Unsymmetrical Diaryl Sulfones through Electrochemical Oxidation of Fast Violet B in the Presence of Aryl Sulfinic Acids

Cyclic voltammetry, controlled-potential coulometry and preparative electrolysis were performed using an Autolab model PGSTAT 20 potentiostat/galvanostat. The working electrode used in the voltammetry experiments was a glassy carbon disc (1.8 mm diameter) and a platinum wire was used as the counter electrode. The working electrode used in controlled-potential coulometry and macroscale electrolysis was an assembly of four carbon rods (31 cm²), while a large platinum gauze constituted the counter electrode. The working electrode potentials were measured vs. Ag/AgCl (all electrodes from AZAR Electrodes). Fast Violet B (FVB), 4-toluenesulfinic acid (pKa = 1.99),⁴¹ benzensulfinic acid (pKa = 1.84)⁴¹ and 4-chlorosulfinic acid (pKa = 1.81),⁴¹ perchloric acid and ethanol were obtained from commercial sources.

The geometries of all species in the gas phase were fully optimized at BP86/def2-SVP level of theory using Gaussian 09.⁴² The frequencies were computed at mentioned level of theory. All minima and transition structures were confirmed to have none and one imaginary frequency, respectively. Transition structures were located using the Berny algorithm. Intrinsic reaction coordinates (IRC)^43,44 calculations were used to confirm the connectivity between transition structures and minima. Also, NBO analyses⁴⁵ were carried out at the mentioned levels of theory. A starting molecular-mechanics structure for the ab initio calculations was obtained using the HyperChem 5.02 program.

An ethanol/aqueous perchloric acid (c = 0.1 M) mixture (80 ml) (30/70, v/v) solution containing FVB (0.25 mmol) and aryl sulfinic acids sodium salt (1a–1c) (0.25 mmol) was electrolyzed in an undivided cell at 0.56 V vs. Ag/AgCl. The electrolysis was terminated when the current decreased by more than 95% (after consumption of about 50 coulombs). After neutralization by addition of sodium bicarbonate, the precipitated solid was washed with water. After washing and drying, the products were characterized by IR, ¹HNMR, ¹³C NMR, and MS.

Isolated yield: 88%. mp 142-144°C. ¹H NMR (400 MHz, DMSO-d₆) δ: 2.31 (s, 3H, methyl), 2.57 (s, 3H, methyl), 4.0 (s, 3H, methoxy), 6.5 (broad, ∼2H, NH₂), 7.14 (s, 1H, aromatic), 7.61 (d, J = 8.4 Hz, 2H, aromatic), 7.68 (t, 2H, aromatic), 7.75 (t, 1H, aromatic), 7.94 (d, J = 8.0 Hz, 2H, aromatic), 8.12 (d, J = 7.2 Hz, 2H, aromatic), 9.95 (s, 1H, NH); ¹³C NMR (100 MHz, DMSO-d₆) δ: 15.1 (C-12), 21.0 (C-1), 56.1 (C-9), 114.3 (C-10), 117.1 (C-11), 124.6 (C-7), 125.4 (C-6), 125.6 (C-13), 127.5 (C-16), 128.3 (C-4), 129.8 (C-17), 131.5 (C-3), 134.1 (C-18), 138.7 (C-15), 140.1 (C-5), 144.0 (C-2), 144.9 (C-8), 165.4 (C-14); IR (KBr) v: 3497, 3393, 3236, 2935, 1640, 1579, 1515, 1476, 1302, 1241, 1139, 1121, 843, 814, 597, 555 cm⁻¹; MS (EI, 70 eV) (m/z) (relative intensity %): 410 (M, 63), 382 (7), 210 (32), 181(11), 105 (100), 77(88), 43 (20).

Isolated yield: 85%. mp 160-163 ^oC. ¹H NMR (400 MHz, DMSO-d₆) δ: 2.32 (s, 3H, methyl), 4.02 (s, 3H, methoxy), 6.5 (broad, ∼2H, NH₂), 7.17 (s, 1H, aromatic), 7.68 (t, 2H, aromatic), 7.76 (t, 1H, aromatic), 7.82 (t, 2H, aromatic), 7.89 (t, 1H, aromatic), 8.06 (d, J = 7.6 Hz, 2H, aromatic), 8.13 (d, J = 7.2Hz, 2H, aromatic), 9.97 (s, 1H, NH); ¹³C NMR (100 MHz, DMSO-d₆) δ: 15.1 (C-11), 56.1 (C-8), 114.5 (C-9), 116.7 (C-10), 124.7 (C-6), 125.5 (C-12), 127.5 (C-5), 128.3 (C-15), 129.5 (C-3), 131.5 (C-16), 133.5 (C-2), 134.1 (C-17, C-1), 138.9 (C-14), 142.9 (C-4), 144.9 (C-7), 165.4 (C-13); IR (KBr) v: 3477, 3380, 3242, 2940, 1635, 1579, 1521, 1475, 1306, 1245, 1143, 720, 694, 607, 555 cm⁻¹; MS (EI, 70 eV) (m/z) (relative intensity %): 396 (M, 78), 331 (6), 196 (30), 149 (14), 105 (100), 77 (62), 43 (38).

Isolated yield: 96%. mp 122-124°C. ¹H NMR (400 MHz, DMSO-d₆) δ: 2.13 (s, 3H, methyl), 3.83 (s, 3H, methoxy), 6.36 (broad, 2H, NH₂), 6.99 (s, 1H, aromatic), 7.50 (t, 2H, aromatic), 7.57 (t, 1H, aromatic), 7.71 (d, J = 8.8 Hz, 2H, aromatic), 7.88(d, J = 8.4 Hz, 2H, aromatic), 7.94 (d, J = 7.2 Hz, 2H, aromatic), 9.79 (s, 1H, NH); ¹³C NMR (100 MHz, DMSO-d₆) δ: 15.1 (C-11), 56.1 (C-8), 114.6 (C-9), 116.2 (C-10), 124.8 (C-6), 125.3 (C-12), 127.48 (C-5), 127.51 (C-1), 128.3 (C-15), 129.6 (C-16), 131.5 (C-3), 134.1 (C-2), 138.3 (C-17), 139.0 (C-14), 141.6 (C-4), 145.0 (C-1), 165.4 (C-13); IR (KBr) v: 3479, 3386, 3371, 3230, 3095, 1635, 1578, 1478, 1308, 1243, 1145, 1121, 755, 696 cm⁻¹; MS (EI, 70 eV) (m/z) (relative intensity %): 430 (M, 16), 382 (79), 266 (53), 152 (16), 105 (60), 43 (100).

The electrochemical oxidation of FVB has been studied at various pH values (Fig. 1, Part I). As can be seen, the cyclic voltammograms of FVB show one anodic peak (A₁) in the positive going scan which can be ascribed to the oxidation of FVB to (Z)-N-(4-imino-5-methoxy-2-methylcyclohexa-2,5-dien-1-ylidene)benzamide N-acetyl-p-benzoquinonediimine (FVB_ox) and corresponding cathodic peak (C₁) which is attributed to the reduction of FVB_ox to FVB (Scheme 1). Under these conditions, the peak current ratio (I_pC1/I_pA1) is close to one, which can be considered as a criterion for the stability of electrogenerated FVB_ox under the experimental conditions. In other words, any side reactions, such as hydrolysis, dimerization,⁴⁶ or oxidative ring cleavage⁴⁷ on electrochemically generated FVB_ox are too slow to be observed at the time scale of cyclic voltammetry.

Zoom In Zoom Out Reset image size

Scheme 1. Part I: Acid/base behaviors of FVBH/FVB and FVB/FVB⁻. Part II: Proposed mechanism for the electrochemical oxidation of FVB in the presence of arylsulfinic acids. Part III: Redox behavior of 3/3_ox.

By increasing pH, E_pA1 and E_pC1 shifts to negative values. This confirms the participation of proton(s) in the redox behavior of FVB/FVB_ox. The variation of E_pC1 with pH is described by:

where m is the number of protons involved in the reaction and E'_{pC1(pH = 0)} is the peak potentials at pH = 0.0. R, T and F have their usual meanings. The Pourbaix diagram for peak C₁ includes three lines with different equations and slopes around pH values 6.5 and 10.6 (Fig. 1, Part II).

It can be concluded that the electrode reaction occurring at the pH lower than 6.5 is a two-electron/three-proton process involving the oxidation of the "protonated-FVB" (FVBH) to FVB_ox and vice versa. On the other hand, at 6.5 < pH > 10.6, the E_pC1 value shifts by −59 mV per pH, indicating that the electrode reaction is a two-electron/two-proton process involving the oxidation of FVB to FVB_ox and vice-versa. Whereas, in the pH > 10.6, the E_pC1 value shifts by −25 mV per pH, indicating that the electrode reaction is a two-electron/one-proton process involving the oxidation of FVB⁻ to FVB_ox and vice versa. The pK_a' for acid/base equilibria, FVBH/FVB and FVB/FVB⁻ is 6.5 and 10.6 respectively (Scheme 1, Part I).

Another noteworthy behavior was observed in the basic solutions (pHs 9 and 10). In this pH region, CV of FVB in water (bicarbonate buffer, c = 0.2 M, pH 9.0)/ethanol (70/30, v/v) mixture show two anodic (A₁ and A₂) and two cathodic (C₁ and C₂) peaks (Fig. 2, Part Ia and inset). In addition, the cyclic voltammogram of FVB in the same conditions, however in water/ethanol (50/50, v/v) mixture is shown in Fig. 2, Part Ib. As shown, with increasing ethanol percentage in the solvent mixture, the voltammogram tends to become regular in shape.

Zoom In Zoom Out Reset image size

In addition, the effect of FVB concentration on the normalized cyclic voltammograms of FVB is shown in Fig. 2, Part II. Normalized cyclic voltammograms are obtained by dividing the current by the FVB concentration (I/c). It indicates that with increasing FVB concentration, the anodic peak current ratio (I_pA1/I_pA2) decreases. These data show the adsorption phenomena in electrochemical oxidation of FVB in the basic solutions.

The cyclic voltammogram of FVB in a water (perchloric acid, 0.1 M)/ethanol (70/30 v/v) mixture is shown in Fig. 3a. The cyclic voltammogram of FVB in the presence of p-toluenesulfinic acid (1a) as a nucleophile at the same conditions, was shown in Fig. 3b. Compared with Fig. 3a, the cathodic peak current (I_pC1) significantly decreased and two new anodic peaks (A₃ and A₄) appears at more positive potentials. The decrease in I_pC1 indicates that the electrochemically generated FVB_ox is consumed by a chemical reaction with 1a.⁴⁸ In this Figure, curve c is the cyclic voltammogram of 1a in the absence of FVB. Comparison of curve b with curve c, shows that peak A₄ is related to the oxidation of 1a.

Zoom In Zoom Out Reset image size

More data were obtained by changing the potential scan rate in a solution of FVB in the presence of 1a (Fig. 4, Part I). The results indicate that the peak current ratio (I_pC1/I_pA1) is dependent on the potential scan rate, so that, it increases with increasing scan rate. The increasing I_pC1/I_pA1 at faster scan rates is related to the decrease of the experimental time scale. At this condition, a small amount of FVB_ox could be reacted with 1a. The same result was obtained by decreasing the concentration of 1a (Fig. 4, Part II). As can be seen, with increasing 1a concentration: (1) peak C₁ has been removed. (2) Peaks A₃ and A₄ have been increased. The increasing of 1a concentration causes an increase in the reaction rate of FVB_ox with 1a and thus decreasing the I_pC1. Since, peak A₃ is related to the oxidation of the product formed after the reaction of 1a with FVB_ox, it increases with increasing 1a concentration (Part II) and decreasing potential scan rate (Part I).

Zoom In Zoom Out Reset image size

Controlled-potential coulometry (CPC) was used to determine the number of electrons involved in the oxidation of FVB in the presence of 1a. CPC was performed in water (perchloric acid, 0.1 M)/ethanol (70/30 v/v) mixture containing FVB (0.25 mmol) and 1a (0.25 mmol) at 0.56 V vs. Ag/AgCl. The electrolysis progress was monitored using cyclic voltammetry (Fig. 5). It was found that, proportional to the progress of coulometry, both anodic peaks (A₁ and A₃) decreased. These peaks disappeared when the charge consumption was 2.3 e⁻ per molecule of FVB.

Zoom In Zoom Out Reset image size

Based on our experience in the electrochemical generation of quinone diimine^49–51 along with the electrochemical and spectroscopic data of the final product, we propose the following mechanism for the electrochemical oxidation of FVB in the presence of 1 (Scheme 1, Part II). In the proposed mechanism, for simplification, the protonation of species is omitted. According to this Scheme, the generation of the FVB_ox is followed by a Michael type addition reaction of 1 producing the 2 which after aromatization to afford the 3 as the final product.

The oxidation of 3 is more difficult than the oxidation of FVB due to the presence of the electron-withdrawing phenylsulfonyl group as well as by the insolubility of the final products (3) in water (perchloricacid, 0.1 M)/ethanol (70/30 v/v) mixture. According to Scheme 1, Part 3, the anodic peak A₃ pertains to the oxidation of 3a to its oxidized form 3a_ox.

The suggested mechanism for the reaction of FVB_ox with 1a has been studied using density functional theory (DFT). In order to study the Michael acceptor (FVB_ox) reactivity toward the reaction with 1a, we have investigated four parameters, natural charge and LUMO coefficient of FVB_ox and thermodynamic and kinetic stability of possible products. FVB_ox can be attacked by 1a from sites A, B,C or D to yield four types of compounds (3a, 4a, 5a and 6a) (Scheme 2). However, ¹H NMR results derived in this work (presence of a singlet aromatic peak at 7.14 ppm) are in contrast with the nucleophilic attacked of 1a to FVB_ox from N situation. Thus, further theoretical studies will be done according to the paths A and B. The geometry of the compounds 3a and 4a investigated here were optimized at BP86/def2-SVP level and are given in Fig. 6.

Zoom In Zoom Out Reset image size

Scheme 2. The structures of possible compounds from the oxidation of NAP with 1a.

The NBO analysis was used for calculation of the partial charge as well as LUMO coefficient of each atom in FVB_ox investigated here at BP86/Def2-SVP level of theory. As shown in Fig. 7, the natural charges of C1 and C4, in FVB_ox are about −0.31 and −0.26e. This result shows that C4 is slightly positive than C1. On the other hand, the calculated LUMO coefficient for latter carbon in excellent agreement with NBO analysis shows that the C4 with LUMO coefficient about 0.25 is more suitable site than C1 with LUMO coefficient about −0.30. Thus it can be concluded that the NBO analysis results as well as LUMO coefficient in conjugation with experimental data confirms that the C4 position, is more favored for latter electrochemical reaction.

Zoom In Zoom Out Reset image size

In continuation the thermodynamic stability of two possible products (3a and 4a) of latter electrochemical reaction were investigated. Fig. 6 shows the structure of two possible products according to proposed mechanism for latter electrochemical reaction. A comparison between the relative Gibbs free energies of two possible products 3a and 4a, are 0.0 and 4.6 kcal/mol, respectively. As can be seen, in good agreement with NBO analysis and calculated LUMO coefficient the relative Gibbs free energies show that the compound 3a is more stable than compound 4a.

The mechanism for the formation compounds 3a and 4a from the electrochemical reaction of FVB_ox with 1a has also been studied kinetically (Scheme 3). On the basis of proposed mechanism here, we could consider one intermediate as well as two transition states for the latter electrochemical reaction. The minimum energy path of the initial reaction starts with the nucleophilic attack of 1a to C4, C1 (Scheme 3, paths i and ii) of FVB_ox which led to the formation of Int-1 and Int-1-2 in paths I and II.

Scheme 3. The proposed pattern for the formation of compound 3a and 4a.

The calculated barrier energies for the formation of Int-1 and Int-1-2 are −7.7 and −3.5 kcal/mol and those for corresponding TS-1 and TS-1-2 are about 46.5 and 91.6, respectively (Scheme 3 and Fig. 8). The calculated barrier energies for the formation of compounds 3a and 4a are −31.8 and −27.2 kcal/mol, respectively and those for corresponding TS-2 and TS-2-2 are about 31.2 and 33.5 kcal/mol (Scheme 3 and Fig. 8). The result confirms that the formation of 3a not only thermodynamically but also kinetically more favorable than 4a, and transition states in the path I is less endothermic than path II.

Zoom In Zoom Out Reset image size

According to these data, it can be declared that final products influenced by four parameters, natural charge and LUMO coefficient of Michael acceptor and thermodynamic and kinetic stability of possible products. Also according to above data we can conclude that the 3a is the most favorable product obtained from the electrochemical reaction.

Among the electrochemical parameters for the synthesis of organic compounds, the current density is one of the most important factors influencing the yield and purity. The synthesis of 3a was performed via the constant current electrolysis under the same conditions as described for controlled-potential electrolysis when the theoretical amount of electricity (50 C) was passed. In this study, the current density varied from 0.01 to 0.75 mA/cm, while the other parameters (temperature = 298 K, FVB amount: 0.25 mmol, and 1a amount: 0.25 mmol) are kept constant. Our results show that the highest yield (88%) was obtained at a current density of 0.05 mA/cm². Fig. 9, Part I, shows the effect of current density on the yield of 3a. In this current density, the electrode potential remained nearly constant at +0.56 V (Fig. 9, part II), which is necessary for the oxidation of FVB. Higher current densities result in an increase in side reactions such as the oxidation of 3a, solvent and ...., resulting in a decrease in product yield.

Zoom In Zoom Out Reset image size