New aspects of electrophylic aromatic substitution mechanism: Computational model of nitration reaction
Charge and Electron Transfer
Considerable redistribution of electron density occurs during interaction of nitronium and aromatic compound, and positive charge in σ-complex concentrates on hydrogen atoms of aromatic ring*. But is there electron transfer (jump) occurrence, and does the system gain biradical character? After all, if encounter of reagents proves to be ineffective, do they part without changes or as products of reduction/oxidation?
Let us evaluate thermochemical possibility of electron transfer, considering reaction:
C6H6 + NO2+ ⇔ C6H6•+ + NO2•
Energy of ET equals the difference of electron affinities (EAs) of cations of nitration agent and of substrate. As Table V shows, in gas-phase approximation (i.e., for nitration in gaseous phase),
NO2+ has higher EA than that of substrates for all methods used, so ET is possible in that case. But an electrophil that is a model of solvated nitronium (nitracidium) is not an oxidant for even active aromatic compounds. Consideration of EA calculated in liquid phase approximation** gives the same conclusion. The reason is obvious: in liquid phase, the ET to nitrogroup causes its desolvation and thence considerable loss of energy (the energy of solvation of aromatic cation is smaller than that of nitronium because its positive charge is delocalized over benzoic ring). So in liquid phase the considered equilibrium may be inclined to its left part, in other words, no ET takes place.
Comparison of isolated reagents gives rise to considerable inaccuracy caused by the differences in basis sets, especially in CI calculations. Therefore, it was better to discuss reciprocal donor–acceptor properties investigating dynamics of interaction. But here we saw that results of calculations of supermolecule depended on the method of calculation: increasing the distance between reagents, we got either nitronium and benzene (RHF) or neutral nitrogen dioxide and radical cation of benzene (unrestricted Hartree–Fock [UHF] and CI). The problem was that UHF and RHF should not be used in such a case, but the required level of CI was not reachable in semiempirical packages. The situation was complicated by the fact that adiabatic approximation ceased to be correct at the
C-N distances over 3.5 Å: there were two close singlet states of supermolecule, the one with closed shell and the charge on nitrogroup, and the other, of biradical nature, with positive charge on substrate. Those states differed only in geometry of
NO2 group: linear in the first case and bent in the second***
Investigation of all these circumstances is far beyond the competence of semiempirical methods, so we return to the chemical aspects of the process, which are usually lost in discussion on ET. Investigators imply that ET is the main stage of reaction, that if ET is realized, the destiny of radical pair is determinate: to recombine. Such is not the case: all used methods, without exception, predict energy growth along with reagent's approach. Radical recombination (i.e., disappearance of configurations with unpaired electrons in wave function) is entirely accomplished at 2.2 Å, whereas TSs lie further on the pathway, at 2 Å. Herein (at 2 Å) lie the grounds of reactivity of substrates: the value of activation energy does not depend on possibility of ET in prereaction stage.
* It is interesting how terminology can change the character of description of this process: in σ-complex the positive charge is located on H-atoms of benzoic ring, and formally this is transfer of positive charge from NO2+ to H-atoms. Those charges (∼0.1 e) do not actually change much, redundant electronic density being transferred from C-atoms to NO2.
** It is worth mentioning that solvated NO2+ and H2NO3+ have equal EA in the SM2.1 method of calculation.
*** Albunia et al.  made an attempt to describe the diabatic state of such a system.