Вы находитесь на странице: 1из 11

Journal of Electroanalytical Chemistry 507 (2001) 124 134 www.elsevier.

com/locate/jelechem

Nickel(I)(salen)-electrocatalyzed reduction of benzyl chlorides in the presence of carbon dioxide


Armando Gennaro *1, Abdirisak Ahmed Isse, Flavio Maran *2
Dipartimento di Chimica Fisica, Uni6ersita di Pado6a, 6ia Loredan 2, 35131 Pado6a, Italy ` Received 3 August 2000; received in revised form 31 October 2000; accepted 13 November 2000

Abstract The electrochemical carboxylation of a series of substituted benzyl and benzylic-type chlorides Y-C6H4CH(Z)Cl (Z= H; Y=H, 4-CF3, 4-OCH3, 3-OCH3; and Y=H; Z= Ph, CH3), catalyzed by nickel(I)(salen), was investigated in acetonitrile by cyclic voltammetry and controlled-potential electrolysis. For all of the chlorides investigated, remarkable catalytic currents were measured under atmospheric CO2 pressure. For most chlorides, controlled-potential electrolysis yields the corresponding carboxylic acid as the main product, the other signicant product being the substituted toluene. The results indicate that the extent of carboxylation is strongly inuenced by the structure of the halide. Electron-withdrawing groups, either on the phenyl ring or on the benzylic carbon, favor the formation of the carboxylate. In the rst step of the electrocatalytic process, the halide reacts with electrogenerated [NiI(salen)] to form the benzyl radical. Further reduction of the radical to the corresponding carbanion is followed by electrocarboxylation in competition with protonation by residual water. The yield of the carboxylic acid is determined by the ease of reduction of the benzylic radical, which in turn reects the nucleophilicity and basicity of the ensuing carbanion. 2001 Elsevier Science B.V. All rights reserved.
Keywords: Benzyl chlorides; Electrocatalysis; Nickel(I)(salen); Carboxylation

1. Introduction The electrochemical carboxylation of organic halides (RX) has been extensively studied as a convenient way to obtain carboxylic acids and carboxylic acid derivatives [15]. Generally, the electrocarboxylation is carried out by reduction in a CO2-saturated aprotic solvent. A major drawback of the direct electroreduction approach, however, is that very negative potentials are often required. This is particularly true for chlorides (RCl) because, unless the RCl molecule is easily reducible, thanks to the presence of an electron-withdrawing substituent, concomitant reduction of CO2 may take place. Since this leads to undesired products and reduced current efciency, indirect electrochemical methods provide a valid alternative. These methods employ catalytic systems, such as those provided by transition-metal complexes [6 16]. For example, several nickel [611] and palladium [12,13] complexes with
1 2

*Corresponding author; a.gennaro@ch.unipd.it *Corresponding author; f.maran@ch.unipd.it

phosphine ligands have displayed good catalytic properties in the electrocarboxylation of various aromatic and benzylic halides. For some nickel complexes, electrocarboxylation proceeds through oxidative addition of the halide by an electrogenerated zerovalent nickel complex [9]. Alternatively, with other nickel complexes, oxidative addition occurs on a nickel(I) intermediate, possibly complexed with carbon dioxide itself [11]. In either case, the carboxylate is then formed by the insertion of CO2 into a nickel carbon bond, although there is still some disagreement about the nature of the actual organometallic intermediate [10]. Square-planar cobalt complexes such as CoII(salen) (H2salen= N,N%bis(salicylidene)-ethane-1,2-diamine) [14,15] and Co(tetraphenylporphyrin) [16] have displayed good catalytic activity in the electrocarboxylation of several alkyl halides. The reaction between electrogenerated CoI and RX yields an organometallic intermediate, [CoIII(L)(R)], which upon further reduction releases the alkyl group as a radical. Reduction of the latter yields a carbanion R that then reacts with CO2 [15,17]. Electrogenerated organic radical anions acting as outer-

0022-0728/01/$ - see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 2 2 - 0 7 2 8 ( 0 1 ) 0 0 3 7 3 - 4

A. Gennaro et al. / Journal of Electroanalytical Chemistry 507 (2001) 124134

125

sphere electron-transfer agents have also been used as catalysts in the electrocarboxylation of some substituted benzyl halides [18]. Macrocyclic nickel complexes have been employed as catalysts in the electrochemical or chemical activation of a variety of organic halides [11,19 47]. The electrocatalytic reduction of some alkyl halides by nickel complexes containing tetraazamacrocyclic or tetradentate Schiff base ligands was investigated by Pletcher and co-workers [1921]. It was suggested that the reaction of the halides with the electrogenerated nickel(I) complexes yields alkylnickel(III) complexes. Homolytic cleavage of the latter to give alkyl radicals (R) and a nickel(II) complex or, alternatively, further oneelectron dissociative reduction of the alkylnickel complex leading to R were considered as possible reaction pathways. More recently, Peters and co-workers, who also studied the electrocatalytic reduction of various alkyl halides by [NiI(salen)], proposed a reduction mechanism based on an initial dissociative electron transfer to RX, leading to the formation of R which undergoes different radical reactions in the solution [2226]. Complex reaction mixtures were also obtained by Fry and Fry [27] who studied the catalytic reduction of benzal chloride. Both carbanion and radical pathways were proposed to take place as a consequence of inefcient reduction of the radical by [NiI(salen)]. Espenson and co-workers [28 31] have shown that some nickel(I) complexes, particularly Ni(1,4,8,11tetramethyl-1,4,8,11-tetraazacyclotetradecane)+, react readily with various alkyl and benzyl halides. The reaction was proposed to occur in two steps, the rst being electron transfer from nickel(I) to RX to give R which is then trapped by the nickel(I) complex in a second rapid step. Stolzenberg and co-workers have used nickel(I) complexes of porphyrin-type ligands to reduce alkyl halides [32 34]. The results could be explained by a mechanism involving the transient formation of alkyl-nickel(II) intermediates, whereas free-radical pathways were ruled out. A different mechanism was proposed for the same system by Helvenston and Castro [35] who attributed the main reduction products to radical reactions. The initial step of the reduction, however, was envisaged as an SN2-like process involving the halide and NiI(L). The formation of radicals or radical-like species is indeed an important issue in the reactions catalyzed by nickel complexes.

Radical addition onto double bonds is a particularly useful reaction in this regard, as was shown for several reactions catalyzed by nickel complexes, bearing different ligands, through stepwise reduction of aromatic or aliphatic halides [11,22,3646]. Ohmori [36,37,39 41], Peters [22], Fukumoto [42] and Dunach [11,4346] and co-workers have described, in particular, several cases of intramolecular cyclizations of suitable halogenated precursors. The radical formed in the mediated reduction or, in those cases in which oxidative addition of the halide is most likely to occur, the radical character induced by the alkylnickel complex is exploited to attack a side-chain double or triple bond. Mononuclear and hetero-dinuclear (NiII/Ba2 + ) nickel complexes, having a salen-type macrocyclic ligand, have been studied as possible catalytic systems for the electroreductive carboxylation of benzyl chloride in N,N-dimethylformamide (DMF) [47]. However, very poor yields (B5%) of phenylacetic acid were obtained. This result led to the conclusion that these nickel complexes are very inefcient catalysts for this type of reaction. In this paper, we report data on the electrochemical carboxylation of a series of benzyl or benzylic-type chlorides. The reaction was catalyzed by using Ni(salen) in acetonitrile (MeCN). To examine the effect of substituents on both the phenyl ring and the a-carbon, we selected a series of six benzyl chlorides (Scheme 1). 1-Bromo-1-phenylethane was also investigated for comparison. The carboxylations were investigated by means of cyclic voltammetry and controlled-potential electrolysis. For most of the systems investigated, we found that the reaction is efciently catalyzed by Ni(salen). High yields (\ 80%) of carboxylic acid were obtained with the benzyl chlorides bearing electronwithdrawing substituents on either the phenyl ring or the a-carbon. Mechanistic analysis suggests that, despite the apparent role played by the reduction potential of the benzyl radical, carbanion reaction pathways dominate the overall carboxylation process.

2. Experimental

2.1. Chemicals
Acetonitrile (BDH) was distilled over CaH2 and stored under an argon atmosphere. Tetra-n-butylammonium perchlorate (TBAP, Fluka) was crystallized twice from ethanol+ water (2:1) and dried in a vacuum oven at 60C. Ni(salen) was prepared as described in the literature [48]. The methyl esters, used as authentic samples to identify and quantify the electrocarboxylation products, were synthesized by reaction of the appropriate carboxylic acid with methanol in the presence of H2SO4 [49]. 1,4-Diphenylbutane was prepared by reduction of freshly distilled styrene with lithium in

Scheme 1.

126

A. Gennaro et al. / Journal of Electroanalytical Chemistry 507 (2001) 124134

3. Results and discussion

3.1. Cyclic 6oltammetry of benzyl chlorides


Cyclic voltammetry experiments were carried out in MeCN + 0.1 M TBAP, using mercury and glassy carbon electrodes. Although the two voltammetric patterns are qualitatively the same, glassy carbon provided better dened and more reproducible peaks. Fig. 1 shows a cyclic voltammogram for the reduction of benzyl choride at the glassy carbon electrode, obtained at a scan rate (6) of 0.2 V s 1. A similar voltammetric pattern was observed for the other chlorides. On both electrodes, all compounds exhibit a single irreversible, broad peak. The values of the peak potentials (Ep/V) measured at 6= 0.2 V s 1 at the glassy carbon electrode are 1.93 (4-CF3-PhCH2Cl), 2.04 (Ph2CHCl), 2.15 (PhCHMeCl), 2.20 (3-CH3O-PhCH2Cl), 2.23 (PhCH2Cl), and 2.34 (4-CH3O-PhCH2Cl). Ep varies linearly with the logarithm of 6 and the average slope, #Ep/# log 6, is ca. 120 mV decade 1. The peak width depends on the substituted benzyl chloride, the values being in the 0.110.25 V range at 6= 0.2 V s 1. These data are typical of a reduction controlled by the kinetics of the heterogeneous electron transfer. The same holds true for 1-bromo-2-phenylethane, whose Ep is 2.55 V at 0.2 V s 1. The transfer coefcient h was calculated from the peak width as well as from #Ep/ # log 6 and h values distinctly lower than 0.35 were obtained for all the compounds investigated (h= 0.21 0.32). This indicates that the actual reduction potential is much more negative than the standard potential of the rate-determining electron-transfer step [5255]. In fact, the reduction of most of the substituted benzyl chlorides follows a concerted dissociative electron transfer mechanism, i.e. a mechanism where electron transfer and cleavage of the carbonhalogen bond are concerted (Eq. (1)) [56,57]. Owing to the very negative potentials required for the actual reduction of the chlorides investigated (EB 1.8 V), fast reduction of the ensuing benzyl radicals takes place at the working potentials (Eq. (2)). In fact, the standard potentials for reduction of the benzyl radicals are in the range from 0.9 to 1.65 V [58]. Owing to the basicity of the so-formed benzyl carbanion, a protonation reaction completes the reduction scheme (Eq. (3)). Although generically indicated as SH, the proton donor is mostly residual water (vide infra). It is worth noting that, for the compounds investigated, no fatherson reaction and more specically no self-protonation reaction (protonation between the electrogenerated carbanion and the starting material) take place in MeCN [56]. Under our experimental conditions, the direct electroreduction of the selected benzyl chlorides is thus a well-dened two-electron process.

Fig. 1. Cyclic voltammetry of 2.17 mM PhCH2Cl in MeCN+ 0.1 M TBAP at the glassy carbon electrode. 6 = 0.2 V s 1.

tetrahydrofuran, as described in the literature [50,51]. The other products were commercially available and were used as received.

2.2. Instrumentation
Electrochemical measurements were carried out by using an EG&G Princeton Applied Research Model 173/179 potentiostat/coulometer, an Amel Model 568 programmable function generator, and a Nicolet Model 2090 digital oscilloscope. For cyclic voltammetry measurements, glassy carbon or mercury was the working electrode [52]. The counter electrode was a Pt wire. The reference electrode was Ag AgI 0.1 M n-Bu4NI in DMF, calibrated after each experiment against the ferrocene/ferricenium couple. All potentials are reported versus the KCl saturated calomel electrode (SCE), using E + =0.450 V versus SCE. The referFc/Fc ence electrode and the counter electrode were separated from the catholyte by glass frits and Tylose-TBAP-saturated bridges. Controlled-potential electrolyses were carried out in a divided cell by using a stirred Hg pool electrode. All experiments were performed at 25C. The electrolysis products were analyzed with either an HPLC PerkinElmer Series 4 liquid chromatograph, equipped with a UV detector and a reversed-phase ODS2 column eluted with MeCN+ H2O, or a Varian 3700 gas chromatograph equipped with a Supelco column GP 10% SP-1200, 1% H3PO4. Carboxylic acids were detected by GC as their methyl esters. For this purpose, a portion of the electrolyzed solution was treated with excess methyl iodide.

A. Gennaro et al. / Journal of Electroanalytical Chemistry 507 (2001) 124134

127

RCl + e R + Cl R + e ? R R + SH RH +S

(1) (2) (3)

3.2. Cyclic 6oltammetric in6estigation of the Ni(salen) -catalyzed reduction of benzyl chlorides
Cyclic voltammetry of [NiII(L)] in MeCN+ 0.1 M TBAP shows a reversible couple with E = 1.62 V, corresponding to the reversible one-electron reduction of NiII to NiI [48]. [NiII(L)] +e ? [NiI(L)] (4)
Fig. 3. Cyclic voltammograms of 1.08 mM Ni(salen) in MeCN + 0.1 M TBAP obtained at the mercury electrode in the: (a) presence; and (b) absence of CO2. 6 = 0.2 V s 1.

Fig. 2. Cyclic voltammograms of Ni(salen) in MeCN + 0.1 M TBAP at a mercury electrode, 6= 0.2 V s 1: (a) 0.99 mM Ni(salen); (b) as (a) in the presence of 9.98 mM 3-MeOPhCH2Cl; and (c) as (b) in the presence of CO2 (0.28 M).

Table 1 Relative current enhancements of the NiII(salen) reduction peak caused by addition of benzyl chlorides RCl in the absence and presence of CO2 a RCl [NiII(L)]/mM [RCl]/mM Ip/I p
b

Ip/I b with p CO2 23.0 20.2 10.3 8.4 4.3 3.5

Ph2CHCl 4-CF3-PhCH2Cl 3-MeO-PhCH2Cl PhCH2Cl 4-MeO-PhCH2Cl PhCH(CH3)Cl


a

0.99 1.03 0.99 1.03 1.02 2.02

9.84 10.14 9.98 9.99 10.32 10.20

2.9 3.0 2.6 2.4 2.3 1.7

In MeCN+0.1 M TBAP; Hg electrode; 6 = 0.2 V s1. I and Ip are the peak currents for the reduction of NiII(salen) p measured before and after addition of RCl.
b

Upon addition of an organic halide to the solution, the voltammetric behavior of [NiII(L)] changes. The current of the cathodic peak increases whereas that of the anodic peak decreases. Fig. 2 illustrates, as an example, the effect caused by the addition of 10 mM 3-MeOPhCH2Cl to a solution of 1 mM [NiII(L)]. Under these conditions, the reduction peak current (Ip) increases by as much as 2.6 times, whereas the anodic peak disappears. The relative current enhancements (Ip/I, where I is Ip in the absence of the halide) p p obtained with the halides investigated are shown in Table 1. Since direct electroreduction of the benzyl chlorides requires E to be more negative than 1.8 V, the observed current enhancement is caused by the reaction between the electrogenerated [NiI(L)] and RCl, leading to the catalytic reduction of the latter. When the solution containing [NiII(L)] and the arylmethyl chloride is saturated with carbon dioxide, the peak current for the reduction of the catalyst increases substantially. An example of the effect of CO2 on the electrocatalytic reduction of benzyl chlorides is illustrated in Fig. 2 (curve c). The Ip/I ratios measured in p the presence of CO2 are also included in Table 1. Since several square-planar nickel(I) complexes have been reported to reduce CO2 catalytically to CO [59 63], we checked whether the catalytic reduction of CO2 by [NiI(salen)] could contribute or not to the observed current enhancement. Fig. 3 shows the cyclic voltammograms of the nickel complex obtained in the absence and then in the presence of CO2. In CO2-saturated acetonitrile, only a slight increase (by a factor of ca. 1.1) of the cathodic peak and a decrease of the anodic peak were observed at 6 =0.2 V s 1. A moderate increase of the scan rate (up to 1 V s 1) fully restores the uncomplicated cyclic voltammetric pattern of [NiII(L)] reduction. Although these results point to some interac-

128

A. Gennaro et al. / Journal of Electroanalytical Chemistry 507 (2001) 124134

tion between the reduced nickel complex and CO2, this reaction is clearly too slow to compete with the reaction between [NiI(L)] and benzyl chlorides. Therefore, we conclude that the remarkable catalytic currents observed for the reduction of NiII(salen) in the presence of RCl and CO2 are due to catalytic reduction of the halide. Our data show that the presence of CO2 is necessary to achieve a good catalytic efciency and stability of the catalyst. Addition of proton donors such as water, phenol, or triuoroethanol was indeed found to cause current enhancements similar to those reported in Table 1 for CO2. When controlled-potential electrolysis of 10 mM PhCH2Cl in argon-saturated MeCN was carried out in the presence of 1 mM NiII(salen), a complete deactivation of the catalyst was observed after a few catalytic cycles. About 42% of the benzyl chloride consumed under these conditions are transformed into toluene. On the other hand, when the same electrolysis was repeated in the presence of 1 M H2O or 15 mM PhOH, the yield of toluene increased to 84 and 90%, respectively, with no appreciable deactivation of the catalyst. No deactivation of the catalyst was observed also in the experiments performed in CO2-saturated MeCN (see below). Rapid deactivation of the catalyst complex with a low yield of toluene has been previously reported for the electrocatalytic reduction of PhCH2Cl by cobalt complexes with salen-type ligands [15,17,64]. Such a low catalytic efciency is attributable to some parasitic reaction, involving a reduction intermediate that destroys the catalyst. The reacting intermediate is most likely the carbanion and thus the role of proton donors or CO2 is to act as carbanion scavengers.

3.3. Electrocatalytic carboxylation of benzyl chlorides


Controlled-potential electrolyses in CO2-saturated acetonitrile solutions ([CO2]= 0.28 M [65]) containing 1 mM [NiII(L)] and about a 10-fold excess of RCl, were carried out at a potential corresponding to the reduction of NiII to NiI. The reduction is catalytic and proceeds with the formation of the appropriate carboxylate and toluene. As a rule, the electrolyses were interrupted after the conversion of the halide was theoretically achieved, i.e. after the passage of two electrons/molecule of RCl. The actual conversion ranged from 90 to 98%. Cyclic voltammograms recorded at the end of the electrolysis showed that no appreciable deactivation of the catalyst took place. The results of these electrolyses are reported in Table 2. Although the main product obtained with most halides is the corresponding carboxylic acid, the yield is strongly dependent on the structure of the halide. Whereas electron-withdrawing substituents on the phenyl ring increase the yield of RCO2H at the expense of that of RH, electron-donating groups produce the opposite effect (compare entries 24). The product distribution is also very sensitive to substituents on the a-carbon of the alkyl chloride. A comparison of entries 1, 3, and 6 shows that whereas a CH3 group causes a decrease of the yield, a phenyl group increases the RCO2H yield considerably. In the case of 1-chloro-1-phenylethane, the catalytic activity of the nickel complex was found to decrease during the electrolysis. Besides RCO2H and RH, which were formed with a total yield lower than 60%, small amounts of 2,3-diphenylbutane and styrene were also obtained. Moreover, the distribution of these products

Table 2 Catalytic reduction of benzyl chlorides (RX) in the presence of carbon dioxide a Entry RX Catalyst [catalyst]/mM [RCl]/mM E1/2 (R/R) b/V % RCO2H c % RH c 8 4 24 37 35 34 31 45 6 16 15

1 2 3 4 5 6 7 8 9 10 11
a

Ph2CHCl 4-CF3-PhCH2Cl PhCH2Cl 4-CH3O-PhCH2Cl 3-CH3O-PhCH2Cl PhCH(CH3)Cl e PhCH(CH3)Cl e,f PhCH2CH2Br g 4-CF3-PhCH2Cl PhCH2Cl 4-CH3O-PhCH2Cl

[NiII(L)] [NiII(L)] [NiII(L)] [NiII(L)] [NiII(L)] [NiII(L)] [NiII(L)] [NiII(L)] MCB MCB MCB

0.99 1.00 1.00 1.01 0.99 2.00 2.02 2.06 1.02 1.02 1.05

9.84 10.14 9.99 9.95 9.98 10.20 10.20 19.84 10.14 9.99 9.95

1.14 0.89 d 1.45 1.49 d 1.39 1.60 1.60 1.6 d 0.89 d 1.45 1.49 d

82 89 62 38 57 21 27 1 82 56 46

Controlled-potential electrolysis in CO2-saturated MeCN+0.1 M TBAP at a Hg pool electrode at E =1.7 V. From Ref. [58]. Reduction potentials are in MeCN, unless otherwise stated. c Yield is calculated with respect to converted RX (see Section 3.3). d DMF. e Small amounts of styrene and 2,3-diphenylbutane were also detected. f Applied potential: 2.0 V. g 1,4-Diphenylbutane (24%) and styrene (4%) were also obtained.
b

A. Gennaro et al. / Journal of Electroanalytical Chemistry 507 (2001) 124134

129

is affected by the electrolysis potential. The ratio RCO2H/RH increased from 0.62 at a reduction potential of 1.7 V to 0.87 at 2.0 V. Also the yields of 2,3-diphenylbutane and styrene decreased to barely detectable amounts when the electrolysis was performed at the most negative potential. In the case of a simple alkyl halide, such as 1-bromo-2-phenylethane, practically no electrocarboxylation takes place. We used the bromide because the reaction between [NiI(salen)] and alkyl chlorides does not take place on the voltammetric time scale. For example, the cyclic voltammogram of Ni(salen) is unaffected even after the addition of a 200-fold excess of 1-chloro-2-phenylethane. As shown in Table 2 (entry 8) the main products of reduction are ethylbenzene (45%) and 1,4-diphenylbutane (24%) together with styrene (4%). Only trace amounts of 3phenylpropionic acid were obtained. Further insight into the mechanism of the Ni(salen)catalyzed electrocarboxylation process was obtained by comparison with the result obtained through catalysis by an organic anion radical. The electrocatalytic reduction of benzyl halides by organic anion radical donors (D) has been previously investigated in some detail [66 69]. After electrogeneration of D (Eq. (5)), the main step of the reaction is a dissociative electron transfer from the anion radical to the organic halide, to yield X and the radical R (Eq. (6)). The anion radicals D may then react with R either by an electron transfer, to give the carbanion R (Eq. (7)), or by radical coupling, leading to alkylation of the mediator (Eq. (8)). Coupling between electrogenerated organic anion radicals and alkyl radicals is well documented in the literature [67,68,70,71]. The rate constant of reaction (7) depends on the difference between the standard reduction potentials of the radical R and the mediator, DE= E /R E . When R D/D DE is large, R is rapidly reduced to R and thus the overall process is a two-electron reduction of the halide to the corresponding carbanion. When DE decreases, reaction (8), which has a rate constant of the order of 109 M 1 s 1 [71], becomes more important. Consequently, since reaction (8) removes the catalyst from the process, the catalytic efciency decreases. In the presence of CO2, rapid formation of the carboxylate RCO 2 through the nucleophilic attack of the carbanion by CO2 is expected to take place (Eq. (9)), provided D itself does not undergo carboxylation. D +e ? D D + RX D+R +X D + R ? D +R D + R DR R + CO2 RCO
2

In the present context, methyl p-cyanobenzoate (MCB) is ideal as a mediator because, besides having the same standard potential as Ni(salen), E= 1.62 V, its radical anion does not undergo carboxylation [72]. The results of electrocarboxylations catalyzed by MCB are also included in Table 2 (entries 9 11). Comparison with the results obtained with the Ni(salen)-catalyzed reactions shows that, regardless of the catalyst, carboxylic acid forms as the main product and the yield depends on RCl. Two differences, however, can be noticed. First, the total yield of the process catalyzed by MCB tends to be lower than that catalyzed by Ni(salen). This may be due to a major contribution of the coupling reaction (Eq. (8)) in the MCB-catalyzed process, especially when reaction (7) becomes less exergonic. Second, when the electronwithdrawing power of the aryl substituent decreases, the ratio between the two products RCO2H and RH changes more rapidly for Ni(salen) than for MCB.

3.4. Mechanism of the nickel-catalyzed electrocarboxylation


Catalytic reduction of organic halides by macrocyclic NiI complexes has been the subject of several investigations [11,1947]. Although different reaction pathways have been proposed, an important result that emerges from these studies is that the initial reaction between nickel(I) complexes and alkyl halides is usually believed to form free radicals. This holds particularly true for the Ni(salen)-catalyzed reduction of a variety of organic halides in aprotic solvents [1927]. Very recently, through a comparative kinetic study, we have shown that the reduction of several benzyl halides by [NiI(salen)] proceeds through a dissociative electron transfer mechanism having some inner-sphere character [57]. This reaction leads to the formation of radical R and [NiII(L)] (Eq. (10)). [NiI(L)] + RCl [NiII(L)]+ R + Cl (10)

(5) (6) (7) (8) (9)

We have shown here that CO2 reacts very slowly with [NiI(L)], despite the large concentration of CO2 at saturation (0.28 M in MeCN [65]). Therefore, the rst step of the electrocarboxylation process is the rather fast reaction between [NiI(L)] and RCl (Eq. (10)) (for most of the benzyl chlorides investigated, log k10 is in the range 3.84.7 [57]). In principle, the so-formed radicals may undergo different reactions, namely radicalradical coupling (Eq. (11)), disproportionation to RH and olen (if b-hydrogen atoms are present) (Eq. (12)), H-atom abstraction from the solvent (Eq. (13)), and reduction either at the electrode (Eq. (2)) or by [NiI(L)] (Eq. (14)) to yield R. 2R RR 2R RH + R( H) (11) (12)

130

A. Gennaro et al. / Journal of Electroanalytical Chemistry 507 (2001) 124134

Fig. 4. Yields of RCOOH and RH as a function of the reduction potential of the benzyl radicals R (see text). R is: a = 4-CF3PhCH2, b = Ph2CH, c=3-CH3OPhCH2, d= PhCH2, e= 4-CH3OPhCH2, and f = PhCH(CH3). The reactions were catalyzed, in MeCN + 0.1 M TBAP at 25C, by either Ni(salen) (RCOOH, ; RH, ) or MCB (RCOOH, ; RH, ). The lines are intended to show the experimental trends.

R + CH3CN RH + CH2CN R +e ? R R + [NiI(L)] ? R +[NiII(L)]

(13) (2) (14)

In most cases, coupling products were either absent or detectable only at trace levels. This indicates that reaction (11) does not play a signicant role in the electrocatalytic process. The major reduction products were RCO and RH. The carboxylate ion, which is the 2 main product obtained with most of the halides investigated, is formed by nucleophilic attack of R on CO2. The hydrocarbon RH may form by protonation of R (by any available proton source in the reaction medium (Eq. (3)) and/or by H-atom abstraction from the solvent (Eq. (13)). A third reaction route which may contribute to the formation of the hydrocarbon is hydrolysis of an organonickel(II) complex formed by coupling of benzyl radicals with [NiI(L)] (Eqs. (15) and (16)). This would be in line with a similar mechanism proposed for the catalytic reduction of alkyl halides by nickel(I) complexes of porphyrin-type ligands [3234]. R + [NiI(L)] [NiII(L)(R)] [NiII(L)(R)] + H2O [NiII(L)] +RH + OH (15) (16)

The formation of organonickel complexes through rapid coupling between R and macrocyclic nickel(I) complexes has been suggested [28 31,34,73]. In particular, intermediate formation of [NiII(salen)(R)] through coupling between R and [NiI(salen)] has been proposed in previous studies on the electrocatalytic reduction of alkyl halides by Ni(salen) [24 26]. In this work,

we have observed a denite dependence of the distribution between the two main products, RCO2H and RH, on the nature of the catalyst (see Table 2, entries 24 and 911). With both catalysts, Ni(salen) and MCB, the ratio RH/RCO2H increases when the electron-withdrawing character of the aryl substituent decreases. However, despite the fact that the two catalysts have the same standard potential, RH formation increases more rapidly with Ni(salen) than with MCB. Although both the mediated processes lead to RH through some identical reaction pathways, the higher RH yield observed with Ni(salen) can be accounted for by the additional reaction route illustrated by Eqs. (15) and (16). A similar radical coupling reaction (Eq. (15)) may occur also with MCB, leading to an alkylation product. However, this reaction causes the deactivation of the catalyst instead of formation of RH in a following step. In principle, since the reduction of R to the corresponding carbanion R is essential for the carboxylation process, the decrease in the yield of RCO2H could be related to a diminished rate of carbanion formation. Carboxylate formation is not expected when the reduction of the radical is endergonic or when the reaction is so slow that it is outrun by more rapid radical reactions. Reduction potentials of transient organic radicals have been measured by photomodulation voltammetry in MeCN [7476]; E1/2 values measured by this technique are quite close to available estimates of E /R, R especially for benzyl radicals [74,75,77]. The reduction potentials of substituted benzyl radicals, obtained in MeCN or DMF, are shown in Table 2. The reduction potential of the 3-methoxybenzyl radical was estimated from the Hammett plot of the experimentally determined E1/2 values of ring-substituted benzyl radicals [75], on the basis of s. Inspection of the E1/2 values shows that, for all benzyl chlorides, reduction of R to R, either directly at the electrode (electrolyses were carried out at 1.7 V) or in solution by [NiI(L)] or MCB (E= 1.62 V), is thermodynamically favored, although with very different driving forces. However, since R is generated in the bulk (through reaction (6) or (10)), reactions other than (2), (7), and (14) may also occur. The experimental yields of RCOOH and RH are plotted in Fig. 4 as a function of E1/2 R/R. At rst glance, the plot suggests that indeed some of the radicals escape reduction for those substrates characterized by more negative E1/2 R/R values. However, there is enough information to rule out radical pathways, particularly for the four benzyl chlorides and Ph2CHCl. This stems from a comparison between the lifetimes of the radical with respect to the possible decay pathways. Reaction (13) is believed to be slow because MeCN is known to be per se a poor H-atom donor, particularly if one considers that benzyl radicals

A. Gennaro et al. / Journal of Electroanalytical Chemistry 507 (2001) 124134

131

react slowly even with good H-donors (for example, the rate constant for H-atom abstraction from PhSH is only 3.1105 M 1 s 1 [78]). Consequently, as far as reaction (13) is concerned, a lower limit for the lifetime of R is likely to be 10 4 s. Coupling of radicals (Eq. (12)), however, is a fast reaction. On the other hand, if we take into account the average steady-state R concentration that is present in solution (note that we are considering indirect electrolyses, where R is generated in the bulk), the lifetime of R can be estimated to be not shorter than 10 5 10 4 s. If, during this lifetime, reduction by another donor molecule is more efcient (Eq. (7), [NiI(L)] or MCB), carbanion chemistry ensues. This calculation can be made by means of the Marcus equation, DG " =DG " (1 + DG/4DG " )2, 0 0 where DG " is the activation Gibbs energy, DG is the Gibbs energy, and DG " is the intrinsic barrier of the 0 electron-transfer step [79]. In our specic case, the contribution by the benzyl radical/carbanion couple to DG " seems to be particularly small [77]. Therefore, we 0 adopt a value of 17 kJ mol 1, which is essentially the same as that of typical self-exchange reactions of aromatic compounds. By assuming the reaction to be adiabatic, we can now estimate the rate constants for electron transfer between R and a donor having E= 1.62 V. The result (pre-exponential Z = 31011 M 1 s 1, diffusional rate kd =1 1010 M 1 s 1) is that for most compounds the rate constant is almost diffusional. Even for PhCH(CH3), whose reduction occurs essentially at zero driving force, the value is still 5108 M 1 s 1. Under our experimental conditions, the average lifetime of R (taking into account the average donor concentration in the reaction layer) may be estimated to be 10 7 10 5 s, depending on the specic radical. Therefore, it appears very likely that essentially all of the benzylic radicals generated in the initial dissociative electron transfer undergo fast reduction. In principle, some radical/anion competition may be present only in the case of PhCH(CH3), which would indeed account for traces of the dimer when the electrolysis is carried out at 1.7 V. Finally, some heterogeneous reduction of the radical is also expected, as indicated by the virtual absence of dimers in the electrolysis of PhCH(CH3)Cl when the applied potential is shifted to 2.0 V. This result further strengthens the importance of carbanion formation over the radical pathways. Carbanions therefore seem to be the main reactive species leading to the two principal products in the mediated reduction-carboxylation of benzyl chlorides. Protonation of the benzyl carbanion by water must be considered as the best way to form RH because of the relatively high concentration (ca. 0.02 M) of residual water in the formally aprotic MeCN. In fact, this proton transfer is a fast reaction [80], the rate constant in MeCN being estimated to be ca. 1 108 M 1 s 1.

We may now apply the Marcus theory [81] to the proton transfer depicted in Eq. (3). The driving force of the reaction is the difference between the pKa values of the two acidbase couples exchanging the proton. The pKa values of the toluenes in MeCN were previously estimated [75]. The pKa of water in the same solvent may be estimated to be ca. 43 (from a correlation between the pKa values in MeCN and DMSO [82]). From the k value, the driving force, and the usual pre-exponential factor, the intrinsic barrier can be estimated to be 39 kJ mol 1. This is a reasonable value since we are considering the proton transfer between a carbon base and an OH-acid in a dipolar aprotic solvent [83,84]. With this intrinsic barrier, it is now possible to estimate the protonation rate constant for the whole series of benzylic carbon bases considered here. Turning now to the rate of the competitive reaction, the carboxylation of the anions, little information is available. However, the following rate constants have been measured for some radical anions: 5.0103 M 1 s 1, anthracene [85], 8.9 105 M 1 s 1, benzophenone [86], 3.1 106 M 1 s 1 phenanthrene [87]. In the case of carbanions, the carboxylation rate constant is expected to be much larger. For example, we can estimate a value larger than 5107 M 1 s 1 for the carbanion-enolate electrogenerated from a-bromo isobutyramides [88,89]. By using the latter estimate (which refers to carboxylation of a tertiary carbon) and taking into account the concentration of CO2, we can safely estimate the lifetime of the benzyl carbanion in reaction (9) to be B 7 10 8 s. At this point, however, it is useful to employ the relative RCOOH/RH yields to address the problem of the product competition better. By using the protonation rate constants (calculated as described above and by taking into account the diffusional limit), the average water concentration, and the relative yields, one can estimate the pseudo-rst order rate constants for the carboxylation of R from the ratio of the RCOOH/ RH yields. As we will see, this hypothesis holds only for MCB, which may lead to RH formation only through reaction (3). By using this procedure, which assumes both processes to be governed by pseudo rst-order kinetics, we may sketch, in principle, an activation/driving force plot for both protonation and carboxylation. Concerning the protonation driving force, we should note that the electrolyses were carried out in CO2-saturated MeCN. Therefore, we expect the proticity of the medium to be enhanced to some extent by the presence of carbon dioxide. For the sake of argument, a minimum value of ca. 23 for the pKa of carbonic acid may be obtained if one considers the difference between the pKa values of OH-acids in MeCN and water [90,91]. However, instead of relying on either of the above very different pKa values, we preferred to use the pKa as an

132

A. Gennaro et al. / Journal of Electroanalytical Chemistry 507 (2001) 124134

adjustable parameter varying between 23 and 43. This is a more practical approach because it also takes into account that MeCN itself might act as a proton donor [92], although we believe that this is a minor contribution for our experimental conditions. Since both the proton-transfer and the carboxylation rate constants ought to be increasing functions of the driving force, we found that meaningful values for the MCB-catalyzed reactions could be obtained only for an effective pKa equal to or larger than 36. On the other hand, the carboxylation rate constant cannot be too small (see above). In fact, given the large difference between the pKas of the corresponding CH-acids, the benzyl carbanion is expected to be a stronger nucleophile than the carbanion-enolate of an isobutyramide. Consequently, even pKa values larger than 36 37 appear unreasonable because the carboxylation rates would be too slow. The pseudo rst-order kinetic data corresponding to pKa = 36 are illustrated in Fig. 5. It is useful to start from the results obtained for MCB, which acts as a simple outer-sphere electron donor. The different sensitivity of the two competitive reactions to the nature of the benzylic carbanion, illustrated by the two solid lines in Fig. 5, can be related to the two intrinsic barriers. In fact, the carboxylation process involves a particularly large inner reorganization because of both its geometric requirement and the need to modify the O C O angle. As a matter of fact, we have found evidence of the relevance of the latter term in carboxylation decarboxylation reactions by

studying the dissociative oxidation of oxalate, an electron-transfer process accompanied by a rather large inner reorganization term [93]. In conclusion, when we increase the driving force of both competitive reactions, it turns out that the basicity is enhanced more than the nucleophilicity, resulting in a relative decrease of carboxylation. On the other hand, the picture is not as simple in the case of Ni(salen), and indeed Fig. 5 shows an apparent decrease of the carboxylation rate (dashed line) when the driving force is increased. This is an artifact, because the carboxylation pseudo rst-order rate constants are calculated with reference to the protonation rates, based on the relative yields. It is evident from the RH yields that the protonation rate is more signicant for the Ni(salen)-catalyzed reaction than for the MCB case. In other words, the solid line of Fig. 5 underlying the protonation activation/driving force relationship is valid only for the uncomplicated MCB-mediated process. As was already mentioned, the mechanism for this increased RH formation may be related to the hydrolysis of the [NiII(L)(R)] intermediate (Eqs. (15) and (16)). Our observations are, once again, in agreement with some specic interaction between R and NiI(salen), by analogy with the observations of other workers [3234]. One nal remark concerns the reduction of 1-bromo2-phenylethane. We found that the catalytic reduction leads to no appreciable carboxylation (Table 2, entry 8). Although the potential of PhCH2CH2 reduction is unknown, it has been shown that for primary alkyl radicals E1/2 is not very sensitive to the length of the alkyl chain [77]. Therefore, we assume that the E1/2 for reduction of PhCH2CH2 corresponds to the average of the values reported for other primary alkyl radicals ( 1.6 V [77]). It should also be noted that, upon reduction, alkyl radicals undergo a more signicant reorganization than benzyl radicals [77,94]. The reduction rate of PhCH2CH2 is thus expected to be signicantly lower than that of a faster R/R couple such as PhCH(CH3)/PhCH(CH3), which has essentially the same E. Thus, although reduction of PhCH2CH2 is thermodynamically feasible in our reaction conditions, the rate is too slow to compete with the radical reaction pathways (11)(13), leading to phenylethane, 1,4diphenylbutane, and styrene.

Fig. 5. Apparent pseudo-rst order rate constants for the protonation ( ) and carboxylation (Ni(salen), ; MCB, ) of benzyl carbanions R. R is: a =4-CF3PhCH2, b= Ph2CH, c= 3-CH3OPhCH2, d= PhCH2, e= 4-CH3OPhCH2, and f= PhCH(CH3). The source data (RH and RCOOH yields) were obtained in MeCN +0.1 M TBAP at 25C. Whereas the protonation values were calculated from the corresponding activation/driving force relationship, the carboxylation values were estimated through the ratio between the RCOOH and RH yields (see text). The lines (solid for protonation and MCB-mediated carboxylation and dashed for the Ni(salen)-mediated carboxylation) are intended to show the experimental trends.

4. Conclusions Electrochemical carboxylation of benzyl chlorides in MeCN is catalyzed by [NiI(salen)]. Very high yields of carboxylic acids are obtained when electron-withdrawing substituents are present on the phenyl ring or on the a-carbon of the organic halide. The benzyl radicals, formed by the dissociative reaction between the electrogenerated nickel(I) complex and RCl, are easily reduced

A. Gennaro et al. / Journal of Electroanalytical Chemistry 507 (2001) 124134

133

to the corresponding carbanions, the key intermediates of the electrocatalytic process. R is then captured by CO2, to yield RCO, in competition with protonation 2 by residual water, to yield RH. A particularly anhydrous solvent is thus necessary to carry out the electrocarboxylation process efciently, at least for the less activated benzyl chlorides, which are particularly powerful probases. Only for the R/R redox couples with the most negative potentials does some radical reaction seem to take place. Halides that give rise to simple alkyl radicals are not carboxylated by the Ni(salen) catalytic system. Our observations also suggest some specic interaction between R and NiI(salen), in analogy with what we found for the reaction between the precursor RCl and NiI(salen) [57]. Although the transient interaction between R and NiI(salen) leads to some decrease of the efciency of the carboxylation, it might be possible to exploit these transient interactions for specic synthetic goals.

Acknowledgements This work was nancially supported by the Consiglio Nazionale delle Ricerche (CNR) and the Ministero dellUniversita e della Ricerca Scientica e Tecnologica ` (MURST).

References
[1] M.M. Baizer, J.L. Chruma, J. Org. Chem. 37 (1972) 1951. [2] J.H. Wagenknecht, J. Electroanal. Chem. 52 (1974) 489. [3] G. Silvestri, S. Gambino, G. Filardo, A. Gulotta, Angew. Chem., Int. Ed. Engl. 23 (1984) 979. [4] G. Silvestri, S. Gambino, G. Filardo, Enzymatic Model Carboxylation Reduction Reactions, Carbon Dioxide Utilisation. In: NATO ASI Series C, vol. 314, Reidel, Dordrecht, 1990, p. 101. [5] N.S. Murcia, D.G. Peters, J. Electroanal. Chem. 326 (1992) 69. [6] J.F. Fauvarque, A. Jutand, M. Francois, New J. Chem. 10 (1986) 119. [7] J.F. Fauvarque, A. Jutand, M. Francois, J. Appl. Electrochem. 18 (1988) 109. [8] J.F. Fauvarque, Y. de Zelicourt, C. Amatore, A. Jutand, J. Appl. Electrochem. 20 (1990) 338. [9] C. Amatore, A. Jutand, J. Am. Chem. Soc. 113 (1991) 2819. [10] K. Osakada, R. Sato, T. Yamamoto, Organometallics 13 (1994) 4645. [11] S. Olivero, E. Dunach, Eur. J. Org. Chem. (1999) 1885. [12] S. Torii, H. Tanaka, T. Hamatani, K. Morisaki, A. Jutand, F. Puger, J.F. Fauvarque, Chem. Lett. (1986) 169. [13] C. Amatore, A. Jutand, F. Khalil, M.F. Nielsen, J. Am. Chem. Soc. 114 (1992) 7076. [14] J.C. Folest, J.M. Duprilot, J. Perichon, Y. Robin, J. Devynck, Tetrahedron Lett. 26 (1985) 2633. [15] A.A. Isse, A. Gennaro, E. Vianello, J. Chem. Soc., Dalton Trans. (1996) 1613. [16] G. Zheng, M. Stradiotto, L. Li, J. Electroanal. Chem. 453 (1998) 79. [17] A.A. Isse, A. Gennaro, E. Vianello, J. Electroanal. Chem. 444 (1998) 241.

[18] O. Scialdone, G. Filardo, A. Galia, D. Mantione, G. Silvestri, Acta Chem. Scand. 53 (1999) 800. [19] C. Gosden, K.P. Healy, D. Pletcher, J. Chem. Soc., Dalton Trans. (1978) 972. [20] C. Gosden, D. Pletcher, J. Organomet. Chem. 186 (1980) 401. [21] J.Y. Becker, J.B. Kerr, D. Pletcher, R. Rosas, J. Electroanal. Chem. 117 (1981) 87. [22] M.S. Mubarak, D.G. Peters, J. Electroanal. Chem. 332 (1992) 127. [23] M.S. Mubarak, D.G. Peters, J. Electroanal. Chem. 388 (1995) 195. [24] C.E. Dahm, D.G. Peters, J. Electroanal. Chem. 406 (1996) 119. [25] A.L. Butler, D.G. Peters, J. Electrochem. Soc. 144 (1997) 4212. [26] D. Bhattacharya, M.J. Samide, D.G. Peters, J. Electroanal. Chem. 441 (1998) 103. [27] A.J. Fry, P.F. Fry, J. Org. Chem. 58 (1993) 3496. [28] A. Bakac, J.H. Espenson, J. Am. Chem. Soc. 108 (1986) 713. [29] A. Bakac, J.H. Espenson, J. Am. Chem. Soc. 108 (1986) 719. [30] M.S. Ram, A. Bakac, J.H. Espenson, Inorg. Chem. 25 (1986) 3267. [31] M.S. Ram, J.H. Espenson, Inorg. Chem. 25 (1986) 4115. [32] A.M. Stolzenberg, M.T. Stershic, J. Am. Chem. Soc. 110 (1988) 5397. [33] G.K. Lahiri, L.J. Schussel, A.M. Stolzenberg, Inorg. Chem. 31 (1992) 4991. [34] G.K. Lahiri, A.M. Stolzenberg, Inorg. Chem. 32 (1993) 4409. [35] M.C. Helvenston, C.E. Castro, J. Am. Chem. Soc. 114 (1992) 8490. [36] S. Ozaki, T. Nakanishi, M. Sugiyama, C. Miyamoto, H. Ohmori, Chem. Pharm. Bull. 39 (1991) 31. [37] S. Ozaki, H. Matsushita, H. Ohmori, J. Chem. Soc., Chem. Commun. (1992) 1120. [38] S. Ozaki, H. Matsushita, H. Ohmori, J. Chem. Soc., Perkin Trans. 1 (1993) 649. [39] S. Ozaki, H. Matsushita, H. Ohmori, J. Chem. Soc., Perkin Trans. 1 (1993) 2339. [40] S. Ozaki, I. Horiguchi, H. Matsushita, H. Ohmori, Tetrahedron Lett. 35 (1994) 725. [41] S. Ozaki, H. Matsushita, M. Emoto, H. Ohmori, Chem. Pharm. Bull. 43 (1995) 32. [42] M. Ihara, A. Katsumata, F. Setsu, Y. Tokunaga, K. Fukumoto, J. Org. Chem. 61 (1996) 677. [43] J.C. Clinet, E. Dunach, J. Organometal. Chem. 503 (1995) C48. [44] S. Olivero, J.C. Clinet, E. Dunach, Tetrahedron Lett. 36 (1995) 4429. [45] S. Olivero, J.-P. Rolland, E. Dunach, Organometallics 17 (1998) 3747. [46] E. Dunach, A.P. Esteves, A.M. Freitas, M.J. Medeiros, S. Olivero, Tetrahedron Lett. 40 (1999) 8693. [47] F.C.J.M. van Veggel, M. Bos, W. Verboom, D.N. Reinhoudt, Recl. Trav. Chim. Pays-Bas 109 (1990) 515. [48] C. Gosden, J.B. Kerr, D. Pletcher, R. Rosas, J. Electroanal. Chem. 117 (1981) 101. [49] A.I. Vogel, Practical Organic Chemistry, Longmans, London, 1964. [50] C.E. Frank, J.R. Leebrick, L.F. Moormeier, J.A. Scheben, O. Homberg, J. Org. Chem. 26 (1961) 307. [51] D.H. Richards, N.F. Scilly, J. Chem. Soc. (C) (1969) 55. [52] S. Antonello, M. Musumeci, D.D.M. Wayner, F. Maran, J. Am. Chem. Soc. 119 (1997) 9541. [53] J.-M. Saveant, in: P.S. Mariano (Ed.), Advances in Electron Transfer Chemistry, vol. 4, JAI Press, Greenwich, 1994, p. 53. [54] S. Antonello, F. Maran, J. Am. Chem. Soc. 119 (1997) 12595. [55] S. Antonello, F. Maran, J. Am. Chem. Soc. 121 (1999) 9668. [56] A. Cardinale, A. Gennaro, A.A. Isse, F. Maran, in preparation. [57] A. Gennaro, A.A. Isse, F. Maran, in preparation.

134

A. Gennaro et al. / Journal of Electroanalytical Chemistry 507 (2001) 124134 [76] B.A. Sim, P.H. Milne, D. Griller, D.D.M. Wayner, J. Am. Chem. Soc. 112 (1990) 6635. [77] H. Lund, K. Daasbjerg, T. Lund, D. Occhialini, S.U. Pedersen, Acta Chem. Scand. 51 (1997) 135. [78] J.A. Franz, N.K. Suleman, M.S. Alnajjar, J. Org. Chem. 51 (1986) 19. [79] R.A. Marcus, N. Sutin, Biochim. Biophys. Acta 811 (1985) 265. [80] B. Bockrath, L.M. Dorfmann, J. Am. Chem. Soc. 96 (1974) 5708. [81] R.A. Marcus, J. Phys. Chem. 72 (1968) 891. [82] R.L. Donkers, F. Maran, D.D.M. Wayner, M.S. Workentin, J. Am. Chem. Soc. 121 (1999) 7239. [83] C.D. Ritchie, J. Am. Chem. Soc. 91 (1969) 6749. [84] O.A. Reutov, I.P. Beletskaya, K.P. Butin, CH-Acids, Pergamon Press, Oxford, 1978, p. 170 (ch. 5). [85] E.A. Ticianelli, L.A. Avaca, E.R. Gonzalez, J. Electroanal. Chem. 258 (1989) 379. [86] A. Gennaro, A.A. Isse, unpublished results. [87] A.V. Nikolaitchik, M.A.J. Rodgers, D.C. Neckers, J. Org. Chem. 61 (1996) 1065. [88] F. Maran, M. Fabrizio, F. DAngeli, E. Vianello, Tetrahedron 44 (1988) 2351. [89] F. Maran, S. Rofa, M.G. Severin, E. Vianello, Electrochim. Acta 35 (1990) 81. [90] K. Izutsu, Acid-Base Dissociation Constants in Dipolar Aprotic Solvents, Blackwell Scientic Publications, Oxford, 1990. [91] Handbook of Chemistry and Physics, 77th ed., CRC Press, Boca Baton, 1996, Ch. 8, p. 45. [92] D.A. Koch, B.J. Henne, D.E. Bartak, J. Electrochem. Soc. 134 (1987) 3062. [93] A.A. Isse, A. Gennaro, F. Maran, Acta Chem. Scand. 53 (1999) 1013. [94] C.P. Andrieux, I. Gallardo, J.-M. Saveant, J. Am. Chem. Soc. 111 (1989) 1620.

[58] K. Daasbjerg, S.U. Pedersen, H. Lund, in: Z.B. Alfassi (Ed.), General Aspects of the Chemistry of Radicals, Wiley, New York, 1999, p. 385. [59] B. Fischer, R. Eisenberg, J. Am. Chem. Soc. 102 (1980) 7361. [60] M. Beley, J.-P. Collin, R. Ruppert, J.-P. Sauvage, J. Am. Chem. Soc. 108 (1986) 7461. [61] D.J. Pearce, D. Pletcher, J. Electroanal. Chem. 197 (1986) 317. [62] F. Abba, G. De Santis, L. Fabbrizzi, M. Licchelli, A.M.M. ` Lanfredi, P. Pallavicini, A. Poggi, F. Ugozzoli, Inorg. Chem. 33 (1994) 1366. [63] E. Fujita, J. Haff, R. Sanzenbacher, H. Elias, Inorg. Chem. 33 (1994) 4627. [64] F. Bedioui, Y. Robin, J. Devynck, C. Bied-Charreton, J. Organomet. Chem. 326 (1987) 117. [65] A. Gennaro, A.A. Isse, E. Vianello, J. Electroanal. Chem. 289 (1990) 203. [66] L.A. Avaca, E.R. Gonzalez, E.A. Ticianelli, Electrochim. Acta 28 (1983) 1473. [67] T. Lund, H. Lund, Acta Chem. Scand. 41 (1987) 93. [68] R. Fuhlendorf, D. Occhialini, S.U. Pedersen, H. Lund, Acta Chem. Scand. 43 (1989) 803. [69] Y. Huang, D.D.M. Wayner, J. Am. Chem. Soc. 116 (1994) 2157. [70] J. Simonet, M.-A. Michel, H. Lund, Acta Chem. Scand. B29 (1975) 489. [71] S.U. Pedersen, T. Lund, K. Daasbjerg, M. Pop, I. Fussing, H. Lund, Acta Chem. Scand. 52 (1998) 657. [72] A. Gennaro, A.A. Isse, J.-M. Saveant, M.G. Severin, E. Vi anello, J. Am. Chem. Soc. 118 (1996) 7190. [73] D.M. Guldi, M. Kumar, P. Neta, P. Hambright, J. Phys. Chem. 96 (1992) 9576. [74] D.D.M. Wayner, D.J. McPhee, D. Griller, J. Am. Chem. Soc. 110 (1988) 132. [75] B.A. Sim, D. Griller, D.D.M. Wayner, J. Am. Chem. Soc. 111 (1989) 754.

Вам также может понравиться