Introduction

Nucleophilic aromatic substitution in electron-deficient arenes, particularly nitroarenes, SNAr, is a fundamental process of organic chemistry widely used in research laboratories and industrial organic synthesis. According to common knowledge, the reaction consists in the replacement of halogens or other leaving groups X by oxygen, nitrogen, sulfur, carbon, etc., nucleophiles [1, 2]. Due to great practical value and wide application, the reaction has been the subject of thorough mechanistic studies [1,2,3] and also ab initio calculations [4]. According to the commonly accepted mechanism (Scheme 1) formulated by Bunnett [3], the substitution proceeds via addition of nucleophiles to the electron-deficient rings at positions occupied by X to form σX adducts, followed by departure of X−. Since the addition is associated with dearomatization, whereas departure of X− recovers the aromaticity of the ring, the addition is the slow, rate-determining step. For instance, replacement of fluorine proceeds faster than chlorine. The addition–elimination mechanism shown in Scheme 1 is presented in textbooks, monographs and numerous review articles [1,2,3,4,5,6,7,8].

Scheme 1
scheme 1

The mechanism of SNAr reaction

Full size image

Recently, on the basis of detailed mechanistic studies and calculations, a conclusion was drawn that in some cases, due to the high rate of the departure of X−, the lifetime of the hypothetical intermediate σX is negligible, and hence the reaction proceeds as a concerted process and the σX adducts are in fact transition states [5,6,7,8].

Thanks to detailed mechanistic experimental and computational studies, the many fine features of the process are clarified; nevertheless, an important question is surprisingly omitted. It is evident that electron deficiency of the ring of halonitrobenzenes is due to the presence of the nitro group; hence, in the halonitroaromatic rings, all positions ortho- and para- to the nitro group are susceptible to the addition of nucleophiles, regardless of whether they are occupied by halogens or hydrogen, and therefore nucleophiles can add to p-chloronitrobenzene at ortho positions as well [12,13,14]. Moreover, reactions that proceed via this addition mode have been known for many years. For instance, the von Richter reaction between p-chloronitrobenzene and cyanide anions reported in 1871 proceeds via the formation of σH adducts that are subsequently converted into m-chlorobenzoic acid [12] (Scheme 2).

Scheme 2
scheme 2

Von Richter reaction proceeds via nucleophilic addition at positions occupied by hydrogen

Full size image

Even more instructive is the reaction of this nitroarene with the carbanion of phenylacetonitrile, which proceeds via addition of the carbanion at the ortho (in protic media) and para (in polar aprotic media) positions (Scheme 3) [14].

Scheme 3
scheme 3

Carbanion of phenylacetonitrile adds to p-chloronitrobenzene at the ortho (b) or para (a) positions

Full size image

Key hypothesis

Addition at the position occupied by chlorine (path a) is an irreversible process; therefore, the reaction shown in path 3b can proceed only when the addition at positions occupied by hydrogen proceeds faster and is a reversible process. Generalization of this observation leads to the hypothesis that this is a general situation, as presented in Scheme 4.

Scheme 4
scheme 4

Hypothetical relation of the rates of nucleophilic addition

Full size image

This hypothesis, when confirmed, shall significantly expand the scope and versatility of the nucleophilic aromatic substitution, which should embrace nucleophilic substitution of hydrogen and halogens. Moreover, it should require revision of the commonly accepted mechanism of this reaction. This hypothesis can be confirmed provided that processes for fast conversion of the initially formed σH adducts into products of substitution of hydrogen are designed and executed. Unlike halogen anions that depart spontaneously from the σX adducts (Scheme 4, path a), spontaneous departure of the hydride anion from the σH adducts does not occur; thus it should be removed in a separate operation. Direct removal of hydride anions from organic molecules proceeds under the action of oxidants, and hence one could expect that treatment of the σH adducts with external oxidants should result in the formation of the products of oxidative nucleophilic substitution of hydrogen (ONSH). Alternatively, the hydrogen can be removed from the anionic σH adducts in the form of protons. This seemingly paradoxical process can be realized by means of base-induced β-elimination.

Regardless of how the conversion of the σH adducts occurs in subsequent processes, it is feasible when the equilibrium 4b ensures sufficient concentration and lifetime of the σH adducts. The position of the equilibrium is a function of the electrophilic activity of arenes and the nucleophilicity of the nucleophiles, as well as the reaction conditions—solvents, and particularly temperature. Due to the entropy effects, low temperature favors the addition; thus many processes of SNArH are realized at low temperatures.

Oxidation of the anionic σH adducts

The removal of the hydride anions from anionic σH adducts via oxidation by external oxidants encounters substantial obstacles. First, the σH adducts are usually short-lived species, produced in reversible processes. Thus, taking into account that nucleophiles themselves are sensitive to oxidation, the addition of oxidants to an equilibrated system nitroarene—nucleophile, e.g. 4b, can result in simple oxidation of the nucleophile. The desired oxidation of the σH adducts can proceed provided that:

  1. a.

    nucleophiles are resistant towards oxidation;

  2. b.

    the addition equilibrium is shifted towards adducts, and in an extreme case the addition proceeds irreversibly; and

  3. c.

    the rates of oxidation of the σH adducts are higher than the oxidation of nucleophiles.

Case A Treatment of p-chloronitrobenzene dissolved in liquid ammonia with KOH, source of OH− anion, a nucleophile resistant towards oxidation and oxygen, results in oxidation of the σH adducts to form 2-nitro-5-chlorophenol in high yield [15] (Scheme 5). As shown in textbooks, SNAr of chlorine by OH− anions proceeds at elevated temperatures.

Scheme 5
scheme 5

Oxidative nucleophilic hydroxylation of p-chloronitrobenzene

Full size image

Liquid ammonia is a moderately active nucleophile and also an excellent solvent for KMnO4 to produce a stable solution. Dissolution of 2,4-dinitrochlorobenzene in such solution results in efficient oxidative amination to give 3-chloro-4,6-dinitroaniline (Scheme 6) [16].

Scheme 6
scheme 6

Oxidative amination of 2,4-dinitrochlorobenzene

Full size image

Case B Methinic, moderately stabilized carbanions add to nitrobenzenes selectively at para positions due to steric effects. At low temperatures, the equilibrium is fully shifted towards the adducts. Oxidation of the σH adducts by potassium permanganate in liquid ammonia or by dichlorodicyanobenzoquinone (DDQ) or dimethyldioxirane (DMD) in DMF/THF results in the formation of products of ONSH, usually in high yields (Scheme 7) [17, 18].

Scheme 7
scheme 7

Oxidative substitution of hydrogen in o-fluoronitrobenzene with methinic carbanion

Full size image

Other kinds of anionic nucleophiles, including anilide anions and potassium diphenylphosphide, are less sterically demanding, so they can add at ortho and para positions to form σH adducts that can be subsequently oxidized to give products of ONSH. Addition at positions occupied by halogens proceeds much slower (Scheme 8) [19, 20].

Scheme 8
scheme 8

Oxidative substitution of hydrogen in p-fluoronitrobenzene with phosphorus and nitrogen nucleophiles

Full size image

Primary alkyl magnesium bromides add to nitroarenes and also to p- or o-chloronitrobenzenes to irreversibly form σH adducts that can be subsequently oxidized, giving products of oxidative nucleophilic alkylation (Scheme 9) [21, 22].

Scheme 9
scheme 9

Oxidative nucleophilic alkylation of p-chloronitrobenzene

Full size image

Sufficiently nucleophilic methylenic carbanions of arylacetonitriles, ethyl arylacetates, some ketones, esters of benzyl phosphonic acids, etc., can add at the ortho and para positions of nitroarenes. The produced σH adducts can be efficiently oxidized by a variety of oxidants. Even in 2,4-dinitrofluorobenzene, the Sanger reagent in which SNAr of fluorine is a very fast process, ONSH with carbanions proceeds faster than the replacement of fluorine (Scheme 10) [23].

Scheme 10
scheme 10

ONSH with methylenic carbanions in the Sanger reagent

Full size image

Particularly interesting and attractive is oxidation with oxygen, which proceeds efficiently in the presence of an excess of base, apparently upon deprotonation of the σH adducts; hence dianions are actually oxidized [24]. For instance, the reaction of esters of arylacetic acids with nitroarenes in the presence of an excess of base and air results in the formation of substituted nitrobenzophenones (Scheme 11) [25].

Scheme 11
scheme 11

Nucleophilic benzoylation of nitroarenes via ONSH

Full size image

Taking into account the scope and versatility of oxidative nucleophilic substitution of hydrogen, ONSH is presently a valuable methodology in organic synthesis. A few examples presented above unambiguously confirm the hypothesis shown in Scheme 4b.

Conversion of the σH adducts into substituted nitrosoarenes

The σH adducts of some carbanions, particularly arylacetonitriles, generated in protic media undergo protonation–dehydration to form substituted nitrosoarenes that are active compounds and usually react further (Scheme 3). A similar reaction can be promoted when σH adducts of methinic carbanions are treated by trimethyl chlorosilane in the presence of triethylamine. O-silylation of the oxygen atoms of the nitro group followed by 1,6-elimination of the silanol results in the formation of stable substituted nitrosoarenes [17], exemplified in Scheme 12.

Scheme 12
scheme 12

Conversion of anionic σH adducts into nitrosoarenes via silylation

Full size image

Similarly, the addition of anilide anions to nitroarenes at low temperature followed by protonation of the σH adducts results in the formation of nitrosodiarylamines, usually in high yields (Scheme 13) [26].

Scheme 13
scheme 13

Synthesis of nitrosodiarylamines in the reaction of anilines with nitroarenes

Full size image

Removal of hydrogen from the anionic σH adducts in the form of protons via base-induced β-elimination

All these reactions of the anionic σH adducts—elimination of water upon protonation or silanol upon silylation—can be considered as a stepwise removal of the hydride anions in the form of protons. Such a process in a direct form is the key step of the most important variant of the nucleophilic substitution of hydrogen—vicarious nucleophilic substitution (VNS) [27]. σH adducts of nucleophiles that contain nucleofugal groups X at the nucleophilic centers, particularly α-chlorocarbanions, undergo base-induced β-elimination of HX at the expense of the ring hydrogen to form nitrobenzylic carbanions, which upon protonation give products of the substitution. The reaction of carbanion of chloromethyl phenyl sulfone with p-fluoronitrobenzene can be considered a model case of VNS (Scheme 14) [28, 29].

Scheme 14
scheme 14

Vicarious substitution of hydrogen versus SNAr of fluorine in p-fluoronitrobenzene

Full size image

When this reaction is carried out at room temperature or lower in the presence of excess t-BuOK, VNS at the ortho position proceeds exclusively (path a). On the other hand, under conditions that favor equilibration and disfavor β-elimination, as for example in the slow addition of a solution of the carbanion to p-fluoronitrobenzene at temperatures above 20 °C, SNAr of fluorine is the main process (path b) [28, 29]. These simple experiments unambiguously confirm that even in the case of p-fluoronitrobenzene, the addition of the carbanion at position ortho occupied by hydrogen is the fast primary and reversible process and that conversion of the σH adducts proceeds via base-induced β-elimination of HCl. Detailed mechanistic studies have provided further pieces of evidence of this picture [30].

VNS is a general reaction that takes place between a great variety of α-chlorocarbanions and a practically unlimited variety of electron-deficient arenes: carbocyclic and heterocyclic nitroarenes and arenes of high electrophilic activity due to electronic configuration—particularly azines [27, 31]. When the arenes contain halogens in activated positions, VNS of hydrogen proceeds faster as a rule than SNAr of halogens. Some examples of VNS with α-chlorocarbanions are presented in Scheme 15 [32,33,34].

Scheme 15
scheme 15

Examples of VNS by α-chlorocarbanions in various halonitrobenzenes

Full size image

α-Chlorocarbanions are rather unstable entities; hence, in some cases, VNS proceeds better with carbanions containing other leaving groups, such as RO or RS (Scheme 16) [32].

Scheme 16
scheme 16

VNS in p-fluoronitrobenzene with carbanion of phenoxyacetonitrile

Full size image

The examples presented in Schemes 13, 14, 15 and 16 indicate that VNS proceeds with similar stoichiometry but opposite polarity to the Friedel–Crafts reaction. Indeed, because the result of both of these reactions is that a hydrogen atom in an aromatic ring is replaced by a carbon substituent—however, via the addition of either nucleophilic or electrophilic agents—it is justified to consider VNS as an umpolung of the Friedel–Crafts reaction. This relation can be convincingly exemplified by nucleophilic benzylation and α-chlorobenzylation of nitroarenes [35], as shown in Scheme 17.

Scheme 17
scheme 17

Nucleophilic benzylation and α-chlorobenzylation of nitroarenes via VNS

Full size image

VNS is not limited to carbanionic nucleophiles. Anions of t-butyl and cumyl hydroperoxides add to nitroarenes to form σH adducts. Subsequent base-induced β-elimination of alcohols provides nitrophenols. This is a general process of nucleophilic hydroxylation of electron-deficient arenes (Scheme 18) [36].

Scheme 18
scheme 18

VNS hydroxylation of 2,4-dinitrochlorobenzene

Full size image

Similarly, N-anions generated via deprotonation of some sulfenamides, hydrazine and hydroxylamine derivatives enter VNS reaction with electron-deficient arenes (Scheme 19) [37, 38].

Scheme 19
scheme 19

VNS amination of p-chloronitrobenzene

Full size image

As a rule, both of these reactions, VNS hydroxylation and amination, proceed faster with halonitroarenes than substitution of chlorine.

Summary: initial formation of anionic σH adducts is confirmed

Numerous examples of the conversion of the σH adducts of nucleophiles to nitroarenes and also halonitroarenes into products of nucleophilic substitution of hydrogen via oxidative and eliminative processes that proceed efficiently with a plethora of halonitroarenes provide unambiguous proof that the hypothesis formulated in Scheme 4 is correct. Therefore, it is necessary to accept that nucleophilic substitution of hydrogen proceeds faster than that of halogens, and to reject the commonly accepted mechanism of the SNAr reaction and formulate the following corrected general mechanism. The initial step of the reaction between nucleophiles and electron-deficient arenes is a fast and reversible addition at positions occupied by hydrogen to form σH adducts. When, due to the type of educts and conditions, further rapid conversion of the σH adducts into products of the substitution of hydrogen does not proceed, the σH adducts dissociate, and slower addition at positions occupied by halogens leads to conventional replacement of halogen. On the other hand, when, thanks to the nature of the educts and conditions, σH adducts can be rapidly converted into products of nucleophilic substitution of hydrogen, this process is dominant. Therefore, nucleophilic substitution of hydrogen is the fast primary process, whereas nucleophilic substitution of halogens is the secondary ipso reaction [39, 40]. Of course, the substitution of halogens is a secondary process with respect to rates, not to the importance of these processes. This mechanistic picture of SNAr reactions formulated on the basis of numerous experimental observations and studies is fully confirmed by ab initio calculations. The calculated rates of the addition of some nucleophiles at positions occupied by hydrogen of p-halonitrobenzenes are higher than the rates of addition at positions occupied by halogens [41, 42]. It should be mentioned that in the quantum chemical calculations of nucleophilic aromatic substitution published in numerous reports [4, 9,10,11], the possibility of the addition of nucleophiles to halonitrobenzenes at positions occupied by hydrogen is totally ignored. In my opinion, calculations of the reactions between two partners should search for a pathway with the lowest energy barrier, the lowest free energy of the transition state; thus these calculations, in spite of the high theoretical level, have a conceptual deficiency.

Does the addition proceed directly or via single-electron transfer (SET)?

To present a complete picture of nucleophilic aromatic substitution, the question of how the addition proceeds should also be discussed. It is well known that nitroarenes are active single-electron acceptors, and the exposure of nitroarenes to some nucleophiles and/or strong bases generates nitroaromatic anion radicals detected by electron spin resonance (ESR) spectroscopy [43]. On the basis of these and related observations, a concept was formulated that the addition of nucleophiles to nitroarenes often proceeds not directly, but as a two-step process—SET—followed by anion-radical–radical coupling [44,45,46,47]. Such a pathway was confirmed, for instance, for addition of the primary alkyl magnesium bromides to nitroarenes to form σH adducts on the basis of the radical clock criterion [47].

An interesting effect of conditions on the reaction pathway between carbanions of diphenylacetonitrile and o-chloronitrobenzene was observed [48].

The carbanions of diphenylacetonitrile in dipolar aprotic solvent DMSO and DMF or protic solvent methanol (Scheme 20a, b) are in the form of loose ion pairs and add directly to the ring, giving products of the substitution of hydrogen and chlorine. In cases a and b, initial addition proceeds at the para position to form σH adducts. In case a, the adduct dissociates, and slower addition at the position occupied by chlorine results in SNAr. In protic media, case b, the σH adduct is converted into the nitrosoarene (as in Scheme 12). The subsequent addition of the carbanion to the nitroso group gives the final product. On the other hand, in nonpolar solvent, the carbanions are in the form of tight pairs with the counterions, and hence they exhibit low nucleophilic activity. Therefore, SET dominates, which leads to dimerization of the carbanion and reduction of the nitroarene (Scheme 20). In order to clarify whether SET participates in the VNS and other reactions of α-chlorocarbanions with nitroarenes, experiments with a fast radical clock were carried out (Scheme 21) [49].

Scheme 20
scheme 20

Effects of conditions on reactions of carbanions with nitroarenes

Full size image
Scheme 21
scheme 21

Experiments with a fast radical clock indicate that VNS does not proceed via SET

Full size image

SET from carbanions of sulfone A to a substituted nitrobenzene should result in the formation of short-lived cyclopropylmethyl radical ·A, which rearranges with the rate constant 109 s−1. As shown in Scheme 21, VNS by this carbanion proceeds without traces of rearrangement; hence intermediacy of the radical ·A is excluded, and the addition proceeds directly and not via SET.

General analysis concerning the participation of SET as a means of nucleophilic addition to nitroarenes is presented in Scheme 22.

Scheme 22
scheme 22

Comparison of direct addition and SET as the pathway of nucleophilic aromatic substitution

Full size image

The addition of nucleophiles to p-chloronitrobenzene, a model chloronitroarene, can proceed directly (paths aH and aCl) to produce σH and σCl adducts, or via SET, which initially produces nitroaromatic anion radical and radical Nu· followed by the combination of these paramagnetic species path bH and bCl to produce σH and σCl adducts identical to those produced via direct addition. According to the microscopic reversibility principle, dissociation of these σ adducts to form educts should proceed via identical transition states (identical pathways). Therefore, in the case of both of the σ adducts formed via SET, pathways bH and bCl, the dissociation of C–Nu and C–Cl bonds, should proceed via homolysis to produce nitroaromatic anion radicals and radical Nu· and chlorine atoms followed by SET from nitroaromatic anion radicals to Nu· and Cl·. Such a pathway of dissociation of the σH adducts, while possible, seems unlikely. However, homolysis of C–Cl bonds in anionic σCl adducts and generation of chlorine atoms is highly improbable. On the other hand, in my opinion, different routes for the conversion of the σCl adducts via dissociation of C‒Nu and C‒Cl bonds in terms of the microscopic reversibility principle are difficult to accept. It appears, therefore, that when the SNAr of halogens in halonitroarenes is preceded by reversible addition of the nucleophiles at positions occupied by hydrogen, and such reversibility can be confirmed, SET as the reaction pathway can be excluded.

Consequences and advantages of the use of the general mechanism of substitution

On the basis of the general mechanism discussed above, it was possible to determine the true effects of substituents on the electrophilic activity of nitroarenes. The effect of substituents on rates of nucleophilic substitution of halogens in halonitroarenes has been the subject of numerous studies [1,2,3,4]. However, the results, although practically useful, cannot be considered as a measure of electrophilicity of nitroarenes, for two main reasons: (a) the rates of the substitution were dependent on the nature of the leaving groups, and (b) the substitution is a secondary process preceded by fast and reversible addition at positions occupied by hydrogen and formation of the σH adducts. The true effects of substituents on the electrophilic activity of nitroarenes were determined by measuring of the rates of nucleophilic addition at positions occupied by hydrogen [50, 51].

The general mechanism of nucleophilic aromatic substitution in electron-deficient arenes and its specific features presented in this paper reveal a much wider scope, versatility and diversity of the reactions between nucleophiles and nitroarenes. First, according to the general mechanism, nucleophilic substitution is not limited to halonitroarenes, but can proceed in a variety of nitroarenes that do not contain a leaving group, and hence so far had not been taken into consideration for reactions with nucleophiles. Moreover, the reactions of nucleophiles with halonitroarenes are not limited to simple replacement of the halogens, but offer a few processes for conversion of the σH adducts into a variety of products of the substitution of hydrogen. It should be pointed out that nucleophilic substitution of hydrogen in nitroarenes and in a variety of other electron-deficient arenes, particularly azines, is presently a rich and important chapter of organic chemistry [27, 31, 52,53,54,55]. Thus a general picture of such reactions is shown in Scheme 23.

Scheme 23
scheme 23

General picture of nucleophilic substitution in halonitroarenes

Full size image

Furthermore, since there are a few variants of conversion of the σH adducts, it is possible to observe situations in which a given nucleophile–nitroarene pair can react in a few ways to give different products [20, 26, 56]. A striking example of such a case is shown in Scheme 24.

Scheme 24
scheme 24

Carbanion of phenoxyphenylacetonitrile can react with o-chloronitrobenzene in five! different ways with high selectivity

Full size image

I certainly hope that the general mechanism presented in this paper will find a way to classrooms and textbooks and will be helpful in solving many synthetic problems. I also hope that the examples and reasoning presented in the paper will promote wider application of these processes in practical organic synthesis and will inspire the design of new reactions.