ANALYST, AUGUST 1992, VOL. 117 1243 Novel Conducting Polymers Incorporating Covalently Bound Metal-Tetraazamacrocycle Complexes* Geoff King, Simon J. Higginst and Nikki Price Department of Chemistry, Donnan and Robert Robinson Laboratories, University of Liverpool, P. 0. Box 147, Liverpool L69 3BX, UK The syntheses of a tetraazamacrocycle functionalized at nitrogen with a pendant 3-thiophene, 1 -[2-(3- thienyl)ethyl]-l,4,8,1 I-tetraazacyclotetradecane (cyclamN-CH2CH2-thiophene), and its complexes with Nil1 and Cull, are described. Electro-oxidation of the pure Nil1 complex in acetonitrile electrolytes afforded only soluble oligomers. Novel conducting polymer-modified electrodes were fabricated, however, by electro- copolymerization of the Nil' complex with 3-methylthiophene. These are unusual in that both the metal centre [Ni1l-NiI1l at +I .I 5 V (saturated calomel electrode)] and the conducting polymer backbone show stable, reversible redox behaviour, with no apparent oxidative degradation of the poly(thiophene) backbone. The Ni1I-Ni1l1 redox behaviour appears unaffected by incorporation into the conducting polymer matrix. However, the NilI-Nil wave, observed at -1.40 V in solution, is entirely suppressed in the polymer, which at that potential is an electronic insulator. Keywords: Poly( thiophene); metal complex modified electrode; cyclam; conducting polymer The implications of electrode modification for analytical applications are now well recognized.' We are interested in the modification of electrodes with various types of metal complexes, both electrocatalytic and catalytic.A particularly useful method of electrode modification is the electropoly- merization of a suitable monomer.2J In principle, precise control of the amount of metal complex deposited per unit area is possible, via control of the charge passed in the electropolymerization. The technique allows individual coat- ing of single electrodes in microelectrode arrays.4 It might also allow the sequential deposition of different polymer layers on the same electrode.5 These considerations could be important in device fabrication.4 Broadly, two approaches to electrochemically generated polymer-modified electrodes, incorporating covalently bound metal complexes, have been used. In the first, reductive polymerization of a metal complex, incorporating a vinyl- substituted bipyridine or pyridine ligand, results in a coating of a redox-conducting polymer on the electrode.In order to obtain a reproducible result, the metal complex concerned must be stable at the cathodic potentials used; polymerization occurs as a result of the ligands being reduced to reactive radical anions.3 In the second approach, oxidative poly- merization of a metal complex, incorporating a suitable ligand functionalized with a pendant thiophene or pyrrole moiety, is used [Fig. l(a)]."10 A poly(heterocyc1e) coating on the electrode is obtained. In this instance, the metal complex must be stable at the anodic potential required to oxidize the heterocycle to the radical cation, which is the key intermediate in the subsequent polymerization. In principle, the latter approach has the advantage that in their oxidized form, simple poly(heterocyc1e) films are elec- tronically conducting.If the functionalized polymers were also conducting, this might be valuable in electrocatalytic applica- tions (eg., in electroanalysis) .I1 Additionally, the recent development of a wide range of other functionalized hetero- cycles for electropolymerization should permit copolymeriza- tion to yield polymers with optimized conductivity, hydrophi- licity and electrocatalytic activity. Some attempts at this, using pyrrole-functionalized metal complexes and simple hetero- cycles (thiophene, pyrrole, bithiophene and 3-methylthio- phene), have already been described.11-'3 * Presented at the meeting on Analytical Applications of Chemi- f- To whom correspondence should be addressed.cally Modified Electrodes, Bristol, UK, January 7-8, 1992. To date, most of the work in this field has been performed with ligands of the type shown in Fig. l(a), as the syntheses of these are fairly straightforward. However, homopolymers prepared from metal complexes of these ligands display poor electronic conductivity.10 Also, the electroactivity of the conducting polymer backbone is often adversely affected on repeated electrochemical cycling.8 For some systems, the redox wave(s) due to the metal complex are virtually unchanged in spite of this; presumably, in these instances, the modified electrode material has sufficient redox conductivity to allow the metal centres to communicate with the electrode even after oxidative degradation of the poly(heterocyc1e) backbone.14 More recently, attempts have been made to overcome the poor conductivity found with N-functionalized poly(pyrro1e)- based materials by using 3-functionalized pyrroles or thio- l a R = (CH2InNC4H4 R = CH3 Ib Fig. 1 (a) Pyridine and 2,2'-bipyridine ligands functionalized with pyrrole moieties; ( b ) l-[2-(3-thienyl)ethyl]-1,4,8,11-tetraazacyclo- tetradecane (cyclamN-CH2CH2-thiophene)1244 ANALYST, AUGUST 1992, VOL. 117 phenes. It is known that, provided that a 'spacer' of suitable length separates a bulky functional group from the hetero- cycle, poly(3-functionalized heterocycle) materials, which display high electronic conductivity and good polymer proper- ties, can be synthesized.15 Few materials of this type incorpor- ating metal complexes have as yet been prepared, but those that have'6.17 show interesting properties.Metal complexes of tetraazamacrocyclic ligands are known to act as electrocatalysts for many significant redox reactions, including carbon dioxide reduction,l8 alkene epoxidation'g and reduction of alkyl halides.20 We wished to prepare conducting polymer-modified electrodes containing covalently bound tetraazamacrocyclic ligands and their com- plexes, but at the same time to preserve the electronic conductivity of the conducting polymer. Therefore, it was decided to use 3-functionalized thiophene chemistry17 and the ligand shown in Fig. l(b). During the course of this work, a preliminary report of similar work appeared 2 1 Our results differ from those described in that report in several respects, and we report these results here. Experimental Apparatus Infrared (IR) spectra were measured as Nujol mulls, with use of a Perkin-Elmer 1720X Fourier transform (FT) spec- trometer (Norwalk, CT, USA), from 400 to 4500 cm-1.Electronic absorption spectra were recorded on solutions, by using a Perkin-Elmer 330 spectrophotometer, from 12 500 to 40000 cm-1. Proton nuclear magnetic resonance (NMR) spectra were recorded with use of a Bruker WM 250 MHz FT spectrometer (Billeria, MA, USA). Samples were dissolved in CDCI3. Tetramethylsilane was used as reference. Mass spectra were recorded with a VG 7070E instrument (Stam- ford, CT, USA) operated in the electron-impact (70 eV) mode (intermediates and free ligands) or positive-ion fast-atom bombardment (FAB) mode (complexes); 3-nitrobenzyl al- cohol was used as solvent, and xenon as bombardment gas in the latter instance.Cyclic voltammetry experiments were performed with laboratory-built potentiostat and a conven- tional three-electrode cell configuration. Working electrodes were polished platinum discs, counter electrodes were plati- num gauzes and the reference electrode was a commercial saturated calomel electrode (SCE) separated from the work- ing compartment by a glass frit and a Luggin capillary. All potentials quoted are referred to the SCE. The ferrocene- ferricinium couple was examined routinely to check the stability of the reference electrode and junction potentials. Chemicals Acetonitrile (BDH, Poole, Dorset, UK; HPLC grade) was dried by reflux over, and distillation from, CaH2.Tetraethyl- ammonium tetrafluoroborate (TEAT) was prepared from the bromide (Lancaster Synthesis, Lancaster, UK) and 50% HBF4 (Fluka, Buchs, Switzerland) in water, and recrystallized three times from hot ethanol. It was dried for 48 h at 0.1 mmHg. The 1,4,8,11-tetraaq1cyclotetradecane (cyclam) was prepared by a literature method.22 p-Toluenesulfonyl chloride (Aldrich, Milwaukee, WI, USA) was purified by a literature method.23 All other chemicals were used as received. Syntheses 2-(3- Thieny1)ethyl-p-toluenesulfonate p-Toluenesulfonyl chloride (1.86 g, 9.76 mmol) was dissolved in anhydrous pyridine (3.0 cm3) and left to stand for 15 min at 0 "C. 2-(3-Thienyl)ethanol (1.00 g, 7.80 mmol) was added dropwise, with stirring, at 0 "C. Stirring was continued for 2 h.Ice (10 g) was added and the product was extracted into dichloromethane (3 X 10 cm3). The extracts were washed with HCI (2 mol dm-3; 3 x 10 cm3), then water (3 x 10 cm3) and dried over Na2S04. The solvent was removed (rotary evapora- tor) to leave a white, crystalline solid. Yield: 1.835 g, 83.4%. Proton NMR spectroscopic data [selected 1H NMR data: 6 4.20 (t, 2 H, S020CH2CH2C4H3S)] showed this to be sufficiently pure for synthetic use. A sample was recrystallized from diethyl ether-light petroleum (boiling-range 60-80 "C) for microanalysis and electrochemical experiments. (Found: C, 55.00; H, 4.95. Calc. for Cl3HI4O3S2: C, 55.32; H, 4.96% .) 2.43 (s, 3 H, CH3C6H4-), 3.05 (t, 2 H, -S020CH2CH2C4H3S), 1 - [ 2-(3- Thienyl)ethyl]- 1,4 ,8,11-tetraazacyclotetradecane (lb) ( cyclamN-CH2CH2-thiophene) To a solution of cyclam (7.4 g, 37.0 mmol) in acetonitrile (350 cm3) and pyridine (150 cm3) was added dropwise 2-(3-thienyl)- ethyl-p-toluenesulfonate (4.76 g, 16.86 mmol), with stirring, under reflux.The mixture was heated under reflux for a further 4 h, the solvent was removed under reduced pressure, and the residue was dissolved in 0.2 mol dm-3 NaOH (100 cm3). The solution was extracted with diethyl ether (5 x 100 cm3). (The excess of cyclam could be recovered from the aqueous phase by extraction into dichloromethane.) The organic phase was dried over Na2S04, and the solvent was removed under reduced pressure, leaving a waxy oil. This was dried in vacuo (0.1 mmHg) for 48 h.Proton NMR spec- troscopy: 6 1.70 (complex m, 4 H, -NHCH2CH2CH2NH-), 2.60 (complex m, 23 H), 6.88 (m) and 7.00 (d of d, total 3 H, -C4RH3S-3). Mass spectrometry [mlz 200 (9), 310 (loo), 420 (S)] indicated the presence of small amounts of unreacted cyclam and difunctionalized cyclam. Attempts to purify the crude material by chromatography proved unsatisfactory, but the material was sufficiently pure to allow complexes to be prepared. { 1-[2-(3- Thienyl)ethyl]-l,4,8,ll-tetraazacyclotetradecane} - nickel(i1) perchlorate { [ Ni(cyclamN-CH2CH2-thiophene)] The ligand (0.70 g, 2.26 mmol) dissolved in methanol (20 cm3) was added dropwise to a stirred, hot solution of [Ni(H20)6]- [C1O4I2 (0.91 g, 2.48 mmol) in methanol (60 cm3).A small amount of material precipitated from the orange solution. After 15 min, the solution was filtered hot, and allowed to cool. An orange precipitate appeared, which was filtered off and dried in vacuo. Yield 0.80 g. (Found: C, 33.60; H, 5.33; N, 9.89. Calc. for C16H30C12N4NiOsS: C, 33.83; H, 5.32; N, 9.86% .) FAB-MS: Nil1 complexes of unfunctionalized cyclam {mlz 256, [Ni(L-H)]+}, monofunctionalized cyclam (467, { [Ni(L)]C104}+; 367, [Ni(L-H)]+) and difunctionalized cyclam (577, {[Ni(L)C104} +; 477, [Ni(L-H)]+). The complex was recrystallized three times from hot methanol-ethanol, whereupon FAB-MS data indicated that it was pure. Obser- ved and calculated isotope patterns for the peak at mlz 467 were in good agreement. Yield 0.62 g; 48%. lO-3E,,,l~rn-~ (~,,,/dm-3 mol-1 cm-1) (CH3CN), 22.03 (21).[C10412} [ l-{2-(3-Thienyl)ethyl}-l,4,8,1l-tetraazacyclotetradecane]- copper(i1) perchlorate [ Cu(cyc1amN-CH2CH2-thiophene)]- [CEO412 This was prepared in the same way as the Ni*' complex, as a purple-red solid. Yield 63%. (Found: C, 33.60; H, 5.33; N, 9.89. Calc. for C16H30C12C~N408S: C, 33.83; H, 5.32; N, 9.86% .) 10-3 E,,,lcm-1 (~,,,/dm-3 mol-1 cm-1) (CH3CN), 14.66 (240); 18.68 sh (96). Electro-copolymerization Electrochemical polymerization of 2-(3-thieny1)ethyl-p- toluenesulfonate A 0.015 mol dm-3 solution of the recrystallized tosyl ester in 0.2 mol dm-3 Et4NBF4-CH3CN was used. The workingANALYST, AUGUST 1992, VOL. 117 1245 electrode had an area of 0.85 cm2. After assembly of the cell, freshly activated molecular sieve (4A) was added to the electrolyte, and the solution was de-oxygenated by bubbling 02-free N2.Cycling from 0.25 to +1.95 V (SCE) at 0.05 V s-1 produced a dark-blue adherent film. Electrochemical polymerization of { 1-[ 2-(3-thienyl)- ethyl]-1,4,8,11-tetraazacyclotetradecane}nickel( I I ) perchlorate The conditions used were as for the above experiment, except that the electrolyte contained 0.03 rnol dm-3 3-methylthio- phene and 0.03 rnol dm-3 NilL complex. Cycling from 0.00 to +2.00 V at 0.10 V s-1 produced an adherent electrochromic film, although formation of some soluble material was also evident. Results and Discussion Commercially available 2-(3-thienyl)ethanol is a valuable starting material for 3-functionalized thiophenes.15 Interest- ingly, it has been found that, under rigorously anhydrous conditions (vacuum oven-dried 4A molecular sieves present in the cell), the tosyl ester can be electropolymerized to a conducting polymer with electrochemical characteristics typical of poly(3-al kyl thiophene)~.15 Electropolymerization was carried out by cyclic voltammetry between +0.25 and +1.91 V (SCE) in 0.2 rnol dm-3 TEAT in CH3CN. Conducting polymers of this type could be useful for subsequent functionalization; a patent report mentions the electrosynthesis of a similar polymer and its subsequent reaction with sodium iodide in acetone.24 We have synthesized 1-[2-(3-thienyl)ethy1]-1,4,8,11-tetra- azacyclotetradecane (lb) (cyclamN-CH2CH2-thiophene) from cyclam and the tosyl ester of 2-(3-thienyl)ethanol. Square planar Nil1 and Cu" complexes of the ligand were prepared by using the metal perchlorates.Although these complexes had the correct microanalyses (C, H and N; see under Experimental), FAB-MS data suggested that this was fortuitous; small and variable amounts of complexes of unfunctionalized cyclam and difunctionalized cyclam were invariably present. It was found to be more convenient to purify the complexes, by recrystallization from hot methanol- ethanol, than to purify the free ligand. Electronic spectro- scopy for the pure complexes showed a slight diminution in the ligand field on functionalizing the ligand at nitrogen. Hence, for the Nil1 complex, the broad band due to the overlapping transitions to dx2-y2 is at 21 370 cm-1. In [Ni(cyclam)]2+, this occurs at 22030 cm-1.Tertiary amine donors are weaker field ligands than secondary amine donors.25 This affects the redox potentials for the NiI-Ni" and Ni"-Ni"l couples. For [Ni(cy- clam)]*+ in CH3CN, these occur at -1.46 and +1.08 V, respectively. For [Ni( cyclamN-CH2CH2-thiophene)]2+ , they occur at - 1.40 and + 1.15 V, respectively. This behaviour has been observed previously for N-functionalized macrocycles .26 Attempts were made to prepare modified electrodes by electro-oxidation of the thiophene moieties in cyclic voltam- metry experiments by using the pure (thrice recrystallized) Nil1 complex. When using 0.2 rnol dm-3 TEAT-CH3CN, a quasi-reversible wave at + 1.15 V was observed, correspond- ing to the Nill-Nilll process, and at more positive potentials, an irreversible oxidation (onset approximately + 1.85 V).Anodic of the latter potential, intensely blue material was observed streaming away from the working electrode. Changes in the conditions (electrolyte concentration 0.01-0.35 rnol dm-3; monomer concentration 0.001-0.045 rnol dm-3; anodic poten- tial limit +1.85 to +2.10 V) did not alter the observed behaviour; no polymer was deposited on the working elec- trode. Attempts were therefore made to copolymerize the [ Ni(cyc1amN-CH2CH2- thiophene)]'+ with 3-me thy1 thiophene in the hope that the solubility of the oligomers evidently being generated at > + 1.85 V would be reduced. This strategy was successful: adherent, stable films were Potential - Fig. 2 Cyclic voltammograms of poly(3-methylthio hene)- copoly { [ (cyclamN-CH2CH2-thiophene)nickel( 11'3 perchPorate)- modified platinum disc electrode in 0.2 mol dm- TEAT-CH3CN at: A, 0.01; B, 0.02; and C, 0.05 V s-l, respectively.Scan range, 0.00 to +1.40 V (SCE) produced. The modified electrodes were removed from the growth solution in their reduced state, washed in pure CH3CN and stored in dry air. Subsequent cyclic voltammetry in 0.2 rnol dm-3 TEAT-CH3CN (Fig. 2) showed several features of interest. First, the conducting polymer backbone is electro- active, although its redox wave is of an unusual type for a 3-functionalized thiophene .15 In particular, two distinct processes are evident. Second, by extrapolating the wave for poly(3-R-thiophene) and calculating the charge due to this process and that due to Nill-NitlL, and assuming that the oxidation of the thiophene moieties corresponds to one electron removed per four rings,15 the composition of the copolymer is approximately 3-methylthiophene-[Ni(cy- clamN-CH2CH2-thiophene)]z+ (10 : 1). Third, repeated cycling from 0.00 to +1.40 V has a negligible effect on either the poly(thiophene) or the NiI1-NiI1l process.Fourth, the modified electrode is highly electrochromic, being orange-red when neutral, green after the poly(thiophene) oxidation wave, but before that of NilL-Ni[ll, and intensely blue after the NiII-Ni"' wave. Finally, the Ni1I-Ni1I1 wave Ed is identical, within experimental error, for the process in solution and that in the polymer. This is particularly significant; on oxidation to N P , the process: [Ni(L)]2+ - e- + 2 CH3CN + [Ni(L)(CH,CN)#+ is occurring.Clearly, for the potential to be unchanged within the polymer, there must be sufficient solvent within the film to permit this process to occur. In complexes of this type, it is known that E; for Nill-Nill' is a sensitive function of the species available as axial co-ligands.26 Ellipsometric data for poly- (thiophene)27 suggested that little solvent was present in the as-grown film. Presumably, the copolymer films prepared here contain more solvent as a result of the presence of highly charged [Ni(L)(CH3CN)#+ moieties within the growing film. The scan range was extended to negative potentials. Interestingly, the Ni"-Nil couple, observed in solution at -1.40 V, was not seen for the polymer-trapped complex. Instead, an irreversible reduction (onset potential approxi- mately -1.7 V) occurred.On the reverse scan, an additional oxidation then occurred (Epa approximately +0.83 V), and the Ni"-Ni"' wave was greatly diminished. If the potential was then kept at 0.00 V for 10 min, and another cycle from 0.00 to rt 1.42 V was then recorded, the poly(thiophene) electro- chemistry was almost unchanged, but the charge under the Ni"-Ni"' wave was diminished by 75%.1246 ANALYST, AUGUST 1992, VOL. 117 The irreversible reduction wave is tentatively assigned to partial reduction of the poly(thiophene) to its anionic conduct- ing form. Clearly, in our film there is insufficient redox conductivity to allow communication of the Nil1 complex with the electrode when the poly(thiophene) is neutral (electronic- ally insulating).This contrasts with observations made for poly(pyrro1e) films N-functionalized by bipyridine com- plexes.7-10 Once the poly(thiophene) film becomes partially conducting again at < - 1.7 V, the Ni" complex is presumably reduced to Nil and cannot be re-oxidized again until > +0.85 V. This would account for the decrease in the Ni"-NilI1 charge; Nil-tetraazamacrocycle complexes are labile, and on the voltammetric time scale, significant decomposition of the Nil complex could occur. Currently, in situ electronic spectros- copy and FTIR studies are in progress to investigate this further. While this work was in progress, a preliminary report appeared describing a homopolymer of [Ni(cyclamN- CH2CH2-thiophene)]2+ and its electrochemistry.21 The poly- mer was prepared with use of 0.05 mol dm-3 solutions in 0.1 mol dm-3 Bu4NBF4-CH3CN by cycling from -0.5 to + 2.0 V (versus ferrocene-ferricinium).In contrast to our copolymer, the poly(thiophene) showed no electroactivity, although the NilI-Nilll process was well defined. Attempts to duplicate this result with the recrystallized Nil1 complex failed: only soluble oligomers were generated. However, on using material that was recrystallized only once,21 variable results were obtained; in one instance, an orange polymer film had grown. This had the same electro- chemical characteristics as those described elsewhere.21 Interestingly, this polymer was not noticeably electrochromic. The FAB-MS data on the monomer used in this experiment showed the presence of the complex of a difunctionalized ligand.It is commonly observed that complexes containing more than one monomer unit are more readily electropoly- merized.11-13,28 We tentatively suggest that the presence of this impurity might assist the formation of a polymer-modified electrode; attempts to characterize this process further are continuing. Attempts at mediated electron transfer with the copolymer- modified electrode are in progress, as are attempts to obtain polymers incorporating the Cull complex. The methylation of the remaining amine donors in 1-[2-(3-thienyl)ethyl]-l,4,8,11- tetraazacyclotetradecane is being examined with a view to the preparation of Ru" complexes, and of modified electrodes for alkene epoxidation. 19 We thank Alan Mills for the FAB-MS measurements.S. J. H. is grateful to the Nuffield Foundation for a grant under the Awards for Newly-Appointed Science Lecturers scheme. References 1 Chemically-Modified Surfaces in Catalysis and Electrocatalysis, ed. Miller, J. S., American Chemical Society, Washington, DC, 1982. 2 3 4 5 6 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Handbook of Conducting Polymers, ed. Skotheim, T. A., Marcel Dekker, New York, 1986. Elliott, C. M., Baldy, C. J., Nuwaysir. L. M., and Wilkins, C. L., Inorg. Chem., 1990,29,389. Thackeray, J. W., White, H. S., and Wrighton, M. S.. J. Phys. Chem., 89, 5133. Iyoda, T., Toyoda, H., Fujitsuka, M., Nakahara, R., Tsuchiya, H., Honda, K., and Shimidzu, T., J. Phys. Chem., 1991, 95, 5215. Collin, J. P., Jouaiti, A., and Sauvage, J.P., J. Electroanal. Chem. Interfacial Electrochem., 1990, 286, 75. De Oliveira, 1. M. F., Moutet, J. C., and Vlachopoulos, N., J. Electroanal. Chem. Interfacial Electrochem., 1990, 291, 243. Daire, F., Bedioui, F., Devynck, J., and Bied-Charreton, C., J. Electroanal. Chem. Interfacial Electrochem., 1987, 224, 95. Cosnier, S., Deronzier, A., and Moutet, J. C., J. Electroanal. Chem. Interfacial Electrochem., 1986,207, 315. Cosnier, S., Deronzier, A., and Roland, J. F., J. Electroanal. Chem. Interfacial Electrochem., 1991, 310, 71. Ochmanska, J., and Pickup, P. G., J. Electroanal. Chem. Interfacial Electrochem., 1991,297, 211. Ochmanska, J., and Pickup, P. G., J. Electroanal. Chem. Interfacial Electrochem., 1991, 297, 197. Ochmanska, J., and Pickup, P. G., Can. J. Chem.. 1990, 69, 653. Eaves, J. G., Munro, H. S., and Parker, D., Inorg. Chem., 1987,26, 644. Roncali, J . , Youssoufi, H. K., Garreau, R., Garnier, F., and Lemaire, M., J. Chem. SOC., Chem. Commun., 1990,414. Inagaki, T., Hunter, M., Yang, X. Q., Skotheim, T. A., and Okamoto, Y., J. Chem. SOC., Chem. Commun., 1988, 126. Mirrazaei, R . , Parker, D., and Munro, H. S., Synth. Met., 1989, 30, 265. Fujihara, M., Hirata, Y., and Suga, K., J. Electroanal. Chem. Interfacial Electrochem., 1990, 292, 199. Che, C. M., Tang, W. T., Wong, W. T., and Lai, T. F., J. Am. Chem. SOC., 1989, 111,9048. Bakac, A., and Espenson, J. H., J. Am. Chem. SOC., 1986,108, 712. Sable, E., Handel, H . , and L'Her, M., Electrochim. Acra, 1991, 36, 15. Barefield, E. K., and Freeman, G., Inorg. Synth., 1980,20,108. Vogel's Textbook of Practical Organic Chemistry, eds. Furniss, B. S., Hannaford, A. J., Rogers, V., Smith, P. W. G., and Tatchell, A. R.. Longman, London, 4th edn., 1978, p. 317. Wudl, F., and Heeger, A. J., PCT Int. Appl. WO 87 05,914; Chem. Abstr., 1988, 109, 38478~. Wagner, F., and Barefield, E. K., Inorg. Chem., 1976,15,408. De Santis, G., Di Casa, M., Mariani, M., Seghi, B.. and Fabbrizzi, L., J. Am. Chem. SOC., 1989, 111,2422. Hamnett, A., and Hillman, A. R., J. Electrochem. Soc., 1988, 135,2517. Beer, P. D., Kocian, O., and Mortimer, R. J., J. Chem. SOC., Dalton Trans., 1990, 3283. Paper 21005 77H Received February 3, 1992 Accepted February 17, 1992