J. CHEM. SOC. FARADAY TRANS., 1994, 90(21), 3241-3244 Theoretical Study of the Electroreduction of Halogenated Aromatic Compounds Part 3.”f-, m-and p-Dibromobenzenes studied by AM1 and PM3 Methods Roberto Andreoli, Giovanna Battistuzzi Gavioli, Marco Borsari and Claudio Fontanesi* Universita degli Studi di Modena , Dipartimento di Chimica, Via Campi 183,4I 100Modena , Italy ~~~ The electroreductive potential values of the o-, rn-and p-dibromobenzenes (DBBs) follow an unusual pattern in that, unlike the structurally related dichlorobenzene (DCB) derivatives, these three isomers exhibit a strong ‘ortho effect’ and are accompanied by a large difference in the €,/2 values. Any interference with the mecha- nism of reduction on the part of the chemical enivironment can be safely ruled out, given the consistency of the €,/* values obtained in four different solvents.The use of theoretical indices calculated by the PM3 method enables the experimental behaviour of the DBBs to be rationalised on the basis of the electronic structure of the neutral isolated molecule. Remarkably, PM3 indicates the formation of a o-type radical anion in the case of the reduction of the bromo benzene derivatives. The theoretical rationalisation of the electroreductive mecha- nism of aromatic compounds bearing one or more halogens bound to the aromatic moiety has long been a matter of debate.’-’’ The influence of a broad set of parameters on the electrochemical reduction of this class of compounds (e.g. solution acidity and the number and type of substituents bound to the aromatic ring) has been extensively examined.The experimental evidence suggests that, as is generally accepted, reduction involves the cleavage of the C-X bond (which turns out to be a multiple cascade cleavage process when more than one halogen is present’ 2,1 3, and is an overall irreversible diffusion-controlled two-electron process. The first step of the reduction mechanism involves a reversible one-electron transfer which produces a radical anion. This latter species, following the cleavage of the C-X bond, yields the aryl radical. A second electron transfer or a dispro-portionation reaction (or, more rarely, other types of interme- diate stepsI4), both followed by protonation, leads to the final aromatic product.”-’ Mo reover, great efforts have been made to rationalise these results theoretically in terms of the molecular structure of the compounds undergoing reduction. In the present study, all the mono-and di-substituted chloro-and bromo-benzene derivatives are investigated under the same experimental conditions in four solvents fea- turing quite different physical and chemical characteristics : dimethyl sulfoxide (DMSO), dimethylformamide (DMF), ace- tonitrile (ACN) and ethanol (EtOH). In particular, the electroreductive behaviour of the dichloro- and dibromo-substituted benzenes is closely com- pared, since the ordering of the reduction potential is reversed when the experimental Eliz values of the ortho, meta and para derivatives of the two series are considered.The dibromobenzenes, unlike the dichlorobenzenes, exhibit an ‘ortho effect’ (i.e. the reduction potential of the ortho isomer is patently more positive than that of the meta and paru The experimental findings were subjected to a process of theoretical rationalisation using the AM1 l9 and the quite recent PM320 parametrisations, both of which belong to the family of NDDO-based calculations. AM1 was particularly effective in the study of the electro- reduction of a series of compounds featuring both bromo- and chloro-substituted aromatics, whereas the original MNDO parametrisation t Part 2: J. Chem. SOC.Faraday Trans., 1993,89, 3931. The PM3 method was selected by virtue of the fact that simultaneous optimisation of the parameters relating to H, C, C1 and Br atoms has been reported.” In the case of the orig- inal MND02’ and AMl19 parametrisation, the H and C atom parameters are determined simultaneously, while the C1 and Br atom parameters are obtained in two successive steps.19.2 1-24 Calculations All molecular orbital data are calculated at the single-determinant level and are of RHF or ROHF type.Open-shell calculations are made by employing the half-electron method2’ as implemented in the AMPAC 2.1 and MOPAC 6 suite of pr~grarns.~~,~~ Geometries of the neutral molecules are fully optimised. Experimental Chloro- and bromo-benzene, and dichloro- and dibromo- benzenes were purchased from Carlo Erba (R.P.E.) and used without any further purification.Polarographic and voltam- metric measurements in the depolariser concentration range (0.5-5) x mol dm-3 were carried out by a PAR 273A Potentiostat/Galvanostat, using DMF (Fluka, <0.01% H,O), DMSO (Fluka, <0.01% H20), ACN (Fluka, <0.01% H,O) and EtOH (Carlo Erba, <0.1% H20) as solvents. The working electrodes were a DME and an HDME. The reference electrodes were: Ag/AgCl, NaCl,,,. in DMF, Ag/ AgC1, KCl,,,. in DMSO, Ag/AgN03 (0.1 mol dmP3) ACN and Hg/Hg2C12, KCl,,,. in EtOH. A Pt sheet was used as the counter electrode. The Eli! values were calibrated against the cobaltocene/ cobaltocenium couple and an aqueous saturated calomel electrode (SCE), with an accuracy of ca. f.10 mV, was used as reference.The ionic strength of all the solutions was kept constant (I = 0.1 mol dmP3), using Bu,NClO, (Fluka) as the supporting electrolyte, as was the cell temperature (T = 298 & 0.1 K). Results and Discussion Electrochemical Behaviour The half-wave potentials of the eight halogenobenzene derivatives in DMF, DMSO, ACN and EtOH, are reported in Table 1. Chloro- and bromo-benzene present only one bielectronic, diffusion-controlled reduction wave, no matter which solvent J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 1 Polarographic reduction potential values in DMSO, DMF, ACN and EtOH, dropping time 1 s no. compound chlorobenzene bromobenzene 1,2-dichlorobenzene 1,3-dichlorobenzene 1,4-dichlorobenzene 1,Zdichlorobenzene 1,3-dichlorobenzene 1,4-dichlorobenzene DMSO DMF ACN EtOH wave I wave I1 wave I wave I1 wave I wave I1 wave I wave I1 2.740 - 2,750 - - - 2.820 - 2.610 - 2.630 - 2.700 - 2.680 - 2.505 2.735 2.525 2.755 2.635 - 2.610 2.810 2.475 2.740 2.485 2.750 2.620 - 2.570 2.810 2.470 2.735 2.490 2.750 2.595 - 2.565 2.820 2.027 - 2.045 - 2.060 - 2.005 2.685 2.155 2.615 2.180 2.635 2.215 2.690 2.140 2.685 2.330 2.615 2.355 2.630 2.385 2.695 2.365 2.680 is used.The semilogarithmic plot indicates that the electron transfer is irreversible, while the a, value is ca. 0.65 in aprotic solvents (DMF, DMSO, ACN) and 0.69 in EtOH. In the case of the dihalogenobenzenes, and with the excep- tion of 1,2-dibromobenzene, which exhibits a single four- electron wave, two bielectronic, diffusion-controlled waves are observed.Again, the electron transfer is invariably irre- versible and the a, values are similar to those obtained for the monohalogenobenzenes, except in the case of 1,2-dibro- mobenzene, for which an = 0.72 and 0.75 in aprotic and protic media, respectively. The product id q1/2/n (where id is the diffusion current, q the solvent viscosity and n the number of electrons involved in the reduction process determined by coulometric measurements) is practically constant for all the polarographic waves in all four solvents. The irreversibility of the overall reduction process was also confirmed by cyclic voltammetry measurements for scan rates up to 1OOovs-'. In particular, for DCBs and DBBs, the value of the first reduction wave depends on the type and position of the halogen atom on the benzene ring, while the ElIz value of the second wave (save for o-DBB) is constant and corre-sponds to the EIl2 value of the monohalogenosubstituted compounds.In general, the observed polarographic behav- iour is in agreement with previously published studies on the electroreduction of chlorobenzenes. '-'vl 2-14918 According to the evidence obtained in the case of the reduction of other halogenoaromatic compounds, and to the results of the UV spectra analysis performed on the products of the coulometry, the overall electrochemical process under- lying the first wave of the dihalogeno-derivatives can be sum- marised as follows : PhX, + 2e-+ H+ -+ XPhH + X-(1) and XPhH is subsequently reduced to PhH by the same mechanism.For o-DBB the same result is obtained in a single polarographic wave. Fig. 1 shows plots of in DMSO, ACN and EtOH us. in DMF, to be linear. The same sign of the slope for all the solvents indicates that the compounds always undergo reduction in the neutral molecular form. In particular, the linearity and the slope of unity suggest that structural effects play the same role in the determination of the values in the different solvents. These findings also indicate the absence of any significant interference by the solvent on the reduction mechanism illustrated above (e.g.rearrangement in the solva- tion shell or direct solvent intervention in the redox process). Note that there is close agreement between the electro- chemical behaviour described here and the data reported both in ref. 18 (concerning compounds 1, 2, 6, 7, 8: El.,; 18 =ref -0.823 + 1.02E;y; work, r = 0.996; besides the good linearity the two series of data are monotonically ordered, i.e. without any inversion) and in ref. 12 (again, a monotonic ordering of the reduction potentials of chlorobenzene and of the DCB isomers is observed). Theoretical Indices In this study the chlorinated aromatic derivatives serve essen- tially as reference compounds, as their electrochemical behav- iour has been satisfactorily described in ref.2 and 3. Attention is drawn rather to the o-, rn-and p-DBB com- pounds, which follow a rather peculiar reductive pattern (Table 1).In particular, there is a strong 'ortho effect', charac- terised by considerable divergence of the El/2 values of the individual isomers, whereas the half-wave potentials of the DCBs are fairly close. Attempts have been made to rationalise the anomalous behaviour of the o-, rn-and p-DBBs by invoking certain par- ticular " In the case of the reduction of the ortho isomer the forma- tion of benzyne has been hypothesised.18 This could account for the single reduction wave shown by the ortho-isomer, but the difficulties encountered in attempting to rationalise the behaviour of the rneta and para isomers are left unresolved (difficulties shown both by the MND06 and AM1 methods.Fig. 2 and 3). Also, although the multivariate fit of the ElIz values using CT*energies and the net charge of the leaving group = a + bE@ + cqleavingX)is able to predict correctly the ortho-rneta ordering of the DBBs, it fails completely to predict that of the DCBS;~ for, in DMF the calculated potentials are -1.740 and -1.88 V for the o-and rn-DCB isomers, respec- 3.0 I I , I 02.8 v 0" 2.6 v) 2 $2.4 . 7 G I d'2*2ii, 12.0 -, , 2.0 2.2 2.4 2.6 2.8 3.0 -Ei/,(DMF)/V VS. SCE Fig. 1 El,, measured in DMF us. El/, measured in DMSO (V),ACN (0)and EtOH (0) J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 -0.1 0 I 4o03 h2 -0.2 05 -0.3106 07 t -0.4 - (38 - -0.5." " ' " ' 1 , l I I I , tively,6 while experimentally they were reported as under- going reduction at the same value of potential (-1.81 V).6 Note also that this latter experimental result is inconsistent both with the data presented in ref. 12 and with our findings. (Table 1, column 2). These two sets of experimental data (ours and that of ref. 12) indicate that the qualitative predic- tion of the ordering of the reduction of the DCBs6 is also wrong. This failure is probably due to the inability of the original MNDO parametrisation to account for variation in the molecular electronic structure when dealing simulta- neously with aromatic compounds bearing Cl, Br and I sub-stituents, at least in the field of electrored~ction.~*~~ Effects induced by the chemical environment (e.g. prefer-ential solvation) can be excluded, as indicated by the absence of any inversion, in the series of the EIl2 values, when the solvent is changed (Table 1, Fig.1) despite the fact that the four different solvents employed are characterised by signifi- cant variations in both physical (relative permittivity, vis- cosity, etc.) and chemical (acidity, DN, AN, etc.) proper tie^.^^ Thus they create reaction environments which are substan- tially different. Against this background, the overall electro- reductive behaviour of DBBs (notably, evidence of the 'ortho effect') still remains unclear. The lowest unoccupied molecular orbital (LUMO) energy (and type) and vertical electron affinity, A, = E,,,, URA 0.0 -0.2 -0.4 2 4--.400 3 -0.6 05 07 06-0.8 08 -1 .o 2.0 2.2 2.4 2.6 2.8 3.0 -EII,(DMF)/V vs. SCE Fig. 3 Calculated electron affinity, A,, us. El/* measured in DMF; AM1 method 3243 -E,,,, ,f are the theoretical indices relating to the char- acterisation of the electrochemical behaviour of the com-pounds studied, they are calculated using AM1 and PM3 methods. The molecular descriptors obtained by the AM1 method give the wrong reduction ordering (Fig. 2 and 3), mainly for the DBB isomers. The use of the AM1 method in the study of the electroreduction of organic compounds appeared very prom- ising, owing to its ability to rationalise, in a single corre- lation, the electroreductive behaviour of organic compounds bearing a Br or a C1 substituent, as well as affording a correct prediction of the reduction product^.^ This failure was there- fore particularly frustrating.Alternatively, the values calculated by means of the PM3 method afford a monotonically ordered and satisfactorily linear pattern when plotting both LUMO energies and verti- cal electron affinity us. El12.(Fig. 4 and 5). This finding sug- gests that the electroreductive behaviour of both the chloro- and bromo-benzene derivatives is governed by the electronic -0.4":, 2.0 2.2 2.4 2.6 2.8 3.0 -EII,(DMF)/V VS. SCE Fig. 4 Calculated LUMO energies us. El/2 measured in DMF; PM3 method -0.2 1 -0.4 2--. q! -0.6 -0.8 -1 .o 21.0 2.2 2.4 2.6 2.8 3.0 -E,/,(DMF)/V vs.SCE Fig. 5 Calculated electron affinity, A,, us. E,/, measured in DMF; PM3 method t URA stands for unrelaxed radical anion; unrelaxed means, in this context, that the spatial nuclear coordinates of the radical coin- cide with those of the neutral parent molecule. 3244 features of the isolated neutral molecule. This is in agreement with results obtained in previous studies. 1-3 In particular, in the case of the o-DBB, the formation of benzyne as a reaction intermediate cannot be completely excluded. For, its LUMO energy is -0.902 eV (PM3), which is lower than the value determined for the o-DBB. Actually, if benzyne were formed following the reduction of the o-DBB, it would undergo sudden reduction.On the other hand, o-DBB fits satisfactorily the correlations in Fig. 4 and 5, thus if we accept the hypothesis of benzyne formation, we must con-clude that it has a negligible effect on the reduction potential value. Note that, beyond the purely structural nature of the EL,,, us. E 1/2 correlation, which enables a clear structure- activity relationship to be established between the com-pounds in question and their observed experimental behaviour, the linear trend also found when plotting A, us. ElI2 (Fig. 5) indicates that the first electron transfer deter- mines the value of El,,; moreover, approximation of the slope to unity in Fig. 5 implies that our El/, values are directly proportional to the E' values (i.e.E' = a + E1/2).3917 In fact, a qualitative systematic difference between the values obtained by the two theoretical methods, appears to constitute the underlying cause of the significantly different outcome.For, the LUMOs calculated by the AM1 method are always of the n-type, and the same result is found not only when dealing with a large series of halogenated com- pound~,~ Thisbut also when using the MNDO meth~d.~.~ would appear to be a general tendency of these NDDO- based methods. However, the LUMOs of all 12 possible chlo- rinated benzene derivatives calculated by CND0/2 have been reported as being a mixture of CT and n virtual molecular orbitals. When calculated by PM3, the LUMOs were found to be of a n nature of the chlorinated compounds and a for the bromo- and dibromo-benzenes. Note that the a-type LUMO is localised on the C-Br bond, so the bromine is predicted to be the leaving group, as is found experimentally.Note that in the calculation of A,, both the LUMO of the neutral parent compound (no. 2, 6, 7, 8) and the SOMO (singly occupied molecular orbital) of the radical anion is of the a-type. This suggests that, in the case of the bromine derivatives, a r~ radical is formed, which is at variance with the findings for the chloro-substituted aromatics (compounds no. 1, 3, 4, 5). The parameters used in the MNDO, AM1 and PM3 methods reveal that the most important change is due to a variation in bromine parametrisation. In particular, the &/&, ratio changes progressively: 0.9, MNDO; 2.2, AM1; 4.6, PM3; where s and p refer to the s and p atomic orbitals, respectively.Indeed, b, and j?, are the two parameters used in the calculation of the Fock matrix elements connecting orbitals on different atoms.,' Consequently, this seems to be the origin of the r~ nature of the LUMO in the case of the bromo-substituted benzenes when calculations are carried out by means of the PM3 method. In any event, the finding that bromo-substituted benzenes featuring a-type LUMO (and SOMO) electronic configu- rations are the most stable neutral closed-shell (and radical anion) species (with respect to the n ones) should be viewed with caution, and verified before any definitive physical meaning is attached to it.This particular point is currently being investigated using ab initio methods. Conclusion In the study of the electroreduction (or oxidation) of organic compounds the use of molecular descriptors/half-wave poten- J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 tial relationships as an aid to the interpretion of the redox mechanism at a molecular level, is widespread. The 'quality' of the theoretical indices employed in such relationships has been improved from Hammett constants to quantum-mechanical molecular descriptors, which can be cal- culated at various levels of sophistication. l~~ On the basis of the present findings the PM3 method appears to be able to rationalise correctly the peculiar redox behaviour of DBBs in terms of the electronic structure of the neutral isolated molecule.Thus, its use is suggested when theoretical indices are employed in the modelling of the redox behaviour of bromo- and chloro-substituted derivatives. This work was supported by a grant from the Consiglio Nazionale delle Ricerche (CNR), Roma, and from the Minis- tero dell' Universita e della Ricerca Scientifica e Tecnologica, Roma, MURST 40%. Calculations were performed at the CICAIA, Universita di Modena. References 1 M. Borsari, C. Fontanesi, R. Andreoli, L. Benedetti, G. Battis- tuzzi Gavioli and G. Grandi, Electrochim. Acta, 1988,33, 1085. 2 L. Benedetti, G. Battistuzzi Gavioli and C. Fontanesi, J. Chem. SOC., Faraday Trans., 1990,86, 329. 3 G. Battistuzzi Gavioli, M.Borsari and C. Fontanesi, J. Chem. SOC., Faraday Trans., 1993,89,3931. 4 F. A. Beland, S. 0. Farwell, P. R. Callis and R. D. Geer, J. Electroanal. Chem., 1977,78, 145. 5 J. Komenda and P. Janderka, Collect. Czech. Chem. Commun., 1979,44,1715. 6 J. Casado, J. Gallardo and M. Moreno, J. Electroanal. Chem., 1987,219,197. 7 C. P. Andrieux, J. M. Saveant and D. Zann, Noun J. Chim., 1984, 8, 107. 8 J. Grimshaw and J. Trocha-Grimshaw, J. Chem. SOC., Perkin Trans. 2, 1975,216. 9 C. P. Andrieux, J. M. Dumas-Bouchiat and J. M. Saveant, J. Electroanal. Chem., 1978,87, 55. 10 F. M'Halla, J. Pinson and J. M. Saveant, J. Am. Chem. SOC., 1980,102,4120. 11 C. P. Andrieux, J. M. Saveant and K. Binh Su, J. Phys. Chem., 1986,90,3815. 12 S. 0.Farwell, F. A. Beland and R. D. Geer, J. Electroanal. Chem., 1975,61,303. 13 S. 0. Farwell, F. A. Beland and R. D. Geer, J. Electroanal. Chem., 1975,61,315. 14 C. P. Andrieux, C. Blocman, J. M. Dumas-Bouchiat and J. M. Saveant, J. Am. Chem. SOC., 1979,101,3431. 15 A. J. Bard and A. Merz, J. Am. Chem. SOC., 1979,101,2959. 16 C. Amatore, M. A. Oturan, J. Pinson, J. M. Saveant and A. Thiebault, J. Am. Chem. SOC.,1985, 107, 3451. 17 C. Amatore, H. Lund and M. M. Baizer, in Organic Electrochem- istry, Marcel Dekker, New York, 1991. 18 S. Wawzonek and J. H. Wagenknecht, J. Electrochem. SOC., 1963,110,420. 19 M. J. S. Dewar, E. G. Zoebisch, E. F. Healy and J. J. P. Stewart, J. Am. Chem. SOC., 1985,107,3902. 20 J. J. P. Stewart, J. Comput. Chem., 1989, 10,209. 21 M. J. S. Dewar and W. Thiel, J. Am. Chem. SOC.,1977,99,4899. 22 M. J. S. Dewar and H. S. Rzepa, J. Comput. Chem., 1983,4,158. 23 M. J. S. Dewar and E. F. Healy, J. Comput. Chem., 1983,4,242. 24 M. J. S. Dewar and E. G. Zoebisch, Theochem., 1988,180, 1. 25 M. J. S. Dewar, J. A. Hashmall and G. C. Venier, J. Am. Chem. SOC.,1968,90,1953. 26 QCPE 506, Quantum Chemistry Program Exchange, Indiana University, Bloomington, U.S.A., 1989. 27 QCPE 455, Quantum Chemistry Program Exchange, Indiana Univeristy, Bloomington, U.S.A., 1990. 28 R. Artioli, Dissertation, University of Modena, 1992. 29 Handbook of Chemistry and Physics, 56th edn. CRC Press, 1975, Cleveland. Paper 4/03983A; Received 1st July, 1994