- Email: [email protected]
Catalytic cathodic cyclodimerization of vinylarylsulfones
Journal of
Electroanalytical Chemistry Journal of Electroanalytical Chemistry 569 (2004) 175–184 www.elsevier.com/locate/jelechem
Catalytic cathodic cyclodimerization of vinylarylsulfones J.F. Bergamini a, J. Delaunay a, P. Hapiot a, M. Hillebrand b, C. Lagrost a, J. Simonet a, E. Volanschi b,* a
Laboratory of Molecular and Macromolecular, Electrochemistry, UMR 6510, University of Rennes1, Campus de Beaulieu, 35042 Rennes, France b Department of Physical Chemistry, University of Bucharest, Blvd Elisabeta 4-12, Bucharest RO-70346, Romania Received 29 August 2003; received in revised form 17 November 2003; accepted 21 February 2004 Available online 7 May 2004
Abstract The cathodic cyclodimerization reaction of two vinylarylsulfones was investigated using cyclic voltammetry and UV-spectroelectrochemical techniques. The results indicate an ~ ECE sequence, where the chemical step is a dimerization anion radical–substrate. The proposed mechanism involved a catalytic cycle (ETC), based on the facts that (i) the reduction potential of the dimer is more negative than that of the monomer; (ii) cleavage rates of sulfone anion-radicals are slower than electron exchange rates. AM1modeling of the reaction routes implying the anion radical–substrate dimerization that leads to the dimer anion radical supports the proposed reaction mechanism. Ó 2004 Elsevier B.V. All rights reserved. ‹
Keywords: Vinylarylsulfones; Cyclobutane; Catalytic electron transfer
1. Introduction (2+2) cycloadditions of a,b-unsaturated sulfones are rare and usually occur under photochemical or electrochemical activation [1–3], the single case of electrochemical activation being that of benzothiophensulfone [2]. The cyclodimerization of vinylarylsulfones, initiated by electron transfer (ET) has been known since 1990 [3]. The importance of this reaction resides in the fact that it easily produces building blocks of interest in cyclobutane and sulfone chemistry [4]. There is no equivalent synthetic route in ‘‘pure’’ organic chemistry, as the reaction could not be reproduced by means of conventional reducing agents, frequently used in organic synthesis (such as dissolved metals in tetrahydrofuran). It was shown that this reaction could occur in high chemical yield when activated by cathodic ET. Very small amounts of electricity are sufficient to permit the
*
Corresponding author. Fax: +40-21-315-9249. E-mail address: [email protected] (E. Volanschi).
0022-0728/$ - see front matter Ó 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jelechem.2004.02.026
achievement of the reaction, which appears to be stereo- and regio-selective. A mechanism based on an electrocatalytic cycle, starting with the monoelectronic reduction of the substrate was suggested for the reduction of the phenylvinylsulfone (1), but the different intermediate steps and the possible interferences of other known reactions characteristic to the aryl-sulfones, (for example, the cleavage of the carbon–sulfur bond) were not discussed [5]. The present work aims to seek the best experimental conditions for this reaction and to obtain a better knowledge about the mechanism of the cyclodimerization. Therefore, six vinylarylsulfone dimers (2, 2a–2e) were synthesized under different experimental conditions (solvents, electrode materials, supporting electrolytes, etc). Two vinylarylsulfones were selected and their reduction investigated in dimethylformamide and acetonitrile, using coupled electrochemical and spectral techniques: b-naphthyl (1a) and p-tolyl (1b) (Scheme 1). Semi-empirical AM1 gas phase and solvent dependent MO-calculations were used in order to rationalize the experimental results and to support the proposed electron-transfer catalytic (ETC) mechanism.
176
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
Ar 2 CH2
SO2
Solid cathode
ArSO2
CH
DMF NR4ClO4 or
ArSO2
1
a Ar = β–naphthyl b Ar = p-tolyl
2
LiClO4
Scheme 1.
2. Experimental 2.1. Chemicals Arylvinylsulfones were prepared according to previously published procedures [4,5] Acetonitrile (ACN) and dimethylformamide (DMF) were from SDS (analytical anhydrous quality grade) and were used without further purification. Dimethylsulfoxide (DMSO) was distilled and stored over molecular sieves. Supporting electrolytes were from Fluka and used as received (puriss, electrochemical grade). 2.2. Electrolysis The cyclodimers were prepared by constant current electrolysis of the corresponding arylvinylsulfone. A simple two-electrode set-up was used in a one-compartment cell made with two platinum wires immersed in a container with the deoxygenated solution containing the vinylsulfone (2 g/10 mL of solution) and the supporting electrolyte. Electrolysis was stopped after the passage of a quantity of electricity equivalent to 0.1 electrons/mol of vinysulfone (see Table 1). Electrolysed solutions were poured into 100 ml of iced water to precipitate the cyclodimers produced. After filtration and efficient rinsing by cold water to eliminate the solvent and LiClO4 , the solid phase (white crystals) was dried and then recrystalized. Chemical structures were determined by NMR and IR spectroscopy. The cyclodimers 2 have the following spectral and physical characteristics:
E-1,2-diphenylsulfonylcyclobutane 2. Yield ¼ 87%, m.p. (ethanol) ¼ 199 °C. IR (KBr, m/cm1 ) mAr 1590, mSO2 1300, 1145. NMR 1 H (CDCl3 ) 7.78 (d, 4H, 3 J ¼ 8:5 Hz), 7.64 (t, 2H, 3 J ¼ 7:5 Hz), 7.51 (t, 4H, 3 J ¼ 7:5 Hz), 4.20 (oct, 2H), 2.52 (m, 2H), 2.35 (m, 2H). NMR 13 C (CDCl3 ) 136.9; 134.2; 129.4; 128.4; 57.0; 19.1. E-1,2-bis(2-naphthylsulfonyl) cyclobutane 2a.Yield ¼ 56%, m.p. (chloroform) ¼ 210 °C. IR (KBr, m/cm1 ) mAr 1580, mSO2 1300, 1145. NMR 1 H (CDCl3 ) 8.25 (s, 2H), 7.65 (m, 8H), 7.50 (m, 4H), 4.30 (oct, 2H), 2.52 (m, 2H), 2.35 (m, 2H). E-1,2-bis(4-methylphenylsulfonyl)cyclobutane 2b. Yield ¼ 87%, m.p. (ethanol) ¼ 170 °C. IR (KBr, m/cm1 ) mAr 1590, mSO2 1300, 1145. NMR 1 H (CDCl3 ) 7.65 (d, 4H), 7.30 (m, 4H), 4.25 (oct, 2H), 2.52 (m, 2H), 2.60 (s, 6H) 2.35 (m, 2H). E-1,2-bis(2-methylphenylsulfonyl)cyclobutane 2c. Yield ¼ 84%, m.p. (ethanol) ¼ 132 °C. IR (KBr, m/cm1 ) mAr 1580, mSO2 1310, 1145. NMR 1 H (CDCl3 ) 7.78 (d, 4H), 7.40 (m, 4H), 4.30 (oct, 2H), 2.52 (m, 2H), 2.50 (s, 6H) 2.35 (m, 2H). E-1,2-bis(4-methoxyphenylsulfonyl) cyclobutane 2d. Yield ¼ 82%, m.p. (ethanol) ¼ 198 °C. IR (KBr, m/cm1 ) mOCH3 2850, mAr 1590, mSO2 1300, 1145. NMR 1 H (CDCl3 ) 7.70 (d, 4H), 6.90 (d, 4H), 4.30 (oct, 2H), 3.85 (s, 6H), 2.52 (m, 2H), 2.35 (m, 2H). E-1,2-bis(2-methoxyphenylsulfonyl) cyclobutane 2e. Yield ¼ 68%, m.p. (ethanol) ¼ 158 °C. IR (KBr, m/cm1 ) mOCH3 2850, mAr 1590, mSO2 1310, 1145. NMR 1 H (CDCl3 ) 7.75 (dd, 2H), 7.50 (dd, 2H), 7.10 (t, 2H), 6.80 (t, 2H), 4.65 (oct, 2H), 3.80 (s, 6H), 2.52 (m, 2H), 2.35 (m, 2H). Cyclic voltammetry experiments, including fast scan voltammetry, were carried out in a one-compartment three-electrode cell, using a Pt wire as a counter electrode and an aqueous saturated calomel electrode (SCE) as the reference electrode. The supporting electrolyte (concentration 0.2 M) was LiClO4 or tetrabutylammonium hexafluophosphate (TBAPF6 ). Ferrocene was ad-
Table 1 Galvanostatic electrolyses of sulfone 1 (2 g in 10 ml of electrolyte); undivided cell; counterelectrode: platinum wire (diameter 0.5 mm, length in contact with solution 2 cm) Entry
Solvent
Supporting electrolyte
Cathode
I (mA)
Q (electron molecule1 )
Yield (%)
1 2 3 4 5 6 7 8 9 10 11
DMSO DMF THF DMSO DMF DMF DMF DMSO DMSO DMSO DMSO
0.01 M LiClO4 0.01 M LiClO4 0.01 M LiClO4 0.05 M LiClO4 0.1 M LiCl 0.1 M LiBF4 0.1 M LiPF6 0.01 M LiClO4 0.01 M LiClO4 0.01 M LiClO4 0.01 M LiClO4
Pt wire Pt wire Pt wire Pt wire Pt wire Pt wire Pt wire Cu wire Graphite bar Pt sheet (4 cm2 ) Pt sheet (4 cm2 )
5 5 5 5 5 5 5 5 5 5 10
0.1 0.1 0.1 0.08 0.1 0.1 0.1 0.1 0.1 0.1 0.2
87 84 0 95 84 90 70 80 75a 61 65b
a b
Efficient area of graphite: 2.5 cm2 . At 0.1 electron/molecule, the reaction was not completed.
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
ded to the electrolyte solution at the end of each series of experiments and the ferrocenium/ferrocene couple (E0 ¼ 0:475 V (SCE) in DMF + 0.2 M TBAPF6 and E0 ¼ 0:405 V (SCE) in ACN + 0.2 M TBAPF6 ) served as an internal probe. The working electrode was a gold disk (1 mm diameter). Electrochemical instrumentation consisted of a Tacussel GSTP4 programmer and a home-built potentiostat equipped with a positive-feedback compensation device [6]. The voltammograms were recorded with a 310 Nicolet oscilloscope. For high scan rate cyclic voltammetry, the ultramicroelectrode was a gold wire (10 lm diameter) sealed in soft glass. The signal generator was an Agilent 33250A, and the data were acquired with an Agilent Infiniium 54810A oscilloscope (500 MHz, 1 G sample/s). The electrode was carefully polished before each voltammetry experiment with 1 lm diamond paste and rinsed ultrasonically in absolute ethanol. Electrolyte solutions were thoroughly purged and kept under a dry argon flow during each run. Experiments were performed at room temperature (20 2 °C). Numerical simulations of the voltammograms were performed with the BAS DigiSim simulator 3.03, [7] using the default numerical options with the assumption of planar diffusion and a Butler-Volmer law for the electron transfer kinetics. The charge-transfer coefficient, a, was taken as 0.5 and the diffusion coefficients were taken equal for all species (D ¼ 105 cm2 s1 ). Spectroelectrochemistry. UV and visible absorption spectroelectrochemistry in ACN was performed using in situ techniques. Semi-empirical gas phase and solvent dependent MO-calculations were performed using the AM1 Hamiltonian and SM5.4A Charge Model in the AMSOL program package [8,9] and the HyperChem realease 7 software (Hypercube, Inc. Gainesville, Florida).
3. Results 3.1. Constant current electrolysis of arylvinylsulfones 1 In preliminary work [3] reporting the cathodic treatment of arylvinylsulfones 1, it was shown that such structures are able to afford cyclodimerization products (dimer 2) in fair yields. Side products such as those coming from the decomposition of dimer 2 by electrogenerated bases or the protonation of the vinyl moiety (protonation of 1: ) were also isolated [3]. The use of carefully dried solvents limits the formation of electrogenerated bases and favours the selectivity of the cyclodimerization reaction [5]. In the first part of this work, our goal was to find the best experimental conditions by varying different parameters (solvents, supporting electrolyte, charge, etc) (see Table 1). Good results were obtained with constant current electrolysis
177
using a simple two-electrode cell made with two Pt wires immersed in a one-compartment cell strictly deaerated by argon bubbling. A better selectivity in the cyclodimer was found with a low consumption of electricity (in most cases 0.1 electron/molecule), when the applied potential remained practically at the threshold of the first cathodic peak observed with the series 1, for the greatest part of the electrolysis [10]. It is worth noting the simplicity and interest of such a procedure, as it does not require the use of a powerful generator or potentiostat. Concerning the solution, the best yields were obtained when using dimethylsulfoxide (DMSO) containing lithium perchlorate as supporting electrolyte (0.05 M LiClO4 ,) with 2 g of sulfone 1 in 10 ml solution (conditions of entry 4 in Table 1). The use of DMF as the solvent gave also fair yields that were much higher than those already published when using a stirred mercury pool as a cathode [3]. The use of tetrahydrofuran as the solvent did not allow cyclodimerization to occur under these conditions. Moreover, the use of copper and graphite cathodes in DMSO were found to be successful in the formation of 2. Considering these as the best experimental conditions, the method was extended and tested in the preparation of a series of substituted E1,2-bis(R-arylsulfonyl)cyclobutanes 2a–2e where R ¼ 2-naphthyl, 4-methylphenyl, 2-methylphenyl, 4-methoxyphenyl, 2-methoxyphenyl, respectively. Except for 2e, the yield of isolated cyclobutane was higher than 80%. 3.2. Cyclic voltammetry In the second stage of this work, we tried to obtain more insight into the mechanism involved in the catalytic process. Cyclic voltammetric experiments were performed, in DMF and ACN with LiClO4 or tetrabutylammonium hexafluorophosphate (NBu4 PF6 ) as supporting electrolytes. Similar general patterns were obtained with these two salts; however due to a more resistive medium with LiClO4 , which increases the difficulty of achieving full ohmic drop compensation, detailed investigations at fast scan rates were performed only with NBu4 PF6 . With both supporting electrolytes, cyclic voltammetry at low sweep rates of 1a in DMF (Fig. 1(a)-1) shows two redox processes for which the relevant electrochemical data are presented in Table 2. The first wave has no anodic counterpart and the second one is less well-shaped. The first wave becomes reversible at sweep rates higher than 20 V s1 and was assigned (by comparison with the peak current of the one electron oxidation of ferrocene), to the monoelectronic reduction of the monomer (1a/1a ). Analysis according to the usual criteria in cyclic voltammetry shows that the ratio of the peak currents Ipa =Ipc increases with sweep rate. Under the sweep rate conditions corresponding to an irreversible voltammogram, the peak potential Epc
178
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
1.5 1.0 0.5
-0.5
wave I
11
-1.0
Ipc v-1/2 / A s1/2 V-1/2
I / µA
0.0
3
-1.5
1
2
-2.0
10 9 8
wave II
7 6 0.00
0.05
0.10
0.15
0.20
v / V s -1
-2.5 -2.2
-2.0
-1.8
(a)
-1.6
-1.4
-1.2
E/V 2 1
I / µA
0 -1
3 -2 -3
-4 -2.4 (b)
1 -2.2
-2.0
-1.8
-1.6
-1.4
-1.2
E/V
Fig. 1. (a) Cyclic voltammetry of 1a in DMF + 0.2 M TBAPF6 (c ¼ 1:6 103 M): 1, v ¼ 50 mV s1 ; 2–after preelectrolysis for 1 min.; 3, v ¼ 200 V s1 (for curve 3 the current scale is I 102 lA). Inset: Dependence of the peak current function Ipc v1=2 on the scan rate for 1a. (b) Cyclic voltammetry of 1a in ACN (first 3 scans), c ¼ 1:11 103 M, v ¼ 200 mV s1 .
varies linearly with the logarithm of the scan rate with a slope of )33 mV/decade. This value is close to the theoretical value expected for a ‘‘DIM2’’ mechanism, i.e., a reaction scheme involving a rapid electron transfer followed by anion-radical-substrate dimerization (29.5 mV) [11]. The standard rate constant for the heteroge-
neous electron transfer, ks , (uncorrected for the double layer effect) was estimated from the forward and reverse peak potential difference using both Nicholson’s W function [12] and the DigiSim program. A value of 5 cm s1 was obtained. The second redox couple was observed only at low sweep rates and an estimate of the standard ET rate gives ks 0:15 cm s1 , with a standard potential E20 value of )1.952 V vs SCE. If the monomer 1a is submitted to preelectrolysis for about 1 min, only the second redox couple is observed (Fig. 1(a-2)), whereas at high scan rates (200 V s1 ) only the first couple is present (Fig. 1(a-3)). Starting with a solution of dimer 2a, only one reversible redox couple was observed in the range 0 to )2.0 V, characterised by a E0 ¼ 1:950 V (SCE) with ks ¼ 0:14 cm s1 and was assigned to the monoelectronic reduction of the dimer. The dependance of the peak current function, Ip v1=2 , on the scan rate, also included in Fig. 1(a) as an inset, shows the decrease for wave I of the monomer at low scan rates, simultaneously with an increase for wave II, assigned to the reduction of the dimer, due to the progressive transformation of the monomer into the dimer. These effects are visible only for the lowest scan rates where natural convection starts to interfere with the voltammetric response. As the influence of convection is to increase the current with respect to the value expected under theoretical pure diffusion conditions, it results that the variation of the electron stoichiometry can be considered as an upper limit. The marked decrease of the reduction wave of 1a, observed on the second sweep in acetonitrile ‹ (Fig. 1(b)), is also indicative of an ~ ECE mechanism [13]. Considering all these data, the second redox couple observed at the reduction of the monomer 1a was assigned to the monoelectronic reduction of the dimer 2a, formed as a consequence of the anion radical–substrate dimerization. A similar anion radical–substrate dimerization was reported as most probable for the electroreductive coupling of vinylpyridines and vinylquin- olines leading to dipyridylcyclobutane derivatives [14], whereas alternative routes, usually radical anion– radical anion dimerization, followed by protonation, would lead mainly to the linear hydrodimer.‹ Therefore, the following simplified ~ ECE sequence may be proposed:
Table 2 Relevant electrochemical data for compounds 1a and 2a I wave
II wave
I wave
1a
2a
v (V s )
)Epc (V)
)Epc (V)
Ipa Ipc
0.05 0.20 200 1000
1.846 1.860 1.884 1.880
– – 1.824 1.813
– – 0.867 1.170
a
I0
calculated with formula [12]: Ipa =Ipc ¼ Ipa þ pc
0 0:485Isp Ipc
a
þ 0:086.
Epc (V)
)Epa (V)
Ipa Ipc
Epc (V)
)Epa (V)
Ipa Ipc
1.997 1.994 – –
1.895 1.937 – –
0.548 0.559 – –
1.899 1.891 1.967 1.976
– – 1.840 1.829
– – 0.622 1.030
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
1a þ e 1a ;
E10 ¼ 1:856 V
ð1Þ
k2
1a þ 1a ! 2a
2a þ e 2a ;
Table 3 Relevant electrochemical data for compounds 1b and 2b
¼ 1:950 V
ð3Þ
The anion radical–substrate dimerization (reaction 2) leads to the anion radical of the dimer. The E0 for the couple 2a/2a is slightly more negative than the standard potential corresponding to the monomer couple 1a/ 1a . Thus, direct oxidation of 2a to 2a at the electrode (at the monomer reduction potential) or by homogeneous electron transfer between the anion radical 2a and 1a are thermodynamically favoured and account for the catalytic character of the process, as indicated from the decrease of the electron stoichiometry at the lowest scan rates (see inset in Fig. 1(a)). However, this simplified mechanism does not allow a detailed explanation of all the cyclic voltammetry data (for example, it does not explain the irreversibility of the first voltammetric peak). For a full description of the experimental data, it is necessary to take into account in the mechanism the different possible end reactions, especially those involving the disappearance of the dimer anion radical, for which we have only scarce information ( see discussion and MO calculation parts below). The cyclic voltammetry results for 1b and 2b are displayed in Fig. 2 and Table 3. At the reduction of 1b, an asymmetric wave consisting of two superimposed waves is observed at low sweep rates. With increasing scan rate, the first peak becomes more intense, but no reversibility was reached even at 1000 V s1 . Similarly to 1a, the peak potential Epc was found to vary linearly with scan rate and a slope of 28.3 mV/decade was measured. This peak was assigned to the monoelectronic
2b
1b
ð2Þ E20
179
I wave
II wave
v (V s )
Epc (V)
Epc (V)
Epc (V)
0.100 0.200 2 200
2.045 2.047 2.082 2.144
2.083 2.101 2.169 2.270
– 2.139 2.170 2.278
reduction of the monomer (1b/1b ), followed by a rapid coupling reaction between the anion-radical 1b and the substrate 1b. Starting with a solution of dimer 2b only one reduction wave was observed, with a peak potential corresponding to the second redox process visible during the reduction of the monomer. Therefore, we propose a similar ECE sequence (DIM2) to that which describes the reduction of 1a to account for the experimental results. 3.3. UV–VIS Spectroelectrochemical experiments UV–VIS spectroelectrochemical reduction of 1a was investigated in ACN + 0.2 M TBAPF6 (Fig. 3) (investigations in DMF are not possible due to the low absorption wavelength of the arylvinylsulfone). At the potential corresponding to the threshold of the first reduction wave, the data show the progressive transformation of the monomer (characterised by two bands with maximum absorption wavelengths kmax ¼ 235 and 214 nm and absorption coefficients e235 > e214 ) into the absorption spectrum of the dimer 2a, (characterized by two bands located at kmax ¼ 233 and 215 nm with a 1.0
40
9
20
0.8
0
Absorbance
I/µA
-20 -40
2
-60 -80
0.6
0.4
1
1
0.2
-100 -120 -2.6
0.0 200
-2.4
-2.2
-2.0
-1.8
-1.6
-1.4
-1.2
-1.0
225
250
275
300
wavelength / nm
E/V Fig. 2. Fast scan cyclic voltammetry in DMF + 0.2 M TBAPF6 (v ¼ 200 V s1 ): 1, monomer 1b and 2, dimer 2b.
Fig. 3. UV-spectra recorded for the electrolytical reduction of 1a at the potential of the first wave; 1, initial spectrum of the monomer, 2–9, evolution in time, about 2 min between spectra.
180
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
reverse intensity ratio, e233 < e215 ). The new band at 245 nm, which decreases in favour of the dimer band, was tentatively assigned to an intermediate in the reaction sequence, presumably the anion-radical 1a . In the same experiments performed with the compound 1b, the spectroelectrochemical results show the progressive transformation of the monomer (kmax ¼ 234 nm) into the dimer 2b (kmax ¼ 227 nm).
4. Discussion The experimental results may be rationalised in terms of the following ‘‘catalytic DIM2’’ mechanism [5], based ‹ on the ~ ECE reaction sequence previously discussed: The first step in this scheme is the monoelectronic reduction of the substrate to the anion radical, followed by a rapid coupling with another neutral sulfone molecule leading to the dimer anion-radical, 2a,2b . The key step in Scheme 2 is the reoxidation of 2a,2b , which is thermodynamically favoured because of a more positive standard potential for the monomer than for the dimer. In the case of 1a, the anion radical lifetime is of the order of several 104 s, i.e., 1a has the time to diffuse far from the electrode before it reacts. Therefore it follows that the major pathway for the oxidation of 2a is the homogeneous electron transfer from the dimer anion radical to the substrate, leading to the neutral dimer and recycling the anion radical of the monomer: 2a + 1a ! 2a + 1a . To account for a fair yield in cyclodimer and low consumption of current, this reaction should be able to compete with the other decay reactions involving 2a , as a carbon–sulphur bond cleavage or the reverse dimerization (cleavage of the dimer) reaction. An estimation of the homogeneous electron transfer kinetic constant k2 can be performed in terms of the Marcus–Hush model that establishes a relation between the reaction Gibbs energy DG0 and the activation Gibbs
activation
energy DG# [15–19]. A DG0 value of )2.21 kcal mol1 (1 kcal ¼ 4.18 kJ) is derived from the experimental reduction potentials. From the relation DG# ¼ k=4 ð1 þ DG0 =kÞ where k is the intrinsic barrier (we do not consider any working terms or electrostatic contribution because the charge of the substrate is zero), and taking for k a value in the range 20–30 kcal mol1 , as previously obtained for similar organic compounds in ACN [15], an estimate of 4–6 kcal mol1 is obtained for DG# . Considering a collision factor Zsol 6 1011 M1 s1 , this gives an estimate of the rate constant k2 in the range 107 –108 M1 s1 . The Marcus–Hush model also provides a relation between the rate constant for the standard homogeneous electron transfer rate constant ðkex Þ and the heterogeneous ðks Þ electron transfer rate constant for the same reactant [12,15–19]. As shown from data on the electrochemical reduction of a series of aromatic molecules, [20] a good estimation of kex is obtained from ks values that are uncorrected for the double-layer effect and using Hush’s model where the image effect is not taken into account [17,18], due to a cancellation of these two effects: ks =Zel ¼ kex =Zsol . Using ks ¼ 5 cm s1 , as determined from cyclic voltammetry for the first reduction wave of 1a, Zel ¼ 5 104 s1 and Zsol ¼ 6 1011 M1 s1 [12], a kex value of 6 107 M1 s1 is derived leading to a k2 value of the order of 108 M1 s1 . For 1b, it is more difficult to obtain a clear cut result, as we were not able to obtain a measurement of its radical anion lifetime, and we may have to take into account, in addition to the homogeneous oxidation pathway, the possible direct oxidation of 2b at the electrode, which also leads to the regeneration of 1b . In conclusion, we can consider that the homogeneous electron transfer between 2a,2b and 1a,1b is fairly rapid in comparison with other possible reactions of the dimer anion radical, like the reverse dedimerization reaction. Other reactions involving the neutral dimer, for example the decomposition of dimer 2 by electrogenerated bases or the cleavage of the C–SO2 bond with the formation of cyclobutene derivatives will also contribute to decrease the yield of cyclodimer. 4.1. MO-modeling
ArSO2 ArSO2
eSO2Ar
SO2Ar
propagation
. ..-
SO2Ar trans
... SO2Ar
-.
ArSO2 SO2Ar
Scheme 2.
SO2Ar SO2Ar
The MO calculations performed were intended to answer two questions: firstly, to account for the electronic structural features implied in redox processes, explaining the difference in the reducibility of the monomers and dimers in both gas phase and solvent, and secondly to furnish theoretical support for the proposed ETC mechanism, by analysing different reaction pathways implied in the mechanism in Scheme 2. The redox properties of the compounds investigated were compared with theoretical predictions in the framework of the semi-empirical AM1 method, in both the gas phase and polar solvents (ACN, DMF). The
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
in the distance between the two molecules. Two attempts were made. The potential energy surface (PES) was built assuming firstly a synchronous approach of both vinyl carbon atoms d110 ¼ d220 ¼ d and, secondly, considering an asynchronous way, i.e., a previous formation of the bond between the b carbon atoms. In the first case, the plot presented in starting with a distance of 3.5 A, Fig. 4 reflects an increasing energy up to a distance of and the presence of a minimum point energy about 2 A As corresponding to the dimer anion radical at 1.54 A. these are gas phase calculations, we cannot consider that the maximum point of energy on the PES can give an estimation of the real energy barrier of the process. The modeling using the second way of approaching the molecules led to unreliable results. As the single restriction between the two molecules was the distance between the b-carbon atoms, d110 , the two fragments were able to adopt different relative positions and the formation of the cyclobutane ring could not be shown. It is likely that even if the dimerization process is an asynchronous process, the use of a single reaction coordinate is not sufficient to describe it.
120 -1
100 Erel / kcal mol
estimation of the solvation effects was performed with the AMSOL program. The redox properties of the compounds investigated as determined by the features of the frontier orbitals outline the fact that in the monomers the lowest empty molecular orbital (l.e.m.o.) is localized mainly on the vinyl moiety and the sulfone group, whereas in the dimer, it is localized on the sulfone and the aryl substituent. Therefore, if the reactivity in the cyclodimerization reaction is mainly determined by the vinyl group, and probably more influenced by the substitution there, the reactivity of the dimer and its reduction potential are mainly determined by the sulfone group and the aryl moiety. The different degrees of localization of the l.e.m.o. orbitals on the different moieties of the molecule are responsible for the order of the reduction potentials of the monomers and dimers and therefore are important for the catalytic process observed experimentally. Standard potentials E0 can be compared with adiabatic electron affinities, calculated as the difference between the formation enthalpies of the anion radical and neutral molecule. A further indication of the reactivity in reduction processes is given by the vertical and adiabatic electronegativities. The calculated electron affinities and electronegativities are presented in Table 4, for both the gas phase and acetonitrile, together with the experimental potentials derived from cyclic voltammetry experiments. For the couple 1a/2a, the calculations predict that E0 of the monomer is very close to that of the dimer (slightly more positive) in agreement with cyclic voltammetry data. However, for the p-tolyl derivatives 1b/2b, although nearly equal electron affinities and vertical electronegativities are found from gas phase calculations, the reverse order of reducibility is predicted in acetonitrile. Probably, more elaborate calculations are necessary to account for this discrepancy. Comparing the results in ACN with the gas phase calculations, higher electron affinities and vertical or adiabatic electronegativities are observed in ACN, attesting to an enhanced reducibility in polar solvents. A first insight on the initial step of the reaction mechanism presented in Scheme 2, the interaction of the substrate with the generated anion radical, was obtained by monitoring the total energy with respect to the change
181
SO2Ar
] -.
[
80
] -.
SO2Ar
[1
2
1'
2'
SO2Ar
SO2Ar
60 40 20 0 1.5
2.0
2.5
3.0
d / Å (d = d11' = d22') Fig. 4. Potential energy surfaces for the anion radical–substrate dimerization of 1b in the gas phase, considering a synchronous approach of the two species. Erel represents the energy with respect to the minimum of the dimer anion radical 2b .
Table 4 AM1-calculated redox properties of the compounds investigated in the gas phase and ACN Gas phase
ACN
Compound
Aad (kcal mol )
vv (eV)
vad (eV)
Aad (kcal mol1 )
vv a (eV)
vad b (eV)
0 Eexp (V)
1a 2a 1b 2b
)42.89 )41.89 )29.51 )30.02
5.04 5.07 5.37 5.40
5.09 4.94 5.21 5.69
)83.96 )83.52 )77.42 )83.92
5.26 5.21 5.54 5.72
5.36 5.28 5.49 5.74
)1.856 )1.950 )2.047 )2.139
a
1
a
b
vv ¼ 1=2ðehomo þ elemo Þ. vad ¼ 1=2ðIad þ Aad Þ; Iad , the adiabatic ionisation potential, is given by the enthalpy of the process, M ! Mþ + e ; Aad , the adiabatic electron affinity, is given by the negative of the enthalpy of the process, M + e ! M . b
182
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
The fully optimized structure of the trans-dimer anion radical in ACN reflects the presence of several conformers with heats of formation of about )181 kcal mol1 , differing by the positions of the sulfone and the aryl groups. Comparing the energy of the minimum with the sum of the energies of the parent compound and the corresponding anion-radical ()163 kcal mol1 ), it can be seen that the dimerization process represents a stabilisation of the system. The most striking feature of the 2b geometry and charge density distribution consists in an unexpected asymmetry, reflected by the fact that the C–SO2 distances are not equal and the charge is mainly localized on one of the dimer moieties. An important problem in Scheme 2 is the further reactivity of the dimer anion radical, consisting in several possible reactions which could compete with the homogeneous electron transfer, like the cleavage of the C–SO2 bond with elimination of the arylsulfonyl group, or the reverse dedimerization process. In the following, these reactions were modeled in DMF using the appropriate reaction coordinates, in order to obtain an estimation of their relative probability. In the case of the cleavage reaction of the carbon– sulfur bond, starting from the fully optimised dimer anion-radical, 2b , one C–SO2 distance was increased progressively, but the energy increased steadily and at reached a maximum and then decreased. An about 2.6 A estimate of the energy barrier gives about 26 kcal mol1 . Therefore, it seems rather unlikely that this reaction could occur from the dimer anion radical. The reverse reaction, the dimer anion radical dissociation process, was modeled considering as a reaction coordinate the distance between the arylsulfone substituted carbons, d220 . The plot in Fig. 5 shows two bran-
(∆H+∆Gsol)rel / kcal mol
-1
25 20 II 15
I
10 5
ches. Starting from the minimum value, the energy increased steadily up to a value of about 2.0–2.1 A (branch I) differing for the chosen conformer. At that point a sudden decrease was noted and the plot continued smoothly towards a broad minimum belonging to the open form of the dimer anion radical (branch II). The further increase of the distance d has no further significance as the free arylsulfone fragments can rotate freely and adopt several conformations. Up to the in spite of the opening of the mentioned value of 2 A, cyclobutane ring, the geometric parameters of the arylsulfone fragments as well as the charge distribution are not significantly modified from those of the closed dimer major changes were noted structure. At about 2 A, which were maintained in some conformations of the open dimer anion radical and reflect an overall symmetry, i.e., equal C–SO2 distances and charges of the two fragments. Starting from the open form minimum and conducting the optimization in the reverse direction, decreasing the d value, a crossing of both branches of the plot was obtained. Considering the crossing region as a transition state, an activation energy of about 20 kcal mol1 can be estimated for the cyclobutane ring opening process. This high value of the activation energy, even in polar solvents like ACN or DMF, seems to indicate that the cyclobutane ring opening reaction is not sufficiently rapid to compete with the homogeneous electron transfer, responsible for the catalytic character of the process. Another possible reaction, which can contribute to lower yields in cyclodimer, is the cleavage of the C–SO2 bond under the influence of electrogenerated bases, B , starting from the neutral dimer and leading to arylsulfonyl cyclobutene derivatives. Therefore we found it interesting to calculate the potential energy surface for this reaction. This was performed in two steps, as shown in Scheme 3: firstly, one proton was removed from a substituted carbon (C3) in the cyclobutane ring and the resulting anion was fully optimised. Secondly, starting with the anion 2b , the C–SO2 bond ðd2–5 Þ was increased. The PES were calculated both in the gas phase and in DMF and are presented in Fig. 6. It may be stated that the energy barrier in DMF (9 kcal mol1 , Fig. 6, curve 2) is about the half that in the
0 5 SO2-Ar
-5 1.5
2.0
2.5 d22' / Å
3.0
2b Fig. 5. Potential energy surface calculated in DMF for the dimer anion–radical 2b dissociation, with respect to the distance between the arylsulfone substituted carbons, d220 . The ðDH þ DGsol Þrel are relative values with respect to the minimum of the dimer anion radical 2b . Branches I and II (see text).
+ B-
1
2 H
4
3
2b-
(-)
-
Ar-SO-2
SO2-Ar 6 Scheme 3.
SO2-Ar
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
24
(∆H+∆Gsol)rel / kcal mol
-1
20 16 1
12 8
2
4 0 1.5
2.0
2.5
3.0
3.5
4.0
d25 / Å Fig. 6. Potential energy surface for the cleavage of the C–SO2 bond ðd2–5 Þ, under the influence of electrogenerated bases, starting from the neutral dimer: 1, gas phase; 2, DMF. The ðDH þ DGsol Þrel are relative values with respect to the minimum of the dimer anion radical 2b .
1.8
double bond 1.6
gas phase (20 kcal mol1 , Fig. 6, curve 1), and about the third of the estimated energy (26 kcal mol1 ) for the same reaction starting from the dimer anion radical. The progress of the reaction is well reflected also by the evolution of the cyclobutane ring bond order p2–5 , as well as by the total charge on the leaving arylsulfonyl group, presented in Fig. 7(a) and (b). It may be observed that the cyclobutane bond evolves from a single to a double bond at the end of the reaction (Fig. 7(a)) and the charge on the arylsulfonyl moiety changes from 0 at the beginning of the reaction, to )1 at the end (Fig. 7(b)). In conclusion, AM1 modelling accounts satisfactorily for the electronic features of the frontier orbitals implied in redox processes for both monomers and dimers. However, these methods appear as not sufficiently reliable, especially when solvent effects have to be considered, to account for the small differences in the experimental reduction potentials. AM1 modeling in polar solvents of the reaction routes implying the anion radical–substrate dimerization indicate that the other possible reactions of the dimer anion radical are slow as compared with the electron transfer to the substrate and this therefore represents support for the proposed reaction mechanism. The cleavage of the C–SO2 bond under the influence of electrogenerated bases leading to cyclobutene seems to proceed rather from the neutral dimer than from the dimer anion radical.
p
1.4
183
1.2
5. Conclusions
1.0
single bond 0.8 1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
d2-5 / Å
(a)
0.0
In conclusion, the electrochemical and UV-spectroelectrochemical data, as well as the MO calculations performed, support the ETC–catalytic mechanism ‹ (Scheme 3), based on an ~ ECE reaction sequence, followed by electron transfer from the dimer anion radical to the substrate. Homogeneous ET from the dimer anion radical to the substrate, analysed according to the Marcus–Hush model seems to be more rapid than other follow-up reactions of the dimer anion radical.
q Ar-SO2
-0.2
References
-0.4
-0.6
-0.8
-1.0 1.6
(b)
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
d2-5 / Å
Fig. 7. Evolution of the p2–3 bond order (a) and of the total charge on the arylsulfone fragment (b) during the cleavage of the C–SO2 bond.
[1] M.S. El Faghi El Amoudi, P. Geneste, J.L. Olive, J. Org. Chem. 46 (1981) 4258. [2] M.S. El Faghi El Amoudi, P. Geneste, J.L. Olive, Nuov. J. Chim. 5 (1981) 251. [3] J. Delaunay, G. Mabon, A. Orliac, J. Simonet, Tetrahedron Lett. 31 (1990) 667. [4] A. Orliac-Le Moing, J. Delaunay, J. Simonet, New J. Chem. 18 (1994) 901. [5] J. Delaunay, A. Orliac-Le Moing, J. Simonet, New J. Chem. 17 (1993) 393. [6] D. Garreau, J.-M. Saveant, J. Electroanal. Chem. 35 (1972) 309. [7] M. Rudolph, D..P. Reddy, S.W. Felberg, Anal. Chem. 66 (1994) 589A.
184
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
[8] G.D. Hawkins, G..D. Giesen, G.C. Lynch, C.C. Chambers, I. Rossi, J.W. Storer, J. Li, T. Zhu, D. Rinaldi, D.A. Liotard, C.J. Cramer, D.G. Truhlar, AMSOL version 6.5.3, University of Minnesota, Minneapolis, 1997. [9] C.C. Chambers, C.J. Cramer, D.G. Truhlar, J. Phys. Chem. 100 (1996) 16385. [10] J. Delaunay, J. Simonet, unpublished results. [11] C.P. Andrieux, L. Nadjo, J.M. Saveant, J. Electroanal. Chem. 42 (1973) 223. [12] A.J. Bard, L.R. Faulkner, Electrochemical Methods, Fundamentals and Applications, second ed., Wiley, New York, 2001, p. 122, 240.
[13] P.H. Rieger, Electrochemistry, Chapman & Hall, New York, 1994, p. 304. [14] R.G. Janssen, M. Motevalli, J.P. Utley, Chem. Commun. (1998) 539. [15] L. Eberson, Electron Transfer in Organic Chemistry, Springer Verlag, Berlin, 1987, p. 29. [16] R.A. Marcus, J. Chem. Phys. 24 (1956) 966, 979. [17] N.S. Hush, J. Chem. Phys. 28 (1958) 962. [18] N.S.J. Hush, Trans. Faraday Soc. 57 (1961) 557. [19] R.A. Marcus, J. Chem. Phys. 43 (1965) 679. [20] H. Kojima, A.J. Bard, J. Am. Chem. Soc. 97 (1975) 6317.
Electroanalytical Chemistry Journal of Electroanalytical Chemistry 569 (2004) 175–184 www.elsevier.com/locate/jelechem
Catalytic cathodic cyclodimerization of vinylarylsulfones J.F. Bergamini a, J. Delaunay a, P. Hapiot a, M. Hillebrand b, C. Lagrost a, J. Simonet a, E. Volanschi b,* a
Laboratory of Molecular and Macromolecular, Electrochemistry, UMR 6510, University of Rennes1, Campus de Beaulieu, 35042 Rennes, France b Department of Physical Chemistry, University of Bucharest, Blvd Elisabeta 4-12, Bucharest RO-70346, Romania Received 29 August 2003; received in revised form 17 November 2003; accepted 21 February 2004 Available online 7 May 2004
Abstract The cathodic cyclodimerization reaction of two vinylarylsulfones was investigated using cyclic voltammetry and UV-spectroelectrochemical techniques. The results indicate an ~ ECE sequence, where the chemical step is a dimerization anion radical–substrate. The proposed mechanism involved a catalytic cycle (ETC), based on the facts that (i) the reduction potential of the dimer is more negative than that of the monomer; (ii) cleavage rates of sulfone anion-radicals are slower than electron exchange rates. AM1modeling of the reaction routes implying the anion radical–substrate dimerization that leads to the dimer anion radical supports the proposed reaction mechanism. Ó 2004 Elsevier B.V. All rights reserved. ‹
Keywords: Vinylarylsulfones; Cyclobutane; Catalytic electron transfer
1. Introduction (2+2) cycloadditions of a,b-unsaturated sulfones are rare and usually occur under photochemical or electrochemical activation [1–3], the single case of electrochemical activation being that of benzothiophensulfone [2]. The cyclodimerization of vinylarylsulfones, initiated by electron transfer (ET) has been known since 1990 [3]. The importance of this reaction resides in the fact that it easily produces building blocks of interest in cyclobutane and sulfone chemistry [4]. There is no equivalent synthetic route in ‘‘pure’’ organic chemistry, as the reaction could not be reproduced by means of conventional reducing agents, frequently used in organic synthesis (such as dissolved metals in tetrahydrofuran). It was shown that this reaction could occur in high chemical yield when activated by cathodic ET. Very small amounts of electricity are sufficient to permit the
*
Corresponding author. Fax: +40-21-315-9249. E-mail address: [email protected] (E. Volanschi).
0022-0728/$ - see front matter Ó 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jelechem.2004.02.026
achievement of the reaction, which appears to be stereo- and regio-selective. A mechanism based on an electrocatalytic cycle, starting with the monoelectronic reduction of the substrate was suggested for the reduction of the phenylvinylsulfone (1), but the different intermediate steps and the possible interferences of other known reactions characteristic to the aryl-sulfones, (for example, the cleavage of the carbon–sulfur bond) were not discussed [5]. The present work aims to seek the best experimental conditions for this reaction and to obtain a better knowledge about the mechanism of the cyclodimerization. Therefore, six vinylarylsulfone dimers (2, 2a–2e) were synthesized under different experimental conditions (solvents, electrode materials, supporting electrolytes, etc). Two vinylarylsulfones were selected and their reduction investigated in dimethylformamide and acetonitrile, using coupled electrochemical and spectral techniques: b-naphthyl (1a) and p-tolyl (1b) (Scheme 1). Semi-empirical AM1 gas phase and solvent dependent MO-calculations were used in order to rationalize the experimental results and to support the proposed electron-transfer catalytic (ETC) mechanism.
176
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
Ar 2 CH2
SO2
Solid cathode
ArSO2
CH
DMF NR4ClO4 or
ArSO2
1
a Ar = β–naphthyl b Ar = p-tolyl
2
LiClO4
Scheme 1.
2. Experimental 2.1. Chemicals Arylvinylsulfones were prepared according to previously published procedures [4,5] Acetonitrile (ACN) and dimethylformamide (DMF) were from SDS (analytical anhydrous quality grade) and were used without further purification. Dimethylsulfoxide (DMSO) was distilled and stored over molecular sieves. Supporting electrolytes were from Fluka and used as received (puriss, electrochemical grade). 2.2. Electrolysis The cyclodimers were prepared by constant current electrolysis of the corresponding arylvinylsulfone. A simple two-electrode set-up was used in a one-compartment cell made with two platinum wires immersed in a container with the deoxygenated solution containing the vinylsulfone (2 g/10 mL of solution) and the supporting electrolyte. Electrolysis was stopped after the passage of a quantity of electricity equivalent to 0.1 electrons/mol of vinysulfone (see Table 1). Electrolysed solutions were poured into 100 ml of iced water to precipitate the cyclodimers produced. After filtration and efficient rinsing by cold water to eliminate the solvent and LiClO4 , the solid phase (white crystals) was dried and then recrystalized. Chemical structures were determined by NMR and IR spectroscopy. The cyclodimers 2 have the following spectral and physical characteristics:
E-1,2-diphenylsulfonylcyclobutane 2. Yield ¼ 87%, m.p. (ethanol) ¼ 199 °C. IR (KBr, m/cm1 ) mAr 1590, mSO2 1300, 1145. NMR 1 H (CDCl3 ) 7.78 (d, 4H, 3 J ¼ 8:5 Hz), 7.64 (t, 2H, 3 J ¼ 7:5 Hz), 7.51 (t, 4H, 3 J ¼ 7:5 Hz), 4.20 (oct, 2H), 2.52 (m, 2H), 2.35 (m, 2H). NMR 13 C (CDCl3 ) 136.9; 134.2; 129.4; 128.4; 57.0; 19.1. E-1,2-bis(2-naphthylsulfonyl) cyclobutane 2a.Yield ¼ 56%, m.p. (chloroform) ¼ 210 °C. IR (KBr, m/cm1 ) mAr 1580, mSO2 1300, 1145. NMR 1 H (CDCl3 ) 8.25 (s, 2H), 7.65 (m, 8H), 7.50 (m, 4H), 4.30 (oct, 2H), 2.52 (m, 2H), 2.35 (m, 2H). E-1,2-bis(4-methylphenylsulfonyl)cyclobutane 2b. Yield ¼ 87%, m.p. (ethanol) ¼ 170 °C. IR (KBr, m/cm1 ) mAr 1590, mSO2 1300, 1145. NMR 1 H (CDCl3 ) 7.65 (d, 4H), 7.30 (m, 4H), 4.25 (oct, 2H), 2.52 (m, 2H), 2.60 (s, 6H) 2.35 (m, 2H). E-1,2-bis(2-methylphenylsulfonyl)cyclobutane 2c. Yield ¼ 84%, m.p. (ethanol) ¼ 132 °C. IR (KBr, m/cm1 ) mAr 1580, mSO2 1310, 1145. NMR 1 H (CDCl3 ) 7.78 (d, 4H), 7.40 (m, 4H), 4.30 (oct, 2H), 2.52 (m, 2H), 2.50 (s, 6H) 2.35 (m, 2H). E-1,2-bis(4-methoxyphenylsulfonyl) cyclobutane 2d. Yield ¼ 82%, m.p. (ethanol) ¼ 198 °C. IR (KBr, m/cm1 ) mOCH3 2850, mAr 1590, mSO2 1300, 1145. NMR 1 H (CDCl3 ) 7.70 (d, 4H), 6.90 (d, 4H), 4.30 (oct, 2H), 3.85 (s, 6H), 2.52 (m, 2H), 2.35 (m, 2H). E-1,2-bis(2-methoxyphenylsulfonyl) cyclobutane 2e. Yield ¼ 68%, m.p. (ethanol) ¼ 158 °C. IR (KBr, m/cm1 ) mOCH3 2850, mAr 1590, mSO2 1310, 1145. NMR 1 H (CDCl3 ) 7.75 (dd, 2H), 7.50 (dd, 2H), 7.10 (t, 2H), 6.80 (t, 2H), 4.65 (oct, 2H), 3.80 (s, 6H), 2.52 (m, 2H), 2.35 (m, 2H). Cyclic voltammetry experiments, including fast scan voltammetry, were carried out in a one-compartment three-electrode cell, using a Pt wire as a counter electrode and an aqueous saturated calomel electrode (SCE) as the reference electrode. The supporting electrolyte (concentration 0.2 M) was LiClO4 or tetrabutylammonium hexafluophosphate (TBAPF6 ). Ferrocene was ad-
Table 1 Galvanostatic electrolyses of sulfone 1 (2 g in 10 ml of electrolyte); undivided cell; counterelectrode: platinum wire (diameter 0.5 mm, length in contact with solution 2 cm) Entry
Solvent
Supporting electrolyte
Cathode
I (mA)
Q (electron molecule1 )
Yield (%)
1 2 3 4 5 6 7 8 9 10 11
DMSO DMF THF DMSO DMF DMF DMF DMSO DMSO DMSO DMSO
0.01 M LiClO4 0.01 M LiClO4 0.01 M LiClO4 0.05 M LiClO4 0.1 M LiCl 0.1 M LiBF4 0.1 M LiPF6 0.01 M LiClO4 0.01 M LiClO4 0.01 M LiClO4 0.01 M LiClO4
Pt wire Pt wire Pt wire Pt wire Pt wire Pt wire Pt wire Cu wire Graphite bar Pt sheet (4 cm2 ) Pt sheet (4 cm2 )
5 5 5 5 5 5 5 5 5 5 10
0.1 0.1 0.1 0.08 0.1 0.1 0.1 0.1 0.1 0.1 0.2
87 84 0 95 84 90 70 80 75a 61 65b
a b
Efficient area of graphite: 2.5 cm2 . At 0.1 electron/molecule, the reaction was not completed.
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
ded to the electrolyte solution at the end of each series of experiments and the ferrocenium/ferrocene couple (E0 ¼ 0:475 V (SCE) in DMF + 0.2 M TBAPF6 and E0 ¼ 0:405 V (SCE) in ACN + 0.2 M TBAPF6 ) served as an internal probe. The working electrode was a gold disk (1 mm diameter). Electrochemical instrumentation consisted of a Tacussel GSTP4 programmer and a home-built potentiostat equipped with a positive-feedback compensation device [6]. The voltammograms were recorded with a 310 Nicolet oscilloscope. For high scan rate cyclic voltammetry, the ultramicroelectrode was a gold wire (10 lm diameter) sealed in soft glass. The signal generator was an Agilent 33250A, and the data were acquired with an Agilent Infiniium 54810A oscilloscope (500 MHz, 1 G sample/s). The electrode was carefully polished before each voltammetry experiment with 1 lm diamond paste and rinsed ultrasonically in absolute ethanol. Electrolyte solutions were thoroughly purged and kept under a dry argon flow during each run. Experiments were performed at room temperature (20 2 °C). Numerical simulations of the voltammograms were performed with the BAS DigiSim simulator 3.03, [7] using the default numerical options with the assumption of planar diffusion and a Butler-Volmer law for the electron transfer kinetics. The charge-transfer coefficient, a, was taken as 0.5 and the diffusion coefficients were taken equal for all species (D ¼ 105 cm2 s1 ). Spectroelectrochemistry. UV and visible absorption spectroelectrochemistry in ACN was performed using in situ techniques. Semi-empirical gas phase and solvent dependent MO-calculations were performed using the AM1 Hamiltonian and SM5.4A Charge Model in the AMSOL program package [8,9] and the HyperChem realease 7 software (Hypercube, Inc. Gainesville, Florida).
3. Results 3.1. Constant current electrolysis of arylvinylsulfones 1 In preliminary work [3] reporting the cathodic treatment of arylvinylsulfones 1, it was shown that such structures are able to afford cyclodimerization products (dimer 2) in fair yields. Side products such as those coming from the decomposition of dimer 2 by electrogenerated bases or the protonation of the vinyl moiety (protonation of 1: ) were also isolated [3]. The use of carefully dried solvents limits the formation of electrogenerated bases and favours the selectivity of the cyclodimerization reaction [5]. In the first part of this work, our goal was to find the best experimental conditions by varying different parameters (solvents, supporting electrolyte, charge, etc) (see Table 1). Good results were obtained with constant current electrolysis
177
using a simple two-electrode cell made with two Pt wires immersed in a one-compartment cell strictly deaerated by argon bubbling. A better selectivity in the cyclodimer was found with a low consumption of electricity (in most cases 0.1 electron/molecule), when the applied potential remained practically at the threshold of the first cathodic peak observed with the series 1, for the greatest part of the electrolysis [10]. It is worth noting the simplicity and interest of such a procedure, as it does not require the use of a powerful generator or potentiostat. Concerning the solution, the best yields were obtained when using dimethylsulfoxide (DMSO) containing lithium perchlorate as supporting electrolyte (0.05 M LiClO4 ,) with 2 g of sulfone 1 in 10 ml solution (conditions of entry 4 in Table 1). The use of DMF as the solvent gave also fair yields that were much higher than those already published when using a stirred mercury pool as a cathode [3]. The use of tetrahydrofuran as the solvent did not allow cyclodimerization to occur under these conditions. Moreover, the use of copper and graphite cathodes in DMSO were found to be successful in the formation of 2. Considering these as the best experimental conditions, the method was extended and tested in the preparation of a series of substituted E1,2-bis(R-arylsulfonyl)cyclobutanes 2a–2e where R ¼ 2-naphthyl, 4-methylphenyl, 2-methylphenyl, 4-methoxyphenyl, 2-methoxyphenyl, respectively. Except for 2e, the yield of isolated cyclobutane was higher than 80%. 3.2. Cyclic voltammetry In the second stage of this work, we tried to obtain more insight into the mechanism involved in the catalytic process. Cyclic voltammetric experiments were performed, in DMF and ACN with LiClO4 or tetrabutylammonium hexafluorophosphate (NBu4 PF6 ) as supporting electrolytes. Similar general patterns were obtained with these two salts; however due to a more resistive medium with LiClO4 , which increases the difficulty of achieving full ohmic drop compensation, detailed investigations at fast scan rates were performed only with NBu4 PF6 . With both supporting electrolytes, cyclic voltammetry at low sweep rates of 1a in DMF (Fig. 1(a)-1) shows two redox processes for which the relevant electrochemical data are presented in Table 2. The first wave has no anodic counterpart and the second one is less well-shaped. The first wave becomes reversible at sweep rates higher than 20 V s1 and was assigned (by comparison with the peak current of the one electron oxidation of ferrocene), to the monoelectronic reduction of the monomer (1a/1a ). Analysis according to the usual criteria in cyclic voltammetry shows that the ratio of the peak currents Ipa =Ipc increases with sweep rate. Under the sweep rate conditions corresponding to an irreversible voltammogram, the peak potential Epc
178
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
1.5 1.0 0.5
-0.5
wave I
11
-1.0
Ipc v-1/2 / A s1/2 V-1/2
I / µA
0.0
3
-1.5
1
2
-2.0
10 9 8
wave II
7 6 0.00
0.05
0.10
0.15
0.20
v / V s -1
-2.5 -2.2
-2.0
-1.8
(a)
-1.6
-1.4
-1.2
E/V 2 1
I / µA
0 -1
3 -2 -3
-4 -2.4 (b)
1 -2.2
-2.0
-1.8
-1.6
-1.4
-1.2
E/V
Fig. 1. (a) Cyclic voltammetry of 1a in DMF + 0.2 M TBAPF6 (c ¼ 1:6 103 M): 1, v ¼ 50 mV s1 ; 2–after preelectrolysis for 1 min.; 3, v ¼ 200 V s1 (for curve 3 the current scale is I 102 lA). Inset: Dependence of the peak current function Ipc v1=2 on the scan rate for 1a. (b) Cyclic voltammetry of 1a in ACN (first 3 scans), c ¼ 1:11 103 M, v ¼ 200 mV s1 .
varies linearly with the logarithm of the scan rate with a slope of )33 mV/decade. This value is close to the theoretical value expected for a ‘‘DIM2’’ mechanism, i.e., a reaction scheme involving a rapid electron transfer followed by anion-radical-substrate dimerization (29.5 mV) [11]. The standard rate constant for the heteroge-
neous electron transfer, ks , (uncorrected for the double layer effect) was estimated from the forward and reverse peak potential difference using both Nicholson’s W function [12] and the DigiSim program. A value of 5 cm s1 was obtained. The second redox couple was observed only at low sweep rates and an estimate of the standard ET rate gives ks 0:15 cm s1 , with a standard potential E20 value of )1.952 V vs SCE. If the monomer 1a is submitted to preelectrolysis for about 1 min, only the second redox couple is observed (Fig. 1(a-2)), whereas at high scan rates (200 V s1 ) only the first couple is present (Fig. 1(a-3)). Starting with a solution of dimer 2a, only one reversible redox couple was observed in the range 0 to )2.0 V, characterised by a E0 ¼ 1:950 V (SCE) with ks ¼ 0:14 cm s1 and was assigned to the monoelectronic reduction of the dimer. The dependance of the peak current function, Ip v1=2 , on the scan rate, also included in Fig. 1(a) as an inset, shows the decrease for wave I of the monomer at low scan rates, simultaneously with an increase for wave II, assigned to the reduction of the dimer, due to the progressive transformation of the monomer into the dimer. These effects are visible only for the lowest scan rates where natural convection starts to interfere with the voltammetric response. As the influence of convection is to increase the current with respect to the value expected under theoretical pure diffusion conditions, it results that the variation of the electron stoichiometry can be considered as an upper limit. The marked decrease of the reduction wave of 1a, observed on the second sweep in acetonitrile ‹ (Fig. 1(b)), is also indicative of an ~ ECE mechanism [13]. Considering all these data, the second redox couple observed at the reduction of the monomer 1a was assigned to the monoelectronic reduction of the dimer 2a, formed as a consequence of the anion radical–substrate dimerization. A similar anion radical–substrate dimerization was reported as most probable for the electroreductive coupling of vinylpyridines and vinylquin- olines leading to dipyridylcyclobutane derivatives [14], whereas alternative routes, usually radical anion– radical anion dimerization, followed by protonation, would lead mainly to the linear hydrodimer.‹ Therefore, the following simplified ~ ECE sequence may be proposed:
Table 2 Relevant electrochemical data for compounds 1a and 2a I wave
II wave
I wave
1a
2a
v (V s )
)Epc (V)
)Epc (V)
Ipa Ipc
0.05 0.20 200 1000
1.846 1.860 1.884 1.880
– – 1.824 1.813
– – 0.867 1.170
a
I0
calculated with formula [12]: Ipa =Ipc ¼ Ipa þ pc
0 0:485Isp Ipc
a
þ 0:086.
Epc (V)
)Epa (V)
Ipa Ipc
Epc (V)
)Epa (V)
Ipa Ipc
1.997 1.994 – –
1.895 1.937 – –
0.548 0.559 – –
1.899 1.891 1.967 1.976
– – 1.840 1.829
– – 0.622 1.030
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
1a þ e 1a ;
E10 ¼ 1:856 V
ð1Þ
k2
1a þ 1a ! 2a
2a þ e 2a ;
Table 3 Relevant electrochemical data for compounds 1b and 2b
¼ 1:950 V
ð3Þ
The anion radical–substrate dimerization (reaction 2) leads to the anion radical of the dimer. The E0 for the couple 2a/2a is slightly more negative than the standard potential corresponding to the monomer couple 1a/ 1a . Thus, direct oxidation of 2a to 2a at the electrode (at the monomer reduction potential) or by homogeneous electron transfer between the anion radical 2a and 1a are thermodynamically favoured and account for the catalytic character of the process, as indicated from the decrease of the electron stoichiometry at the lowest scan rates (see inset in Fig. 1(a)). However, this simplified mechanism does not allow a detailed explanation of all the cyclic voltammetry data (for example, it does not explain the irreversibility of the first voltammetric peak). For a full description of the experimental data, it is necessary to take into account in the mechanism the different possible end reactions, especially those involving the disappearance of the dimer anion radical, for which we have only scarce information ( see discussion and MO calculation parts below). The cyclic voltammetry results for 1b and 2b are displayed in Fig. 2 and Table 3. At the reduction of 1b, an asymmetric wave consisting of two superimposed waves is observed at low sweep rates. With increasing scan rate, the first peak becomes more intense, but no reversibility was reached even at 1000 V s1 . Similarly to 1a, the peak potential Epc was found to vary linearly with scan rate and a slope of 28.3 mV/decade was measured. This peak was assigned to the monoelectronic
2b
1b
ð2Þ E20
179
I wave
II wave
v (V s )
Epc (V)
Epc (V)
Epc (V)
0.100 0.200 2 200
2.045 2.047 2.082 2.144
2.083 2.101 2.169 2.270
– 2.139 2.170 2.278
reduction of the monomer (1b/1b ), followed by a rapid coupling reaction between the anion-radical 1b and the substrate 1b. Starting with a solution of dimer 2b only one reduction wave was observed, with a peak potential corresponding to the second redox process visible during the reduction of the monomer. Therefore, we propose a similar ECE sequence (DIM2) to that which describes the reduction of 1a to account for the experimental results. 3.3. UV–VIS Spectroelectrochemical experiments UV–VIS spectroelectrochemical reduction of 1a was investigated in ACN + 0.2 M TBAPF6 (Fig. 3) (investigations in DMF are not possible due to the low absorption wavelength of the arylvinylsulfone). At the potential corresponding to the threshold of the first reduction wave, the data show the progressive transformation of the monomer (characterised by two bands with maximum absorption wavelengths kmax ¼ 235 and 214 nm and absorption coefficients e235 > e214 ) into the absorption spectrum of the dimer 2a, (characterized by two bands located at kmax ¼ 233 and 215 nm with a 1.0
40
9
20
0.8
0
Absorbance
I/µA
-20 -40
2
-60 -80
0.6
0.4
1
1
0.2
-100 -120 -2.6
0.0 200
-2.4
-2.2
-2.0
-1.8
-1.6
-1.4
-1.2
-1.0
225
250
275
300
wavelength / nm
E/V Fig. 2. Fast scan cyclic voltammetry in DMF + 0.2 M TBAPF6 (v ¼ 200 V s1 ): 1, monomer 1b and 2, dimer 2b.
Fig. 3. UV-spectra recorded for the electrolytical reduction of 1a at the potential of the first wave; 1, initial spectrum of the monomer, 2–9, evolution in time, about 2 min between spectra.
180
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
reverse intensity ratio, e233 < e215 ). The new band at 245 nm, which decreases in favour of the dimer band, was tentatively assigned to an intermediate in the reaction sequence, presumably the anion-radical 1a . In the same experiments performed with the compound 1b, the spectroelectrochemical results show the progressive transformation of the monomer (kmax ¼ 234 nm) into the dimer 2b (kmax ¼ 227 nm).
4. Discussion The experimental results may be rationalised in terms of the following ‘‘catalytic DIM2’’ mechanism [5], based ‹ on the ~ ECE reaction sequence previously discussed: The first step in this scheme is the monoelectronic reduction of the substrate to the anion radical, followed by a rapid coupling with another neutral sulfone molecule leading to the dimer anion-radical, 2a,2b . The key step in Scheme 2 is the reoxidation of 2a,2b , which is thermodynamically favoured because of a more positive standard potential for the monomer than for the dimer. In the case of 1a, the anion radical lifetime is of the order of several 104 s, i.e., 1a has the time to diffuse far from the electrode before it reacts. Therefore it follows that the major pathway for the oxidation of 2a is the homogeneous electron transfer from the dimer anion radical to the substrate, leading to the neutral dimer and recycling the anion radical of the monomer: 2a + 1a ! 2a + 1a . To account for a fair yield in cyclodimer and low consumption of current, this reaction should be able to compete with the other decay reactions involving 2a , as a carbon–sulphur bond cleavage or the reverse dimerization (cleavage of the dimer) reaction. An estimation of the homogeneous electron transfer kinetic constant k2 can be performed in terms of the Marcus–Hush model that establishes a relation between the reaction Gibbs energy DG0 and the activation Gibbs
activation
energy DG# [15–19]. A DG0 value of )2.21 kcal mol1 (1 kcal ¼ 4.18 kJ) is derived from the experimental reduction potentials. From the relation DG# ¼ k=4 ð1 þ DG0 =kÞ where k is the intrinsic barrier (we do not consider any working terms or electrostatic contribution because the charge of the substrate is zero), and taking for k a value in the range 20–30 kcal mol1 , as previously obtained for similar organic compounds in ACN [15], an estimate of 4–6 kcal mol1 is obtained for DG# . Considering a collision factor Zsol 6 1011 M1 s1 , this gives an estimate of the rate constant k2 in the range 107 –108 M1 s1 . The Marcus–Hush model also provides a relation between the rate constant for the standard homogeneous electron transfer rate constant ðkex Þ and the heterogeneous ðks Þ electron transfer rate constant for the same reactant [12,15–19]. As shown from data on the electrochemical reduction of a series of aromatic molecules, [20] a good estimation of kex is obtained from ks values that are uncorrected for the double-layer effect and using Hush’s model where the image effect is not taken into account [17,18], due to a cancellation of these two effects: ks =Zel ¼ kex =Zsol . Using ks ¼ 5 cm s1 , as determined from cyclic voltammetry for the first reduction wave of 1a, Zel ¼ 5 104 s1 and Zsol ¼ 6 1011 M1 s1 [12], a kex value of 6 107 M1 s1 is derived leading to a k2 value of the order of 108 M1 s1 . For 1b, it is more difficult to obtain a clear cut result, as we were not able to obtain a measurement of its radical anion lifetime, and we may have to take into account, in addition to the homogeneous oxidation pathway, the possible direct oxidation of 2b at the electrode, which also leads to the regeneration of 1b . In conclusion, we can consider that the homogeneous electron transfer between 2a,2b and 1a,1b is fairly rapid in comparison with other possible reactions of the dimer anion radical, like the reverse dedimerization reaction. Other reactions involving the neutral dimer, for example the decomposition of dimer 2 by electrogenerated bases or the cleavage of the C–SO2 bond with the formation of cyclobutene derivatives will also contribute to decrease the yield of cyclodimer. 4.1. MO-modeling
ArSO2 ArSO2
eSO2Ar
SO2Ar
propagation
. ..-
SO2Ar trans
... SO2Ar
-.
ArSO2 SO2Ar
Scheme 2.
SO2Ar SO2Ar
The MO calculations performed were intended to answer two questions: firstly, to account for the electronic structural features implied in redox processes, explaining the difference in the reducibility of the monomers and dimers in both gas phase and solvent, and secondly to furnish theoretical support for the proposed ETC mechanism, by analysing different reaction pathways implied in the mechanism in Scheme 2. The redox properties of the compounds investigated were compared with theoretical predictions in the framework of the semi-empirical AM1 method, in both the gas phase and polar solvents (ACN, DMF). The
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
in the distance between the two molecules. Two attempts were made. The potential energy surface (PES) was built assuming firstly a synchronous approach of both vinyl carbon atoms d110 ¼ d220 ¼ d and, secondly, considering an asynchronous way, i.e., a previous formation of the bond between the b carbon atoms. In the first case, the plot presented in starting with a distance of 3.5 A, Fig. 4 reflects an increasing energy up to a distance of and the presence of a minimum point energy about 2 A As corresponding to the dimer anion radical at 1.54 A. these are gas phase calculations, we cannot consider that the maximum point of energy on the PES can give an estimation of the real energy barrier of the process. The modeling using the second way of approaching the molecules led to unreliable results. As the single restriction between the two molecules was the distance between the b-carbon atoms, d110 , the two fragments were able to adopt different relative positions and the formation of the cyclobutane ring could not be shown. It is likely that even if the dimerization process is an asynchronous process, the use of a single reaction coordinate is not sufficient to describe it.
120 -1
100 Erel / kcal mol
estimation of the solvation effects was performed with the AMSOL program. The redox properties of the compounds investigated as determined by the features of the frontier orbitals outline the fact that in the monomers the lowest empty molecular orbital (l.e.m.o.) is localized mainly on the vinyl moiety and the sulfone group, whereas in the dimer, it is localized on the sulfone and the aryl substituent. Therefore, if the reactivity in the cyclodimerization reaction is mainly determined by the vinyl group, and probably more influenced by the substitution there, the reactivity of the dimer and its reduction potential are mainly determined by the sulfone group and the aryl moiety. The different degrees of localization of the l.e.m.o. orbitals on the different moieties of the molecule are responsible for the order of the reduction potentials of the monomers and dimers and therefore are important for the catalytic process observed experimentally. Standard potentials E0 can be compared with adiabatic electron affinities, calculated as the difference between the formation enthalpies of the anion radical and neutral molecule. A further indication of the reactivity in reduction processes is given by the vertical and adiabatic electronegativities. The calculated electron affinities and electronegativities are presented in Table 4, for both the gas phase and acetonitrile, together with the experimental potentials derived from cyclic voltammetry experiments. For the couple 1a/2a, the calculations predict that E0 of the monomer is very close to that of the dimer (slightly more positive) in agreement with cyclic voltammetry data. However, for the p-tolyl derivatives 1b/2b, although nearly equal electron affinities and vertical electronegativities are found from gas phase calculations, the reverse order of reducibility is predicted in acetonitrile. Probably, more elaborate calculations are necessary to account for this discrepancy. Comparing the results in ACN with the gas phase calculations, higher electron affinities and vertical or adiabatic electronegativities are observed in ACN, attesting to an enhanced reducibility in polar solvents. A first insight on the initial step of the reaction mechanism presented in Scheme 2, the interaction of the substrate with the generated anion radical, was obtained by monitoring the total energy with respect to the change
181
SO2Ar
] -.
[
80
] -.
SO2Ar
[1
2
1'
2'
SO2Ar
SO2Ar
60 40 20 0 1.5
2.0
2.5
3.0
d / Å (d = d11' = d22') Fig. 4. Potential energy surfaces for the anion radical–substrate dimerization of 1b in the gas phase, considering a synchronous approach of the two species. Erel represents the energy with respect to the minimum of the dimer anion radical 2b .
Table 4 AM1-calculated redox properties of the compounds investigated in the gas phase and ACN Gas phase
ACN
Compound
Aad (kcal mol )
vv (eV)
vad (eV)
Aad (kcal mol1 )
vv a (eV)
vad b (eV)
0 Eexp (V)
1a 2a 1b 2b
)42.89 )41.89 )29.51 )30.02
5.04 5.07 5.37 5.40
5.09 4.94 5.21 5.69
)83.96 )83.52 )77.42 )83.92
5.26 5.21 5.54 5.72
5.36 5.28 5.49 5.74
)1.856 )1.950 )2.047 )2.139
a
1
a
b
vv ¼ 1=2ðehomo þ elemo Þ. vad ¼ 1=2ðIad þ Aad Þ; Iad , the adiabatic ionisation potential, is given by the enthalpy of the process, M ! Mþ + e ; Aad , the adiabatic electron affinity, is given by the negative of the enthalpy of the process, M + e ! M . b
182
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
The fully optimized structure of the trans-dimer anion radical in ACN reflects the presence of several conformers with heats of formation of about )181 kcal mol1 , differing by the positions of the sulfone and the aryl groups. Comparing the energy of the minimum with the sum of the energies of the parent compound and the corresponding anion-radical ()163 kcal mol1 ), it can be seen that the dimerization process represents a stabilisation of the system. The most striking feature of the 2b geometry and charge density distribution consists in an unexpected asymmetry, reflected by the fact that the C–SO2 distances are not equal and the charge is mainly localized on one of the dimer moieties. An important problem in Scheme 2 is the further reactivity of the dimer anion radical, consisting in several possible reactions which could compete with the homogeneous electron transfer, like the cleavage of the C–SO2 bond with elimination of the arylsulfonyl group, or the reverse dedimerization process. In the following, these reactions were modeled in DMF using the appropriate reaction coordinates, in order to obtain an estimation of their relative probability. In the case of the cleavage reaction of the carbon– sulfur bond, starting from the fully optimised dimer anion-radical, 2b , one C–SO2 distance was increased progressively, but the energy increased steadily and at reached a maximum and then decreased. An about 2.6 A estimate of the energy barrier gives about 26 kcal mol1 . Therefore, it seems rather unlikely that this reaction could occur from the dimer anion radical. The reverse reaction, the dimer anion radical dissociation process, was modeled considering as a reaction coordinate the distance between the arylsulfone substituted carbons, d220 . The plot in Fig. 5 shows two bran-
(∆H+∆Gsol)rel / kcal mol
-1
25 20 II 15
I
10 5
ches. Starting from the minimum value, the energy increased steadily up to a value of about 2.0–2.1 A (branch I) differing for the chosen conformer. At that point a sudden decrease was noted and the plot continued smoothly towards a broad minimum belonging to the open form of the dimer anion radical (branch II). The further increase of the distance d has no further significance as the free arylsulfone fragments can rotate freely and adopt several conformations. Up to the in spite of the opening of the mentioned value of 2 A, cyclobutane ring, the geometric parameters of the arylsulfone fragments as well as the charge distribution are not significantly modified from those of the closed dimer major changes were noted structure. At about 2 A, which were maintained in some conformations of the open dimer anion radical and reflect an overall symmetry, i.e., equal C–SO2 distances and charges of the two fragments. Starting from the open form minimum and conducting the optimization in the reverse direction, decreasing the d value, a crossing of both branches of the plot was obtained. Considering the crossing region as a transition state, an activation energy of about 20 kcal mol1 can be estimated for the cyclobutane ring opening process. This high value of the activation energy, even in polar solvents like ACN or DMF, seems to indicate that the cyclobutane ring opening reaction is not sufficiently rapid to compete with the homogeneous electron transfer, responsible for the catalytic character of the process. Another possible reaction, which can contribute to lower yields in cyclodimer, is the cleavage of the C–SO2 bond under the influence of electrogenerated bases, B , starting from the neutral dimer and leading to arylsulfonyl cyclobutene derivatives. Therefore we found it interesting to calculate the potential energy surface for this reaction. This was performed in two steps, as shown in Scheme 3: firstly, one proton was removed from a substituted carbon (C3) in the cyclobutane ring and the resulting anion was fully optimised. Secondly, starting with the anion 2b , the C–SO2 bond ðd2–5 Þ was increased. The PES were calculated both in the gas phase and in DMF and are presented in Fig. 6. It may be stated that the energy barrier in DMF (9 kcal mol1 , Fig. 6, curve 2) is about the half that in the
0 5 SO2-Ar
-5 1.5
2.0
2.5 d22' / Å
3.0
2b Fig. 5. Potential energy surface calculated in DMF for the dimer anion–radical 2b dissociation, with respect to the distance between the arylsulfone substituted carbons, d220 . The ðDH þ DGsol Þrel are relative values with respect to the minimum of the dimer anion radical 2b . Branches I and II (see text).
+ B-
1
2 H
4
3
2b-
(-)
-
Ar-SO-2
SO2-Ar 6 Scheme 3.
SO2-Ar
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
24
(∆H+∆Gsol)rel / kcal mol
-1
20 16 1
12 8
2
4 0 1.5
2.0
2.5
3.0
3.5
4.0
d25 / Å Fig. 6. Potential energy surface for the cleavage of the C–SO2 bond ðd2–5 Þ, under the influence of electrogenerated bases, starting from the neutral dimer: 1, gas phase; 2, DMF. The ðDH þ DGsol Þrel are relative values with respect to the minimum of the dimer anion radical 2b .
1.8
double bond 1.6
gas phase (20 kcal mol1 , Fig. 6, curve 1), and about the third of the estimated energy (26 kcal mol1 ) for the same reaction starting from the dimer anion radical. The progress of the reaction is well reflected also by the evolution of the cyclobutane ring bond order p2–5 , as well as by the total charge on the leaving arylsulfonyl group, presented in Fig. 7(a) and (b). It may be observed that the cyclobutane bond evolves from a single to a double bond at the end of the reaction (Fig. 7(a)) and the charge on the arylsulfonyl moiety changes from 0 at the beginning of the reaction, to )1 at the end (Fig. 7(b)). In conclusion, AM1 modelling accounts satisfactorily for the electronic features of the frontier orbitals implied in redox processes for both monomers and dimers. However, these methods appear as not sufficiently reliable, especially when solvent effects have to be considered, to account for the small differences in the experimental reduction potentials. AM1 modeling in polar solvents of the reaction routes implying the anion radical–substrate dimerization indicate that the other possible reactions of the dimer anion radical are slow as compared with the electron transfer to the substrate and this therefore represents support for the proposed reaction mechanism. The cleavage of the C–SO2 bond under the influence of electrogenerated bases leading to cyclobutene seems to proceed rather from the neutral dimer than from the dimer anion radical.
p
1.4
183
1.2
5. Conclusions
1.0
single bond 0.8 1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
d2-5 / Å
(a)
0.0
In conclusion, the electrochemical and UV-spectroelectrochemical data, as well as the MO calculations performed, support the ETC–catalytic mechanism ‹ (Scheme 3), based on an ~ ECE reaction sequence, followed by electron transfer from the dimer anion radical to the substrate. Homogeneous ET from the dimer anion radical to the substrate, analysed according to the Marcus–Hush model seems to be more rapid than other follow-up reactions of the dimer anion radical.
q Ar-SO2
-0.2
References
-0.4
-0.6
-0.8
-1.0 1.6
(b)
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
d2-5 / Å
Fig. 7. Evolution of the p2–3 bond order (a) and of the total charge on the arylsulfone fragment (b) during the cleavage of the C–SO2 bond.
[1] M.S. El Faghi El Amoudi, P. Geneste, J.L. Olive, J. Org. Chem. 46 (1981) 4258. [2] M.S. El Faghi El Amoudi, P. Geneste, J.L. Olive, Nuov. J. Chim. 5 (1981) 251. [3] J. Delaunay, G. Mabon, A. Orliac, J. Simonet, Tetrahedron Lett. 31 (1990) 667. [4] A. Orliac-Le Moing, J. Delaunay, J. Simonet, New J. Chem. 18 (1994) 901. [5] J. Delaunay, A. Orliac-Le Moing, J. Simonet, New J. Chem. 17 (1993) 393. [6] D. Garreau, J.-M. Saveant, J. Electroanal. Chem. 35 (1972) 309. [7] M. Rudolph, D..P. Reddy, S.W. Felberg, Anal. Chem. 66 (1994) 589A.
184
J.F. Bergamini et al. / Journal of Electroanalytical Chemistry 569 (2004) 175–184
[8] G.D. Hawkins, G..D. Giesen, G.C. Lynch, C.C. Chambers, I. Rossi, J.W. Storer, J. Li, T. Zhu, D. Rinaldi, D.A. Liotard, C.J. Cramer, D.G. Truhlar, AMSOL version 6.5.3, University of Minnesota, Minneapolis, 1997. [9] C.C. Chambers, C.J. Cramer, D.G. Truhlar, J. Phys. Chem. 100 (1996) 16385. [10] J. Delaunay, J. Simonet, unpublished results. [11] C.P. Andrieux, L. Nadjo, J.M. Saveant, J. Electroanal. Chem. 42 (1973) 223. [12] A.J. Bard, L.R. Faulkner, Electrochemical Methods, Fundamentals and Applications, second ed., Wiley, New York, 2001, p. 122, 240.
[13] P.H. Rieger, Electrochemistry, Chapman & Hall, New York, 1994, p. 304. [14] R.G. Janssen, M. Motevalli, J.P. Utley, Chem. Commun. (1998) 539. [15] L. Eberson, Electron Transfer in Organic Chemistry, Springer Verlag, Berlin, 1987, p. 29. [16] R.A. Marcus, J. Chem. Phys. 24 (1956) 966, 979. [17] N.S. Hush, J. Chem. Phys. 28 (1958) 962. [18] N.S.J. Hush, Trans. Faraday Soc. 57 (1961) 557. [19] R.A. Marcus, J. Chem. Phys. 43 (1965) 679. [20] H. Kojima, A.J. Bard, J. Am. Chem. Soc. 97 (1975) 6317.