Eickman, N. C. ; Himmelwright, R. S. ; Solomon, E. I. Proc. Natl. Acad. Sci. USA, ~ ' 76, 2094-2098. Brown, 138 References (continued) J. M.; Powers, L.; Kincaid, B.; Larrabee, J. A.; Spiro, T. G. J. Amer. Chem. Soc. !fil!Q, 102, 4210-4216. Larrabee, J. A.; Spiro, T. G. J. Amer. Chem. Soc. !fil!Q, 102, 4217-4223. (8) Jocelyn, P. C. J. Chem. Soc. ,,..,.,__.._ 1957, 3305-3307. (9) Kornfeld, E. C.; Wolf, L.; Lin, T. M.; Slater, I. H. J. Med. Chem. ~ ' 11, 1028-1031. (10) Ellen, P.; Braldey, D.; Hursthouse, M.; Meek, D. Coord. Chem. Rev. JJU.1, 24, 1-95. Jardine, F. Adv. Inorg. Chem. Radiochem. ~ ' 17, 115-163. (11) Ullmann, F.; Brittner, K. Chem. Ber. ~ ' 42, 2539-2548. (12) Spiro, C. L.; Ph.D. Dissertation, California Institute of Techn.ology, 1981, p. 86. (13) Hathaway, B. I.; Holah, D. G.; Postlethwaite, J. D. J. Chem. Soc. ~ ' 3215-3218. (14) Niemers, E.; Hiltmann, R. Synthesis~' 593-595. (15) Secrist, J. A.; Logue, M. W. J. Org. Chem. 1.m, 37, 335-336. (16) Koepp, H. M.; Wendt, H.; strehlow, H. Z. Electrochem. 1960, ~ 64, 483-491. (17) Gagne, R.R.; Koval, C. A.; Lisensky, G. C. Inorg. Chem. 1980, 19, in press. ~ 139 CHAPTER V Summary and Conclusions This thesis has presented the design and synthesis of a number of binuclear copper(!) complexes as potential models for binuclear - copper protein sites. Indeed, system 1 may represent the best synthetic "models" available for the reduced protein site. The most R=J),~ ~' , etc. 1 X= , \N-NI , I() I / fNIIN\ ~ recent physical studies on hemocyanin point to two or three coordinate copper(!) ions bound to imidazole ligands. 1 Also, a bridging ligand such as a tyrosine oxygen appears to be present in oxyhemocyanin. l, 2 - Thus, the series of complexes, 1, presented in Chapter II, closely • 140 approximates the proposed ligand environment for hemocyanin. The positive reduction potentials exhibited by the compounds (ca. 0 V vs. nhe) approach the value recorded for tyrosine (+O. 36 V). This system fails, however, to mimic· the dioxygen and carbon monoxide reactivity of these proteins. Reversible oxygenation was not observed, and this may have been due to the presence of dissociable ligands and activated positions (toward oxidation) as well as the lack of steric hindrance to prevent the four-electron reduction of oxygen to water. While hemocyanin appears to bind one CO per binuclear site, 3 synthetic com- - plexes (1) are unreactive toward CO. . The three-coordinate arrangement of nitrogen and oxygen ligands in 1 appears to provide a very stable environment for copper(!) (displayed in the rather positive reduction potentials). This ligand environment also fosters an intermolecular Cu-Cu interaction. The - crystal structure of 1 where R = 2-(2' -pyridyl) ethyl suggested a much stronger interaction bet~een the nitrogen ligands and the copper(!) centers compared to the interaction between the oxygen and the coppers. Hence, the stability and unusual reactivity of these compounds may result from the nearly linear arrangement of nitrogen ligands. The apparent preference of copper(!) for a two-coordinate - environment of nitrogen ligands is evident in the crystal structure of 2. \ Again, a direct Cu-Cu interaction was observed. This complex con- - - trasts the earlier system (1), and 2 reacts with carbon monoxide - to yield the four-coordinate copper(!) centers in 3. 141 OC-Cu- c? 2 3 - Based on the chemistry observed for system 1 as well as recent protein studies, l, 2 complexes 4 and 5 were designed as possible - - 4 5 improved models for the binuclear copper active sites. Although the I - - synthesis of the ligands for 4 and 5 was accomplished, the isolation of the cuprous complexes was not realized. This may reflect a thermo- - - dynamic instability of the desired products (4 and 5) or the difficulties involved in the synthesis of such copper(!) species. Future research in this area should involve the synthesis of new 142 copper(!) complexes for the investigation of their dioxygen reactivity, taking into account the results presented above. Binuclear copper(!) - - systems such as 6 and 7 are proposed by this author. These ligands 2+ 6 7 could provide a stable, two-coordinate environment for the Cu1 ions, and the ligand structure should be somewhat rigid due to the phenylene bridge and steric repulsions between the heterocyclic rings. This latter factor may prevent major ligand reorganization upon dissolution or reaction 'with 0 2 or CO. Finally, the ligand can provide steric bulk around a dioxygen molecule bound in the "pocket" between the two coppers. 143 References ~ (1) Brown, J. M.; Powers, L.; Kincaid, B.; Larrabee., J. A.; Spiro, T. G. J. Amer. Chem. Soc. ~ ' 102, 4210-4216. (2) Eickman, N.; Himmelwright, R. S.; Solomon, E. I. Natl. Acad. Sci. USA~' 76, 2094-2098. Proc. Larrabee, J. A.; Spiro, T. G. J. Amer. Chem. Soc.,~' 102, 4217-4223. (3) Alben, J. O.; Yen, L.; Farrier, N. J. J. Amer. Chem. Soc. filQ, 92, 4475-4476. Fager, L. Y.; Alben, J. O. Biochemistry 1972, -11, 4786-4792. ~ 144 APPENDIX I Investigation of a Reported One-step Two-Electron Reduction - The binuclear copper complex, 1, has been reported by Lintvedt et al. 1 to undergo a one-step, two-electron reduction at E1. = -0. 61 V 2 versus the standard caloµiel electrode (see). 1 approximately -0. 34 V vs. nhe. This corresponds to Cu~ (PAA);i. en This result is quite surpri.s ing since the complex is reported to be completely diamagnetic at room temperature, showing no EPR signal and sharp NMR signals. The electrochemistry observed here is in contrast to that expected for two metals which are interacting to such a large extent (indicated by the diamagnetism). Usually, the reduction of one of the metals results in a more negative environment around the remaining unreduced metal (compared to the environment about the first metal, prior to reduction), and this second metal is reduced at a more negative potential. 2 Two possible explanations for this anomalous behavior were postulated by our laboratory. 3 a) The reduction of the first Cun to Cu1 and its concomitant adoption of tetrahedral geometry could force a more tetrahedral geometry around the remaining Cun, making its reduction 145 easier. b) Due to the asymmetric nature of this complex, access to the electrode may be hindered at one of the coppers, and if the electrons are forced to enter via the cull which is inherently more difficult to reduce, one potential may be observed for the reduction of both metals. A one-step, two-electron reduction of a binuclear cupric complex was of great interest due to its analogy with the electrochemistry exhibited by the binuclear copper protein sites. 4 For this reason, we set out to ascertain if, indeed, the -r _eductioil of 1 was a one-step, two-electron process. Complex 1 was ~ynthesized by published methods (Scheme 1). 1 , 5 - The triketone was synthesized by a modified Claisen type condensation, Scheme 1 ~ + '1rY 0 0 0 0 0 H-.PAA E+OH 1 Ha_N"VNHa., ( 1 \ 146 and this was followed directly by a Schiff base condensation with ethylenediamine. Treatment of the ligand with cupric acetate gave the desired complex, !.· Ct½(PAA) 2 en (H 2 PAA =pivaloylacetylacetone and en =ethy lenediamine). This complex (!.) appears to be only slightly paramagnetic, showing a fairly well resolved nuclear magnetic resonance (NMR) spectrum (see Figure 1). Note that the two downfield peaks due to the methine protons (A) and the methylene protons (B) are quite broad due to the possible influence of paramagnetic copper centers. It was found that this complex does give an electron paramagnetic resonance (EPR) spectrum, although the signal is quite weak (Figure 2). Note the multitude of superhyperfine splitting (vide infra). - Cyclic voltammetry of 1 at a hanging mercury drop electrode is shown in Figure 3. The formal reduction potential was ca. -1. 04 V vs. nhe. The measured potential was converted to vs. nhe using Ef = +O. 40 V vs. nhe f~r ferrocene. 6, 7 Cyclic v~ltammetry with see as reference electrode yielded Ef = -0. 93 V vs. see. However this latter value is rather uncertain due to the unknown junction potential which exists between the aqueous calomel electrode and the dimethylformamide solti: ion. 8 (The uncertainty which arises from the use of this electrode in a non-aqueous solvent such as DMF is evident in the 300 mV difference between the value attained here and that from the previous study. ) 1 The peak separation (Ep -Ep ) at a scan rate, v, of 50 m V per a C second was 80 mV. Unfortunately this peak separation is not directly © ®· ~t 111 'Q I \,k,., ..... ~ -:t Cllel3 ~ TMS (j CJ 8 7 ~ 6 s i 3 2. NMR spectrum of!, in CDC1 3 , (34°C). 0 148 ;1:· :i :I ::,:· •• L · :' • • • • I • , , • • . l '. . : .j • I . . . ... . :: :.. , .. . -: '. _O;': j - •• - _j · • : : • • • • • - • • ·• • I i : :I. :: -- I Figure 2. (microwave frequency = 9. 172 GHz). -J.6 - EPR spectrum of Cl½(PAA) 2 en (1) in C~Cli at 77°K -0.8 -0.6 - Figure 3. Cyclic voltammogram of Cl½(PAA\en (1) in DMF solution containing 0. 1 M TBAP as the electroJyte. A hanging mercury drop electrode was used and the scan rate was 100 mV/sec. 149 indicative of the number of electrons involved in the process (n value). It is also related to other factors such as the scan rate, resistance between the electrodes and of course the rate of electron transfer. The value obtained here is larger than that expected for a reversible one-electron process (58 mV) or a reversible two-electron process ( 29 mV). 9 For this reason, constant pctential electrolysis (cpe) was - performed ·on a DMF solution of 1 to determine the n value. After passing the number of coulombs corresponding to a !-electron reduction, the current dropped to a very slow rate, and copper metal plated out on the glass walls of the electrochemical cell as further electrolysis was done. It appears that the binuclear Curr complex was being reduced by one electron to produce an unstable mixed valence species. This complex may then slowly disproportionate to give copper metal and Cun complexes which can undergo further electrolysis. For instance, consider Scheme 2. Scheme 2 ~ CuIIcuIIL 2CuIIcu1L - +le-> cpe CuIICu1L - > Cu0 + CuIILs- + CuIIcuIIL further reductions (L =(PAA) 2 en) A sample of the electrolyzed solution {after one equivalent of electrons was passed and the current had slowed to a minimum) was removed and analyzed by EPR, and an interesting result was obtained. The spectrum, including super-hyperfine peaks was identical to 150 Figure 2, obtained for the original binuclear Cun complex, but the signal was ca. 200 times stronger (for identical concentrations). Due to. the observed super-hyperfine peaks, the reduction product appears to be the mixed valence species, ~. Very similar EPR spectra 2 were observed by Lintvedt et al. 1 for both the mononuclear cull[H 2 (PAA) 2 en] and the mixed metal CuIIznII(PAA) 2 en. The Curr is believed to be nitrogen coordinated in both of these compounds. The observed EPR signal after reduction may also be due to the mononuclear Cun complex following a disproportionation reaction (see - Scheme 2). Attempts to isolate 2 were unsuccessful due to further disproportionation reactions. The fact that the original binuclear - cull complex, 1, exhibited the same spectrum may be due to mononuclear impurities. The investigation of this system was halted at this point since the complex does not appear to be reduced by a reversible two-electron process, as reported. In addition, the reduction potential (ca. -1. 04 V vs. nhe) is very negative, indicating the destabilization of Cu1 by this ligand (which is in agreement with the observed disproportionation 151 upon reduction of!.). This is not surprising due to the fairly rigid, planar ligand which utilizes relatively "hard" donor atoms. The formal _negative charge on the ligand ( -4) should also resist the addition of electrons to the system. Neither the negative reduction potentials nor the type of ligands about the coppers are conducive to a model for copper proteins. 152 All chemicals were reagent grade and were used as received unless ctherwise noted. Tetrabutylammonium perchlorate (TBAP, Southwestern Analytical Chemicals) was dried exhaustively under ' vacuum (25 ° C) before use. Reagent grade N, N-dimethylformamide (DMF) was dried over MgS04 and 4A molecular sieves for 24 hrs and then vacuum distilled. Cl.½(PAA) 2 en, 1. The ligand, HiPAA) 2 en, was synthesized by the literature methods 1 ' 5 without isolation of the intermediate, H PAA. ~ 2 Cupric acetate monohydrate (0. 40 g, 2 mmoles) was dissolved in 50 ml of hot methanol. To this solution was added HiPAA) 2 en (0. 37 g, 1 mmole), resulting in a green solution. After cooling to 5 ° C, green - crystals of 1 were collected, and the product was recrystallized from chloroform and washed with methanol. After drying under vacuum overnight, the yield was 0. 45 g (87%) of yellow-green needles. These needles decompose at ca. 244 ° C. Anal. Calcd. for C22H32N20 4 Cl.½: C, 51.25; H, 6.26, N, 5.43; Cu, 24.65. Found: C, 50.88; H, 6.07; N, 5. 34; Cu, 24. 3. The NMR spectrum was recorded on a Varian EM 390 spectrophotometer at 90 mHz (34 ° C). The solvent utilized was CDC:Is containing Me 4 Si as the reference. The X-oond EPR spectra were recorded on a Varian E-line spectrophotometer. - The EPR spectrum of the reduced species, 5, was recorded under an atmosphere of helium. Elemental Analysis was performed by the Caltech Microanalytical Laboratory. 153 ~ All electrochemical measurements were performed under helium in a Vacuum Atmospheres Dri-Lab inert atmosphere chamber. A Princeton Applied Research (PAR) Model 174A polarographic analyzer was used for cyclic voltammetry. For display purposes, a Hewlett-Packard 7004B X-Y recorder was utilized. The apparatus used for constant potential electrolysis (cpe) consisted of a PAR Model 173 pctentiostat-galvanostat coupled with a Model 179 digital coulometer. Cyclic voltammetry was done in a single compartment cell with a volume of ca. 5 ml of DMF. The supporting electrolyte was 0.1 M TBAP. The working electrode consisted of a platinum button electrode or a hanging mercury drop electrode. In all cases, the auxiliary electrode was a coiled platinum wire and the reference electrode consisted of a silver wire immersed in an acetonitrile solution containing AgNO3 (0. 01 M) and TBAP (0.1 M), all contained in a 9 mm glass tube fitted on the bottom with a fine porosity sintered glass frit. The apparatus employed for constant potential electrolysis was similar, except a two-compartment H-cell was used to isolate the auxiliary electrode from the working compartment. A mercury pool was used as the working electrode for cpe. All potentials are reported versus the normal hydrogen electrode, nhe. This was accomplished by the use of an internal reference redox couple, namely ferrocene, for which the formal potential is reported to be +0. 400 V versus nhe in water. 6 It has been proposed that the ferrocene reduction potential changes very little in different solvents, and hence it is a good solvent-independent redox couple. 7 154 Experimentally, small amounts (ca. 10-3 M) of ferrocene were added to solutions containing the compounds of interest,and formal potentials for both couples were measured under the same conditions. 155 References ~ (1) Lintvedt, R. L.; Tomlonovic, B.; Fenton, D. E.; Glick, M. D. Adv. Chem. Series 1975, 150, 407-425. ' """"""' (2) For example see: Creutz, C.; Taube, H. J. Amer. Chem. Soc. ~ ' 95, 1086-1094. Gagne, R. R.; Koval, C. A.; Smith, T. J.; Cimolino, M. C. J. Amer. Chem. Soc. ill!!., 101, 4571-4580. (3) Gagne, R. R. , personal communication. (4) See the discussion in Chapter I and references cited therein. (5) Miles, M. L.; Harris, T. M.; Hauser, C. R. J. Org. Chem. lli.Q, 30, 1007-1011. (6) Koepp, H. M.; Wendt, H.; Strehlow, H. Z. Electrochem. 1960, ~ 64, 483-491. (7) For a discussion of the use of ferrocene as an internal standard for electrochemistry see: Gagne, R. R. ; Koval, C. A. ; Lisensky, G. C. Inorg. Chem. lliQ_, 19, in press. (8) Badoz-Lamblingand, J.; Bardin, J. C. Compt. Rend. Acad. Sci. 1973, 276C, - 583-586. A1fenaar, M.; de Ligny, C.; Remynse, A. ~ Rec. Trav. Chim. Pays-Bas ~ ' ~ 986-992. (9) Anson, F. C. "Electroanalytical Chemistry'~ A. C. S. Audio Course, 1976, Section G. 156 APPENDIX II An Approach to the Synthesis of Binuclear, Tetrahedral Copper(I) Complexes The quest for reversible oxygenation in synthetic copper(!) complexes has met with little success. 1 This is disappointing, since the protein hemocyanin employs copper in its function to bind diaxygen reversibly. 2 Careful design of the ligand system may be necessary for the realization of this behavior in model compounds. Complexes such - as 1 are proposed by this author as viable candidates for reversible • + / \ u I "N I - N --N N__. Cu,,, I \d ~ N-- b 1oxygenation. This complex includes a number of factors, based on protein and synthetic copper(!) investigations, which may be important for obtaining clean, reversible oxygenation. (a) A binuclear complex is suggested by the fact that hemocyanin appears to contain a binuclear copper active site. 2 (b) A tetrahedral ligand environment has been found to be favorable for copper(!) 3 leading to high reduction potentials, 4 157 which may be desirable for reversible dioxygen binding. 5 (c) The four imidazole ligands may provide steric bulk around a dioxygen molecule bound between the two coppers. This may be necessary to prevent further reactions of bound dioxygen. 6 (cl) The ligands provided - by the polydentate chelate in 1 are almost identical to those proposed for hemocyanin. 7 (e) The system avoids dissociable ligands and obvious activated positions (which might be subject to autoxidation). 8 The synthesis of the binuclear copper complex, !., requires the prior synthesis of the primary amine, a, a-bis- [2-(1-methylimidazolyl)- - methyl] benzylamine, 2, which could then be condensed with 2-hydroxy- - 5-methylisophthalaldehyde to generate the polydentate ligand in 1. An attempt to synthesize 2 is presented in Scheme 1. Unfortunately, this approach failed at the final step. Apparently, the large amount of steric hindrance adjacent to the amide prevents its degradation to the - desired amine, 2. This result was surprising since the conversion of amides to amines has been affected at hindered and heterocyclic amides. 9 The reaction was also attempted using methanol as the solvent, but again starting material was recovered. Although this attempt did not yield the desired amine, ~' slight alterations in the reaction route or in the target amine may yield a bis-imidazolyl primary amine. This method may then afford interesting binuclear copper(!) complexes for investigation of their diaxygen reactivity. 158 Scheme 1 -~-~ flN"'- l)CH 1 0 > [N~CI LN 1 HO I I\ ~ oH \ :i>Hc, <t)CH 2 CN SOCl 2 r.-Nh·HCJ NaH >~ -DM_F_~> ~ Br,_ NaOH 2 ' 159 All chemicals were reagent grade and were used as received - unless otherwise noted. N ,_N-dimethylformamide (DMF) was dried over MgS0 4 and 4A molecular sieves for 24 h and then vacuum distilled. The preparation of 2-hydroxymethyl-1-methylimidazole hydrochloride and 2- chloromethyl-1-methylimidazole hydrochloride is described in Chapter IV. Proton magnetic resonance spectra were recorded on a Varian EM390 spectrophotometer at 90 MHz (34 ° C, CDClg). Elemental analyses were performed by the Caltech Microanalytical Laboratory. . . 10 Sodium hydride (36g, O. 75 moles) was added to 500 ml of dry (MgS04) DMF under an atmosphere of nitrogen. Phenylacetonitrile (17. 2 ml, 0.15 moles) was added dropwise from a dropping funnel while stirring the mixture. When the evolution of H2 became slow, 2-chloromethyl1-methylimidazole hydrochloride (50 g, O. 3 moles) was added in small portions over 90 min. , keeping the temperature at 40-50 ° C by the heat from the exothermic reaction. The mixture was then heated at 80-90 ° C for 2 hrs and cooled to ambient temperature. Water (5 ml) was added to quench the excess NaH. The solvent was removed by rotary evaporation and 300 ml of water was added. The solution was rendered acidic with cone. HCI and extracted with 2 x 50 ml of chloroform. The chloroform extracts were discarded, and the aqueous phase was rendered alkaline by the addition of aqueous NaOH. The aqueous phase was then extracted with 4 x 50 ml of chloroform. The combined chloroform extracts were dried (MgSOJ, filtered,and the solvent was 160 removed by rotary evaporation. Addition of diethyl ether (300 ml) with stirring yielded off-white solid which was collected by filtration and washed with diethyl ether. Recrystallization from toluene gave offwhite crystals of a, a-bis- [2-(1-methylimidazolyl)methyl] benzylcyanide (28 g, 61 % yield). The product was washed with diethyl ether and dried overnight under vacuum. Anal. Calcd. for C18H19 N5 : C, 70. 82; H, 6.23; N, 22.95. Found: C, 70.71; H, 6.27, N. 22.53. . . . 11 a-a-bis-[2-(1-methylimidazolyl)methyll benzylcyanide (12 g, O. 04 moles) was mixed with 60 ml of 80% H2 S04 (aq) under N2 at ambient temperature. Most of the solid dissolved and the mixture was stirred and heated at 100° C for 90 min. The reaction mixture was then cooled in an ice bath and added to a stirred solution of 450 ml of 4N NaOH which was also cooled in an ice bath. The aqueous mixture was extracted with 4 x 50 ml of CHCis. The combined chloroform extracts were dried (MgS04), and evaporation of the solvent resulted in an oil which crystallized upon addition of a solution of 10 ml CH 2 C~ and 20 ml of diethylether. After cooling to O° C, the white crystals of a, a-bis-[2-(1-methyl-imidazolyl)methyl] phenylacetamide (7. 8 g, 61 %) were collected by filtration, washed with diethyl ether and dried under vacuum overnight. Anal. Calcd. for C18H21 N50: C, 66. 87; H, 6. 50; N, 21. 67. Found: C, 66. 88; H, 6. 42; N, 21. 99. Attempted Synthesis of a, a-bis-[2-(1-methyl-imidazoyl)methyl]benzylamine, 2. 9 Sodium hydroxide (0. 095 g, 2. 4 mmoles) was ~ dissolved in 2 ml of water. The solution was cooled to 0° C and 161 bromine (0.12 g, 1. 5 mmoles) was added followed by a, a-bis-[2-(1methylimidazolyl)methyl] phenylacetamide (0. 2 g, 0. 6 mmoles). The mixture was stirred at O° C for 30 min. , then at 80 ° C for one hour. The mixture was cooled, acidified (cone. HCl) and washed with CHC!s (5 ml). The aqueous layer was then made basic (NaOH) and extracted with 2 x 5 ml of CHCis. The combined chloroform extracts were dried (MgS04) and the solvent was removed by rotary evaporation. NMR of the resulting oil indicated mainly starting material. starting material was also recovered in an analogous procedure using methanol as the solvent. 162 References ~ (1) See: Simmons, M. G.; Wilson, L. J. J. Chem. Soc., Chem. Commun. 1978, 634-636, and references cited therein. ~ (2) Lontie, R. in "Inorganic Biochemistry", Eichhorn, G. I., Ed.; American Elsevier: New York, 1973, p. 344. (3) Jardine, F. H. Adv. Inorg. Chem. Radiochem. fil.§_, 17, 115-163. (4) Patterson, G. S.; Holm, R. H. Bioinorg. Chem. ~ ' .!, 257-275. (5) Reversibility of oxygenation in cobalt complexes was shown to be related to reduction potentials: Carter, M. J.; Rillema, D. P.; Basolo, F. J. Amer. Chem. Soc. 1974, 96, 392-400. ---- ( 6) This strategy was successful in preventing the irreversible decomposition of dioxygen adducts with iron porphyrins. See: Collman, J. P.; Gagne, R. R., Reed, C. A.; Halbert, T. R.; --- - Lang, G.; Robinson, W. T., J. Amer. Chem. Soc. 1975, 97, 1427-1439. (7) See: Brown, J.M.; Powers, L.; Kincaid, B.; Larrabee, J. A.; Spiro, T. G. J. Amer. Chem. Soc. ~ ' 102, 4210-4216, and references cited thereino (8) See Chapter IV of this thesis. (9) Wallis, E. S.; Lane, J. F. Org. React. ~ ' ~ 267-306. (10) Zugg, H. E.; Dunnigan, D. A.; Michaels, R. J.; Swett, L. R.; Wang, T. S.; Sommers, A.H.; DeNet, R. W. Jo Org. Chem. ~ ' 26, 644-651. (11) Sperber, N.; Papa, D.; Schwenk, E. J. Amer. Chem. Soc. 1948, -70, 3091-3097. ~ 163 PROPOSITIONS Abstracts ~ Proposition 1. The synthesis of transition metal complexes derived from "face-to-face" 1, 10-phenanthroline macrocycles is proposed. The compounds will be investigated as catalysts for the reduction of oxygen to water .. Proposition 2. The synthesis of gold metallocycles and an investigation of their stability and reactivity patterns are proposed. Proposition 3. The synthesis of models for blue copper prcteins is proposed. Proposition 4. The use of macrocyclic copper(!) complexes for the photosensitization of norbornadiene to quadricyclene is proposed. Proposition 5. An investigation into the mechanism of the reaction between transition metal formyl complexes and carbon electrophiles is proposed. 164 _PROPOSITION 1 Metal Complexes Derived from Face-to-Face 1, 10-Phenanthroline Macrocycles The development of fuel cell technology would represent a significant advance towards more efficient, less polluting energy. 1 A fuel cell has the potential of efficiently and cleanly converting the energy stored in fuels into useful electrical energy by the electrochemical oxidation of these fuels (e.g., H2 , hydrocarbons) with oxygen. For the fuel cell to become an efficient device for energy conversion, a catalyst must be found for the reduction of oxygen to water (equation 1). This four-electron process is (1) likely to require a catalyst which can deliver four electrons to the oxygen molecule and avoid high-energy peroxide and superoxide intermediates. 2 Polynuclear transition metal complexes are good candidates for such multi-electron transfer. 3 In this light it is interesting to note that the reduction of oxygen to water is catalyzed by the natural enzyme, lac case, which contains four copper atoms per molecule, all of which appear to be necessary for enzymatic activity o 4 This rationale has led to the synthesis of novel binucleating ligands capable of holding two metal ions in close proximity, such that these ions can react with dioxygen and possibly deliver two electrons per metal to the bound dioxygen moiety. 2, 5 ' 6 The following "face-toface" cobalt porphyrin has been synthesized by Collman et al., and 165 indeed it was found to catalyze the reduction of oxygen to water at a graphite disk electrode. 6 Electrons were supplied by the electrode 0~ \ N-H • ~o I H-N 1 ,..,, at a potential of ca. O. 6 volts vs. nhe, and molecular oxygen was reduced catalytically to .water. This was performed at a rotating ring-disk electrode, 7 and the production of (unwanted) hydrogen peroxide was monitored at the outer ring. By this technique, it was determined that very little H2 0 2 was produced, and that oxygen was reduced to water at a turnover rate of 30/site/sec! The exact mechanism of this reaction is not known with certainty. These results are very encouraging, but certain problems need to be overcome. The reduction of oxygen does not commence until the electrode potential is more negative than +O. 7 volts vs. nhe, and this potential is well below the standard reduction potential for the four-electron reduction of oxygen to water (1. 23 volts vs. nhe). For optimum voltage output from a fuel cell, the catalyst should 166 operate at a potential as close to 1. 23 volts as possible, while still reacting rapidly with 0 2 • The potential at which oxygen is reduced by· 3 2 this system is probably limited by the Co + /Co + reduction potential of ca. O. 6 volts for these cobalt porphyrin complexes. Catalysts which exhibit more positive reduction potentials are ultimately desirable. In addition, the catalyst suffers from a gradual loss in activity, possibly due to oxidative degradation of the ligand by H2 0 2 or desorption of the catalyst from the electrode surface. Finally, the synthesis of these "face-to-face" porphyrins is a very tedious procedure involving many steps. A more simple ligand synthesis will likely be required for practical use of such catalysts. In an effort to surmount some of these problems inherent in the "face-to-face" porphyrin system, I propose the synthesis of "face-toface" 1, 10-phenanthroline macrocycles. The synthesis of such ligands could be accomplished by condensation of 2, 9-dichlorophenanthroline and a diamine such as 1, 3-diaminopropane or _Q-phenylene diamine (equation 2). 8 Polymerization reactions will certainly be a complication 4 167 in such a synthesis, but high dilution may decrease the extent of polymerization. Surprisingly, polymerization was not a problem in - the synthesis of the prototype macrocycle, 3, which was synthesized 8 in 94% yield! Cl NH 2 K2C0 3 + ni trobenzene ) 180° Ct NH 2 3 ,...., The variability of the R-groups which connect the two macro- - cycles in 2 allows for systematic changes in the metal-to-metal distance. For the species, M-0-0-M, the optimum metal-to-metal distance is ca. 4. 8 A, 2 and molecular models indicate a M-M distance of 4-5 A for R =propylene in the proposed system. - This "face-to-face" phenanthroline ligand system, 2, possesses several advantageous features: The neutral ligand will lead to a system with a +4 overall charge including two divalent metals (such as Co2 +). This will certainly raise the reduction potentials for the metals relative to the overall neutral "face-to-face" metal porphyrins" 9 Since the reactivity of cobalt complexes with oxygen seems to be proportional to the amount of negative charge on the metal center, 1 O a more positively charged system may show reduced oxygen affinity. 168 For this reason, a variety of metals such as Co2+ , Fe 2+ , Ru2+ , and Cu+ will be investigated in an effort to produce the desired combination of high reduction potentials and high oxygen affinity. Metal complexes of during oxygen reduction. Also, the overall +4 charge of the bimetallic the aromatic ligands in 2 should be resistant to oxidative degradation system should increase its solubility in polar solvents and thereby aid in the characterization of these species. These highly charged compounds are likely to be water soluble, as opposed to the neutral "faceto-face" metal porphyrins. Finally, the synthesis of the proposed - binucleating ligand, 2, appears straightforward and the required reactants are commercially available, or have been made in good yields by published procedures. 8 Fortunately, there are no possible isomers of proposed system i, contrary to the ~ and anti isomers which are formed in the synthesis of the "face-to-face" porphyrins. - (Compound 1 is the anti isomer.) -- - - As an alternative synthesis of 2, the macrocycle 3 could be • synthesized by the published method, 8 followed by alkylation with a dihalide such as 1, 3-dibromopropane to yield the desired product, i. An early indication of the viability of the proposed type of synthesis can be obtained from the reaction depicted in equation 4. Since 1, 2-diaminobenzene is commercially available, and 2-chlorophenanthroline can be synthesized in three steps by published methods, the below reaction could be consummated with relative ease. their own right, since molecular models indicate that these metal Bimetallic complexes of the ligand, 4, would be quite interesting in complexes would assume a face-to-face geometry due to the steric 11 169 4 base repulsion present in any other geometry. Hence, the synthesis of the - ligand, 4, presents a convenient starting point for this project, and it may not even be necessary to synthesize the more complicated, doubly- - bridged ligand system, 2. The syntheses proposed herein should involve standard organic and inorganic laboratory techniques. No special precautions need be taken except in the handling of the low-valent metal complexes, which should be done under an inert atmosphere to prevent air oxidation. The compounds will be fully characterized by techniques such as elemental analysis, IR spectroscopy, Nl\1R spectroscopy (for the ligands and diamagnetic metal complexes), ESR spectroscopy (for the paramagnetic metal complexes), electronic spectroscopy (for the metal complexes), and mass spectroscopy. The reduction potentials for the metal complexes will be determined by either cyclic voltammetry or differential pulse voltammetry. The catalytic activity of these com- 170 plexes towards dioxygen reduction will be studied at a rotating platinum ring-disk electrode consisting of a graphite disk and a concentric platinum ring. The metal complex will be adsorbed onto the graphite disk, and then the rate of oxygen reduction at the disk can be monitored electrochemically. At the same time, the amount of undesirable hydrogen peroxide can be measured at the platinum ring by reoxidation to oxygen at a potential of ca. 1. 4 volts vs. nhe. 171 References 1. J. Bockris and S. Shrinivasan, "Fuel Cells: Their Electro- chemistry", McGraw Hill, New York (1969). 2. J. Collman, C. Elliott, T. Halbert and B. Tovrog, Proc. Natl. A cad. Sci. USA, 11, 18 (1977). 3. J. P. Collman, M. Marocco, P. Denisovich, F. C. Anson, and C. Koval, J. Electroanal. Chem., 101, 117 (1979). "'"""" 4. a) J. Fee, structure and Bonding, ~' 1 (1975); b) Vanngard, "Biological Applications of Electron Spin Resonance Spectroscopy", H. Swartz, J. Bolton, and D. Borg, eds., J. Wiley, New York, p. 411 (1972). 5. R.R. Gagne, C. A. Koval and T. J. Smith, J. Amer. Chem. Soc., ~' 8367-8368 (1977). 6. J. Collman, private communication (1979). 7. "Ring-Disk Electrodes", W. J. Albery and .M . L. Hitchma·n, Clarendon Press, -Oxford, 1971. 8. a) S. Ogawa and T. Yamaguchi, J. Chem. Soc., Perkin, 976 (1974). b) S. Ogawa, J. Chem. Soc., Perkin, 214 (1977). 9. G. Patterson and R. Holm, Bioinorganic Chem., _1, 257 (1975). 10. a) F. Basolo, Chem. Comm., 810 (1973). b) R. D. Jones, D. A. Summerville, andF. Basolo, Chem. Rev., -- 79, 142 (1979). 11. B. Halcrow and W. Kermack, J. Chem. Soc., 155 (1946). 172 PROPOSITION 2 Gold Metallocycles Metallocyclic compounds have been postulated as intermediates _ in several reactions catalyzed by transition metals. Included in these are the [2 + 2] cycloaddition of olefins, olefin metathesis, and isomerization of strained carbocyclic rings. For example, an iridiocyclic compound has been isolated and demonstrated as the intermediate in the iridium catalyzed dimerization of norbornadiene (equation 1). 1 ) / Cl NBD ) ·\ H80 (1) An indication that metallocyclic compounds may be involved in olefin metathesis was demonstrated by the reaction of 1, 4-dilithiobutane with tungsten hexachloride (equation 2). 2 Possibly the most novel example wc1 6 + :c~: ~ [ °f;wci~J ~ ( not oc.tua fI y obse.rved) CH 2 : c.HD 85'°fo CHz.:::.CHz ( IJ.ol CHO= CHO) (2) 10 173 of all is the valence isomerization of cubane by rhodium shown in equation 3. 3 Here again the metallocyclic complex was not observed Rh(l)) (3) directly, but its probable existence is indicated by the isolation of the following acyl-rhodium adduct (equation 4). 3 Note that both the [2 +2] (4) '1 cycloaddition of olefins and the valence isomerization are thermally "forbidden" by Woodward-Hoffman orbital symmetry conservation rules. 4 In addition to the above examples, metallocyclic complexes have been observed with titanium, 5 tantalum, 6 iron, 7 nickel, 8 and platinum. 9 The most complete studies involve nickel and platinum. A series of platinocycles have been prepared by Whitesides et al. 9 for comparison with acyclic platinum alkyls. The platinocycle, - 2, was found to be thermally much more stable compared to the acyclic - - analogue, 1. Complex 1 yields ethylene and ethane (1 :1) upon decom- position, while complex 2 yields butene (primarily 1-butene). The 174 ........ mechanism suggested for beth decompositions is /3-elimination to form a metal hydride complex, followed by reductive elimination as shown in equations 5 and 6. Presumably, {3-elimination of the acyclic O"'=-CH2 + Ol3C.H3 + P+(o) (5) ) P+(o) (6) - complex, 1, follows dissociation of a triphenylphosphine ligand since the rate of decomposition is greatly suppressed by excess triphenylphosphine. lO Strangely enough, the rate of decomposition of the - metallocycle, 2, is accelerated by excess triphenylphosphine. Hence, a different mechanism may be involved for the metallocycle. Nevertheless, the metallocyclic complex is much more stable than the acyclic analogue, indicating that the facile pathway to {3-elimination - observed for 1, is not as readily available for the metallocyclic 175 - complex 2. It is postulated that these {3-eliminations occur most readily when the M-C-C-H dihedral angle is 0°. For this reason, the - tetramethylene metallocycle, 2, in which M-C-C-H dihedral angles would be greater than 90° ,: is expected to show a decreased rate of /3-elimination, as observed. 9 The rate of decomposition is also lowered for the analogous pentamethylene and hexamethylene platinocyclic complexes. Since these metallocycles exhibit increased resistance to thermal decomposition via {3-elimination, other pathways of decomposition may become available for metallocyclic compounds. These include reductive elimination with concomittant carbon-carbon bond breaking and/or carbon-carbon bond formation (i.e., equation 7). The only platinocycle D < Q > which gave a significant amount of decomposition other than - /3-elimination was the tri-n-butylphosphine analogue of 2 (equation 8). 9 - ) (8) The multitude of decomposition pathways available to metallocycles is demonstrated by nickelocyclopentanes,in which different 176 major products are observed depending upon the concentration of tricyclohexylphosphine (L) (Scheme 1). 8 The three-coordinate complex, +L~ LN1.:) '-L L2NO 3 +L :.ii '-L L3No s- ---- Lf ,,...,._ J, J t \_fl D ~C~=C'12 ~ Scheme 1 ~ - 3, decomposes via J3-elimination. This is not surprising since a vacant coordination site is available for coordination of the ~-hydrogen. The mechanism of decomposition of the four- and five-coordinate complexes is unknown. In all three cases, a mixture of products is observed, which may be a result o.f the equilibrium present between the three-, four-, and five-coordinate complexes. A very clean product was obtained from the bidentate ligand, diphos, which should exist almost - entirely as the four-coordinate complex, 6 (equation 9). Clean products (9) 177 are also obtained by oxidation with dioxygen, followed by decomposition (equation 10). In each case, cyclobutane was formed as the major □ product ( >85%). Photochemical decomposition resulted in a higher yield of ethylene in all cases as shown in equation 11. The results D (11) '-ti% from oxidation and photolysis have not been explained, 8 and further study is certainly warranted. To augment the ~ormation known about the formation and decomposition of metallocycles, I propose a study of gold{III) metallocycles. These metallocycles would be of special interest for comparison to the metallocycles discussed above, especially those of nickel and platinum. The modes of decomposition of such gold metallocycles may be sensitive to changes in geometry, coordination number, oxidation state, and orbital occupancy as was observed with the nickelocyclic compounds. 8 A comparison of the stability and modes of decomposition between aurocycles and their acyclic analogues may provide additional information about the nature of metallocyclic compounds (vide infra). Gold metallocycles could be synthesized in a 178 manner analogous to known acyclic alkyl gold compounds (Scheme 2). 12, 13 > X::. 6F1 >PF6 , ClOi, > etc. L: PPh., , PMe.3 >As Ph3 J etc. Scheme 2 ~ - Acyclic alkyl gold complexes such as 9, decompose exclusively by reductive elimination to give alkyl-coupled products (e.g., equation 12). 13 + 179 A number of factors influence the stability of these dialkylgold com- - plexes. Cationic complexes such as 9 eliminate faster than neutral analogues. Bulky ligands increase the rate of reductive elimination, while excess ligand {L) and bidentate ligands decrease the rate of elimination. This and other mechanistic information indicate that the mechanism involves dissociation of a ligand, L, followed by intramolecular, c~certed reductive elimination from an unstable threecoordinate intermediate. A similar investigation of the mechanism for decomposition of metallocyclic gold compounds is proposed. A mechanism similar to that described above for the acyclic complex, !!_, may be expected; however, platinocyclic compounds were shown to decompose by a different mechanism than their acyclic analogues, although the overall process appeared to be the same (vide supra). 9 For this reason, the dependence of rate on phosphine (L) concentration will be examined to determine if dissociatio~ of ligand is involved in the rate-determining step. A free radical process can be discovered by use of a radical trap such as di-t-butyl nitroxide, radicals of which could be observed by EPR. 13 Intermolecular processes would be evidenced by the formation of larger rings as shown in Scheme 3. Intermolecular decompositions would not be expected for cationic - aurocycles such as 9, since the dialkyl acyclic analogues seem to proceed via an intramolecular mechanism. However, an intermolecular mechanism has been suggested for neutral trialkyl gold(ill) compounds such as 10. 14 The mechanism, however, is solvent.,._.. 180 + [ CAu(~ C Al.l..,. L 2. □ '> infro.r1\?(ec.ular + > • 'L infer mol ec.<.da.r 8 ,...,_ Scheme 3 ~ dependent, showing only intramolecular coupling in donor solvents such as dimethyl sulfoxide (DMSO) (Scheme 4). Yet this result contra- CH 3 I oc-Au.-L 3 ' cH 3 t-&Ph > DJV1SO . CD3 C:D3 ~ co3 cH3 + CH3 CH3 6o/o > CD3 C.H..3 75°/c SO% 4'1¥o + CH3 Cf/3 25¾> /0 ,._, Scheme 4 ~ diets an earlier study by the same group in which the decomposition of trialkyl gold; complexes in tert-butyl benzene was claimed to be intramolecular. 15 The inconsistency here may be due to difficulty in quantitatively analyzing the amount of deuterated products, which was done by mass spectroscopy. It should be easy to differentiate between the intramolecular and intermolecular routes for the aurocycles, 181 since the products differ greatly as shown above. Tri-alkyl substituted aurocycles should be prepared for comparative purposes. These cyclic derivatives could be prepared as sho.vn in Scheme 5, which is similar to the synthesis of acyclic analogues such as 10. 15 , 16 , 17 -- All (P?h3) Cl + Li "\/v Ll CA1.tPPh3-Li + • + RI Scheme 5 ~ -- Derivatives such as 11 may undergo an intermolecular mechanism - more readily than cationic complexes such as 9, which will exhibit electrostatic repulsions when approaching one another. Another interest~ mechanistic application would be the - -- synthesis of labeled derivatives of 9 and 11, proposed in Scheme 6. • ~Li Lyu D D Scheme 6 ~ 182 Scrambling of the deuteriums in the decomposition products would indicate more complicated rearrangements rather than simple concerted reductive elimination. For instance, Scheme 7 is a possibility: + ~ o: 0 1l D + Aull + A/" L DP~ + 'L DD AuL2.+ Scheme 7 ~ In contrast to nickel 11 and platinum 10 acyclic dialkyl complexes, which tend to decompose via ,S-elimination as a major route, acyclic trialkyl gold(III) complexes decompose exclusively via reductive elimination and coupling of alkyl groups (equation 13). 13 , 15 For this -~> CH 3Cl-lzCl-llCJ.13 + so% 183 reason, the decomposition of gold metallocycles may not be complicated with the ~-elimination products as in the nickel and platinum cases. Accordingly, aurocyclic complexes may give the first good comparison between the stabilities of acyclic and cyclic transition metal alkyl complexes for which the exclusive mode of decomposition is reductive elimination and concerted carbon-carbon bond formation. Complexes such as ll. are very intriguing in this respect. The products from decomposition of ll.. should indicate whether elimination of two cyclic groups or one cyclic and one acyclic group is faster (Scheme 8). These G Au/R 'PPh 11 □ + RAu.PPh3 OMSO 3 Scheme 8 ~ two routes should easily be distinguished by analysis of the gaseous products and by NMR of the resulting solution. Alkyl coupling between alkyl groups in the Au(I) products shown above is a possibility, but this reaction is slower and can be completely eliminated by a slight excess oftriphenylphosphine. 14 As was true with the nickelocycles, ethylene may be a decompo- - -- sition product from aurooycles_ such as 9 and 11. An attempt could be made to induce the formation of ethylene by photolysis. (Recall that photolysis of nickelocycles favors production of ethylene.) Excess ligand also resulted in the formation of ethylene (reason unknown). 8 184 Hence, it would be useful to study the effect of excess triphenylphosphine on the product distribution. Since Au(ID) complexes are rarely 5-coordinate, 18 this effect should be slight, if any; however, the potentially three-coordinate ligand, 1, 1, 1-tris-( dimethylphosphinomethyl) ethane 19 as well as the pctentially bidentate diphos ligand may provide interesting results. For instance, consider the following possible complexes, 12 and 13. """"' """"' + 12 13 Finally, the effect of oxidation on these complexes should be investigated. As was tl)e case with nickel, 8 rapid reductive elimination to form cyclobutane would be expected, but the situation is somewhat different here in that oxidation of Au(III) complexes by one electron followed by reductive elimination would yield an Au(II) complex. Gold(II) complexes are extremely unstable 20 and generally lead to disproportionation. Two rea_sonable possibilities for the oxidative decomposition of these aurocycles are shown on Scheme 9o The second case, which shows disproportionation of the A 1:1(II) intermediate accompanied by alkyl transfer, was observed in the oxidative decomposition of tetramethylaurate(Ill) compowids (equation 14). 21 In the above reaction, triphenylphosphine was required, presumably to 185 [CA~~+~ [ou~~ZT~ B ,,.,.,._ " Scheme 9 ~ stabilize the Au(II) intermediate; otherwise reductive elimination 0 directly to Au was observed. The effect of triphenylphosphine (L) - concentration on the oxidative decomposition of 11 merits investigation. Oxidation of these complexes could be affected by adding molecular oxygen, but this presents an added variable due to the possibility of 0 2 complexing to the metal. Complexation of dioxygen to nickel may be involved in the decomposition of nickelocycles by 0 • 8 , 22 Olefins 2 with electron-withdrawing substituents were shown to bind to dialkyl nickel compounds and promote reductive coupling of the alkyl groups. 11 A similar process could occur with dioxygen. Therefore, in addition to decomposition by dioxygen, decomposition could be generated by 186 electrochemical oxidation, in which the oxidation is most likely a simple electron-transfer. Also, the effect of electrochemical reductions could be easily studied electrochemically, and cyclic voltammetry will show the relative stability of any oxidized or reduced species. Finally, the rate constants of follow-up chemical reactions (i.e., reductive elimination) can possibly be determined by double pctential-step chronocoulometry and/or current reversal chronopctentiometry. 23 187 References 1. A. R. Fraser, P. H. Bird, S. A. Bezman, J. R. Shapley, R. White, J. A. Osborn, J. Amer. Chem. Soc., ~' 597 (1973). 2. --------- • (1972). 3. -- R. Grubbs and T. Brunck, J. Amer. Chem. Soc., 94, 2538 L. Cassar, P. Eaton, and J. Halpern, J. Amer. Chem. Soc., -- 92, 3515 (1970). 4. R. B. Woodward and R. Hoffmann, Angew. Chem., Int. Ed. Engl., ~' 781 (1969). 5. G. Whitesides and J. McDermott and M. Wilson, J. Amer. -- Chem. Soc., 98, 6259 {1976). 6. S. J. McLain, C. D. Wood, and R. R. Schrock, J. Amer. Chem. -- -Soc., 99, 3519 (1977). 7. M. Lo H. Green, Adv. Organomet. Chem., !!_, 29 (1970). 8. R. H. Grubbs, A. Miyashita,- M. Liu, and P. Burk, J. Amer. Chem. Soc., 100,. 2418 {1978). ~ 9. G. Whitesides, J. White, and J. McDermott, J. Amer. Chem. , 10. Soc., ~' 6521 (1976). G. Whitesides, J. Gaasch, and E. stedronsky, J. Amer. Chem. Soc., ~' 5258 {1972). 11. T. Yamamoto, A. Yamamoto, and S. Ikeda, J. Amer. Chem. -- -Soc., 93, 3350 (1971 ). 12. F. H. Brain and C. S. Gibson, J. Chem. Soc., 762 (1939). 13. P. Kuch and R. Tobias, J. Organomet. Chem., ffi, 429 {1976). 14. S. Kamiya, T. A. Albright, R. Hoffmann and J. K. Kochi, -- J. Amer. Chem. Soc., 98, 7255 (1976). 188 References (continued) 15. A. Tamaki, S. Magennis, and J. Kochi, J. A mer. Chem. Soc. , 96, 6140 (1974). "" 16. A. TamakiandJ. Kochi, J. Organomet. Chem., fil:__, 441 (1973). 17. A. Tamaki and J. Kochi, J. Organomet. Chem., g, C39 (1973). 18. F. Cotton and G. Wilkinson, "Advanced Inorganic Chemistry", Inters.c ience, Newt ork, 1972, p. 1054. 19. G. Whitesides, C. Casey, and J. Krieger, J. Amer. Chem. Soc., -- 93, 1379 (1971). 20. J. Huheey, "Inorganic Chemistry", Harper & Row, New York, 1972, p. 334. 21. S. Komiya, T. A. Albright, R. Hoffman, J. K. Kochi, J. Amer. Chem. Soc., 99, 8440 (1977). "" 22. M. Almemark and B. Akerkmark, J.C. S. Chem. Comm., 67 (1978). • 23. R. W. Murray in "Physical Methods of Chemistry, Part IIA: Electrochemical Methods", A. Weissberger and B. Rossiter, Eds., Wiley Interscience, New York, 1971, Ch. VIII. 189 PROPOSITION 3 Models for Blue Copper Proteins The Type 1, or "blue" copper, protein site is present in a large number of proteins including azurin, stellacyanin, plastocyanin and lac case. 1 This site is characterized by an intense absorption envelope centered around 600 run aqd an E PR spectrum having an unusually small hyperfine coupling constant (A 11). These properties suggest a novel coordination environment for copper(II) in these protein sites, and a considerable amount of research has been directed toward the full characterization of the Type 1 site. The blue copper site in poplar plastocyanin has been characterized by X-ray diffraction, and the copper was found in a distorted tetrahedral geometry, the ligands being a cysteine thiolate, a methionine thioether, and two histidine imidazole nitrogens. 2 An X-ray structural analysis for P. aeruginosa azurin was also perfor~ed and aH:hough the resolution was poor, the probable ligands are the same as in the plastocyanin site. 3 This would seem to indicate a common copper(II) coordination environment in all the Type 1 sites, but at least stellacyanin must be different since it does not contain methionine. 4 In addition, the large range of reduction ' potentials (200-800 mV vs. nhe ) exhibited by the Type 1 binding sites suggests a variation in structure and/or ligands. 1 There could be a common feature (e.g., mercaptide coordination) which is responsible for the chara·c teristic Type 1 properties, while other features may vary among the different protein sites. 190 Numerous spectroscopic studies are in agreement with a near tetrahedral arrangement of sulfur and nitrogen ligands around copper in the Type 1 sites. 5 The intense absorption at ca. 600 nm in these blue copper proteins has been assigned to a S( cysteine) - Cu(II) a -1.MCT. 6 The small hyperfine splitting (AJ pobserved in the EPR spectra of these sites has been accounted for by assuming a tetragonally flattened tetrahedral copper(II) geollfetry. 5 While the novel properties and structures of the blue copper protein sites are beginning to be understood, these sites have yet to be effectively modeled by synthetic complexes. The difficulty in obtaining such model complexes is largely due to the tendency of copper(II) compounds to instantaneously oxidize mercaptides to disulfides. Mercaptide coordination to copper(II) has been observed in some complexes at low temperature, and indeed, strong absorption bands attributable to S(thiolate) - Cu(II) LMCT are observed for these compounds. 6 , 7 However, tp.e other ligands in these complexes bear little resemblance to the proposed prctein sites. The electronic spectra of a number of copper(II) complexes employing thioether coordination have been examined, and S(thioether) - Cu(II) LMCT absorptions have been identified. 7 , 8 The other ligands as well as the geometries in these complexes fail to mimic the blue copper prctein sites. The few cases of mercaptide to copper(II) coordination which are stable at room temperature hav,e rather complex structures, and again bear little overall resemblance to the proposed Type 1 site. 9 ' 1O In no case has this blue copper site been mimicked completely in terms of coordination geometry and ligands, electronic absorption spectrum, 191 reduction potential and stability at room temperature. As an effective model for the blue copper prcxein site, I propose the synthesis of the copper complex,!.• In addition to providing a tetra- 1 hedral arrangement of appropriate ligands to mimic the Type 1 site, this complex contains considerable steric bulk around the thiolate sulfur in an effort to prevent the oxidative coupling of the mercaptides. A possible synthesis of the ligand is given in Scheme 1. All of the proposed reactions have ample precedent. ll-lS If the dimethylthiocarbamate group fails as a prcxecting group, the p-toluenesulfonyl group may be used on 3, 5-di-!_-butyl-2-hydroxybenzyl alcohol and .. removed after formation of the thioether, prior to the reaction with dimethylthiocarbamyl chloride. 14 - Once the ligand is in hand, the desired compound, 1, could be synthesized directly from copper(II) starting materials, or, if . necessary, the copper(!) complex could be synthesized, followed by 192 -... r ~ 'z~z c,, :c z ~i) C, ( / , z :z:~ . ~✓ <:J tn _.re z I "" % ~ o~z U) u~ o; :c " c,,, 2: <n=u · ~ ~ 0~2 ':r0 - zN 7 ~ V) ~1z: u '~ ~~ ~~ :z: II V) ~i ~,., ~ ~ :ra Cl)::~~ ,z zJ : u - "' - ir~ ~J z ~5 £' % +:c u _u -=-N' f'f) ~ (l) s (l) ..c: C.) 00 I z t( ~ ~ 0<.J~ · ::z:fz: o I ~ ~ _,,,,u tt ~ 0 ~ ~ 0 z'v> ~, N ii 1 z -oN 1- -::- ~ ,::,. 1 ::C <:1 o:z: ti)~/ ~ V') I z z~ ~ 193 electrochemical oxidation to yield the desired compound. Initial studies on this complex should include electronic absorption, EPR, and electro chemistry for comparison to the blue copper protein sites. It would be of interest to investigate the effect of slight mcx:lifications - in the geometry and type of ligands around copper(II). The ligand, 2, appears to be amenable to systematic variations of the donor atoms. For instance, consider cqmplexes 1 and .1 which could be synthesized - along with 1. The effects of these modifications on the electronic absorption spectrum, EPR spectrum and redax potentials could yield .• information about the possible factors responsible for the unique features of the Type 1 site. + + 'c /~ s '\ 'cu/ / ., N- N--- N / N- ~N---- 4:vN,W - ~N~N-...._ -3 4 - It is possible that complex 1 may exist as Cu(I) and a sulfur radical at room temperature, rather than the desired Cu(II) and thiolate anion. This formulation, in which the mercaptide has been 194 oxidized instead of the Cu(II) ion, has been suggested for the "visible" copper site (Cua) of cytochrome .£ oxidase. 17 Modification of the ligand to include "harder" nitrogen or oxygen donors (~. , complex 1) should stabilize copper(II) relative to copper(!) and favor the Cu(II)-Sformulation. These studies may in fact yield informative comparisons to the Cua site in cytochrome .£ oxidase. If additional steric pulk is necessary to prevent the oxidative - coupling of the mercaptides, the binuclear complex, 5, could be - synthesized from p-cresol in a manner analogous to the synthesis of 1. - In this binuclear system, the best model for the blue copper sites 2 would probably be the mixed valence species Cu(II)Cu(I) L + ( L = the binucleating ligand in .§_ ). The two copper(II) ions in the fully oxidized species are likely to be antiferromagnetically coupled, possibly leading to a diamagnetic system. 18 The extent of such coupling is relevant to both cytochrome .£ axidase and hemocyanin, since both proteins contain pairs of metals which a~e strongly coupled, presumably through a bridging ligand. 19 ' 20 In neither case is the identity of the bridging ligand known, and a mercaptide bridge from cysteine is a possibility. - The binuclear complex, 5, or slight variations of it, may indeed be reasonable models for the binuclear copper site in hemocyanin. 20 In conclusion, the study of a series of copper complexes such as !., ~' 1 and .§_ could yield informative comparisons with a number of copper containing proteins. The necessary ligands could be synthesized by standard organic techniques. Subsequent studies of the copper complexes should include characterization by elemental analysis, I. R. spectroscopy, NMR spectroscopy (for the ligands and 195 h+ 5 diamagnetic metal complexes), EPR spectroscopy, electronic absorption spectroscopy, and perhaps mass spectroscopy. The reduction potentials for the metal complexes could be determined by cyclic voltammetry, differential pulse voltammetry or by classical polarography. More s~phisticated techniques such as circular dichroism and magnetic circular dichroism spectroscopy, 5 photoelectron spectroscopy, 21 variable temperature magnetic susceptibility, 22 and measurement of electron transfer kinetics 5 could be employed for further characterization and comparison to the copper protein sites., 196 References 1. Fee, J. A. Struct. Bonding (Berlin) ~ ' 23, 1. 2. Colman, P. M.; Freeman, H. C.; Guss, J. M.; Murata, M.; Norris, V. A.; Ramshaw, J. A. M.; Venkatappa, M. P. Nature (London) ~ ' 272, 319. 3. Adman, E. T. ; Stenkamp, R. E. ; Sieker, L. C. ; Jensen, L. H. J. Mol. Biol. ~.'123, 35. 4. Bergman, C.; Gandvik, E. K.; Nyman, P. O.; strid, L.; Biochem. Biophys. Res. Commun . .!!!TI, 77, 1052. 5. Solomon, E. I.; Hare, J. W.; Dooley, D. M.; Dawson, J. H.; Stephens, P. J.; Gray, H.B. J. Amer. Chem. Soc.~' 102, 168. 6. Thompson, J. S.; Marks, T. J.; Ibers, J. A. Proc. Natl. Acad. - Sci. U.S.A. 1977, 74, 3114. ~ 7. Amundsen, A . . R.; Whelan, J.; Bosnich, B. J. Amer. Chem. Soc . .!!!TI, 99, 67·30. 8. Miskowski, V. M.; Thich, J. A.; Solomon, R.; Schugar, H. J.; J. Amer. Chem. Soc. ~ ' 98, 8344. 9. Schugar, H. J.; Ou, C.; Thich, J. A.; Potenza, J. A.; Lalancette, R. A. ; Furey, W. J. Amer. Chem. Soc. ~ ' 98, 3047: 1 O. Hughey, J. L. ; Fawcett, T. G.; Rudich, S. M.; Lalancette, R. A.; Furey, W. J. Amer. Chem. Soc. .,.._,..._,.._,_ 1979, 101, 2617 . 11. Tang, C. C.; Davalian, D.; Huang, P.; Breslow, R. J. Amer. Chem. Soc. ~ 1978, 100, 3918. 197 References 12. Gagne, R. R.; Kreh, R. P.; unpublished results. 13. Bruson, H. Org. Reactions ~ ' 5, 79. 14. Ullmann, F.; Brittner, K. Chem. Ber. ~ ' 42, 2539. 15. Newman, M. S.; Karnes, H. A. J. Org. Chem. !,lli, ll_, 3980. 16. Gillis, R. G.; Lacey, A. B.; Org. Synthesis~, Coll. Vol. IV, I 396. 17. Chan, S. I.; Bocian, D. F.; Brudvig, G. W.; Morse, R. H.; Stevens, T. H. in"CytochromeOxidase", King, T. E. etal. eds.; Elsevier/North-Holland Biomedical Press: Amsterdam, 1979, p. 177. 18. Copper(II) ions bridged by oxygen ligands are often strongly coupled. For example, see: Lintvedt, R. L.; Tomlonovic, B.; Fenton, D. E.; Glick, M. D. Adv. Chem. Series~' 150, 407. 19. Blumberg, W. E.; Peisach, J. in "Cytochrome Oxidase", King, T. E. et al. eds.; Elsevier /North Holland Biomedical Press: Amsterdam, 1979; p. 153. 20. Lontie, R. ; Witters, R. in "Inorganic Biochemistry", Eichorn, G. I. ed.; American Elsevier: New York, 1973; p. 344. 21. Wurzbach, J. A.; Grunthaner, P. J.; Dooley, D. M.; Gray, H. B.; Grunthaner, F. J.; Gay, R. R.; Solomon, E. I. J. Arner. Chem. Soc. 1977, 99, 1257. ~ 22. - Solomon, E. I.; Dooley, D. M. ; Wang, R.; Gray, H. B.; Cerdonio, M.; Magno, F.; Romani, G. L. J. Amer. Chem. Soc. 1976, 98, 1029. ~ 198 PROPOSITION 4 Macrocyclic Cuprous Complexes as Photosensitizers The valence isomerization of norbornadiene {NBD) to quadricyclene (Q) has been investigated as a method for photochemical energy I hv ·~ Se.tlsiti z.e..r NBD t > h Q catalyst I storage. l, 2 The reaction is endoth~rmic by 21-26 kcal • mole- which 1 corresponds to a storage capacity of 240 calories per gram of quadricyclene produced. 3 While quadricyclene is indefinitely stable at ambient temperatures, ~his energy can be released by exposure to appropriate catalysts. 4 Since norbornadiene does not absorb in the wavelength region ( > 300 nm) of available solar energy, a "sensitizer" is necessary for absorption of light at longer wavelengths. A variety of copper(!) complexes have been investigated as sensitizers. Cuprous chloride has been found to sensitize the conversion of NBD to Q when irradiated with light at a wavelength of 313 nm. 5 A quantum yield 6 of ca. O. 39 with catalytic factors 7 up to 394 have been ol:tained for CuCl in neat NBD. The proposed mechanism is outlined in Scheme 1. The initially formed ClCu-NBD complex exhibits an 3 1 1 intense charge transfer band ( Amax 248 nm, e = 6. 3 x 10 M- cm- in 199 Scheme 1 ~ 1 ~ l. /)..__ '1 + CuCI 3 - b--cuCI 7 ~ ~ CuCI + h ~CuCI ~ CuCI > 200 ethanol) which increases the absorption of the system out to 350 nm. The direction of this charge transfer was not assigned (MLCT, or LMCT), but in either case a weakening of the bonding between C1 -C 2 and C3 -C 4 with a strengthening of the bonding between C1 -C 3 and C2 -C 4 is expected. 8 Relaxation of this excited state can then lead to either Q or NBD. Upon formation of Q, it is lost from the cuprous coordina- tion sphere and replaced ~ NBD to begin the catalytic cycle again. The conversion of NBD to Q is also sensitized by copper , • phosphine compounds, but the mechanism appears to be different (Scheme 2). 9 The copper complex, 1, appears to be four-coordinate - Scheme 2 hv Cu(PPh) 2 BH 4 -~> Cu(PPh3) 2 BH 4 * 1 - (with a bidentate BH 4 ligand), and there is no evidence for ground state - coordination of NBD to copper. Compound 1 exhibits an intense absorption centered at 257 nm, and it arises from the transfer of a lone pair electron on phosphorus to an empty antibonding TT orbital centered on a phenyl ring. 9 Excitation of this transition by incident light at 313 nm produces an excited Cu(I) complex which apparently interacts with NBD to generate Q. Quantum yields up to 0. 56 were observed with excess NBD in benzene solutions. Low catalytic factors ("' 10) were observed due to photodegradation of Cu(PPh) 2 BH4 .. 201 This sensitization can also be accomplished by a four-coordinate Cu(I) complex containing a single substitutionally labile ligand, e.g. , i. lO The carbon monoxide ligand is replaced by NBD,and the reaction is proposed to proceed by the mechanism outlined in Scheme 1 (for CuCl). The complex Cu[HB(pz) 3] (NBD) exhibits a charge transfer absorption which extends out to 370 nm. Irradiation (313 nm) of a solution of ~ and excess NBD produces Q with a quantum yield of 0. 6 and catalytic factors > 65. The present proposition suggests the investigation of fourcoordinate macrocyclic copper(!) complexes as sensitizers. for the photochemical conversion of NBD to Q. The cuprous complexes shown in Figure 1 have recently been prepared electrochemically from the copper(II) analogues. 11 The majority of these compounds bind carbon monoxide as a fifth ligand (Table 1). 11 - 13 While the binding of these compounds to olefins has not been investigated, it is likely that they will also hind NBD as a fifth ligand. Thus, these macrocyclic copper(!) complexes are candidates for the sensitization of the conversion of NBD to Q by a mechanism similar to Scheme 1. 202 n cti::(~ XV X9iA ~ c~,: y ~ I '9 / E_ J= ~J= r8'9 u \ 3 I I 4 5 ~\CuI A I \ ~ I v-0 6 ~ \ 66+ ~ 7 8 10 11 Macrocyclic copper(!) complexes which are proposed as sensitizers for the conversion of norbornadiene to quadricyclene. 203 Table 1 ** E* f Keo 11 -0.762 6.1 X 10 4 11 9 - -0.683 2. 5 X 10 4 11 8 -0.616 4. 7 X 10 15 10 .,.._,..._ -0.594 < 10 11 3 - -0.456 1. 2 X 10 5 15 4 -0.451 1. 5 X 10 3 13 7 - -0.438 8. 8 X 10 5 15 5 -0.270 1. 2 X 10 3 15 - -0.169 5. 3 X 10 3 13 Complex """ 6 Reference *on pt or Hg electrode (no significant differences noted) in dimethylformamide. Potentials converted to values versus nhe using ferrocene as an internal standard. 16 ** By the method in reference "-15. 204 The complexes shown in Figure 1 present several possible advantages as sensitizers over the previously studied systems. A major problem with the systems presented thus far is the lack of light absorption in the visible region (400-800 nm) where the intensity of sunlight is greatest. The complexes shown in Figure 1 are all colored - due to visible or near U. V. absorptions. Most nctably, complex 3 is blue, due to an absorptiofi at 677 nm. 12 Als~ the charge transfer absorptions, Cu __. NBD MLCT, are likely to be at longer wavelengths than those observed for ClCu-NBD and Cu [ HB(pz) 3] (NBD). This is predicted for the (presumed) more electron-rich copper(!) centers which result from the high a -donating character of the "hard" imine nitrogens. The high a -donating character of these ligands also results in a relative destabilization of Cu(I) vs. Cu(II) when bound to these ligands. This is manifested by rather negative reduction potentials (Table 1). In the event that these macrocyclic cuprous compounds do not bind NBD, they may still sensitize the conversion of NBD to Q by an excited state interaction with NBD (see Scheme 2). In this case, the visible absorptions (~., at 677 nm for ~ may be useful. The binding constants for NBD to these Cu(I) compounds can be determined spectraphotometrically or electrochemically. 12 • These macrocyclic cuprous compounds present two additional attractive features as sensitizers. First, the pres~nce of a single coordination site favors the desired intramolecular rearrangement rather than oligimerization or fragmentation processes which appear to be favored by multiple coordination sites on the rnetaL 14 Finally, 205 the organic framework of these macrocyclic ligands provides positions which can be functionalized for attachment to a solid support. This is desirable for the isolation of the sensitizer from the catalyst (to prevent cross-contamination) and to facilitate the replacement of the 3, 4 • • a wor k.mg so la r energy st orage dev1ce. • sens1·t izer m 206 References 1. Kutal, C. Adv. Chem. Ser. 1978, 168, 158. 2. Kutal, C.;Grutsch, P.A. Adv. Chem. Ser.~' 173,325. 3o ----------- - • Kutal, C. ; Schwendiman, D. P. ; Grutsch, P. Sol. Energy !J!TI, 19, 651. 4. Sweet, E. M.; King, R. B.; Hanes, R. M.; lkai, S. Adv. Chem. Ser. ~ ' 173, 344/ 5. Schwendiman, D. P.; Kuta!, C. J. Amer. Chem. Soc. !J!TI, 99, 5677. 6. Quantum yield ( <P) is defined by <P = number of molecules undergoing a particular change number of quanta absorbed 7. Catalytic factor • =moles of Q produced per mole of sensitizer present. 8. Schwendiman, D. P.; Kutal, C. Inorg. Chem. !J!TI, 16, 719. 9. Grutsch, P. A.; Kutal, C. J. Amer. Chem. Soc. 1979, 101, 4228. 10. sterling, R. S.; Kutal, C. Inorg. Chem. ~ ' .!.!!_, 1502. 11. Lisensky, G. C., Ph.D. Disseration, California Institute of ~ - Technology, 19 81. 12. Allison, J. L., Ph.D. Dissertation, California Institute of Technology, 1979. 13. Ingle, D. M., Ph.D. Dissertation, California Institute of Technology, 1980. 14. Salomon, R. G.; Kochi, J. K. J. Amer. Chem. Soc. ~ 1974, 96, 1137. Salomon, R. G.; Sinha, A.; Salomon, M. F. J. Amer. Chem. Soc. !JU!!, 100,520; Evers, J. T. M;Mackor, A. 207 References (continued) Tetrahedron Lett. 1978, 2317. ~ 15. Gagne, R. R., Allison, J. L., Ingle, D. M., Inorg. Chem. ~ ' 18, 2767. 16. Gagne, R. R., Koval, C. A., Lisensky, G. C., Inorg. Chem. ~ ' 19, in press. 208 PROPOSITION 5 Transition Metal Formyl Chemistry The current energy shortage has prompted much interest in the Fischer Tropsch synthesis reaction. 1 This process involves the metal catalyzed reduction of CO by H 2 to yield mixtures of paraffins, olefins, alcohols, etc. Metal for~yl complexes, !., have been proposed as 0 II LnM-C-H 1 intermediates in this process, and as a result they are the subject of numerous current investigations. These studies have involved the synthesis and reactivity of transition metal formyl complexes. The most general preparative route is given in equation 1. 2 - 4 0 HB(OR)3 + ½:M(CO) -, B(OR)3 + ½:MAH (1) M = Cr, W, Re, Fe The most stable formyl complexes have been formed with rhenium and iron. The normal method of decomposition involves loss of CO, followed by hydride migration to yield a metal hydride complex ~ - , equation 2). 5 Alternatively, these formyl compounds can act as 209 0ti -co 0II _ _ (CO) 4 Fe-CH ~ (CO) 3 Fe-C-H ~ (CO) 4 FeH (2) hydride donors to a number of electrophiles (e.g. , Scheme 1). 4 Scheme 1 ~ 0 II Li+ [CO 5ReRe(CO) 4 CH] - I -50° C OLi I C (CO) 5 ReRe(CO) 5 + C6H5 -CH 2 Fe(CO) 5 - In contrast to these reactions, the iron formyl compounds, 2 and 2a, react differently toward . acid. 3 - 5 The reason for this different - N(Pc/) 3) ; [(CO) 4 Fe(CHO)] 2 H+ > 0 HAH (3) 0 H+ ti Et 4 N+ [ ( <p()) 3 P(CO) 3 Fe-CH ~ CH 3 OH (4) 2a reactivity is not obvious, but it may be related to the presence of an - available coordination site on iron (for H+). The reactivity of these compounds apparently does net directly reflect the relative stability of 210 the metal formyl complexes which was determire d to be: [(CJI 5) 3 P] 2 N+(C0) 4 FeCHO- > Et 4 N+ cis-(CO) 9 Re 2 CHO- > Et4 N+ trans[CJI O] P(CO) FeCHO-. 3 The mechanisms for these processes are 5 3 3 unknown, but two proposed routes for the production of methanol are outlined in Scheme 2. Scheme 2 --~" - - The analogous reactions of 2 and 2a with carbon electrophiles do not yield aldehydes as would be predicted from the above acid reactions. For instance, complex 2 reacts with 1-iodooctane to give - octane and only a trace of nonan al. This result is consistent with direct nucleophilic displacement of iodide by the formyl hydrogen atom 3 or with the more complicated mechanism shown in Scheme 3. 5 211 Scheme 3 ~ 0 - II (C0) 4 Fe CH Cs H11 I + k, ' k_, 2 I 0 -CO,ki <: + CO k-2 I II k3 ~ (CO\ ~eCH ' k-3 CaH,1 ~ 0fl ( C0)4 FeCH C' 8 H,1 3 (C0)4 ~eH CaH,1 4 + Ca H,a An investigation into this mechanism is proposed. In particular, the dependence of the reaction rate on carbon monoxide pressure will be investigated. While the direct hydride transfer mechanism should show no dependence on the CO pressure, the mechanism in Scheme 3 indicates that the reaction should be inhibited by CO. Assuming that ~ » k_ 3 and that processes k4 and ks are fast and irreversible leads to the following rate expression (equation 5) when ~eady-state kinetics 212 = (5) -d[~] (5a) dt are applied to 1 and .1_. 6-~ A plot of k~~s versus [CO]- should yield 1 a straight line. Even if the assumptions used here are invalid, the rate should be affected by [CO] , contrary to the direct hydride attack mechanism. The reaction can be conveniently monitored by the disappearance of the metal formyl proton magnetic resonance (pmr) signal at 14. 95, o versus TMS in THF. 5 Low temperature and/or high CO pressure - may slow down the reaction such that intermediate 4 can be observed by pmr. Evidence supporting the mechanism in Scheme 3 would be significant for several reasons. It suggests the direct attack of electrophile on the metal center, with possible ramifications for the reactions of iron formy l complexes with acids (equations 3 and 4). Also the mechanism would seem to suggest either a greater tendency for reductive elimination with hydride and formyl substituents compared to alkyl and formyl substituents or a favoring of CO extrusion (reverse migratory insertion) from the formyl ligand when an alkyl ligand is bound to the iron. To supplement this study, an analogous investigation into the - products and mechanism of the reaction between 2 and a smaller 213 carbon electrophile ( ~ , CH 3 I) may be of interest. Also, the effect of [CO] on the reaction rates for the processes in Scheme 1 (for [(C0) 5ReRe(C0) 4 C(O)H] -) could be examined for comparative purposes. 214 References 1. Olive, G. H.; Olive, S. Angew. Chem., Int. _Ed. Engl. ~ ' 15, 136. 2. Casey, C. P.; Neumann, .S. M. J. Amer. Chem. Soc. 1976, ~ 98, 5395. 3. Casey, C. P.; Neumann, S. M. J. Amer. Chem. Soc. 1978, ~ 100, 2544. / 4. Gladyz, J. A.; Tam, W. J. Amer. Chem. Soc. llll, 100, 2545. 5. Collman, J. P.; Winter, S. R. J. Amer. Chem. Soc., 1973, ~ 95, 4089. 6. It is reasonable to assume Is « :k_ 3 since migratory insertion of CO into a M-H bond is unknown. (See reference 5.) 7. Reductive elimination of an alkyl and a hydride from a transition metal (k4 ) is usually fast and almost never reversible. See: --- Abis, L.; Sen, A.; Halpern, J. J. Amer. Chem. Soc., 1978, 100, 2915. 8. The unsaturated 16-electron species, Fe(CO) 4 , should quickly bind ancther CO ligand (ks), and the equilibrium should lie far to the right. See: Basolo, F.; Pearson, R. G.; "Mechanisms of Inorganic Reactions"; John Wiley and Sons, Inc., New York, 1967, pp. 533-578.