Anodic Methoxylation of Cinnamate Esters

Este trabalho descreve a oxidação eletroquímica do 4-metóxi-cinamato de metila e de outros ésteres metílicos do ácido cinâmico die trimetoxilado, em solução metanólica neutra ou alcalina. Nos estudos eletroquímicos foram empregadas voltametria cíclica, bem como eletrólises preparativas a potencial ou corrente controlados, em celas de um ou dois compartimentos. A influência de parâmetros tais como densidade de corrente, tipo de cela, material do eletrodo, solvente, concentração do substrato e base foi investigada para o 4-metóxi-cinamato de metila. Os produtos observados foram resultantes de metoxilações nucleares e/ou na cadeia lateral, entretanto com o aumento do número de metoxilas no anel aromático foram observados principalmente produtos referentes à metoxilação no anel aromático. As estruturas dos produtos também dependem do tipo de procedimento empregado no seu isolamento.


Introduction
The oxygenation of aromatic rings by non-electrochemical methods usually involves several steps.Anodic methoxylation of aromatic rings not only furnishes novel addition products of synthetic utility but also serves as a direct method for introducing methoxy groups into aromatic substrates.A great variety of substituted 1,4-dimethoxybenzenes 1 , naphthalenes 2 , and benzothio-phenes 3 were converted into the corresponding quinone bisketals in high yields under single-cell and constant current conditions.The quinone monoketal, obtained through mild acid hydrolysis of the bisketal, has found extensive use in organic synthesis 1a,4,5 .
A synthetic application of this method is the anodic methoxylation of methyl eugenol where 1,2-and 1,4dimethoxy addition products are formed.Reaction of this product mixture with acid affords a convenient synthetic route to γ-asarone: 5-(2-propenyl)-1,2,4-trimethoxy-benzene 6 .A reaction competing with the oxidative addition of methanol to the aromatic ring is the oxidation of benzylic positions, especially those para to the methoxy group.The anodic oxidation of methyl eugenol in methanol without added base gave the benzylic substitution product in 60% yield 6b .
Here we report the results of the electrochemical oxidation of cinnamates 1 -6 and the dihydro derivative 7 in methanolic solution.

Results and Discussion
Table-1 shows the oxidation peak potentials (Epa) for 1 -7, determined by cyclic voltammetry.The oxidation waves were irreversible and a considerable current decrease was observed during the second cycle.The Epa values were measured in acetonitrile.
Electrochemical oxidation of the aromatic ethers will involve electron transfer from the π-system as normally observed for oxidation of aromatic compounds to radical cations 7 .According to Table 1, the ease of oxidation, related to Epa1 is in the following order: e-mail: vpardini@iq.usp.br In the mono methoxylated compounds series, the meta substituted compound 2 shows a more positive oxidation potential than methoxybenzene 8 , but for the ortho and para substituted substrates 1 and 3 the first oxidation peak potentials are less positive.This difference can be explained considering the presence of the unsaturated side chain which allows a more effective electron delocalization in the intermediate radical cation when ortho or para in relation to the methoxy substituent whereas in the meta position its electron-withdrawing character seems to predominate, destabilizing the intermediate.A similar stabilizing effect is also evident when the oxidation potentials of compounds 1 and 7 are compared.
The presence of two methoxy groups in substrates 4 and 5 causes a negative shift of the first oxidation peak potentials when compared with compounds 1 and 3. A similar effect is observed in the case of the dimethoxybenzenes in relation to methoxybenzene, but being about 100 mV larger for the 1,4-derivative than for the two other isomers.In our case although the relative position of the methoxy substituents is para in compound 5, the negative shift of its oxidation potential is ca 50 mV smaller than that observed for 4. It seems reasonable to suggest that in compound 5 the carbon-carbon double bond destabilizes the radical cation formed, thus the oxidation potential is shifted to a more positive value than that expected for 1,4-dimethoxybenzene.This result could be accounted for by a steric ortho effect although for compounds 1 and 3 this effect seems not to be operative.In the case of substrate 4 the negative shift of the first oxidation peak potential shows to be similar to that observed when 1,2-dimethoxybenzene is compared with methoxy benzene 8 .On the other hand a comparison of the second oxidation peak potential of substrates 4 and 5 is noteworthy and shows a reverse situation, the latter being oxidized at a potential ca 150 mV less positive than 4. Compound 6 under cyclic voltammetry conditions (Table 1) presents its oxidation peak potential shifted to a less positive value when compared with substrate 1 and the observed difference is approximately the same as reported for methoxybenzene and 1,2,3-trimethoxybenzene 8 .
The preparative electrolyses were performed in MeOH containing sodium methoxide.The reason for choosing this condition was the observation that in MeOH or MeCN containing solely NaClO 4 the electrolyses afforded untreatable mixtures of products.
Compound 1 was electrolyzed under controlled current conditions in a methanolic solution containing base.After passage of the appropriate amount of electricity, the cell contents were worked-up and the products isolated by column chromatography or by preparative TLC.The results of these electrolyses are shown in Table 2. Conditions: 0.2 mol L -1 NaClO 4 in MeOH solution; undivided cell; Pt foil anode (2.5 x 2.5 cm); W wire cathode and charge = 4 F mol -1 .a-S.M.: Recovered Starting Material; b-the methanolic fraction obtained from chromatographic column (polymeric material) varied from 14 to 28%; cchromatographic yields using biphenyl ether as internal standard.
After the passage of 4 F mol -1 (current density: 8 mA cm -2 ) in an undivided cell, 1 (entry 1, Table 2) afforded 7, formed by reduction of 1 at the cathode, and 10 and 11, obtained through the olefinic moiety methoxylation, besides recovered starting material.Upon increasing the current density (entry 2, Table 2), product 9 was observed, together with small amounts of p-anisaldehyde 8, formed probably through hydrolysis of 9 during work-up.It is worth mentioning that these results are different from those observed in an experiment conducted in a divided cell.Under these conditions 7 was not formed, being 8 (29%) and 9 (49%), Cα-Cβ bond cleavage products 9 of 1, the main products.A disadvantage of this procedure was a decrease of pH of the reaction medium in the anodic compartment during the electrolysis, and it was necessary to add base to the solution at regular intervals.When substrate concentration was decreased three times (entry 3, Table 2), competition between nuclear and side chain methoxylation was detected.The yield of the nuclear methoxylated product, 12, was increased with larger base/ substrate ratio (entry 3 and 4, Table -2), indicating that these conditions favor the aromatic ring methoxylation.Compound 13 was isolated after the crude reaction mixture passed through the column chromatography, indicating an acidic hydrolysis of compound 12.
The results obtained in the electrochemical oxidation of compound 1 deserve some further comments.It is well known 10 that the anodic oxidation of 4methylanisole occurs preferentially at the aromatic ring under conditions where formation of methoxyl radicals also takes place (platinum anode, supporting electrolyte NaOMe).This is observed when low current densities and a 1:1 base/substrate ratio or higher current densities and a 4:1 base/substrate ratio are used, but in the presence of LiBF 4 as supporting electrolyte the side chain methoxylation becomes more competitive.In our case NaClO 4 was used besides variable concentrations of NaOMe.It seems that these conditions favor the formation of side chain substitution and cleavage products (entries 1 and 2, Table 2), the first most probably precursors of the latter as with higher current densities (entry 2) the cleavage products are formed.When higher base/substrate ratios were employed (entries 3 and 4, Table 2) nuclear methoxylation became a competitive pathway in relation to the side chain substitution.It is reasonable to suggest that the electron delocalization in the intermediate radical cation involving the side chain carbon-carbon double bond turns the latter more prone to react with methoxyl radicals therefore competing efficiently with the aromatic nucleus.
If the electrolyses were conducted without base, black polymeric materials were obtained, suggesting that the presence of methoxy radicals formed during the oxidation play an important role in the methoxylation of the intermediate like radical cations as pointed out previously 10 .Without base the intermediates seem to react with each other leading to compounds with higher molecular weight.
To verify the effect of the electrode material on the course of the reaction, platinum was substituted for graphite but under these conditions a dark solution containing a complex mixture of products was obtained.Probably the substrate is preferentially adsorbed at the electrode surface producing intermediates with reactive sites at the aromatic ring and the side chain, which react with each other producing a complex mixture 10 .
The oxidation of 2 and 3 was conducted under similar conditions used for 1, and in all cases complex mixtures containing a tar like material together with starting material were obtained.Even when the charge was increased from 4 to 8 F mol -1 , the starting material was recovered with 2 (32%) and 3 (20%).A different behavior was observed with the dimethoxylated 4 and 5 when compared to ortho 3 and meta 2 derivatives.Table 3 shows the oxidation products of 4 and 5 obtained under different experimental conditions.
Compound 4 afforded the trimethoxyderivative 14 after acidic work-up conditions (entries 1 and 2, Table 3).When the work-up was conducted under neutral conditions compounds 15 and 16, the 1,2-and 1,4addition products, were isolated (entry 3, Table 3).This difference of products observed when different work-up procedures were employed became clear when it was shown that 15 and 16 were transformed into 14 when treated with MeOH/H 2 SO 4 solution.Similar results were previously reported for the anodic methoxylation of methyl eugenol 6 where 1,2-and 1,4-addition products were also isolated under basic work-up conditions 6b and a trimethoxybenzene derivative under acidic work-up 6a .Wang and Swenton 6b showed that the 1,2-and 1,4addition products are transformed into the trimethoxybenzene derivative when treated with acid.
Compound 5 is a 1,4-dimethoxybenzene derivative and the electrolysis conducts to the corresponding quinone bisketal like 17. Product 17 was previously obtained in 46% yield by Swenton et al. 2 electrolyzing 5 at controlled potential in CH 3 OH/KOH solution and undivided cell.According to Table 3 (entries 4, 5 and 6) the electrolysis product was mainly 17 and after its isolation, the product was stirred in an acid solution affording 18, an aromatic compound.
When 6 was electrolyzed at constant current, a very complex mixture was obtained together with polymeric material.This compound is the easiest oxidizable compound (1.80 V) of the series studied and presumably a further oxidation is occurring, leading to a complex mixture.
To evaluate the conjugated double bond effect of the side chain, 7 was electrolyzed at constant current, undivided cell and basic conditions to afford the methoxylated product at the Cα position, 19 (30%) and the 1,4-nuclear addition product 20 (41%).

Conclusions
The results presented here indicate that depending on the electrolysis conditions, the anodic oxidation of 1 led to both aromatic ring and side chain methoxylated compounds.The best yields of nuclear methoxylated product 12 were obtained when larger base/substrate ratios were used.The presence of base showed to be important otherwise black polymeric materials were obtained.These findings agree with the EECrCp mechanism suggested by Dolson and Swenton 11 in which the involvement of methoxy radicals plays an important role.For 7, although less reactive than 1, methoxylation at the aromatic ring and/or side chain was also observed.Compound 2 has been shown to be less reactive than other monomethoxylated compounds and 6 was more reactive but afforded a complex mixture of products.Dimethoxylated compounds 4 and 5 yielded mainly aromatic ring methoxylated products whose structures depend upon the work-up procedures.Under basic or neutral conditions cyclohexadiene derivatives (15, 16 or 17) are isolated whereas in acidic medium trimethoxybenzene derivatives (14 or 18) are obtained.It is noteworthy that products 15 and 16 can be converted into 14, and also compound 18 can be easily obtained from product 17, in good yields stirring them in MeOH/H 2 SO 4 solution at room temperature.

Experimental
NMR spectra ( 1 H and 13 C) were recorded on Brukers AC-200, DPX-300 and DRX-500 CDCl 3 solutions.Chemical shifts are reported as δ values relative to   tetramethylsilane (TMS) as internal reference.GLC analyses were performed on a HP5890 instrument, (fused silica capillary column SE-30, 12m x 0.2mm) and mass spectra were recorded on a Hewlett Packard HP 5988A GLC/MS instrument at 70 eV (EI).Some of the products have been identified by comparison to authentic samples.

Electrochemical Experiments
Cyclic voltammetry experiments were run in a three compartment cell using a Princeton Applied Research (PAR) 173 potentiostat/galvanostat.The reference electrode was Ag/AgI (a silver wire immersed in an acetonitrile solution 0.1mol L -1 in NaClO 4 and 0.04mol L -1 in tetrabutylammonium iodide), platinum bead as working electrode and platinum foil as counter electrode.Solvent supporting electrolyte was 0.1 mol L -1 NaClO 4 in acetonitrile, the substrate concentration was 3 mmol L -1 , and the sweep rate was 200 mV s -1 .

General procedure for the electrolyses
The compounds 1 -7 (3.5 -5.5 mmol) were electrolyzed in two different cells: undivided and divided cell made of glass (35 or 70 mL) and equipped with a platinum foil (2.5 x 2.5 cm) or a graphite plate (3.5 x 4.5 cm) as working electrode and a tungsten wire as counter electrode.The electrolyses were conducted at constant current or controlled potential (see Tables 2 and 3).The solvent was dry methanol containing NaClO 4 as supporting electrolyte or sodium methoxide in methanol (see details in Tables 2 and 3).After the passage of the appropriate amount of electricity the cell contents were worked up using 3 different procedures: i) The solvent was evaporated until dryness under reduced pressure and a solution of 0.1mol L -1 H 2 SO 4 in methanol was added until pH 7.0.The mixture was extracted with dichloromethane or ethyl ether (3x50 mL), washed with water (3x10mL) and dried over magnesium sulfate (neutral procedure).ii) The solvent was evaporated and the residue was extracted directly, washed and dried (basic procedure).iii) After the evaporation of the solvent, the mixture was acidulated (pH=2) with diluted HClO 4 , and then extracted, washed and dried.
In all cases the solvent was stripped in vacuum and the mixture was passed through to a chromatographic column (SiO 2 Merck 60G or alumina) using hexane/ethyl acetate in varying proportions as elluent.Yields were calculated from gas chromatography using biphenyl ether as the internal standard.

Electrolysis products identification
The products 8 and 9 were characterized by comparison of their MS spectra with those of the authentic samples.