Synthesis of 5,9-Diaza[5]helicenes

: A new method for the synthesis of 5,9-diaza[5]helice-nes is presented using 2,3-bis(acylamino)-substituted ortho -ter-phenyls as precursors. Activation of the amide groups and electrophilic substitution at the ortho positions of the adjacent phenyl groups leads to the 5,9-diaza[5]helicenes. A stepwise reaction including protection of the first amino group, amide formation at the second amino group with subsequent cyclization, followed by deprotection, amide formation and cyclization at the first amino group ensures that both electrophilic substi-Helicenes, i.e. polycyclic aromatic compounds with ortho fused benzenes, have special optical and electronic properties, which are similarly present in derivatives bearing heteroatoms or being constructed from other aromatic or heteroaromatic moieties. [1–3] According to the IUPAC definition, [4] the helicenes consist of at least five rings (pentahelicene, [5]helicene). [5]Heli-cenes are non-planar and thus chiral, albeit with a very low racemization barrier (Figure 1). [5,6]

A first considered retrosynthetic approach would cleave the 5,9-diaza [5]helicenes A in two concomitant retro-Friedel-Craftstype substitutions into a 2,3′-bis(acylamino)terphenyl B, which could be traced back in a retro-cross coupling to a suitably metalated building block D and a 2-halo-3-nitrobiphenyl E. The latter should again be accessible by cross coupling (Scheme 1).
In the here reported first investigations within this strategy we considered it useful to aim at first for 3,12-dimethoxy-substituted diaza [5]helicenes to facilitate the synthetic approach (including the final electrophilic acylations), to furthermore improve the solubility of the final products, and to allow for a later attachment of further functionalities at possibly liberated hydroxy groups.
The nucleophilic coupling partner for the cross coupling to terphenyl 11 was prepared from commercially available 4methoxy-2-nitroaniline (8), which was converted into an iodo derivative 9. It turned out that this could not be transferred to a boronate (possibly due to the ortho-nitro group), preventing the planned Suzuki coupling. We alternatively considered a Stille coupling and prepared aryl stannane 10 by reaction with bis(tributyltin) in the presence of bis(triphenylphosphine)palladium(II) dichloride as catalyst (57 % over two steps). Stille cross coupling of building blocks 7 and 10 was achieved with tetrakis(triphenylphosphine)palladium(0) and copper(I) iodide as catalysts and yielded terphenyl derivative 11 (88 %). Reduction to diamine 12 was successfully performed with potassium borohydride/copper(I) chloride. [29] Nevertheless, all tested protocols for an amide formation with concomitant cyclization towards J failed (Scheme 3). Reac-tion with tert-butyl isocyanide in the presence of cobalt(II) acetylacetonate and oxygen (as recently reported by Tobisu et al. for the synthesis of phenanthridines [30] ) did not lead to any well-defined product. Amide coupling with propionic acid in the presence of propylphosphonic anhydride (T3P®) to bis-(amide) 13 was successful, but a subsequent cyclization with Eaton's reagent (P 2 O 5 in MeSO 3 H) [31] did only lead to the formation of phenanthridine derivative 14 (albeit with poor yields). Obviously, the fastest aromatic substitution is that of the eastern amide group at the activated (i.e., OMe-substituted) benzene. The acylamino-substituted benzene seems to be less reactive; its substitution proceeds with lower rates. Once the phenanthridine 14 is formed, the second substitution at this electron-deficient arene is obviously too sluggish -even at elevated temperatures. Scheme 3. Attempted acylation towards diaza [5]helicenes.
Consequently we changed our strategy and decided to pursue a stepwise strategy (Scheme 4). This would allow to perform both electrophilic aromatic substitutions with activated arenes and would furthermore give access to products F with different substituents R 1 and R 2 introduced with the amide moieties. We planned to build the western amide function, where the other amino group is still suitably protected (I); electrophilic acylation towards phenanthridine H would now proceed at an activated substrate. Deprotection and amide formation at the eastern amino group would furnish a substrate G, which again should be sufficiently reactive towards a substitution. For this approach we again started with 4-methoxy-2-nitroaniline (8), which was subjected to a Sandmeyer reaction (now furnishing the respective bromo compound 15 [32] ) and reduced to aniline 16. [33] Acylation furnishing acetanilide 17 was here performed with acetic acid in the presence of propylphosphonic anhydride (T3P®) as coupling agent. [34] This transformation yielded amides with excellent yields (further derivatives have been prepared but are not part of this publication), but could in principle be similarly performed with simpler protocols (e.g. by coupling with the respective acid chlorides). Borylation to a suitable nucleophile 18 for a pursued Suzuki coupling was achieved by reaction with bis(pinacolato)diboron in the presence of [1,1′-bis(diphenylphosphino)ferrocene]dichloropalladium(II) as catalyst (Scheme 5). [35] The rather low yield of 59 % for this step was due to the poor solubility of the boronate, which led to losses during the work-up and purification process. Nevertheless, boronate 18 was thus obtained in four consecutive steps with a total yield of 41 %. A stannylated substrate which could have been used in a Stille coupling was similarly synthesized, but since a non-identified side product could not be separated by conventional methods, this variation was thus no longer pursued. Eur. J. Org. Chem. 2019, 6697-6701 www.eurjoc.org The nitrobiphenyl 7 (Scheme 2) was reduced to amine 19 by a standard protocol. [28] Cross coupling with boronate 18 to terphenylamine 20 was achieved in 87 % yield using a proven method with palladium(II) acetate as precatalyst and SPhos (2-dicyclohexylphosphino-2′,6′-dimethoxybiphenyl) as ligand (Scheme 6). [36] We considered allyl groups suitable as protecting groups for the amino group, [37] which can be introduced with standard protocols. [38] They should not deactivate the arene in the electrophilic substitution, and they can be cleaved with rhodium catalysis not affecting any of the other functional groups in the substrates. A double allylation of terphenylamine 20 towards 21 was achieved with 61 % yield. Dehydrating ring closure to phenanthridine 21 and (in the further course of the synthesis similarly to the diaza[5]helicene 25) was performed with triphenylphosphine oxide and trifluoromethanesulfonic anhydride [Hendrickson reagent, (Ph 3 P + ) 2 O·2F 3 CSO 3 -], [39] which proved to be superior to a more conventional protocol using phosphoryl chloride (Scheme 7). Phenanthridine derivative 22 was thus obtained in 81 % yield. The second cyclization towards diaza [5]helicenes 25 was achieved after cleavage of the allyl groups (→ 23) [40] and amide  Table 1). formation (→ 24) with T3P® (vide supra). The obtained substrates together with yields for the last two steps are summarized in Table 1. Table 1. Synthesis of 5,9-diaza [5]helicenes.
[a] Helicene formation yielded product with a non-separable side product.
This strategy was applied to three 5,9-diaza [5]helicenes (Table 1) with different substituents at positions C-6 and C-10. To reduce the synthetic effort, we used only one common terphenyl precursor 21 bearing a methyl substituent at the position later being C-6 in the diazahelicenes. Different substituents were introduced with the second amide moiety, an isopropyl, an (E)-2-propenyl, and a 3-(methoxycarbonyl)propyl group. The propenyl group would allow for the introduction of further functionalities, while the latter could be especially useful for the attachment of structures via an ester or an amide linkage. Introduction of a 4-heptanyl group was not successful: The respective amide 24 was only isolated with a poor 20 % yield, possibly due to steric reasons; it was not used in a 5,9-diaza [5]helicene synthesis. No amide was obtained, when amine 23 was treated with 4-bromobutanoic acid. Cyclization of amide 24a (R = Et) led to the formation of helicene 25a, but the presence of unidentified side products and problems during the purification process prevented an isolation of the product. Its synthesis was thus not included in Table 1. The 5,9-aza [5]helicenes 25 were additionally characterized by their UV/Vis spectra; they show distinct absorptions around 237 and 310 nm. Detailed data and spectra are given in the Supporting Information. 5,9-Aza [5]helicenes 25 were thus obtained in total yields of 2-4 % over 11 linear steps (total 15 steps) starting with commercially available 2,6-dinitroaniline and 4-methoxy-2-nitroaniline. The herein presented method allows for variable substituents at positions C-6 and C-10, but could be applicable with further substituted starting materials (compounds 25 bear methoxy groups at positions C-3 and C-12). Slight modifications of this protocol should furthermore give access to the related 5,10-and 6,10-diaza [5]helicenes ( Figure 1); their synthesis is now investigated in our laboratories, as is the extension of these methods to the synthesis of higher helicene-type compounds.