One-pot synthesis of 4′-alkyl-4-cyanobiaryls on the basis of the terephthalonitrile dianion and neutral aromatic nitrile cross-coupling

A convenient one-pot approach to alkylcyanobiaryls is described. The method is based on biaryl cross-coupling between the sodium salt of the terephthalonitrile dianion and a neutral aromatic nitrile in liquid ammonia, and successive alkylation of the long-lived anionic intermediate with alkyl bromides. The reaction is compatible with benzonitriles that contain methyl, methoxy and phenyl groups, fluorine atoms, and a 1-cyanonaphthalene residue. The variety of ω-substituted alkyl bromides, including an extra bromine atom, a double bond, cyano and ester groups, as well as a 1,3-dioxane fragment are suitable as alkylation reagents.


Scheme 2:
Para-cyanophenylation of substituted benzonitriles 2 by dianion 1 2− with the formation of a long-lived cyanocyclohexadienyl anion 3, followed by oxidation or alkylation, leading to dicyanobiphenyls 4 or alkylcyanobiphenyls 5. excellent yields. Nevertheless, the methodology has some disadvantages: the catalytic systems are quite expensive, the ligands are often hardly accessible and require special synthetic efforts, and preactivation of the substrates via introduction of an organometallic group is necessary. The last disadvantage is overcome to some extent by the direct C-H arylation protocols [18], but usually they are catalytic and thus the resulting crosscoupling products contain traces of transition metals difficult to remove that limits their use in areas such as pharmacy [22].
terphenyls [25] formed in good yields. The study of probable pathways for the 1 2− interaction with benzonitriles 2 demonstrated that the cyanocyclohexadienyl anion 3 was a long-lived reaction intermediate [23,24]. This finding allowed us to obtain not only dicyanobiphenyls 4 via anion 3 oxidation, but also alkylcyanobiphenyls 5 by treating 3 with an alkyl halide. Earlier, we used butyl bromide for anion 3 trapping and obtained 4-butyl-4'-cyanobiphenyl (5aa) in 56% yield (Table 1, entry 1) [23]. In order to expand the scope of the synthetic utilization of the cross-coupling reaction under investigation, as well as to work out a short and convenient approach to a number of universal carbonitrilic structural blocks with different both aryl and alkyl moieties [16][17][18][19], we varied the nature of the neutral participant 2 in the coupling with dianion 1 2and also the reagent for anion 3 alkylation. The presented results not only support the approach based on anionic forms of cyano-arenes as effective cross-coupling reagents toward neutral unactivated substrates [26,27], but also expand the knowledge about nucleophilic mechanisms such as aromatic substitution of a hydrogen atom (S N H) [28] and the vicarious substitution [29].

Results and Discussion
The suggested one-pot synthesis of alkylcyanobiaryls 5 comprises consecutive generation of terephthalonitrile dianion (1 2-) [30][31][32] by the addition of metallic sodium to a suspension of dinitrile 1 in liquid ammonia, treating of the thus formed 1 2− salt with a twofold excess of nitrile 2, stirring the reaction mixture for ca. 1.5 h, which is necessary for cross-coupling, and final quenching by the addition of an excess of alkyl halide 6.
According to the total reaction scheme, these byproducts arise via alkylation of the anionic forms of nitriles 1 and 2 [26]. The latter appear in the reaction media due to the electron transfer from 1 2− to neutral cyanoarene 2 (Scheme 2). Of note, 4,4'dicyanobiphenyl 4 (formed due to anion 3 oxidation) in all cases except one (see below) is detected in the reaction mixtures only in trace amounts (up to 5%). Thus, the alkylation step can be assumed to be quite fast, and the target product yield mostly depends on the efficiency of the cross-coupling stage. Relatively volatile components 2 and 7-9 are easily distilled off in vacuo, and alkylcyanobiphenyls 5 are additionally purified by preparative TLC leading to 31-67% isolated yields.
The structures of the obtained products 5ab-af were consistent with the proposed reaction scheme (Scheme 2). It corresponds to the introduction of the cyanophenyl fragment of dianion 1 2− into the para-position of nitrile 2a with the formation of cyclohexadienyl anion 3 after rapid decyanation of the primary crosscoupling dianionic product. Next, the alkylation of intermediate 3 occurs at the position ipso to the cyano group. This orientation is in agreement with the electronic structure of anion 3, which is characterized by HOMO localization predominantly at the ipso position to the cyano group [24]. This type of HOMO is typical for cyclohexadienyl anions generated from monocyanoarenes [42,43]. The highest yield of 5 (5ab, 70%) was obtained for 5-bromopent-1-ene (6b). Reactions with ω-substituted alkyl bromides 6d and 6e provided lower product yields, which depended on the nature of the alkylating reagent. In the reaction with dibromopentane 6c we found no disubstitution product -the corresponding 1,5-bis(4-cyanobiphenyl-4-yl)pentane -but the reaction mixture contained traces of compounds with molar masses of 325 and 352 (<1.5% according to GC-MS), which can be attributed to 4'-(5-phenylpentyl)-[1,1'biphenyl]-4-carbonitrile and 4'-(5-(1-cyanocyclohexa-2,5-dien-1-yl)pentyl)-[1,1'-biphenyl]-4-carbonitrile, respectively. The detection of these byproducts indicates that the bromine-containing product 5ac is drawn into consecutive reactions with the anionic forms of 2 as an alkylating reagent and that the rate is comparable or higher than the rate of the interaction of dibromide 6c with anion 3 (compared with [34]). In the case of alkylating reagents 6d,e that contain quite acidic protons in the α-position to the electron-withdrawing substituent, we suppose that partial protonation of the anionic reaction intermediates takes place together with the alkylation, thereby reducing the yield of the target product. The decreased yield in the reaction with acetal 6f is probably due to the latter's lower reactivity compared to other alkylating reagents, what was indicated by increased yield of dicyanobiphenyl 4 (14% vs 5% formed in reactions with 6b-e).
To demonstrate the effects of varying the aromatic moiety, a set of aromatic nitriles containing substituents of different natures and positions were used in the reaction (Table 2). We chose the nitriles that had earlier effectively underwent a cross-coupling with dianion 1 2− [24,25], in particular, 2-and 3-methyl (2b,c), 2-and 3-methoxy-(2d,e), and 2-fluorobenzonitrile (2f), as well as 4'-methyl-2-cyanobiphenyl (2h). In addition, two new neutral carbonitriles -2,6-difluorobenzonitrile (2g) and 1-cyanonaphthalene (2i) -were tested as cross-coupling participants and found to react with dianion 1 2− . Alkylation of intermediate anions 3b-i in all these reactions was provided by butyl bromide (6a). The composition of the reaction mixtures and the yields of butylcyanobiaryls 5ba-ia were similar to those obtained in the reactions of 1 2− with 2a and alkyl bromides 6a-f.  The highest yield (70-75%) was obtained from the reactions with tolunitriles 2b,c. In general, the product yields matched the previously reported [24] dependence of the cross-coupling efficiency on the substituent identity and position in the neutral reaction participant 2. The reactions of 1 2− with 2-substituted benzonitriles provided higher product yields than in the case of the 3-substituted analogs, apparently due to the decrease in spatial hindrance in course of the cross-coupling. The yields depended on the nature of the substituent in 2b-f as follows: yield for methoxy < for fluorine < for methyl. Most likely this order reflects the influence of the substituent electronic effect on the two competing mechanisms within cross-coupling: heterolytic addition of dianion 1 2− (S N H) to nitrile 2, and/or electron transfer from 1 2− to 2 followed by recombination or re-  laxation in the primary pair of radical anions 1 ·− and 2 ·− (Scheme 2). These probable reaction mechanisms were studied in [23,24] and are beyond consideration of the present paper. In this context, 2,6-difluorobenzonitrile (2g) and 4'-methylbiphenyl-2-carbonitrile (2h) [25] (which can be considered as benzonitrile, substituted with a p-tolyl fragment at the 2 position) show generally the same behavior as all the explored benzonitriles.
The result obtained in the reaction with 1-naphthonitrile (2i) deserves separate discussion because the formation of 4-(4butylnaphthalen-1-yl)benzonitrile (5ia) is the first evidence that dianion 1 2− can effectively undergo the cross-coupling with an annulated aromatic nitrile ( Table 2, entry 8). The structure of 5ia shows that a new bond forms between the carbon atoms at position 1 of dianion 1 2− and position 4 of 2i, leading to the formation of the dimeric dianion 10 (Scheme 3). Further transformation of this primary product into butylbiaryl 5ia is via the pathways described above, i.e., a rapid and irreversible decyanation of dianion 10 with the formation of long-living monoanion 3i, and the alkylation of anion 3i at the position ipso to the cyano group providing alkyldihydro product 11, which undergoes fast dehydrocyanation in basic reaction media thus forming final aromatic product 5ia. The high ipso-regioselectivity of the bulylation of monoanion 3i is typical for an alkylation with primary alkyl halides of cyanocyclohexadienyl anions derived by two-electron reduction of aromatic mononitriles in liquid ammonia, which was studied earlier and was found to proceed via an S N 2 mechanism [26,35].

Conclusion
We propose a one-pot method of the transition-metal-free biaryl cross-coupling for the preparation of 4'-alkyl-4-cyanobiaryls as potentially valuable building blocks [3][4][5][6][7][8][9]20,44] with a variable structure of aromatic moieties (biphenylic, m-terphenylic or phenylnaphthylic), as well as an alkyl moiety. The main advantages of this method are experimental simplicity, utilization of commercially accessible reagents (terephthalonitrile, metallic sodium, substituted aromatic monocarbonitriles, alkyl bromides), together with a recyclable solvent -liquid ammoniawhich is now considered as a green solvent [45][46][47]. Our approach is compatible with a variety of different synthetically valuable substituents both in the aromatic part (methyl and methoxy groups, and fluorine atoms) and the side-chain (double bond, the bromine atom, ester and cyano groups, and 1,3dioxane fragment). The presence of such substituents in the products opens up the possibility of further synthetic modifications.

Supporting Information
Supporting Information File 1 Experimental section and 1 H, 13 C and 19 F NMR spectra of all synthesized compounds.