The Syntheses and Characterization of Binuclear Clathrochelate Complexes

Citation

Marritt, William Alan (1983) The Syntheses and Characterization of Binuclear Clathrochelate Complexes. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/xpqf-ns89. https://resolver.caltech.edu/CaltechTHESIS:10302018-160040560

Abstract

A series of binuclear clathrochelate complexes, M _A (II)M _B (II)L ⁺ , have been synthesized and characterized for the purpose of examining electrostatic, magnetic, and electronic delocalization effects in binuclear complexes. The clathrochelating ligand, L ^3- , is a bicyclic Schiff-base compound derived from 3 equiv of 2-hydroxy-5-methylisophthalaldehyde and 2 equiv of 2,2',2"-triaminotriethylamine (tren). The following M _A (II)M _B (II)L ⁺ complexes were prepared by combinations of metal-ion template, metathesis, and insertion reactions: homonuclear complexes with M _A = M _B = Mn, Fe, Co, Cu, Zn, and Cd; heteronuclear complexes with M _A = Mn, M _B = Co, Ni, Zn, Cd, and Mg; M _A = Fe, M _B = Mn, Co, Ni, Cu, Zn, Cd, and Mg; M _A = Cu, M _B = Mn, Co, Ni, Zn, Cd, and Mg. In addition, syntheses and characterization are reported for the binuclear Cu(II)Cu(II)LH ²⁺ complex and mononuclear complexes of the non-deprotonated ligand, MLH ₃ ⁿ⁺ , with M = Na(I), Mn(II), Mn(III), Fe(II), Fe(III), Co(II), Ni(II), Cu(II), Zn(II), Cd(II), and Mg(II).

The M _A redox potentials of the Mn(III)/Mn(II), Fe(III)/Fe(II), and Cu(II)/Cu(I) couples were determined by cyclic voltammetry and differential pulse voltammetry for three series containing both homonuclear and heteronuclear complexes, Mn(II)M _B (II)L ⁺ , Fe(II)M _B (II)L ⁺ , and Cu(II)M _B (II)L ⁺ . Within experimental error, the average of the M _A redox potentials for each series of heteronuclear complexes was the same as that for the corresponding homonuclear complex. This result suggests that there is no measurable stabilization of Mn(III)Mn(II)L ²⁺ , Fe(III)Fe(II)L ²⁺ , and Cu(I)Cu(II)L with respect to mixed-valent heteronuclear analogues.

The magnetic properties of a series of four homonuclear M _A (II)M _B (II)L ⁺ complexes were examined by variable temperature susceptibility measurements (M _A = M _B = Mn, Fe, Co, and Cu). In all cases, weak antiferromagnetic coupling between the metal centers was observed (-33.5 cm ^-1 < J < -0.8 cm ^-1 ). The weakness is attributed to limited overlap of metal orbitals with those of the bridging atoms.

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):

Grubbs, Robert H.

Thesis Committee:

Bercaw, John E. (chair)
Gray, Harry B.
Gagne, Robert R.
Grubbs, Robert H.

Defense Date:

2 August 1982

Funders:

Funding Agency	Grant Number
Caltech	UNSPECIFIED

Record Number:

CaltechTHESIS:10302018-160040560

Persistent URL:

https://resolver.caltech.edu/CaltechTHESIS:10302018-160040560

DOI:

10.7907/xpqf-ns89

Default Usage Policy:

No commercial reproduction, distribution, display or performance rights in this work are provided.

ID Code:

11252

Collection:

CaltechTHESIS

Deposited By:

Benjamin Perez

Deposited On:

31 Oct 2018 18:50

Last Modified:

06 Aug 2025 23:33

Thesis Files

Preview

PDF - Final Version
See Usage Policy.
48MB

Repository Staff Only: item control page

THE SYNTHESES AND CHARACTERIZATION OF BINUCLEAR CLATHROCHELATE COMPLEXES Thesis by William Alan Marritt In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1983 (Submitted August 2, 1982) ii ACKNOWLEDGMENTS While it is impossible to thank all of the people who have contributed to the success of my four year struggle at Caltech, it likewise is impossible not to mention a special few whose support and friendship have been especially meaningful. Much credit and thanks are gratefully given to Bob Gagne roth for giving me the freedom to explore my own research ideas as well as for offering me good, creative ideas when on occasion mine were somewhat lacking. Scientific discussions wi±h the other members of the Gagne research group were always enjoyable, especially the interactions with Larry Henling, John Dodge, and Dave Marks. The generous financial support of the Caltech Department of Chemistry for 1982 is appreciated, as is the interest and support shown for me during these last months by the Inorganic faculty. Some close friendships were formed during these four years. The companionship of Greg Hillhouse made life outside of the lab more fun and enjoyable. He was always ready to use his scientific brilliance to help with a variety of inorganic and organometallic problems, spent many hours obtaining my FT NMR spectra, and, fortunately, never shut me out in one-on-one basketball. Tom Shih, my first friend in Pasadena, has been a major source of answers for my organic chemistry questions; more importantly, he has been invaluable in helping to keep my fine car in running condition. Good times with Don Tilley will be long remembered, especially his introducing me to the challenging sport of frisbee golf and showing me the iii quickest way to LAX. As a sometimes Hankie, a big "Thank You" goes to Scotty, Tommy, Ricky, Jonathan, and Chester for many enjoyable (especially when I played) Thursday nights at Hazel's. Members of the chemistry faculty at Ohio Wesleyan University, especially Harold Wilcox and Larry Wilson, deserve credit for fueling my interest in chemistry. Sincere thanks also go to Henriette Wymar for an excellent job done in typing this thesis and for putting up with my delaying tactics. Finally and most importantly is a loving thanks to my parents and family for their unwavering support (both moral and financial!) during my twenty-one years of schooling. iv ABSTRACT A series of binuclear clathrochelate complexes, MA (II)MB(II)L +, have been synthesized and characterized for the purpose of examining electrostatic, magnetic, and electronic delocalization effects in binuclear 3 complexes. The clathrochelating ligand, L - , is a bicyclic Schiff-base compound derived from 3 equiv of 2-hydroxy-5-methylisophthalaldehyde and 2 equiv of 2, 2', 2" -triaminotriethylamine (tren). The following MA (II)MB(II)L + complexes were prepared by combinations of metal-ion template, metathesis, and insertion reactions: homonuclear complexes with MA = MB = Mn, Fe, Co, Cu, Zn, and Cd; heteronuclear complexes with MA = Mn, MB =Co, Ni, Zn, Cd, and Mg; MA =Fe, MB = Mn, Co, Ni, Cu, Zn, Cd, and Mg; MA = Cu, MB = Mn, Co, Ni, Zn, Cd, and Mg. In addition, syntheses and characterization are reported for the binuclear Cu(II)Cu(II)LH 2+ complex and mononuclear complexes of the non-depro- tonated ligand, MLH~+, with M = Na(I), Mn(II), Mn(III), Fe(II), Fe(III), Co(II), Ni(II), Cu(II), Zn(II), Cd(II), and Mg(II). The MA redox potentials of the Mn(III)/Mn(II), Fe(III)/Fe(II), and Cu(II) /Cu(I) couples were determined by cyclic voltammetry and differential pulse voltammetry for three series containing both homonuclear and heteronuclear complexes, Mn(II)MB(II)L +, Fe(II)MB(II)L +, and Cu(II)MB(II)L +. Within experimental error, the average of the MA redox potentials for each series of heteronuclear complexes was the same as that for the corresponding homonuclear complex. This result 2+ suggests that there is no measurable stabilization of Mn(III)Mn(II)L , 2 Fe(III)Fe(II)L +, and Cu(I)Cu(II)L with respect to mixed-valent heteronuclear analogues. v The magnetic properties of a series of four homonuclear MA (IT)MB(II)L + complexes were examined by variable temperature susceptibility measurements (MA = MB = Mn, Fe, Co, and Cu). In all cases, weak antiferromagnetic coupling between the metal centers was observed (-33.5 cm- 1 < J < -0.8 cm- 1 ). The weakness is attributed to limited overlap of metal orbitals with those of the bridging atoms. vi TABLE OF CONTENTS Page ~ 1 Introduction ~ 12 The Syntheses and Electrochemical Properties of Binuclear Clathrochelate Complexes ~ 87 Magnetic Properties of Binuclear Clathrochelate Complexes ~ 112 Mononuclear and Binuclear Complexes of 1, 3-Bis(2-pyridylimino)isoindolines ~ Synthesis, Characterization, and Dioxygen Reactivity Studies of Mononuclear and Binuclear Iron Complexes of 1, 3-Bis(2-pyridylimino)isoindolines 157 CHAPTER 1 Introduction 2 Many of the most interesting, primary enzymatic processes involve multinuclear metal centers.1 ' 2 Processes such as substrate binding, substrate activation, and electron transfer often require the presence of two or more active metals to supplement the bioorganic "backbone" of the enzyme. Indeed, arguments can be made that the major role of the organic portion of the protein in enzymes such as cytochrome c oxidase, a metalloprotein with two iron and two copper centers, 2 is to position and rigidly hold the metal ions in the appropriate interacting environment that is required for the proper functioning of the enzyme. In this respect, one of the major challenges of modern synthetic inorganic chemistry is to prepare simple models of the enzyme's active metal sites which mimic the behavior of the metalloproteins. At a more fundamental level, many research groups are preparing systems containing two or more metal atoms in structurally well defined environments to study the nature of basic metalmetal interactions. An understanding of these processes, such as electrostatic interactions, magnetic interactions, and electronic delocalization between metals is essential to elucidating mechanisms of the more complex phenomena of enzyme action. An understanding of these metal-metal interactions has equally important applications for non-enzymatic systems. For example, multimetal systems are used in industrially important catalytic reactions, 3 in some fuel-cell electrode materials, 4 and in artificial photosynthetic devices used for solar energy conversion. 5 A typical example of the first system is the bismuth-molybdate conglomerate which is used in the SOHIO process for the catalytic ammoxidation of 3 propylene to acrylonitrile. 6 Co-facial bimetallic porphyrin systems show some promise as possible fuel-cell electrode components for the 7 electrocatalytic reduction of 0 2 to H20. Typical examples of the 8 third system include colloidal Ru02 /Pt/Ti02 particles and molybdenumhalogen clusters. 9 A few multimetallic systems yielding information relating to the fundamental nature of intermetallic interactions have been particularly well studied. Foremost among these systems are the ruthenium dimer complexes which have been examined extensively by Taube,10 Meyer, 11 and Sutin, 12 among others. These systems have provided considerable information on electronic delocalization phenomena for mixed-valent states. Another well studied system of this type includes complexes of type ,.... 1. These studies have addressed the nature of both magnetic 1 interactions 13 and electronic delocalization. 14 A final example, sy~tems containing two or more iron atoms studied by Ludi15 and Brown, 16 show electronic delocalization similar to the ruthenium dimers 4 mentioned previously. However, since iron is a common enzymatically 17 active meta1 (in contrast to ruthenium) the results from these systems have added significance in light of their biological applicability. Our approach towards understanding metal-metal interactions in a well-defined ligand environment entailed the preparation of binuclear 3 complexes of the ligand, L -, and examination of their properties by electrochemical and magnetic techniques. Complexes of this type in N 0- N 3 L 3- which metal ions are completely encapsulated by a rigid polydentate ligand are known as clathrochelates. Such complexes were first 18 described and named in 1964 by Busch; however, an example was not 19 forthcoming untill968. Since there are relatively few examples of these clathrochelate complexes, a brief comprehensive review is given below. The first example, 2, reported by Rose, is a mononuclear Co(III) "' 20 complex prepared by the reaction of Co(dimethylglyoxime)!+ and BF3 • 5 This reaction method, addition of a Lewis acid to alkoxy substituents, 21 proved to be useful for the preparation of related clathrochelates. Holm and co-workers prepared a series of the mononuclear clathrochelates (3) by a multistep synthesis using Rose's Lewis acid "" addition in the last step. 22 X-ray structural studies for a series of + 3 "" 6 these complexes provided useful information for correlating metal-ion electronic configuration with the metal's preference for trigonal prismatic vs. trigonal antiprismatic coordination geometry. Sargeson and co-workers have prepared several mononuclear clathrochelates incorporating a bicyclic octaamine ligand which they have dubbed sepulchrates (4). 23 The Co(III) complex is prepared from ,... 4 the reaction of Co( ethylenediamine):+ with ammonia and formaldehyde. This particular complex has found application in electron-transfer studies with the ligand remaining fully associated with the metal ion in 24 the normally substitution-labile +2 oxidation state. In the course of studies of planar, macrocyclic Ni(II) complexes, Goedken and co-workers isolated mononuclear clathrochelates (5) ,.. as minor byproducts . 25 The synthesis of these clathrochelates has since been improved and generalized for metal ions other than Ni(II) . 26 7 n+ 5 " In a recent report, Busch and co-workers detail the syntheses and characterization of a new family of clathrochelates (6). 27 A unique " 2+ 6 " aspect of this system is that these complexes are prepared by a base- 8 induced rearrangement of an originally tetradentate complex. Additionally, the authors suggest the possibility that this rearrangement reaction might be chemically reversible. The new clathrochelates of L s- described herein are unique with respect to the other known clathrochelates in that they include two metal ions within a rigid ligand environment. Syntheses and characterization of both homonuclear and heteronuclear complexes of L s- are presented in Chapter 2. In addition, electrochemical studies of these complexes are reported and the results are discussed in terms of their relationship to metal-metal interactions. In Chapter 3, a variable temperature magnetic study for a series of these complexes is detailed. This work was done in collaboration with Professor D. N. Hendrickson and Mark Tim ken at the University of Illinois. Additionally, the syntheses and characterization of several mononuclear and binuclear complexes of 1, 3-bis(2-pyridylimino)isoindoline ligands are described in Appendices 1 and 2. 9 REFERENCES ~ 1. (a) Lehninger, A. L. "Biochemistry"; 2nd Ed., Worth Publishers, Inc.: New York, 1976, p. 488; (b) Mason, H. S. ~ Ann. Rev. Biochem. ~' 35, 595; (c) Lontic, R. "Inorganic Biochemistry"; Eichorn, G. I., Ed., Elsevier: Amsterdam, 1973, p. 344; (d) Klotz, I. R. Bioi. Macromol. lin., §_, 105; (e) Carrico, R. J.; Deutsch J. Biol. Chern. !11Q, 245, 723. !111, ~' 389. 2. Malmstrom, B. G. Quart. Rev. Biophys . 3. (a) Muetterties, E. L.; stein, J. Chern. Rev. ~' 79, 479; (b) Vidal, J. L.;Walker, W. E. Inorg. Chern.~' 19,896. 4. "Fuel Cell Systems-II"; Gould, R. F., Ed., Am. Chern. Soc. Publications: Washington, D. C., 1969. 5. Gratzel, M. Ace. Chern. Res. ill!_, 14, 376. 6. Burrington, J. D.; Grasselli, R. K •. J. Catalysis, illQ, 59, 79; 1980, 63, 235. ~- 7. (a) Collman, J. P. Marrocco, M.; Denisevich, P.; Koval, C. A.; Anson, F. C. J. Electroanal. Chern.~' 101, 117; (b) Collman, J. P.; Denisevich, P.; Konai, Y.; Marrocco, M.; Koval, C.; Anson, F. C. J. Am. Chern. Soc. ~' 102, 6761. 8. Kiwi, J.; Borgadello, E.; Pelizzetti, E.; Visca, M.; Gratzel, M. Angew. Chern., Int. Ed. Engl. illQ, 19, 646. 9. (a) Gray, H. B.; Maverick, A. W. Science ill!_, 214, 1201; (b) Maverick, A. W.; Gray, H. B. J. Am. Chern. Soc. ~' 103, 1298. 10. (a) Magnuson, R. H.; Taube, H. J. Am. Chern. Soc . .!1.ll, 94,7213. 10 REFERENCES (continued) ~ (b) Tom, G. M.; Creutz, C.; Taube, H. J. Am. Chern. Soc. li71., 96, 7827; (c) Creutz, C.; Taube, H. J. Am. Chern. Soc. ![G, 95, 1086; (d) Sutton, J. E.; Taube, H. Inorg. Chern . .!Jjg_, 20, 3126; (e) Stein, C. A.; Taube, H. Inorg. Chern. 1981, 20, ~- 3126. 11. Callahan, R. W.;Keene, F. R.;Meyer, T. J.;Salman, D. J. J. Am. Chern. Soc. 1977, 99, 1064. ~- 12. Sutin, N.; Creutz, C. Pure App. Chern . .!J!Q, 52, 2717. 13. (a) Gagne, R. R.; Spiro, C. L.; Smith, T. J. ; Hamann, C. A.; Thies, W. R.; Shiemke, A. K. J. Am. Chern. Soc. ~' 103, 4073; (b) Long, R. C.; Hendrickson, D. N. submitted to J. Am. Chern. Soc. 14. (a) Lambert, S. L.; Hendrickson, D. N. Inorg. Chern. ~' 18, 2683; (b) Spiro, C. L. ; Lambert, S. L.; Smith, T. J.; Duesler, E. N.; Gagne, R. R.; Hendrickson, D. N. Inorg. Chern. 1981, 20, 1229; (c) Spiro, C. L.; Lambert, S. L.; Gagne, R. R . ; ~- Hendrickson, D. N. Inorg. Chern. ~' 21, 68. 15. (a) Felix, F.; Ludi, A. Inorg. Chern. ~' 17, 1782; (b) Sullivan, B. P.; Ludi, A. J. Am. Chern. Soc. 1977, 99, 7368. ~- 16. Dziobkowski, C. T.; Wrobleski, J. T.; Brown, D. B. Inorg . .Chern. ~' 20, 679. 17. Nielands, J. B. in "Metal Ions in Biological Systems"; Dahr, S. K., Ed., Plenum: New York, 1972, p. 13. 18. Busch, D. H. Rec. Prog. Chern. ~' 25, 107. 11 REFERENCES (continued) ~ 19. Boston, D. R.; Rose, N. J. J. Am. Chern. Soc. 1968, 90, 6859. 20. (a) Boston, D. R.; Rose, N. J. J. Am. Chern. Soc. ll71, 95, ~- 4163; (b) Jackels, S. C.; Duerdorf, S. S.; Rose, N. J.; Zehtzer, J. J. Chern. Soc., Chern. Comm. ~' 129_2; (c) Jackels, S. C.; Rose, N. J. Inorg. Chern. llll, 12, 1232. 21. Parks, J. E.; Wagner, B. E.; Holm, R. H • . J. Am. Chern. Soc • .!&lQ, 92, 3500; In or g. Chern. !.!tl!, 10, 2472. 22. Larsen, E.; La Mar, G. N.; Wagner, B. E.; Parks, J. E.; Holm, R. H. Inorg. Chern. ~' .!.!., 2652. 23. Creaser, I. I.; Harrowfield, J. McB.; Herlt, A. J.; Sargeson, A. M.; Springborg, J.; Gener, R. J.; Snow, M. R. J. Am. Chern. Soc. 1m, 99, 3181. 24. Ferraudi, G.J.; Endicott, J. F. Inorg. Chim. Acta 1m, ~ 219. 25. Goedken, V. L.; Peng, S. M. J. Chern. Soc., Chern. Commun. 1973, 62. ~ 26. Goedken, V. L. "Inorganic Syntheses"; Vol. XX, Busch, D. H., Ed.; John Wiley and Sons: New York, 1980, p. 87. 27. Herron, N.; Grzybowski, J. J.; Matsumoto, N.; Zimmer, L. L.; Christoph, G. G.; Busch, D. H. J. Am. Chern. Soc.~' 104, 1999. 12 CHAPTER 2 The Syntheses and Electrochemical Properties of Binuclear Clathrochelates 13 INTRODUCTION ~ Binuclear complexes of type 1 have been studied by electro"' 5 1 3 2 chemical, ' magnetic, - and spectroscopic 2' 6 methods. These studies have been aimed at understanding the nature of interactions between metal ions. One general approach towards studying metal-metal interactions in complexes of type 1 is to examine physical, chemical, and/or redox "' properties of these complexes while varying the metal ions systematically . 1 ' 3 - 5 Differences in the measured properties are attributed to variations in metal-metal interactions, if it can be established that metal-ligand geometry is nearly identical for the series examined. In an electrochemical study of type .!. complexes with MA = Cu(II) and MB = Mn(II), Fe(II), Co(II), Ni(II), Cu(II), and Zn(II), there is evidence 14 which suggests that structural differences as a function of these MB ions are minimal. However, since axial-ligand geometry cannot be rigorously controlled in type 1 complexes, structural variations "' dependent on the nature of both metal ions are certainly possible. Indeed, there is structural evidence for differing axial-ligand geometries for several type 1 complexes. 7 - 9 "' Our approach towards examining metal-metal interactions in binuclear complexes was to prepare complexes related to those of type 1 for which metal-ligand geometries are likely to be less variable. "' s- The ligand which was chosen for this purpose, L , is a bicyclic Schiffbase compound derived from 3 equiv of 2-hydroxy-5-methylisophthaldehyde and 2 equiv of 2, 2', 2" -triaminotriethylamine (tren). The term N 0- I N-{ CH -+---+- N 22 3 L s- """"" clathrochelate has been used to describe complexes of ligands such as 3 L - in which the metal ions are totally encapsulated by a ligand. 10 15 Two conceptual illustrations of binuclear M(ll) clathrochelates of this ligand are shown in Figures 1a and 1b. Based on structural studies of mononuclear complexes with remotely similar ligands, 11 the 3 geometry for each metal ion in binuclear clathrochelates of L - is expected to be nearly trigonal antiprismatic 12 ' 13 (see Figure 1b). Since clathrochelating ligands have limited structural freedom, it was expected that metal-ligand geometries of M(ll)M(II)L + complexes would not vary significantly as a function of the metal ions. The synthesis, characterization, and electrochemical properties of M(Il)M(II) L + complexes are reported here. Also included is a discussion of the differences between the electrochemical behavior of M(II)M(II)L +complexes and that of type!. complexes with MA = Cu(II) and MB = Mn(II), Fe(II), Co(II), Ni(II), Cu(II), and Zn(II). Additionally, synthesis and characterization of mononuclear complexes of the nondeprotonated ligand, LH 3 , are reported. 16 + M( li ) M( JI ) L+ (a) (b) ~· Illustrations of M(II)M(II)L + showing (a) the chemical structure (the dotted line represents the third 2-hydroxy-5-methylisophthaldehyde fragment) and (b) the expected trigonal antiprismatic geometrical structure. 17 RESULTS ~ Synthesis. ~ L 3- Binuclear complexes of the clathrochelating ligand were first prepared by a metal-template reaction. Condensation reactions of 3 equiv of 2-hydroxy-5-methylisophthalaldehyde and 2 equiv of tren, in the presence of certain metal (II) salts and base, yielded cationic binuclear complexes, M(ll)M(ll)L + (eq 1). Thus far + 2 tren + 2 M(II) base > M(II)M(II)L + (1) M(II)= Mn and Cd 0 OH 0 the success of this method has been limited to Cd(II) and Mn(II) salts. Intractable mixtures, which probably contain polymeric Schiff-base compounds, were obtained from similar reactions using salts of Cr(II), Fe(II), Co(II), Ni(II), Cu(II), and Zn(II). The binuclear Cd(II) complex, Cd(II)Cd(II)L + has been isolated as No;, BPh;, c1o;, PF;, and BF; salts. Only the BF; salt is appreciably soluble in solvents such as acetonitrile, N, N-dimethylformamide (DMF), dimethylsulfoxide, and methylene chloride. For this reason, Mn(II)Mn(II)L +and related binuclear complexes, prepared by alternate methods (vide infra), were isolated as BF; slats. For the preparation of M(II)M(ll)L + complexes with metals other 18 than Cd(II) and Mn(II), it was necessary to devise synthetic strategies other than the one-step template synthesis. A multistep synthesis has been developed by which both binuclear and mononuclear clathrochelate complexes have been prepared. The primary step is the preparation of the Schiff base ~ derived from 1 equiv of tren and 3 equiv of 2-hydroxy- 0 N H I 0 3 2 5 -methylisophthalaldehyde. A difficulty in preparing ,... 2 is the high probability for the multifunctional reactants to form Schiff-base condensation polymers instead of ,... 2. It is likely, for this reason, that several attempts to prepare ,... 2 by methods reported for related Schiff bases 14 yielded intractable mixtures. The polymerization problem was circumvented, however, and ,...2 has been isolated in high yield by reacting tren with excess 2-hydroxy-5-methylisophthalaldehyde under conditions which favor the rapid crystallization of ,...2 (eq 2). . The second step in this multistep synthesis is the formation of mononuclear clathrochelate complexes using template reactions of ~ with both tren and a metal salt. Several template reactions were screened after initial attempts to prepare the free ligand, LH 3 , from 19 tren + > 2" XS 0 OH (2) 0 high dilution reactions of tren and 2 failed to yield identifiable products. -'"" We found that mononuclear complexes in which the ligand is not deprotonated, M(II)LH 3 (BF 4 ) 2 • H20, were isolable from reactions of~ with 1 equiv of tren and 1 equiv of certain metal(II) salts (eq 3). ~ + tren + M(II) - Thus far 2+ (3) M(II)LH 3 M(II) = Mn, Fe, Co, and Cd only Mn(II), Fe(II), Co(IT), and Cd(II) mononuclear complexes have been prepared by this approach; similar reactions using Cr(IT), Ni(II), Cu(II), and Zn(II) salts yielded intractable mixtures. This template method, although useful, has several drawbacks in addition to being metal-ion specific. First, to obtain material of reasonable purity, the complexes must be slowly crystallized over a period of several days. Second, despite slow crystallization, samples of M(II)LH 3 (BF4 ) 2 • H2 0 frequently contain detectable amounts of the corresponding binuclear complexes, M(II)M(II)LBF4 • xH 2 0. Levels of binuclear contamination are highest for the Mn(II) and Co(II) complexes. Third, product yields are relatively low, being less than 50%. For 20 these reasons, efforts were made to develop alternative synthetic 2+ routes to M(II)LH 3 complexes. As a result of these efforts, a general synthesis for M(II)LH~+ complexes has been developed which utilizes an alkali-cation template reaction. Treatment of~ with 1 equiv of both tren and NaN03 affords the mononuclear complex, NaLH 3 N0 3 , in high yield (eq 4). Yellow (4) crystalline material is obtained which is moderately soluble in solvents such as methanol, acetonitrile, and methylene chloride. Based on thermodynamic and kinetic studies of (alkali-cation)-cryptate complexes 15 by J. M. Lehn and co-workers, it seemed reasonable that exchange of divalent metal ions for Na + in NaLH 3 N03 would be thermodynamically favored and would occur rapidly. Indeed, upon addition of 1 equiv of a M(II) salt to a methanolic solution containing 1.1 equiv of NaLH 3 N0 3 , the initially orange-yellow solution changes color almost immediately to that characteristic of the divalent metal complex, M(II)LH~+. By 2+ this exchange method (eq 5), M(II)LH3 complexes of Mn, Fe, Co, Ni, M(II) + NaLH 3 N03 - 2+ M(II)LH3 + NaN0 3 (5) Zn, Cd, and Mg have been prepared and isolated as monohydrated BF; salts. These complexes are soluble in acetonitrile, DMF, and dimethylsulfoxide. Mononuclear complexes prepared by this method contain insignificant amounts of binuclear contamination. Apparently, for most metal ions, insertion of a second metal ion into M(II)LH~+ does not 21 occur as facilely as the initial exchange reaction. For Cu(ll), however, insertion of Cu(II) into Cu(ll)LH~+ appears to be competitive with the initial exchange of Cu(ll) for Na +. As a result, to prepare Cu(II)LH~+ uncontaminated by binuclear species, it is necessary to use a two-fold excess of NaLH 3 N0 3 relative to the Cu(II) salt. By this modification, a Cu(II) mononuclear complex has been prepared and isolated as Cu(II)LH 3 (BF 4 ) 2 • %H 2 0. Mononuclear complexes of trivalent metals also have been prepared both directly by exchange of the trivalent metal cations as well as indirectly by exchange of a divalent metal cation followed by electrochemical oxidation to the trivalent state. Treatment of NaLH 3 N03 with hydrated ferric chloride afforded Fe(m)LH~+ which was isolated as a BF; salt. The analogous Mn(III) complex was prepared by controlled potential electrolytic oxidation of Mn(II)LH:+ prepared from NaLH 3 N0 3 and Mn(II). In this case, Mn(III)LH~+ was isolated as a Clo; salt. The last step in the multistep synthesis of binuclear clathrochelate complexes, M(II)M(II)L +, is the insertion of a second metal ion into the mononuclear complexes, M(II)LH 3 (BF4 ) 2 • xH 20. In general, we found that addition of MB(II) salts to asolutionof M(II)LH 3 (BF4 ) 2 • xH 20, in the presence of triethylamine, yielded the expected binuclear complexes, M A (II)MB(II)LBF4 • xH 2 0 (eq 6). Us.ing this approach, homobase MA(II) + MB(ll)LH 3 (BF4 ) 2 • xH 20 -~) MA(II)MB(ll)LBF4 • xH 20 (6) binu_c lear complexes (MA = MB) of Mn(II), Fe(II), Co(II), Zn(II), and Cd(II) have been prepared. It is also possible to prepare these homo- 22 binuclear complexes more directly by reacting 2 equiv of a M(ll) salt with 1 equiv of Nal.JI 3 N03 under basic conditions. With the exception of Fe(II)Fe(IT)L +, however, this approach is unfruitful since a large fraction of the binuclear product precipitates as an insoluble No; salt. For Fe(II)Fe(IT)L +prepared directly from Nal.JI 3 N0 3 , only a minimal amount of Fe(II)Fe(IT)LN03 precipitates, which after removal leaves the bulk of the product for isolation as a BF; saH:. Several attempts to prepare similar homobinuclear complexes of Ni(II), Mg(ll), and Cu(II) by the insertion methods described above were unsuccessful. From reactions of either Ni(ll)LH~+ or Mg(II)LH~+ with an excess of the appropriate M(II) salt, in the presence of various bases, only unreacted mononuclear complexes have been isolated. Such lack of reactivity is not a limitation in the case of Cu(II). In describing the preparation of Cu(II)LH~+, it was noted that Cu(II) inserts facilely into Cu(II)LH:+ even in the absence of added base. The difficulty in preparing Cu(II)Cu(II)L +by the insertion methods described above is that one phenolic oxygen remains protonated and the binuclear complex Cu(II)Cu(II)LH(BF4 ) 2 • 2H 2 0 is isolated. We found, however, that Cu(II)Cu(ll)L +is isolable as a monohydrated BF; salt if, instead of triethylamine, diisopropylethylamine is used as a co-solvent. Optimal yields have been obtained by first isolating Cu(II)Cu(II)LH(BF4 ) 2 • 2H 2 0 under nearly neutral conditions and subsequently deprotonating the complex. The primary utility of the insertion reaction (eq 6) is the preparation of heterobinuclear complexes (MA ~ MB). Using this reaction the following heteronuclear complexes have been prepared: 23 Mn(II)M(II)L +, M = Co, Ni, Zn, Cd, and Mg; Fe(II)M(ll)L +, M = Mn, Co, Ni, Cu, Zn, Cd, and Mg; and Cu(II)M(II)L +, M = Mn, Co, Ni, Cd, and Mg. For Ct(II)ZnL +, it is necessary to use deprotonating conditions similar to those utilized in preparing Cu(II)Cu(II)L +. In general, a minimal amount of scrambling (i.e., 2 [MA (ll)LH:+] + 2MB(II) --+ MA (II)MA (II)L + + MB(II)MB(II)L +)was observed (vide infra) in the preparation of all these heteronuclear complexes. In many cases, however, there is a preferred method (i.e., MA(II) + MB(II)LH~+ vs. MB(II) + M A(II)LH~+) for preparing heteronuclear complexes which contain lesser quantities of homonuclear impurities. Details of the preferred methods are given in the Experimental Section. In general, both homobinuclear and heterobinuclear complexes can be obtained as microcrystalline materials. However, only in the case of Fe(II)Mn(II)LBF4 and Fe(II)Co(II)LBF 4 have crystals been obtained of sufficient quality for X-ray structural determination. Structural studies are presently being undertaken in collaboration with D. N. Hendrickson and co-workers. 24 Characterization ~ Infrared The infrared spectra of all binuclear complexes, M(II)M(II)L +, are nearly indistinguishable and show features characteristic of metal-imine complexes. 16 Complete imine formation is indicated by the absence of aldehydic carbonyl-stretching .bands in the 1660-1680 cm- 1 region. The characteristic imine-stretching vibrations typically appear as three bands in the 1590-1640 cm- 1 region. Typical separations and relative intensities of these bands are shown in Figure 2(a); Table I contains vC=N data for the homobinuclear compounds. In addition to imine-stretching bands, numerous unassigned ligand vibrations are observed in the 700-1640 em - 1 region. For all other binuclear and mononuclear complexes of this type, infrared spectra are obtained which show features similar to those exhibited by M(II)M(II)L + complexes. For example, the infrared 2+ spectra of the M(II)LH3 complexes show no significant changes as a function of the metal, but generally exhibit bands which are less well resolved than the corresponding M(II)M(II)L + complexes. An example of typical resolution is shown for imine-stretching bands in Figure 2(b). Magnetic Susceptibility. Magnetic moments, 1-Leff (295 K), of 2+ + M(II)LH 3 and homonuclear M(II)M(ll)L complexes show that each metal ion is high spin in the solid state (Table I). Additionally, for Fe(III)LH 3 (BF4 ) 3 and selected heteronuclear complexes, magnetic moments also are consistent with high spin electronic configurations. In comparing 1-Leff/M(II) values of homonuclear M(II)M(II)L + complexes with those of the corresponding M(II)LH~+ complexes, the binuclear lleff/M(II) values are slightly lower in all cases. This observation is 25 (A) (B) I I • v I 1700 1600 1500 1700 ' I I 1600 1500 WAVENUMBER CM-1 ~ Infrared spectra showing v C=N stretching bands for (a) Mn(II)Mn(II)LBF4 and (b) Mn(II)LH 3 (BF4 ) 2 • H2 0. 26 ~ Selected Infrared, Magnetic, and Conductivity Data for Binuclear and Mononuclear Complexes of L -1 s- , LH 2- , and LH 8 • 2 h Complex vC=N' em Mn(TI)Mn(II)L + 1639,1615, 1602 (sh) 5.59 72 Fe(II) Fe(II)L + 1628,1609, 1594 (sh) 5.18 66 Co(II)Co(ll)L + 1631,1608, 1596 4.58 64 Cu(II)Cu(II)L + 1643 ~sh), 1630, 1611 sh),1600(sh) 1. 77 74 Zn(II)Zn(II)L + 71 Cd(TI)Cd(TI)L + 72 2 Cu (II) Cu(II) LH + 2+ Mn(II)LH 3 2+ Fe(IT)LH 3 2+ Co(TI)LH3 2+ Ni(II)LH 3 1. 83 153 5.99 303 5.28 276 5.06 307 3.19 276 Cu(TI)LH:+ 1.80 316 Zn(II)LH~+ 306 Cd(II)LH~+ 275 2+ Mg(II)LH 3 301 Fe (III) LH: + -1 lleff/M (p.B) xMcm •O m ·mol 5.88 389 .. 1 27 consistent with some degree of antiferromagnetic metal-metal interaction in the binuclear species. Variable temperature suscer.tibility measurements confirm the antiferromagnetic nature of these interactions.17 ~· Of the homobinuclear complexes, M(II)M(II)L +, only the diamagnetic Cd(II) complex is sufficiently soluble in several solvents to obtain suitable NMR spectra. Both 1 H and 13 C { ~} NMR spectra are consistent with a symmetrical clathro- chelating solution structure (Table II). Two features of the 1 H NMR spectrum are particularly noteworthy. First, although the ethylene proton resonances are overlapping at 90-MHz, four inequivalent protons are resolved at 500-MHz (Figure 3). This observation of four signals is indicative of a static structure with rigid ethylene linkages. Geminal and vicinal coupling constants (Table II) were obtained by selective decoupling experiments and are comparable to values 18 observed for similar rigid systems. Also noteworthy is the splitting of one-fourth of the imine resonance into a satellite doublet. This observation is consistent with unresolved 111 113 ' Cd-H coupling (for Cd, I = i, and relative abundances are 12. 75% and 12. 26%, 19 111 113 respectively). Based on limited reports of ' Cd-H coupling and 111 Cd and 113 • similar behavior observed for Cd(II)LH:+, the measured J value of 36 Hz is ascribed to three-bond coupling. While the solubility of MLH~+ complexes is nat a limitation in 1 obtaining their H NMR spectra, the spectra of the paramagnetic complexes are either extremely broad or complex; thus,tentative peak assignments have not been made. In contrast, spectra of diamagnetic 28 Table II. NMRa Data for C~LBF4 • ~ + CH:3 H H H,c II II N CH 0 HbCHcHd-N'Cd..-O'C0 N-CHdHcCHbHa N 3 3 3 Type Assignment 1Hb, c CH 3 2. 20 s aryl 7.10 s imine 8. 23 sd Jlu, 113CdH = 36 Ha 2. 61 m JHH =13.4 JH H = 2. 7 Chemical shift, 6 Coupling a b a c 2. 85 m 3.02 m 3. 59 m CH 3 19.2 CHaHb, CHcHd 57.6,59.0 aryl 119. 9, 121. 8, 140.9,169.3 imine 171.1 JH H = 12.3 a d JH H = 1.1 b c JH H = 2.4 b d JHH =11.0 c d JHdH. . = 1. 0 unme 29 Table II. (continued) ~ ~R spectra were obtained at ambient temperature. Chemical shifts are reported in ppm relative to internal Me4 Si; coupling constants are reported in hertz. bspectrum was taken in acetonitrile-.9a. cThe alkyl region of this spectrum is interpretable only at 500 MHz; see Figure 2. Tentative assignments were made by selective decoupling experiments (coupling constants± 1 Hz). dA satellite doublet due to 111 113 ' Cd-H coupling also was observed. espectrum taken in dimethylsulfoxide-~. 6.0 _l__ ~0 PPM J 4.0 I ./ 3.0 I 20 I \...~-• ..J '-' \...) ~ ~ 500 MHz 1H NMR spectrum of Cd(II)Cd(II)LBF4 • The four inequivalent -CH 2 CH 2 - I 7.0 I 8.0 .)\..... solvent impurities. protons are resolved and appear at 2. 6, 2. 8, 3. 0, and 3. 6 ppm (6). Asterisks denote ~ ...... ~ • w 0 31 n+ MLH 3 complexes are interpretable and provide structural information on these complexes. The~ NM:R for all diamagnetic MLH:+ complexes show similar features and are consistent with a symmetrical clathrochelating structure, containing a metal ion in one site and three phenolic protons in the other site (Table ID). The phenolic pr<tons (He) typically appear as a single broad (>100Hz fwhm) downfield resonance which is resolvable as a doublet only at high Rf power. By deuterium exchange, it was established that the splitting of the He resonance is due to coupling to an imine proton (Hd, Table III). These observations are consistent with intracavity hydrogen bonding of the phenolic protons to proximal imine nitrogens, with the broadness of the He resonance being attributed to 14 N quadrupolar broadening. 2+ • • • For Cd(TI)LH 3 , additional couplmg of 111 Cd and 113 • Cd nuclei to the other set of three imine pr<tons (Ha) is observed. The J value of 37 Hz for this unresolved 111 113 ' Cd-H coupling is similar to that observed for Cd(II)Cd(II)L + (36Hz) . 1 The H NMR of NaLH 3 N03 have been examined in several different solvents. In methylene chloride-~, the spectrum of NaLH 3 N03 is similar to those of diamagnetic M(II)LH~+ complexes in acetonitrile-3s, exhibiting inequivalent aryl (Hb and He) and imine (Ha and Hd) resonances (Table III). As these spectra have been interpreted, the metal ions are chelated in one of two available sites and are not undergoing either site-site exchange or dissociative metal ligand processes on the NMR time scale. In contrast, the spectrum of NaLH 3 N03 in methanol-g4 or acetonitrile-~ shows two broad resonances at ,. ._ 7.1 and"' 8. 5 ppm corresponding to the aryl and imine pr<tons, respectively. This c...-Hd II e fCH2t2 CH 3 ,s He,br 1. 8-4.1 2.25 13.5 1. 8-4.1 2.12 13.9 8.34 JH H = 7 d e 8.25 Hd,d Hc,d 7.50 JH H =14 d e JH H = 2 b c 7.33 1. 7-4.0 2.19 13.9 8.36 7.51 HbHc- JH H =1 4 d e 1.9-4.1 2.36 13.6 8.40 7.35 7.55 7.34 -2 8.23 JH H =14 d e JH H = 2 b c Mg(IOLH3 (BF4 ) 1 • H1 0c 8.29 J Zn(II)LH 3 (BF4 ) 2 • H 20c Complex 7.09 J -2 7. 22 HbHc- d Cd(II)LH3 (BF4 ) 2 • H20 c Hb,d b 3 n• 8. 33 Jm, UliCdH = 37 NaLH3N03 3 CH 2)2""N'~~N-( CH 2 II Ha'c !-\,*He Q CH3 a n+ H NMR Data for MLH 3 Complexes 8. 21 1 Ha,s Assignment ~ c., 1:\:) 33 Table III (continued) ~ a1 H NMR spectra were taken at ambient temperature except where indicated. Chemical shifts (0) are reported in ppm relative to internal Me4 Si; coupling constants are reported in hertz. bSpectrum was taken at 1 o· c in methylene chloride-~. cSpectra were taken in acetonitrile-9s. dA satellite doublet due to 111 113 ' Cd-H coupling also was observed. e A range of chemical shifts is given since the 8 inequivalent alkyl protons are unresolved at 90 MHz. 34 observation strongly suggests that under these conditions the complex is undergoing rapid, reversible metal-ligand dissociation and/or the Na +ion is exchanging between the two sites without dissociating from the ligand. This type of NMR behavior is similar to that observed by J. M. Lehn and co-workers for 2:1 cryptate/alkali-cation-salt mixtures 20 in aqueous solutions. ~ Molar conductivities, AM, for mononuclear and binuclear complexes of this type are generally within the expected ranges (Table 1). 21 For example, the homonuclear complexes, M(II)M(II)LBF4 • xH 20, in DMF solutions have AM values (64-74 cm 2 • moC1 • ohm - 1 ) consistent with those of 1:1 salts 1 1 (60-80 cm 2 • mol- • ohm- ). Also determined in DMF, the anomalous homobinuclear species, Cu(II)Cu(II)LH(BF4 ) 2 • 2H 2 0, gives a AM value 2 (153 cm • mol- 1 • ohm- 1) within the normal range for 2:1 electrolytes 2 1 1 (130-170 cm • mol- • ohm - ). For mononuclear complexes, M(II)LH 3 (BF4 ) 2 • xH 2 0, in acetonitrile solutions, molar conductivities 2 1 vary from 275 to 316 cm • mol- 1 • ohm- • This range is slightly higher than that expected for 2:1 electrolytes in acetonitrile (220-300 cm2 • moC 1 • ohm- 1 ), although values as high as 1 336 cm2 • mol- 1 • ohm- have been observed previously in several 21 cases. For Fe(III)LH 3 (BF 4 ) 3 in acetonitrile, the AM value (389 cm 2 • moC 1 • ohm - 1 ) falls within the expected range for 3:1 1 electrolytes (340-420 cm2 • mol- 1 • ohm - ). Elemental Analysis. All of the complexes reported here have been analyzed for C, H, and N, and gave satisfactory analyses (Table IV). A ddit ionally, most complexes have .b een analyzed for the 12. 54(12. 56) 5. 88(5. 94) 6. 22(6. 04) 6.10(6. 35) 6.13(6.33) 6. 49(6. 75) 6. 76(7. 00) 11. 06( 11. 4 2) 11.17(11.0?) 12. 35(12.13) 12. 27(12.12) 12. 38(12. 08) 12. 07(12. 08) 11. 97(11. 90) 12. 05(11. 99) 11. 35(11. 42) 5. 02( 4. 98) 5. 26(5. 45) 5. 21(5. 45) 5.40(5.43) 5.16(5.44) 5. 36(5. 46) 5.26(5.07) 5.24(5.14) 46.64(46.31) 50. 70(50. 73 50. 66(50. 67) 50. 47(50. 50) 50. 32(50. 52) 49. 87(49. 77) 50.1 0(50.16) 47. 93( 47. 75) Mn(ll)LH 3 (BF4 ) 2 • H2 0 Fe(II)LH 3 (BF4 ) 2 • H20 Co(II)LH3 (BF4 ) 2 • H2 0 Ni(II)LH3 (BF4 ) 2 • H20 Cu(ll)LH3 (BF4 ) 2 • %H 20 Zn(ll)LH 3 (BF4 }2 • H20 Cd(II)LH 3 (BF4 ) 2 • H20 Cu(ll)Cu(II)LH(BF 4 ) 2 • 2H2 0 5.14(5. 09) 22. 88(22. 81) 47.44(47.53) Cd(II)Cd(II)LBF4 13. 84(14. 03) 11. 44(11. 37) 52. 46(52. 55) Zn(IT)Zn(IT)LBF4 12. 05(12. 37) 13. 08(13.14) 4. 69( 4. 60) 5.49(5.23) 51. 78(51. 71) Cu(IT)Cu(II)LBF4 • H20 12. 59(12. 49) 13.10(12. 80) 12. 99(12. 62) MA 13. 87(14. 67) 5.11(5. 28) 52. 24(52. 24) Co(IT)Co(IT)LBF4 • H20 13.17(12. 84) 12. 74(12. 87) N 12. 7 0( 12. 57) 5.29(5.20) 53. 82(53. 70) Fe(II) Fe(II) LB F 4 5.24(5.21) H 53. 96(53. 80) c Found (Calcd.) (%) Analytical Data for MA(II)MA(ll)L +, MA(II)MB(II)L +,and MALH~+ Complexes. Mn(ll)Mn(II) LBF4 Complex ~ Table IV. MB CJ1 (A) Table IV. (continued) 6. 02{6. 64) 13. 53(13. 32 12. 51(12. 49) 12. 58(12. 56) 12. 35(12. 34) 12. 01(11. 96) 5. 56{5. 39) 5. 20{5. 28) 5.21(5.19) 5. 23(5.19) 5.23{5.22) 4. 89( 4. 84) 55. 48(55. 71) 51.95{52.21) 52. 21(52. 51) 52. 41(52. 52) 51. 57(51. 61) 49. 75(50. 01) Fe(II)Mg(II)LBF4 Cu(TI)Mn(TI)LBF4 • H20 Cu{II)Co(II)LBF4 • ~H 20 Cu(II)Ni(II)LBF4 • !H2 0 Cu(II)Zn(II)LBF 4 • H20 Cu{II)Cd(TI)LBF 4 12. 56{12. 56) 5. 98(6. 01) 12. 26{12. 06) 5. 03(4. 88) 50. 82{50. 43) Fe(II)Cd{II)LBF4 11. 88(12. 00) 7. 50(7. 20) 6. 30(7. 00) 6. 46(6. 78) 6. 41(6. 58) 5. 78(6.12) 2.59(2.89) 12. 05(12.1 0) 7.11(7.26) 6. 84(7. 22) 6. 55(6. 70) 6. 34(6. 30) MB 6.83(7.12) 6. 82(7. 08) 5. 87(6. 20) 12. 56{12. 45) 5.26(5.26) 52. 01(52. 05) Fe(II)Zn(II)LBF4 • H20 Fe{II)Cu(II)LBF4 6. 38(6. 37) 13. 02(12. 80) 6. 28(6. 34) 53. 36(53. 52) Fe(II)Ni(II)LBF4 5. 16(5 .18) 6.16(6. 40) 12. 90(12. 85) 13. 04(12. 73) 53. 56(53. 50) Fe(II)Co(II)LBF4 5. 28(5. 20) 5. 52(5. 64) 11.11(11. 31) 5.16(5.15) 53. 73(53. 75) Fe(TI) Mn{II) LB F 4 5.02(4.89) 1 o. 99(1 o. 88) 2. 58(2. 72) MA 53. 42{53. 23) 47. 34(47. 31) Fe(III)LH3(BF4)3 4. 87(4. 70) 12.45(12.55) N 13.11(12. 80) 45. 80( 45. 47) Mn(III)LH 3(ClO4 ) 3 5. 56(5. 64) Found (Calcd.) (%) H 5.18(5.18) 52. 81(52. 46) c Mg(mLH3(BF4)2 • H20 Complex ~ w 0) Table IV. (continued) 12. 29(12.11) 2. 87(2. 89) 12. 32(12. 08) 13. 38(13. 34) 4. 96( 4. 88) 5. 44(5. 40) 50. 83(50. 47) 55. 63(55. 77) Mn(ll)Cd(II)LBF4 Mn(ll)Mg(II)LBF4 7. 35(7. 42) 6.16(6. 23) 12. 87(12. 71) 5.19(5 .14) 53. 28(53.17) Mn(II) Zn(II)LBF4 6. 68(6. 72) 13. 06(12. 81) 5.35(5.19) 53. 26(53. 57) 2. 58(2. 78) MB Mn(IT)Ni(II)LBF4 7. 07(7. 26) MA 5.34(5.53) 12. 88(12. 80) N 53. 45(53. 50) c Found (Calcd.) (%) H Cu(II)Mg(II)LBF4 • %H 20 Complex ~ C-" -3 38 appropriate metals. For the heteronuclear complexes, metal analyses generally are consistent with the expected formulations. It should be noted, however, that these data cannot be used to distinguish between pure heteronuclear complexes, MA(II)MB(II)L +, and samples which contain significant equimolar amounts of homonuclear complexes, MA(II)M A (II)L + and MB(II)MB(II)L +. For the purpose of determining the heteronuclear purity of these samples, the electrochemical methods described below have been used. • The 4. 2 K 57 Fe Mossbauer spectrum ~~~~~~~~~~~ of Fe(II)Fe(II)L + shows only one quadrupole-split doublet for which a least squares fit yields a quadrupole splitting of AEQ = 0. 9601(19) mm/sec and an isomer shift of 6 = 0. 4870(9) mm/sec vs. iron metal. Contamination of the sample by Fe(III) ions is not observed. The AEQ value is lower than those values typically observed for six-coordinate, high22 spin Fe(II) complexes (2. 0-3.0 mm/sec). Abnormally low AEQ values generally are indicative of either metal-ligand covalent bonding 8 23 or highly symmetric ligand environments. For Fe(II) ions in L - , the latter explanation appears to be more plausible. Electrochemistry. Cyclic voltammetry and differential pulse ~ voltammetry were used to characterize the electrochemical behavior at a platinum electrode of both mononuclear and binuclear complexes. Formal potentials (Ef) were measured and are reported with respect to 24 an internal reference standard, the ferricenium/ferrocene couple (Fe+ /Fe). For complexes with limited solubility, potentials could be measured with greater precision using differential pulse voltammetry instead of cyclic voltammetry. For this reason, formal potentials 39 cited in the text are the differential pulse voltammetric values (EP). It should be noted, however, that potentials obtained by one technique differ by 15 mV or less with respect to the other technique. The electrochemical results for all of the binuclear complexes reported here were obtained in DMF solutions. For the series of homonuclear complexes, M(II)M(IT)L + (M = Mn, Fe, Co, Cu, Zn, and Cd), only the Mn(II), Fe(IT), and Cu(II) complexes yielded cyclic voltammograms with quasi-reversible metal waves. Several overlapping, irreversible waves were observed for Co(II)Co(II)L +, and only ligand reduction waves at very negative potentials (<-2. 0 V) appeared in cyclic voltammograms of Zn(II)Zn(II)L +and Cd(II)Cd(II)L +. Of the three complexes exhibiting quasi-reversible metal waves, only for Fe(II) Fe(II)L + were two waves observable corresponding to sequential one-electron processes at the two interacting metal centers (Figure 4). These waves, ascribed to successive oxidations of + 2+ 3+ Fe(II)Fe(II)L to Fe(II)Fe(III)L and then to Fe(III)Fe(IIT)L , have measured formal potentials of -0.553 V and +0.133 V vs. Fe+ /Fe, respectively. For anodic scans of Mn(II)Mn(II)L +, the oxidation wave of one Mn(II) at +0.120 V is followed by an irreversible wave at a potential ..... 0. 5 V more positive. Since this second wave occurs at similar potentials in scans of Mn(II)M(II)L + (M = Ni, Zn, Cd, and Mg), it is ascribed to an oxidative ligand process. For cathodic scans of Cu(II)Cu(ll)L +, a reduction wave at -1.083 V attributed to the Cu(II)Cu(II)L +/Cu(I)Cu(II)L couple is succeeded at a potential"' 0. 85 V more negative by an irreversible wave. This second wave is absent in scans of heteronuclear Cu(II)M(II)L + complexes and probably -1.0 -0.8 -0.6 -0.2 E, \bits vs. Fe+/Fe -0.4 0.0 0.2 0.4 2 . Fe(II)Fe(II)L + wave, while the Fe(Ill)Fe(III)L3+/ Fe(III)Fe(II)L2+ couple appears at .... 0.1 V. ~ Cyclic voltammogram of Fe(IT)Fe(II)LBF4 • The wave at ,...-0.6V is the Fe(ITI)Fe(II)L +/ a ~.... ..L T ro #J.A ~ 0 41 corresponds to the reduction of Cu(I)Cu(II)L to Cu(I)Cu(I)L-. That all the quasi-reversible metal waves were one-electron processes was verified by coulometry (n values = 1. 0 ± 0. 05). Electrochemical methods were invaluable for establishing the purity of heteronuclear MA (II)MB(II)L +complexes, as has been 1 reported previously. In general, it is observed that the redox pctentials for electroactive metals in heteronuclear complexes differ measurably from those in the potentially contaminating, corresponding homonuclear complexes. For example, for Fe(II)Cu(II)L +, a single Fe(III)/Fe(In wave at -0.505 V and a single Cu(II)/Cu(I) wave at -1. 099 V are observed (Figure 5).. A lack of contamination by Fe(II)Fe(II)L +is indicated by the absence of a shoulder at -0.553 V. Similarly, although less conclusively since the potentials differ only slightly, purity with respect to Cu(II)Cu(II)L + contamination is established by the absence of a shoulder at -1. 083 V. In general, for heteronuclear complexes prepared by methods described in the Experimental Section, homonuclear contamination is estimated at less than 5% using this electrochemical testing method. The reduction potentials of the Mn(III) /Mn(II), Fe(III) /Fe(II), and Cu(II)/Cu(I) as a function of remote metal, MB(II), in MA (II)MB(II)L + complexes are reported in Tables V, VI, and VII, respe.c tively. A description of the characteristic electrochemical . behavior for each series follows: M(III)/Mn(II). As examined by cyclic voltammetry, the Mn(III/II) waves are quasi-reversible for a range of scan rates (20-500 mV /sec). A typical wave shape is shown for this couple in -1.2 -0.8 -0.6 E, Volts vs. Fc~Fc -1.0 -0.4 -0.2 Cyclic voltammogram of Fe(IT)Cu(II)LBF 4 • The wave at"' -1.1 Vis the Cu(II)/Cu(I) -1.4 ~ wave, while the Fe(III)/Fe(II) couple appears at"' -0.55 V. ~ u ~ '- ~ c - 7f 'S p.A ~ t-:) 43 Table V. Reduction Potentials for the Mn(III)/Mn(II) Copple as a ~ Function of Remote Metal MB(II) (measured in DMF vs. Fe+/Fe). Differential Pulse Voltammetry Cyclic Voltammetry b Complex Ef(V)a Ep - Ep (mV) a c Mn(II)Mn(II)L + 0.120 71 0.114 Mn(II)Ni(II)L + 0.363 god 0.370 Mn(II)Cu(II)L + -0.062 81 -0.053 Mn(II)Zn(II)L + 0. 018 87 0.028 Mn(II)Cd(II)L + 0.036 70 0.048 Mn(II)Mg(II)L + 0.281 86 0.285 Ep(V)c aFormal potentials are reported using Ef = (Ep - Ep )/2 and were a c measured at a scan rate of 200 mV /sec. bAnodic peak to cathodic peak separations are reported for the minimum observed separations in the scan-rate range of 20-500 mV /sec. cPeak potentials were measured at a scan rate of 1 mV /sec. ~his value can only be estimated because of the low solubility of the complex and the proximity of the metal wave to a ligand wave. 44 Table VI. Reduction Potentials for the Fe(ITI) /Fe(ll) Couple as a ~ Function of Remote Metal MB(ll) (measured in DMF vs. Fe+ /Fe). Cyclic Voltammetrya Differential Pulse Voltammetry Complex Ef(V) Ep - Ep (mV) a c Ep(V)b Fe(ll)Mn(II)L + -0.506 68 -0. 511 Fe(II)Fe(ll)L + -0.557 68 -0.553 Fe(ll)Co(II)L + -0.548 68 -0.557 Fe(ll)Ni(II)L + -0.357 71 -0.365 Fe(II)Cu(ll)L + -0.499 69 -0.505 Fe(II) Zn(II) L + -0.586 70 -0.590 Fe(II)Cd(II)L + -0.555 75 -0.564 Fe(II)Mg(ll)L + -0.397 68 -0. 401 aSee footnotes (a) and (b) for Table V. bPeak potentials were measured at 0. 5 mV /sec. 45 ~ Reduction Potentials for the Cu(II) /Cu(I) Couple as a Function of Remote Metal MB (measured in DMF vs. Fe+ /Fe). Cyclic Voltammetrya Differential Pulse Voltammetry Complex Ef(V) Cu(II)Mn(II)L + -1.070 83 -1.084 Cu(II) Fe(II)L + -1.089 73 -1.099 Cu(II)Co(II)L + -1.001 81 -1.015 Cu(II)Ni(II)L + -0.938 79 -0.942 Cu(II)Cu(II)L + -1.077 101 -1.083 Cu(II)Zn(ll)L + -1.086 99 -1.092 Cu(II)Cd(II)L + -1. 186 85 -1.188 Cu(II)Mg(II)L + -0.997 81 -1.001 Ep -Ep (mV) a c a See footnotes (a) and (b) for Table V. bPeak potentials were measured at 0. 5 mV /sec. 46 Figure 6. Peak potential separations (Epa- Epc) show only a slight scan-rate dependence. For the series, the average minimum Epa- Epc is 80 mv 25 in comparison with the Nernstian vahle of 59 mV. 26 Formal potentials vary over a range of "'425 mV. Mn(III)/Mn(II) potentials have not been observed and therefore are not reported for Mn(II)Fe(II)L + and Mn(II)Co(II)L +. For these complexes only the more negative potential M(III)/M(II) waves of iron and cobalt (irreversible) have been observed. Fe(III)/Fe(II). Scan-rate independent Fe(III) /Fe(II) waves with an average minimum Epa- EPc of 71 mV have been observed by cyclic voltammetry. These Fe(III)/Fe(II) waves, exhibiting nearly reversible behavior, have potentials which vary over a range of,...., 225 mV. A typical cyclic voltammogram for this redox couple is shown in Figure 7. Cu(II)/Cu(I). In comparison with Mn(III)/Mn(II) and Fe(III)/Fe(II) waves, Cu(II)/Cu(I) waves are the least reversible with a minimum Epa- EPc of 85 mV. A typical wave shape for the Cu(II)/Cu(I) couple is shown in Figure 8. For Cu(II)M(II)L +complexes (M = Mn, Fe, and Co), waves are observable for both metal centers. Characteristics of the Mn(III)/Mn(II) and Fe(III)/Fe(II) waves are included in the previously described series; the Co(III)/Co(II) wave is quasi-reversible with Ef = -0.246 V and Epa- EPc =90 mV. The Cu(II)/Cu(I) potentials vary over a range of "'250 mV. Cyclic voltammograms obtained for the series of mononuclear 2+ . complexes M(II)LH 3 (M = Mn, Fe, Co, N1, Cu, Zn, Cd, and Mg), featured broad ligand waves for which both anodic and cathodic peak currents were not observed. In acetonitrile solutions, non-irreversible -0.2 -ll -0.1 0.1 E, \k>lts vs. Fe+ /Fe 0.0 02 0.3 0.4 ~ Typical cyclic voltammogram for Mn(III)/Mn(II) couple in Mn(II)M(II)L +complexes. u i T f IJ.A ~ -l -0.8 -0.7 -05 -0.4 E, Vo Its vs. Fe+/Fe -0.6 -0.3 -0.2 ~· Typical cyclic voltammogram for Fe(III)/Fe(II) couple in Fe(II)M(II)L +complexes. a~ c - 'f '5 f.!.A ~ ~ 00 -1.4 -1.3 -1.2 ~ -to E, Volts vs. Fc+/Fc -1.1 -0.9 -0.8 -0.7 ~ Typical cyclic voltammogram for Cu(II)/Cu(I) couple in Cu(II)M(II)L +complexes. d .... -i 'r 10 JJA ~ co 50 metal waves could be observed only for the Mn(II), Fe(II), and Co(II) complexes, corresponding to a M(ll)/M(II) couple (Table VITI). Of these complexes, a quasi-reversible metal wave with a minimum Ep - Ep less than 130 mV was observed only for Fe(II)LH~+. Despite a c substantial broadness of the Mn(Ill)/Mn(II) and Co(Ill)/Co(II) waves, coulometric experiments indicated that the waves represent chemically reversible, one-electron processes (!!values = 1. 0 ± 0. 05 ). 51 ~Electrochemical Data for the M(III)/M(ll) Couple of M(II)LH!+ Complexesa Cyclic Voltammetryb Differential Pulse Voltammetry Complex Ef(V) Ep - Ep (mV) a c EP(V)c Coulometricd n value Mn(II)LH:+ 0.309 130 0.301 1. 05 Fe(II)LH:+e -0.283 75 -0.287 0.96 Co(II)LH~+ 0.085f >400 NDg 1. 02 a All measurements were taken with acetonitrile as the solvent. Formal potentials are reported vs. Fe+ /Fe. bsee footnotes (a) and (b) for Table V. cPeak potentials were measured at a scan rate of 1 mV /sec. d.rhese values were determined by controlled potential electrolysis at potentials 20-30 mV positive of Epa· eThe electrochemical results reported for Fe(II)LH:+ are an average of . 2+ 3+ the data obtamed for bath Fe(II)LH 3 and Fe(III)LH 3 • f.rhis value only can be estimated since (Epa- EPc) is > 400 mV. gND = not determined. 52 DISCUSSION ~ Clathrochelates and Tern late Reactions. Two general types of clathrochelates have been reported prior to the preparation and characterization of the mononuclear and binuclear clathrochelates described in this report. The first type is flexible mononuclear and binuclear complexes which can be prepared directly by adding metal ions to solutions containing clathrochelating polyether /polyamine ligands. These ligands, referred to as both cryptates and cryptands, have been prepared and their coordination chemistry studied extensively by 15 Cryptates and cryptands are synJ. M. Lehn and co-workers. thesized without the use of metal-ion template reactions using, instead, multistep organic syntheses which include high-dilution reactions. For most complexes of this type, the metal ions are highly labile and metal-exchange reactions occur facilely. The second general type of clathrochelates characteristically have transition metal ions encapsulated by rigid ligand structures. For these complexes, the metal ions cannot be displaced by other metal ions or removed with the ligand remaining intact. Two general methods have been used to prepare these clathrochelates. In the more common method, a six-coordinate complex with purposefully functionalized ligands is treated with reagents which, through ring closure reactions with ligand functionalities, interconnect the ligands. Com27 plexes prepared by this approach have been reported by Rose, . 28 29 30 . Holm, Sargeson, and Goedken. The second method mvolves base-induced rearrangement of a metal-coordinated tetradentate ligand 53 which has a non-coordinated appendage containing additional donor atoms. Busch and co-workers have recently reported complexes prepared by this method. 31 The MLH 3 n+ and M(II)M(II)L +complexes described in this report have features in common with both types of clathrochelates described above including ligand flexibility and method of preparation. Comparing the former feature is relatively subjective owing to a lack of data for the previously reported complexes. Nevertheless, there is evidence which suggests that L s- 2+ and LH 3 in M(II)M(II)L +and M(II)LH 3 com- plexes, respectively, have some degree of rigidity comparable to the 3 ligand flexibility of known rigid clathrochelates. In the case of L -, 1 the H NMR spectrum of Cd(II)Cd(II)L +at ambient temperatures exhibits resonances and couplings characteristic of a ligand structure with inflexible -CH 2 CH2 - subunits. From similar studies of M(II)LH~+ complexes, the rigid nature of LH3 is suggested by evidence for intramolecular hydrogen bonding. It is noteworthy, however, that the ligands in M(II)M(II)L +and M(II)LH~+ complexes are not so inflexible that they cannot adopt slightly different geometries required for different metals. Indications for this adjustable rigidity are slight differences in the infrared spectra of these complexes (Table I) and by the anomalous basicity of Cu(IT)M(IT)L +complexes (M(II) = Cu and Zn). In contrast to the general ligand inflexibility of M(II)LH:+ and M(ll)M(II)L + complexes, the evidence for NaLH3 N0 3 strongly suggests ligand flexibility analogous to the cryptate/cryptand complexes. For these complexes, ligand flexibility is characterized by metal-ligand dissociative and/or internal site-site exchange processes occurring on the NMR time sc~le. 54 Preparative methods used for both general classes of clathrochelates discussed above are utilized in the syntheses of MLH~+ and M(II)M(II)L +complexes. Analogous to methods used for cryptate/ 2+ 3+ cryptand complexes, MLH 3 ' complexes can be prepared by the direct substitution of an appropriate metal ion for a complexed alkalication (eq 5). Similarly, M(II)M(II)L +complexes can be synthesized by direct addition of a metal ion to a mononuclear clathrochelate with an available chelating site (eq 6). Analogous to the preparation of rigid clathrochelates, MLH~+ and M(II)M(II)L + complexes also can be prepared by reacting initially coordinated ligand constituents with other ligand components (eq 1, 3, and 4). In these reactions, tren is the ligand constituent which, in a quadridentate manner, 32 initially coordinates to a metal ion. As is invariably the case for known syntheses of rigid clathrochelates, subsequent clathrochelate-forming reactions . ·t·1a I comp Iex are me t a I-1on . spec1·r·1c. 27 - 31 A s a resu It , of th e m1 MLH~+ and MA (II)MB(II)L + complexes have been prepared by rigid clathrochelate methods only for M = Na +, Mn(II), Fe(II), Co(II), and Cd(II) and MA (II) = MB(II) = Cd(II) and Mn(II). The preparative methods for rigid clathrochelates described above can be included in a more general classification, metal-template reactions. Such reactions have been used extensively for preparing 33 macrocyclic complexes as well as clathrochelates. Although this classification is not normally subdivided, metal-template reactions are basically of two types: simple and complex. Simple metal-template reactions are characterized by ligand constituents which, if they are even capable of coordination, do not form initial complexes structurally 55 resembling the final product. Generally, E;uch ligand constituents contribute at most three ligating atoms per metal to the final product. The preparation of phthalocyanine complexes from 1, 2-dicyanobenzene and metal salts is an example of this simple metal-template reaction. 34 In contrast, complex metal-template reactions are characterized by one or more ligand constituents forming initial complexes which are structurally similar to the final product. In all cases, the coordinating ligand constituents are multidentate. The metal-ligand product results from ring closure reactions of non-coordinated ligand constituents with the coordinated ones. All of the preparations for previously reported rigid clathrochelates are examples of complex metal-template reactions. In this regard, the metal-template reactions detailed here are exceptional in that the clathrochelates are formed by simple metal2+ template reactions. For example, in formation of the M(II)LH3 complexes, the tren is initially coordinated to the metal in a quadridentate manner (through its four nitrogen atoms); this ligand then undergoes a structural change in coordination geometry and is finally incorporated as a tridentate species. An even more remarkable example of a simple metal-template reaction is that which produces the Mn(II) and Cd(IT) homonuclear complexes, M(II)M(ll)L +. In these cases, five ligand constituents are brought together around two metal ions to form a single, -bicyclic ligand (>50% yield). The only literature example of a simple metal-template reaction which approaches the M(II)M(ll)L + system in complexity is the formation of superphthalocyanine complexes from uo~+ and 5 equiv of 1, 2-dicyanobenzene (<20% yield). 35 • 56 ~ Of the homobinuclear complexes (M(II)M(ll)L l with electroactive metal ions, only Fe(II)Fe(ll)L +shows two reversible, sequential redox processes, which in this case are oxidations. These redox proces~es are separated by an amount 6E = 0. 686V, a quantity related thermodynamically to the comproportionation equilibrium as shown in eq. 7-10. Using eq 10, a Kcom = (4. 0 ± 0.1) x 10 [Ox-Ox] + e [Ox-Red] + e 11 can be f El - - [Ox-Red] (7) Ef2 - - [Red-Red] [Ox-Ox] + [Red-Red] Kcom (8) 2 [Ox-Red] (9) E! - E~ = 6E = (RT /nF) ln Kcom (1 0) calculated for Fe(II)Fe(II)L +. Although both Cu(II)/Cu(I) reductions are not electrochemically reversible, a lower limit for a Kcom of 13 Cu(II)Cu(II)L +can be estimated as approximately 10 • In comparison, the corresponding constant for Fe(II) and Cu(II) homobinuclear com4 6 1 36 plexes of type !. are 2. 5 x 10 and 4. 0 x 10 , respectively • ' For comparison with a system which is lmown to exhibit significant elec. s+/s+/1+ tronic delocalization, (NH 3 ) 5Ru-pyrazme-Ru(NH 3 ) 5 has a reported 6 37 Kcom = 4 x 10. Those interactions between metal centers which contribute significantly to a large 6E, and thus a large Kcom' are difficult to quantify, since Kcom may reflect interactions which destabilize the fOx-Ox] and [Red-Red] species and/or stabilize the [Ox-Red] species. Several 57 factors which affect the relative energetics of these three species have 1 been cited previously and are recounted briefly here: 1) structural variations associated with oxidation state changes such as altered metal-ligand bonding distances and geometries are known to affect Kcom values significantly. 38 2) Electrostatic repulsion between proximate metal centers both through-orbital and through-space result in the addition of a second electron being energetically more difficult than the first process. 3) Electronic delocalization enhances the stability of the mixed-valent state with respect to the is ova lent states. 4) Magnetic superexchange interactions are of potential significance for those oxidation states with coupled paramagnetic centers; however, for weakly coupled systems the net effect on Kcom is usually negligible. 5) A statistical factor, which for totally non-interacting metal centers yields a Kcom = 4 (~E = 36 mV), contributes negligibly to Kcom for systems for which the other factors affect Kcom substantially . 1 In a previous study, type ! complexes with MA = Cu(II) and MB = Mn(II), Fe(II), Co(II), Ni(II), Cu(II), and Zn(ll), were examined electrochemically and magnetically in an attempt to quantify some of 1 the factors contributing to ~E for the Cu(II) homonuclear complex. The Cu(II) /Cu(I) reduction pctentials were measured as a function of MB and were found to be relatively invariant for MB ¢ Cu(II). In contrast, E; (which is the analogous pctential for MB = Cu(II)) is positive of the average heteronuclear value by 137 ± 14 mV, after minimal corrections were applied to account for magnetic superexchange. It was hypothesized that the average Cu(II) /Cu(I) reduction potential for the heteronuclear complexes was an accurate measure of E; due only 58 to electrostatic repulsion. Based on this hypothesis, the 137 mV = 3. 2 ± 0. 3 kcal/mole difference between E! (observed) and E~ (electro- static repulsion only) was ascribed to electronic delocalization which stabilizes the mixed-valent Cu(II)-Cu(I) complex relative to the isovalent Cu(II) -Cu(II) complex. For E~ (electrostatic repulsion only) to be accurately represented by the average Cu(II)/Cu(I) reduction potential of the heteronuclear complexes, two conditions must be satisfied. First, the mixed-valent heteronuclear complexes must be negligibly delocalized. Since these complexes have relatively inaccessible MB(II)/MB(I) reduction potentials, this condition is met. Second, since electrostatic repulsion depends on the structural relationship between MB and MA, as MA is reduced, structural differences between the homonuclear and heteronuclear complexes must be negligible or, at least, compensative. With the Cu(II)/Cu(I) reduction potentials being relatively insensitive towards MB for MB ¢ Cu(II), it is argued that structural differences in these complexes due to the nature of MB are minimal. Assuming that this argument can be extended to MB = Cu(II), the second condition is satisfied. It should be noted, however, that the axial ligand geometry of these complexes in solution is not rigorously controlled. Thus, it is conceivable that the homonuclear complex might be structurally anomalous with respect to the heteronuclear series. If this is the case, a reliable estimate cannot be obtained for the stabilization due to electronic delocalization. The binuclear clathrochelate complexes, MA (II)MB(II)L +, were examined to further evaluate the factors which contribute significantly 59 to .6-E, especially electrostatic repulsion and electronic delocalization. There were several reasons for studying complexes with this clathrochelating ligand, L s- • First, such a ligand has the potential for rigorously controlling metal-ligand geometry and minimizing structural variations. Second, solvent and anion interactions with the cationic metal centers most likely would be hindered. Third, with three bridging atoms the magnitudes of the interactions between MA and MB might be greater than those of complexes with fewer bridging atoms. That some form of interaction between metals is enhanced in these complexes is strongly suggested by the exceptionally large .6.E values for Fe(II) Fe(II)L +and Cu(II)Cu(II)L +. To evaluate the principal contributions to .6-E for MA (II)MB(II)L + complexes, the same approach as in the study of type 1 complexes was "' used. For three series of complexes, the potential of a MA redox couple was determined as a function of a variable MB. Specifically, the Mn(III) /Mn(II), Fe(III)/Fe(II) and Cu(II)/Cu(I) reduction potentials were measured, if observable, with respect toMB = Mn(II), Fe(II), Co(II), Ni(II), Cu(II), Zn(II), Cd(II), and Mg(II). The potentials were not adjusted to account for magnetic superexchange, since such interactions for these complexes are very weak. In contrast to the results observed for type ! Cu(II) heteronuclear complexes, the measured MA redox pctentials of MA (ll)MB(II)L +complexes are not invariant with respect toMB for MB ¢ MA. The Fe(III)/Fe(II) and Cu(II)/Cu(I) reduction potentials, which have been determined for seven heteronuclear complexes, span ranges of 225 mV and 250 mV, respectively (Figure 9). The potential range of the less extensive 60 (a) Co Zn Cd Cu Mg Fe Mn II I I -0~5 I -0.6 (b) Zn -d.4 Fe Mn Cd ·co Cu Ni Mg -1.0 -1.1 -1.2 Ni (c) Mn Cu -0.1 I I Tr II 0.0 +0.1 Mg Ni I +0.2 II I I +0.3 +0.4 Volts vs. Fe+/Fe ~· MA Redox potentials in MA (II)MB (ll)L+ complexes as a function of MB for (a) Fe(ill)/(II) couple, (b) Cu(ll)/(I) couple and (c) Mn(III)/(n) couple. 61 five-membered Mn(III)/Mn(TI) series is 425 mV. In comparing the three series, the relative ordering of potentials from negative to positive appears to depend randomly on the nature of MB, except for MB = Ni(II) and Mg(II). For heteronuclear complexes containing these metals, the two most positive MA redox potentials were consistently observed. If these systematically positive potentials are excluded from the Fe(nn/Fe(II) and Cu(II)/Cu(I) series, the range of pctentials which depend non-systematically on the nature of MB are reduced to 85 mV and 175 mV, respectively. In spite of the considerable variance in the MA redox potentials forMA (II)MB(II)L +complexes (MA ¢ MB) described above, the data have been evaluated by the approach used in the study of type 1 com"' plexes. Specifically, the MA potentials for a given series of MA (II)MB(II)L + heteronuclear complexes are averaged to give an estimated value of E~ or E! due to electrostatic repulsion only. These estimated values are then compared with the corresponding observed E! or E! values. Using this approach on the restricted series of Fe(II)MB(II)L + complexes (MB Mg(II) or Ni(In) yields E! (electro¢ static repulsion only) = -0. 545 ± 0. 036 V for comparison with E! (observed) = -0. 553 V. Similarly for the analogous series of Cu(ll)MB(IT)L +complexes, a value for E~ (electrostatic repulsion only)= -1. 09~ ± 0. 062 V is calculated for comparison with E! (observed) = -1.083 V. If the MB = Mg(II) and Ni(II) values are not excluded, E! (electrostatic repulsion only) for Fe(II)MB(II)L+is -0. 499± 0. 085 V, while E~ (electrostatic repulsion only) for Cu(II)MB(II)L + is -1. 060± 0. 081 V. For the complete series of Mn(II)MB(II)L +complexes, 62 a value for E! (electrostatic repulsion only) = +0.136± 0.182 Vis calculated for comparison with E! (observed) = +0.114 V. Two differences between these results and those for type ,..... 1 complexes (MA = Cu(II); MB = Mn(II), Fe(II), Co(II), Ni(TI), Cu(TI), and Zn(II)) are apparent. First, the uncertainties associated with the estimated values of E! or E! (electrostatic repulsion only) are con- siderable, since the MA potentials for these MA (TI)MB(TI)L +complexes vary substantially. Second, there is no significant difference between E~ or E~ (electrostatic repulsion only) and E! or E! (observed) within the determined error limits. Provided this approach is valid, the apparent conclusion for these M A (InMB(InL + complexes is that there is negligible stabilization, within experimental error, for the mixedvalent homonuclear complexes relative to the analogous mixed-valent heteronuclear complexes. Specifically, for the restricted series of 2 Fe(II)MB(II)L +complexes, the stabilization of Fe(ITI)Fe(Il)L + relative 2 to Fe(III)MB(TI)L + (MB -.r: Fe(II)) is 0. 2± 0. 8 kcal/mole. Similarly, for the restricted series of Cu(II)MB(II)L + complexes, the stabilization of Cu(I)Cu(II)L relative to Cu(I)MB(II)L (MB -.r: Cu(II) is 0.3 ± 1. 4 kcal/mole. Using the complete data sets for Fe(II)MB(TI)L +and Cu(II)MB(II)L + complexes gives stabilization values of 1.2± 2. 0 kcal/mole and -0. 5± l. 9 kcal/mole, respectively. For the less extensive Mn(II)MB(II)L + series, the stabilization of Mn(III)Mn(II)L + relative to Mn(III)MB(II)L + (MB "# Mn(II)) is 0. 5 ± 4. 2 kcal/mole. Before discussing factors which might account for the lack of stabilization for mixed-valent homonuclear complexes, the validity in 63 using the above approach for estimating the stabilization energies of these complexes should be considered. The approach is justifiable only if the average of the MA redox potentials for a series of similar heteronuclear complexes provides a reasonable estimate forE; orE~ (electrostatic repulsion only). The reasonableness of the estimate depends on 1) the mixed-valent heteronuclear complexes being negligibly electronically delocalized and 2) structural differences between the homonuclear and heteronuclear complexes being minimal. Condition 1 is satisfied since MA and MB redox potentials differ by at least 0. 4 V in all of the MA (II)MB(II)L + complexes examined. As has been discussed previously for type 1 complexes, it is " difficult to establish conclusively that the second condition is met. Based on structural studies of other clathrochelates in which only slight metal-dependent structural changes are observed, 13 it seems reasonable that homonuclear and heteronuclear MA (II)MB(II)L +complexes also will be structurally similar. Additionally, the near indistinguishability of infrared spectra for a given series of MA (II)MB(II)L +complexes is consistent with minimal structural differences. In apparent contradiction with these arguments, however, is the variability of redox potentials for a given series which suggests that structural differences in MA (II)MB(II)L +complexes might not be negligible. Although a degree of structural uncertainty has been introduced, it is still likely, based on the previous arguments, that a homonuclear complex will structurally resemble analogous heteronuclear complexes. Owing to 1) the general structural similarities between M(II)M(II)L + (M = Fe and Cu) and type 1 complexes and 2) the observation that large " 64 Kcom values often reflect a degree of electronic delocalization, it seemed reasonable that stabilization of mixed-valent states due to electronic delocalization for M(II)M(II)L +complexes might be of a measurable magnitude. Thus, it was surprising that the electrochemical results for both Fe(II)Fe(II)L +and Cu(II)Cu(II)L + complexes showed, within experimental error, no apparent stabilization of the mixed-valent states due to electronic delocalization. Plausible explanations for the lack of measurable stabilization include 1) the absence of metal-ligand orbital overlap of the appropriate magnitude or symmetry, 2) that even slight structural differences between 1 complexes might result in very M(II)M(II)L + complexes and type ,.... different metal-metal communication, and 3) that these stabilization energies are inherently smaller than errors associated with the electrochemical method employed here. Independent magnetic studies yielded results consistent with the first explanation. Results from other studies of electronic delocalization in binuclear complexes show very small, but measurable, stabilization energies consistent with the third explanation given above. By a method described by Taube et al., 39 which avoids the necessity of preparing structurally similar homonuclear and heteronuclear complexes, electronic de localization energies of Ru(TI, ill) dime rs can be estimated. For [((NH 3 ) 5 Ru) 2 -(4, 4' -bipyridine)] and (NH 3 ) 5Ru(pyrazine)Ru(edta)+, these energies are 0. 05 39 and 0. 8 kcal/mole, 40 respectively. Bot.h of these values are less than the uncertainty in the approach used to study M(II)M(II)L + complexes and the 0. 05 kcal/mole values is less than the corresponding uncertainty for type 1 complexes. Unfortunately, "' 65 neitherM(II)M(II)L + nor type! complexes are conducive to study by Taube's approach. Since many complexes, including the M(ll)M(II)L + series, cannot be studied by Taube's method, it is worth comparing and contrasting the results for known systems amenable to an electrochemical determination of delocalization energies. The complexes, (NH 3 ) 5Ru(4, 4'6 bipyridine)Ru(NH 3 ):+ and (NH 3 ) 5Ru(4, 4' -bipyridine)Rh(NH 3 ) +, have 41 Ru(III)/Ru(ll) pctentials of +0. 79 V and +0. 74 V, respectively. In this case, the 50 mV (1. 2 kcal/mole) potential difference has been ascribed to de localization stabilization of the Ru(III) -Ru(II) mixed-valent complex. 1 A similar system, (NH ) Ru(pyrazine)Ru(edta)+ and 3 5 (NH 3 ) 5Ru(pyrazine)Rh(edta) +, has Ru(III)/Ru(II) potentials of -0.56 V 40 and -0. 61 V, respectively. This result is counterintuitive since it suggests that the Ru(II)-Ru(III) mixed-valent complex is destabilized by 50 mV (1. 2 kcal/mole) with respect to the Ru(II)-Rh(III) mixedvalent complex. This contrasting behavior is indicative of deficiencies in this electrochemical method for estimating a de localization energy, especially when a series of structurally similar heteronuclear complexes are net examined. An interesting recent report details the electrochemistry of type~ complexes with MB = Fe(II), Co(II), NI(II), 42 Cu(II), Zn(ll), Mg(II), Ca(II), and Ba(II). These studies showed that the Cu(II)/Cu(I) reduction potential is invariant with respect toMB for MB = Fe(II), Co(II), Ni(ll), Zn(II), and Mg(II), analogous to the behavior observed for type ,.., 1 complexes. The surprising result of this study is that for MB = Cu(II) the Cu(II)/Cu(I) potential is 480 mV 66 0 N Ph 0 0 "'cu I ""- MI8 I """ 0 I " 0 I N 0 3 "" positive of the average heteronuclear value. This 480 mV difference translates into an incredible 11 kcal/mole stabilization energy for the mixed-valent state, when, in fact, a negligible stabilization is expected for such an unsymmetrical complex. In light of these above results, a reassessment of the validity of the electrochemical method for determining delocalization energies is warranted, especially in the cases of type 1 and 3 complexes. In the "" "" former case, there is additional experimental evidence which appears to contradict the interpretation of the measured stabilization as due to electronic de localization. Specifically, CO binding studies for type 1 "" complexes (MA = Cu(I) and MB = Mn(II), Fe(II), Co(II), Ni(II), Cu(II), and Zn(II)) indicate that CO binds most strongly to the mixed-valent Cu(II)-Cu(I) complex, and does so in a nonsymmetrical, terminal 67 1 2 fashion. ' Bonding of CO to Cu(I) in the homonuclear system in an unsymmetrical fashion would certainly disrupt any delocalization and, thus, a preference for CO binding to Cu(I) in the homonuclear species should be disfavored with respect to binding to Cu(I) in the analogous heteronuclear species. In the case of type ~complexes, single crystal X-ray structural results for similar complexes establish that the homonuclear complex (MB = Cu(II)) is structurally different from heteronuclear analogues 43 44 These results are relevant to the elec(MB = Co(II) and Mg(II)). ' trochemical studies assuming that the solid-state structures are good representations of the solution structures. In the case of the homonuclear complex, the mean planes of the two four-coordinate sites 43 intersect with a dihedral angle of 191. a•. In contrast, the corresponding mean planes of the heteronuclear complexes intersect with a 44 dihedral angle of 180 •. Although the difference is only "' 12 •, this increased twisting contributes to the shift of the Cu(II)/(I) potential (for the metal ion in N2 0 2 site) by 480 mV. It is noteworthy that arguments used in the electrochemical study of tYPe 1 complexes, based "' on the invariance of potentials for MB ¢ Cu(II), would have suggested that structural differences between the homonuclear and heteronuclear complexes are negligible. It seems, therefore, impossible to predict whether there are important structural changes for these types of complexes in the absence of X-ray studies. In light of the above considerations, it might be tempting to question the fundamental premise on which the electrochemical method for determining electronic delocalization is based, namely that 68 structural changes among a series of homonuclear and heteronuclear complexes is minimal. Certainly it appears treacherous to make any assumptions of isostructuralism for a series of complexes of types ...... 1 and ...... 3 (see above). However, since for M(II)M(II)L +complexes the same general results are observed for three different series (Fe, Cu, Mn), it is unlikely that the homonuclear species are uniformly structurally different with respect to their heteronuclear analogues. Thus, the application of the electrochemical method for these systems seems probably more valid than for the previously examined systems . 69 CONCLUSIONS ~ In an effort to study metal-metal commWlication in a clathro- chelating environment, homonuclear and heteronuclear MA (II)MB(II)L + complexes of the ligand Ls-, a deprctonated Schiff-base compound derived from 3 equiv of 2-hydroxy-5-methylisophthala.ldehyde and 2 equiv of 2, 2', 2" -triaminotriethylamine (tren), were synthesized and characterized. Two synthetic routes to these complexes were used. In the first method, a metal-ion specific template reaction, Mn(II)Mn(II)L +and Cd(II)Cd(II)L +were prepared directly from the reaction of 2 equiv of tren with 3 equiv of 2-hydroxy-5 -methylisophthalaldehyde in the presence of Mn(TI) or Cd(TI) salts. In the second, more general method, a mononuclear complex of the non-deprctonated ligand, NaLH 3 N03 , was prepared by the condensation reaction of ~ (a Schiff-base compound derived from 1 equiv of tren and 3 equiv of 2-hydroxy-5-methylisophthala.ldehyde) with 1 equiv of both tren and 2+ NaN03 • By using metathesis reactions, mononuclear M(II)LH 3 complexes (M = Mn, Fe, Co, Ni, Cu, Zn, Cd, and Mg) were prepared from NaLH 3 N03 • Insertions of the same type of metal ion into M(II)LH~+ complexes, in the presence of base, give the homonuclear M(II)M(II)L + products (M = Mn, Fe, Co, Cu, Zn, and Cd). By this method, the following heteronuclear complexes also have been prepared: Mn(II)M(II)L +, Fe(II)M(II)L +, and Cu(II)M(II)L +with M = Mn, Fe, Co, Ni, Cu, Zn, Cd, and Mg. All homonuclear and heteronuclear MA (II)MB(II)L + complexes were examined by electrochemical methods (cyclic voltammetry and 70 differential pulse voltammetry). For the three series, the Mn(III)/Mn(II), Fe(III)/Fe(II), and Cu(II)/Cu(I) couples were determined as a function of adjacent metal ion. A comparison of the electrochemical results for these homonuclear and heteronuclear complexes shows that, within experimental error, there is no stabilization of the mixed-valent 2+ 2+ Mn(III)Mn(II)L , Fe(III) Fe(II)L , or Cu(I)Cu(II)L complexes with respect to mixed-valent heteronuclear analogues. These results are compared and contrasted with those obtained for other related binuclear systems. 71 EXPERIMENTAL SECTION General Considerations. Unless noted otherwise, all solvents and reagents were purchased from commercial sources and were used without further purification. Tetrabutylammonium tetrafluoroborate, TBABF 4 (Southwestern Analytical Chemicals), was dried exhaustively in vacuo (25 • C) before use. Reagent grade N, N-dimethylformamide, DMF, was dried over Mg004 and then doubly distilled under reduced pressure from activated 4-A. molecular sieves. 2-Hydroxy-5-methylisophthalaldehyde was prepared by a published modification 1 of the literature method. 45 It was purified by flash chromatography 46 on silica gel; the compound was loaded on and eluted from the column with toluene. The dialdehyde was isolated by precipitation with petroleum ether (30-60• C) from a concentrated solution of the eluate. 2, 2', 2" -Triaminotriethylamine trihydrochloride, tren. 3HC1, prepared by the method of Kimura et al. ,47 was neutralized and distilled according to the procedure of Glerup et al. 48 to yield the free tetraamine, tren. Preparation of air-sensitive compounds and the electrochemical experiments were carried out under helium in a Vacuum Atmospheres Dri-Lab inert atmosphere chamber. Ph sical Measurements. 1 H NMR spectra were recorded with Varian EM390, JEOL FX90Q, or Bruker WM500 spectrometers. 13 C NMR spectra were obtained on a JEOL FX90Q spectrometer. Infrared spectra were recorded on a Beckman 4240 spectrometer. Samples were examined as KBr pellets. Magnetic susceptibility measurements were obtained at ambient temperature using a Cahn Instruments Faraday balance. The 72 instrument was calibrated using HgCo(SCN) 4 and diamagnetic corrections were made using Pascal's constants. 49 Mossbauer data were collected on a previously described instrument. 50 Elemental analyses were determined by Galbraith Laboratories, Inc.; the Spang Microanalytical Laboratory; and the California Institute of Technology's analytical facility. ~· Both a Princeton Applied Research (PAR) Model 174A polarographic analyzer and a PAR Model 173 potentiostat galvanostat coupled with a ramp generator of our own design were used for cyclic voltammetry. The PAR Model 174A was also utilized for differential pulse voltammetry. Results were displayed on either a Hewlett-Packard 7004B X-Y recorder or on a storage oscilloscope. Coulometric measurements were obtained by controlled potential electrolyses using the PAR Model 173 in conjunction with a PAR Model 179 digital coulometer. For all electrochemical measurements the supporting electrolyte was 0.1 M TBABF4 and a Ag+ /Ag reference electrode was used. The reference electrode consisted of a silver wire immersed in an acetonitrile solution containing AgN03 (0. 01 M) and tetrabutylammonium perchlorate (0.1 M). The Ag+ solution and silver wire were contained in an 8 mm glass tube fitted at one end with a fine porosity sintered glass frit. For the auxilliary electrode, either a coiled platinum wire or a compacted platinum gauze was used. Cyclic and differential pulse voltammetry were carried out in a single compartment cell containing approximately 5 mL of solution. A platinum button electrode was used as a working electrode. 73 Controlled potential electrolyses were carried out in a three compartment H cell. The cell consisted of two 25 mL outer compartments connected to a 5 mL inner compartment by medium porosity sintered glass frits. A platinum gauze working electrode and the Ag+ /Ag reference electrode were situated in the outer compartment containing the electroactive solution. The auxilliary electrode was isolated in the other outer compartment. By a method described elsewhere, 24 reduction potentials for the complexes were measured with respect to an internal standard which for most cases was ferrocene. When the ferricP.nium/ferrocene (Fe+ /Fe) wave overlapped with the wave under investigation, cobalticenium hexafluorophosphate was used; the cobalticenium / cobaltocene wave appears at -1.336 V vs. Fe+ /Fe. All potentials are reported versus the Fe+ / Fe couple. Cd(II)Cd(II)LBF 4 , Template Method. A warm solution of 133 mg (0. 81 mmol) of 2-hydroxy-5-methylisophthalaldehyde and 8 drops of triethylamine (TEA) in 25 mL of ethanol was added dropwise to a warm, stirred solution of 137 mg (0. 51 mmol) of tren in 25 mL of ethanol. The resulting yellow solution was heated at reflux for 20 min and then cooled to ambient temperature. A solution of 170 mg (0. 51 mmol) of TBABF4 in 5 mL of ethanol was added resulting in immediate precipitation .of a lemon-yellow solid. The product (121 mg, 48%) was collected, washed with ethanol, washed with diethyl ether, and dried in vacuo. Mn(II)Mn(II)LBF4 , Template Method. Under a helium atmos- phere, a solution of 200 mg (0. 82 mmol) of Mn(OAc) 2 • 4H 20 and 74 120 mg (0. 82 mmol) of tren in 25 mL of ethanol was added dropwise to a warm, stirred solution of 210 mg (1. 28 mmol) of 2-hydroxy-5-methyl ... isophthalaldehyde and 8 drops of TEA in 25 mL of ethanol. The resulting yellow-orange solution was heated at reflux for 15 min. After the solution had cooled to ambient temperature, a solution of 140 mg (0. 42 mmol) of TBABF4 in 5 mL of ethanol was added. A yellow solid (0. 21 g, 59%) precipitated immediately and was collected, washed with ethanol, washed with diethyl ether, and dried in vacuo. 3, 3', 3"-Tris(2-hydroxybenzaldehyde-3-carboxaldimino)nitrilo~ A slurry of 1. 0 g (6. 7 mmol) of 2-hydroxy-5-methyl- isophthalaldehyde in 5 mL of ethanol was stirred at 40• C until all the solid just dissolved. The heating was discontinued and ,...., 10 mL of a solution of 0. 2 g (1. 4 mmol) of tren in 25 mL of ethanol was added dropwise to the stirred dialdehyde solution. The interior of the reaction flask was then scratched several times, and after stirring was resumed,yellow microcrystalline solid began to precipitate. To this stirred mixture, the remaining portion of the tren solution was added dropwise. After being cooled to ambient temperature, the mixture was filtered through a coarse frit. 'The collected solid (0. 57 g, 70o/o) was washed with ethanol (2 x 10 mL) and diethyl ether (2 x 20 mL), and dried in vacuo. 1 H NMR (CDC~, chemical shifts, <'5, in ppm relative to internal Me 4 Si) 2.04 (3H, s, CH 3 ), 2.88 (2H, m, N-CH 2 -R), 3.58 (2H, m, imine-CH 2 -R), 6.39 and7.54(2H, d, aryl, JHH=2Hz), 7.98(1H, s, imine), and 10.05 (1H, s, aldehyde). 75 M(InLH 3 (BF4 ) 2 • H2 0 Template Method; M(In=Mn(II), Fe(II), Co(II) and Cd(II). Preparations of Mn(II), Fe(II), and Co(II) complexes ~ were conducted under a helium atmosphere. To a solution of 35 mg (0. 24 mmol) of tren in 10 mL of ethanol was added 0. 24 mmol of the appropriate hydrated metal(II) acetate and 10 mL of acetonitrile. The mixture was stirred until the metal salt dissolved and then added dropwise to a stirred slurry of 154 mg (0. 26 mmol) of 2 in 15 mL of aceto"" nitrile. After being stirred for 1 h, the resulting solution was filtered, and a solution of 238 mg (0. 72 mmol) of TBABF4 in 30 mL of ethanol was added. If solid precipitated, acetonitrile was added dropwise until the solid just dissolved. Slow evaporation of the solution over a period of 4-5 days resulted in the product crystallizing from solution. The crystals (30-50% yield) were collected, washed with ethanol, washed with diethyl ether, and air dried. The colors of the complexes are as follows: Mn(II) orange-yellow, Fe(II) green, Co(II) yellow-orange, and Cd(II) yellow. NaLH 3 N0 3 • A mixture of 87.7 mg (0. 60 mmol) of tren and ~ 51.0 mg (0. 60 mmol) of sodium nitrate in ethanol (23 mL) and acetonitrile (15 mL) was stirred until all of the sodium nitrate dissolved. This solution was added to a stirred slurry of 370 mg (0. 63 mmol) of 2 "" in 25 mL of acetonitrile. The resulting yellow solution was stirred for 1 h, filtered, and diluted with 40 mL of ethanol. After reducing the solution volume to 25 mL using a rotary evaporator, 30 mL of diethyl ether was added. Yellow microcrystalline needles formed immediately which upon standing overnight recrystallized as orange-yellow parallelepipeds (380 mg, 83%). The product was collected, washed with 76 diethyl ether, and dried in vacuo. M(II)LH3 (BF4 ) 2 • H2 0 Exchange Method; M=Mn(ll), Fe(II), Co(II), Ni(II), Zn~II), Cd(ll), and Mg(ll). For all metals except Fe(II), a solution of 0. 36 mmol of the appropriate hydrated metal(II) acetate in 20 mL of methanol was prepared. For Fe(II), 0. 36 mmol of FeC~ · 2THF 51 was used instead. The metal solution was added dropwise to a stirred solution of 305 mg (0. 40 mmol) of NaLH 3 N03 in 40 mL of methanol. The solution was stirred for 15 min and then filtered. To the stirred filtrate was added a solution of 500 mg (1. 5 mmol) of TBABF4 in 20 mL of methanol. Solid precipitated and was collected, washed with 1:1 methanol/diethyl ether (2 x 15 mL) and diethyl ether (2 x 20 mL), and dried in vacuo. Yields of complexes were generally greater than 85%. The colors of those complexes which can be prepared only by this method are as follows: Ni(II) gold, Zn(II) yellow, and Mg(II) yellow. Cu(II)LH 3 (BF4 ) 2 • 3/2H 2 0. To a stirred slurry of 600 mg (0. 79 mmol) of NaLH 3 N0 3 in 60 mL of methanol was added 125 mg (0. 63 mmol) of Cu(OAc) 2 • H20. The resulting brown solution was stirred for 15 min and then filtered. To the stirred filtrate was added a solution of 1. 0 g (3. 0 mmol) of TBABF4 in 15 mL of methanol. A gold-brown solid (210 mg, 35%) crystallized which was collected, washe~ with 1:1 methanol/diethyl ether, and dried in vacuo. Fe(III)LH 3 (BF4 ) 3 • To a slurry of 400 mg (0. 53 mmol) of ~ NaLH 3 N03 in 80 mL of methanol was added a solution of 216 mg (0. 80 mmol) of FeCis • 6H 2 0 in 15 mL of methanol. After being stirred for 15 min, the resulting dark-blue solution was filtered. To the 77 filtrate was added 1. 5 g (4. 5 mmol) of TBABF4 • The volume of the solution was reduced to 10 mL using a rctary evaporator. The darkblue solid (380 mg, 72%) which crystallized from the solution was collected, washed with methanol, washed with diethyl ether, and dried in vacuo. MnLH 3 (Cl04 ) 3 • Under a helium atmosphere, a solution of 110 mg (0.12 rnmol) of Mn(II)LH 3 (BF4 ) 2 • H2 0 in 30 mL of acetonitrile (0.1 M in tetrabutylarnmonium perchlorate) was electrolyzed at 0. 7 V vs. a Ag+ /Ag reference electrode. After passing 12.1 coulombs, a solution of 1. 0 g (2. 9 mmol) of tetrabutylammonium perchlorate in 75 mL of methanol was added to the electrolytic solution. Using a rotary evaporator, the volume of the solution was reduced to 15 mL resulting in the crystallization of a dark-brown solid. The product (66 mg, 55%) was collected, washed with methanol, washed with diethyl ether, and dried in vacuo. Fe(II)Fe(II)LBF4 • Under a helium atmosphere, a solution of ~ 150 mg (0. 86 mmol) of Fe(OAc) 2 in 20 mL of warm methanol was added to a solution of 300 mg (0. 39 mmol) of NaLH 3 N03 in 50 mL of methanol. After being stirred for 15 min, the solution was filtered and 8 drops of TEA was added. The solution was warmed to 50° C and stirred for 15 min. After being cooled to ambient temperature, the solution was stirred for an additional 12 h. A small amount of microcrystalline Fe(II)Fe(II)LN03 was removed by filtration, and a solution of 300 mg (0. 91 mmol) of TBABF4 in 15 mL of methanol was added to the filtrate. A brown, microcrystalline solid (192 mg, 56%) was collected, washed with methanol, washed with diethyl ether, and dried under a stream of helium. 78 Co(II)Co(II)LBF4 • H20. Under a helium atmosphere, 8 drops of TEA and a filtered solution of 87 mg (0. 35 mmol) of Co(OAc) 2 • 4H 20 in 20 mL of methanol were added to a filtered solution of 300 mg (0. 32 mmol) of Co(II)LH 3 (BF4 ) 2 • H2 0 in 70 mL of 1:1 methanol/acetonitrile. The solution was stirred for 12 h during which time an orange solid precipitated. The solid (198 mg, 69%) was collected, washed with 1:1 methanol/acetonitrile (2 x 20 mL) and diethyl ether (2 x 20 mL), and dried in vacuo. Cu(rncu(rnLBF4 • 2H 2 0. To a stirred solution of 300 mg (0. 39 mmol) of NaLH 3 N03 in 80 mL of methanol was added 173 mg (0. 87 mmol) of Cu(OAc) 2 • H2 0. The resulting green solution was filtered, and 400 mg (1. 2 mmol) of TBABF4 was added to the filtrate. The volume of the solution was reduced to 20 mL using a rotary evaporator. A darkgreen solid crystallized from the solution and was collected, washed with 1:1 methanol/diethyl ether, and dried in vacuo. Cu(II)Cu(II)LBF4 • H2 0. To a solution of 325 mg (0. 33 mmol) of Cu(II)Cu(II) LH(BF 4 ) 2 • 2H 20 in 30 mL of acetonitrile was added 3 mL of diisopropylethylamine and 80 mL of methanol. The resulting orange-red solution was concentrated to a volume of 15-20 mL using a rotary evaporator. Upon standing for 4-6 h, the solution yielded maroon-red crystals (136 mg, 45%). This hygroscopic product was collected, washed with 100:1 methanol/diisopropylethylamine (2 x 20 mL) and diethyl ether (2 x 25 mL), and dried in vacuo. MA (II)MB(II)LBF4 • xH 2 0 Insertion of MB(II) into MA (II)LH 3 (BF4 ) 2 • xH 2 0; General Method. The followirig general method can be used to prepare most of the homonuclear and heteronuclear MA (II)MB(II)L + 79 complexes. Specific details are reported following the general procedure for those complexes which were prepared by this method. If either MA (II) or MB(II) was Mn, Fe, or Co, the preparation was carried out under a helium atmosphere. A filtered solution of 0. 49 mmol of MB(II) acetate (hydrate except for Fe(II)) in 25 mL of methanol (heated to 50• C for Fe(II)) and 8 drops of TEA were added to a filtered solution of 0.49 mmol of MA(II)LH 3 (BF4 ) 2 • xH 2 0 in acetonitrile (30-35 mL) and methanol (40-45 mL). While the solution was being stirred for 12 h, a microcrystalline solid precipitated. The product was collected, washed with methanol (2 x 20 mL) and diethyl ether (2 x 20 mL), and dried in vacuo. The mononuclear complex into which Mn(II) and Co(II) were inserted was Mn(II)LH 3 (BF4 ) 2 • H2 0. Yellow Mn(II)Mn(II)LBF 4 and orange Mn(II)Co(II)LBF4 were isolated in 65-70% yields. M=Mn(II complex into which Mn(II), Co(II), and Cu(II) were inserted was Fe(II)LH3 (BF4 ) • H2 0. Dark-green Fe(II)Mn(II)LBF 4 , brown Fe(II)Co(II)LBF4 , and black Fe(II)Cu(II)LBF4 were isolated in 50-65% yields. into :Which Mn(II) and Fe(II) were inserted was Ni(II)LH 3 (BF4 ) 2 • H2 0. Light-brown Ni(II)Mn(II)LBF4 and red-brown Ni(II)Fe(II)LBF4 were isolated in 60-70% yields. M = Mn(ll) and Co(II). The mononuclear complex into which Mn(II) and Co(II) were inserted was 80 Cu(II)LH 3 (BF4 ) 2 • 3/2H 20. Orange-brown Cu(II)Mn(IT)LBF4 • H2 0 and brown Cu(II)Co(II)LBF4 • !H 2 0 were isolated in 65-70% yields. M(IT)Zn(II)LBF • xH 0 M = Mn(II), Fe(II) and Zn(IT). The mononuclear complex into which Mn(II), Fe(II), and Zn(II) were inserted 'W'aS Zn(II)LH 3 (BF4 ) 2 • H2 0. A second crop of Mn(II)Zn(II)LBF4 was isolated from the filtrate after the volume had been reduced to ,..., 30 mL using a rotary evaporator. Yellow Mn(II)Zn(II)LBF4 , brown Fe(IT)Zn(II)LBF4 • H20, and yellow Zn(II)Zn(II)LBF4 were isolated ih 70-85% yield. and Cd(II). The mono- ~~"""'~,.,.,..._."""'..,..._,..,_"""'"'w'"""""~w""""'"'.,...,..._"""~........,....~'"'"'""""""''"""' nuclear complex into which Mn(II), Fe(II), Cu(IT), and Cd(II) were inserted was Cd(II)LH 3 (BF 4 ) 2 • H20. A second crop of Cd(II)Cd(II)LBF 4 was isolated from the filtrate after the volume had been reduced to ,..., 20 mL using a rotary evaporator. Yellow Mn(II)Cd(II)LBF4 , brown Fe(II)Cd(II)LBF4 , light orange-brown Cu(II)Cd(II)LBF4 , and light yellow Cd(IT)Cd(II)LBF 4 were isolated in 75-85% yields. M(II)Mg(II)LBF · M=Mn II), Fe(II). The mononuclear complex into which Mn(II) and Fe(II) were inserted was Mg(II)LH 3 (BF 4 ) 2 • H2 0. Yellow Mn(II)Mg(IT)LBF4 and light maroon Fe(II)Mg(II)LBF 4 were isolated in 55-65% yields. ~BF · M =Ni(II 450 mg (0. 48 mmol) of Cu(II)LH 3 (BF4 ) 2 • 3/2H2 0 in acetonitrile (30 mL) and methanol (55 mL) were added 8 drops of TEA and a filtered solution of 0. 49 mmol of the appropriate hydrated metal(II) acetate in 25 mL of methanol. The solution was stirred for 2 hand then 25 mL of methanol was added. The solution was concentrated to 15 mL using a rotary evaporator. The brown microcrystals ( 40-55% yield) which 81 had formed were collected, washed with 2:1 methanol/diethyl ether (2 x 20 mL) and diethyl ether (2 x 20 mL), and dried in vacuo. Cu(II)Zn(ll)LBF 4 • H2 0. To a filtered solution of 300 mg (0. 32 mmol) of Cu(II)LH 3 (BF4 ) 2 • 3/2H2 0 in acetonitrile (25 mL) and methanol (35 mL) was added a filtered solution of 70 mg (0. 32 mmol) of Zn(OAc) 2 • 2H 20 in 35 mL of methanol. After adding 2 mL of diisopropylethylamine) th'= solution was concentrated to 15-20 mL using a rotary evaporator. Upon standing for 4-6 h, the solution yielded redbrown crystals (121 mg, 42%). The product was collected, washed with 100:1 methanol/diisopropylethylamine (2 x 20 mL) and diethyl ether (2 x 25 mL), and dried in vacuo. 82 REFERENCES ~ 1. Gagne, R. R. ; Spiro, C. L. ; Smith, T. J. ; Hamann, C. A. ; Thies, W. R.; Shiemke, A. K. J. Am. Chern. Soc. ill!, 103, 4073. 2. Gagne, R. R.; Koval, C. A.; Smith, J.; Cimolino, M. C. J. Am. Chern. Soc. ~' 101, 4571. 3. Lambert, S. L.; Hendrickson, D. N. Inorg. Chern. 1979, 18, ~ 2683. 4. Spiro, L. L. ; Lambert, S. L. ; Smith, T. J. ; Duesler, E. N. ; Gagne, R. R.; Hendrickson, D. N. Inorg. Chern. ~' 20, 1229. 5. Spiro, C. L. ; Lambert, S. L. ; Gagne, R. R.; Hendrickson, D. N. Inorg. Chern. .!1.§1, 21, 68. 6. Long, R. C.; Hendrickson, D. N. submitted to J. Am. Chern. Soc. 7. Gagne, R. R.; Henling, L. M.; Kistenmacher, T. J. Inorg. Chern. 1980, 19, 1226. ~ 8. (a) Hoskins, B. F.; Williams, G. A. Aust. J. Chern. 1975, 28, ~- 2607; (b) Hoskins, B. F.; McLeod, N. J.; Schaap, H. A. Aust. J. Chern. ~' 29, 515. 9. (a) Hoskins, B. F.; Robson, R.; Williams, G. A. Inorg. Chim. Acta !.ill, 16, 121; (b) Hoskins, B. F.; Williams, G. A. Aust. J. Chern. 1975, 28, 2593. ~- 10. Busch, D. H. Rec. Prog. Chern.~' 25, 107. 11. Wilson, L. J.; Rose, N. J. J. Am. Chern. Soc. ~' 90, 6041. 12. Mealli, C.; Lingafelter, E. C. J. Chern. Soc., Chern. Commun. 1970, ,...,..,_ 885. 83 REFERENCES (continued) ~ 13. Larsen, E.; LaMar, G. N.; Wagner, B. E.; Parks, J. E.; Holm, R. H. Inorg. Chern. 1972, 11, 2652. ~- 14. (a) Dwyer, F. P.; Gill, N. S.; Eleonora, C. G.; Lions, F. J. Am. Chern. Soc. llll, 79, 1269; (b) Broomhead, J. A.; Robinson, P. J. Aust. J. Chern.~' 21, 1365. 15. Lehn, J. -M. Ace. Chern. Res. !m, 11, 49. 16. Krause, R. A.; Colthup, N. B.; Busch, D. H. J. Phys. Chern. _!1g, 65, 2216. 17. The J values for M(II)M(II)L +, with M = Mn, Fe, Co, and Cu, 1 are -2. 8, -0. 8, -0. 9, and -33.5 cm- , respectively. 18. Bothner-By, A. A. Adv. in Mag. Res. _!W, !, 195. 19. (a) Jensen, C. F.; Deshmukh, S. ; Jakobsen, H. J. ; Ilhmers, R. R.; Ellis, P. D. J. Am. Chern. Soc. ~' 103, 3659 (J for 113 Cd-N-C-H = 12.1 Hz); (b) Cardin, A. D.; Ellis, P. D.; Odom, J. D.; Howard, Jr., J. W. J. Am. Chern. Soc. llll, 97, 1672 (J for 20. 113 Cd-C-C-H =,...50 Hz). Lehn, J. -M.; Sauvage, J. P.; Dietrich, B. J. Am. Chern. Soc. 1970, 92, 2916. ~- 21. Geary, W. J. Coord. Chern. Revs. ill!, ']_, 81. 22. Merrill, P. H.; Goedken, V. I.; Busch, D. H.; stone, J. A. J. Am. Chern. Soc. !Ql§_, 92, 7590. 23. Fluck, E. in "Chemical Applications of Mossbauer Spectroscopy", Goldanski, V. I.; Herber, R. H.; Eds.; Academic Press, New York, 1968. 84 REFERENCES (continued) ~ 24. Gagne, R. R.; Koval, C. A.; Lisensky, G. C. Inorg. Chern. 1980, 19' 2854. ~- 25. These EPa - Epc values were not corrected for uncompensated solution resistance. 26. Nicholson, R. S.; Shain, I. Anal. Chern. J!M, 36, 706. 27. (a) Boston, D. R.; Rose, N. J. J. Am. Chern. Soc. ~' 90, 6859; li71, 95, 4163; (b) Jackels, S. C.; Duerdorf, S. S.; Rose, N. J.; Zehtzer, J. J. J. Chern. Soc., Chern. Commun. !1ll, 1292; (c) Jackels, S. C.; Rose, N. J. Inorg. Chern. ~' 12, 1232. 28. Parks, J. E.; Wagner, B. E.; Holm, R. H. J. Am. Chern. Soc. 1970, 92, 3500. ~,.._,..._ 29. Creaser, I. I.; Harrowfield, J. McB.; Herit, A. J.; Sargeson, A. M.; Springboard, J.; Gener, R. J.; Snow, M. R. J. Am. Chern. Soc. ]111, 99, 3181. 30. Goedken, V. L.; Peng, S. M. J. Chern. Soc., Chern. Commun. 1973, 62. ~ 31. Herron, N.; Grzybowski, J. J.; Matsumoto, N.; Zimmer, L. L.; Christoph, G. G.; Busch, D. H. J. Am. Chern. Soc. ~' 104, 1999. 32. Hoskins, B. F.; Williams, F. D. Coord. Chern. Rev • .!Q.ll, ~' 365. 33. (a) Lindoy, L. F. Chern. Soc. Rev. ~' !, 421; (b) Busch, D. H.; Farmery, K.; Katovic, V.; Melayk, A. C.; Sperati, C. R.; Tokel, N. E. Adv. Chern. Ser. ~' 100, 44. 85 REFERENCES (continued) ~ 34. Linstead, R. P. J. Chern. Soc. 1934, 1016. 35. Day, V. W.; Marks, T. J.; Wachter, W. A. J. Am. Chern. Soc. ~ W§_, 97' 4519. 36. Spiro, C. L. Ph. D. Thesis, California Institute of Technology, 1981. 37. Creutz, C.; Taube, H. J. Am. Chern. Soc. 1973, 95, 1086. 38. Phelps, J.; Bard, A. J. J. Electroanal. Chern. 1976, 68, 313. 39. Sutton, J. E.; Sutton, P. M.; Taube, H. Inorg. Chern. ~' ~- ~- 18, 1017. 40. Creutz, C.; Kroger, P.; Matsubara, T.; Netzel, T. L.; Sutin, N. J. Am. Chern. Soc. 1979, 101, 5442. ~- 41. Taube, H. Ann. N.Y. A cad. Sci. 1978, 481. 42. Kanda, W.; Nakamura, M.; Okawa, H.; Kida, S. Bull. Chern. ~ Soc. Jap. ~' 55, 1982. 43. Galy, J.; Jaud, J.; Kahn, 0.; Tola, P. Inorg. Chim. Acta~' 36, 229. 44. (a) Mikuriya, M. ; Okawa, H. ; Kida, S.; Ikuhiko, U. Bull. Chern. Soc. Jap. ~' Q!., 2920; (b) Beale, J. P.; Cunningham, J. A.; Philips, D. J. Inorg. Chim. Acta ~' 33, 113. 45. Ullmann, F.; Brittner, K. Chern. Ber. 1909, 42, 2539. 46. . still, W. C. ; Kahn, M.; Mitra, A. J. Org. Chern. 1978, 43, 2923. 47. Kimura, E.; Young, S. ; Collman, J. P. Inorg. Chern. lll.Q, .§., ~- ~ 1183. 48. Glerup, J.; Josephsen, J.; Michelsen, K.; Pedersen, E.; Schaffer, C. E. Acta Chern. Scand. lll.Q, 24, 247. 86 REFERENCES (continued) ~ 49. Earnshaw, A. "Introduction to Magnetochemistry"; Academic Press, London, 1968, p. 9. 50. Munck, E.; Debrunner, P. G.; Tsibris, J. C.; Gunsalvs, I. C. Biochemistry 11:G, !!, 855. 51. LeVanda, C.; Bechgaard, K.; Cowan, D. 0.; MuellerWesterhoff, U. T.; Eilbracht, P.; Candela, G. A.; Collins, R. L. J. Am. Chern. Soc. ~' 98, 3181. 87 CHAPTER 3 Magnetic Properties of Binuclear Clathrochelate Complexes 88 INTRODUCTION ~ Binuclear complexes with magnetically interacting metal centers have been examined in efforts (1) to understand the mechanisms of spinspin coupling 1 - 3 and (2) to develop model systems for magnetic interactions in linear chains, clusters, and paramagnetic aggregates. Certainly the most extensively studied series of such binuclear complexes are those with Cu(IT) metal centers and two bridging oxygendonor ligands. 4 - 6 From these studies a theoretical relationship between J -values and metal-ligand-metal bridging angle has been developed. 6b Two general classes of magnetically interacting binuclear complexes which have received less attention are (1) those with three symmetrically bridging ligands 7 ' 8 and (2) those in which the metal . . d syst emat•1ca11y. 9 - 11 wns are vane Such studies of the latter type are essential for obtaining an understanding of the relationships between exchange coupling and the number of unpaired electrons. These studies have been limited because of the difficulty in preparing structurally similar complexes for a series of different metal ions. The general syntheses of binuclear complexes of the clathro3 chelating ligand, L -, allow the examination of magnetic interactions N s- L ,...,..., 89 for several different pairs of metal ions. In complexes of Ls- the metal ions are likely to be symmetrically bridged by three phenolic oxygens. Although metal-ligand geometry may differ depending on the nature of the metals, a clathrochelating ligand is likely to minimize these structural variations. The results of variable-temperature magnetic susceptibility studies for a series of four of these complexes are reported here. RESULTS ~ Variable temperature (4. 2 to> 220 K) magnetic susceptibility data were collected for Mn{II)Mn{TI)LBF4 , Fe(II)Fe(II)LBF4 , Co(II)Co(II)LBF4 • H2 0, and Cu(TI)Cu(II)LBF4 • H2 0. For each set of data a least-squares fit to an appropriate equation was calculated. These equations include a theoretical expression (from Van Vleck's equation) derived for isotropic exchange interactions (H = -2JS1 • ~) and a correction term for paramagnetic impurities (PARA); details are given in the Experimental Section. Both data and least-square results for the complexes are given in Tables I-IV and are depicted graphically in Figure 1-4. The molar susceptibility (x) for Mn(II)Mn(II)LBF4 as a function of temperature (T) does not appear to differ significantly from Curie-law behavior (Figure 1) . . There are, however, two obvious indications for an antiferromagnetic interaction in this complex. First, the effective moment per metal ion (J.leff/Mn) at ambient temperature, 5. 6(1) 1-LB' is lower than the expected spin-only value (S1 = S2 = 5/2) of 5. 92 1-LB· Second, 1-leff/Mn decreases significantly from 5. 51 1-LB at 224 K to 1. 62 at 4. 2 K (Table I). The least-squares fitting of the data gives 90 values indicating weak antiferromagnetic coupling with J = -2. 8 em -l, g = 1. 94, and PARA = 0. 0961 cgsu. A likely impurity, which would necessitate a paramagnetic correction, is Mn(II)LH3 (BF4 ) 2 • H20. The X vs. T curve for Fe(II)Fe(II)LBF4 more closely resembles an ideal Curie-law X vs, T curve than that of Mn(IT)Mn(II)LBF4 (Figure 2). Indeed, prior to a least-squares fitting of the data, the only evidence for a magnetic interaction is the substantial decrease of 1-Leff/Fe from 5. 58 1-LB at 286 K to 3.12 1-LB at 4. 2 K (Table IT). This behavior is consistent with weak antiferromagnetic coupling, and is supported by a least-squares fitting of the data which gives J = -0. 82 em - l and g = 2. 22. There is no evidence for paramagnetic impurities since PARA = 0. 00. The behavior of x as a function of T for Co(II)Co(II)LBF4 • H20 is very similar to that for Fe(II)Fe(II)LBF4 (Figure 3). Weak antiferromagnetic coupling again is suggested by the decrease of I-Leff/Co from 4. 81 JlB at 286 K to 3. 54 1-LB at 4. 2 K (Table III). Values for J and g of 0. 93 cm- 1 and 2. 49, respectively, have been obtained from a leastsquares fitting of the data. As was the case for samples of Fe(II)Fe(II)LBF4 , those of Co(II)Co(IT)LBF4 • H20 were not contaminated by paramagnetic impurities (PARA = 0. 00). Of the four clathrochelates examined by variable temperature magnetic techniques, only Cu(II)Cu(II)LBF4 • H20 has a maximum molar susceptibility at an intermediate temperature (,...,. 65 K) characteristic of an antiferromagnetic interaction (Figure 4). Additionally, Jleff/Cu decreases from 1. 83 JlB at 245 K to 0. 20 Jl B at 4. 84 K (Table IV). A least-squares fitting of the data gives J = -33.5 em-\ g = 2.19, and 91 PARA =0. 0418 cgsu. Possible paramagnetic impurities are mononuclear Cu(IT) species. DISCUSSION ~ The magnetic exchange parameters (J-values) for the four binuclear clathrochelate complexes are summarized in Table V. The data indicate that the metals in all four complexes are weakly antiferro- magnetically coupled. For certain geometries and d-electron configurations, weak antiferromagnetic interactions can result from the cancellation of considerable antiferromagnetic interactions by roughly compensatory ferromagnetic interactions. 12 Since the weak interactions in these M(II)M(II)L + complexes are generally independent of the spin state of M(In, an alternative explanation for these weak interactions must be invoked. A reasonable explanation for these interactions is that the overlap of metal d-orbitals with the s- and p-orbitals of the bridging phenolic oxygens is small. 13 Structural studies which are in progress should clarify the nature of the metal-oxygen bonds in these complexes. There are few related first-row transition-metal complexes having three bridging ligands with which to compare the results obtained for the M(II)M(IT)L +complexes reported here. Several [M(IIT) 2 C~] s- complexes with three bridging chloride ligands have been observed to exhibit weak antiferromagnetic interactions ( -25 cm-1 < J < -1 em - 1 ). 7 In a few of these cases, the weak couplings also have been ascribed to poor orbital overlap. 7ab However, since both metal ion charges and bridging ligands differ from those of M(InM(!DL +complexes, these results are not comparable. It is clear 92 that additional studies are needed to unambiguwsly define the nature of these weak interactions. 93 E rimental Section All compounds were prepared and characterized as described in Chapter 2. A PAR Model150A vibrating sample magnetometer operated at 13.5 KG was used to obtain the variable-temperature susceptibility data. The instrument was standardized with respect to CuS04 • 5H2 0 in the temperature range, 4. 2-286 K, using the reported molar susceptibility values. 14 Sample temperature was determined using a calibrated GaAs diode. The data for all samples and at all temperatures were corrected for the diamagnetism of the sample container and th~ background. Additionally, the data were corrected for diamagnetism of each compound using estimated values from Pascal's constants. 15 The resulting paramagnetic susceptibility data were least squares fit, using an available computer program, 16 to the appropriate equations which include a theoretical expression and a correction term for paramagnetic impurities. The theoretical expressions are based on a spin Hamiltonian for isotropic magnetic exchange, H = -2JS1 • ~· The general form for these expressions is given in eq 1 as has been reported by Earnshaw et al. 17 A 11 of the constants have their usual 2 XM = (Ng l/3kT){~ S'(S' +1)(2S' +1)n(S')exp(-Es'/kT)}/ {~ (2S' + 1) n(S')exp (Es' /kT)} (1) meanings. For homonuclear complexes Sis defined as S1 = S2 • The term, S', is the quantum number of the coupled energy levels and assumes values of 28,28-1, 28-2, .•. , 0 in the summation. The 94 energy term, Es', is equal to -J[S' (S' +1)- 2S(S + 1)], and O(S') is the degeneracy of a given energy level. The variable term for paramagnetic impurities is included in the appropriate equations as (4. 2/T) PARA. 95 Table I. Experimental and Calculated Magnetic Susceptibility Data for ~ Mn(II)Mn(II)LBF4 1 XM X 10, cgsu T •K P. eff /Mn, P.B ' obsd calcd obsd calcd 224. 0.3389 0.3472 5.51 5.58 203. 0.3727 0.3778 5.51 5.55 183. 0.4055 0.4143 5.45 5.51 163. 0.4496 0.4584 5.41 5.46 143. 0. 5023 0.5119 5.35 5.40 122. 0.6063 0.5784 5.45 5.32 103. 0.6701 0.6816 5.25 5.21 94.2 0.6974 0.7051 5.12 5.15 86.6 0.7351 0.7480 5.04 5.09 76.9 0.7532 0.8098 4.81 4.99 67.0 0.8582 0.8820 4.80 4.86 61.6 0.9056 0.9255 4.72 4.77 55.9 0.9479 0.9746 4.60 4.67 45.2 1.024 1. 074 4.30 4.41 41.9 1.056 1.106 4. 21 4.30 38.6 1.086 1.138 4.10 4.19 35.3 1.120 1.169 3.98 4.06 32.0 1.156 1.199 3.85 3.92 28.1 1.208 1.253 3.68 3.72 24.2 1.262 1.264 3.49 3.50 22.2 1.290 1.280 3.38 3.37 96 Table I (continued) ~ 21.2 1. 298 1.287 3.32 3.30 20.2 1.303 1.295 3.24 3.23 18.9 1. 312 1.305 3.15 3.14 17.7 1.322 1. 315 3.06 3.05 16.1 1.345 1.330 2.94 2.92 14.5 1.369 1.347 2.82 2.79 12.6 1.411 1. 373 2.66 2.62 10.6 1.464 1.408 2.49 2.44 10.2 1.476 1.418 2.45 2.40 9.71 1.488 1.429 2.40 2.36 9.38 1.502 1.437 2.37 2.32 8.95 1.519 1. 449 2.33 2.28 8.44 1. 537 1.465 2.28 2.22 7.77 1. 562 1.489 2.20 2.15 7.29 1. 589 1.508 2.15 2.10 6.32 1.612 1.556 2.02 1. 98 5.64 1.626 1.598 1. 91 1.90 5.24 1.655 1.628 1.87 1.85 4.46 1.550 1.697 1. 66 1.74 4.20 1.573 1.725 1.62 1. 70 97 Table II. "'"""""~ Experimental and Calculated Magnetic Susceptibility Data for Fe(II)Fe(II)LBF4 T •K XM X 10\ cgsu 1-Leff /Fe, 1-LB ' obsd calcd obsd calcd 286. 0.2724 0.2547 5.58 5.39 265. 0.2806 0.2740 5.45 5.39 745. 0.2980 0.2964 5.40 5.38 224. 0.3207 0.3229 5.36 5.38 204. 0.3620 0.3546 5.43 5.37 183. 0.4252 0.3931 5.58 5.36 163. 0.4566 0.4410 5.45 5.35 143. 0.5113 0.5012 5.41 5.35 123. 0.5927 0.5803 5.39 5.33 103. 0.6923 0.6861 5.33 5.31 94.2 0.7585 0.7455 5.34 5.30 86.1 0.8072 0.8115 5.27 5.28 76.9 0.8985 0.9023 5.26 5.27 67.0 1. 018 1.026 5.22 5.24 61.4 1. 098 1.110 5.19 5.22 55.9 1.194 1.210 5.17 5.20 51.4 1.273 1.305 5.12 5.18 43.3 1.470 1.521 5.04 5.13 39.8 1.583 1.638 5.02 5.10 36.2 1. 700 1.773 4.96 5.07 32.8 1.839 1. 931 4.91 5.03 98 Table ll (continued) ~ 29.3 2.011 2.119 4.85 4.98 25.6 2.198 2.354 4.75 4.91 22.0 2.457 2.644 4.65 4.82 20.2 2.618 2.814 4.60 4.77 19.6 2.706 2.884 4.59 4.74 18.8 2.819 2.958 4.50 4.72 17.0 2. 961 3.158 4.49 4.64 15.3 3.126 3.362 4.37 4.55 13.7 3. 205 3. 601 4.26 4.45 12.2 3.295 3.857 4.08 4.32 10.1 3.979 4.176 4.00 4.10 7.73 4.534 4.530 3.74 3.74 7.14 4.543 4.603 3.60 3.62 6.55 4.639 4.665 3.49 3.50 5.96 4.742 4.709 3.36 3.35 5.38 4.872 4.733 3.24 3.19 4.79 5.109 4.732 3.13 3.01 4.49 5.364 4.720 3.10 2.91 4.20 5.799 4. 701 3.12 2.81 99 Table III. Experimental and Calculated Magnetic Susceptibility Data ~ for Co(II)Co(II)LBF4 • H2 0 1 T ' •K xM x 10, cgsu lleff /Co, J..LB obsd calcd obsd calcd 286. 0.2024 0.2008 4. 81 4.79 265. 0.2175 0.2161 4.80 4.78 245. 0.2305 0.2339 4.83 4 78 224. 0.2603 0.2547 4.83 4.78 204. 0.2863 0.2800 4.83 4.78 183. 0.3181 0.3107 4.83 4.77 163. 0.3555 0.3489 4.81 4.77 143. 0.4051 0.3970 . 4. 81 4.76 122. 0.4519 0.4605 4.70 4.75 103. 0.5487 0.5458 4.75 4.74 94.2 0.6047 0.5938 4.77 4.73 86.1 0.6438 0.6474 4.70 4.72 76.9 0.7093 0.7213 4.67 4.71 67.0 0.8050 0.8222 4.64 4.69 61.6 0.8661 0.8901 4.62 4.68 55.9 0.9484 0.9751 4.60 4.67 50.2 1.046 1.078 4.58 4.65 47.5 1.103 1.135 4.58 4.64 44.8 1.166 1.198 4.57 4.63 41.8 1.241 1.276 4.55 4.62 38.8 1.319 1.364 4.53 4.60 100 Table III (continued) ~ 35.3 1. 426 1.485 4.49 4.58 31.9 1.567 1.626 4.47 4.56 28.6 1.721 1. 788 4.44 4.52 24.9 1.919 2.012 4.37 4.48 22.8 2.065 2.168 4.34 4.44 21.6 2.152 2.263 4.31 4.42 20.5 2.249 2.365 4.29 4.40 17.8 2.513 2.641 4.23 4.34 15.0 2.667 2.994 4.15 4.24 13.2 3.135 3.281 4.07 4.16 11.1 3.526 3.601 3.96 4.04 10.7 3.631 3.751 3.94 4.00 10.2 3.711 3.841 3.90 3.97 10.0 3.755 3.888 3.88 3.95 9.81 3.784 3.935 3.85 3.93 9.59 3.839 3.984 3.84 3.91 9.38 3.883 4.030 3.82 3.89 9.16 3.964 4.080 3.81 3.86 8.94 4.005 4.130 3.78 3.84 8.74 4.057 4.176 3.77 3.82 8.54 4.110 4.223 3.75 3.80 8.35 4.180 4.267 3.74 3.77 7."95 4.279 4.368 3.67 3.72 7.55 4.421 4.454 3.65 3.67 101 Table Ill (continued) ~ 7.18 4.543 4.545 3.61 3.61 6.78 4.670 4.635 3.55 3.54 6.37 4,842 4.720 3.51 3.47 5.71 4.993 4.851 3.38 3.33 5.12 5. 301 4.944 3.29 3.18 4.20 5.464 5. 017 3.54 2.98 102 Table IV. Experimental and Calculated Magnetic Susceptibility Data ~ for Cu(II)Cu(II)LBF4 • H20 1 XM X 10, cgsu T •K I-Leff /Cu, 1-LB ' obsd calcd obsd calcd 245. 3.382 3.473 1. 82 1. 84 204. 4.039 4.042 1. 82 1.81 183. 4.510 4.400 1. 82 1. 80 143. 5.274 5.322 1.73 1.74 123. 5.773 5.914 1. 68 1. 70 103. 6.438 6.590 1. 63 1.64 94.2 6.822 6.908 1.60 1. 61 86.1 6.982 7.211 1.55 1. 58 76.9 7.270 7.538 1. 50 1. 52 67.0 7.502 7.817 1.42 1.45 61.6 7.517 7. 901 1. 36 1.40 55.9 7.484 7.899 1. 29 1. 33 50.2 7.395 7.753 1. 21 1.25 44.8 6.944 7. 418 1.12 1.15 41.8 6.667 7.126 1. 06 1.00 38.8 6.358 6.741 0.993 1. 02 35.3 5.917 6.161 0.914 0.933 31.9 . 5.409 5.444 0.831 0.833 28.6 4.655 4.636 0.745 0.728 24.9 4.096 3.648 0.639 0.603 22.8 3.757 3.076 0.585 0.529 103 Table IV (continued) ~ 21.6 3. 575 2.785 0.556 0.491 20.5 3. 316 2.512 0.525 0.454 17.8 2.915 1.986 0.456 0.376 15.0 2.469 1.676 0.386 0.318 13.2 2.391 1.631 0.355 0.293 11.1 2.384 1.753 0.326 0.279 10.7 2.363 1.803 0.318 0.278 10.2 2.394 1.860 0.313 0.276 9.81 2.362 1. 928 0.304 0.275 9.38 2.376 2.004 0.299 0.274 8.94 2.341 2.091 0.289 0.273 8.54 2.335 2.180 0.282 0.273 8.15 2.351 2.276 0.277 0.272 7.75 2.252 2.386 0.264 0.272 7.36 2.292 2. 505 0.260 0.272 6.96 2.222 2.642 0.249 0.271 6.37 2.335 2.875 0.244 0.271 4.84 2.145 3.745 0.204 0.269 104 Table V. Magnetic Exchange Parameters ~ Complex J, em Mn(II)Mn(II)LBF4 -2.8 Fe(II) Fe(II)LBF4 -0.8 Co(II)Co(II)LBF 4 • H2 0 -0.9 Cu(II)Cu(II)LBF 4 • H2 0 -33.5 -1 105 Figure 1. Molar paramagnetic susceptibility, xM' and effective magnetic moment, 1-Leff/Mn, vs. temperature and leastsquares fit of the data for Mn(II)Mn(II)LBF4 • Figure 2. Molar paramagnetic susceptibility, xM' and effective magnetic moment, 1-Leff/Fe, vs. temperature and leastsquares fit of the data for Fe(II)Fe(II)LBF4 • Figure 3. Molar paramagnetic susceptibility, xM, and effective magnetic moment, 1-Leff/Co, vs. temperature and leastsquares fit of the data for Co(II)Co(II)LBF4 • H2 0. Figure 4. Molar paramagnetic susceptibility, xM' and effective magnetic moment, 1-Leff/Cu, vs. temperature and leastsquares fit of the data for Cu(II)Cu(II)LBF4 • H2 0. 106 IC ~ · ~----------------------------------------------~ c> \11 " IC ~ - Ill' :::1. s:: - ~ Cl) :::1. c. ..; c "' ~ c c) o.c . !iO . O BC. 0 l20 . C l60 . 0 200 . (1 240.0 280.0 3410.0 240.0 21'!0.0 320. 0 Temperature (K) .,c 0 f]; 0 ~ -;::s l'1.l - 0 .., c b.C c:i ...0 ..... "" C) X ::;s >( 0 c .., w 0 ID N 0 c:i c 0 c c;)• ().0 ~ 1&0.0 8C.C l2C.D J6C.O 200.0 Temperature (K) 107 c .0 * c 4- ~ c:::. - ::; ~ - ::i. 0 ..; (I) - ~ ............ Q) :i "'..,;· 0 .- . CJ o· D.P. 'iC.I:' J20.C Bl' . C' 160.0 200.0 2ll;J.(l 28('.0 320 . 0 2llo.r. 2SC.(l 320.(1 Temperature (K) ::t' :'1' L) 0 N L, '-"' 0 0 ,.,. 1.!. 0 . ~ m b.O () - + CD ""r1 0 .-4 0 It r- "' 0 X ~ >< \~ -· ~ ~. ..,r: "'~ 0 C> 0 0- o.r ~ . llCl.O 80.(' :20 . 0 160.1" 200 . 0 Temperature (K) 108 0 - ..; ~ - :::1. 0 ~'o-4 Cl N 'o-4 ~ :::1. ~ Cl ci c.o "0. (l so.c J20.[' J60.C 200.Cl 240. [' 2BO.O 320.1:' ~o . o 280. 0 320.0 Temperature (K) ro 110 0 ;:: c ., "'c N ==' Cf.l btl r.: ::I' c (.) ...0 .-4 \11 v,.,. c X ~ >< ""~ c ~ 0 0 c c; 0 ~ o.c &&o.o 10.0 J20. Cl J60 . 0 200.0 Temperature (K) 109 c. N,---------------------------------------------~ - ~ -~ ::1. .::. ~ ( I) ::1. j :{ c;. c:: o.c "o.r B:. r J~c.r lEO . C ~oc.c 21oc.r 2SO.P 3:.:; .. r 24D.C 2DO.C 32:.:- Temperature (K) r. , c.. c. ~T e: c.. c.. ci I.: c. c;. ::s t1.l b.O t) C'lo ~ X ~ X ci ~.J c:; ci { c. c.. J J 0 .:tf I +/ e; : 'k/ r .. c.. .:: · 5 c. ci o.r ~ I{C.O BC. C !20. 0 J60 . 0 ~ur.r Temperature (K) 110 REFERENCES 1. Hatfield, W. E. in "Theory and Applications of Molecular Paramagnetism"; Boudreaux, E. A.; Mulay, L. N., Eds., Wiley-Interscience: New York, 1976, p. 349. 2. Henning, J. C. M.; van den Boom, H. Phys. Rev. 1973, ~ B8, 2255. 3. 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. 4. J. Am. Chern. Soc. 1m, 97, 3039. Morishita, T. ; Hori, K. ; Kyuno, E. ; Tsuchiya, R. Bull. Chern. Soc. Jap. ~' 38, 1276. 5. (a) McWhinnie, P. J. Inorg. Nucl. Soc. 6. Ja~ J. Chern. Soc. 1964, 2959; McWhinnie, P. Chern.~ 27, 1063; (b) Okawa, H. Bull. Chern. WQ, 43, 3019. (a) McGregor, K. T.; Watkins, N. T.; Lewis, D. L.; Hodgson, D. J.; Hatfield, W. E. Inorg. Nucl. Chern. Lett.~ ~ 423; (b) Hodgson, D. J. Progress in Inorg. Chern.; Vol. 19; Lippard, S. J., Ed., Interscience: New York, 1975, pp. 173-187. 7. (a) Grey, I. E.; Smith, B. W. Aust. J. Chern. 1971, 24, 73; ~- (b) Ginsberg, A. P.; Robin, M. B. Inorg. Chern. ~ ~' (c) Saillant, R.; Wentworth, R.A. D. Inorg. Chern. 1968, ~ 8. 817; !J 1606. S_aillant, R. ; Jackson, R. B. ; Streib, W. E. ; Folting, K. ; Wentworth, R.A. D. Inorg. Chern. 1971, 10, 1453. ~ 9. (a) Lambert, S. L. ; Hendrickson, D. N. Inorg. Chern . ..!:JlQ, ~ 2683; (b) Spiro, C. L.; Lambert, S.L.; Smith, T.J.; Duesler, E.N.; 111 REFERENCES (continued) Gagn~, R. R. ; Hendrickson, D. N. Inorg. Chern. ~ 20, 1229; (c) Lambert, S. L.; Spiro, L. L. ; Gagn~, R. R. ; Hendrickson, D. N. Inorg. Chern. ~' _!!, 68. 10. Kahn, 0.; Tola, P.; Coudanne, H. J. Chern. Phys. J.m, 42, 355. 11. (a) O'Connor, C.J.; Freberg, D.P.; Sinn, E. Inorg. Chern. 1979, ~ 18, 1077; (b) Torihara, N.; Okawa, H.; Kida, S. Chern. Lett. 1978, 185. ~ 12. Ginsberg, A. D. Inorg. Chern. Acta Rev. 1971, §._, 45.. ~ 13. Coffman, R. E.; Buettner, G. R. J. Phys. Chern. 1979, 83, 2387. ~ ~' 14. Reekie, J. Proc. Royal Soc. (London) 173, 367. 15. Figgis, B. N.; Lewis, J. in "Modern Coordination Chemistry"; Lewis, J.; Wilkins, R. G., Eds., Interscience:. New York, 1960, p. 403. 16. Chandler, J.P. Program 66, Quantum Chemistry Program Exchange; Indiana University: Bloomington, IN. 17. Earnshaw, A.; Lewis, J. J. Chern. Soc.1961, 396. ~ 112 APPENDIX I Mononuclear and Binuclear Metal Complexes of 1, 3-Bis(2-pyridylimino)isoindolines Robert R. Gagne, William A. Marritt, David N. Marks, and Walter 0. Siegl 113 Mononuclear and Binuclear Metal Com lexes of 1 3-Bis(2- ridylimino isoindolines Robert R. Gagne, 1 a William A. Marritt, la David N. Marks, 1 a and Walter 0. Siegl1b Contribution No. 6297 from the Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91125, and the Ford Motor Company, Engineering and Research staff, Dearborn, Michigan 48121 Abstract ~ Condensation of 1, 2-dicyanobenzene and 2-amino-4-methylpyridine resulted in the formation of 4' -MeLH, which in its anionic form functions as a tridentate chelating ligand. Metal complexes were prepared with metal to ligand ratios of 1:1 and 1:2. The 1:1 complexes also contain acetate as a ligand while in the 1:2 complex the two tridentate ligands provide a pseudooctahedral environment about the metal ion. An analogous binucleating ligand was prepared by reaction of 1, 2, 4, 5-tetracyanobenzene and 2-amino-4-sec-butylpyridine. This ligand, as a dianion, is capable of binding two metal ions, providing three coordination sites for each. Complexes were prepared in which the remaining coordination sites are either occupied by 4' -MeLH, resulting in two six-coordinate metal ions, or by an acetate ligand. Mononuclear and binuclear complexes were prepared with Mn(II), Fe(II), C o(II), Co( III), Ni(II), Cu(II), and Zn(II). Magnetic, spectral and electrochemical properties of these molecules were investigated. 114 INTRODUCTION ~ stable metal complexes of the ligand .,... 1 have been prepared with ligand to metal ratios of 2:1 and 1:1. 2 The ligand functions as an LH, 3 R=H 4' -MeLH, R = 4' -CH 3 1 anionic tridentate chelate capable of occupying three coplanar sites about the metal ion and forming a pseudooctahedral environment around the metal ion in the bis-{ligand) complex, M~. 3 Metal complexes with M = Fe(II), Co(TI), NI{TI), Zn(II) and Cd{II) have been reported, but little had been done to characterize these molecules. These complexes became of greater interest after the synthesis of the conjugated organic molecule 2, which may function as a bridging ligand between two metal "' 4 ions. This molecule made it possible to prepare not only binuclear metal -complexes but also polymeric metal species. The polymeric complexes were of great interest for two reasons: polynuclear complexes offer a chance to study metal-metal interactions via electrochemistry and magnetic measurements, and complexes of this type are 115 HL-LH, 3 R = 4' -sec-butyl 2 interesting candidates for reactions involving multielectron transfer. We report here the preparation and certain physical properties of mononuclear and binuclear metal complexes with Mn(II), Fe(II), Co(II), Co( III), Ni(II), Cu(II) and Zn(II). RESULTS AND DISCUSSION nthesis and Characterization of the Li ands. Two general syntheses of mononucleating, 1, 3-bis(arylimino)isoindoline ligands RLH (1) involving metal ion-assisted condensation of phthalonitrile with "' a 2-aminopyridine or other 2-amino heterocycles were described earlier 5 (eq 1). The binucleating ligand 2 was prepared in a similar manner by "' these methods as shown in eq 2 and 3. The method of choice afforded the binucleating ligand in good yield from a one-flask synthesis utilizing an alkaline-earth salt (anhydrous CaCl 2 ) to catalyze the condensation of tetracyanobenzene with 4 equiv of aromatic amine. A second method, 116 ~CN ~CN + 2 d catalyst ~N~\tf) n-BuOH • (1) 1 a template-type synthesis, utilized Ni(II) or Cu(ll) acetate to facilitate the condensation, affording initially the metal complex~' which was subsequently treated with KCN to produce free binucleating ligand, 2. "' Attempts to prepare ligand 2 via alkoxide catalysis resulted in the "' production of significant amounts of blue-purple pigment, possibly phthalocyanine impurities. NC)§(CN NC CN + 4 ~N J6J N R CoCI 2 ~ n-BuOH, 6 2 117 R NC:©(CN NC 0 + CN d.._ M(0Ac)2 M=Cu,NI N CH30H,Il 4~\;1) HtJ (3) A series of binucleating ligands was prepared from tetracyano- benzene and various alkyl-substituted 2-aminopyridines for the purpose of obtaining a ligand with acceptable solubility in organic solvents. The bridging ligands are all high melting, yellow crystalline solids and with the exception of the 4' -g-propyl and 4' -tert-butyl derivatives, which are somewhat hygroscopic, have long shelf lives. Solubilities vary over a range of two orders of magnitude as shown in Table I, with the 4' -sec-butyl derivative being the most soluble; unless indicated otherwise, all work described in this paper was carried out with the 4' -sec-butyl substituted ligand, HL-LH. The rather complex infrared spectra of the bridging ligands contain characteristic bands in the 1650-1500 cm- 1 region, ca. 1640 (m-s) and 1590 (s) (also a band at ca. 1540 (m) observed for the 4' -alkylsubstifuted ligands), which undergo changes upon chelation of a metal ion. The infrared spectrum of the chelated ligand exhibits only a much weaker band at 1640 em - 1 and a shift of the 1590 em - 1 band to lower energy; two new bands appear at ca. 1610 and 1525-1515 cm- 1 • These 118 Table I. ~ Yields, Melting Points, and Relative Solubilities for the Binucleating Ligands, 2. "' Cpd. Derived from (amine) Yield Melting Relative Point ( •c) Solubility 2a 2-aminopyridine 90 324-6 2.6 2b 2-amino-4-methylpyridinea 59 340-1 1 2c 2-amino-5-methylpyridine 79 345-8 (dec) ND 2d 2-amino-3, 5-dimethylpyridine NDb 375-7 ND 2e 2-amino-4-ethylpyridine 45 353-5 72 2f 2-amino-4-g-propylpyridinea 58 357-9 52 2g 2-amino-4-sec-butylpyridine 83 369-71 105 2h 2-amino-tert-butylpyridinea 44 435-7 1.7 2i 2-amino-4-g-amylpyridine 49 317-18 11 aLigand isolated as a monohydrate. bND = not determined. 119 generalizations appear to be true for a variety of substituted ligands and metal ions. Proton NMR spectra were obtained only for the most soluble of the 4' -alkyl-substituted ligands. The benzo protons appear as a singlet at 8. 60-8. 70 ppm (CDC!s) and the pyridyl protons are shifted in the general order of H 5 < H 3 < H 6 as reported earlier for the mononucleating ligands 1. 5 "' Synthesis and Characterization of Mononuclear Metal Com !exes. Mononuclear divalent metal complexes of the type ML(OAc) 3 (M = Mn, Fe, Co, Ni, Cu, and Zn) are readily formed upon treatment of the free ligand with excess metal acetate in alcohol (eq 4). The reaction is rapid at ambient temperature, affording highly colored crystalline solids. All were air and thermally stable in the solid state, although Mn(II), Fe(II), and Co(II) complexes were, to some extent, air sensitive in solution; the latter were prepared under an inert atmosphere. The Co(II), Ni(II), Cu(II), and Zn(II) complexes were also prepared by a template reaction directly from phthalonitrile, amine, and metal salts (eq 5). 6 -->~ ML(OAc) (4) ~CN ~CN (5) 120 With the exception of the Zn(TI) complex, all of the complexes, ML(OAc), were paramagnetic. The infrared spectra are dominated by ligand bands; changes observed in the spectrum of the mononucleating ligand upon coordination to a metal ion are analogous to those described above for the bridging ligand. Treatment of a divalent transition metal salt (Mn, Fe, Co, Ni, Cu, Zn) with 2 equiv of chelating ligand afforded the neutral complex ML 2 in high yields, as indicated in eq 6. :MX 2 + 2LH - ML2 + 2HX. (6) For metal salts with very poor bases as counterions, e. g., c1o;, an amine base was usually added to facilitate deprotonation of the ligand; however, with Cl- or OAc-, no external base was required. Alternatively, the M~-type complexes could be prepared in a two-step process by subsequently treating the product of eq 4 with a second equivalent of chelating ligand, as shown in eq 7. The two-step ML(OAc) + LH --+ M~ (7) approach allows the preparation of complexes with two different ligands, as indicated in eq 8. ML(OAc) + L'H --+ MLL' • (8) The synthesis of such unsymmetrically substituted (chelated) complexes was.usually carried out under conditions whereby the product precipitated from solution, thus minimizing the possiblity of ligand scrambling. Evidence that asymmetric synthesis (eq 8) could be carried out for 121 Co(II) without ligand scrambling was obtained from cyclic voltammetric measurements on the mixed ligand complex Co(4' -MeL)(5'-ClL). The complex Co(4' -MeL)(5' -ClL) exhibited a single wave at 0. 081 V vs. NHE, whereas a mixture would have given two waves at -0.094 V vs. NHE for Co( 4' -MeL) 2 and at 0. 235 V vs. NHE for Co(5' -ClL) 2 • The observation that ligand exchange could occur in solution, even at room temperature, was made from a series of metal-exchange experiments. Cupric ion readily replaces substitution-labile metal ions such as Co(II), Ni(II), and Zn(II) from complexes of types ML(OAc) and ML 2 • However, no exchange occurs with cupric ion and substitution inert Co(III) in Co( 4' -MeL) 2 PF 6 • The ML 2 complexes of metal ions with unfilled d shells are all dark, intensely colored crystalline solids. In general, theM~ complexes have much greater solubilities in organic solvents than the corresponding ML(OAc) complexes; typically, they are soluble in solvents of low polarity such as cyclohexane or toluene, but only sparingly soluble in very polar solvents such as methanol. 7 Infrared spectra are essentially identical to spectra of the corresponding ML(OAc) complexes. The NMR spectra, magnetic properties, and electrochemistry are discussed in a later section. The Co(II) complex, Co(4' -MeL) 2 , was oxidized with eerie ion to the corresponding Co(III) complex, diamagnetic Co( 4' -MeL):, and isolated as its BF;, BPh;, and PF; salts. Attempts to similarly prepare Mn( 4' -MeL): were unsuccessful. Complexes of Fe(IIT) could be prepared by ferricenium oxidation of Fe(4' -MeL) 2 ; however, they 122 were unstable in polar solvents, apparently dissociating one of the tridentate ligands. Synthesis and Characterization of Binuclear Com !exes. The preparation of homonuclear complexes with the binucleating ligand HL-LH, ~' parallels the preparation of mononuclear complexes with the mononucleating ligand !· The simplest binuclear complexes were obtained by treatment of the bridging ligand with an excess of metal acetate, as shown in eq 9 [M = Mn(II), Fe(II), Co(II), Ni(II), Cu(II), and Zn(II)]. 3 The same binuclear complexes of Ni(II) and Cu(II) were HL-LH +excess M(OAc) 2 - (AcO)ML-LM(OAc) (9) obtained via the template route shown in eq 3, although the purity of complexes obtained via the template route was lower. The relatively low solubility of the (AcO)ML-LM(OAc) complexes made purification using solution techniques impractical, and accordingly eq 9 vras the preferred route. The new complexes are intensely colored and have high thermal stability. The characteristic changes observed in the infrared spectrum upon coordination of ligand 2 with metal ions was "' discussed earlier. The binuclear complexes, (OAc)ML-LM(OAc) (3), react readily "' with mononucleating ligand, LH (1), as shown in eq 10, a procedure "' analogous to the formation of MLL' described by eq 8. The preparation (AcO)ML-LM(OAc) + 2LH 3 "' LML-LML 4 "' (10) 123 HL-LH + 2ML(OAc) - LML-LML (11) 4 of this new series of binuclear complexes, 1_, can also be approached by treating the bridging ligand with 2 equiv of ML(OAc) according to eq 11. The two metal ions in the complexes LML-LML (4) are encap" sulated by organic ligands similar to M~ and accordingly have significantly increased solubility in organic solvents when compared with the binuclear complexes ~' allowing the application of solution techniques for characterization. The monomeric nature of these complexes is suggested by a molecular weight determination (in toluene) on (( 4' -sec-butyl)L)NiL-LNi(( 4' -sec-butyl)L) which yielded a value of 1699 compared with a calculated value of 1680. Attempts to prepare (4'-MeL)FeL-LFe(4' -MeL) by the approaches outlined above (eq 10 and 11) were unsuccessful. The mononuclear metal complex, Fe(4' -MeLH)Br 2 (coordinated 4'-MeLH is protonated presumably at one of the imine nitrogens), obtained from the reaction of 4' -MeLH with excess ferrous bromide in alcohol, proved a useful starting material for preparing (4'-MeL)FeL-LFe(4'-MeL). Treatment of deprotonated binucleating ligand with an excess of Fe( 4' -MeLH)Br2 yielded the desired binuclear Fe(II) complex. 'The infrared spectra of binuclear M(II) complexes are very similar to those of M~. Proton NMR spectra and magnetic and electrochemical properties are discussed below. The cobalt complex ( 4' -MeL)CoL-LCo( 4' -MeL) was successfully 124 oxidized with eerie ion to the corresponding Co(III) complex and isolated as its BPh; and PF; salts. Magnetic Susceptibility. Table II gives values of lleff for both mononuclear and binuclear M(ll) complexes, all of which are high spin. Magnetic measurements have been made on (AcO)CuL-LCu(OAc) over a range of temperatures from liquid helium to room temperature. The results of these measurements show a very slight metal-metal 1 interaction with J = -1 cm- • 8 In general, the magnetic moments of binuclear complexes are slightly lower per metal ion than the moments of the corresponding mononuclear complexes. It is unlikely that these lower values result from magnetic coupling but are probably attributable to trace impurities. ectra. ~~~~~~~~~~~~~ The visible spectra of all mono- nuclear and binuclear complexes show intense (E '""" 20-50, 000) ligand absorptions in the range 400-450 nm. All of the complexes are yellow in dilute solution as is the ligand. For the metal complexes, the high energy ligand absorptions are fairly broad so that concentrated solutions ('""" 10 mM) are nearly black, and ligand field absorptions could not be identified. The mononuclear iron (II) complex shows two broad absorptions at 650 nm and 740 nm (E '""" 800) in addition to the ligand absorptions. These bands may be assigned, due to their position and intensity, as Fe(II) - L charge transfer bands. Proton NMR S ectra. Proton NMR spectra were obtained for the mononuclear Fe(II), Co(II), Co(III), Ni(II), and Zn(II) complexes, 125 M( 4' -MeL) 2 , to identify the complexes and to test their purity (Table III). Spectra of the Mn(II) and Cu(II) complexes are severely broadened due to the slow electron-relaxation times of these metals. In the diamagnetic complexes, the protons could be assigned with no ambiguity from the splitting pattern of the peaks. In contrast, peak widths for the paramagnetic complexes of approximately 0. 2 ppm totally overwhelmed any splitting. For this reason assignment was more difficult and less reliable. In general, N1\1R spectra of all the paramagnetic complexes are very similar. The protons in the paramagnetic species were assigned by using integration data and by trends observed in the diamagnetic complexes and in other similar paramagnetic complexes. g The magnitude of the shift in the paramagnetic molecules was found to be dependent on the metal involved. Among the paramagnetic complexes, the protons a to the nitrogen of the pyridine ring, Hd' were observed only for the Co(II) complex and appear far downfield at 106 ppm. This resonance is extremely broad due to the nearness of these protons to the paramagnetic center. Similar behavior was 2+ 10 reported for Co(phen) 3 • The binuclear products gave very complex spectra. The diamagnetic Co(III) complex, however, gave a particularly well resolved spectrum (Figure 1), permitting analysis. Integrations of sec-butyl proton_s and methyl protons gave the expected ratio for one binucleating ligand and two mononucleating ligands per complex, helping to substantiate that this complex is binuclear and not a higher order polymer. The aromatic protons were difficult to assign due to overlapping peaks in this region of the spectrum. 126 Table II. Magnetic Susceptibility Data for Mononuclear and ~ Binuclear Complexes per Metal Ion. Complex P.eff (B. M. 298 K) Mn( 4' -MeL) 2 5.92 Fe( 4' -MeL) 2 5.12 Co( 4' -MeL) 2 4.85 Ni( 4' -Me L) 2 3.10 Cu( 4' -sec-butylL) 2 1. 85 (4' -MeL)MnL-LMn(4' -MeL) 5.66 (4' -MeL)FeL-LFe(4' -MeL) 5.13 (4' -MeL)CoL-LCo(4' -MeL) 4.65 (4' -MeL)NiL-LNi(4' -MeL) 3.07 (4' -MeL)CuL-LCu(4' -MeL) 1. 83 127 ~ Proton NM:R Data for Mononuclear Complexes in CDC!s. Values are given as 5, ppm, vs. TMS at 34 •c. Hd He CH 3 4'-MeLH Zn(II) Co(III) Co(II) Ni(II) Fe(II) Ha 8.01 8.04 8.13 26.5 13. 0 11. 0 Hb 7.54 7.55 7.76 26.0 11. 9 9.0 He 13.97 Hd 8. 40 7.93 7.37 106 He 6.83 6.35 6.62 35.6 33. 1 31. 6 Hr 7.23 7.07 7.10 43.7 51. 1 41.5 CH 3 2.30 2.07 2.31 -15.7 -6. 3 -24.8 Proton 128 ~ Proton NMR spectrum of (4'-MeL)CoL-LCo(4'-MeL)(Cl04 ) 2 in CDCJ.s at 34•c. Integration of the sec-butyl protons (0. 8 and 1.1 ppm) and the methyl protons (2. 3 ppm) provide evidence for describing the complex as binuclear. 129 0 C\J ,__ ro - E ~ Q. Q. ._,__ -+- '+-..c U) 0 (.) c.o -- E Q) ..c (_) 2. 04 1.90 100 160 o. 182 -0.045 (4' -MeL) FeL-LFe(4' -MeL) (4' -MeL)CoL-LCo(4' -MeL) -1.664 -1.435 -1. 477 -1. 455 -1.494 -1.418 -1.650 -1.640 -1.642 -1.600 -1.660 -1.680 -1.420 -1.470 -1.680 -1.652 -1. 460 -1.438 Ligand waves, Ef .£Potentials were obtained by differential pulse voltammetry due to low solubility. ~Number of electrons involved in metal oxidation determined by constant potential electrolysis • ~Anodic peak to cathodic peak separation of metal wave in cyclic voltammetry given in millivolts. --- -1. 190 2.04 240 0. 323 (4'-MeL)MnL-LMn(4'-MeL) 2.10 -- -- -- Zn(4' -MeL) 2 o. 976.£ -1. 185 1. 01 90 0.954 Ni(4'-MeL) 2 (4'-MeL)NiL-LNi(4'-MeL) -1.204 0.95 90 -0.094 Co(4'-MeL) 2 -1.173 0.99 n~ 65 ~ 0.130 Pc Fe(4'-MeL) 2 -E 1. 00 Pa 90 E 0.310 Ef Mn(4'-MeL) 2 Complex Measured in DMF (vs. NHE). M(II/III) -1.758 Table IV. Electrochemical Potentials for Mononuclear and Binuclear Complexes as ~ 1-.£ 0 (,.) 131 ~· Cyclic voltammograms of DMF solutions: upper figure Co(4' -MeL) 2 ; lower figure Ni(4' -MeL) 2 • 132 ;rA 1 1 Ni (4 -MeL)2 -1.2 -0.4 Volts vs. NHE 0.4 133 The use of NMR spectroscopy to investigate intramolecular electron transfer in the mixed-valent Con Colli complex "WaS attempted. Equilibrium concentrations of (4' -MeL)CoL-LCo(4' -MeL)+ were generated by mixing equimolar amounts of (4' -MeL)CoL-LCo(4' -MeL) 2 and (4' -MeL)CoL-LCo( 4' -MeL) +. The mixture gave a spectrum that is very similar to the composite spectrum of (4'-MeL)CoL-LCo(4'-MeL) plus (4' -MeL)CoL-LCo( 4' -MeL) 2 + but with several additional peaks. One of the peaks, which is very broad, sharpens into a single peak at -1.5 ppm at higher temperatures (- so·c) and resolves into two peaks at 4. 3 ppm and -10. 7 ppm as the temperature is lowered (...... -so•c). This behavior is typical of a process whose rate is on the order of the NMR time scale. However, due to the complexity of these spectra and the chemical structures involved, it was not possible to assign this process to intramolecular electron transfer. Electrochemistry. Electrochemical measurements have been ~ made on all mononuclear and binuclear complexes (Table IV). Most complexes exhibited quasi-reversible behavior except for Cu( 4' -sec-butyl L) 2 and all complexes having a single tridentate ligand, M( 4' -MeL)OAc, for which both anodic and cathodic cyclic voltammetric waves could not be obtained. Cyclic voltammograms of mononuclear complexes, M(4'-MeL) 2 (e.g., Figure 2) consist of a metal oxidation wave apd two reductions attributable to the ligand. The anodic and cathodic peak currents for the M(III)/M(II) couple are equal but the peak pctential separation is larger than the expected 58 mV for a reversible one-electron process. 11 Coulometry at potentials anodic of 134 the metal oxidations verify that these are one-electron processes with n. = 1. 0 ± 0. 05 . The metal reduction potentials of the binuclear complexes are very similar to potentials measured for the mononuclear complexes. The metal oxidation waves in the binuclear complexes are quite broad (Figure 3) and constant potential electrolyses anodic of these waves yields !!. values of 2. 0 ± 0.1. This behavior is consistent with two closely spaced one-electron oxidations resulting in one broad wave. Similar behavior has been observed for some ruthenium dimers. 12 In general, the factors determining the pctential difference between two consecutive redox processes in binuclear complexes are complex. 13 A separation of the two one-electron waves by 100 mV or less sets a range for the constant describing the comproportionation equilibrium (eq 12; 4 < Kcom ~ 50). The lower limit of 4 corresponds to comM(II) M(II) + M(III) M(III) + ~ 2M(II) M(III)+ 2 (12) pletely noninteracting metal centers, in which case one would expect a 14 single electrochemical wave with EPa - EPc =58 mV. The electrochemical results indicate that the binuclear complexes undergo a two-electron metal oxidation with the two electrons being transferred at nearly the same potential. One would then expect longer chain polymers to undergo multielectron oxidation with all of the electrons transferred at approximately the same potential. This may be a very desirable property for utilization of these complexes as multielectron-transfer catalysts. 135 ~ Cyclic voltammograms of DMF solutions: upper figure Co( 4' -MeL) 2 ; lower figures (4' -MeL)CoL-LCo( 4' -MeL). 136 Co (4'-MeL)2 1 (4 -Mel) CoL L Co (4'-Mel) -1.2 -0.4 Volts vs. NHE 0.4 137 Cyclic voltammograms of the binuclear complexes also show three ligand reductions. Two of these reductions appear at approximately the same potential as reductions observed for the mononuclear complexes. The third ligand reduction appears at a more positive reduction pctential and is presumably due to reduction of the bridging ligand. The binuclear Fe(II) complex shows a fourth reduction at even more negative potentials than the ligand reductions. The product of this reduction has not been investigated. A series of mononuclear Co(II) complexes containing various substituents on the pyridine ring of the ligand were prepared in order to determine the effect of these substituents on the metal reduction potential (Table V). The results of these measurements show that the effect of the substituents is significant. A shift of nearly 700 mV is observed in going from a 4' -methyl to a 5' -nitro substituent. A plot of the reduction potential vs. the Hammett parameter of the substituent is approximately linear (Figure 4), as has been observed in certain other ligand systems. 15 Thus, by changing the substituents, a large shift in reduction potential can be produced, making it possible to design a complex having specific redox chemistry. Accordingly, the reactivity of these complexes with various substrates might be significantly altered by the effects of these substituents on the reduction potential. 138 ~ Cobalt (III)/Cobalt (II) Reduction Potentials, Ef, for Several Co(RL) 2 • a Complex Co(III) /Co(II) Co(4' -MeL) 2 -0.094 v CoL2 -0.031 v Co( 4' -MeL)(5' -ClL) 0. 081 v Co(5' -ClL) 2 o. 235 v Co(5' -N0 2 L) 2 0. 589 v aPotentials were measured in DMF and are reported vs. NHE. 139 ~ Half v;ave potentials vs. ap for mononuclear Co(II) complexes containing various substituents on the isoindoline ligand. 0.0 0.2 Ef 0.4 0.6 -0.2 • 0.0 Co (4'- Mel) 2 e/ 0.2 a- 0.4 0.6 Co (5'-CI L} 2 Co(4'-MeL)(5~CI L} • Co (5'-N02 L}2 • 0 ...... ~ 141 EXPERIMENTAL SECTION Synthesis. All organic ligands gave satisfactory carbon, ~ hydrogen and nitrogen analyses, while metal complexes gave satisfactory carbon, hydrogen, nitrogen, and metal analyses. Mononuclear Chelati physical and spectral data for all the mononuclear chelating isoindoline ligands, RLH, employed in this study were reported earlier. 5 Binucleatin L · ands. All of the binucleating ligands reported in Table I were prepared by the general procedure illustrated below for the 4' -sec-butylpyridyl derivative ,.._,.... 2g. Ligands with significantly lower solubility were purified by recrystallization from nitrobenzene or quinoline. 1, 3, 5, 7 -Tetra(2-( 4-~-butylpyridyl)imino)benzodipyrrole (2g). A lL roundbottom flask was charged with 5. 81 g (0. 033 mol) of 1, 2, 4, 5-tetracyanobenzene, 22.5 g (0.15 mol) of 2-amino-4-sec-butylpyridine, 5 2. 61 g of anhydrous calcium chloride, and 500 mL of methanol. The mixture was stirred at ambient temperature for 7 days and then warmed gradually to reflux temperature. After 2 days at reflux, the methanol was allowed to distill off and the solvent volume was maintained by the gradual addition of n-butyl alcohol. The suspension in butyl alcohol was heated at reflux for an additional 7 days. The mixture was allowed to cool and then filtered. The green crystalline solid was washed with methanol and dried in vacuo to a weight of 21. 35 g. The crude solid was dissolved in methylene chloride, treated with activated charcoal, Norit A, overnight, and passed through a 142 column of celite (top) and silica gel (bottom). After the volume of eluate was reduced, 17.10 g of yellow crystalline product, 369-71•c, was obtained. With further volume reduction and addition of hexane, an additional 1. 70 g of less pure ligand was also obtained for a yield of 76%. NMR analysis of the product gave the following peaks: sec-butyl protons - 0. 90 (t), 1. 31 (d), 1. 65 (quintet), 2. 64 (sextet); pyridine protons - 6. 92 (d), 7. 31 (s), 8. 5 (d); benzene protons - 8. 77 (s); pyrrole protons - 14. 08 (br, s). Template Synthesis of the Binucleating Ligand (2). ~~~~~~~~~~~~~~~ A mixture of 89 mg (0. 5 mmol) of 1, 2, 4, 5-tetracyanobenzene, 330 mg (2. 2 mmol) of 2-amino-4-sec-butylpyridine, 200 mg (1. 0 mmol) of cupric acetate hydrate, and 10 mL methanol was stirred at ambient temperature for 5 days followed by 3 days at reflux. After the solution was cooled, the solvent was allowed to evaporate and the residue was washed with water and dried to afford 430 mg of green powder. To this was added 5 mmol of KCN (325 mg), 20 mL of ethanol, and 10 mL of chloroform. Gentle heating, below the reflux temperature, was applied for 2 days. Upon being cooled, the green suspension was filtered and the insoluble material washed with chloroform. The combined filtrate washes were evaporated to dryness and the residue extracted with chloroform. The chloroform extract was passed through a minicolumn of alumina which removed most of the green pigment. From the eluant, a 45% yield (167 mg) of greenish yellow crystals was obtained which were identical spectroscopically to the ligand obtained via the CaC12 -catalyzed route. 143 Solubilit of Binucleati An excess of binucleating ligand was added to 25 mL of benzene and warmed gently. The supersaturated solutions were allowed to stand at ambient temperature (23 ± 0. 5•c) for at least 1 week before analysis. Aliquots were removed periodically, passed through a filter (Millipore LS, 5. 0 /.lM), and, after appropriate dilution, were analyzed spectrophotometrically on a Cary 17D spectrophotometer. Using previously determined molar extinction coefficients, concentrations were calculated. Analyses were repeated periodically until concentrations had stabilized; relative solubility values are reported in Table I. Pre aration of (Acetato) 1, 3-bis(2- yridylimino isoindolinato - for the preparation of M(RL)OAc complexes of divalent Mn, Fe, Co, Ni, Cu, and Zn. The preparation of the Mn(II), Fe(II), and Co(II) complexes was carried out under an argon atmosphere; no effort was made to exclude air from the other preparations. One millimole of chelating ligand and 2 mmol of metal(II) acetate hydrate in 15 mL of methanol was stirred at ambient temperature for 24 h. After this time the suspension was filtered and the solid washed with methanol and dried in vacuo. Yields of complexes were generally 90% or higher. The colors of the M( 4' -MeL)OAc complexes are as follows: Mn(II) tan, Fe(II) green, Co(II) gold-tan, Ni(II) green, Cu(II) brown, and Zn(II) yellow. Preparation of Fe(4' -MeLH)Br2 • Under a helium atmosphere, a solution of 2. 0 g (6.1 mmol) of 4'-MeLH in 40 mL of warm methanol was added to a solution of 2. 6 g (9. 0 mmol) of FeBr2 • 4H 2 0 in 10 mL of 144 methanol. The reaction mixture was heated with stirring for 10 min. After the mixture was cooled to ambient temperature, 20 mL of diethyl ether was added to the mixture and stirred for 5 min. The green crystalline product (65% yield) was collected by vacuum filtration, washed with diethyl ether, and dried under a stream of helium. Preparation of Mononuclear Metal(II) Complexes, M(RL) • The following general procedure was used to prepare mononuclear complexes of Mn(II), Fe(II), Co(IT), Ni(II) and Zn(II) with RL = 4' -MeL. Preparations of iron and cobalt complexes were conducted in an inert atmosphere. A solution of 0. 7 mmole of the metal(II) perchlorate in 5 mL of methanol was added to a solution containing the mononucleating ligand (1. 5 mmoles) and 0. 5 mL of triethylamine in 40 mL of hot methanol. The reaction was heated at reflux for 30 min during 'Which time a dark microcrystalline solid had formed. After being cooled, the mixture was filtered and the product was washed with hot methanol. The complexes CoL2 and Co(5' -N02 L), were prepared in the same fashion. The yields were: Mn(4'-MeL) 2 74%, Fe(4'-MeL) 2 77%, Co(4'-MeL) 2 7g%, Ni( 4' -MeL) 2 75%, Zn( 4' -MeL) 2 83%, Co~ 91%, and Co(5' -N02 L) 2 12%. Preparation of Bis(1, 3-bis({5-chloro-2-pyridyl)imino) isoindolinato)cobalt(II). To 368 mg (1. 0 mmol) of ligand and 119 mg (0. 5 mmol) of cobaltous chloride hexahydrate under argon was added 10 mL of methanol and 1 mL of triethylamine. After being stirred for 16 hat 25•c, the mixture was filtered in the air and the product was 145 washed with methanol and dried in vacuo. A 97% yield (383 mg) of red-brown microcrystals was obtained, m. p. > 350• C. Preparation of Bis(1, 3-bis(( 4-sec-butyl-2-pyridyl)imino)iso[Cu(4' -sec-butylL) • A solution of 34 mg (0. 25 mmol) of anhydrous cupric chloride in 10 mL of methanol was added to 210 mg (0. 51 mmol) of the chelating ligand, 4' -sec-butylLH, and stirred at ambient temperature. 7 The initial green solution changed to a yellow-brown suspension after a few minutes. After 15 min, 0. 2 mL of triethylamine was added and the stirring was continued. After 45 min the suspension was filtered and the solid was washed with methanol and dried in vacuo. An 84% yield (186 mg) of gold microcrystals was obtained, mp 263-264. 5•c. Conversion of Co( 4' -MeL)OAc to Co( 4' -MeL) . A Schlenk tube was charged with 110 mg (0.25 mmol) of Co(4'-MeL)OAc, 88 mg (0. 27 mmol) of chelating ligand, and 5 mL of pyridine. The dark red solution was stirred under argon at ambient temperature for 24 h. Water (4 x 5 mL) was added, and the resulting precipitate was collected, washed with water, and dried in vacuo to afford a quantitative yield of red-brown crystalline Co(4' -MeL) 2 , mp 337-8• C. Pre aration of [1 3-bis((4-methyl-2- ridyl)imino isoindolinato - [1, 3-bis((5-chloro-2-pyridyl)imino) isoindolinato] cobalt(ll) [Co( 4' -MeL) (5'-ClL)]. To 184 mg (0.5 mmol) of 5'-ClLH and 222 mg (0.5 mmol) ~ of Co(4' -MeL)OAc in a Schlenk tube under argon was added 10 mL of methanol and 1 mL of triethylamine. The mixture was stirred at ambient temperature for 2 days and then filtered in the air. The solid 146 was washed with methanol and dried in vacuo to afford a 92% yield (348 mg) of brown microcrystalline powder. Pre ridyl)imino iso- indolinato)cobalt(III) Hexafluorophosphate [Co(4' -MeL) 2 PF ] . To 71 mg (0.1 mmol) of Co( 4' -MeL) 2 and 57 mg (0.1 04 mmol) of eerie ammonium nitrate, (NH 4 ) 2 Ce(N03 ) 6 , in a Schlenk tube under argon was added 5 mL of methanol. After a few minutes a clear red-brown solution was obtained. After stirring at ambient temperature for 16 h, the solution was transferred to an open beaker and the solvent was allowed to evaporate. The solid residue was washed with water. Methanol (5 mL) and 25 mg (0.15 mmol) of NH 4 PF 6 were added to the residue and the mixture was stirred until the solvent had evaporated. The residue was extracted with methylene chloride, and heptane was added in small amounts to the filtered extract. On standing dark crystals of the hexafluorophosphate salt were obtained as the methylene chloride solvate, Co( 4' -MeL) 2 PF 6 • 2CH 2 Cl2 , mp 256.5-257.5 •c, 89% yield. Anal: Calcd for: C 42H 36 N10 CoC~PF 6 : C, 49.14; H, 3. 53; N, 13. 65; Cl, 13. 82; P, 3. 02. Found: C, 49. 5; H, 3. 45; N, 13. 7; Cl, 13, 75; p' 3. o. Preparation of Bis(1, 3-bis(( 4-methyl-2-pyridyl)imino)isoindolipato)cobalt(ITI) Tetraphenylborate [Co(4' -MeL) 2 (Ph 4 B)]. A mixture of 44 mg (0.1 mmol) of Co(4' -MeL)OAc, 36 mg (0.11 mmol) of chelating ligand, 4' -MeLH, 41 mg (0.12 mmol) of sodium tetraphenylboron, and 6 mL of methanol was stirred for 24 h under argon 147 at ambient temperature. Then 77 mg (0.14 mmol) of eerie ammonium nitrate was added and the red-brown solution immediately changed to a brown suspension. After an additional 5 h at 25 •c, the suspension was filtered in the air and the solid was washed with methanol and with water. After drying in vacuo, an 84% yield (87 mg) of gold-brown microcrystals was obtained. Metal Exchange studies. (a) Ni( 4' -MeL)OAc and Cu(OAc) · H 0. To 55 mg (0.125 mmol) of Ni(4'-MeL)OAc and 250 mg (1. 25 mmol) of cupric acetate hydrate was added 3 mL of chloroform and 2 mL of methanol. After a few minutes a clear dark green solution was obtained and was stirred at ambient temperature for 24 h. The solvent was then allowed to evaporate, and the residue was washed with water and dried. The solid ·was extracted with methylene chloride and heptane was added in small aliquots to the extract. On standing, beautiful brown needles deposited. The crystals were collected, washed with heptane and ether, and dried to afford an 88% yield of Cu( 4' -MeL)OAc. (b) Ni( 4' -MeL and Cu(OAc · H 0. To 71 mg (0.1 mmol) of Ni( 4' -MeL) 2 and 200 mg (1. 0 mmol) of cupric acetate hydrate was added 3 mL of chloroform and 2 mL of methanol. The resulting dark green solution was stirred at ambient temperature for 24 h and then evaporated under a stream of argon. The residue was extracted with methyiene chloride and the solvent removed from the extract. The extract residue was extracted with benzene. From the benzene extract, 100 mg of green powder was obtained. The IR (KBr) spectrum was 148 identical to that of Cu( 4' -MeL)OAc. Analysis by thin layer chromatography (Si02 , EtOAc) of the powder showed no remaining Ni(4' -MeL) 2 • The yield of Cu(4' -MeL)OAc was 93%. (c) [Co( 4' -MeL) 2 ] PF6 and Cu(OAc) 2 • H2 0. A solution of 86 mg (0.1 mmol) of [Co(4' -MeL) 2 ] PF 6 and 200 mg (1. 0 mmol) of Cu(OAc) 2 • H2 0 in 3 mL of chloroform and 2 mL of methanol was stirred at ambient temperature for 24 h. The solvent was evaporated under a stream of argon and the residue was extracted with methylene chloride. After the methylene chloride was evaporated from the extract, the new residue was extracted with tetrahydrofuran. After addition of heptane to the tetrahydrofuran extract, dark crystals gradually deposited. The crystals (85 mg) were collected and dried in vacuo. The IR (KBr) spectrum was identical to that of the starting material. Thin layer chromatographic analysis (Si02 ; CHC~:EtOH, 10:1) of the product showed the absence of any Cu(4' -MeL)OAc, the potential exchange product. Preparation of (AcO)ML-LM(OAc) (3). The general procedure for the preparation of (AcO)ML-LM(OAc) complexes (~) is illustrated below for the binuclear Cu(II) complex. The preparation of the Mn(II), Fe(II) and Co(II) complexes was carried out under an inert atmosphere. To 745 mg (1. 0 mmol) of binucleating ligand, HL-LH, and 1 g (5. 0 mmol) of cupric acetate hydrate was added 50 mL of methanol, followed by 10 mL of chloroform. The suspension was stirred at ambient temperature for 72 h. The suspension was filtered and the solid was washed repeatedly with methanol until the washes were no 149 longer green. After drying in vacuo, a 97% yield (959 mg) of greenish yellow powder was obtained. Preparation of LML-LML (4). Binuclear complexes of Mn(II), Co(II) and NI(II) were prepared by two different methods. The two different preparations are illustrated below for Co( II). Preparation of (4' -MeL)CoL-LCo(4' -MeL). To 89 mg (0. 2 mmol) of Co( 4' -MeL)OAc, 74 mg (0. 1 mmol) of HL-LH, and 5 mL of methanol in a Schlenk tube under argon was added 0. 5 mL of triethylamine. After the dark brown suspension was stirred at ambient temperature for 20 h, an additional 10 mL of methanol was added. The suspension was filtered in the air and the solid was washed with methanol and dried in vacuo to afford a 91% (137 mg) yield of red-brown powder which exhibited a single spot on thin layer chromatography (Si02 , EtOAc). The Co(II) binuclear complex gave a 1 H NM:R spectrum in CDC!s with the following peaks (vs. TMS at 34.C): aromatic protons: 48. 6, 45.7, 37.0, 30.6, 29.1, and 27.5; sec-butyl protons: -4.3 and -7.6; and methyl protons: -13.7. A Schlenk tube was charged with 98 mg (0.1 mmol) of (AcO)CoL-LCo(OAc), 65 mg (0. 2 mmol) of 4' -MeLH, and 5 rnL of methanol; the mixture was stirred under argon at ambient temperature. After 15 min, 0. 5 mL of triethylamine was added to the brown suspension and the stirring was continued for an additional 16 h. An additional 10 rnL of methanol was added and the supension was filtered in the air. The dark brown powder was washed with methanol and dried in vacuo to afford a 77% 150 yield (116 mg). The m (KBr) spectrum was identical to that of material prepared by the first method. Pre Under a helium atmosphere, 160 mg (0. 79 mmol) of AgC10 4 was added to a slurry of 430 mg (0. 79 mmol) of Fe( 4' -MeLH)Br2 in 35 mL of methanol. After stirring briefly, the mixture was filtered to remove precipitated silver bromide. Toluene (15 mL) and a solution of 43 mg (0. 79 mmol) of sodium methoxide in 20 mL of methanol were added to the filtrate resulting in a green to reddish brown color change. To this solution was added slowly a solution of 200 mg (0. 27 mmol) of HL-LH and 29 mg of (0. 54 mmol) sodium methoxide in 5 mL of methanol and 15 mL of toluene. After stirring for 1 h, the dark green solid was collected by vacuum filtration, washed with several portions of methanol, and dried under a stream of helium; 79% yield. Preparation of [(4'-MeL)CoL-LCo(4'-MeL)](PF6 ) 2 • To a suspension of 45 mg (0. 03 mmol) of (4' -MeL)CoL-LCo( 4' -MeL) in 5 mL of methanol was added sufficient methylene chloride to produce a homogeneous solution. To this solution was added 33 mg (0. 06 mmol) of eerie ammonium nitrate, and the mixture was stirred for 0. 5 h at ambient temperature after which time the solvent was evaporated under a stream of argon. The residue was washed with water, dried in vacuo, and then extracted with methylene chloride. To the CH 2 C~ extract was added 11 mg (0. 07 mmol) of NH 4 PF 6 in 3 mL of methanol; after 0. 5 h of st~rring, the mixture was filtered to remove a small amount of brown precipitate and the solvent was evaporated under a stream of 151 argon. The residue was washed with water and dried in vacuo to afford 45 mg of brown powder. The powder was extracted with ethyl acetate, and heptane was added in small aliquots to the extract. On standing, dark brown microcrystals deposited. The crystals were collected, washed with toluene, and dried in vacuo to afford a 67% yield (26 mg) of ((4' -MeL)CoL-LCo(4' -MeL)] (PF 6 ) 2 • Preparation of (4' -sec-butylL)NiL-LNi(4' -~-butylL) and Molecular Wei ht Determination. To 245 mg (0. 25 mmol) of (AcO)NiL-LNi(OAc) and 226 mg (0. 55 mmol) of 4' -sec-butylLH was added 20 mL of chloroform and the resulting solution was stirred at ambient temperature for 72 h. Heptane was added in small amounts until crystals deposited. A 60% yield (250 mg) of dark red crystals was obtained. The solution molecular weight determination in toluene was carried out on the red crystals by Dr. M. Zinbo of the Ford Scientific Research Labs, using a Hitachi-Perkin Elmer molecular weight apparatus. This vapor-phase osmometry method gave an observed value of 1699 compared with a calculated value of 1680. (0. 033 mmol) of (AcO)CuL-LCu(OAc) and 22 mg (0. 066 mmol) of 4' -MeLH was added 5 mL of methanol and 0. 2 mL of triethylamine. The resulting suspension was stirred at 25 •c, as the color gradually changed from olive-green to yellow-gold. After 3 h the suspension was filtered and the solid was washed with methanol and dried in vacuo. A 48% yield (27 mg) of gold-colored powder was obtained; mp > 39o•c. 152 Attemrted Preparation of (4'-MeL)ZnL-LZn(4'-MeL). To 450 mg (1 mmol) of Zn(4' -MeL)OAc and 272 mg of (0. 5 mmol) of binucleating ligand, HL- LH, was added 200 mL of methanol and 20 mL of triethylamine; the resulting suspension was stirred at ambient temperature for 7 days. The reaction mixture was filtered, washed with methanol, and dried in vacuo to afford a 60% yield (458 mg) of yellow-orange microcrystalline powder; mp > 390eC. Carbon, hydrogen and nitrogen analysis of the product was satisfactory for the binuclear Zn(II) complex, however, the proton NM:R spectrum of the product gave a mononucleating ligand to binucleating ligand ratio of less than 2:1 expected for a discrete binuclear complex. This suggests that the product isolated contains a higher order polymer. ~ Tetrabutylammonium perchlorate, TBAP (Southwestern Analytical Chemicals), was dried in vacuo before use. Spectroquality acetonitrile and N, N-dimethylformamide, distilled under reduced pressure over 4-A molecular sieves, were used for electrochemical measurements. A Princeton Applied Research Model 173 potentiostat galvanostat coupled with a Model 179 digital coulometer and a ramp generator of our own design were used for constant potential electrolysis and cyclic voltammetry. A storage oscilloscope and an X-Y recorder were used to display the results. A Princeton Applied Research 174A Polarographic Analyzer was used in conjunction with an X-Y recorder for differential pulse voltammetry. 153 Constant potential electrolyses were carried out in a threecompartment H cell. The cell consisted of 25 mL sample and auxiliary compartments separated by a small center compartment. Each compartment was separated by a medium porosity sintered glass frit. A platinum gauze was used as the working electrode for electrolyses. Cyclic voltammetry and differential pulse voltammetry were carried out in a single compartment cell containing approximately 5 mL of solution. The working electrode was a platinum button electrode. For all electrochemical measurements the supporting electrolyte used was 0.1 M TBAP. The Ag/Ag + reference electrode consisted of a silver wire immersed in an acetonitrile solution containing 0. 01 M AgN03 and 0.1 M TBAP. The Ag+ solution and silver wire were contained in an 8 mm glass tube fitted on the bottom with a fine porosity sintered glass frit. The auxiliary electrode consisted of a coiled platinum wire. All measurements were made in a helium atmosphere. Small amounts of ferrocene were added to electrochemical solutions as an internal standard. Potentials for the complexes were measured vs. ferrocene. 16 The formal potentials were then adjusted to potentials vs. NHE with the assumption of a value of +0. 400 V for the ferricenium/ferrocene couple. Formal reduction potentials, Ef, were measured by cyclic voltammetry using the formula Ef = (EPa + EPc)/2. The potentials determined in this way are approximate in that the systems examined did not display strict reversibility and corrections were net made for diffusion coefficients. Reduction potentials measured by cyclic voltammetry and differential pulse voltammetry agreed to within± 10 mV. 154 Ph sical Measurements. Magnetic susceptibility measurements were obtained on samples at the ambient temperature using a Cahn Instruments Faraday balance, with HgCo(SCN) 4 as a calibrant. Diamagnetic corrections were made using Pascal's constants. Proton N1ffi spectra were obtained on Varian EM-390 and JEOL FX90Q NMR spectrometers with CDC~ as the solvent and tetramethylsilane as a reference. Electronic spectra were recorded on a Cary 14 spectrophotometer. Solution spectra were obtained using one centimeter matched quartz cells. 155 REFERENCES ~ 1. (a) California Institute of Technology, Pasadena, CA. (b) Ford Motor Company, Dearborn, MI. 2. (a) Elvidge, J. A.; Linstead, R. P. J. Chern. Soc. }1g, 5000. (b) Robinson, M. A. ; Trctz, S. I.; Hurley, T. J. Inorg. Chern. 1967, §., 392. ~ 3. Mononucleating ligands, !, will be denoted as RLH; substituents on the pyridine ring precede the abbreviation, e.g., 4'-MeLH, where the prime designates the position of the pyridine ring. Mononuclear metal complexes are of two types: complexes with metal to ligand ratio of 1:1, ML(OAc), where L is the anion of the ligand LH and OAc also is bound to the metal, and complexes with metal to ligand ratio 1:2, ML 2 , with two anionic tridentate ligands. HL-LH denotes the binucleating ligand, ~' with 4' -secbutyl groups on the pyridine rings. Binuclear complexes containing the dianion of the ligand are denoted (OAc)ML-LM(OAc) or(4'-MeL)ML-LM(4'-MeL) depending on the other ligands boond to the metals. 4. Siegl, W. 0. Inorg. Chim. Acta !_W, 25, L65. 5. Siegl, W. 0. J. Org. Chern. liTI., 42, 1872. 6. Siegl, W. 0. Inorg. Nucl. Chern. Lett. li7.1, 10, 825. 7. The Cu(II) complex, Cu(4' -MeL) 2 , is considerably more soluble in organic solvents than other M( 4' -MeL) 2 complexes. Solid samples of Cu( 4' -MeL) 2 usually contained small amounts of the ligand, 4' -MeLH, because the solubility of the ligand was similar 156 REFERENCES (continued) ~ to that of the metal complex. For this reason, the more soluble sec-butyl substituted ligand, 4' -sec-butylLH, was used in order to avoid coprecipitation of the Cu(II) complex and unreacted ligand. 8. Hendrickson, D. N., unpublished results. 9. Cramer, R. E.; Drago, R. S. J. Am. Chern. Soc. !JUQ, 92, 66. 10. LaMar, G. N.; VanHecke, G. R. Inorg. Chern. !JU.Q, ~ 1546. 11. Nicholson, R. S.; Shain, I. Anal. Chern. 1964, 36, 706. 12. (a) Tom, G. M.; Creutz, C.; Taube, H. J. Am. Chern. Soc. ~- !ill, 96, 7827; (b) Powers, M. J.; Meyer, T. J. Inorg. Chern. !,m, 17, 1785. 13. Gagne, R. R.; Spiro, C. L. J. Am. Chern. Soc. 1980, 102, 1443. 14. Flanagan, J. B. ; Margel, S.; Bard, A. J.; Anson, F. C. ~- J. Am. Chern. Soc. 1978, 100, 4248. ~- 15. For example, Addison, A. W.; Stenhouse, J. H. Inorg. Chern. 1978, 17' 2161' ~- 16. Gagne, R. R.; Koval, C. A.; Lisensky, G. C. Inorg. Chern. ~' 19, 2854. 157 APPENDIX 2 Synthesis, Characterization, and Dioxygen Reactivity Studies of Mononuclear and Binuclear Iron Complexes of 1, 3-Bis(2-pyridylimino)isoindolines 158 Synthesis, Characterization, and Dioxygen Reactivity studies of Mononuclear and Binuclear Iron Com !exes of 1, 3-Bis(2-pyridylimino)isoindolines lliTRODUCTION ~ Reactions in which two or more electrons are transferred from a metal complex to a small molecular substrate are of considerable interest. Such multielectron reactions are potentially useful in chemical synthesis, solar energy conversion, and fuel cells. One reason for studying binuclear complexes of a binucleating analog of 1, 3-bis(2-pyridylimino)isoindoline was their potential use as multielectron reactants. In preparing and characterizing mononuclear and binuclear complexes of 1, 3-bis(2-pyridylimino)isoindolines, it was observed that preparations of a binuclear Fe(II) complex appeared to react with dioxygen. 1 Evidence for such reactivity was a distinct color change by solutions of the complex (N, N-dimethylformamide (DMF) or dichloromethane/methanol) upon exposure to dioxygen. Under similar conditions, the analogous Fe(II) mononuclear complex was unreactive. Additionally, related mononuclear and binuclear complexes of other transition metals were unreactive towards dioxygen. Although dioxygen reactivity of the binuclear Fe(II) complex was seemingly anomalous with respect to the behavior of related complexes, further investigation was warranted because of the possibility that the reaction involved mu"!tielectron transfer. As a result of this investigation, it was determined that the 159 original preparations of the binuclear Fe(II) complex were not sufficiently pure for reliable dioxygen reactivity studies. An alternative preparation was developed by which analytically pure samples of the complex could be prepared. Synthesis, magnetic susceptibility, and electrochemical behavior of this complex have been reported previously. 2 Some additional characterization is reported here as well as syntheses and characterization of several unreported mononuclear and binuclear iron complexes of 1, 3-bis(2-pyridylimino)isoindolines. Experiments were conducted to examine the dioxygen reactivity of solutions of analytically pure samples of the binuclear Fe(TI) complex. No evidence for such reactivity was observed. Experimental details and a discussion of the results are also reported here. 160 RESULTS AND DISCUSSION Synthesis and Characterization. The binuclear Fe(II) complex which was purported to react with dioxygen incorporates both mononucleating and binucleating 1, 3-bis(2-pyridylimino)isoindoline ligands, !. and ~· The original samples of this complex, N~ ©(NH N~ 1 ,... 4' -MeLH, 3 R = 4-CH 3 2 "' HL-LH, 3 R =4' -sec-butyl (4' -MeL)FeL-LFe(4' -MeL), were prepared by the reaction of HL-LH with 2 equiv of Fe(4' -MeL)(OAc) 2 under basic conditions (eq 1). HL-LH + 2 (Fe( 4' -MeL)(OAc)] base) (4' -MeL)FeL-LFe( 4'MeL) + 2H+ (1) Magnetic susceptibility measurements yielded values significantly lower than the expected value of 5.12 B. M. 2 Additionally, microanalytical data for these samples were not in agreement with the 161 calculated values for (4' -MeL)FeL-LFe(4' -MeL). Instead, the best fit for the data was a mixture of (4' -MeL)FeL-LFe{4' -MeL), Fe(4'-MeL)(OAc), and HL-LH. Because these experimental data were indicative of a product of unacceptable purity, an alternative synthetic approach was developed. The complex, (4' -MeL)FeL-LFe(4' -MeL), was successfully prepared by the reaction of HL-LH with 2 equiv of Fe( 4' -MeLH)Br2 which was first dehalogenated with Ag+ (eq 2 and 3). 2 For products of [Fe(4'-MeLH)Br2 l + Ag+ ___, [Fe(4'-MeLH)Br] + AgBr (2) HL-LH + 2(Fe(4'-MeLH)Br] + 4NaOCH 3 -+ {4'-MeL)FeL-LFe(4'-MeL) + 4CH 3 0H + 4Na+ + 2Br- (3) this reaction, the average I-Leff(298 K)/Fe(II) value, 5.13 B.M., and microanalytical data are within experimental error of the expected 1 values. Additionally, the H NMR spectrum (Table I and Figure 1) is consistent with the expected formulation. Specifically, the 3:1 ratio of methyl protons (Hg) to the methine protons (Hg') establishes that the complex is a discrete binuclear species. A lower ratio would be indicative of polynuclear contamination. The binuclear Fe(II) complex, (4' -MeL)FeL-LFe(4' -MeL) was oxidiz~d with 2 equiv of ferricenium tetrafluoroborate ((Cp 2 Fe)BF 4 ) in methylene chloride, with the corresponding binuclear Fe(Ill) complex being isolated as a methylene chloride sesquisolvate of the bis-tetrafluroborate salt (eq 4). The same method was also used to oxidize Fe( 4' -MeL) 2 to Fe( 4' -MeL) 2 BF 4 (1)· Both mononuclear and binuclear 3 ~ 4'-MeL CH 'Q N II N 2 3 CH§HCH ~ ~N N fi CH 3CHCHt:~ aValues are given in ppm vs. Me 4 Si. bNO = not observed. N N ~:a Hd 'o''He ~ Hg L-L ~ . Ha' N 'He• II CH3CHCHi=H3 NV N ~·:.o-Hd' ~ '' l CH 3CHCH 2CH3 Hf' Hg' Hh' ~· Table I. Proton NMR Data for (4' -MeL)FeL-LFe(4' -MeL) in CDC13 ~ 2.27 0.48 Hh' fl., 1 -7.54 Hg' -25.33 Hg 42.14 He 0.48 32.80 Hd' Hf' 32.80 Hd 42.14 NO He' He' .... NOb He 0) N 11.31 9.99 9.29 - Chemical shift a Hb Ha' Ha Assignment 40 1 ~ 30 --' 20 I -- -r- to ppm, 0 -----~-.----- _ l \ il"''cHC~ Hh' I r o Hg' -to I_ 1 - -2o I - ---1 Hg H NMR spectrum of (4'-MeL)FeL-LFe(4'-MeL) in CDCts. 1 Hd and Hd' ...__ _ _ _ _ _ _ I He and He' Hb Ha' KTMS +- Hr and H 1, -3o l ~ ..... 164 CH C~ (4' -MeL)FeL-LFe(4' -MeL) + 2(Cll2Fe)BF4 - - -2- - . . [(4' -MeL)FeL-LFe(4' -MeL)] (BF4 ) 2 • iCH 2 C~ + 2Cll2Fe (4) Fe(III) complexes are low spin with I-Leff(298 K)/Fe(Ill) values in the 1 range of 2.1-2.3 B.M. In the H NMR spectrum of Fe(4'-MeL) 2 BF4 , three broad signals attributable to the complex and seven sharp signals characteristic of the free ligand, 4' -MeLH, are observed. This latter observation is suggestive of some ligand dissociation by the Fe(III) complex. The electrochemical behavior of freshly prepared DMF solutions of both the mononuclear and binuclear Fe(III) complexes was identical with that of the analogous Fe(II) complexes. However, over a period of ,..., 20 min, irreversible waves appeared in cyclic voltammetric scans of solutions of these complexes. This result also is suggestive of ligand dissociation with the irreversible waves being attributed to related complexes in which a 1, 3-bis(2-pyridylimino)isoindoline ligand has been replaced by solvent molecules. A third Fe(ITI) complex, Fe(4' -MeL)C~ (~), was prepared and isolated from a reaction of 4' -MeLH with FeCis (eq 5). In the infrared FeCis + 4' -MeLH - Fe( 4' -MeL)Cl 2 + HCl (5) spectrum of this complex, two equal intensity absorptions are 1 observed at 348 and 363 cm- • These absorptions are attributed to Fe(In)-Cl stretching; 4 with two such absorptions being expected for a complex with C2 v symmetry. 5 This Fe(III) complex is high spin with a 1-Leff (298 K)/Fe(III) value equal to the expected spin-only value, 5.92 B.M. There are four extremely broad signals in the 1H NMR 165 spectrum of this complex; only the methyl protons at 9. 52 ppm could be definitively assigned. Dioxygen Reactivity Studies. Solutions of (4' -MeL)FeL-LFe(4' -MeL) ("" 1 mM) in methylene chloride, 1:1 methylene chloride/methanol, and DMF exposed to dioxygen for 1 week showed no reactivity by electronic absorption spectroscopy. The intensity of the absorption at 755 nm (E = 1450) decreased only slightly, and the appearance of an absorption at 843 nm (E = 1220) characteristic of the expected oxidized product, 2 (4'-MeL)FeL-LFe(4'-MeL) +, was not observed. Solutions of Fe( 4' -MeL) 2 behaved similarly. The analogous Co(II) and Ru(II) binuclear complexes, with M(III)/M(II) reduction potentials similar to that of the Fe(II) binuclear complex, 2 ' 6 also were unreactive towards dioxygen in apr otic or weakly prot ic solvents. 1 In such solvents with very low proton concentrations, this lack of multielectron dioxygen reactivity is not completely surprising, since the formation of free 0~- is thermodynamically unfavorable. Only if 0~- is somehow stabilized by an interaction with the complex would there be a driving force for this reaction in aprotic or weakly protic solvents. Thermodynamically more favorable conditions for the two-electron reduction of dioxygen would be provided by acidic media. However, sine e the binuclear Fe(II) complex decomposes under these conditions, dioxygen reactivity of this complex in thermodynamically favorable media could not be studied. The related Ru(II) binuclear complex, (4' -MeL)RuL-LRu( 4' -MeL), 6 which is not decomposed in acidic media, showed no dioxygen reactivity in aqueous hydrochloric acid solutions. 1 166 The original samples of (4' -MeL)FeL+LFe(4' -MeL) for which 0 2 reactivity was observed most likely contained significant quantities of unreacted starting materials, HL-LH and Fe(4' -MeL)(OAc) (vide supra). This latter species, a five-coordinate mononuclear Fe(II) complex, does react with dioxygen in solution. It is likely that the apparent reactivity of the original samples of (4' -MeL)FeL-LFe(4' -MeL) was due to Fe{4' -MeL)(OAc) which was a significant component of these samples. 167 EXPERIMENTAL SECTION General Considerations. Unless noted otherwise, solvents were reagent grade and were used without further purification. Methylene chloride was distilled from calcium hydride and diethyl ether was distilled from sodium. The ligand, 4' -MeLH, 7 and the complexes, Fe(4' -MeL) 2 and (4' -MeL)FeL-LFe(4' -MeL), 2 were prepared by published methods. Microanalytical data are reported here for the latter complex: Anal. Calcd. for C 8 Jf 82 Fe 2 N20 :C, 68.52;H, 5.48; N, 18.58; Fe, 7.41. Found: C, 68.58; H, 5.52; N, 18.52; Fe, 7.2. Ph sical Measurements. Magnetic susceptibility measurements were obtained at ambient temperature using a Cahn Instruments Faraday balance. The balance was calibrated using HgCo(SCN) 4 and diamagnetic corrections were made using Pascal's constants. 8 Infrared spectra were recorded on a Beckman 4140 spectrometer. Samples were examined as KBr pellets. 1 H NMR spectra were recorded on a JEOL FX90Q spectrometer with CDCls as the solvent and Me 4 Si as an internal reference. Electrochemical procedures have been described previously. 2 Electronic absorption spectra were recorded on a Cary 14 spectrophotometer using matched quartz 1-cm cells. Elemental analyses were determined by the California Institute of Technology's analytical facility. ((4' -MeL)FeL-LFe(4' -MeL)] (BF4 ) 2 • ~CH 2 Cl (3). All manipulations in this synthesis were carried out using high-vacuum line techniques. A solution of 50 mg (0. 033 mmol) of (4' -MeL)FeL-LFe(4' -MeL) and 18 mg (0. 066 mmol) of ferricenium tetrafluoroborate 9 in 20 mL of CH 2 C~ was stirred for 5 min. After carefully removing all but 3 mL 168 of the solvent, diethyl ether was slowly distilled into the reaction flask until a reddish solid just began to precipitate. The mixture was then quickly filtered. Deep red 3 (29 mg, 49%) was crystallized from the "" filtrate, washed with diethyl ether, and dried in vacuo. Anal. Calcd. for C 67 • 5 H 85 ~ClsF8 Fe 2 N20 : C, 58.11; H, 4.73; N, 15.49; Fe, 6.17. Found:C, 57.76; H, 4.88; N, 15.54; Fe, 6.3. All manipulations in this synthesis were carried out using high-vacuum line techniques. A solution of 198 mg (0. 28 mmol) of Fe( 4' -MeL) 2 and 77 mg (0. 28 mmol) of ferricenium tetrafluoroborate in 30 mL of CH2 Cl2 was stirred for 5 min. All but 4 mL of the solvent was removed and 20 mL of diethyl ether was distilled into the reaction flask. Dark red 4 (131 mg, 59%) was "" crystallized from this solution, washed with diethyl ether, and dried invacuo. Anal. Calcd. forC 40 H32 BF4 FeN10 : C, 60.40; H, 3.98; N, 17.61; Fe, 7.02. Found: C, 59.47; H, 3.91; N, 17.48; Fe, 7.0. ~A solution of 490 mg (1.5 mmol) of 4'-MeLH in 30 mL of warm toluene was added dropwise to a filtered solution of 275 mg (1. 7 mmol) of FeC1 3 in 40 mL of methanol. The resulting dark red solution was stirred and heated at reflux for 5 min. The volume of solution was then reduced to 40 mL by evaporating the solvent at boiling. Upon being cooled to ambient temperature, the solution yielded. red-purple crystals of "" 5 (380 mg, 56%). The product was collected, washed with diethyl ether, and dried in vacuo. Anal. Calc d. for (: 2 oH16Cl12 FeN5 : C, 53. 01; H, 3. 56; N, 15. 46. Found: C, 53.49; H, 3.65; N, 15.66. 169 REFERENCES ~ 1. Marks, D. N., unpublished results. 2. Gagne, R. R.; Marritt, W. A.; Marks, D. N.; Siegl, W. 0. Inorg. Chern. lli!., 20, 3260. 3. For the mononucleating ligands, !, position(s) and type(s) of the pyridine ring substituent(s) precede the abbreviation, LH. The binucleating ligand, ...... 2, with (4' -sec-butyl)pyridine rings is denoted HL-LH. 4. Nakamoto, K. "Infrared Spectra of Inorganic and Coordination Compounds;' 2nd ed., Wiley Interscience: New York, 1970. 5. Cotton, F. A. "Chemical Applications of Group Theory"; 2nd ed.; Wiley Interscience: New York, 1970. 6. Marks, D. N. Ph. D. Dissertation, California Institute of Technology, Pasadena, CA, 1982. 7. Siegl, W. 0. J. Org. Chern. !lTI_, 42, 1872. B. Earnshaw, A. "Introduction to Magnetochemistry"; Academic Press: London, 1968, pp. 9-12. 9. Hendrickson, D. N.;Sohn, Y.S.; Gray, H. B. Inorg. Chern. 1971, -10, 1559 . ...............